\v.cM mmmi. -:hhhv i := is ts :.\ V. Vit;} % viv.j j t|.f H V? |1 ttt«£HWttK»i}ttt»*5 ; t \ < * c $ • ■• '; '• v-’: h\v .’t' v.'J-^V ?->V. :•:%•-■ MWM Mm0ii 'iWvVv'i ’& 4rhfci*.H ,v?u! sH i \ ^iSft U.S. Department of Commerce Volume 99 Number 3 July 2001 Fishery Bulletin U.S. Department of Commerce Donald L. Evans Secretary National Oceanic and Atmospheric Administration Scott B. Gudes Acting Under Secretary for Oceans and Atmosphere National Marine Fisheries Service William T. Hogarth Acting Assistant Administrator for Fisheries D) Scientific Editor Dr. John V. Merriner Editorial Assistant Sarah Shoffler Center for Coastal Fisheries and Habitat Research, NOS 101 Pivers Island Road Beaufort, NC 28516 The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fish- eries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115-0070. Periodicals postage is paid at Seattle, WA, and at additional mailing offices. POST- MASTER: Send address changes for sub- scriptions to Fishery Bulletin, Superin- tendent of Documents, Attn.: Chief, Mail List Branch, Mail Stop SSOM, Washing- ton, DC 20402-9373. Although the contents of this publica- tion have not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this peri- odical is necessary according to law for the transaction of public business 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: $55.00 domestic and $68.75 foreign. Cost per single issue: $34.00 domestic and $42.50 foreign. See back for order form. Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C15700 Seattle, Washington 98115-0070 Editorial Committee Dr. Andrew E. Dizon Dr. Harlyn O. Halvorson Dr. Ronald W. Hardy Dr. Richard D. Methot Dr. Theodore W. Pietsch Dr. Joseph E. Powers Dr. Harald Rosenthal Dr. Fredric M. Serchuk National Marine Fisheries Service University of Massachusetts, Boston University of Idaho, Hagerman National Marine Fisheries Service University of Washington, Seattle National Marine Fisheries Service Universitat Kiel, Germany National Marine Fisheries Service Fishery Bulletin web site: fishbull.noaa.gov The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. It began as the Bulletin of the United States Fish Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. 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 99 Number 3 July 2001 Fishery Bulletin Contents Articles 389-398 Andrews, Allen H., Erica J. Burton, Kenneth H. Coale, Gregor M Cailliet, and Roy E. Crabtree Radiometric age validation of Atlantic tarpon. Megalops atlanticus 399-409 Colura, Robert L., and Britt W. Bumguardner Effect of the salt-box catch-bycatch separation procedure, as used by the Texas shrimp industry, on short-term survival of bycatch 410-419 Gillanders, Bronwyn M. Trace metals in four structures of fish and their use for estimates of stock structure 420-431 Hanrahan, Brian, and Francis Juanes Estimating the number of fish in Atlantic bluefin tuna ( Thunnus thynnus thynnus) schools using models derived from captive school observations The conclusions and opinions expressed in Fishery Bulletin are solely those of the authors and do not represent the official position of the National Marine Fisher- ies Service (NOAA) or any other agency or institution. The National Marine Fisheries Service (NMFS) does not approve, recommend, or endorse any proprietary product or pro- prietary material mentioned in this pub- lication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales pro- motion which would indicate or imply that NMFS approves, recommends, or endorses any proprietary product or pro- prietary 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. 432-442 Lucena, Flavia M., and Carl M. O'Brien Effects of gear selectivity and different calculation methods on estimating growth parameters of bluefish, Pomatomus saltatrix (Pisces: Pomatomidae), from southern Brazil 443-458 McBride, Richard S., Timothy C. MacDonald, Richard E. Matheson Jr., David A. Rydene, and Peter B. Hood Nursery habitats for ladyfish, Elops saurus, along salinity gradients in two Florida estuaries 459-464 McFarlane, Gordon A., and Jacquelynne R. King The validity of the fin-ray method of age determination for lingcod (Ophiodon elongatus ) Fishery Bulletin 99(3) 465-474 Rabe, Jessica, and Joseph A. Brown The behavior, growth, and survival of witch flounder ( Glyptocephalus cynoglossus ) larvae in relation to prey availability: adaptations to an extended larval period 475-481 Schubarf, Christoph D., Jesus E. Conde, Carlos Carmona-Suarez, Rafael Robles, and Darryl L. Felder Lack of divergence between 16S mtDNA sequences of the swimming crabs Callinectes bocourti and C. maracaiboensis (Brachyura: Portunidae) from Venezuela 482-501 Sturdevant, Molly V., Audra L. J. Brase, and Leland B. Hulbert Feeding habits, prey fields, and potential competition of young-of-the-year walleye pollock (Thercigra chalcogramma) and Pacific herring (Clupea pcillasi ) in Prince William Sound, Alaska, 1994-1995 502-509 Sun, Chi-Lu, Chien-Lung Huang, and Su-Zan Yeh Age and growth of the bigeye tuna, Thunnus obesus, in the western Pacific Ocean 510-515 Note Takahashi, Kazutaka, and Kouichi Kawaguchi Nocturnal occurrence of the swimming crab Ovalipes punctatus in the swash zone of a sandy beach in northeastern Japan 516 Errata 517 Subscription form 389 Abstract— An improved radiometric aging technique was used to examine annulus-derived age estimates from otoliths of the Atlantic tarpon, Meg- alops atlanticus. Whole otoliths from juvenile fish and otolith cores, repre- senting the first 2 years of growth, from adult fish were used to determine 210Pb and 226Ra activity; six age groups con- sisting of pooled otoliths and nine indi- vidual otolith cores were aged. This unprecedented use of individual oto- lith cores to determine age was possible because of improvements made to the 226Ra determination technique. The dis- equilibria of 210Pb:226Ra for these sam- ples were used to determine radiometric age. Annulus-derived age estimates did not agree closely with radiometric age determinations. In most cases, the pre- cision (CV<12%) among the otolith readings could not explain the differ- ences. The greatest radiometric age was 78.0 yr for a 2045-mm-FL female, where the radiometric error encompassed the annulus-derived age estimate of 55 yr by about 4 yr. The greatest radio- metric age for males was 41.0 yr for a 1588-mm-FL tarpon, where the radio- metric error encompassed the annulus- derived age estimate of 32 yr by 1 yr. Radiometric age determinations in this study indicated that the interpretation of growth zones in Atlantic tarpon oto- liths can be difficult, and in some cases may be inaccurate. This study provides conclusive evidence that the longevity of the Atlantic tarpon is greater than 30 years for males and greater than 50 years for females. Manuscript accepted 26 January 2001 . Fish. Bull. 99:389-398 (2001). Radiometric age validation of Atlantic tarpon. Megalops atlanticus Allen H. Andrews Erica J. Burton Kenneth H. Coale Gregor M. Cailliet Moss Landing Marine Laboratories 8272 Moss Landing Road Moss Landing, California 95039-9647 E-mail address (for A H Andrews): andrews@mlml.calstate.edu Roy E. Crabtree National Marine Fisheries Service 9721 Executive Center Drive North St. Petersburg, Florida 33702-2439 The Atlantic tarpon ( Megalops atlanti- cus) is believed to be a long-lived fish. The strongest evidence is from a cap- tive female Atlantic tarpon that lived to at least 63 years when it died in 1998 at the John G. Shedd Aquarium in Chicago, Illinois.1 Life in the aquarium, however, is difficult to compare with life in the natural environment. A simi- lar longevity of 55 years was estimated from growth zones in sagittal otoliths of a wild female tarpon (Crabtree et al., 1995). The annual periodicity of the growth zones used to estimate age was validated by using oxytetracycline up to an age of 9 yr in captive fish (Crab- tree et al., 1995), but extrapolation of these results to older, wild fish, how- ever, is fallible. In addition, wild Atlan- tic tarpon are highly migratory and inhabit waters that vary considerably in salinity (0 to 43 ppt) and tempera- ture (17° to 37°C; Zale and Merrifield, 1989; Nichols2). It is uncertain what affect these changing conditions have on the formation of growth zones in the otoliths, but periods of stress have been shown to halt otolith growth (Cam- pana, 1983). Furthermore, growth slows with age and growth zones become increasingly compressed, rendering oto- lith growth zones difficult to interpret. To circumvent the potential problems of otolith interpretation and to inde- pendently determine age, an applica- tion of the radiometric aging technique by using the 210Pb:226Ra disequilibria in sagittal otolith cores (Campana et al., 1990) of Atlantic tarpon was per- formed, and the results are discussed in the context of existing estimated growth parameters. Materials and methods Pooled and individual otolith cores from Atlantic tarpon were analyzed for 210Pb and 226Ra to determine age. Annulus- derived age estimates were based on six independent otolith readings deter- mined by Crabtree et al. (1995) and resulted in a coefficient of variation of less than 12%. For radiometric age determination, otoliths were pooled into six groups based on annulus-derived age estimates, sex, and collection date. In addition, nine individual otoliths were analyzed. The age groups were selected to cover the full range of annulus-derived age estimates. Selected age groups had a narrow age range and members of the group were collected within a 6-month period. Otoliths of adult age groups were cored to the first two years of growth and otoliths of the youngest age groups 1 Pamper, K. 1999. Personal commun. John G. Shedd Aquarium, 1200 S. Lake- shore Dr., Chicago, IL 60605. 2 Nichols, K. M. 1994. Age and growth of juvenile tarpon. Megalops atlanticus , from Costa Rica, South Carolina and Venezu- ela. Senior thesis, Univ. South Carolina, Columbia, SC 29208, 52 p. 390 Fishery Bulletin 99(3) were analyzed whole because they were similar in size to cores from adult fish. Because 2-yr cores were large and the activity of 226Ra was relatively high (~10 to 100 times higher than usual; Andrews et al., 1999b), single otoliths were analyzed with the radiometric aging technique. Coring the adult otoliths required establishing a target size and weight that closely approximated that of an oto- lith from a 2-year-old fish. Dimensions and weights of otoliths from four juvenile fish aged 2 years were record- ed and averaged; the resultant dimensions were approxi- mately 12 mm long by 6 mm high by 1 mm thick and a weight of 0.1 g. Each otolith from each adult age group was sculpted into this shape and weight by being hand- ground on a Buehler Ecomet® III lapping wheel. All sam- ples were cleaned of any adhering contamination by fol- lowing specific procedures described elsewhere (Andrews et al., 1999b). These clean samples were placed in acid- cleaned 100-mL Teflon® PFA Griffin beakers and dried at 85°C for 48 h. A detailed protocol describing sample preparation, chro- matographic separation of 226Ra from barium and calci- um, and analysis of 226Ra using thermal ionization mass spectrometry (TIMS) is described elsewhere (Andrews et al., 1999b). Only an overview of the 226Ra procedures is given here with details on the determination of 210Pb ac- tivity. Because the levels of 226Ra and 210Pb typically found in otoliths were extremely low (from femtograms [10~15 g] for 226Ra and attograms [10~18 g] for 210Pb) and because of the great potential for contamination from calcium, bar- ium, and lead, trace-metal clean procedures and equip- ment were used throughout sample preparation, separa- tion, and analysis. All acids used were double distilled (GFS Chemicals®) and dilutions were made with Milli- pore® filtered Milh-Q water (18 MQ/cm). To determine 226Ra activity with thermal ionization mass spectrometry (TIMS), the sample must be clean of naturally occurring organics (such as otolin). Organic resi- dues elevate background counts in the 226Ra region and increase the analytical uncertainty during TIMS analy- sis. Dried and weighed samples were dissolved in beakers on hot plates at 90°C by adding 8N HN03 in 1-2 mL ali- quots. Alternation between 8N HN03 and 6N HC1, with an aqua regia transition, several times resulted in com- plete sample dissolution. The dried sample, after dissolu- tion, formed a yellowish foam. To further reduce any re- maining organics, and to put the residue into the chloride form required for the 210Pb activity determination proce- dure, the samples were redissolved in 1 mL 6N HC1 and taken to dryness five times at ~90°C. A whitish residue in- dicated that a sufficient amount of the organics had been removed. These samples were used to determine 210Pb ac- tivity prior to TIMS analysis. Determination of 2:cPb Activity To determine 210Pb activity in the otolith samples, the alpha-decay of 210Po was used as a daughter proxy for 210Pb. To ensure that activity of 210Po was due solely to ingrowth from 210Pb, the time elapsed from capture to 210Pb determination was greater than 2 yr. Samples pre- pared for 210Po analysis were spiked with 208Po, a yield tracer. The amount of 208Po added was estimated on the basis of observed 226Ra levels in otoliths of juvenile tarpon. This amount was adjusted to five times the expected 21C)Po activity in the otolith sample to reduce error in the 210Pb activity determination. The spiked samples were redis- solved in approximately 50 mL of 0.5N HC1 on a hot plate at 90°C covered with a watch glass. The 210Po and 208Po- tracer were autodeposited for 4 hours onto a silver plan- chet (Flynn, 1968). The activities of these isotopes were determined by using alpha-spectrometry on the plated samples. Quantification of the 210Po was made by subtract- ing a detector blank and reagent counts from each peak region-of-interest, by multiplying the 2iopo;208p0 count ratio by the known 208Po activity, and by correcting for decay back to the time of plating. To attain sufficient counts, samples were counted for 21-23 d. The solution remaining after polonium plating was dried and saved for 226Ra analyses. Determination of 226Ra Activity To prepare the samples for 226Ra activity determination with TIMS, each sample was spiked with 228Ra, a yield tracer, and a newly developed ion-exchange separation technique was used to isolate radium from calcium and barium (Andrews et al., 1999b). The final samples were processed by using TIMS and the measured ratios of 226Ra:228Ra were used to calculate 226Ra activity. The anal- ysis of unspiked otolith samples indicated that 228Ra was not present in measurable quantities in otoliths of juve- nile fish, and no adjustment was necessary for the mea- sured 226Ra:228Ra ratio in spiked samples. Radiometric age determination To assess the feasibility of applying the radiometric aging technique to Atlantic tarpon, uptake of 210Pb and 226Ra was assessed in otoliths from juvenile fish. Because the age of juvenile Atlantic tarpon is better constrained than that of adults, age was determined by using 210Pb:226Ra disequilibria in whole otoliths from juvenile fish. For the juvenile otolith samples, the age determined would be higher than expected if a significant amount of exogenous 210Pb was incorporated into the otolith. Age was estimated from the measured 210Pb and 226Ra activities (Eqs. 1 and 2). Because the activities were mea- sured from the same sample, the calculation was indepen- dent of sample mass. For adult samples, where estimated age was greater than that of the 2-year-old core, radiomet- ric age was calculated as follows with an equation derived from Smith et al. (1991) to compensate for the ingrowth gradient of 210Pb:226Ra in the otolith core, Andrews et al.: Radiometric age validation of Megalops atlanticus 391 where t = A210Pb,(. = A226Ra TIMS = R0 = X = T = the radiometric age at the time of capture; the 21l,Pb activity corrected to time of cap- ture; the 226Ra activity measured with TIMS; the activity ratio of 210Pb:226Ra initially incorporated; the decay constant for 210Pb (ln(2)/22.26 yr); and the core age (2 yr). The radiometric age calculation for the juvenile age-group was determined by iteration of an equation derived from Smith et al. ( 1991), A210Pb„ A226 RaT„ = 1- (1 -Rn) 1-e -Maet \ Xt (2) where all equation components were as defined above. A radiometric age range, based on the analytical uncer- tainty, was calculated for each sample by applying the cal- culated error for 210Pb and 226Ra activity determinations to the measured 210Pb:226Ra. Calculated error included the standard sources of error (i.e. pipetting, spike, and cali- bration uncertainties), alpha-counting statistics for 210Pb (Wang et al., 1975), and an analysis routine used to run 226Ra samples on the thermal ionization mass spectrom- eter (Andrews et al., 1999b). Age estimate accuracy To compare annulus-derived age with radiometric age, a plot of the annulus-derived age estimate and measured 210Pb:226Ra activity ratio was compared graphically to the expected 210Pb:226Ra activity ratio from ingrowth. This comparison included a graphical compensation for the 210Pb:226Ra gradient in the core sample. This model assumed a linear mass-growth rate for the first two years of growth. Annulus-derived age range and the analytical uncertainty of the activity ratio were plotted with each data point. A direct comparison of annulus-derived age and radiometric age was made in a plot where a regression of the data was compared to a line of agreement or slope of one. A paired two-sample /-test was used to determine if a significant difference existed between the age estimates. Von Bertalanffy growth functions were fitted to the ra- diometric ages with FISHPARM software (Saila et al., 1988) and plotted with the annulus-derived ages and growth functions from Crabtree et al. (1995) for a visual comparison. High and low radiometric age and the size range of each sample were plotted for each data point. No statistical comparison was made between the growth functions because of the low number of samples and wide confidence interval for each parameter. Results Six age groups and nine individual otolith cores were se- lected for radiometric analyses (Table 1). Of the age group samples, three male samples and three female samples were selected to span the estimated age range of each sex. Male age groups were 2-4 yr, 18-21 yr, and 31-36 yr. Female age groups were 3—4 yr, 15-24 yr, and 48-50 yr. For each age group the capture dates were within a 6-month period, except the 2-4 yr male age group where the period spanned 7 months. Individual otolith core samples ranged in annulus-derived age from 13 yr to 32 yr for males and 35 yr to 55 yr for females. The greatest annulus-derived ages were for females; the oldest female was estimated to be 55 yr and the oldest male was 36 yr. Fork length was lowest for the juvenile age groups and did not overlap with older age groups. There was some overlap between the middle and old age groups. Males were typically smaller than females; the largest male was 1620 mm FL and the largest female was 2045 mm FL. The number of otoliths used in each age group ranged from 4 to 1 1 with the fine- cleaned sample weight ranging from 0.3314 to 1.0718 g. The individual core samples ranged in weight from 0.0884 g to 0.1366 g. These samples are the lowest weights and the first individual otolith cores ever used for radiometric age determination. Activities of 210Pb and 226Ra were determined and com- bined to form an activity ratio for each sample (Table 2). The 210Pb activity spanned a wide range and increased from juveniles to adults by as much as 88 times. The low- est activities were for the juvenile samples (0.003 and 0.005 disintergrations per gram |dpm/g] ) and the highest activity was 0.265 dpm/g (±6.0%) for the largest male ( 1620 mm FL). The activity of 226Ra varied by an order of magnitude and ranged from 0.044 dpm/g (±1.53%) to 0.401 dpm/g (±1.02%). Calculated 210Pb:226Ra ranged, as predicted, between 0 and 1; the lowest activity ratios were for the juvenile samples and the highest were for large adults. All low and high ratios were within the limits of 0 to 1 (values >1 are mathematically undefined; Eqs. 1 and 2) except for the largest female (2045 nun FL), which had an upper limit that exceeded 1. Radiometric age of the ju- venile samples was very close to the expected age (Table 3) . Exogenous 210Pb was, therefore, either not present or present in negligible quantities. This was inferred to be true for the core, or juvenile region, of adult Atlantic tar- pon otoliths. Comparison of annulus-derived ages and radiometric ag- es indicated there were differences in the aging results for each technique (Table 3 ). In most cases, the precision among the otolith readings could not explain the differences. In four cases, the different age estimates overlapped, but the extent of overlap was at the extreme of the age range. The lowest radiometric ages were one year for the samples from juvenile fish and the highest was 78.0 yr for the larg- est adult female. A graphical comparison of the expected 210Pb:226Ra dis- equilibria from ingrowth with the measured 210Pb:226Ra in- dicated there was variation in the measured results above and below what was expected (Fig. 1). The low radiomet- ric age for the juvenile age-groups indicated that uptake of exogenous 210Pb (R0; Campana et al., 1990) by juveniles was insignificant. Therefore, the best model for ingrowth of 210Pb from 226Ra in tarpon otoliths was for R0 = 0.0. 392 Fishery Bulletin 99(3) Table 1 Detailed summary of data for pooled otolith age groups and single otoliths from Megalops atlanticus. Age range of age groups is based on annulus-derived age estimates. Capture dates and average fork length are listed for comparison. Age groups consisted of 4 toll otolith cores and amounted to -1 g. Single otolith cores weighed close to 0.1 g. Age-group or single otolith age (yr) Capture date or range Fork length (mm ±SD) Number of otoliths Sample weight (g) Male 2-47 7 Mar-16 Oct 1989 568 ±80 7 0.6786 13 22 Jun 1989 1499 1 0.1129 17 4 Jul 1989 1452 1 0.1124 18-21 1 May-25 Jul 1989 1442 ±78 10 1.0498 31-36 12 Jun— 7 Jul 1990 1532 ±73 4 0.3480 32 6 Jun 1989 1588 1 0.1279 32 19 Jul 1992 1620 1 0.1220 Female 3-4 7 24 Jan-15 Jul 1989 583 ±59 9 0.8775 15-24- 6 Jun- 12 Aug 1989 1641 ±92 11 1.0718 35 4 Jul 1989 1613 1 0.1002 36 12 May 1992 1780 1 0.1331 37 12 Jun 1993 1950 1 0.1167 44 2 Sep 1991 2040 1 0.1366 48-50 20 Apr-20 Jun 1989 1708 ±110 4 0.3314 55 4 Mar 1991 2045 1 0.0884 1 Whole juvenile otoliths. 2 Ten otoliths ranging in age from 21 to 24 yr; one 15-yr-old otolith was erroneously included in this age group. Table 2 Radiometric results for each Megalops atlanticus age group and single otolith samples. Activities are expressed as disintegrations per minute per gram (dpm/g). Age-group or single otolith age (yr) Sample weight (g) 210Pb (dpm/g) ± % error7 226Ra (dpm/g) ± % error2 210Pb:226Ra activity ratio 210Pb:226Ra low 210Pb:226Ra high 2-4-3 0.6786 0.005 ±6.7 0.258 ±1.03 0.017 0.016 0.019 13 0.1129 0.043 ±12.1 0.085 ±1.52 0.507 0.439 0.577 17 0.1124 0.035 ±13.6 0.061 ±1.61 0.568 0.483 0.656 18-21 1.0498 0.072 ±3.4 0.246 ±1.01 0.292 0.280 0.305 31-36 0.3480 0.094 ±5.3 0.181 ±1.25 0.520 0.486 0.555 32 0.1279 0.031 ±13.5 0.044 ±1.53 0.712 0.607 0.821 32 0.1220 0.265 ±6.0 0.401 ±1.02 0.663 0.617 0.710 Female 3-45 0.8775 0.003 ±5.7 0.217 ±1.06 0.015 0.014 0.016 15-24 1.0718 0.059 ±3.9 0.163 ±1.07 0.362 0.345 0.381 35 0.1002 0.119 ±8.2 0.299 ±1.35 0.397 0.360 0.435 36 0.1331 0.086 ±8.5 0.144 ±2.24 0.597 0.534 0.663 37 0.1167 0.126 ±8.0 0.243 ±1.18 0.518 0.471 0.567 44 0.1366 0.113 ±7.4 0.170 ±1.33 0.667 0.609 0.726 48-50 0.3314 0.148 ±4.5 0.266 ±1.19 0.555 0.524 0.587 55 0.0884 0.086 ±12.2 0.094 ±1.42 0.909 0.787 1.035 7 Calculation based on standard deviation of 210Pb activity; Wang et al, 1975. 2 Calculation based on TIMS analysis routine (±1 SE). 3 Whole juvenile otoliths. Andrews et al.: Radiometric age validation of Megalops atlanticus 393 Figure 1 Expected 210Pb:226Ra ingrowth curves and observed 210Pb:226Ra activity ratios for male and female Atlantic tarpon ( Megalops atlanticus) individuals and age groups. Expected ingrowth curves repre- sent initial uptake ratios (R0) of 0.0, 0.1, and 0.2. Expected 210Pb:226Ra ingrowth during early oto- lith growth (until core-age) is based on a linear mass-growth model. After core growth, the secular equilibrium model for 210Pb:226Ra resumes and continues to unity. Data points are annulus-derived ages (average age for age groups) plotted against measured 210Pb:226Ra activity ratios. Vertical bars represent analytical uncertainty of 210Pb and Z26Ra measurements. Horizontal bars represent the range of annulus-derived ages for age groups. The upper limit of the 210Pb:226Ra ratio for the oldest female exceeded 1.0; therefore, high radiometric age is mathematically undefined (Eq. 2). Table 3 Comparison of annulus-derived ages and radiometric ages for Megalops atlanticus. The mean age of each group is based on annu- lus-derived age estimates. The radiometric age range is based on low and high activity ratios from analytical uncertainty calcula- tions. The radiometric age calculations are based on the measured ratio of 21uPb:226Ra. Age-group or single otolith age (yr) Mean annulus- derived age (yr) Radiometric age range (yr) Radiometric age (yr) Age-group or single otolith age (yr) Mean annulus- derived age (yr) Radiometric age range (yr) Radiometric age (yr) Male Female 2-4 3.0 1. 1-1.2 1.1 3-4 3.4 0.9-1. 1 1.0 13 13 19.6-28.7 23.7 15-24 22.6 14.6-16.4 15.5 17 17 22.2-35.3 28.0 35 35 15.3-19.3 17.2 18-21 19.0 11.5-12.7 12.1 36 36 25.5-35.9 30.2 31-36 33.8 22.4- 27.0 24.6 37 37 21.5-27.8 24.5 32 32 31.0-56.2 41.0 44 44 31.2-42.6 36.3 32 32 31.8 - 40.8 35.9 48-50 48.8 24.8-29.4 27.0 55 55 50.6-undefined7 78.0 1 Upper radiometric age limit for the 78-yr female is undefined because the 210Pb:226Ra ratio exceeded 1.0. 394 Fishery Bulletin 99(3) Hence, radiometric age was determined on the basis of measured activity ratios and no adjustment for exogenous 210Pb was necessary. When the radiometric and annulus reading (±12% CV) age ranges were considered graphical- ly, most of the data points still did not agree with the ex- pected ingrowth curve (2?0=0.0). A direct comparison of annulus-derived age with radio- metric age indicated that the ages were evenly distributed on either side of a line of agreement and that the slope of the regression (slope=0.915) was close to 1 (Fig. 2). Radio- metric age was not statistically different from the annu- lus-derived age estimates (paired two-tailed t-test, df=14, t=0.4181, P=0.6822). The coefficient of determination was low (adjusted /-2=0.55; Kvalseth, 1985) and the power of the test to detect significant differences was low because of the variability of the data and the small sample size. The analytical uncertainty associated with radiometric age determination or the reading range of the annulus- derived age encompassed the line of agreement for only four samples. The sample with the closest agreement was a female (1780 mm FL) whose annulus-derived age was 36 years. The annulus-derived age estimates of the juve- nile age groups were high by 1.9 and 2.4 years. For older age groups and individual cores, annulus-derived age was higher than radiometric age by as much as 21.8 years and lower than radiometric age by as much as 23 years. Otolith sections that had radiometric and annulus-derived age es- timates that differed considerably were photographed to il- lustrate the difficulty in aging otoliths (Fig. 3). Von Bertalanffy growth functions fitted to the radiomet- ric ages of males and females (Figs. 4 and 5) were similar to annulus-derived growth functions determined by Crab- tree et al. (1995). However, the low number of radiometric data points resulted in large confidence intervals (±95%) for growth model parameters. Radiometric-age growth pa- rameters for males indicated an asymptotic length (LJ of 1550 ±83 mm FL and a growth coefficient ( k ) of 0.19 ±0.12 (Fig. 4). Radiometric-age growth parameters for females indicated an of 2030 ± 227 mm FL and a growth coef- ficient ( k ) of 0.08 ±0.04 (Fig. 5). Discussion In four cases, the radiometric age range encompassed the annulus-derived age estimate. The oldest female tarpon had an estimate (55 yr) that was lower than the radio- metric age by 23 yr. The radiometric age range, however, encompassed the annulus-derived age by 4 years. This may indicate that the annulus-derived age estimate was correct in this case. The radiometric age val- idation, by itself, confirms the longevity of female Atlantic tarpon to at least 50.6 yr, but it may indicate that age can meet or exceed 78.0 yr (Table 3). Similarly, the longevity of male Atlantic tarpon was confirmed to at least 31.8 yr, but may exceed 41.0 yr. When the CV for annulus-derived age was included in the range of age estimates, some additional samples encompassed the radio- metric age estimates, but most deviant ages could not be explained by this variation. The wide dispersion of residuals and the low coefficient of determination indicated that there were differences between radio- metric age and annulus-derived age that could be explained in the interpretation of otolith growth zones (Fig. 2). According to Crabtree et al. (1995), aging otoliths of the Atlantic tarpon was confusing and many were not readable. Although Crabtree et al. (1995) validated the annual periodicity of growth zone formation for 12 out of 18 young fish up to an age of 9 yr in captivity, the pattern of growth zone formation in older, wild fish may not be annual. Because the tarpon is a migratory fish that inhab- its inshore and estuarine waters of vary- ing salinity and temperature and spawns offshore (Zale and Merrifield, 1989; Crab- tree et al., 1992), there is a high potential for irregular growth zone formation from the stresses of extreme changes in habitat (Pannella, 1980; Campana, 1983). Suban- nual growth zones may explain annulus- 100 80 -- 60 -- o 40 -- 20 -- S _o •3 c4 Upper limit undefined Line of agreement Regression • Male o Female y = 0.9 15r -0.085 Adj r2 = 0.55 20 40 60 Annulus-derived age (yr) 80 100 Figure 2 Comparison of male and female Atlantic tarpon (Megalops atlanticus) annulus-derived ages and radiometric ages. A linear regression and a line of agreement are plotted for comparison. Vertical bars represent low and high radiometric age estimates based on the analytical uncertainty of 210Pb and 226Ra measurements. Horizontal bars represent the range of annulus- derived ages for age groups. The upper limit of the radiometric age esti- mate for the oldest female is undefined because the analytical uncertainty of 210Pb:226Ra exceeds 1.0. Andrews et al.: Radiometric age validation of Megalops atlantlcus 395 Figure 3 Transverse otolith sections from three Atlantic tarpon ( Megalops atlanticus) illustrating the inherent difficulty in counting annuli. Annulus-derived ages were estimated by counting visible growth increments for each section shown (Crabtree et al. 1995). One sagitta was used for counting annuli and the other for radiometric aging. (A) An otolith section for which the number of growth increments clearly exceeds the radiometric age estimate. The annulus-derived age estimate was 37 yr and the radiometric age estimate was 24.5 yr. (B) An otolith section with fewer growth increments than the radiometric age estimate, but which could have had a higher annulus-derived age estimate if the reader had taken a more aggressive approach. The annulus-derived age estimate was 17 yr and the radiometric age estimate was 28.0 yr. (C) An otolith section with a well-developed sulcal region and numerous growth increments that rapidly become compressed toward the margin at the lower left. ( D ) Close up of the region in the same section (C) where growth increments are compressed and narrowly spaced. Note that growth increments in the region are partially obscured by what appears to be a proteinaceous inclusion. The annulus-derived age estimate was 55 yr and the radiometric age estimate was 78.0 yr. derived age estimates that were higher than radiometric age. Poorly defined annual growth zones, or growth zones that become too small or compressed to be quantified, may explain annulus-derived ages that were lower than radiometric ages (Fig. 3). An additional consideration was that the age groups were composed of otoliths pooled to- gether on the basis of annulus-derived age. This means that the radiometric age determined for these groups was the average age of the otoliths in the sample. Hence, a mixture of otoliths of differing ages would create an unac- countable error in the radiometric age of the age groups. This variability, however, was minimized by using oto- liths of similar weight. Von Bertalanffy growth functions fitted to the radio- metric ages had growth parameters that were similar to the annulus-derived growth parameters (Crabtree et al., 1995). This determination, however, is subjective because 1) there were a low number of radiometric samples, 2) the confidence intervals were large, and 3) three out of seven male samples and three out of eight female sam- 396 Fishery Bulletin 99(3) Figure 5 Von Bertalanffy growth curves for female Atlantic tarpon (Megalops atlanticus ) with radiometric age ( yr ) and fork length (mm; average fork length for age groups) and with annulus-derived age (yr) and fork length. Vertical bars represent fork length range for age-groups. Horizontal bars represent low and high radiometric age estimates. Andrews et ai.: Radiometric age validation of Megalops atlcinticus 397 pies were an average age and average fork length for each age group. The growth coefficient for male radio- metric ages (£=0.19 ±0.12) was similar to the annulus-de- rived estimate (£=0.12) according to the margin of error and was driven largely by one middle-age sample (Fig. 4). The growth coefficient for female radiometric ages (£=0.08 ±0.04) was also similar to the annulus-derived estimate (£=0.10). The male asymptotic length for radiometric ages (L„- 1550 ±83 mm FL) encompassed the result from annu- lus counts (Loc=1567 mm FL). For females, the asymptotic length was greater (L„= 2030 ±227 mm FL) but still en- compassed the result from annulus counts (LTC=1818 mm FL). The closeness of the radiometric estimate of to the maximum size was probably driven by the exceptionally large tarpon used in our study (Fig. 5). The large size of the 2-year-old otolith core, coupled with the relatively high 226Ra activity and high sensitivity of the TIMS technique, permitted age determination of smaller samples than previously possible (Kastelle et ah, 1994; Fenton and Short, 1995). In a recent radiometric aging study of the blue grenadier ( Macruronus novaezelandiae ), six pooled otolith cores were needed to attain a sample weight of approximately 1 g with measurable 226Ra activ- ity (Fenton and Short, 1995). Although the 226Ra activity for the samples in that study were a factor of 10 lower than the 226Ra activities observed in our study, the technique used was less sensitive than TIMS and resulted in rela- tively high analytical uncertainties (12-21%). In a recent study of sablefish (Anoplopoma fimbria ), where radium ac- tivities were similar to the activities observed in our study, the margin of error was relatively low (~4%), but otolith cores still needed to be pooled to attain approximately 1 g (Kastelle et ah, 1994). The use of TIMS to determine 226Ra in our study made it possible to age age-groups that were approximately one third of a gram and individual otolith cores that were less than 0. 1 g with analytical uncertain- ties typically less than 2%. Because of this advance, the greatest contribution to the aging error was no longer from 226Ra determination, but from the determination of 210Pb by means of a-spectrometry (Andrews et ah, 1999b). In our study, the determination of 210Pb activity contributed from 77% to 90%' of the aging error. The radiometric aging technique is well supported as a valid aging tool in numerous fish aging studies (Bergstad, 1995; Burton et ah, 1999). In our study, 226Ra activities varied by approximately an order of magnitude, but the 210Pb activities never exceeded the activity of 226Ra. There- fore, it is Highly unlikely that 210Pb was not the result of ingrowth from 226Ra incorporated during otolith formation. By performing a complete analysis on individual otoliths, we found that we could use radiometric age estimates with more confidence than estimates derived from pooled otolith samples. More recent successes with this technique (An- drews et ah, 1999a) suggest that it is a reliable means to obtaining accurate age estimates. Radiometric age deter- minations in our study, therefore, indicate that the inter- pretation of growth zones in Atlantic tarpon otoliths can be difficult, and in some cases can be inaccurate. In addition, our study provides conclusive evidence that the longevity of the Atlantic tarpon is greater than 50 years. Conclusions Accurate age determination of heavily harvested fish spe- cies is critical to the formulation of responsible man- agement strategies. The typical growth-zone validation techniques have limited applicability to long-lived fishes (Mace et ah, 1990; McFarlane and Beamish 1995). Under- estimated longevity and overfishing were factors that led to the decline of the Pacific ocean perch (Sebastes alutus) and the orange roughy (Hoplostethus atlanticus'. Beamish, 1979; Smith et ah, 1995). Because of the highly variable growth zone patterns (Fig. 3) and irregular life history of the Atlantic tarpon, conventional aging methods appear to be problematic for this species. The improved radio- metric aging technique used in our study incorporated a growth-zone independent chronometer that, when judi- ciously applied, will enable accurate age determination of many threatened species. At a time when many fish species are suffering from fishing pressure and changing oceanographic conditions, new methods need to be applied to help ascertain appropriate management strategies. The radiometric aging technique, which has been success- fully applied to the sablefish, whose longevity has been validated with other techniques (Kastelle et al., 1994, Beamish and McFarlane 2000), promises to be a valuable tool for aging species with unknown or difficult-to-inter- pret otolith growth patterns. Acknowledgments This work was supported in part by funding from the Department of the Interior, U.S. Fish and Wildlife Ser- vice, Federal Aid for Sportfish Restoration, project number F-59. Measurement of radium samples with TIMS was performed by colleagues in the Department of Earth Sci- ences at University of California, Santa Cruz. We espe- cially thank Craig Lundstrom, Zenon Palacz, and Pete Holden, and the University of California, Santa Cruz. This work was presented at the First International Tarpon Symposium held at the University of Texas Marine Sci- ence Institute in Port Aransas, Texas, 15-16 February 2001. We also thank Joan Holt and the organizers of the symposium, including Paul Swacina of Tarpon Tomorrow, a nonprofit organization aimed at tarpon restoration and conservation. Additional funding was provided by a grant from the National Sea Grant College Program, National Oceanic and Atmospheric Administration, U.S. Depart- ment of Commerce, under grant number NA36RG0537, project number R/F-148 through the California State Resources Agency. Literature cited Andrews, A. H., G. M. Cailliet, and K. H. Coale. 1999a. Age and growth of the Pacific grenadier ( Coryphae - noides acrolepis ) with age estimate validation using an improved radiometric ageing technique. Can. J. Fish. Aquat. Sci. 56:1339-1350. 398 Fishery Bulletin 99(3) Andrews, A. H., K. H. Coale, J. L. Nowicki, C. Lundstrom, Z. Palacz, E. J. Burton, and G. M. Cailliet. 1999b. Application of an ion-exchange separation technique and thermal ionization mass spectrometry to 226Ra deter- mination in otoliths for radiometric age determination of long-lived fishes. Can. J. Fish. Aquat. Sci. 56:1329-1338. Beamish, R. J. 1979. New information on the longevity of Pacific ocean perch (Sebastes alutus). J. Fish. Res. Board Can. 36:1395-1400. Beamish, R. J., and G. A. McFarlane. 2000. Reevaluation of the interpretation of annuli from oto- liths of a long-lived fish , Anoplopoma fimbria. Fish. Res. 46:105-111. Bergstad, O. A. 1995. Age determination of deep-water fishes: experiences, status and challenges for the future. In Deep-water fish- eries of the North Atlantic oceanic slope, NATO ASI series, series E: applied sciences, vol. 296 (A. G. Hopper, ed.), p. 267-283. Kluwer Academic Press, Netherlands. Burton, E. J., A. H. Andrews, K. H. Coale, and G. M. Cailliet. 1999. Application of radiometric age determination to three long-lived fishes using 210Pb:226Ra disequilibria in calcified structures: a review. In Life in the slow lane: ecology and conservation of long-lived marine animals (J. A. Musick, ed. ), p. 77-87. Am. Fish. Soc. Symp. 23, Bethesda, MD. Campana, S. E. 1983. Calcium deposition and otolith check formation during periods of stress in coho salmon, Oncorhynchus kisutch. Comp. Biochem. Physiol. 75A:215-220. Campana, S. E., K. C. Zwanenburg, and J. N. Smith. 1990. 210-Pb/226-Ra determination of longevity in redfish. Can. J. Fish. Aquat. Sci. 47:163-165. Crabtree, R. E., E. C. Cyr, R. E. Bishop, L. M. Falkenstein, and J. M. Dean. 1992. Age and growth of tarpon, Megalops atlanticus, larvae in the eastern Gulf of Mexico, with notes on relative abundance and probable spawning areas. Environ. Biol. Fishes, 35:361-370. Crabtree, R. E., E. C. Cyr, and J. M. Dean. 1995. Age and growth of tarpon. Megalops atlanticus, from South Florida waters. Fish. Bull. 93:619-628. Fenton, G. E., and S. A. Short. 1995. Radiometric analysis of blue grenadier, Macruronus novaezelandiae, otolith cores. Fish. Bull. 93:391-396. Flynn, W. W. 1968. The determination of low levels of polonium-210 in environmental materials. Anal. Chim. Acta 43:221-227. Kastelle, C. R., D. K. Kimura, A. E. Nevissi, and D. R. Gunderson. 1994. Using Pb-210/Ra-226 disequilibria for sablefish, Ano- plopoma fimbria, age validation. Fish. Bull. 92:292-301. Kvalseth, T. O. 1985. Cautionary note about R2. Am. Stat. 39:279-285. Mace, P. M., J. M. Fenaughty, R. P. Coburn, and I. J. Doonan. 1990. Growth and productivity of orange roughy ( Hoploste - thus atlanticus) on the north Chatham Rise. N.Z. J. Mar. Freshwater Res. 24:105-109. McFarlane, G. A., and R. J. Beamish. 1995. Validation of the otolith cross-section method of age determination for sablefish (Anoplopoma fimbria) using oxytetracycline. In Recent developments in fish otolith research (D. H. Secor, J. M. Dean, and S. E. Campana, eds.), p. 319-329. The Belle W. Baruch Library in Marine Sci- ence 19, Univ. South Carolina Press, Columbia, SC. Pannella, G. 1980. Growth patterns in fish sagittae. In Skeletal growth of aquatic organisms (D. C., Rhoades and R.A. Lutz, eds.) p. 519-560. Kluwer Academic/Plenum Publishing Corp., New York, NY. Saila, S. B., C. W. Recksiek, and M. H. Prager. 1988. Basic fisheries science programs: a compendium of microcomputer programs and manual of operation. Devel- opments in aquaculture and fisheries science (18). Else- vier Science Publishers, Amsterdam, 230 p. Smith, D. C., G. E. Fenton, S. G. Robertson, and S. A. Short. 1995. Age determination and growth of orange roughy (Hoplostethus atlanticus ): a comparison of annulus counts with radiometric ageing. Can. J. Fish. Aquat. Sci. 52:391- 401. Smith, J. N., R. Nelson, and S. E. Campana. 1991. The use of Pb-210/Ra-226 and Th-228/Ra-228 disequi- libria in the ageing of otoliths of marine fish. In Radionu- clides in the study of marine processes (P. J. Kershaw and D. S. Woodhead, eds.), p. 350-359. Elsevier Applied Sci- ence, New York, NY. Wang, C. H„ D. L. Willis, and W. D. Loveland. 1975. Radiotracer methodology in the biological, environ- mental, and physical sciences. Prentice Hall, Englewood Cliffs, NJ, 480 p. Zale, A. V., and S. G. Merrifield. 1989. Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (South Florida) — ladyfish and tarpon. U.S. Fish Wildl. Serv. Biol. Rep. 82(11.104). U.S. Army Corps of Engr. Report TR EL- 82-4, 17 p. 399 Abstract— Experiments were conduct- ed to determine the effect of the salt-box catch-bycatch separation procedure, as used by the Texas shrimp industry, on short-term survival of bycatch. Bio- assays were conducted on five eco- nomically important bycatch species: spotted seatrout ( Cynoscion nebulo- sus ); red drum (Sciaenops ocellatus ); Atlantic croaker ( Micropogonias undu- latusY, southern flounder ( Paralichthyes lethostigma ); and blue crab iCallinectes sapidus). Red drum were most affected by hypersalinity, requiring 17 minutes exposure to a 70%c salt water solution to kill 50% of the test specimens within 48 hours. For samples collected from commercial boats and Texas Parks and Wildlife (TPW ) trawl samples, we found that neither initial nor final percent survival was significantly different for bycatch removed with or without the aid of a salt-box. Bycatch mortality was high regardless of the method used to separate bycatch from the target catch. At the conclusion of catch sepa- ration, bycatch survival averaged 76% (±22%) for commercial samples and 48% (±40%) for TPW trawl samples separated with salt-boxes. Survival at the conclusion of catch separation with- out a salt-box averaged 56% (±35%) for commercial samples and 43% (±39%) for TPW trawl samples. Bycatch sur- vival 21-27 h after catch separation averaged 13% (±6%-) for commercial samples and 5% (±9%) for TPW trawl samples separated with salt-boxes and 34% (±29%) for commercial samples and 10% (±19%) in TPW trawl samples sepa- rated without a salt-box. Mortality rates ( M ) for bycatch separated with a salt- box averaged 0.08 (±0.03) for commer- cial samples and 0.10 (±0.04) for TPW trawl samples. For bycatch separated without the salt-box, M averaged 0.48 (±1.23) for commercial samples and 0.10 (±0.05) for TPW trawl samples. Results of an exploratory analysis with stepwise multiple regression suggested that final percent survival of bycatch was most affected by trawling time. The salt-box had little or no effect on bycatch sur- vival; therefore, regulating the use of salt-boxes in shrimp trawling operations is not necessary. Manuscript accepted 9 March 2001. Fish. Bull. 99:399-409 (2001). Effect of the salt-box catch-bycatch separation procedure, as used by the Texas shrimp industry, on short-term survival of bycatch Robert L. Colura Britt W. Bumguardner Perry R. Bass Marine Fisheries Research Station Texas Parks and Wildlife Department HC 02 Box 385 Palacios, Texas 77465 E-mail address (for R L. Colura): bob.colura@tpwd. state. tx. us Shrimp ( Penaeus spp. ) are the most important commercial seafood product in Texas, accounting for more than 80% by weight and 93%- of the value of all Texas commercial fisheries land- ings (Robinson et al., 1996). Texas shrimp landings from 1986 through 1995 were about 32,000-44,000 metric tons (t) annually. About 6400-8000 t of the annual catch during the 10-year period came from Texas bays. Texas bay shrimpers are reported to catch 2. 4-6. 8 kg of nontarget species (bycatch) for each kg of shrimp landed (Fuls and McEachron, 1998). Removal of the bycatch is time consuming and thus costly to fishermen. To reduce the time spent removing bycatch, Texas bay shrimpers frequently use salt-boxes to assist in rapid separation of bycatch from shrimp (Bumguardner and Colura, 1997). Salt-boxes are onboard tanks that hold a hypersaline solution of sea- water and food-grade salt. The catch is placed in the solution, where most fish species float and shrimp sink. Floating bycatch is skimmed from the surface and discarded. Shrimp are then dipped from the tank and any remain- ing bycatch removed and discarded. Although use of salt-boxes is known to fisheries managers, the effect of its use on the survival of discarded organisms is unknown. The objectives of our study were to assess the effects of the use of the salt- box on bycatch survival. This was ac- complished by first conducting bioas- says to determine the exposure time to hypersaline conditions that would af- fect survival of five economically impor- tant species known to occur as a part of the bycatch (Fuls and McEachron, 1998). The combined effects of trawl- ing and exposure to hypersaline con- ditions in the salt-box were evaluated by comparing survival and mortality rates of bycatch (all species) and three major components of the bycatch (At- lantic croaker, Micropogonias undula- tus, sand seatrout, Cynoscion arena- rius, and spot, Leiostomus xanthurus ) separated with and without the aid of a salt-box. This comparison was made for a series of samples collected from Texas bay shrimp fishermen and a series of samples collected during control exper- iments conducted by Texas Parks and Wildlife (TPW) personnel using exper- imental trawls. The TPW control ex- periments were conducted to remove some of the variation caused by the differing fishing methods used by each bay shrimper (i.e. different lengths of time the trawl was fished). Materials and methods Bioassays Bioassays were conducted on spotted seatrout (Cynoscion nebulosus), red drum (Sciaenops ocellatus), Atlantic croaker, southern flounder (Pciralich- thys lethostigma), and blue crab (Cal- linectes sapidus). The species were selected because of their importance to Texas’ recreational or commercial fish- eries, or both (Weixelman et al., 1992; Robinson et al., 1996) and because they are known to occur as a portion of the bycatch (Fuls and McEachron, 1998). Red drums and blue crabs were col- 400 Fishery Bulletin 99(3) Table 1 Exposure times, salinity, temperature, and number of individuals per replicate used in high-salinity exposure bioassays. Three replicates were used at each exposure time. Species Exposure times (min) Salinity (' %o ) Temperature (°C) Bay water salinity (%o) Fish/ replicate Red drum 4, 8, 16, 32, 64, 128 70 24 19 10 Atlantic croaker 4,8, 16,32,64, 128 70 23 16 5 Spotted seatrout 4, 8, 16, 32, 64; 69 26 24 5 Southern flounder 4, 8, 16, 32, 64, 128 71 27 18 5 Blue crab 4,8, 16,32,64, 128 70 26 24 5 ' Specimens did not survive to 128 min of exposure. lected from TPW hatchery ponds; Atlantic croaker and southern flounder were collected from Matagorda Bay with a trawl; spotted seatrout were collected form East Matagorda Bay with a bag seine. Fish were held 7-10 days in 1850-L tanks at ambient bay salinities and tem- peratures and fed frozen brine shrimp or chopped shrimp (or both) three times daily. Blue crabs were placed in hold- ing tanks for one day prior to bioassay initiation. Juve- nile blue crabs molt at frequent intervals and because of their aggressive nature, any soft crabs in a common tank are almost immediately cannibalized. Therefore, blue crab trials were held before substantial mortality occurred in holding tanks and without an initial acclimation period. Bioassay trials were conducted at salinities of 70%c (±1 %c) and temperatures from 23 to 27°C. A salinity of 70%z =34). Re- maining species generally averaged <1 per sample. Some differences were observed in the frequency and length of some of the by catch species that composed the NSB and SB samples. However, differences were generally small and were probably biologically insignificant. Atlantic croaker, bay anchovy, hardhead catfish ( Arius felis), and brown shrimp were found in significantly greater numbers in SB samples than in NSB samples (Table 5). Bay whiff (Ci- tharichthys spilopterus), bluntnose jack ( Hemicaranx am- blyrhynchus), Atlantic spadefish ( Chaetodipterus faber ), and inshore lizardfish ( Synodus foetens ) were found in sig- nificantly greater numbers in NSB samples than in SB samples, although they were rare in NSB samples. Mean number of specimens collected for all other species were similar in both sample types. Mean length of Atlantic croaker, Gulf menhaden, spot, sand seatrout, and pinfish (Lagodoii rhomboides) were significantly greater in NSB samples (Table 5). Blue crabs collected in SB samples had a significantly greater mean carapace width than those collected in NSB samples. Mean lengths of all other spe- cies were similar in both sample types. In general, all three measures of salt-box effect on by- catch were highly variable (Table 6). Mean (±SD) initial Colura and Bumguardner: Effect of the salt-box catch-bycatch separation procedure, as used by the Texas shrimp industry 403 Table 4 Species collected from bycatch of Texas bay shrimpers and in Texas Parks and Wildlife (TPW) samples. X indicates that the species was collected. Common name Species Bay shrimpers TPW Vertebrates Atlantic bumper Chloroscombrus chrysurus X Atlantic croaker Micropogonias undulatus X X Atlantic cutlassfish Trichiurus lepturus X X Atlantic midshipman Porichthys plectrodon X X Atlantic spadefish Chaetodipterus faber X X Atlantic stingray Dasyatis sabina X Atlantic threadfih Polydactyl us octonemus X bay anchovy Anchoa mitchilli X X bay whiff Citharichthys spilopterus X X bighead searobin Prionotus tribulus X X blackcheek tonguefish Symphurus plagiusa X X bluefish Pomatomus saltatrix X bluntnose jack Hemicaran x am blyrhyn ch us X X gafftopsail catfish Bagre marinus X X gulf butte rfish Peprilus burti X gulf menhaden Brevoortia patronus X X hardhead catfish Alius fells X X harvestfish Perprilus alepidotus X X hogchoker Trinectes maculatus X X inshore lizardfish Synodus foetens X X least puffer Sphoeroides parvus X X lookdown Selene vomer X pinfish Lagodon rhomboides X X sand seatrout Cynoscion arenarius X X silver jenny Eucinostomus gula X silver perch Bairdiella chrysoura X X skipjack herring Alosa chrysochloris X southern flounder Paralichthys lethostigma X X southern kingfish Menticirrhus americanus X spot Leiostomus xanthurus X X spotted seatrout Cynoscion nebulosus X star drum Stellifer lanceolatus X threadfin shad Dorosoma petenense X Invertebrates Atlantic brief squid Lolliguncula brevis X X blue crab Callinectes sapidus X X brown shrimp Penaeus aztecus X X white shrimp Penaeus setiferus X survival rates were 76.1% (±21.7%) and 55.5% (±35.1%), respectively for SB and NSB samples. Final survival rates were 12.9% (±6.0%) for SB samples and 34.5% (±28.9%) for NSB samples. Means of initial and final survival were not significantly different. Average (±SD) M of organisms separated by use of a salt-box (0.08 [±0.03]) were signifi- cantly lower than M of bycatch separated without the aid of a salt-box (0.48 [±1.23]). For individual species, comparison of the effect of use and non-use of a salt-box on M , and on initial and final survival variables could only be determined for Atlantic croaker. Other species were either found infrequently in individual samples, or in insufficient numbers within the sample type. Estimates of initial and final survival for At- lantic croaker were obtained from 19 samples (6 NSB sam- ples and 13 SB samples). Estimates of M were obtained from 16 samples (4 NSB samples and 12 SB samples). Samples not used bad too few Atlantic croaker (i.e. only one or two specimens in the sample or no deaths) to make estimates of percent survival or M. Means of the three es- timates used to compare the two separation techniques were highly variable for Atlantic croaker (Table 6). Means 404 Fishery Bulletin 99(3) Table 5 Species which differed significantly in either mean numbers or total length (mm) from comparative samples collected from the commercial fishery and separated with or without the aid of a salt-box, and results of Wilcoxon signed rank tests used to compare means. Species Salt-box Mean +SD n No salt-box Mean ±SD n Z P Number of specimens bay anchovy 2.7 ±4.9 15 0.3 ±0.6 15 3.159 0.002 Atlantic croaker 28.8 ±14.5 15 6.9 ±5.0 15 4.030 <0.001 hardhead catfish 2.5 ±3.3 15 0.7 ±1.4 15 2.150 0.032 bay whiff 0.1 ±0.3 15 1.3 ±1.4 15 -3.097 0.002 bluntnose jack 0 15 1.3 ±1.7 15 3.440 0.001 Atlantic spadefish 0 15 0.5 ±0.9 15 -2.366 0.018 inshore lizardfish 0 15 0.4 ±0.8 15 -2.073 0.038 brown shrimp 2.0 ±1.8 15 1.1 ±1.8 15 3.397 0.001 Total length Atlantic croaker 120 ±15 419 104 ±19 93 8.129 <0.001 gulf menhaden 110 ±47 120 119 ±22 57 3.460 <0.001 spot 98 ±23 157 109 ±20 60 3.910 <0.001 sand seatrout 103 ±30 45 129 ±25 44 4.830 <0.001 pinfish 75 ±13 9 89 ±20 82 -3.348 0.001 blue crab 82 ±23 60 66 ±28 16 -2.268 0.023 Table 6 Means of bycatch (all species) and Atlantic croaker initial percent survival, final percent survival, and mortality rate (M) and results of Wilcoxon signed rank tests comparing the means. All samples were collected from Texas bay shrimpers and separated with or without the aid of a salt-box. Variable Salt-box Mean ±SD n No salt-box Mean ±SD n Z P Bycatch initial survival 76.1 ±21.7 15 55.5 ±35.1 15 1.763 0.078 final survival 12.9 ±6.0 15 34.5 ±28.9 15 -1.389 0.093 M 0.08 ±0.03 15 0.48 ±1.23 15 2.605 0.009 Atlantic croaker initial survival 74.4 ±25.7 13 81.8 ±25.1 6 0.746 0.046 final survival 28.6 ±19.9 13 62.3 ±30.7 6 2.330 0.020 M 0.23 ±0.12 12 0.26 ±0.19 4 0.364 0.716 of M were similar between separation techniques. Means of initial (81.8% [±25. 1%] ) and final (62.3% | ±30.7%] ) sur- vival of fish separated without a salt-box were significant- ly greater than initial (74% [+25.7%]) and final (28.6% [±19.9%]) survival of those separated with a salt-box. By examining Spearman correlation coefficients that were significantly greater than 0, we found a relationship between the three measurements and factors other than use or non-use of a salt-box (Table 7). Initial and final survival estimates were negatively correlated to trawling time and catch separation time, whereas M was positive- ly correlated to the two variables. Only estimates of M (rs=0.488) were found to be related to the use or non-use of a salt-box. Final estimates of survival and M were also associated with some bycatch species. M was positively correlated to numbers of Atlantic croaker (rs=0.526) and sand seatrout (/'s=0.458) in the samples, whereas final sur- vival was inversely correlated to the two species (Atlantic croaker rs =-0.391, sand seatrout rs=-0.564). Other species to which M and final survival were associated occurred in fewer than half the samples or averaged <1 specimen per sample. Initial survival estimates were not related to Colura and Bumguardner: Effect of the salt-box catch-bycatch separation procedure, as used by the Texas shrimp industry 405 Table 7 Spearman correlation coefficients (r ) and probability (P) that the coefficient values were greater than 0, of variables that dem- onstrated relationships with mortality rate (M), initial percent survival, and final percent survival for bycatch (all species) and Atlantic croaker. Samples were collected from the catch of Texas bay shrimpers and separated with or without a salt-box. A space indicates the probability that the variables rs value was greater then 0 was >0.05. Variable M Initial survival Final survival rs P P rs P Bycatch separation time 0.638 <0.001 -0.586 0.001 -0.749 <0.001 trawling time 0.650 <0.001 -0.414 0.023 -0.665 <0.001 bay temperature -0.459 0.011 bay anchovy 0.477 0.008 Atlantic croaker 0.526 0.003 -0.391 0.033 gulf menhaden 0.445 0.014 -0.461 0.010 sand seatrout 0.458 0.011 -0.564 0.001 pinfish -0.434 0.017 bighead searobin -0.368 0.046 gafftopsail catfish -0.398 0.029 bay whiff 0.379 0.039 Atlantic cutlassfish -0.373 0.042 salt-box 0.488 0.006 Atlantic croaker separation time 0.637 0.008 0.459 0.016 least puffer 0.994 <0.001 hardhead catfish -0.592 0.001 blue crab -0.463 0.015 brown shrimp -0.566 0.002 harvestfish 0.401 0.038 0.400 0.038 Atlantic brief squid -0.833 0.039 salt-box -0.629 0.004 Table 8 Independent variables that entered stepwise multiple regression models describing bycatch (all species), initial percent survival, final percent survival, and mortality rate (M), and Atlantic croaker M resulting from Texas bay shrimpers separating them from shrimp with or without the aid of a salt-box. Included are y-intercepts±SE, coefficients±SE of the selected independent variables, partial r2 of each variable in parenthesis, and model r2. A space indicates the variable did not enter the model. Dependent variable Independent variable y-intercept Separation time Trawl time Bay temperature Salt-box Model R2 Bycatch initial survival 217.4 ±65.3 -120.5 ±21.8 (0.52) 0.52 final survival 60.0 ±6.9 -20.2 ±4.1 (0.48) -15.5 ±5.8 (0.11) 0.59 M 2.3 ±1.2 5.1 ±0.4 (0.79) -0.3 ±0.1 (0.03) -0.1 ±0.04 (0.03) 0.85 Atlantic croaker final survival 51.2 ±10.1 61.9 ±27.0 (0.11) -31.22 ±10.0(0.38) 0.49 any bycatch species. The initial survival (rs=0.459) and M (rs=0.637) of Atlantic croaker were related to catch separa- tion time. Atlantic croaker final survival (rs=-0.629) was negatively correlated to use of a salt-box. Stepwise multiple regression identified four variables, which explained most of the variation associated with the three measurements (Table 8). Catch separation time (total time that bycatch was on the vessel), trawling time, and 406 Fishery Bulletin 99(3) Table 9 Species collected from samples of bycatch separated with or without the aid of a salt-box in control experiments whose mean total length (mm) differed significantly and results of Wilcoxon signed rank tests used to compare the means. Species Salt-box No salt-box Z P Mean total length ±SD n Mean total length ±SD n Bay anchovy 50 ±11 191 53 ±13 219 -2.977 <0.001 Spot 94 ±18 138 89 ±24 194 2.818 0.005 Silver perch 72 ±37 4 111 ±22 18 -2.605 0.009 Atlantic brief squid 82 ±39 53 58 ±39 63 4.177 <0.001 bay temperature explained 85% of the variation associated with estimates of bycatch M. Catch separation time was the most important of the three, explaining 79% of the varia- tion. Catch separation time explained 52% of the variation observed in initial survival and was the only variable to meet the P< 0.05 significance level required for entry into the model. Trawling time explained 48% of the variation ob- served in bycatch final survival and use or non-use of a salt- box explained the remaining 11%. For the dependent vari- able, Atlantic croaker final survival, the use or non-use of a salt-box explained 38% of the variation explained by the model and catch separation time explained an additional 11%. The five independent variables did not meet the P<0.05 significance level required to enter or stay in stepwise re- gression models for Atlantic croaker initial survival and M. Effect of salt-box use on bycatch survival: control experiments Twenty-eight species were collected for the control experi- ments: 24 fishes and 4 invertebrates (Table 4 ). Bay anchovy (n=3794), Atlantic croaker (n=2048), spot (n=949), sand seatrout (n=291), and Gulf menhaden (n=148) were the most numerous organisms and were observed in more than half the samples collected. Remaining fishes aver- aged <1 individual per sample. Brown shrimp (n=494) and Atlantic brief squid ( Lolliguncula brevis, n = 133) were the most numerous invertebrates caught. By species, no sig- nificant differences were found between numbers of indi- viduals in SB and NSB samples. Bay anchovy, spot, silver perch ( Bairdiella chrysoura), and Atlantic brief squid col- lected in NSB samples were significantly larger than those collected in SB samples, but differences for the most part were small or the organism was present in small numbers (Table 9). Remaining species were similar in size in both NSB and SB samples. Catch separation time averaged 12 min (±6 min) for SB samples and 18 min (±18 min) for NSB samples and was significantly different (Z=2. 05399, P=0.04). At the time samples were collected, bay tempera- ture averaged 26.0° (±3.1°)C, bay salinity, 18,0%t>. (±7.2%e), and salt-box salinity 66%e (±8%o). Survival rates and mortality rates were highly variable for bycatch separated by both methods. Mean initial sur- vival rates were 48.4% (±39.7%) and 43.6% (±39.3%), re- spectively for SB and NSB samples. Final survival rates were 5.4% (±8.9%) for SB samples and 9.5% (±18.7%) for NSB samples. Mortality rates averaged 0.1 (±0.04) for SB samples and 0.1 (±0.05) NSB samples. All estimates of survival and mortality were statistically similar (Table 10). By species, estimates of initial survival, final survival, and M for Atlantic croaker, spot, and sand seatrout were also highly variable and were statistically similar between the two methods of bycatch separation. Spearman correlation coefficients significantly greater than 0 demonstrated a relationship between bycatch (all species) catch separation time and the variables initial survival (rs=0.445) and M(rs=0.522). (Table 11). Initial sur- vival was found to have an inverse relationship with tem- perature. All three measurements of the effect of salt-box use on bycatch were associated with the frequency of oc- currence of some species in the bycatch. The variable M was positively correlated to Atlantic croaker (rs=0.765), spot (rs=0.681), and sand seatrout (rs=0.792) numbers. Ini- tial survival was positively correlated to Atlantic croaker (;’s=0.630), spot(?'s=0.530), and blue crab (rs=0.494) num- bers and was inversely related to the numbers of bay an- chovy (rs=- 0.795) and Gulf menhaden (rs=-0.466). Final survival was positively correlated to blue crab (rs=0.631) numbers and inversely related to bay anchovy numbers (rs=-0.655). Other species to which M, and initial and fi- nal survival were related occurred in fewer than half the samples and averaged <1 organism per sample. Atlantic croaker M was related to bay temperature (rs=0.828), bay salinity (rs=-0.513), and to the presence of Atlantic brief squid (rs=-0.833). Neither Atlantic croaker initial nor final survival estimates were correlated to other variables. Spot and sand seatrout M, initial survival, and final survival were not correlated to any variable. For all species, stepwise multiple regression identified catch separation time, bay salinity, and bay temperature at the time of sample collection as significant variables which explained some of the observed variation in M and final survival (Table 12). The salt-box variable did not enter any model. Catch separation time explained 27% of the varia- tion that was accounted for by the model for the dependent variable M, and an additional 11% was explained by bay salinity at time of sample collection. Bay temperature and salinity at the time of collection explained 34% of variation observed in the final survival model. No variable met the P<0.05 significance level for entry into regression models Colura and Bumguardner: Effect of the salt-box catch-bycatch separation procedure, as used by the Texas shrimp industry 407 Table 10 Mean (±SD) of initial percent survival, final percent survival, and mortality rate (M) and results ofWilcoxon signed rank tests used to compare the means of bycatch (all species), Atlantic croaker, spot, and sand seatrout from bycatch samples collected in control experiments and separated with or without the aid of a salt-box. Variable Salt-box Mean ±SD n No salt-box Mean ±SD n Z P Bycatch initial survival 48.4 ±39.7 14 43.6 ±39.3 14 -0.460 0.643 final survival 5.4 ±8.9 14 9.5 ±18.7 14 -0.322 0.748 M 0.10 ±0.04 14 0.10 ±0.05 14 -0.277 0.782 Atlantic croaker initial survival 41.1 ±38.2 8 56.0 ±40.1 8 0.790 0.430 final survival 18.3 ±24.1 8 16.8 ±34.1 8 -0.373 0.709 M 0.41 ±0.77 8 0.82 ±1.27 8 0.105 0.916 Spot initial survival 70.9±30.3 12 71.7 ±27.3 12 -0.206 0.837 final survival 24.9±34.8 12 21.5 ±36.4 12 -0.695 0.519 M 1.30±2.71 12 2.07 ±3.26 12 0.799 0.424 Sand seatrout initial survival 85.7 ±28.8 11 72.3 ±26.7 11 -1.242 0.214 final survival 29.2 ±38.0 11 14.1 ±20.0 11 -0.774 0.439 M 3.58 ±6.31 11 4.27 ±6.38 11 0.417 0.677 Table 11 Spearman correlation coefficients (rs) and probability (P) that the coefficient values were greater than 0, of variables which demon- strated relationships with bycatch (all species) mortality rate (M), initial percent survival and final percent survival and Atlantic croaker M. Samples were collected in Texas Parks and Wildlife trawls and separated from shrimp with or without the aid of a salt- box during control experiments. A space indicates the probability that the variables rs value was greater then 0 was >0.05. Variable M Initial survival Final survival rs P '7 P rs P Bycatch separation time 0.522 0.004 0.445 0.018 Atlantic croaker 0.765 <0.001 0.630 <0.001 spot 0.681 <0.001 0.530 0.004 sand seatrout 0.792 <0.001 least puffer 0.692 0.039 bay temperature -0.4249 0.025 salt-box salinity -0.680 <0.001 -0.635 0.015 bay anchovy -0.795 <0.001 -0.655 <0.001 gulf menhaden -0.466 0.012 bay whiff 0.444 0.018 brown shrimp 0.433 0.021 brief squid -0.588 0.001 blue crab 0.494 0.008 0.631 <0.001 silver perch -0.386 0.042 -0.480 0.010 bluntnose jack -0.568 0.002 Atlantic croaker bay temperature 0.829 0.042 bay salinity -0.514 0.028 Atlantic brief squid -0.833 0.039 408 Fishery Bulletin 99(3) Table 12 Independent variables which entered stepwise multiple regression models describing bycatch (all species), Atlantic croaker, spot, and sand seatrout short-term final percent survival, or mortality rate (M) (or both) resulting from their separation from shrimp with or without the aid of a salt-box during controlled experiments. Included are y-intercepts (±SE), coefficients (±SE) of the independent variables, partial R2 of each variable in parenthesis, and model R2. A space indicates the variable did not enter the model. Independent variable Dependent variable v- intercept Separation time Bay salinity Bay temperature Model R2 Bycatch final survival 216.59 ±58.069 -2.71 ±0.88 (0.13) -4.68 ±2.08 (0.21) 0.34 M 0.12 ±0.03 0.09 ±0.03 (0.27) -0.002 ±0.001 (0.11) 0.38 Atlantic croaker M -5.36 ±203 2.25 ±0.45 (0.75) 0.21 ±0.08 (0.08) 0.83 Spot M Sand seatrout final survival M -1.44 ±1.41 213.31 ±67.36 -39.75 ±12.90 0.18 ±0.07 (0.20) -7.08 (0.29) 1.61 ±0.47(0.37) 0.20 0.29 0.37 for the dependent variable initial survival. Time required to separate catch accounted for 75% of the variation ob- served in Atlantic croaker M, whereas bay temperature at time of sample collection accounted for the remaining 8% variation explained by the model. Bay salinity at time of sample collection explained 20% of the variation observed in M of spot. Bay temperature at time of sample collection explained 37% and 29%, respectively, of the dependent variables M and initial survival of sand seatrout. No vari- able met the P<0.05 significance level for entry into models to explain initial survival of Atlantic croaker, spot, or sand seatrout, or final survival of Atlantic croaker or spot. Discussion Use of a salt-box to separate bycatch, as practiced by Texas shrimpers, had little or no effect on short-term survival of bycatch. Measurements of bycatch (all species) initial and final survival were not significantly different for the bycatch separation methods in samples collected from bay shrimpers and from TPW trawls. Significant differences between SB and NSB Atlantic croaker initial and final survival of samples collected from the commercial fishery were not corroborated by the results of the TPW control experiments. Exposure time to hypersaline conditions in a salt-box is short, generally less than two min, and the no-observed-effect exposure for selected species suggests that longer exposure (at least eight min for red drum and Atlantic croaker) would probably be required to signifi- cantly affect survival of most bycatch species. For these reasons, regulation of salt-box use as currently practiced by Texas bay shrimpers is unnecessary. The use of a salt-box had little impact on bycatch M, and initial or final survival. Salt-box use was correlated to M, for samples collected from the fishery, but not to initial or final survival. Salt-box use was not correlated to any of the three variables for bycatch in the control experiments. Atlantic croaker final survival was negatively correlated to the pres- ence of a saltbox in samples collected from the fishery but the correlation was not present in the control experiments. The salt-box variable entered the final survival regression models for combined catch and Atlantic croaker from the fishery but entered no other regression models. Environmental factors at the time of sample collection played a role in determining survival of bycatch, although not as great as other variables associated with bycatch separation. Both bay temperature and salinity at the time of sample collection affected survival and M variables in the control experiments as demonstrated by the correla- tion and regression analyses. Bay temperature also ex- plained a small portion of the variation (3%) associated with M in trials conducted on samples collected from the fishery. For individual species, bay temperature and salin- ity, either individually or combined, explained part of the variation associated with the regression models for M of Atlantic croaker, spot, and sand seatrout, and for final sur- vival of sand seatrout. The effect of temperature on M in samples taken from the fishery and survival and M in con- trol experiments suggest there may be a seasonal effect on survival of bycatch. Failure to observe a greater effect in trials with the fishery samples probably is due to the effect of other factors associated with the fishery, such as long trawl times and catch separation times that mask the ef- fect of environmental factors. Catch separation time was one of the most important variables associated with bycatch survival. It was corre- lated to M, initial survival, and final survival for samples collected from the fishery and to M and initial survival in the control experiment. The variable entered multiple re- Colura and Bumguardner: Effect of the salt box catch-bycatch separation procedure, as used by the Texas shrimp Industry 409 gression models for the variables M and initial survival of bycatch in experiments conducted with samples collected from the fishery. It also entered multiple regression models for the dependent variable M of bycatch and Atlantic croak- er in the control experiments. The use of the salt-box sig- nificantly reduced catch separation time in control experi- ments, suggesting that the use of the salt-box could possibly reduce M by speeding catch separation time. The impor- tance of reducing catch separation time has been corrobo- rated by Ross and Hokenson (1997) who reported a posi- tive correlation between mortality and on-deck sorting time for three species of bycatch (American plaice, Hippoglossoi- des platessoides\ witch flounder, Glyptocephalus cynoglos- sus; and pollock, Pollachius virens ) associated with the Gulf of Maine northern shrimp (. Pandalus borealis ) fishery. Trawling time was also an important variable affecting M and survival. It was correlated to M, initial survival, and final survival of bycatch and entered the final survival re- gression model for bycatch from the fishery. This variable may affect survival in several ways. Longer trawl times in general result in larger catches. This would presum- ably increase catch separation time thus contributing to increased bycatch mortality. Long trawl times would also increase the time bycatch is in the net, thus increasing the chance of fatal injuries. In the absence of a satisfactory method to reduce the catch of nontarget species, survival of bycatch could probably be improved by limiting the time a trawl is fished and by re- quiring the catch to be separated prior to resumption of fish- ing. These variables were the only ones found in our study that had significant impact on bycatch survival and that could easily be altered to improve survival. Enforcement of either of these proposals, however, would be difficult. Use of percent survival estimates at the conclusion of the observation periods in our study to predict long-term survival of bycatch released into the bay is not recom- mended. Stress in fishes is characterized by hypersecretion of epinephrine, norepinephrine, and corticosteroids. These hormones induce secondary effects resulting in changes in metabolism, osmoregulation, and the immune systems (Mazeaud et al., 1977) that last for several days and can re- sult in increased susceptibility to disease (Wedemeyer and McLeay, 1981). Furthermore, stress can lead to behavioral changes (Wedemeyer, 1974) that might make the fish more susceptible to predation. Specimens can be maintained un- der identical conditions (in the laboratory or field) to com- pare survival and mortality rates of individuals treated dif- ferently at capture. However, it would be speculative to project survival estimates of specimens maintained in any form of captivity to those released into the wild because it is impossible to account for predation and possibly disease. Acknowledgments We thank Captains Ted Bates Jr., Dennis Williams, and Bobby Pash for allowing us to board their boats and collect samples. We also thank TPW personnel Norman Boyd, Karen Meador, and Dennis Pridgen for their assistance in finding cooperating shrimpers; Eric Young, David Westbrook, Roberta Vickers, and Valentin Flores III for their assistance in sample collection; and Cynthia Gibbs for her assistance in manuscript preparation. Lastly, we thank Rocky Ward, Mark R. Fisher, Lawrence W. McEachron, Rebecca Hensley, Robin K. Riechers, David Abrego, and William J. Karel of TPW, and three anonymous reviewers for their critical review of the manuscript. This project was jointly funded by Texas Parks and Wildlife and the U. S. Department of Commerce, National Marine Fisheries Service, Marine Fisheries Initia- tive (MARFIN) award number NA57FF0047. Literature cited Bumguardner, B. W., and R. L. Colura. 1997. A description of salt-box use by the Texas bay shrimp fishery. Management Data Series 140, Texas Parks and Wildlife, Coastal Fisheries Division, Austin, TX, 9 p. Fuls, B. E., and L. W. McEachron. 1998. Evaluation of bycatch reduction devices in Aransas bay during the 1997 spring and fall commercial bay-shrimp season. Corpus Christi National Estuary Program, CCB- NEP-33, Texas Natural Resource Conservation Commis- sion, Austin, TX, 33 p. Kana, J. C., J. A. Dailey, B. Fuls, and L. W. McEachron. 1993. Trends in relative abundance and size of selected fin- fishes and shellfishes along the Texas coast: November 1975-December 1991. Management Data Series 103, Texas Parks and Wildlife Department, Fisheries and Wild- life Division, Austin, TX, 92 p. Mazeaud, M., F. Mazeaud, and E. M. Donaldson. 1977. Primary and secondary effects of stress in fish: some new data with a general review Trans. Am. Fish. Soc. 106: 201-212. Rand, G. M., and S. R. Petrocelli. 1985. Fundamentals of aquatic toxicology: methods and appli- cations. Hemisphere Publishing, Washington, D.C., 666 p. Robinson L., P. Campbell, and L. Butler. 1996. Trends in Texas commercial fishery landings, 1972- 1995. Management Data Series 127, Texas Parks and Wildlife, Fisheries and Wildlife Division, Austin, TX, 169 p. Ross, M. R., and S. R. Hokenson. 1997. Short-term mortality of discarded finfish bycatch in the Gulf of Maine fishery for northern shrimp Pandulus borealis. North Am. J. Fish. Manage. 17:902-909. SAS Institute, Inc. 1990. SAS/STAT user’s guide, version 6, vols. 1 and 2. SAS Institute Inc., Cary, NC, 1674 p. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, 2nd ed. Freeman, San Francisco, CA, 859 p. Wedemeyer, G. A. 1974. Stress as a predisposing factor in fish diseases. Fish Disease Leaflet-38, U. S. Department of the Interior, Fish and Wildlife Service, Division of Cooperative Research. Washington, D.C., 8 p. Wedemeyer, G. A., and D. J. McLeay. 1981. Methods for determining the tolerance of fishes to environmental stressors. In Stress and fish (A. D. Picker- ing, ed.), p. 247-275. Academic Press, London, U.K. Weixelman, M., K. W. Spiller, and P. Campbell. 1992. Trends in finfish landings of sport-boat anglers in Texas marine waters, May 1974-May 1991. Management Data Series 85, Texas Parks and Wildlife Department, Fisheries and Wildlife Division, Austin, TX, 226 p. 410 Trace metals in four structures of fish and their use for estimates of stock structure Bronwyn M. Gillanders School of Biological Sciences A08 University of Sydney New South Wales 2006, Australia Present address: Department of Environmental Biology University of Adelaide, South Australia 5005, Australia E-mail:bronwyn. gillanders@adelaide.edu.au Abstract— Trace elements in calcified tissues have been suggested as one of the most powerful means for stock dis- crimination yet developed. The struc- ture of choice for determining elemental composition or fingerprints is the oto- lith, although other structures also incorporate trace elements into their matrix. The aim of this study was to compare the elemental fingerprints of four structures (otoliths, scales, eye lenses, and spines) of a territorial reef fish to determine whether there were correlations between otoliths and each of the other structures. Elemental fin- gerprints of juvenile (<3 years of age) and adult fish (which may reach a maximum age of 37 years) were also compared for each structure to deter- mine whether there may be differences between size classes of fish. All struc- tures were analyzed by solution-based inductively coupled plasma-mass spec- trometry (ICP-MS). Otoliths, scales, spines, and eye lenses differed in com- position. Calcium dominated otoliths, scales, and spines but was not detected in eye lenses. Some elements, for exam- ple barium, showed significant correla- tions between the otolith data and that of scales and spines of both juvenile and adult fish. A multivariate test of matrix correspondence (Mantel's test) detected significant relationships between the otolith data and the data matrices for each of scales and spines of both juve- nile and adult fish. Significant rela- tionships between the otolith data and the eye lens data were detected only for juvenile fish. Significant differ- ences were also found between juvenile and adult fish for all structures. The BIOENV multivariate analyses showed that the highest rank correlation was found between the otolith data and the scale or spine data for both juvenile and adult fish. These data suggest that the use of scales and spines may pro- vide a nonlethal alternative to the use of otoliths for future stock discrimina- tion studies. Manuscript accepted 16 February 2001. Fish. Bull. 99:410-419 (2001). Studies of the population dynamics of marine fishes depend on the ability to distinguish separate “stocks” that may make up the total population of a single fish species. Although much of fisheries management assumes that a single pop- ulation is being monitored, measures of growth, natural mortality, recruitment, and reproduction are often made for the total population because data on move- ment or stock separation or fidelity (or the combination of all three) are inad- equate to allow separation of stocks. A variety of methods exist for iden- tifying stocks (e.g. population param- eters, capture-mark-recapture studies, physiological and behavioral characters, morphometric and meristic characters, calcareous characters, cytogenic charac- ters, and biochemical characters; Ihssen et al., 1981), but few methods can pro- vide a reliable measure of stock identity. Genetic techniques are frequently used but fail to differentiate stocks because a small amount of larval or adult mixing among populations makes differences undetectable (Hartl and Clark, 1989). Recently, many studies have focused on the use of “elemental fingerprints” ( sensu Campana et al., 1994), or the ele- mental composition of the otoliths, as a measure of stock identity. Initial microchemistry studies tend- ed to focus on diadromous species and distinguished between freshwater and marine life history phases of fish (e.g. Bagenal et al., 1973; Belanger et al., 1987; Kalish, 1990; Coutant and Chen, 1993; Rieman et al., 1994). The elec- tron microprobe was widely used for such studies but because of the “typi- cal” concentrations of Ca, Na, Sr, K, S, and Cl found in otoliths, electron probes were effectively limited to detecting these six elements. Recently, there have been dramatic advances in analytical techniques and studies now indicate that other elements also exist in otoliths (e.g. Edmonds et al., 1989, 1991, 1992; Sie and Thresher, 1992), many of which are thought to be reflective of the envi- ronment (Fowler et al., 1995). Stock dis- crimination capabilities have improved with the sensitivity of recent instrumen- tation and the inductively coupled plas- ma-mass spectrometer (ICP-MS) has now become the instrument of choice for simultaneously quantifying the con- centration of multiple elements and iso- topes (Houk, 1986; Date, 1991). ICP-MS has recently been used for assays of various fish tissues including otoliths (e.g. Edmonds et al., 1991; Cam- pana and Gagne, 1995; Campana et al., 1994, 1995; Dove et al., 1996; Gillanders and Kingsford, 1996), scales (e.g. Cout- ant and Chen, 1993, Wells et al., 2000), eye lenses (e.g. Dove and Kingsford, 1998) and soft tissues (e.g. Beauchemin et al., 1988; Ishii et al., 1991; Hellou et al., 1992). Sample-specific differences in ele- mental concentrations in tissues other than otoliths have also been reported (e.g. vertebrae — Mulligan et al., 1983; Behrens Yamada et al., 1987; scales — Bagenal et al., 1973; Lapi and Mulli- gan, 1981; eye lenses — Dove and Kings- ford, 1998; soft tissue — Hellou et al., 1992). Such sample-specific differences in trace elements provide evidence for the nonmixing and segregation of pop- ulations. Otoliths are an ideal natural marker of elemental composition be- cause they grow throughout the life of the fish and once the elements are de- Gillanders: Trace metals in four structures of fish and their use for estimates of stock structure 411 Figure 1 Map showing sampling locations along the coast of New South Wales, Australia. posited, they are unlikely to be resorbed or altered (Campana and Neilson, 1985). Recently, Dove and Kingsford (1998) found that eye lenses may be suit- able for differentiating populations of fish because the eye lens has no efficient mechanism for remov- ing ions from the tissue. Eye lenses also grow by the addition of protein-rich cells to their outer sur- face. Bone, scales, and other soft tissues (e.g. gills, liver, muscles) will reflect a variety of factors includ- ing composition during growth, but resorption and remineralization may occur in some species (Gauldie and Nelson, 1990). The temporal stability of the el- emental concentrations in other tissues is therefore questionable (Campana and Gagne, 1995). In some species, however, the market value of the fish is like- ly to be reduced by removing otoliths and eye lenses; therefore, structures such as scales or spines may be easier to obtain. If scales and spines show similar el- emental fingerprints (for fish caught at different lo- cations) to otoliths and eye lenses, then their use in stock discrimination studies may be warranted. The use of scales and spines would also allow sampling without the need to kill the fish, with the result that broodstock could be kept alive and individuals from rare or endangered stocks could be sampled. With the exception of two studies (Dove and Kingsford, 1998; Wells et al., 2000), who compared elemental fingerprints of otoliths with eye lenses and scales, respectively, there have been no comparative studies of more than two structures. The aim of my study was to compare the elemental fingerprints obtained from different structures (oto- liths, scales, eye lenses, and dorsal spines) to determine if there were correlations among structures. Because otoliths are not subject to resorption (but see Mugiya and Uchimu- ra, 1989), their use in studies of stock discrimination ap- pears justified; therefore all comparisons were made be- tween otoliths and the other three structures. These four structures chosen for the present study are also very dif- ferent in terms of their composition. Otoliths are com- posed primarily of calcium carbonate, whereas scales and spines are composed primarily of calcium phosphate, and eye lenses, largely of water and structural proteins. There may, therefore, be differences in the affinities of ions for the different structures and this difference could have implica- tions for the elemental chemistry of each structure. The damselfish Parma microlepis is a territorial species, both juveniles (Moran and Sale, 1977) and adults (territo- ry sizes 2-17 m2, Tzioumis, 1995) remain attached to their natal sites. Therefore it is an excellent model species for this kind of research. Otoliths and eye lenses of P. microle- pis have been shown to contain microconstituents that can be detected with ICP-MS, some elements of which may be site-specific (Dove et al., 1996; Dove and Kingsford, 1998). Juvenile fish are distinguished from adults on the basis of coloration and are generally less than 3 years of age (Tzi- oumis and Kingsford, 1999). By comparison adults reach a maximum age of 37 years, and fish larger than 120 mm SL may represent a wide range of age classes (Tzioumis and Kingsford, 1999). Adults will therefore have longer and potentially more variable age- integrated elemental finger- prints than those of juvenile fish. Materials and methods Parma microlepis was collected from each of ten locations along the coast of New South Wales, Australia (Fig. 1). Fish were not collected from multiple sites within each location because small-scale differences in trace elements of otoliths and eye lenses (sites separated by 50 to 200 m) were not found in a previous study (Dove and Kingsford, 1998). All fish were collected by divers using SCUBA and a hand-spear between January and April 1998. Between five and seven replicate adult and juvenile fish were collected from each location. Juveniles and adults were distinguished from each other on the basis of differences in patterns of coloration. Fish were stored on ice during transportation to the lab- oratory. In the laboratory, the standard length of each fish was measured (Table 1). Otoliths, eye lenses, and approx- imately five non-regenerated scales were removed from each fish and rinsed in MilliQ water. Otoliths and scales were stored dry at room temperature, and eye lenses were stored at -70°C in Eppendorf microcentrifuge tubes in preparation for ion analysis. Dorsal spines were removed and frozen (-4°C) prior to removal of surrounding flesh. After flesh was removed, spines were also stored dry at room temperature. 412 Fishery Bulletin 99(3) Table 1 Summary of collections of Parma microlepis. Sample locations are shown in Figure 1. Sample sizes were n= 5 for juveniles and adults at all locations, except for adult fish at Henry Head where n= 7. Lengths refer to standard length in mm. Sample no. Location Date of collection Juvenile mean length (range) Adult mean length (range) 1 Norah Head Mar 1998 72.0 (56-89) 125.6 (117-133) 2 Terrigal Mar 1998 83.0(61-95) 122.6 (105-132) 3 Barrenjoey Head Feb 1998 64.0 (49-89) 118.8(106-129) 4 Middle Head Feb 1998 88.6 (83-94) 125.6(115-136) 5 Henry Head Jan 1998 83.6 (71-94) 122.6 (114-136) 6 Sutherland Point Jan 1998 85.4(76-93) 118.6(107-125) 7 Jibbon Head Mar 1998 83.2 (65-95) 126.4(114-136) 8 Flinders Islet Mar 1998 76.0 (73-80) 110.4(105-124) 9 Jervis Bay Apr 1998 81.2 (75-95) 117.8 (112-125) 10 Ulladulla Apr 1998 80.8 (76-90) 111.2(102-120) All samples and standards were prepared for ion analy- sis in a laminar flow cabinet. Otoliths and eye lenses, and scales and spines were individually weighed on an ana- lytical balance (to 0.0001 g) and a microbalance (to 0.001 mg), respectively. All samples were rinsed in 1% nitric ac- id for 15 s prior to dissolution. Structures were digested in nitric acid (Aristar) either overnight (otoliths and scales) or for 36 h (eye lenses and spines). For eye lenses, sul- phuric acid (Aristar) was added following dissolution and each sample was heated to ~90°C for 1 hour. For all struc- tures varying amounts of MilliQ water were then added followed by additional dilution with 1% nitric acid where necessary. This was necessary so that concentrations of the sample solutions were standardized before analysis with ICP-MS. Samples were analyzed by solution-based ICP-MS (Per- kin Elmer SCIEX ELAN 5000). Initially, the four struc- tures of fish collected from one location (Henry Head, Fig. 1) were analyzed in “totalquant II” mode {n= 7 fish). This procedure provided a rapid-survey-analysis to semiquanti- tatively determine the elements present (and their range) within each of the four structures. The instrument was calibrated at six points spanning the mass range from 9 to 209. Accuracy and drift of the machine were determined from spiked samples. Samples were then analyzed in “quantitative analysis” mode after the instrument was calibrated for either exter- nal standardization or addition calibration, depending on the structure and expected concentration range of each el- ement. Measurement parameters for the ICP-MS were as detailed in Gillanders and Kingsford (1996). Seven blank solutions were run at the beginning of each session to de- termine detection limits, which were calculated from the concentration of analyte, yielding a signal equivalent to three times the standard deviation of the blank signal. Blank solutions were 1% HN03 for otoliths, scales, and spines, and 2% acid (1%HN03 and 1%H2S04) for eye lens- es. Standards were run through the machine at the begin- ning of each session. Spiked samples were run every ten samples to measure sensitivity changes to the machine through use and for recovery verification. All samples were initially analyzed by using addition calibration mode. The data were then reprocessed in exter- nal calibration mode to obtain concentrations of microele- ments. Each sample took around 3 to 4 min to analyze, in- cluding a delay before starting to read a sample (70 s) and the rinse ( 1% HN03, 0.1% Triton X) between samples of 60 s. Samples from each structure (otoliths, scales, spines, and eye lenses) and size class (juveniles and adults) were ana- lyzed together in blocks, but all samples within each block were randomized to ensure that resulting patterns did not reflect variation in the performance of the instrument. Univariate and multivariate techniques were used to test hypotheses concerning individual elements and multi- element fingerprints of Parma microlepis. Pearson’s cor- relations were used to determine the relation between otoliths and each of the other structures for individual ele- ments (Mn, Sr, Ba, and Pb). Separate analyses were per- formed for each size class and structure. The match between the otolith microchemistry data and the microchemistry data for each of the other three struc- tures was examined by using two types of nonparametric multivariate analysis. For both types of analysis, the data were standardized by subtracting the mean and by divid- ing by the standard deviation of the variable, so that each element would have equal weight. A matrix of Euclidean distances among all pairs of samples was then calculated. Mantel’s test (from the R package) and the BIOENV pro- cedure (included in the PRIMER computer program; copy- right M.R. Carr and K.R. Clarke, Marine Biological Labo- ratory, Plymouth, UK) were used to determine the degree of correlation between pairs of distance matrices (otoliths and each of the other three structures). Mantel’s test describes the relation between two dis- tance matrices with Pearson’s correlation coefficient, r (Mantel, 1967; Legendre and Fortin, 1989). A test of sig- nificance for r is obtained by random permutations of the replicate sample units for one of the matrices. A total of Gillanders: Trace metals in four structures of fish and their use for estimates of stock structure 413 999 permutations were performed for each test and the pi-obability calculated as the number of values equal to, or larger than, the observed value of r divided by the total number of permutations. If significant relationships were found with Mantel’s test, the BIOENV procedure was then used to determine which elements were likely to be important in describing the correlation between the distance matrices (Clarke and Ainsworth, 1993; Clarke and Warwick, 1994). BIOENV calculates rank correlations between a dissimilarity ma- trix derived from the otolith microchemistry data and ma- trices derived from various subsets of the elements for ei- ther the scale, spine, or eye lens data matrices, thereby defining suites of elements that “best explain” the otolith microchemistry data (Clarke and Ainsworth, 1993). The harmonic or weighted Spearman correlation pw was used. Results Otoliths, scales, spines, and eye lenses of Pcinna microl- epis differed in composition, both in terms of the actual elements present and the concentration of individual ele- ments. Otoliths, scales, and spines were dominated by calcium, whereas no calcium was detected in eye lenses. The microelement Sr was present in otoliths, scales, and spines at concentrations of 100’s to 1000’s pg/g of struc- ture, whereas it occurred at low concentrations in eye lenses. Other elements (e.g. Mn, Ba, Pb) were typically found in low-to-trace levels in all structures. Mercury was found in detectable concentrations only in eye lenses. Univariate analyses Differences in the concentration of manganese among structures varied over several orders of magnitude (range: 0.05 for eye lenses to 100 pg Mn/g structure for spines). Concentrations of Mn in otoliths showed significant corre- lations with concentrations of Mn in both eye lenses and scales of juvenile fish, but only with scales from adult fish (Fig. 2, Table 2). For juveniles, fish from one site may have influenced the correlation analyses; therefore fish from this site were removed and the correlations recalcu- lated. This adjustment resulted in a significant correla- tion between Mn in otoliths and Mn in scales, as found previously. The relation between Mn in otoliths and Mn in spines was also significant when fish from this one site were removed (r=0.302, P<0.05). However, there was no longer a significant correlation for eye lenses: fish from one site may therefore have influenced the correlation result for this structure. Concentrations of Sr in otoliths showed significant cor- relations with concentrations of Sr in all the other struc- tures for juvenile fish; however the relation was only posi- tive for otoliths and spines (Fig. 3, Table 2). For scales and spines, fish from one site did not unduly influence correla- tions; when these fish were removed from analyses, cor- relations were still significant. However, as found for Mn, the correlation between Sr in otoliths and Sr in eye lenses was no longer significant when fish from this site were re- Table 2 Correlations between concentration of elements in otoliths and concentration of elements in eye lenses, scales, and spines of juvenile and adult fish collected from ten loca- tions along the coast of New South Wales. Shown are Pearson’s correlation coefficients. Significance levels are indicated by * (P<0.05) and ** (P<0.01), df=48. No correc- tions were made for experiment-wise error rate; therefore 1 in 20 tests would be expected to be significant by chance alone. Otolith Mn Sr Ba Pb Juveniles Eye lenses 0.3410* -0.3787** 0.1260 Scales 0.4018** -0.6237** 0.8810** Spines 0.2466 0.6586** 0.9001** Adults Eye lenses 0.0429 0.2353 0.0239 -0.2564 Scales 0.4717** -0.0051 0.8839** -0.1487 Spines 0.2696 0.8446** 0.9121** -0.0732 moved. The only significant relationship for Sr in adult fish was found between otoliths and spines (Fig. 3, Table 2). Significant positive relationships were found between the amount of Ba in otoliths and the amount of Ba in scales and spines of both juvenile and adult fish (Fig. 4, Table 2). Lead was usually at or below detection limits of the instrument for juvenile fish and therefore correlations were not made. Correlations between the amount of Pb in otoliths of adult fish and the other structures showed no significant relationship (Fig. 5, Table 2). Multivariate analyses Mantel’s test detected a significant relationship between the Euclidean distances among replicates based on the otolith data and the distances based on data from eye lenses, scales, or spines for juvenile fish (Table 3). For adult fish, however, a significant relation was detected between the otolith data and the scale and spine data (Table 3); thus there may be either changes in assimilation of elements with age or resorption and remineralization may occur in some structures, for example in eye lenses. The BIOENV analyses showed that the highest rank correlation was found between the otolith data and the scale data for juvenile fish and involved Sr and Ba (Table 3). For adult fish the highest correlation was between the otolith data and the spine data and involved three ele- ments (Mn, Sr, and Ba; Table 3). Comparisons of juvenile and adult fish for each struc- ture showed no significant relationships (Table 3). The Mantel test, however, was marginally nonsignificant for otoliths (Mantel r statistic 0.2000, P=0.060). 414 Fishery Bulletin 99(3) 0 080 0 065 0.050 0 035 0.020 Juveniles I 0 080 0 065 0.050 0 035 0.020 Adults Eye lenses 30 25 20 15 10 r r 60 50 40 30 20 10 pg Mn / g otolith Figure 2 Relation between concentration of manganese in otoliths (mean ±SE) and in eye lenses, scales, and spines of juvenile and adult fish (mean ±SE) collected from ten locations along the coast of New South Wales. Discussion Otolith elemental fingerprints are reported to be among the most important tools for stock discrimination for some species offish (Campana et al., 1995). In the current study, adult fish showed a significant relationship between the otolith elemental fingerprints and the scale and spine elemental fingerprints. In contrast, juvenile fish showed significant relationships between comparisons of otoliths and each of scales, spines, and eye lenses. Relationships among some structures were also seen for some individ- ual elements. There have been two other studies involving comparisons between elemental fingerprints of otoliths and other structures (Dove and Kingsford, 1998; Wells et al., 2000). Wells et al. (2000) compared trace elements in otoliths and scales of juvenile weakfish by using correla- tion analyses on individual elements. They found similar results to those in the current study, namely a significant correlation between Mn, Sr, or Ba in otoliths and that in scales. Dove and Kingsford (1998) compared elemental fingerprints of otoliths and eye lenses. In their study dif- ferences between locations were compared by using anal- ysis of similarity (ANOSIM) permutation tests. Where comparisons between locations gave the same response Gillanders: Trace metals in four structures of fish and their use for estimates of stock structure 415 0.25 0 22 0.19 0.16 0.13 Juveniles -J I I L_ i.5 r io - 0 5 - 0.0 Adults Eye lenses {hJ.' >-f- 0.10 L 1,500 1,700 1,900 2,100 2,300 2,500 2,400 2,800 3,200 3,600 3,500 r “ 2,500 CO cn 1,500 500 1,500 r 1,300 - 1,100 ■i 4-ji j -I I I I I 900 Scales 1,500 1,700 1,900 2,100 2,300 2,500 2,400 2,800 3,200 3,600 900 r 1,400 r Spines 800 700 600 1,200 1,000 800 1,500 1,700 1,900 2,100 2,300 2,500 2,400 2,800 3,200 3,600 pg Sr / g otolith Figure 3 Relation between concentration of strontium in otoliths (mean ±SE) and in eye lenses, scales, and spines of juvenile and adult fish (mean ±SE) collected from ten locations along the coast of New South Wales. for both otoliths and eye lenses (either both significant or both nonsignificant), then the two structures were viewed as being similar. With this approach, approximately 73% (11/15) of pair-wise comparisons showed agreement (Dove and Kingsford, 1998). Their approach is, however, indirect and does not show evidence of correlation between data matrices for otoliths and those for eye lenses. I am aware of no other studies that have compared multiple struc- tures for fish obtained from a number of locations. Otoliths have been considered the structure of choice for determining elemental fingerprints and for relating these to movements or stock structure of fish because otoliths grow throughout the life of the fish and the material of the annual growth increment, once deposited, is unlikely to be resorbed or altered (Campana and Neilson, 1985). Otoliths therefore have the potential to record endogenous and ex- ogenous factors permanently within their calcium-protein matrix. Other structures may also grow throughout the life of the fish (e.g. eye lenses — Dove, 1997; scales — Mitani, 1955; spines — Hill et al., 1989). Although other structures have been used for chemical analyses and for aging fish, some structures are susceptible to resorption and remin- 416 Fishery Bulletin 99(3) Table 3 Summary of results from Mantel’s test and the BIOENV procedure where otolith microchemistry data were compared with the microchemistry data obtained from eye lenses, scales, and spines. Comparisons are also made between juvenile and adult fish for each structure. The BIOENV procedure (see general text) was done only if significant correlations were found with Mantel’s test. Only the combination of elements that contributed to the maximum p overall, as measured by the weighted Spearman correlation, are shown for each structure comparison. The significance of the Mantel test statistic is shown by * (P< 0.05) or ** (PcO.Ol). Comparison Mantel r statistic BIOENV-maximum p Combination of elements contributing to maximum p Juveniles Otoliths and eye lenses 0.3402* 0.127 Mn, Ba, Hg Otoliths and scales 0.5757** 0.587 Sr, Ba Otoliths and spines 0.4097** 0.403 Sr, Ba Adults Otoliths and eye lenses -0.0646 Otoliths and scales 0.3680** 0.454 Mn, Ba Otoliths and spines 0.4440** 0.512 Mn, Sr, Ba Adults and juveniles Otoliths 0.2000 Eye lenses -0.1357 Scales -0.0136 Spines 0.0676 eralisation (Simkiss, 1974). Despite the possibility of re- sorption of some elements, otoliths and scales, and otoliths and spines showed a significant correspondence between data matrices for both juvenile and adult fish, which may indicate that scales and spines provide estimates of stock structure that are similar to those obtained from otoliths. Differences in elemental fingerprints among structures may be due to different l'outes of ion uptake and differen- tial abilities of each structure to incorporate elements into the organic and inorganic matrix. Calcium and strontium in otoliths are primarily taken up by the gills (Simkiss, 1974), but the route of uptake of trace elements has not been identified (Campana and Gagne, 1995). There may also be some Sr uptake in otoliths through the diet be- cause fish that were fed a Sr-enriched diet showed a de- tectable increase in Sr/Ca ratios (Gallahar and Kingsford, 1996). Other structures, such as scales and bone, may in- corporate ions by diffusion across the gills and through the skin or by ingestion of food and water (Simkiss, 1974). Whether some structures show differential abilities to in- corporate ions through different methods of absorption and whether ions are resorbed from different structures in equal proportions is not known. Otoliths and eye lenses showed the greatest differences in elemental fingerprints between structures. Otoliths are predominantly CaC03 (Ca constitutes from 30% to 39% of otoliths. Thresher et al., 1994; Dove et al., 1996), although small amounts of protein also occur (Degens et ah, 1969) and therefore otoliths are likely to incorporate ions that are able to substitute for Ca in the CaC03 matrix or bind to proteins in the organic matrix (Gunn et ah, 1992; Sie and Thresher, 1992). The spaces between these matrices may also trap ions. In contrast, eye lenses are composed largely of water and structural protein, the latter of which may be soluble or insoluble crystallin (Nicol, 1989). The proteins of eye lenses are rich in sulfhydryl groups that may covalently bind with metals and the proteins also have specific sites for binding with cations (Sharma et ah, 1989). Scales and bone are more similar to otoliths in that the dominant ion is Ca (e.g. in bone Ca may constitutes from 24% to 37%; Hamada et ah, 1995), but they vary considerably in that otoliths are pri- marily carbonate structures, whereas scales and bone are primarily phosphate (or hydroxyapatite) structures. Some differences in elemental composition of otoliths and scales or spines may therefore be expected. Because otoliths, scales, and bone are composed predominantly of a mineral, rather than an organic matrix, it is not surprising that these struc- tures were similar in their elemental fingerprints and that eye lenses presented a different fingerprint. Differences in elemental fingerprints between juvenile and adult fish were also found for all structures. Eye lenses are thought to have no efficient mechanism for removing ions, but there may be changes in structural proteins with age that possibly alter affinities of proteins for specific ions (Dove, 1997). Such ontogenetic effects may result in differ- ent elemental fingerprints between juvenile and adult fish for eye lenses. Both scales and spines show some evidence of resorption or remineralization over time in some species. For example, early growth increments in spines of several species are known to be destroyed as the core-matrix ex- pands (Hill et al., 1989; Gillanders et al., 1997) and there is evidence for resorption in scales of fish living in a stressed environment (Sauer and Watabe, 1989). If an alternative structure to otoliths is required for stock identification, as it may be for broodstock, and for rare or endangered species, or if removing otoliths decreases Gillanders: Trace metals in four structures of fish and their use for estimates of stock structure 417 Juveniles Adults 0 25 0.20 - 0.16 - 0.11 0.07 0.02 3 - CD CD J I I I I 0 07 0 06 0 05 0.03 0 02 Eye lenses + _i i I 1 I H t- 2 " 12 3 4 Scales 4- J L_ 7 " 6 - 9 Spines 6 - 5' fH 6 | 4 - 2- C 1 - n 1 l i 3 1 1 rt 2 - U 0 12 3 0 4 5 0 12 3 4 o L |ig Ba / g otolith Figure 4 Relation between concentration of barium in otoliths (mean ±SE) and in eye lenses, scales, and spines of juvenile and adult fish (mean ±SE) collected from ten locations along the coast of New South Wales. the market value of the fish, then it is recommended that either scales or spines are the best alternative structure because these were the structures that were significantly correlated with the otolith elemental data for both juve- nile and adult fish. However, before using either of these structures, a number of fish (e.g. 30-50) should be collect- ed from three to four locations and the alternative struc- ture, as well as otoliths, should be analyzed. Classification models should then be developed by using data from each structure and the error rates should be compared. This was beyond the scope of the present study because sample sizes from each location were relatively small. 0.060 r 0.045 0030 0.015 0 000 .o CL 2 0 1.5 1.0 0.5 0.0 J I L. 0.0 0 1 02 0 3 04 0 5 Scales -4- J I I L. 0 0 0.1 0.2 0.3 0 4 0.5 Spines 0.0 0.1 0.2 0.3 0.4 0.5 pg Pb / g otolith Figure 5 Relationship between concentration of lead in otoliths (mean ±SE) and in eye lenses, scales, and spines of adult fish ( mean ±SE ) collected from ten locations along the coast of New South Wales. Juveniles are not shown because concentrations of lead were below detection limits. Scales and spines may offer certain advantages over otoliths because they provide a nonlethal alternative. In addition, they can be collected relatively quickly and eas- ily. Scales may also require less preparation for analyses than either spines or otoliths. Thus, scales and spines may provide an alternative structure for determining stock identity. 418 Fishery Bulletin 99(3) Acknowledgments I thank Belinda Curley and Ben Stewart for their assis- tance with collection of fish, Marti Jane Anderson for discussions regarding statistical analyses, and Robert McQuilty (RPA hospital) for use of the ICP-MS. I also thank Marti Jane Anderson and Sophie Dove for useful comments on the manuscript. The study was conducted while BMG held an ARC postdoctoral fellowship and was funded by an ARC grant and research funding from the University of Sydney. Literature cited 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. Beauchemin, D., J. W. McLaren, S. N. Willie, and S. S. Berman. 1988. Determination of trace metals in marine biological reference materials by inductively coupled plasma mass spectrometry. Anal. Chem. 60:687-691. Behrens Yamada, S., T. J. Mulligan, and D. Fournier. 1987. Role of environment and stock on the elemental com- position of sockeye salmon ( Oncorhynchus nerka) verte- brae. Can. J. Fish. Aquat. Sci. 44:1206-1212. Belanger, S. E., D. S. Cherry, J. J. Ney, and D. K. Whitehurst. 1987. Differentiation of freshwater versus saltwater striped bass by elemental scale analysis. Trans. Am. Fish. Soc. 116:594-600. Campana, S. E., A. J. Fowler, and C. M. Jones. 1994. Otolith elemental fingerprinting for stock discrimina- tion of Atlantic cod ( Gadus morhua ) using laser ablation ICPMS. Can. J. Fish. Aquat. Sci. 51:1942-1950. Campana, S. E., and J. A. Gagne. 1995. Cod stock discrimination using ICPMS elemental assays of otoliths. In Recent developments in fish otolith research (D. H. Secor, J. M. Dean, and S. E. Campana, eds.), p. 671-691. Univ. South Carolina Press, Columbia, SC. Campana, S. E., J. A. Gagne, and J. W. McLaren. 1995. Elemental fingerprinting of fish otoliths using ID- ICPMS. Mar. Ecol. Prog. Ser. 122:115-120. Campana, S. E., and J. D. Neilson. 1985. Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. Clarke, K. R., and M. Ainsworth. 1993. A method of linking multivariate community struc- ture to environmental variables. Mar. Ecol. Prog. Ser. 92: 205-219. Clarke, K.R., and R.M. Warwick. 1994. Change in marine communities: an approach to statis- tical analysis and interpretation. Natural Environment Research Council, UK, 144 p. Coutant, C. C., and C. H. Chen. 1993. Strontium microstructure in scales of freshwater and estuarine striped bass ( Morone saxatilis ) detected by laser ablation mass spectrometry. Can. J. Fish. Aquat. Sci. 50: 1318-1323. Date, A. R. 1991. Inductively coupled plasma mass spectrometry. Spec- trochimica Acta Review 14:3-32. Degens, E. T., W. B. Deuser, and R. L. Haedrich. 1969. Molecular structure and composition of fish otoliths. Mar. Biol. 2:105-113. Dove, S. G. 1997. The incorporation of trace metals into the eye lenses and otoliths of fish. Ph.D. diss., Univ. Sydney, Sydney, 201 p. Dove, S. G., B. M. Gillanders, and M. J. Kingsford. 1996. An investigation of chronological differences in the deposition of trace metals in the otoliths of two temperate reef fishes. J. Exp. Mar. Biol. Ecol. 205:15-33. Dove, S. G., and M. J. Kingsford. 1998. Use of otoliths and eye lenses for measuring trace- metal incorporation in fishes: a biogeographic study. Mar. Biol. 130:377-387. Edmonds, J. S., N. Caputi, and M. Morita. 1991. Stock discrimination by trace-element analysis of oto- liths of orange roughy (Hoplostethus atlanticus), a deep- water marine teleost. Aust. J. Mar. Freshwater Res. 42: 383-389. Edmonds, J. S., R. C. J. Lenanton, N. Caputi, and M. Morita. 1992. Trace elements in the otoliths of yellow-eye mullet ( Aldrichetta forsteri) as an aid to stock identification. Fish. Res. 13:39-51. Edmonds, J. S., M. J. Moran, and N. Caputi. 1989. Trace element analysis of fish sagittae as an aid to stock identification: pink snapper ( Chrysophrys auratus ) in Western Australian waters. Can. J. Fish. Aquat. Sci. 46:50-54. Fowler, A. J., S. E. Campana, C. M. Jones, and S. R. Thorrold. 1995. Experimental assessment of the effect of tempera- ture and salinity on elemental composition of otoliths using laser ablation ICPMS. Can. J. Fish. Aquat. Sci. 52: 1431-1441. Gallahar, N. K., and M. J. Kingsford. 1996. Factors influencing Sr/Ca ratios in otoliths of Girella elevata: an experimental investigation. J. Fish Biol. 48: 174-186. Gauldie, R. W., and D. G. A. Nelson. 1990. Otolith growth in fishes. Comp. Biochem. Physiol. 97A:119-135. Gillanders, B. M., D. J. Ferrell, and N. L. Andrew. 1997. Determination of ageing in kingfish (Seriola lalandi) in New South Wales. Fisheries Research Development Corporation (ISBN 0 7310 9406 9), 103 p. Gillanders, B. M., and M. J. Kingsford. 1996. Elements in otoliths may elucidate the contribution of estuarine recruitment to sustaining coastal reef pop- ulations of a temperate reef fish. Mar. Ecol. Prog. Ser. 141:13-20. Gunn, J. S., I. R. Harrowfield, C. H. Proctor, and R. E. Thresher. 1992. Electron probe microanalysis of fish otoliths — evalu- ation of techniques for studying age and stock discrimina- tion. J. Exp. Mar. Biol. Ecol. 158:1-36. Hamada, M., T. Nagai, N. Kai, Y. Tanoue, H. Mae, M. Hashimoto, K. Miyoshi, H. Kumagai, and K. Saeki. 1995. Inorganic constituents of bone of fish. Fish. Sci. 61: 517-520. Hartl, D. L., and A. G. Clark.. 1989. Principles of population genetics. Sinauer Associ- ates Inc, Sunderland, MA, 682 p. Hellou, J., W. G. Warren, J. F. Payne, S. Belkhode, and P. Lobel. 1992. Heavy metals and other elements in three tissues of cod, Gadus morhua from the Northwest Atlantic. Mar. Poll. Bull. 24:452-458. Hill, K. T., G. M. Cailliet, and R. L. Radtke. 1989. A comparative analysis of growth zones in four calci- fied structures of Pacific blue marlin, Makaira nigricans. Fish. Bull. 87:829-843. Gillanders: Trace metals in four structures of fish and their use for estimates of stock structure 419 Houk, R. S. 1986. Mass spectrometry of inductively coupled plasmas. Anal. Chem. 58:97A-105A. Ihssen, P. E., H. E. Booke, J. M. Casselman, J. M. McGlade, N. R. Payne, and F. M. Utter. 1981. Stock identification: materials and methods. Can. J. Fish. Aquat. Sci. 38:1838-1855. Ishii, T., M. Nakahara, M. Matsuba, and M. Ishikawa. 1991 . Determination of 238U in marine organisms by induc- tively coupled plasma mass spectrometry. Nippon Suisan Gakkaishi 57:779-787. Kalish, J. M. 1990. Use of otolith microchemistry to distinguish the prog- eny of sympatric anadromous and non-anadromous salmo- nids. Fish. Bull. 88:657-666. Lapi, L. A., and T. J. Mulligan. 1981. Salmon stock identification using a microanalytic tech- nique to measure elements present in the freshwater growth region of scales. Can. J. Fish. Aquat. Sci. 38:744-751. Legendre, P, and M. Fortin. 1989. Spatial pattern and ecological pattern. Vegetatio 80: 107-138. Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Res. 27:209-220. Mitani, F. 1955. Studies on the fishing conditions of some useful fishes in the western region of Wakasa Bay — 11. Relation between the scale of the yellowtail, Seriola quinqueradiata T. & S., and fork length and age. Bull. Jap. Soc. Scient. Fish. 21:463^166. Moran, M. J., and P. F. Sale. 1977. Seasonal variation in territorial response, and other aspects of the ecology of the Australian temperate poma- centrid fish Parma microlepis. Mar. Biol. 39:121-128. Mugiya, Y., and T. Uchimura. 1989. Otolith resorption induced by anaerobic stress in the goldfish, Carassius auratus. J. Fish Biol. 35:813-818. Mulligan, T. J., L. Lapi, R. Kieser, S. B. Yamada, and D. L. Duewer. 1983. Salmon stock identification based on elemental compo- sition of vertebrae. Can. J. Fish. Aquat. Sci. 40:215-229. Nicol, J. A. C. 1989. The eyes of fishes. Oxford Univ. Press, New York, NY, 308 p. Rieman, B. E., D. L. Myers, and R. L. Nielson. 1994. Use of otolith microchemistry to discriminate Oncor- hynchus nerka of resident and anadromous origin. Can. J. Fish. Aquat. Sci. 51:68-77. Sauer, G. R., and N. Watabe. 1989. Temporal and metal-specific patterns in the accumu- lation of heavy metals by the scales of Fundulus heterocli- tus. Aquat. Toxicol. 14: 233-248. Sharma, Y., C. M. Rao, M. L. Narasu, S. C. Rao, T. Somasundaram, A. Gopalakrishna, and D. Balasubramanian. 1989. Calcium binding tod- and /3-crystallins. J. Biol. Chem. 264:12794-12799. Sie, S. H., and R. E. Thresher. 1992. Micro-PIXE analysis of fish otoliths: methodology and evaluation of first results for stock discrimination. Int. J. PIXE 2:357-379. Simkiss, K. 1974. Calcium metabolism of fish in relation to ageing. In The ageing of fish (T. B. Bagenal, ed.), p. 1-12. Unwin Brothers Ltd, London. Thresher, R. E., C. H. Proctor, J. S. Gunn, and I. R. Harrowfield. 1994. An evaluation of electron-probe microanalysis of oto- liths for stock delineation and identification of nursery areas in a southern temperate groundfish, Nemadactylus macropterus (Cheilodactylidae). Fish. Bull. 92:817-840. Tzioumis, V. 1995. Aspects of the reproductive biology and behaviour of the temperate damselfish Parma microlepis Gunther (Pisces: Pomacentridae). Ph.D. diss., Univ. Sydney, Sydney, 172 p. Tzioumis, V., and M. J. Kingsford. 1999. Reproductive biology and growth of the temperate damselfish Parma microlepis. Copeia. 1999:348-361. Wells, B. K., S. R. Thorrold, and C. M. Jones. 2000. Geographic variation in trace element composition of juvenile weakfish scales. Trans. Am. Fish. Soc.. 129: 889- 900. 420 Abstract— Aerial photographic assess- ment is a promising technique that could be structured to yield a fishery- independent index of abundance for Atlantic bluefin tuna, Thunnus thyn- nus thynnus ( ABTl.The accuracy of this approach may be increased by incor- porating the relationship between the surface characteristics of a school and the total number of individuals. Our objective was to develop models to facil- itate the estimat ion of number of fish in ABT schools from aerial photographs. Video cameras were used to observe 74 incidences of schooling for 50 cap- tive ABT approximately one meter in length. Relationships between the sur- face characteristics of ABT schools and the number of fish in the school were explored by using least-squares regres- sion. The schools ranged in number from 2 to 45 individuals. A weighted regres- sion model incorporating the number of fish in the school at the surface as the independent variable and the number of fish in the remaining por- tion of the school yielded an r 2 of 0.74. A second weighted multiple-regression model incorporating the number of fish in the school at the surface and in the second depth interval (0-25% school depth below surface layer) of the school as independent variables, and the number of fish in the remaining portion of the school as the dependent variable, with 1/variance as the weight, achieved an r2 of 0.70. A third model using the length and width of the sur- face layer of the school as the indepen- dent variables and the number of fish in the school as the dependent vari- able had an r2 of 0.86. One data point from a wild school is currently avail- able to verify model predictions. This school of 125 individuals is well outside the range of school sizes used to con- struct the model (2-45 individuals), yet differs from model predictions by only 7%. We believe that these models have the potential to improve an abundance index based on aerial photographs by estimat- ing the number of individuals in wild ABT schools from surface characteris- tics observed in aerial photographs. Manuscript accepted 26 January 2001. Fish. Bull. 99:420-431 (2001). Estimating the number of fish in Atlantic bluefin tuna ( Thunnus thynnus thynnus) schools using models derived from captive school observations Brian Hanrahan Francis Juanes Department of Natural Resources Conseivation University of Massachusetts Amherst, Massachusetts 01003-4210 E-mail address (for F. Juanes, contact author): |uanes@forwild. umass.edu The bluefin tuna ( Thunnus thynnus ) is distributed worldwide in temperate and subtropical seas. It has a limited dis- tribution in the southern hemisphere. Endothermy by means of vascular heat exchangers allows bluefin tuna to inhabit a wide thermal niche and therefore wide geographic and depth ranges (Carey and Teal, 1969; Carey and Lawson, 1973). In the western Atlantic Ocean, the Atlantic bluefin tuna (ABT), Thunnus thynnus thynnus, is distributed from Labrador to Brazil, including the Gulf of Mexico and Caribbean Sea. Adult ABT occur throughout the entire range, but smaller bluefin tuna (less than 45 kg) are not observed frequently above the latitude of Cape Cod, Massachusetts. The Atlan- tic bluefin tuna is epipelagic and usu- ally oceanic but appears near the coast seasonally (Squire, 1962; Collette and Nauen, 1983) to feed on concentrated assemblages of prey. Adult ABT may attain a length of four meters and a body mass of 680 kg. Large medium (178-195 cm FL) and giant (>195 cm FL) bluefin tuna are targeted by com- mercial purse-seine, long line, and hook- and-line fisheries (Mather, 1974; Figley, 1984). A recreational hook-and-line fish- ery (Mather, 1974; Figley, 1984) targets all sizes of bluefin tuna as they appear along the east coast of the United States and Canada from June to Octo- ber (Mather, 1962). The combination of changes in the spatial distribution over time and as- sociated uncertainty regarding the in- dependence of eastern and western At- lantic stocks makes the estimation of ABT stock size particularly problemat- ic. The stock assessment for this species has been based upon landings data and abundance indices (Scott et al., 1993). Use of ABT landings data to generate an abundance index may lead to bias due to variability in effort, improve- ments in fishing technology (Lo et al., 1992), and variability in annual geographic distribution linked to prey distribution.1 These characteristics of the northwest Atlantic bluefin tuna fishery, in conjunction with popula- tion-level behavioral characteristics ob- served for similar tuna species, suggest that the use of catch-per-unit-of-effort (CPUE) data to evaluate tuna popula- tion trends could lead to inaccurate es- timates (Clark and Mangel, 1979). The accuracy of CPUE-based assessments in estimating the abundance of bluefin tuna in the northwest Atlantic remains controversial (Clay, 1991; Suzuki and Ishizuka, 1991; Safina, 1993). An exten- sive discussion of the issues involved in Atlantic bluefin tuna assessment can be found in the National Research Council report by Magnuson et al. ( 1994). Recent investigations have focused on the feasibility of using aerial pho- tographic assessment of large medium and giant bluefin tuna in New England waters and in the Straits of Florida (Lutcavage and Kraus, 1995; Lutcav- age et al., 1997) as an alternative fish- ery-independent method of obtaining indices of abundance. The aerial survey 'Chase, B. C. 1995. Preliminary report of the Massachusetts bluefin tuna investiga- tion: the diet of bluefin (Thunnus thynnus ) off the coast of Massachusetts. Massa- chusetts Division of Marine Fisheries. Salem, MA 01947, 39 p. Hanrahan and Juanes: Estimating the school size of Thunnus thynnus thynnus 421 method has been used to determine the relative abun- dance for other pelagic fisheries worldwide, including En- graulis mordax (Lo et al., 1992), Engraulis mordax, Sar- da chiliensis, Trachurus symmetricus, etc. (Squire, 1972), Trachurus decliuis, Katsuwonus pelamis, Arripis trutta, Thunnus maccoyii (Williams, 1981), Mugil spp. (Scott et ah, 1989), and Squire (1993) has reported aerial survey data for Thunnus thynnus oi'ientalis and other species. Abundance estimates derived from an aerial assessment are based on biomass or number of individuals per unit of area. Lutcavage and Kraus (1995) concluded that the aerial method could provide area-specific minimum abundance and distribution data for large medium and giant Atlan- tic bluefin tuna under good viewing conditions. However, many difficulties associated with aerial photographic as- sessment of ABT remain to be resolved. Sea state, light- ing conditions, and turbidity all play an important role in the ability to detect and produce useful photographs of schools (Lutcavage and Kraus, 1995). Rough seas, sun glare, and high turbidity may all result in reduced detec- tion of schools, limiting the days on which this type of sur- vey method is effective. Visual counts of individuals at the surface derived from aerial photographs are difficult to in- terpret without a verification count and information on behavioral factors such as surfacing frequency (Lo et al., 1992) and the proportion of the school visible at the sur- face (Lutcavage and Kraus, 1995). In addition, variability in population movements and distribution could lead to an inaccurate abundance estimate if an intensive, spatially expansive sampling scheme is not employed. We propose a technique to address the problem of es- timating the number of fish in a school (NFS) from the surface characteristics of a school. If the relationship be- tween the surface structure or the surface number of fish and number fish in total school was known, school sur- face counts from aerial photographs or visual observa- tions could be adjusted to include an estimate of total NFS, facilitating an improvement of area-specific mini- mum abundance estimates based on visual or photograph- ic data sources. Atlantic bluefin tuna are believed to exhibit the most rig- idly defined spatial structure of schooling fishes (Partridge et al., 1983). Distinct two- and three-dimensional school structures have been described by previous authors (Par- tridge et al., 1983; Lutcavage and Kraus, 1995). Parabolas and echelons are the shapes of commonly observed sur- face-oriented two-dimensional schools, whereas the densely packed dome is the shape of a frequently observed three- dimensional school configuration (see Partridge et al., 1983 and Lutcavage and Kraus, 1995 for illustrations). The num- ber of fish observed in two-dimensional surface schools is generally less than 15, whereas three-dimensional schools such as those forming densely packed domes usually have greater than 15 individuals (Partridge et al., 1983). Al- though the three-dimensional component of bluefin tuna school structure has been observed (Lutcavage and Kraus, 1995), quantitative description and analysis is lacking and little is known of the relationship between the two-dimen- sional surface structure and three-dimensional structure (e.g. total count, biomass) of schools (Partridge et al., 1983; Lutcavage and Kraus, 1995). In addition, the behavioral and environmental factors that may influence tuna school structure and dynamics remain poorly described (Mather, 1962; Clark and Mangel, 1979; Partridge et al., 1983). Our study presents a functional relationship between the surface characteristics of and the total number of indi- viduals in ABT schools. We analyzed video-taped footage of 74 incidences of schooling in a group of captive ABT to quantify the relationship between the number of fish vis- ible at the surface and the total number of individuals in the school (NFS), the relationships between school dimen- sions (e.g. length, width) and NFS, and to explore the effect of environmental conditions within the net-pen enclosure on school size and dimensions. We also analyzed the verti- cal distribution of individuals within schools across school size, and propose a mechanistic explanation for the limited size of the two-dimensional schools observed by Partridge et al. (1983) and Lutcavage and Kraus (1995). We then ap- ply the predictions from one of the resultant models to the single open-ocean school size estimate available. Methods Field methods We employed a 30.5-m diameter, 15.3-m deep, cylindrical floating net-pen enclosure (Fig. 1) to hold the tuna used in our study. This enclosure is similar to those used in tuna research and culture operations around the world. Its low cost, large internal volume (11,128.5 m3), and its resiliency to dynamic and often damaging effects of the offshore envi- ronment make this enclosure the most appropriate type for observing the behavior of large pelagic fish in captivity. The enclosure proved to be very resilient to the damaging effects of a close pass of a hurricane and a tropical storm. A white, one-inch, straight-hung mesh net constituted the vertical walls and bottom of the enclosure. The enclosure was anchored 32.2 km offshore of Wacha- pregue, VA, on the southwest corner of 20 Mile Hill — a bathymetric feature that rises within 33.5 m of the ocean surface in deeper surrounding waters. This location is rel- atively near shore and close to a temporally and spatially reliable aggregation of small (~1 m) Atlantic bluefin tuna regularly targeted by recreational fishermen. The vertical temperature profile (°C), dissolved oxygen (mg/L), pH, suspended solids (NTU), Secchi depth (m), and conductivity (ppt) were monitored twice daily inside and outside the enclosure at 3-m intervals to a minimum depth of 15 m. A pattern of seven, single-hook trolling lures were fished from a 18. 3-m commercial vessel on 13-kg or 22-kg class trolling gear to capture fifty bluefin tuna in the vicinity of the study enclosure in June and July 1996. Tuna were subdued as quickly as possible and landed in a special- ized cloth stretcher. We recorded the fork length (cm), ap- proximate weight (kg), and general condition of each fish and released the fish into a 2400-liter elliptical transport tank. Compressed, bottled oxygen was employed to elevate 422 Fishery Bulletin 99(3) the levels of dissolved oxygen within the transport tank to ease the physiological stress associated with strenuous activity. Fresh seawater was continuously pumped into the tank to eliminate metabolic waste and maintain wa- ter quality. The tank had a padded top that minimized the potential for injury to fish during transport and re- duced water loss resulting from boat movement. Viewing ports in the padded top allowed observation of the spec- imens while in transport. Physical contact with individu- al tuna was minimized, and when contact was necessary, only well-padded devices were employed. Transport time of individual fish was variable, but generally less than 3 hours. The number of fish simultaneously transported in the tank was controlled to avoid crowding and the deple- tion of dissolved oxygen. The tuna were recovered from the transport tank by using a cloth stretcher and released in- to the net-pen enclosure. Divers in the enclosure released the tuna individually, ensuring that they recovered prop- er spatial orientation upon release. Released specimens were assimilated quickly into the existing shoals of cap- tive fish. Video cameras (Hi8) in waterproof housings were used to record school structure of the captive tuna. A shutter speed of 1/2000 second was used to optimize the resolution of still frames within the constraint of available subsur- face (<15.3 m depth) light. The automatic focus feature of the camera was disabled to avoid rapid fluctuations in fo- cal depth from the intended subject to particulate matter suspended in the water column. Observations of schooling tuna were recorded with stand-alone cameras mounted inside the enclosure and with hand-held cameras during observation dives. Cameras mounted to the floating net pen on specialized polyvinyl chloride pipe structures were stabilized with elastic compensators to lessen movement caused by wave energy. Mounted cameras were positioned to provide bead-on and subsequent perpendicular views in relation to the axis of motion of a school. This filming strategy allowed a more accurate observation of the char- acteristics of the school in three dimensions. Schooling was recorded more efficiently by using hand-held camer- as during dives than through use of mounted cameras. During a dive, the entire internal volume of the enclosure was often visible from a given point, allowing a diver to anticipate the path of travel of a tuna school and to re- position the camera to attain the best possible images. Regardless of the apparent ease of tuna schools, divers po- sitioned themselves against the enclosure’s external wall to ensure minimal behavioral modification in the filmed schools. Video recordings from both mounted cameras and from held cameras were used in analysis. Laboratory video analysis Video recordings of ABT schools were reviewed, and 74 incidences of schooling in which an entire target school was visible were used for further analyses. All observa- tions took place during daylight hours (0900 and 1600 h) and none took place during or within one hour of feeding events. An image analysis system that allowed digitiza- tion of points directly from a (paused) video source was employed for more precise quantification of school charac- teristics (Fig. 2). The system employed a video scan con- verter that overlaid the image output from a computer video source (640x480 pixel resolution) upon the video source image. The video scan converter allowed the user to select a color in the computer video overlay to be made transparent, revealing the underlying video signal (Fig.2). When employed in conjunction with an image analysis software package (SigmaScan Pro) on a personal com- puter, all the features of the image analysis software could be used on any still (paused) video source image without the use of a video frame-grabber. The position of each indi- vidual fish and school depth intervals were delineated by using the graphic capabilities and the Cartesian coordi- nate system of the image analysis software. The positions of individual fish in the school could be marked while the recording was advanced or reversed frame by frame, allowing the identification of poorly illuminated fish or fish that may have been hidden by individuals in the fore- ground of the school. Hanrahan and Juanes: Estimating the school size of Thunnus thynnus thynnus 423 Solid color to be replaced by video source image TelevEyes Pro Video Scan Convertor Computer Monitor Desktop computer Video out RGB in RGB out Video monitor Video in a_ I^SigmaScdnPfo - (1 .BMP (lull — ' : RnE3| Itf) Efo< . . Pom* &iaph : Tifloalom# tieiPyl-j-, ; H18 video source Figure 2 A schematic representation of the configuration of the video equipment used for our analyses. Data analysis School size Groups of tuna were considered to be schools and were included in the analyses if they were a polarized group (multiple individuals maintaining lateral proximity to neighbors and actively maintaining the same direction of travel during an observation period). The total number of fish in each school was counted and the frequency distri- bution of school size was determined. Because the obser- vations were assumed to be independent and the total sample size was less than 2000, the normality of the dis- tribution was tested by using the Shapiro-Wilks W-test. Least-squares regression was then used to evaluate the relationship between school size (TV.) and each environ- mental variable. Predicting number of fish in school from surface counts Our video footage was filmed at an oblique perspective to the upper boundary of schools occurring within two meters of the water’s surface. Measurement of school characteristics in body lengths or meters was not pos- sible because of the camera angle or because of poor image resolution due to low light or high turbidity level. The total number of individual fish (Ns) was determined and the distribution of individuals within the school was described in terms of five depth intervals (Fig. 3). The surface interval included fish that were at the immedi- ate surface of the school or fish that overlapped other fish at the surface on the horizontal plane. Each of the four subsequent intervals encompassed 25% of the remaining depth of the school. The number of fish per interval was designated as N: (;=the number of tuna in the zth depth interval). Fish positioned at an interval boundary were assigned to the interval in which the greater portion of their body volume was positioned. Analysis of covari- ance (ANCOVA) (Sokal and Rohlf, 1995) was employed to detect differences in the slopes of each of the regressions of Nj on Ns in order to determine whether the distribution of individuals into school depth intervals changed in pro- portion to school size. Three individual least-squares regression models were used to predict school size. The relationship between the number of individuals in the surface interval (Ny, inde- pendent variable) and the number of individuals in the re- mainder of the school (Ns\ dependent variable) was first explored using simple least-squares regression (Sokal and Rohlf, 1995). The distribution of Ns given Nt was het- eroscedastic necessitating the use of a weighted least- squares regression model by using a weight of 1/variance (Kleinbaum et ah, 1988). A similar-weighted multiple-lin- ear-regression relationship between Nv N2 (independent variables), and Ns (dependent variable) was developed be- cause fish below the immediate surface of the school are sometimes seen and counted in photographs. 424 Fishery Bulletin 99(3) Figure 3 Representation of a school of 50 Atlantic bluefin tuna. Individuals have been separated into depth intervals according to the mean values for all school sizes. Layers are arbitrarily separated in the figure to allow visualization of school structure. Predicting number of fish in school (NFS) from school dimensions School length, width, and depth in number of individual fish were recorded as indicators of school shape (Fig. 3). As in the prediction of NFS from surface counts, measurements were recorded in number of fish, not metric distance, because a precise spatial scale could not be estab- lished consistently. The lack of an accurate spatial scale also precluded the measurement of fine-scale school struc- ture such as interindividual distances. School dimension measures were recorded according to the movement axis of a school. Length in number of fish was measured along the axis of school motion (x); depth was measured verti- cally (y) and width (2) was measured perpendicular to x on the horizontal plane (Fig. 3). School length and width were analyzed in relation to school size by least-squares regression. Regression models employed each one or both dimensions (i.e. length, width, length and width) of school shape as independent variables, and Ns as the dependent variable. Depth data were not analyzed in relation to Ns because these data may not be collected practically from wild schools. The relat ionship between dimensions of each school and selected environmental variables was exam- ined by using least-squares regression. Results Behavior of the specimens The captive specimens used in our investigation exhibited a high degree of awareness of the walls of the enclosure, even during periods of excited behavior. No collision with or brushing of the net wall was observed from above the surface or in the analysis of diurnal activities from under- water video footage, and no evidence of nocturnal colli- sions was observed. After a brief period of acclimation, the tuna did not actively avoid divers in the enclosure; they reacted only to avoid collision. Visualizing the model A three-dimensional model of the typical structure of a school of ABT was constructed based on the mean charac- teristics of the schools analyzed and on qualitative obser- vations of school structure (Fig. 3). The proportionate distribution of individuals within school depth intervals varied little (see ANCOVA results below), suggesting that a single model adequately describes the mean vertical dis- tribution of individuals for schools of varying size. Number of schools and NFS When a single school comprised the entire group of 50 tuna, less than approximately 20% of the enclosure volume was involved in containing such a school (senior author, pers. obs.). Fish swimming within such a school were observed to travel along a slowly arcing path around the entirety of the enclosure without making sharp turns. Smaller schools (up to 25 individuals) occupied only a very small portion of the volume of the enclosure. Single large schools separated into two or more smaller schools and joined back together with fluidity. When more than one school was observed simultaneously, each school exhibited movement independent of another. If two schools came close to one another in the enclosure, they would either pass by, move through the other group, or join together to Hanrahan and Juanes: Estimating the school size of Thunnus thynnus thynnus 425 form a larger group. While schools were remixing during such encounters, portions of one group would sometimes join another, maintaining the same number of indepen- dently acting groups, but changing the number of individ- uals within each group. Particularly in larger schools, individual fish positions were observed to be dynamic, yet the overall shape of the school remained relatively constant. The mean size of schools observed was 18.88 individuals (n= 74, SD=13.90). The smallest schools observed had 2 ( /? =3 ) individuals, and the largest had 45 (n=l). The frequency distribution of school sizes was not normal (Shapiro- Wilks’ W, P<0.0001) and had two prominent modes centered at 5-10 individu- als and 35-40 individuals (Fig. 4). No statistically significant relationships between envi- ronmental variables and Ns were observed (Table 1). How- ever, low power due to small sample sizes may have re- duced our ability to detect significant effects. Predicting NFS from surface counts The relationship between the number of individuals in each depth interval and school size was linear in all cases. Although the r2 values for each Ni-Ns regression were relatively low, their slopes appeared to be similar and were within a narrow range (0.14-0.23). However, ANCOVA revealed that the slopes were significantly dif- ferent (P<0.001). Although the slopes were significantly different, the number of individuals in each interval remained in the same proportion except at low school sizes ( < 15 individuals). The regression model incorporating the number of fish in the surface interval of the school as the independent variable and the number of individuals in the remaining portion of the school as the dependent variable had an r2 of 0.67 (P<0.0001) (Eq. 1, Table 2). However, this regres- sion model is likely biased owing to heteroscedasticity in the dependent variable. A second least-squares regression model, incorporating a weight of 1/variance, achieved an r 2 of 0.74 (PcO.0001) (Eq. 2, Table 2). The third regression model incorporated the number of fish in the surface in- terval (N1) and the second interval (N2) of the school as independent variables, the number of fish in the remain- ing portion of the school as the dependent variable, and 1/variance as the weight. This model had an r2 of 0.70 (P<0.0001) (Eq. 3, Table 2). Partial P-tests for these models could not be executed because the dependent variable N — (Nt+ . . . Nt) changed depending on the number of school depth intervals used to predict NFS. Three least-squares regression models were used to pre- dict school size from school length and width. The model us- ing length as the independent variable and N as the depen- dent variable had an r2 of 0.74 (P<0.0001) (Eq. 4, Table 2). The second model predicted Ns from school width, achiev- 426 Fishery Bulletin 99(3) Table 1 Mean, minimum, maximum, and standard deviation of environmental data (pH; °C=degrees Celsius; mg/L=milligrams/liter; ntu=turbidity units; ppt=parts per thousand) recorded during schooling observation periods. The r2 and P-values for linear regres- sions of each school measure on each environmental variable are provided. Measurements of dissolved oxygen were discontinued due to instrument failure. Date Observation start time pH Temp. (°C) Dissolved oxygen (mg/L) Total suspended solids (ntu) Conductivity (ppt) 27 Jun 1996 10:00 AM — — — — — 21 Jul 1996 3:30 PM 8.51 22.90 9.37 — 27.45 6 Aug 1996 10:11AM 8.42 22.18 8.62 0.12 28.65 7 Aug 1996 8:45 AM 8.43 24.18 8.39 0.23 28.98 15 Aug 1996 3:45 PM 8.44 23.84 — 0.60 28.06 17 Aug 1996 10:48 AM 8.46 23.86 — 0.77 28.42 22 Aug 1996 11:55 AM 8.47 23.40 — 0.05 27.63 23 Aug 1996 11:15AM 8.50 23.50 — 0.45 29.63 Mean 8.46 23.41 8.79 0.37 28.40 Minimum 8.42 22.18 8.39 0.05 27.45 Maximum 8.51 24.18 9.37 0.77 29.63 Standard deviation 0.04 0.68 0.52 0.28 0.76 Ns 0.1* 0.15* 0.84* 0.09* 0.04* Coefficient of determination School length 0.01* 0.19* 0.91* 0.02* 0.01* School width 0.09* 0.22* 0.46* 0.1* 0.03* : /'>(). 10. Table 2 Regression equations employed to predict school size. The results of partial F-tests indicated that model 6 school size UVS). Nv N2, school length, and school width were measured in individual fish. is the best predictor of Model Equation n r2 partial F 1. N j vs. Ns Ns = (0.0337867 + 3.3173368 x NJ+ Nl 74 0.67* 2. N j vs. Ns (weight= 1/variance) Ns = (0.2042865 + 3.1849359 x NJ+ N1 74 0.74* 3. Nv N2 vs. Ns (weight= 1/variance) = (-1.412557 + 1.8516536 xNx + 0.4190831 x N2)+ AT,+ N2 74 0.70* 4. Length vs. Ns Ns = -6.769737 + 5.5334126 x length 74 0.74* 5. Width vs. Ns Ns = -6.70494 + 5.7894971 x width 74 0.79* 6. Length, width vs ,NS Ns = -9.788046 + 3.6463236 x width + 2.7083604 x length 74 0.86* (6 vs. 4) * (6 vs. 5) * * P<0.0001. ing an r2 of 0.79 (P<0.0001) (Eq. 5, Table 2). The final, multiple regression model used school length and width to predict N$ (Eq. 6, Table 2, Fig. 5) with an r2 of 0.86 (P<0.0001). Partial F- tests revealed that the final multiple regression model explained significantly more of the vari- ation in school size than either length or width separately. No significant relationships were seen when school dimen- sions were considered in terms of environmental factors (Table 1). Discussion School structure The non-normal frequency distribution of NFS and the appearance of modes in our data (Fig. 4) suggest the formation of elective groups of particular number. Previ- ous investigations in school size show that fishes actively assemble into elective group sizes dependent upon the Hanrahan and Juanes: Estimating the school size of Thunnus thynnus thynnus 427 interaction of predation risk, food availability, migratory status, and the species’ life history (Hager and Helfman, 1991; Pitcher and Parrish, 1993). Elective group size for a species differs continuously under natural conditions with varying abiotic and biotic factors. At a discrete point in time when two schools were present in the enclosure, the size of one group determined the maximum number of individuals that could be in the other. However, there were no barriers to formation of groups numbering any- where from 2 to 50 individuals. Therefore, the modes in group number that we observed should have resulted from underlying behavioral tendencies rather than enclosure- induced limitation of group size. If elective group sizes form in response to environmental conditions, then the range of environmental conditions during the study period was probably insufficient to detect environmental influ- ences on NFS, and therefore elective group sizes. The range of school sizes that we observed in the enclo- sure was limited by the number of individuals available to join schools. Photographs of ABT in the northwest Atlan- tic Ocean over a 50-day period in 1993 revealed that sur- face school counts ranged from 5 to 1294 individuals, and that the median school size was 84 individuals (Lutcav- age and Kraus, 1995). The median value is well in excess of the maximum number of fish observed in the surface layer of our schools, emphasizing the importance of verify- ing the accuracy of our model predictions for larger schools with field data. When very large schools occur in relatively shallow water, the vertical depth of the school would be confined by the maximum water depth. A similar effect could be imposed by physical and chemical barriers such as vertical stratification in temperature and dissolved ox- ygen. Tagging studies may reveal more of the individual and group behaviors of this species in relation to the en- vironment and assist in further understanding the way in which school structure may be affected by the physical and chemical environment. Our results show that the vertical distribution of fish (in intervals) varies little across the range of observed school sizes. The ANCOVA of Nt on Ns illustrated that the slope of each regression was significantly different from all others (all P<0.001), but the biological importance of this differ- ence is questionable because the slopes varied by less than 10%. Although the rate at which individuals are added to each interval varies as school size increases, there is no consistent trend in slopes among intervals. Furthermore, the statistical significance of the difference in slopes may be driven by the small standard error for each regression and extremely high power (>0.99). The shape of large schools of bluefin tuna was less vari- able than that of small schools. Schools of less than 15 in- dividuals are less vertically expansive, and generally one to three fish deep; larger (> 15 individuals) schools are more than three individuals deep (Fig. 6). The pattern of in- creasing vertical depth continues to the largest school sizes observed, which are nearly always more than five individ- uals deep. The weakly cone-shaped profile of the model school depicted in Figure 3 is representative of the shape of most schools with more than 3 intervals. Smaller schools tend to be distributed in a vertically shallow, loosely oval profile. Our findings related to school structure are consis- tent with the observations of other investigators who have 428 Fishery Bulletin 99(3) 6 i 5 - CD 4 >, ra o o -C 1 - • • • • • • • • • • • • 0 4 1 1 1 T 1 0 19 20 30 40 50 Number in school (A/s) Figure 6 School depth in number of individuals plotted against total school number (Ns). The dotted line is drawn to illustrate the increase in the vertical expanse of schools coincident with the maximum number of individuals observed in parabola and echelon-shaped schools. The maximum number of depth intervals in our analysis was 5. Schools greater than 5 individuals in total depth are described by a surface layer and four 25% depth intervals. observed that small schools have a strong horizontal as- pect, and little vertical expanse. For example, Partridge et al. (1983) and Lutcavage and Kraus (1995) observed that small ( < 15 individuals) parabola- and echelon-shaped schools of noncaptive giant ABT vary little in shape. Inter- estingly, parabola and echelon shapes are not observed for large (>15) ABT schools, coinciding with changes in inter- individual orientation between groups of less than 10 and 10-20 individuals (Partridge et al., 1983). This difference is similar to the shift in the depth of schools that we ob- served with approximately 15 individuals (Fig. 6). It is pos- sible that Partridge et al. ( 1983) and Lutcavage and Kraus (1995) observed small schools in these configurations be- cause larger schools expand vertically and adopt the semi- conical shape that we describe. These results suggest that there may be a critical minimum number of individuals that must be present in a horizontal layer before schools begin to expand vertically, which would explain the limited size of two-dimensional ABT schools. It is possible that ontogenetic variation in school struc- ture exists, but understanding how such changes occur is critical in determining how variability in school structure could affect our modeling approach. Because our models use numerical relationships rather than distance metrics to predict school size, neither ontogenetic shifts in in- terindividual spacing or packing density changes related to school size should substantially affect the predictive ability of our models. However, if ABT schooling behav- iors change at a more basic level due to enclosure or changing fish size, then our estimation techniques may be invalid outside captivity or with larger fish. Basic changes, such as vertical distribution of individuals within schools, the three-dimensional shape of schools, and strong school structure responses in relation to environmental factors, could all have serious effects on our modeling approach. Further, we feel that schools similar in form to those described as “densely-packed domes” by Lutcavage and Kraus (1995) could be described well by our models, but that numerical estimation of other school types might re- quire the application of a different estimation technique. Environmental effects on school formation The physical environment may play a role in determining the vertical position of tuna in the water column (Holland et al., 1990; Block et al., 1997) or school structure (Par- tridge et al., 1983; Lutcavage and Kraus, 1995). No rela- tionship between environmental variables and the shape of schools (quantitatively determined) or the vertical posi- tion of fish (in qualitative observations) were observed in our study, perhaps because of the small range and lack of vertical structure in salinity, pH, and dissolved oxygen Hanrahan and Juanes: Estimating the school size of Thunnus thynnus thynnus 429 measurements (Table 1), the time when school structure data were collected, or low statistical power from small sample sizes. However, on a very limited number of occa- sions, the thermocline became situated within the enclo- sure at approximately 10 m. Only a few individual fish traversed the thermal boundary, and such excursions were very brief. Entire schools were not observed to cross to the cold side of the thermal boundary. The thermal profile is an important factor in determining the vertical distribu- tion of Pacific yellowfin (Thunnus alhacares) and bigeye (T. obesus ) tuna (Holland et ah, 1990). It is reasonable to assume that the thermal profile is important in the verti- cal distribution of Atlantic bluefin tuna as well, and that this effect was not detected in our study because of the periodicity of observations, the limited variation of envi- ronmental conditions within the enclosure, or the spatial constraints imposed by the enclosure. Predicting NFS from surface counts Our least-squares regression model predicts the number of fish in ABT schools from the number of individuals at the surface of the school without attempting to describe the fine-scale structure within schools in terms of interindi- vidual spacing and orientation. The number of fish in the surface interval alone accounts for 74% of the variation in school size of three-dimensional schools similar to a densely packed dome. Because the only school type that we observed was similar to the densely packed dome, the applicability of our model to three-dimensional schools of other configura- tions is questionable and may be determined only by study- ing ocean schools of other configurations. Furthermore, the application of our model to two-dimensional schools such as parabolas, echelons, and surface sheets is not appropriate or necessary given the apparent lack of a three-dimensional component in their structure. The only data point available to verify our model pre- dictions is from a school described as “dome-shaped” that was photographed and subsequently captured by purse- seine (Lutcavage and Kraus, 1995). Thirty-two fish were counted at the surface of this school from an aerial photo- graph, and 125 “large giant” ABT were subsequently cap- tured by the purse-seine vessel. We applied our model that predicts NFS from the number at the surface of the school to this datum to produce a NFS prediction. The prediction with our model of 134 individuals differed by 7% from the purse-seine capture of 125 individual fish, and was well within the 95% confidence intervals (Fig. 7) of model pre- 430 Fishery Bulletin 99(3) dictions. When ABT schools are captured by purse seine, it is assumed that nearly the entire school is captured.2 If the net intersects an edge of the school while it is being de- ployed, the entire school will change its direction of travel in unison resulting in either the entire school being encir- cled in the net and captured or in the entire school escap- ing into open water. As a result, the estimate of total NFS from this data point is likely to be an accurate count of the number of individuals in the school. It is encouraging that our model so closely estimated the NFS of large gi- ants considering that it was constructed from data for age 2+ fish that are a fraction of the size of large giant bluefin tuna. The accuracy of our prediction indicates the poten- tial for generality of ABT school structure across both tu- na size and NFS. However, substantial verification of our models is necessary before they may be applied to abun- dance estimation. Predicting NFS from school dimensions An alternative to using N1 and N2 to predict Ns is to use maximum school length and width. Identifying the longest and widest axis of the surface of a school and counting the number of individuals along these axes may yield more accurate estimates of the total NFS. The maximum dimen- sions of bluefin schools (length, width) had greater power as predictors of Ns than N1 and N.2 ( model 3 versus models 4, 5, 6, Table 2). Moreover, a combination of length and width to predict Ns produced a more confident estimate of school size than models using either variable individually. Irregularity in length and width at small school sizes likely introduced variation that reduced the individual predic- tive power of these variables. Inclusion of both length and width in the model to predict Ns could allow accurate pre- diction of small schools that are elongate or wide. The regression diagnostics for the model using school length and width to predict school size (model 6, Table 2) suggest that it is the more reliable model for estimat- ing NFS with aerial photographs. For the single open- ocean estimate available in the literature (Lutcavage and Kraus, 1995), the length and width of the school in num- ber of individuals could not be determined. Model 6 may have the potential to yield a more accurate estimate of school size with a wider range of photograph qualities (as affected by sea state, water clarity, sunlight, etc. ) because of its ability to predict NFS from partial surface counts. However, the utility of school length and width to predict N s will remain uncertain until field data are available for thorough evaluation. Enclosure effects The effects of capture and captivity on school structure and behavior were points of concern in our study. Evaluat- ing the effects that the enclosure had on school structure is problematic because the same factors that led to the use 2 Genovese, M. (captain). 1995. Personal commun. FV White Dove Too , 600 Shunpike Rd., Cape May, NJ 08210. of an enclosure to make the observations preclude a direct in situ comparison. Because tuna are highly mobile and unpredictable in their movements, it would be difficult to obtain a number of school observations comparable to that collected from the captive fish in our study. Further- more, the ability to approach schools in nature without significantly disturbing them is questionable, and effec- tive observation at a distance that would not cause dis- turbance would be unlikely because of turbidity and the normal movement of schools. In this respect, a group of tuna that has become comfortable with the presence of human observers may be a better source of accurate school structure observations than a school of noncaptive fish that may perceive a human or mechanical presence as a predatory threat and react accordingly. Evidence of the acclimation of the study specimens to the enclosure was seen in their active and aggressive feeding behavior (Han- rahan and Juanes3) that was similar to the available anec- dotal accounts of their open-water feeding behavior. The relation between fish length and enclosure dimensions may cause the impression that school formation was con- strained heavily by captivity and that multiple schools could not achieve meaningful separation from one another (Fig. 1 illustrates a school of tuna at a scale of 1:1 to the enclosure). However, the relatively low density of fish in the enclosure (0.05 kg/m3) allowed individuals to move in an uninhibited manner within the enclosure. Although the extent of enclosure-induced behavioral modification cannot be quantified in our study, the ability of our simple linear model to predict accurately the NFS for a noncap- tive school is very encouraging. Acknowledgments We thank S. Belle, P. Sylvia, and the volunteers of the New England Aquarium bluefin tuna research project for invaluable collaborative support. This manuscript was greatly improved by comments provided by J. Boreman, E. Brainerd, R. Rountree, K. Friedland, and four anonymous reviewers. This work is the result of research sponsored by NOAA National Sea Grant Office, Department of Com- merce, under grant NA 46RG0470, Woods Hole Oceano- graphic Institution Sea Grant project no. 22800058 to F. J. Additional support was provided by NOAA’s Cooperative Marine Education and Research Program, the Department of Natural Resources Conservation (UMass, Amherst), and the Manasquan River Marlin and Tuna Club. Literature cited Block, B. A., J. E. Keen, B. Castillo, H. Dewar, E. V. Freund, D. J. Marcinek, R. W. Brill, and C. Farwell. 1997. Environmental preferences of yellowfin tuna ( Thun - 3 Hanrahan, B., and F. Juanes. In prep. Atlantic bluefin tuna (Thunnus thynnus thynnus) prey size-selectivity and feeding behavior. Hanrahan and Juanes: Estimating the school size of Thunnus thynnus thynnus 431 nus albacares ) at the northern extent of its range. Mar. Biol. 130:119-132. Carey, F. G., and K. D. Lawson. 1973. Temperature regulation in free-swimming bluefin tuna. Comp. Biochem. Physio. 44(A):375-392. Carey, F. G., and J. M. Teal. 1969. Regulation of body temperature by the bluefin tuna. Comp. Biochem. Physio. 28:205-213. Clay, D. 1991. Atlantic bluefin tuna (Thunnus thynnus) (1): a review. In World meeting on stock assessment of bluefin tunas: strengths and weaknesses, p. 91-179. Inter-Am. Trop. Tuna Comm. Spec. Rep. 7. Clark, C. W., and M. Mangel. 1979. Aggregation and fishery dynamics: a theoretical study of schooling and the purse seine tuna fisheries. Fish. Bull. 77:317-329. Collette, B. B., and C. E. Nauen. 1983. FAO Species catalogue, vol. 2: scombrids of the world, an annotated and illustrated catalogue of tunas, macker- els, bonitos and related species known to date. FAO Fish. Synop. 125, 137 p. FAO, Rome. Figley, W. 1984. Recreational fishery for large offshore pelagic fishes of the Mid-Atlantic Coast. NJ Div. Fish Game Wildl. Tech. Ser. 84-1, Trenton, NJ, 66 p. Hager, M. C., and G. S. Helfman. 1991. Safety in numbers: shoal size choice by minnows under predatory threat. Behav. Ecol. Sociobiol. 29:271-276. Holland, K. N„ R. W. Brill, and R. K. C. Chang. 1990. Horizontal and vertical movements of yellowfin and bigeye tuna associated with fish aggregating devices. Fish. Bull. 88:493-507. Kleinbaum, D. G., L. L. Kupper, and K. E. Muller. 1988. Applied regression analysis and other multivariable methods. Duxbury Press, Belmont, CA, 718 p. Lo., N.C.H., I.D. Jacobson, and J.L. Squire. 1992. Indices of relative abundance from fish spotter data based on delta log-normal models. Can. J. Fish. Aquat. Sci. 49:2515-2526. Lutcavage, M., and S. Kraus. 1995. The feasibility of direct photographic assessment of giant bluefin tuna, Thunnus thynnus in New England waters. Fish. Bull. 93:495-503. Lutcavage, M., S. Kraus, and W. Hoggard. 1997. Aerial survey of giant bluefin tuna, Thunnus thyn- nus, in the Great Bahama Bank, Straits of Florida, 1995. Fish. Bull. 95: 300-310. Magnuson, J. J., B. A. Block, R. B. Deriso, J. R. Gold, W. S Grant, T. J. Quinn II, S. B. Saila, L. Shapiro, and E. D. Stevens. 1994. An assessment of Atlantic bluefin tuna. National Academy Press, Washington, D.C., 148 p. Mather, F. J., III. 1962. Tunas (genus Thunnus) of the western North Atlan- tic. Part III: Distribution and behavior of Thunnus species. In Symposium on scombroid fishes, part 1, p.1-16. Symp. series 1, Marine Biological Association of India, Mandapam Camp., India. 1974. The bluefin tuna situation: sixteenth annual interna- tional game fish research conference, p. 93-106. Woods Hole Oceanographic Institution, Contribution 3304, Woods Hole, MA. Partridge, B. L., J. Johansson, and J. Kalish. 1983. The structure of schools of giant bluefin tuna in Cape Cod Bay. Environ. Biol. Fishes 9:253-262. Pitcher, T. J., and J. K. Parrish. 1993. Functions of shoaling behavior in teleosts. In The behavior of teleost fishes, 2nd ed. (T. J. Pitcher, ed.), p. 363- 439. Chapman and Hall, New York, NY. Safina, C. 1993. Bluefin tuna in the west Atlantic: negligent manage- ment and the making of an endangered species. Conserv. Biol. 7:229-233. Scott, G. P, M. R. Dewey, L. J. Hansen, R. E. Owen, and E. S. Rutherford. 1989. How many mullet are there in Florida Bay? Bull. Mar. Sci. 44:89-107. Scott, G. P, S. C. Turner, C. B. Grimes, W. J. Richards, and E. B. Brothers. 1993. Indices of larval bluefin tuna, Thunnus thynnus, abundance in the Gulf of Mexico; modeling variability in growth, mortality and gear selectivity. Bull. Mar. Sci. 53: 912-929 Sokal, R. R„ and J. F. Rohlf. 1995. Biometry. W.H. Freeman and Company, New York, NY, 887 p. Squire, J. L. Jr. 1962. Distribution of tunas in oceanic waters of the north- western Atlantic. Fish. Bull. 62:323-341. 1972. Apparent abundance of some pelagic marine fish off the southern and central California coast as surveyed by an airborne monitoring program. Fish. Bull. 70:1005-1019. 1993. Relative abundance of pelagic resources utilized by the California purse-seine fishery: results of an airborne monitoring program, 1962-1990. Fish. Bull. 93:348-361. Suzuki, Z., and Y. Ishizuka. 1991. Comparison of population characteristics of world bluefin stocks, with special reference to the West Atlantic bluefin stock. Coll. Vol. Sci. Pap. ICCAT Reel. Doc. 35(2): 240-245. Williams, K. 1981. Aerial survey of pelagic fish resources of South East Australia 1973-1977. CSIRO Div. Fish. Oceanogr. Rep. 130, 82 p. 432 Effects of gear selectivity and different calculation methods on estimating growth parameters of bluefish, Pomatomus saltatrix (Pisces: Pomatomidae), from southern Brazil Flavia M. Lucena Carl M. O'Brien CEFAS Lowestoft Laboratory Pakefield Road Lowestoft, Suffolk NR33 OHT, United Kingdom Present address (for F. M. Lucena): Umversidade Federal Rural de Pernambuco Departamento de Pesca. Av. Dom Manoel s/n Dois Irmaos. Recife, PE. Cep: 52171-900, Brazil E-mail address (for F M. Lucena): flavialucena@hotmail.com Abstract— We examined the effects of gear selectivity and the con- sequences of a chosen method for estimating growth parameters of bluefish, Pomatomus saltatrix, from southern Brazil. Samples were obtained from commercial landings at Rio Grande ( 1992-97 ). Age deter- mination for 1159 fish indicated that gill-net and purse-seine fish- eries caught 1-5 year-old and 1-7 year-old fish, respectively. Trawl- ers caught fish from ages 1 to 7. Because of their high degree of selectivity, the gill nets caught the larger individuals of ages 1 and 2, as well as the smaller individu- als of ages 3-5. Faster-growing fish captured with gill nets had smaller scales than fish of a similar length caught with purse seines. By con- trast, the slower-growing fish cap- tured with gill nets had larger scale size than fish of a similar length caught with purse seines. For gill nets and purse seines, there were differences between the von Ber- talanffy growth estimates derived from both mean values of back- calculated length-at-age and indi- vidual back-calculated length-at- age. We also recorded differences in growth parameters obtained from back-calculated length-at-age derived from measurements to the last annulus only and from mea- surements of all annuli up to the sampling age. Selectivity-related bias was incorporated in the esti- mation of growth parameters, yield- ing unrealistic estimates of and k for the gill-net growth curve. Manuscript accepted 18 december 2000. Fish. Bull. 99:432-442 (2001). The bluefish, Pomatomus saltatrix (L.) (Pomatomidae), is a highly mobile pelagic predator (Haimovici and Krug, 1996), widely distributed along the con- tinental shelf in the temperate and warm waters of the Atlantic, Pacific, and Indian oceans (Wilk, 1977). Abun- dant in Brazilian waters, this species is commercially exploited by purse-seine and gill-net fleets operating in the sub- tropical coastal waters of Rio Grande do Sul between Conceigao (31°42'S) and Chut (33°43'S), at depths between 8 and 100 m (Lucena and Reis, 1998) (Fig. 1). Purse-seine boats are approximately 23 m in length and have 220-330 HP engines, whereas gill-net boats are usu- ally 14-16 m in length and have 90-150 HP engines (Boffo and Reis, 1997). Gill nets used in the bluefish fishery are drift gill nets 1800 m in length and that have a stretched mesh size of 90 mm. Commercial catches of bluefish were first monitored from 1976 to 1983 on the fishing grounds from Sarita (32°37'S) to Conceigao at depths up to 35 m (Krug and Haimovici, 1991). Bluefish were then primarily exploited by purse seine but were also taken as bycatch in a gill- net fishery targeting wreckfish (Poly- prion arnericanus ) and elasmobranchs. Commercial catches were dominated by fish between 1 and 3 years old for the purse-seine fleet and by fish aged 2-4 years old for the gill-net fleet ( Krug and Haimovici, 1989). As catches from the estuarine Patos Lagoon declined, beginning in 1982, the artisanal fleet began exploiting new fishing grounds and a fishery developed in shallow coastal waters (Reis, 1992). Since 1990, bluefish have been caught primarily by gill net (up to 72% of land- ings by weight) and the fishing grounds now extend southward into waters up to 100 m (Lucena and Reis, 1998). For the period 1991-95, commercial catches have included individuals to age 10 (Lu- cena, 1997). Since the emergence of the gill-net fleet, the exploitation of the bluefish has been increasing. From 1991 to 1996, there has been an 87% increase in the length of netting deployed and a 166% increase in the soak time of the gear (Lucena and Reis, 1998). In addition, there has been a decrease in mean length of fish caught by purse seines and gill nets from 1992 to 1995 (Luce- na, 1997). Average annual landings of the bluefish decreased from 3100 (dur- ing the period 1990-95) to 1700 tonnes (during 1996-98) (IBAMA1). The growth parameters of the south- ern Brazilian bluefish population were estimated by Krug and Haimovici (1989), using scales from fish landed commercially by several gears (gill nets, purse seines, trawlers, long lines, and beach netting) from 1976 to 1983. Their study excluded individuals larger than 63 cm TL because the larger fish lived 1 IBAMA ( Instituto Brasileiro do Meio Ambi- ente e dos Recursos Naturais Renovaveis). 1990-1998. Unpubl. data. Centro de Pesquisa do Rio Grande (CEPERG)-RS, IBAMA, Brazil. Cep: 96201-900. Lucena and O'Brien: Effects of gear selectivity and different calculation methods on growth parameters of Pomcitomus sciltcitrix 433 54° 53° 52° 5f at depths greater than 35 m, the operation- al limit of the fleet at that time. Given to- day’s increased exploitation of bluefish, and the lack of older fish in the earlier study, we decided to examine the age structure and growth parameters of the bluefish stock off southern Brazil. New data are compared to previously published information and the effects of gear selectivity on the estimation of growth parameters are also evaluated. Annulus formation occurs in winter (June to September) and is associated with lower water temperature (Haimovici and Krug, 1996). Rings do not become visible until growth resumes in January or February. During peak spawning, January-February, the true age of individuals with one ring is 7 months on average (Haimovici and Krug, 1992). Even though most sampled bluefish have usually been caught during the short period of the fishing season (June to September), the real age of a fish may vary by between 4 and 9 months from that notionally adopted because the bluefish is a multiple spawner and because this species’ scale ring forma- tion is associated with lower temperatures (the exact timing of ring formation varies from year to year and can also be influenced by environmental phenomena such as El Nino events). Hence, because sample col- lections were spaced unevenly throughout the fishing season for both within-year and between-period comparisons, growth rates based on changes in observed lengths were deemed inappropriate and the estimation of growth parameters was based on back- calculated lengths. Many early writers, using the back-calcu- lation technique, made no mention of the reason for choosing a particular procedure (e.g. Johnson and Saloman, 1984; Barger, 1990; Vieira and Haimovici, 1993). Valuable reviews on back-calculation were carried out by Hile (1970), Tesch (1971), Casselman (1987) and Francis (1990) and, although this technique is widespread, it still does not appear to be well understood. Moreover, there is a common tendency to plot only mean, rather than indi- vidual, back-calculated length-at-age to produce the growth curve (Hilborn and Walters, 1992). The effects of different approaches on growth parameter estimates are discussed in this paper. Materials and methods Sampling, age structure, and scale reading We sampled commercial landings from three fleets at Rio Grande from 1992 to 1997: catch from the gill-net fleet. catch from the purse-seine fleet, and bycatch from the trawl (pair and otter) fleet (Table 1). We sampled 13 addi- tional bluefish from the gill-net fishery targeting wreck- fish (Polyprion americanus ), choosing larger (older) fish in order to investigate the size of the bluefish that move along the shelf break at depths >200 m, where the wreck- fish fishery operates. We collected an additional 57 individ- uals (75-135 mm in TL) with experimental trawls in the estuary of Patos Lagoon in order to calibrate the weight- length relationship for juveniles. Total length (TL) in mm of 1159 fish (260 to 711 in TL), measured to the end of the extended tail were recorded. The total weight (TW, in grams) and sex (macroscopically determined) of 580 bluefish, ranging in size from 75 to 711 mm TL, were determined to calculate the weight-length relationship, by using nonlinear least-squares regression 434 Fishery Bulletin 99(3) Table 1 Depth, sampling period, length and age range, and number of boats and individuals examined as part of the bluefish sampling in southern Brazil during the 1992-97 period. Bluefish target species Bluefish bycatch Gill net Purse seine Gill net Trawl (pair and otter) net Depth (m) 10-35 8-100 270 8-68 Sampling period Jun-Aug May-Sep Sep Jan-Mar, May Length range (mm) 301-602 265-711 650-770 262-711 Age range (years) 1-5 1-7 5-10 1-7 Number of boats examined 33 11 1 7 Number of individuals examined 853 233 13 60 (SPSS, 1998). T- tests were used to test for gear- and sex- specific length-weight relationships. Scales of captured bluefish were removed from under the pectoral fin and stored in a coin envelop. The scales of 1159 fish were mounted between glass slides and fish age was determined by following Krug and Haimovici (1989). Fish were considered to be 1 year old after the formation of the first winter ring and a further year was added to the age of the fish for each subsequent ring (Krug and Haimovici, 1989). Two different readers assigned ages for the fish, with no knowledge of fish length or month and gear of capture. If two readings agreed, then the age of that reading was ad- opted as definitive. If not, the scale was reread after two months, concurrently by both readers. If two of the three readings were the same, the age of these two readings was assumed as definitive. If all three readings differed, the scale was omitted from further analysis. On some scales, a single reader assigned ages. This method was followed with a time lag (2 months) between readings. Estimation of growth parameters Sample sizes were sufficient for estimation of growth parameters from gill-net and purse-seine samples. Ages from the direct reading of scales were obtained for 816 bluefish from gill-net samples and for 212 bluefish from purse-seine samples. Ricker (1992) pointed out that the inclusion of sampled fish taken at different times during the growing season should be avoided and that preferably fish collected be- tween the formation of an annulus and the start of the next year’s growth should be used in analysis. Only data obtained during the fishing season were included for esti- mation of growth parameters derived from direct reading, and data derived from purse-seine samples (the less selec- tive gear) were used to compare the growth parameters derived from mean observed lengths-at-age and back-cal- culated lengths-at-age. For back-calculation of growth parameters, 283 speci- mens ( 174 from gill-net samples and 109 from purse-seine samples) were used. Fish were processed by sex and gears used in their capture to avoid bias in results. We determined the relationship between fish length (TL) and scale radius (S). Krug and Haimovici (1989) pre- viously identified a nonlinear relationship between fish length and scale radius for the southern Brazilian bluefish ( TL=aSh ). Back-calculated total lengths at age were deter- mined by using the formula of Monastyrsky (see Bagenal and Tesch, 1978; Francis, 1990): Ln = (Sn/S) b x TL, where Ln = the length of fish when annulus “n” was formed; Sn - the radius of annulus “n” (at fish length Ln)\ and b = the slope from the body-scale relationship TL on S. This nonlinear method for back-calculation assumes that the relationship TL on S is of the form TL = aS,b where a and b are parameters to be estimated usually by nonlinear least squares. The periodicity of growth increments was examined by us- ing analysis of marginal increments [(S - Sn )/(Sn - from fish 2-7 years old (mainly 2-3 years old) (following Krug and Haimovici, 1989). Because of the limited temporal range of samples, no marginal increments analysis was done for the months March-April and November-December. We derived theoretical growth parameters by fitting back-calculated lengths-at-age to the von Bertalanffy (1934) growth equation L, = ( 1 - exp where Lt = LTC = k = the length at age t\ the asymptotic length; the brody growth coefficient; and the age when length would theoretically be zero. Estimation of growth parameters was based upon two criteria. Criterion I identified the type of summary statis- tic to be used: Lucena and O'Brien: Effects of gear selectivity and different calculation methods on growth parameters of Pomatomus saltatnx 435 la corresponded to mean values of back-calcu- lated lengths-at-age; and lb represented individual back-calculated lengths- at-age. Criterion II identified the type of data to be used: Ha corresponded to the back-calculated length- at-age derived from the last annulus only; and lib represented the back-calculated lengths-at- age derived from all the annuli up to the sampling age. We calculated growth parameters by using four methods for each gear: method 1 (where criteria la and Ha were used), method 2 (where criteria la and lib were used), method 3 (where criteria lb and Ha were used), and method 4 (where criteria lb and lib were used). The growth curves were estimated with the program AD Model Builder (Otter Research, 1996), which uses an au- tomatic differentiation algorithm to estimate the param- eters of a nonlinear function with an appropriate objective function. The software allows the calculation of confidence intervals for parameter estimates with asymptotic normal approximations. The objective function considered is the nonlinear least- squares regression criterion, R(.), either given by n=816 n= 10 n= 228 n= 57 Figure 2 Relative proportion (%) of bluefish caught by various fisheries oper- ating off Brazil’s southern coast. l R{KXt0\O) = ^Ot-L,? 3 yr) and purse-seine fisheries caught fish of ages 1-7 (mode=2-3 yr). Trawlers caught fish of ages 1-7. Individ- uals of ages 5 to 7 represented less than 2% of samples (individuals of 5 to 7 years old were underrepresented also in the back-calculation procedures). The deep-water (to 200 m) gill-net fishery targeting P. americanus caught individuals of age 5-10 and the only fish older than age 7. There was a considerable overlap in the range of observed lengths-at-age (Table 2), nevertheless significant differ- ences between gears were detected for some ages. Gill nets caught larger individuals of ages 1 and 2 and smaller individuals of ages 3-5 than purse seines. Trawls, on the other hand, appear to catch smaller individuals in all age classes. when mean values are considered, or given by T n, R(LM l°,) = £X(0'I-L')2’ t=l f=l when individual values are considered. Note that Ot is the mean back-calculated length-at-age t, Oti is the ith individual back-calculated length-at-age t, T denotes the maximum number of distinct ages, and nt is the number of observations at age t. To compare growth curves between gears, Hotelling’s t-test (1979) was used. Results Age structure of the stock We examined a total of 1159 fish and 94% of samples were collected during the fishing season (June to September) — the remaining numbers represented bycatch during the off-peak season. Age determination (Fig. 2) indicated that the gill-net fisheries caught fish of ages 1-5 (mode= Weight-length relationship We found no statistically significant differences in length- weight relationships between gear types or between sexes (f-test, P>0.05). The relationship for both sexes and all gears combined is illustrated in Figure 3. Body-scale relationship Gear-specific regressions of TL on S (Fig. 4) showed that gill nets tended to catch the fast-growing fish of younger ages and the slow-growing individuals of older ages. These faster-growing younger fish (<360 mm TL, 1-2 years) have smaller scales than similar-size purse-seine-caught fish U-test; P<0.05). By contrast, the slower-growing fish (>410 mm TL, 4—5 yr) had larger scales than fish of a similar length caught by the purse-seine fleet. For the length range that the gill nets target most efficiently (360-410 mm TL, age 3), the TL-S relationship for both gill net and purse seine was similar. Periodicity of growth increments Our limited data showed marginal increment to be lowest in January-February, corresponding to ring deposition. 436 Fishery Bulletin 99(3) Figure 3 Weight-length relationship for bluefish (sexes and gears combined). Table 2 Mean observed length-at-age for gill-net (targeted catch), purse-seine, and trawl catches. The symbol SD denotes the standard deviation, n denotes the number of individuals examined and “ — “ denotes that a value could not be calculated. ANOVA was per- formed only for the age classes with n greater than 10. Age (yr) Gill net length at age (mm) Purse-seine length at age (mm) Trawl length at age (mm) Average Range SD {n) Average Range SD ( n ) Average Range SD (n) 1 354.3'2 301-417 30.8 (33) 313.5' 276-385 23.4(31) 278.5 265-292 19.0 (2) 2 408.7' 293-495 32.6 (286) 385.0' 290-485 44.5 (49) 372.8 285-409 47.0 (6) 3 438.9' 328-543 31.0 (431) 460.4'"3 364-520 40.7 (61) 419.43 361-498 30.6 (24) 4 467.7' 410-602 42.3 (65) 518. 17’3 451-622 38.3(62) 451. 53 342-554 54.7 (12) 5 498.0 — (1) 552.5 418-711 54.4 (17) 474 — — (1) 6 622.7 556-661 38.7 (7) 507 477-661 89.3 (4) 7 664.5 657-672 10.6 (2) 651 642-660 12.72 (2) ' Significant difference between lengths-at-age for gill-net and purse-seine catches (ANOVA, P<0. 05). 2 Significant difference between lengths-at-age for gill-net and trawl catches (ANOVA, P<0.05). f Significant differences between lengths-at-age for purse-seine and trawl catches (ANOVA, P<0.05). Lucena and O'Brien: Effects of gear selectivity and different calculation methods on growth parameters of Pomatomus sa/tatnx 437 Table 3 Mean back-calculated total lengths (mm) of bluefish for the purse-seine and gill-net fishery. Lc = Mean observed lengths-at-age. Mean back-calculated lengths-at-age derived from the last annulus are shown in bold. Age-group Lc SD n (last annulus only) Method 1 2 3 4 5 6 7 Purse-seine fishery 1 313.5 22 27 202.2 2 385.0 26 31 189.0 315.5 3 460.4 23 36 227.9 358.0 437.3 4 518.1 26 35 234.5 373.1 462.4 504.3 5 552.5 8 72 225.4 366.9 455.4 504.5 547.1 6 622.7 3 30 219.1 352.5 447.2 502.0 552.7 607.9 7 664.5 1 — 247.9 352.6 512.1 574.6 615.9 636.5 647.4 Mean (all annuli) 109 214.6 350.6 451.3 506.1 554.4 615.4 647.4 SD 36 47 45 46 64 28 — Gill-net fishery 1 354.3 18 27 236.2 2 408.7 59 46 230.4 346.1 3 438.9 75 30 239.8 358.2 430.3 4 467.7 21 41 244.9 366.2 445.5 480.8 5 498.0 1 — 196.3 350.3 403.7 426.5 460.0 Mean (all annuli) 174 236.7 354.9 433.3 483.4 460.0 SD 41 39 58 41 — Growth occurs from May to October and peaks in August. From these data, we believe that the growth increments are annuli. Back-calculated length-at-age and estimation of growth parameters Back-calculated lengths-at-age calculated by using the last annulus only ( Ila) were smaller than lengths-at- age calculated by using all annuli (lib) (Table 3). There were also significant differences in lengths-at-age for these two criteria between gear types (7-test, P<0.05 for all age classes). Sexes were combined because there was no significant difference between them (7-test, P<0.05). Back-calculated lengths-at-age for gill-net samples showed a decreasing pattern for older fish (5 years old), indicating the presence of Lee’s phenom- enon and the selective effects of this gear. Estimates of theoretical growth parameters indi- cated differences by gear and method of estimation (Table 4, Fig. 5). For criterion I, estimates with mean back-calculated lengths-at-age (criterion la) produced higher values of (and smaller values of k ) for purse-seine samples than estimates with individual back-calculated lengths-at-age (criterion lb). The op- posite trend was observed for the gill-net samples. On the other hand, for criterion II, especially for purse- seine samples, estimates of mean back-calculated lengths- at-age derived from the last annulus only (criterion Ila) produced smaller values of L ^ (and higher values of k) than estimates with mean back-calculated lengths-at-age for all annuli (criterion lib) (Table 4). Furthermore, for the purse seine, the growth parameters derived from methods 3 (lb, Ila) and 4 (lb, lib) were distinct (Table 4). The at- 438 Fishery Bulletin 99(3) < o o o o o o O 00 CD 400 --- - - - Purse seine fit Gill net fit 200 % X Purse seine observations Gill net observations 1000 800 600 400 200 0 B 12 1000 800 600 400 200 0 0 c 2 4 6 8 10 Age (years) 12 Figure 5 Growth curves for bluefish caught by purse seines and gill nets calculated by (A) method 1 (mean lengths — last annulus only) (B) method 2 (mean lengths — all annuli) and (C) method 3 (individual lengths — last annulus only). tempt to fit the von Bertalanffy growth curve by method 4 for the gill net failed because no feasible solution could be found for the parameters of the curve. For all methods there were significant differences in growth parameter estimates between the gill-net and the purse-seine samples (Hotelling’s t-test, P<0.05). The comparison between mean back-calculated (espe- cially criterion Ha) and mean observed length-at-age showed that the latter is systematically larger than the former (Table 3). This difference was expected because the mean back-calculated lengths-at-age were obtained for length at time of annulus formation (where spawning season and time of annulus formation are considered uni- form for all fish). The differences between both means may be attributed to growth after the growth mark for- mation. Moreover, the mean observed length-at-age cal- culated from sampled fish was obtained from individuals that may have spawned at different times of the year or whose ring formation may have been distinctly visible in time, resulting in fish with distinct real ages. Hence, growth parameters obtained from mean observed lengths- at-age of purse-seine catches during the fishing season (L„= 985 mm, £=0.12 and t0= 2.17) were significantly dif- ferent from the growth parameters derived from back- calculated lengths-at-age (also from purse-seine catches) (Hotelling’s t-test, PcO.001 for all methods). Lucena and O'Brien: Effects of gear selectivity and different calculation methods on growth parameters of Pomatomus saltatrix 439 Table 4 Growth parameters and correlation between k and for different methods (1-4), and respective criteria (la, lb, Ha, lib) of back- calculation. Standard errors are given in parenthesis. “ — “ denotes that parameter estimates did not converge to a solution. Parameters Purse seine Gill net [1] Ia-IIa [2] Ia-IIb [3] Ib-IIa [4] Ib-IIb [1] Ia-IIa [2] Ia-IIb [3] [4| Ib-IIa Ib-IIb Lx 773 (32.6) 743 (26.6) 754(53.8) 670(55.2) 496(21.3) 491 (17.1) 589 (45.3) k 0.25 (0.03) 0.27 (0.03) 0.26 (0.04) 0.35 (0.07) 0.65 (0.14) 0.71 (0.13) 0.39 (0.08) ^0 -0.21 (0.12) -0.27 (0.12) -0.15 (0.12) -0.09 (0.13) 0.03 (0.20) 0.09 (0.17) 0.27 (0.2) r(k,LJ -0.97 -0.97 -0.98 -0.98 -0.89 -0.87 -0.98 Discussion Back-calculation of length is widely used to obtain growth curves, to estimate length-at-age of individuals that are rarely observed, to compare growth differences among populations or sexes of the same species, and even to illus- trate gear selectivity (Francis, 1990). However, this tech- nique is poorly understood (Francis, 1990; Rijnsdorp et al., 1990) and some of the sources of bias in using the technique are 1) inaccurate counts of annuli and incor- rect estimation of time of formation of the growth mark; and 2) an erroneous choice of the mathematical function to describe the body-scale relationship. In respect to the first source of bias, our results corroborate the findings of Krug and Haimovici (1989). In respect to the body- scale relationship, the main concern of Francis ( 1990) was to decide whether the appropriate regression was TL on S (body proportional) or S on TL (scale proportional). The small difference between the back-calculation derived from the two approaches for the bluefish can be a mini- mum measure of precision of the back-calculation proce- dure (Francis, 1990). Campana (1990) and Wright et al. (1990) pointed out that slow-growing fish have larger bony structures (scales or otoliths) than fast-growing fish of the same size. For the bluefish body-scale relationship, we attribute this effect to gear selectivity. The fast-grow- ing fish (ages 1 and 2 from the gill-net catches) tend to occur above the purse-seine curve and the slow-growing fish (ages 4 and 5 from the gill-net catches) tend to occur below the purse-seine curve in a TL-S relationship. An additional source of bias may be introduced depend- ing on whether mean back-calculated lengths-at-ages or individual lengths-at-age are used to fit the von Berta- lanffy equation. The estimated growth parameters derived from the two methods may differ considerably (Hilborn and Walters, 1992). Use of mean values would ignore indi- vidual variability in length-at-age, giving the same weight for possibly uncertain ages because of low sample sizes. We suggest that back-calculated lengths-at-age derived from the last annulus only (criterion Ilahbe used rather than the back-calculated lengths-at-age derived from all annuli (criterion lib). Many investigators continue to use mean-weighted data (all annuli) (e.g. Krug and Haimov- ici, 1989; Barger, 1990) — a procedure in violation of the least-squares assumption of independence of sample ele- ments (Draper and Smith, 1966). This assumption is vio- lated when multiple measures from a single fish are used. Use of the last annulus only does not use all information available about growth of all cohorts and may result in an incorrect estimate of lengths-at-age for age classes absent or infrequently caught. However, if representative sam- ples are available for younger age groups, the use of back- calculated lengths-at-age derived from the last annulus only is to be preferred. Ricker (1969) noted that where differential mortality exists, the calculated average length of a particular age class calculated from the last annulus differs from the cal- culated average length of an age class calculated from pre- vious annuli — the latter representing the former size of the fish that has survived to the sampling age. The differ- ential mortality may be due to natural or fishing causes but under these circumstances, the average size of fish in the year class becomes different as time passes and the frequency distribution of the survivors would become pro- gressively skewed. Gutreuter (1987) suggested that back- calculation should be restricted to the most recent annu- lus to avoid bias from size-selective sampling. Gear selectivity can influence estimates of growth para- meters (Ricker, 1969; Potts et al., 1998), and we found that gear-related differences in parameters were significant. Gill nets, on the basis of their mesh size, are a selective gear. Trawls are selective for smaller fish that may not be able to avoid the nets because of slower swimming speed (Hilborn and Walters, 1992). Purse seines are probably a less selective gear (Cushing, 1968). Observed lengths- at-age reflect the differences in selectivity between gears. Gill nets catch the faster-growing fish at age 1 and 2 and the slower-growing fish at age 4 and 5, and trawls catch the smaller individuals for each age. The high selectivity of gill nets is visible in the body- scale relationship, the mean observed length-at-age, and in the back-calculated lengths-at-age which indicate the presence of Lee’s phenomenon (Lee, 1920). Size-selective mortality — caused by the differing catchabilities of fish at different sizes — is the probable reason for Lee’s phenom- enon in back-calculation of lengths from gill-net catches. 440 Fishery Bulletin 99(3) Selectivity-related bias from gill-net back-calculation is thus incorporated in the growth parameter estima- tion. The growth curve calculated from gill-net back- calculation reports unrealistic and k. Method 4 (in- dividual back-calculated lengths-at-age derived from all annuli) for the gill net failed because no feasible so- lution could be found for the parameters of the curve. High variability of back-calculated lengths-at-age de- rived from all annuli (see standard deviations in Table 3) in the presence of Lee’s phenomenon may have been the reason for the unsuccessful attempt to obtain esti- mates of the parameters of the von Bertalanffy growth curve with this method. In our study, the growth parameters estimated from individual back-calculated length-at-age derived from the last annulus only (method 3) on purse-seine sam- ples were considered to be the most appropriate for future assessment of the southern Brazilian bluefish stock. This gear is less selective than others and data comprise a wider range of total length groups and age groups. Changes in Brazilian commercial fishing operations for the bluefish over the last decade have led to chang- es in the length and age structure of commercial land- ings. From 1976 to 1983 the maximum age and total length of bluefish in the fishery were 7 years and 630 mm TL, respectively, and less than 5% of fish were 4 years and older (Krug and Haimovici, 1989). Cur- rently the fishery lands fish up to age 10, and 16% of fish are aged 4 and older. The inclusion of older in- dividuals in recent catches is due to the expansion of the fishery into deeper areas. In 1990, the purse- seine fishery began moving to deeper waters, catching a wider range of sizes of bluefish. Hilborn and Walters (1992) have pointed out the necessity to understand fleet dynamics in order to assess the population dy- namics of a species. Our mean back-calculated lengths-at-age (calculat- ed from the last annulus only) are similar to those found by Krug and Haimovici (1989) (except for ages 1 and 3) (Table 5), but estimated growth parameters for the two studies are different. We believe these dif- ferences are due to differences in fishing operation be- tween the two periods (fleet operated in shallower wa- ter in 1977-83 than in 1992-97) rather than to fishing pressure. Also, the use of individual back-calculated lengths-at-age for the last annulus only (method 3) rather than the mean back-calculated lengths-at-age for all annuli (weighted mean, method 2, Krug and Haimovici, 1989) could have led to such a difference. Data from the literature indicate considerable vari- ation in length-at-age of bluefish between areas (Ta- ble 5). Terceiro and Ross (1993) attributed this varia- tion to a sampling bias in available fish and to the differential proportion of spawned fishes in collected samples (e.g. due to multiple spawning). Champagnat (1983) suggested that differences in size at first ma- turity could lead to differences in growth parameters. The reported size at first maturity of bluefish varies from 250 mm TL in South Africa (Van der Elst, 1976) Lucena and O'Brien: Effects of gear selectivity and different calculation methods on growth parameters of Pomatomus saltctrix 441 to 430 mm TL in Senegal (Conand, 1975, Champagnat, 1983). In southern Brazil, length at first maturity for blue- fish is between 350 to 400 mm TL (Haimovici and Krug, 1992) and this stock is considered to be relatively fast growing. It is difficult to obtain samples representative of the population’s age structure when using a single gear be- cause of the size selectivity of most gears (Bagenal and Tesch, 1978). Also, the ranges of fish length and age rep- resented in a study may affect the estimation of popu- lation parameters if they are not representative of the stock (Goodyear, 1995). Our study suggests that the data used for the estimation of the growth parameters for a fish stock should allow for the possible effects of gear selec- tivity. Data from different gears should be analyzed and should cover as wide an area of the stock distribution as possible. Moreover, the technique applied for estimation of growth parameters should be carefully assessed and the one chosen, strongly just ified. Verification of the valid- ity of the back-calculation method and validation of the annual nature of increment formation are also strongly recommended. Acknowledgments The authors are grateful to Gladimir Barenho for tech- nical field assistance in Brazil and to Jim Ellis, Richard Millner, Simon Jennings, Paul Kinas, and Verena Tren- kel for their comments on earlier drafts of the paper. The authors are also grateful to Manoel Haimovici for provid- ing additional sample data for our analysis. This study was partially financed by CAPES, Brazil, through a grant to the first author (FML) and by the Ministry of Agri- culture, Fisheries and Food, UK, with funding support (contracts MF0310 and MF0316) provided to the second author (CMO’B). Literature cited Bagenal, T. B., and F. W. Tesch. 1978. Age and growth. In Methods for assessment of fish production in fresh waters (T. B. Bagenal, ed.), p. 101-136. Blackwell Scientific Publications, Oxford. Barger, L. E. 1990. Age and growth of the bluefish Pomatomus saltatrix from the northern Gulf of Mexico and U.S. South Atlantic coast. Fish. Bull. 88:805-809. Bernard, D. R. 1981. Multivariate analysis as a mean of comparing growth in fish. Can. J. Fish. Aquat. Sci. 38: 233-236. Boffo, M., and E. G. Reis. 1997. Estrutura da pesca artesanal costeira no extremo sul do Brasil. In VII Congresso Latino-Americano de Cien- cias del Mar, p. 88-90. Instituto Oceanografico da Uni- versidade de Sao Paulo, Sao Paulo, Brazil. Campana, S. E. 1990. How reliable are growth back-calculation based on otoliths? Can. J. Fish. Aquat. Sci. 47: 2219-2227. Casselman, J. M. 1987. Determination of age and growth. In The biology of fish growth (A. H. Wheatherley and H. S. Gill, eds.), p. 209-242. Academic Press, New York, NY. Champagnat, C. 1983. Peche, biologie et dynamique du tassergal (Pomato- mus saltator , Linnaeus, 1766) sur les cotes senegalomauri- taniennes. Travaux et Documents du l’ORSTOM (Office de la Recherche Scientifique et Technique Outre Mer) 168: 1-279. Conand, C. 1975. Maturite sexuelle et fecondite du tassergal (Poma- tomus saltator ) (L., 1766), Poissons: Pomatomidae. Bull. L’LF.A.N. (l’lnstitut francais de l’Afrique noir), tome 37, ser. A, p. 2. Cushing, D. H. 1968. Fisheries biology. Univ. Wisconsin Press, London, 199 p. Draper, N. R., and H. Smith. 1966. Applied regression analysis. John Wiley and Sons Press, London, 407 p. Francis, R. I. C. C. 1990. Back-calculation offish length: a critical review. Fish. Biol. 3:883-902. Goodyear, C. P. 1995. Mean size at age: an evaluation of sampling strate- gies with simulated red grouper data. Trans. Am. Fish. Soc. 124:454-456. Gutreuter. S. 1987. Consideration for estimation and interpretation of annual growth rates. In Age and growth in fish (R. C. Sommerfelt and R. C. Hall, eds.), p. 115-126. Iowa State Univ. Press, Ames, IA. Haimovici, M., and L. C. Krug. 1992. Alimentagao e reprodu^ao da enchova Pomatomus saltatrix no litoral sul do Brasil. Rev. Bras. Biol. 52(3): 530-513. 1996. Life history and fishery of the enchova Pomatomus saltatrix in southern Brazil. Mar. Freshwater Res. 47: 357-363. Hilborn, R., and C. Walters. 1992. Quantitative fisheries stock assessment: choice, dynamics and uncertainty. Chapman and Hall, New York, NY, 570 p. Hile, R. 1970. Body-scale relation and calculation of growth in fishes. Trans. Am. Fish. Soc. 99:468-474. Johnson, A. G., and C. H. Saloman. 1984. Age, growth, and mortality of gray triggerfish, Bali- ster capriscus, from the northeastern Gulf of Mexico. Fish. Bull. 82(3):485-492. Kolarov, P. 1963. Narastvane na lefera (Pomatomus saltatrix). Izv. Inst. Rib. Varna 3:103-126. Krug, L. C., and M. Haimovici. 1989. Crescimento da enchova Pomatomus saltatrix do sul do Brasil. Atlantica 11(1 ):47 — 6 1 . 1991. A pesca da enchova Pomatomus saltatrix no Sul do Brasil. Anais do simposio da FURG sobre pesquisa pesqueira. Atlantica 13 ( 1 ): 119—130. Lee, R. M. 1920. A review of methods of age and growth determination in fishes by means of scale. Fish. Invest. Lond. Ser. 4:1-32. Lucena, F. M. 1997. Pesca da anchova Pomatomus saltatrix (Pisces: Poma- tomidae) na costa do Rio Grande do Sul: estrutura do estoque e seletividade da rede de emalhar. MSc. thesis, Univ. Rio Grande, Brazil, 153 p. 442 Fishery Bulletin 99(3) Lucena, F. M., and E. G. Reis. 1998. Estrutura e estrategia de pesca da anchova Poma- tomus saltatrix (Pisces: Pomatomidae) na costa do Rio Grande do Sul. Atlantica. 20:87-103. Otter Research. 1996. An introduction to AD model builder for use in non- linear modelling and statistics. Otter Research Ltd, Van- couver, British Columbia, Canada, 40 p. Potts, J., C. S. Manooch III., and D. S. Vaughan. 1998. Age and growth of vermilion snapper from the southeastern United States. Trans. Am. Fish. Soc. 127: 787-795. Reis, E. G. 1992. An assessment of the exploitation of the white croaker Micropogonias furnieri (Pisces, Scianidae) by the arti- sanal and industrial fisheries in coastal waters of southern Brazil. Ph.D. diss., Univ. East Anglia, UK, 219 p. Richards, S. W. 1976. Age, growth and food of bluefish Pomatomus saltatrix from east-central Long Island Sound from July through November 1975. Trans. Am. Fish. Soc. 105:523-525. Ricker, W. E. 1969. Effects of size-selective mortality and sampling bias on estimates of growth, mortality, production and yield. J. Fish. Res. Board Can. 26:479-541. 1992. Back-calculation of fish lengths based on proportion- ality between scale and length increments. Can. J. Fish. Aquat. Sci. 49:1018-1026. Rijnsdorp, A. D., P. I. van Leeuwen., and T. A. M. Visser. 1990. On the validity and precision of back-calculation of growth from otoliths of the plaice, Pleurcmectes platessa L. Fish. Res. 9:97-117. SPSS. 1998. SPSS ( Statistical Package for Social Sciences). SPSS Inc., Chicago, IL. Terceiro, M., and J. L. Ross. 1993. A comparison of alternative methods for the estima- tion of age from length data for Atlantic coast bluefish ( Pomatomus saltatrix). Fish. Bull. 91:534-549. Tesch, F. W. 1971. Age and growth. In Methods for assessment of fish production in fresh waters (W. E. Ricker, ed.), p. 98-130. IBP Handbook 3, Blackwell, Oxford. Van der Elst, R. 1976. Game fish of the east coast of Southern Africa. I. The biology of the elf Pomatomus saltatrix (Linnaeus), in the coastal waters of Natal. South African Assoc. Mar. Biol. Res. Invest. Report 44:1-59. Vieira, P. C., and M. Haimovici. 1993. Idade e crescimento da pescada-olhuda Cynoscion striatus (Pisces, Sciaenidae) no sul do Brasil. Atlantica 15: 73-91. von Bertalanffy, L. 1934. Untersuchungen iiber die Gesetzlichkeiten des Wach- stums. 1. Allgemeine Grundlagen der Theorie Roux Arc. Entwicklungsmech. Org. 131:613-653. Wilk, S. J. 1977. Biological and fisheries data on bluefish, Pomatomus saltatrix (Linnaeus). Tech. Ser. Rep. 11, Sandy Hook Lab., Northeast Fish. Science Cent., Natl. Mar. Fish. Serv., NOAA, Highlands, NJ, 56 p. Wright, P. J., N. B. Metcalfe., and J. E. Thorpe. 1990. Otolith and somatic growth rates in Atlantic salmon parr, Salmo salar L: evidence against coupling. J. Fish. Biol. 36:241-249. 443 Nursery habitats for ladyfish, Elops saurus, along salinity gradients in two Florida estuaries Richard S. McBride Timothy C. MacDonald Richard E. Matheson Jr. David A. Rydene Peter B. Hood Florida Marine Research Institute 100 Eighth Avenue SE St. Petersburg, Florida 33701-5095 E-mail address (for R. S McBride): richard.mcbride@fwc state. fl. us Abstract— Ladyfish, Elops saurus , are recognized as an estuarine-dependent species, although no published study has described how ladyfish use estua- rine habitats. This study found ladyfish to be common throughout Tampa Bay and Indian River Lagoon, Florida. In both estuaries, metamorphosing larvae were collected during several months of the year, but they were most abundant in spring. Length-frequency analyses suggested that age-0 ladyfish grew from 20-30 mm to 200-300 mm standard length during their first year and that at least three age classes were present throughout the year. Age-0 ladyfish fol- lowed an ontogenetic migration with regard to salinity. They entered estu- aries as metamorphosing larvae and became concentrated in waters of lower than median salinity for both estuar- ies (23-25 ppt). In Tampa Bay, which had a greater range of salinity than the Indian River Lagoon, age-0 ladyfish were found principally in mesohaline and oligohaline areas; in the Indian River Lagoon, age-0 ladyfish were found in mesohaline and polyhaline waters. In autumn, age-0 ladyfish moved back to higher salinities, into lower parts of the estuaries, and even out to beaches along the Gulf of Mexico. These field observa- tions are consistent with the hypothesis that ladyfish depend on estuaries, spe- cifically positive estuaries, i.e. where freshwater input exceeds evaporative processes. However, published studies also demonstrate that larval ladyfish can metamorphose and juveniles can survive in hypersaline waters; therefore negative estuaries may also serve as suitable nursery habitat. It is not clear how salinity affects ladyfish growth and mortality, and further research should clarify how different types of estuaries (i.e. positive versus negative) contrib- ute to maintaining populations of this fishery species. Manuscript accepted 19 December 2001 Fish. Bull. 99:443-458 (2001). Many fish species use estuaries of the southeastern United States as nurser- ies (e.g. Skud and Wilson, 1960; Gunter, 1967), but coastal habitat degradation threatens many of the economically important fisheries that rely on estuar- ies ( Gilmore, 1995 ). The re-enacted Mag- nuson-Stevens Fishery Conservation and Management Act (MSFCMA) was developed to protect or enhance fish- eries habitats, by first requiring infor- mation regarding the value of coastal habitats to the survival of marine organisms (e.g. Schmitten, 1996). A simple approach to ranking relative habitat value is to compare intraspe- cific fish distributions with respect to habitat in different estuaries. Associa- tions between abundance and habitat can assist in predicting the response of coastal fish populations to changes in these habitats. Remarkably, such infor- mation is rarely available except for the most economically valuable species (Haedrich, 1983). Freshwater inflows to estuaries of the southeast United States have been severely altered in the last 150 years and coastal development continues to divert more water away from estuaries (Stickney, 1984). The MSFCMA’s Es- sential Fish Habitat mandate provides a policy framework for identifying the effects of reduced freshwater inflows on estuarine-dependent species. Yet re- searchers and managers often charac- terize species as estuarine-dependent more on intuition than rigorous exami- nation of data. Able and Fahay (1998) outlined three criteria for defining es- tuarine-dependence: 1) predictable use of estuaries, 2) non-use of suitable al- ternative habitats, and 3) demonstra- ble effect on a fish population from a loss of estuarine habitat. The first two criteria are best addressed with field studies, but even such simple descrip- tions of habitat use at a landscape level are lacking for most estuarine fish spe- cies (Hoss and Thayer, 1993). One example of this type of data gap is that for ladyfish, Elops saurus, a fish- ery species that inhabits coastal waters of Florida (Hildebrand, 1963; Murray et ah, 1987; FMRI1). Ray (1997) listed la- dyfish as an estuarine-dependent spe- cies but little is known about its use of habitat in coastal waters. Ladyfish are probably considered to be estua- rine-dependent because they spawn in offshore waters and metamorphosing larvae and juveniles are found inshore (Hildebrand, 1943; Gehringer, 1959; El- dred and Lyons, 1966). However, la- dyfish tolerate a wide range of salini- ties (Alikunhi and Rao, 1951; Gunter, 1956; Gehringer, 1959; Bayly, 1972). Numerous studies have reported the occurrence of small juveniles in ineso- haline or lower-salinity waters (<18 ppt), which is consistent with an estua- rine-dependent life history (Tagatz and Dudley, 1961; Gunter and Hall, 1965; Zilberberg, 1966; Tagatz, 1968; Tagatz and Wilkens, 1973; Sekavec, 1974; Gov- oni and Merriner, 1978 [and references within]; Thompson and Deegan, 1982; Peterson and Ross, 1991). Moreover, large ladyfish are present throughout 1 FMRI ( Florida Marine Research Institute ): www.floridamarine.org 444 Fishery Bulletin 99(3) A B 82°50' 82°40' 82°30' 82°20' 81°00' 80°50' 80°40' 80°30' 80°20' Figure 1 Sampling areas and locations where at least one ladyfish, Elops saurus, was collected for (A) Tampa Bay and (B) the Indian River Lagoon. Different symbols are used to represent collections in Tampa Bay: the bay-wide program (circles), the Little Manatee River program (triangles), and the Gulf Beach program (squares). Lines and letters (A-F) designate the sampling zones in each system. See text for further definition of zones and Tables 1-2 for details of all collections. the year in polyhaline and marine ( >18 ppt) waters such as Florida Bay, Biscayne Bay, and the Indian River La- goon (Low, 1973; Sogard et al., 1989a, 1989b; Tremain and Adams, 1995). Other studies, however, have reported presumptive age-0 juveniles (Harrington and Harrington, 1961; Kristensen, 1964), older subadults (Carles, 1967), or both (Simmons, 1957; Roessler, 1970; Brockmann, 1974) in hypersaline waters ( >35 ppt). Thus, age-0 ladyfish may be seeking some condition other than specific salinities. The purpose of our study was to determine whether age-0 ladyfish make predictable size-specific seasonal movements that would identify them as an estuarine-de- pendent species. Because salinity is part of the definition of an estuary (Pritchard, 1967), a landscape approach was used to follow ladyfish movements with respect to salinity within Tampa Bay and the Indian River Lagoon. If age-0 ladyfish select low-salinity waters, then we predicted that they would be found in lower than average salinity wa- ters for each estuary. Furthermore, if age-0 ladyfish select low-salinity habitats, then we predicted that they would be more abundant in the lower salinity areas of Tampa Bay because this bay has a wider salinity range to select than that for the Indian River Lagoon. We also examined length-frequency data to make preliminary assessments of age-class composition and growth rates of ladyfish with- in estuaries. Although several publications have discussed the ecology of ladyfish in estuaries (e.g. Springer and Woodburn, 1960 and citations above), to our knowledge the data treatment in our study is the most comprehen- sive for this species. Methods Study sites Data from Tampa Bay, on central Florida’s west coast, were compared with data from the Indian River Lagoon, on central Florida’s east coast (Fig. 1). Both water systems are located at similar latitudes, so they are subject to simi- lar temperature cycles. Both systems are positive estuar- ies (i.e. freshwater input exceeds evaporative processes; Pritchard, 1967), but they have distinctive salinity regimes (Fig. 2). Tampa Bay is a drowned river system with con- siderable freshwater input and separate satellite barrier- island embayments, whereas the Indian River Lagoon is largely a series of barrier-island embayments with few inlets and no major rivers (Comp and Seaman, 1985). Bay-wide survey— Tampa Bay We examined survey data for Tampa Bay fishes collected from 1989 to 1995 by staff of the Florida Marine Research Institute (FMRI). The sam- pling design incorporated both fixed sampling stations and locations assigned in a stratified random manner (McMi- chael et al., 1995). Tampa Bay was stratified into six major zones (Fig. 1A): the upper bay (A), the western bay (B), McBride et al.: Nursery habitats for Elops saurus 445 the eastern bay (C), the satellite embayments (D), the lower bay (E), and the principal rivers (F). Each zone was stratified further by a 1-square-nautical-mile grid system. Sampling at fixed stations occurred monthly, but stratified random sampling was conducted principally in spring and autumn when many species recruit to Florida estuaries. Sampling gear included seines, trawls, block nets, and gill nets (Table 1). A 21.3-m center-bag seine was deployed in one of three different ways: 1) it was hauled along the shoreline across an area of about 340 m2 (“beach sets”); 2) it was deployed from a boat in a semicircular pattern in river zones and the mean area swept was 70 m2 (“boat sets”); 3) or it was set away from the shoreline and hauled into the current across an area of about 140 m2 (“offshore sets”). A much larger seine (183-m) was also deployed from a boat in a semicircular pattern. A 61-m block net was set against seawalls or mangroves of inundated shorelines at high tide, and fish were collected at the ensuing low tide. Otter trawls were towed for 10 min in bay zones (zones A-E) and for 5 min in river zones (zone F) at an approximate speed of 0.6 m/s. A 184-m gill net with four 46-m panels (75-, 100-, 125-, and 150-mm mesh) and a similar 198-m gill net that included a 15-m section of 50-mm mesh were used. These gill nets were set perpendicular to shore, four nets at a time, so that two nets with the larger mesh were oriented inshore and two nets were oriented in the opposite direction. Little Manatee River— Tampa Bay In addition to sampling the Little Manatee River as part of the bay-wide survey, FMRI staff completed an independent survey of this river 446 Fishery Bulletin 99(3) Table 1 Sources and details of data examined from Tampa Bay: sampling gear used (with mesh size); geographic zones sampled (see Fig. 1A); diel periods (D=day, N=night, C=crepuscular); months (l=January); years sampled during 1989-95; number of ladyfish collected (C (cl7); and total number of sets with each net (f). The bay-wide survey of Tampa Bay included a stratified, random survey design and a fixed-station sampling design. The special survey of the Little Manatee River and the Gulf Beach survey used fixed stations. Gear Mesh (mm) Zones Period Months Years C (c)4 f Bay-wide fixed-station survey 21.3-m seine (340 m2)7 3.2 A, (B-D),2 F D 1-12 1989-95 60 377 21.3-m seine ( 70 m2)7 3.2 F D 1-12 1989-95 942 (6) 1818 21.3-m seine (140 m2)7 3.2 A-E D, N, C2 1-12 1989-95 22 2081 183-m seine 38.5 B-F D 1-12 1992,2 1993-95 1585 458 61-m block net 3.2 D D, N, C2 1-12 1990 2 1991-92 148 210 6.1-mtrawl 3.23 Bay-wide, stratified random station survey B-F D 1-12 1989 2 1990-95 32(2) 2249 21.3-m seine (340 m2)7 3.2 A-E,F2 D, N, C 3-6, 9-12 1989-95 35 1146 21.3-m seine (70 m2)7 3.2 (A-E ),2 F D, N, C 3-6, 9-11, 122 1989-95 31 639 21.3-m seine (140 m2)7 3.2 A-F D, N, C 3-5, 6,2 9-12 1989-95 16 1464 6.1-m otter trawl 3.23 A-F D, N, C 3-6, 9-12 1989-95 3(5) 2538 184-m gill net 75-150 A-E, F2 N, C 3-6, 9-12 1989-93 952 427 198-m gill net 50-150 A-E N, C 3-5,9-11 1994-95 583 160 Little Manatee River survey 9.1-m seine 3.2 — D 1-12 1989-91 15 77 22.9-m seine 3.2 — D 1-12 1988-91 115 1460 120-m seine 3.2 — D 1-12 1990-91 148 (1) 89 Gulf Beach survey 22.9-m seine Total 3.23 D 1-12 1992-94 156 4843(14) 435 1 Value in parentheses indicates area swept by each haul; see text for further details. 2 Less than 30 tows for this sampling unit. 3 Minimum mesh size used. 4 C = late-metamorphic, juvenile, and older stages; (c) = early- or mid-metamorphic (leptocephalus) larvae. between January 1988 and December 1991 (Table 1; Fig. 1A). Samples were collected biweekly with a 22.9-m seine at six fixed shoreline stations located between the river mouth and the freshwater zone, with 2-3 seine hauls per station. Supplemental samples were collected with 9.1- and 22.9-m seines at two additional stations from January 1989 to June 1991 and with a 120-m seine at five fixed sites near the river mouth from March 1990 to November 1991. Gulf beaches— Tampa Bay In a third independent survey by FMRI staff, two beach sites were sampled along the Gulf of Mexico coast of Pinellas County, FL (Table 1; Fig. 1A). Samples were collected biweekly with a 22.9-m seine from September 1992 to November 1994 at Treasure Island and from August 1993 to November 1994 at Indian Shores. Five hauls were made in the surf zone at each site during a single day; each haul began 50 m from the water’s edge and proceeded perpendicular to shore. Lagoon-wide Survey— Indian River Lagoon We also exam- ined data from an FMRI survey of the Indian River Lagoon fishes. The same general sampling design and gear were used in this survey and the bay-wide survey of Tampa Bay (see above; Tables 1-2; Tremain and Adams, 1995). Although the Indian River Lagoon survey started slightly later than the Tampa Bay survey, they were largely con- temporaneous (1990-1995). The northern Indian River Lagoon system is a complex of the Indian River basin (zones A-C) and the Banana River basin (zones D-E; Fig. IB). The sampling program in the Indian River Lagoon dif- fered slightly from that in Tampa Bay. In the Indian River Lagoon, neither the 183-m seine nor block nets were used; gill nets were used at fixed stations and stratified-random locations, and fewer total hauls were made with most types of sampling gear because portions of this program started one or two years later than Tampa Bay’s program. Abundance and size of ladyfish Monthly relative abundance was calculated as the mean number of ladyfish per haul (including hauls with no lady- fish) for each gear type used in each survey. Data from the stratified random programs were not included in cal- culations of monthly relative abundance because stratified McBride et al.: Nursery habitats for Elops saurus 447 Table 2 Sources and details of data examined from the Indian River Lagoon: sampling gear used (with mesh size); geographic zones sampled (see Fig. IB); diel periods (D=day; N=night; C=crepuscular); months ( l=January); years sampled during 1989-95; number of ladyfish collected (C (cF); and total number of sets with each net (f). Gear Mesh (mm) Zones Period Months Years C (c)4 f Lagoon-wide fixed-station survey 21.3-m seine (340 m2)7 3.2 A, C-D D 1-12 1991-95 273 (367) 619 21.3-m seine ( 70 m2)7 3.2 B 2 C, E2 D ( 1-12)2 1991-95 403(67) 235 21.3-m seine (140 m2)7 3.2 A, C-E D, C2 1-12 1991-95 42 (96) 668 6.1-m trawl 3.23 C-E D 1-3 (4-5), 2 6-12 1991, 2 1992-95 7(1) 617 184-m gill net 75-150 A, D C (1-12)2 1991-93, 19942 95 202 198-m gill net 50-150 A-D C, N2 (1-12)2 1994-95 153 126 Lagoon-wide stratified, random station survey 21.3-m seine (340 m2)7 3.2 A-E D, N,C 3-5,9-11 1990-95 232 (218) 977 21.3-m seine ( 70 m2)7 3.2 D2 D,2 C2 52 19902 0 6 21.3-m seine (140 m2)7 3.2 A-E D, N,C 3-5, 9-11 1990-95 46 (61) 1286 6.1-m otter trawl 3.23 A-E D, N, C 3-5,9-11 1990-95 1 (5) 1501 184-m gill net 75-150 A-E N, C 3-5,9-11 1990-93 392 319 198-m gill net 50-150 A-E N, C 3 2 4, 5 2 9, 2 10, 2 11 1994-95 457 168 Total 2101 (815) 1 Value in parentheses indicates area swept by each haul; see text for further details. 2 Less than 30 tows for this sampling unit. 3 Minimum mesh size used. 4 C = late-metamorphic, juvenile, and older stages; (c) = early- or mid-metamorphic (leptocephalus) larvae. random sampling did not occur in all months of the year. Geographic distributions of age-0 ladyfish were also plot- ted for each of four seasons. For such maps, relative abun- dance was calculated as the number of age-0 ladyfish per haul. Values for all positive catches were categorized into four quartile classes (<25th, 26-50th, 51-75th, and >75th per- centile) before plotting, to standardize the data among gear types and estuary. Maximum size criteria were adjusted for each season to exclude fish older than age 0. Fish size was measured and reported as standard length (SL) in mm. At least 20 randomly selected ladyfish per sample were measured and unmeasured fish were pro- portionally adjusted for the length-frequency plots to re- flect the size structure of the entire sample! s). Sizes from the literature are also reported as SL and were converted when necessary by using the equations SL = -0.772 + 0.787 (total length), or SL = -2.46 + 0.943 (fork length)', each equation was based on least squares regressions of measurements from 75 ladyfish that ranged from 39 to 475 mm SL (r2=0.99). Mean salinity at capture was cal- culated for each 25-mm-SL interval, and for each estuary separately, by using the weighted formula Y( = the salinity measured for collection i; and n = the total number of collections with fish in that 25-mm-SL interval for that estuary. We identified early life stages of ladyfish from their size and appearance. The following criteria and general termi- nology are from Gehringer (1959). Early-metamorphic lar- vae have a leptocephalus form, with a clear and laterally compressed body. Early-metamorphic larvae shrink as they develop from about 45 to 25 mm SL. Mid-metamorphic lar- vae are generally less than 25 mm SL. Mid-metamorphic lar- vae shrink to about 18 mm SL and then grow to about 25 mm SL; they lose the leptocephalus form by the end of this stage. Late-metamorphic larvae have a juvenile form and grow from about 25 mm to 60 mm SL. After 60 mm the age-0 fish are referred to as juveniles. Older age classes (i.e. age-1 or age-2+) are defined by size, as inferred from length-fre- quency analysis. The complete size and age range of ladyfish in estuaries is not defined in Gehringer or other published literature but appears to be largely restricted to immature fish. Y. - Results Tampa Bay where wt = the number of ladyfish per 25-mm-SL inter- val for collection i\ In the bay-wide survey, 4422 ladyfish were captured in 7525 21.3-m-seine hauls, 458 183-m-seine hauls, 210 448 Fishery Bulletin 99(3) Figure 3 Monthly length-frequency of ladyfish, Elops saurus, in Tampa Bay. Early- and mid-metamorphic larvae were excluded because length decreases as they develop. Bay-wide data for all years (1989-95) from all zones (A-F) are plotted, n = number of fish sampled. blocknet sets, 4787 trawl hauls, and 587 gillnet sets (Table 1). Generally, a new cohort of late-metamorphic larvae first appeared in April (Figs. 3 and 4A). Only 13 early- or mid-metamorphic larvae were collected (26-34 mm SL) in the bay-wide survey {n = 14 for all Tampa Bay samples). Although the arrival of metamorphosing larvae (largely late-metamorphic stages) was concen- trated in April, isolated individuals 26-41 mm SL were collected as early as February (1989) or March (1994, 1995) or as late as September (1994), October (1993), or McBride et al.: Nursery habitats for Elops saurus 449 03 SZ Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 4 Monthly catch-per-unit-of-effort (mean number per haul ±2 standard error bars) for ladyfish, Elops saurus, collected with 21-m and 23-m seines at fixed stations during 1989-1995 in Tampa Bay or 1990-95 in the Indian River Lagoon. (A) Little Manatee River (early- and mid-metamorphic larvae are excluded); (B) Gulf of Mexico beaches; (C) Indian River Lagoon leptocephali (i.e. only early- and mid-met- amorphic larvae); (D) Indian River Lagoon late-metamorphic and juvenile ladyfish (i.e. early- and mid-metamorphic larvae excluded). November (1990). All ladyfish ranged in length from 20 to 550 mm SL (Fig. 5). Length-frequency analyses sug- gested that there were three age classes: age-0 fish over- wintering at 200-300 mm SL, age-1 fish overwintering at 300-400 mm SL, and age-2+ at sizes of 400 mm SL or larger (Fig. 3). Nearly all ladyfish <200 mm SL were collected in meso- haline and oligohaline areas of rivers (zone F; Fig. 6). Ear- ly- and mid-metamorphic larvae were collected at poly- haline salinities (mean=20 ppt, Fig. 7); these larvae had probably entered Tampa Bay recently. In Little Manatee River, ladyfish were most common (59.7% of the total in- dividuals) in the mesohaline zone (5.1-18 ppt), less com- mon in the oligohaline zone (31.9%; 0. 5-5.0 ppt), uncom- mon in the polyhaline zone (6.7%; 18.1-30 ppt), and rare in fresh water (1.7%; <0.5 ppt). Abundance in oligohaline waters peaked during June, about one month later than peak abundance in mesohaline waters. 450 Fishery Bulletin 99(3) During autumn, when age-0 fish were <300 mm SL, they began to move out of the rivers (Fig. 8). Fish mea- suring 170-360 mm SL (5-95 percentile; n= 55 fish) were found at the mouth of the Little Manatee River during the winter. Large concentrations of age-0 ladyfish were also found at other river mouths, near the dredged (>12 m) portions of the Alafia River, and near the mouth of Tam- pa Bay. There was a sudden pulse of ladyfish along Gulf of Mexico beaches during September-October (Fig. 4B). These were juvenile age-0 ladyfish measuring 219-260 mm SL (5-95 percentile; ft =126). Only 2 of 158 individu- als collected at Gulf beaches were smaller than 178 mm SL. The absence of larval ladyfish in shallow Gulf beach habitats contrasted strongly with the abundance of larval ladyfish in riverine habitats of Tampa Bay. Indian River Lagoon A total of 2916 ladyfish were captured in 3791 21.3-m seine hauls, 2118 trawl hauls, and 815 gillnet sets (Table 2). The seasonal pattern of arrival was similar to that in Tampa Bay. Early- and mid-metamorphic larvae were much better represented (28% or n=815) in the Indian River Lagoon than in Tampa Bay (0.2% or n=14), and these leptocephali were present from at least December to May (Fig. 4C). Late-metamorphic larvae were found in many months throughout the year, but they were most abundant during spring (Figs. 4D and 9). All ladyfish col- lected in the Indian River Lagoon ranged from 20 to 600 mm SL. Age-0 fish reached a modal length during winter (i.e. 250-270 mm SL), similar to that observed in Tampa Bay. Age-1 ladyfish appeared to overwinter at a mode of about 350 mm SL, and a small number of age-2+ were probably present as postulated for Tampa Bay. Ladyfish <100 mm SL were more common in the Indian River basin than in the Banana River basin, and fish >300 mm SL were more common in the Banana River basin than in the Indian River basin (Fig. 10). Juveniles first occupied upper mesohaline and lower polyhaline salinities at sizes of about 75 mm SL, and they remained at such salinities until reaching about 300 mm SL (Fig. 7). In the Indian Riv- er Lagoon, the mean salinity occupied by ladyfish 75-125 mm SL was about 5 ppt higher than in Tampa Bay, but mesohaline and oligohaline habitats are less common in McBride et al.: Nursery habitats for Elops saurus 451 Cl) cr- Upper Tampa Bay Zone A (n= 326) Zone B (n=345) l\ \ l Zone C (n=621) 0 50 100 150 200 250 300 350 400 450 500 550 600 Standard length (mm) Figure 6 Length frequency of ladyfish, Elops saurus, plotted by sampling zones in Tampa Bay. Early- and mid-metamorphic larvae were excluded. Data for all years (1989-95) of bay-wide sampling were plotted, n = number of fish sampled. Indian River Lagoon than in Tampa Bay. Overall, however, ladyfish were broadly distributed in both the Indian River and Banana River basins throughout the year (Fig. 11). Discussion Ladyfish were common, widely distributed, and fast grow- ing in both Tampa Bay and the Indian River Lagoon. Metamorphosing larvae and overwintering juveniles were linked together in a single study by using multiple sam- pling gears. Ladyfish ages inferred from length frequencies indicated that few fish older than 2-3 years were present in either estuary. Over a thousand ladyfish gonads from Tampa Bay and the Indian River Lagoon were examined macroscopically, and nearly all fish were found to be imma- ture (McBride, pers. obs.). Overall, our observations agree with previous reports that ladyfish arrive in coastal embay- ments as metamorphosing larvae and leave after about 2-3 years to mature and eventually spawn at sea. Carles’ ( 1967) samples from a hypersaline Cuban lagoon contained only immature age 1-3 fish (115-375 mm SL). Others who have suggested that ladyfish mature at sea, where they reach a maximum age of about 6 years and a maximum length of 570-660 mm SL, include Hildebrand (1963), Palko (1984), and Santos-Martinez and Arboleda (1993). 452 Fishery Bulletin 99(3) Standard length (mm) Figure 7 Density-weighted mean salinity at capture for different sizes and stages of ladyfish, EIops saurus, in Tampa Bay (1989-95) and Indian River Lagoon (1990-95). Early- and mid-metamorphic larvae are indicated as “Lepto.” Error bars represent 2 standard errors. The solid horizon- tal line represents the median salinity value, and the dotted horizontal lines represent the 10th and 90th salinity percentiles, based on all sam- ples (with or without ladyfish) from each system. Ladyfish in both Florida estuaries followed a similar se- quence of size-specific movements with respect to salin- ity. Soon after metamorphosis and for much of their first year, ladyfish occupied waters that were lower than the median salinity of each estuary. Following their first win- ter, ladyfish occupied higher than median salinities. The available salinity range was wider in Tampa Bay and la- dyfish were found in a wider range of salinity in Tampa Bay than ladyfish in the Indian River Lagoon. Some de- tails of these movements deserve further investigation. For example, as juveniles grew to about 150 mm SL during summer they were found in progressively lower salinities. However, they may have passively remained in shallow- water habitats where salinity was decreasing because of summer rains, instead of actively selecting lower salinity waters. We also did not examine other factors that may co- vary with salinity; therefore we cannot rank the effect of salinity in relation to other variables. In one such simul- taneous analysis, Friedland et al. (1996) showed that juve- nile menhaden ( Brevoortia tyrannus ) select waters of high chlorophyll-a levels more consistently than a specific sa- linity value. Temperature presumably affects ladyfish distributions as well. During colder than average winters, ladyfish have experienced hypothermal mortality in both Tampa Bay (Springer and Woodburn, 1960) and the Indian River La- goon (Snelson and Bradley, 1978). We observed ladyfish moving into deeper water, to river mouths, to the lower McBride et al.: Nursery habitats for Elops sourus 453 o 1 - 25th percentile O 26- 50th percentile O 51 - 75th percentile O 76 - 100th percentile Figure 8 Seasonal distribution plots for age-0 ladyfish, Elops saurus, collected in Tampa Bay, 1989-95. Catch-per-unit-of-effort data from all positive catches of age-0 ladyfish were plotted as four quartile classes to standardize the data among gear types; increasing symbol size indicates higher catch per unit of effort. Smaller-size, age-1 ladyfish were excluded by increasing the size range for age-0, as indicated in each plot, from spring to winter. part of Tampa Bay, and out into Gulf of Mexico waters during colder months. Ladyfish may be selecting these ar- eas for overwintering because such areas are less affected by atmospheric cold fronts. The critical temperatures for ladyfish survival have not been examined in the labora- tory, but the above field studies suggest that they could be as low as 10°C. Further work remains on the processes of habitat selection by ladyfish within estuaries. This description of ladyfish early life history satisfies the first two of three predictions for an estuarine-depen- dent species: 1) predictable use of estuarine habitats, 2) absence of fish in suitable alternative habitats, and 3) demonstration of a negative population impact caused by the loss of estuarine habitats. Age-0 ladyfish followed an ontogentic migration along a salinity gradient and lar- val stages of ladyfish were absent from Gulf beaches. The third prediction, that of a negative impact on the popula- tion by the absence of habitat, was not supported by find- ings in our study. We anticipated that salinity differences between Tampa Bay and the Indian River Lagoon might affect growth rates, either due to osmoregulatory stress or perhaps to other correlated factors. Metamorphosing larvae moved into both estuaries at about the same time (i.e. in spring) but age-0 fish attained similar lengths 454 Fishery Bulletin 99(3) Figure 9 Monthly length frequency of ladyfish, Elops saut'us, in the Indian River Lagoon. Early- and mid-metamorphic larvae were excluded because length decreases as they develop. Data from all years (1990-95), all gear types, and sampling zones (A-E) were plotted, n = number of fish sampled. (250-300 mm SL) by winter. In Florida Bay, age-0 lady- fish entered estuaries at similar stages during spring and reached similar overwintering sizes as well (Roess- ler, 1970). In constrast, Carles (1967) estimated much slower growth rates for ladyfish residing in hypersaline lagoons of Cuba, with sizes at annulus formation of 130 mm SL at annulus I, 195 mm SL at annulus II, and 247 mm SL at annulus III. If the growth increments observed by Carles (1967) are indeed annual (he used unvalidated scale annuli), then hypersaline ( >35 ppt) conditions may McBride et al.: Nursery habitats for Elops saurus 455 Indian River Zone A (n=346) I I Zone B (n= 358) L \ ~H Zone C (n=814) 50 100 150 200 250 300 350 400 450 500 550 600 Standard length (mm) Figure 10 Length frequency of ladyfish, Elops saurus , plotted by sampling zones in the Indian River Lagoon. Banana River and Indian River basins are identified in Figure IB. Early- and mid-metamorphic larvae were excluded. Data from all years (1990-95) and gear types were plotted. n = number of fish sampled. reduce ladyfish growth rates. The timing of ingress into the Cuban estuaries by metamorphosing larvae was not reported by Carles (1967); therefore it is not clear if over- wintering sizes are comparable (i.e. if they represent the same seasonal growth period). If ingress into the Cuban estuaries occurred later than spring (i.e. when ladyfish entered Florida estuaries), then this would shorten the length of the growing season and could explain Carles’ (1967) results. This preliminary attempt to link salinity to growth rate, although suggestive, requires verification of reduced growth rates in hypersaline conditions. Exper- imental studies to determine the optimal salinity for la- dyfish growth and survival will clarify whether ladyfish benefit by actively selecting low-salinity habitats. Such information would be the last step in demonstrating that ladyfish depend on estuaries and for showing the rela- tive value of positive estuaries (with low salinity areas) to negative estuaries (with hypersaline conditions) for la- dyfish populations. It would also be the next step for de- fining the essential fish habitat of ladyfish and for pre- dicting the effects of changes in estuarine salinity on this fishery species. Acknowledgments Sampling throughout Tampa Bay and the Indian River Lagoon was supported in part by funding from the Depart- ment of the Interior, U.S. Fish and Wildlife Service, Fed- eral Aid for Sport Fish Restoration, project F-43. Sampling was also funded in part by sales of Florida’s saltwater fishing licenses and special State of Florida appropria- tions to the Florida Department of Environmental Protec- tion (FDEP). Sampling in the Little Manatee River was made available through grants CM-254 and CM-280 from the Department of Environmental Regulation, Office of Coastal Management, with funds made available through the National Oceanic and Atmospheric Administration under the Coastal Zone Management Act of 1972, as amended. Sampling of Gulf of Mexico beaches was funded 456 Fishery Bulletin 99(3) o 1 - 25th percentile O 26- 50th percentile O 51- 75th percentile O 76- 100th percentile Figure 11 Seasonal distribution plots for age-0 ladyfish, Elops saurus, collected in the Indian River Lagoon, 1991-95. Banana River and Indian River basins are identified in Figure IB. Catch-per-unit-of-effort data from all positive catches of age-0 ladyfish were plotted as four quartile classes to standardize the data among gear types; increasing symbol size indicates higher catch per unit of effort. Smaller-size, age-1 ladyfish were excluded by increasing the size range for age-0, as indicated in each plot, from spring to winter. by special State of Florida appropriations to the FDEP. Dozens of anonymous and not so anonymous hands helped collect, sort, and identify fish. Valuable discussions or com- ments on earlier drafts were provided by R. Crabtree, K. Guindon-Tisdel, S. Kaiser, J. Levesque, K. Peters, R. Ruiz- Carus, and D. Winkelman. Editorial assistance was pro- vided by L. French, J. Leiby, and J. Quinn. We thank all of the above. McBride et al.: Nursery habitats for Elopssciurus 457 Literature cited Able, K. W., and M. P. Fahay. 1998. The first year in the life of estuarine fishes in the Middle Atlantic Bight. Rutgers Univ. Press, New Bruns- wick, NJ, 342 p. Alikunhi, K. H., and S. N. Rao. 1951. Notes on the metamorphosis of Elops saurus Linn, and Megalops cyprinoides (Broussonet) with observations on their growth. J. Zool. Soc. India 3:99-109. Bayly, I. A. E. 1972. Salinity tolerance and osmotic behavior of animals in athalassic saline and marine hypersaline waters. Annu. Rev. Ecol. Syst. 3:233-263. Brockmann, F. W. 1974. Seasonality offishes in a south Florida brackish canal. FI. Sci. 37:65-70. Carles, C. A. 1967. Datas sobre la biologia del banano Elops saurus Lin- naeus (Teleostomi: Elopidae). Inst. Nac. de la Pesca Cuba, Contribution 27, 53 p. Comp, G. S., and W. Seaman Jr. 1985. Estuarine habitat and fishery resources of Florida. In Florida aquatic habitat and fishery resources (W. Seaman Jr., ed.), p. 337-435. Florida Chapter of the American Fisheries Society, Kissimmee, FL. Eldred, B., and W. G. Lyons. 1966. Larval ladyfish, Elops saurus Linnaeus 1766, (Elopi- dae) in Florida and adjacent waters. FI. Board Conserv. Mar. Res. Lab. Leafl., ser. IV, part 1 (Pisces), no. 2. Friedland, K. D., D. W. Ahrenholz, and J. F. Guthrie. 1996. Formation and seasonal evolution of Atlantic men- haden juvenile nurseries in coastal estuaries. Estuaries. 19:105-114. Gehringer, J. W. 1959. Early development and metamorphosis of the ten- pounder Elops saurus Linnaeus. Fish. Bull. 155:618-647. Gilmore, R. G. 1995. Environmental and biogeographic factors influencing ichthyofaunal diversity: Indian River Lagoon. Bull. Mar. Sci. 57:153-170. Govoni, J. J., and J. V. Merriner. 1978. The occurrence of ladyfish, Elops saurus, larvae in low salinity waters and another record for Chesapeake Bay. Estuaries 1:205-206. Gunter, G. 1956. A revised list of euryhaline fishes of North and Middle America. Am. Midi. Nat. 55:345-54. 1967. Some relationships of estuaries to the fisheries of the Gulf of Mexico. In Estuaries (G. H. Lauff, ed.), p. 621- 638. Publication 8 of AAAS (Am. Assoc. Advancement Sci. ), Washington, D.C. Gunter, G., and G. E. Hall. 1965. A biological investigation of the Caloosahatchee estu- ary of Florida. Gulf Res. Rep. 2:1-71. Haedrich, R. L. 1983. Estuarine Fishes. In Estuaries and enclosed seas: ecosystems of the world (B. H. Ketchum, ed.), vol. 26, p. 183-207. Elsevier, New York, NY. Harrington, R. W. Jr., and E. S. Harrington. 1961 . Food selection among fishes invading a high subtropi- cal salt marsh: from onset of flooding through the progress of a mosquito brood. Ecology. 42:646-666. Hildebrand, S. F. 1943. Notes on the affinity, anatomy, and development of Elops saurus Linnaeus. J. Wash. Acad. Sci. 33:90-4. 1963. Family Elopidae: fishes of the western North Atlan- tic. Mem. Sears Found. Mar. Res. Yale Univ., part 3: 123-131. Hoss, D. E„ and G. W. Thayer. 1993. The importance of habitat to the early life history of estuarine dependent fishes. Am. Fish. Soc. Symp. 14: 147-158. Kristensen, I. 1964. Hypersaline bays as an environment of young fish. Proc. Gulf Caribb. Fish. Inst. 63:139-142. Low, R. A., Jr. 1973. Shoreline grassbed fishes in Biscayne Bay, Florida, with notes on the availability of clupeid fishes. M.S. thesis, Univ. Miami, Coral Gables, FL, 145 p. McMichael, R. H. Jr., R. Paperno, B. J. McLaughlin, and M. E. Mitchell. 1995. Florida’s marine fisheries-independent monitoring program: a long-term ecological dataset. Bull. Mar. Sci. 57:282. Murray, J. D., J. C. Johnson, and D. C. Griffith. 1987. Encouraging the use of underutilized marine fishes by southeastern U.S. anglers. Part II: Educational objec- tives and strategy. Mar. Fish. Rev. 49:138-142. Palko, B. J. 1984. An evaluation of hard parts for age determination of pompano ( Trachinotus earolinus), ladyfish ( Elops saurus), crevalle jack ( Caranx hippos ), gulf flounder ( Paralichthys albigutta ), and southern Flounder (Paralichthys letho- stigma). U.S. Dep. Commer., NOAA Tech. Memo NMFS- SEFC 132, 11 p. Peterson, M. S., and S. T. Ross. 1991. Dynamics of littoral fishes and decapods along a coastal river-estuarine gradient . Estuar. Coast. Shelf Sci. 33:467-483. Pritchard, D. W. 1967. What is an estuary: physical viewpoint? In Estuar- ies (G. H. Lauff, ed.), p. 3-5. Publication 83 of AAAS. (Am. Assoc. Advancement Sci.), Washington, D.C. Ray, G. C. 1997. Do the metapopulation dynamics of estuarine fishes influence the stability of shelf ecosystems? Bull. Mar. Sci. 60:1040-1049. Roessler, M. A. 1970. Checklist of fishes in Buttonwood Canal, Everglades National Park, Florida, and observations on the seasonal occurrence and life histories of selected species. Bull. Mar. Sci. 20:860-93. Santos-Martinez, A., and S. Arboleda. 1993. Aspectos biologicos y ecologicos del macabi Elops saurus Linnaeus (Pisces: Elopidae) en la Cienaga Grande de Santa Marta y costa adyacente, Caribe Colombiano. An. Inst. Investig. Mar. Punta Betin. 22:77-96. Schmitten, R. A. 1996. National Marine Fisheries Service: seeking partners for its National Habitat Plan and identifying essential fish habitats. Fisheries 21( 12 ):4. Sekavec, G. B. 1974. Summer foods, length-weight relationship, and con- dition factor of juvenile ladyfish, Elops saurus Linnaeus, from Louisiana coastal streams. Trans. Am. Fish. Soc. 103:472-476. Simmons, E. G. 1957. An ecological survey of the upper Laguna Madre of Texas. Inst. Mar. Sci. 4:156-200. Skud, B. E., and W. B. Wilson 1960. Role of estuarine waters in Gulf fisheries. In Trans. 458 Fishery Bulletin 99(3) 25th North Am. Wildlife Conf. March 8-19, 1960, 320-326 p. Wildlife Management Inst., Washington, D.C. Sogard, S. M., G. V. N. Powell, and J. G. Holmquist. 1989a. Utilization by fishes of shallow, seagrass-covered banks in Florida Bay: 1. Species composition and spatial heterogeneity. Env. Biol. Fish. 24:53-65. 1989b. Spatial distribution and trends in abundance of fishes residing in seagrass meadows on Florida Bay mud- banks. Bull. Mar. Sci. 44:179-199. Snelson, F. F., Jr., and W. K. Bradley Jr. 1978. Mortality of fishes due to cold on the east coast of Florida, January, 1977. Fla. Sci. 41:1-12. Springer, V. G., and K. D. Woodburn. 1960. An ecological study of the fishes of the Tampa Bay area. Fla. Board Conserv. Mar. Res. Lab. Prof. Pap. Ser. 1, 104 p. Stickney, R. R. 1984. Estuarine Ecology of the Southeastern United States and Gulf of Mexico. Texas A & M Univ. Press, College Sta- tion, TX, 310 p. Tagatz, M. E. 1968. Fishes of the St. Johns River, Florida. Q. J. FI. Acad. Sci. 30:25-50. Tagatz, M. E., and D. L. Dudley. 1961. Seasonal occurrence of marine fishes in four shore habitats near Beaufort, N.C., 1957-1960. U.S. Fish Wildl. Spec. Sci. Rep. Fish. 390, 19 p. Tagatz, M. E., and E. P. H. Wilkens. 1973. Seasonal occurrence of young Gulf menhaden and other fishes in a northwestern Florida estuary. U.S. Dep. Commer., NOAA Technical Report NMFS SSRF 672, 14 p. Thompson, B. A., and L. A. Deegan. 1982. Distribution of ladyfish ( Elops saurus ) and bonefish (Albula vulpes ) leptocephali in Louisiana. Bull. Mar. Sci. 32:936-939. Tremain, D. M., and D. H. Adams. 1995. Seasonal variations in species diversity, abundance, and composition of fish communities in the northern Indian River Lagoon, Florida. Bull. Mar. Sci. 57:171-192. Zilberberg, M. H. 1966. Seasonal occurrence of fishes in a coastal marsh in northwest Florida. Publ. Inst. Mar. Sci. Univ. Tex. 11:126- 134. 459 Abstract— Oxytetracycline (OTC) in- jections were used in a mark-recap- ture experiment undertaken to validate the fin-ray method of age determina- tion of lingcod ( Ophiodon elongatus). In most cases, the number of annuli that formed beyond the OTC mark corre- sponded to the number of years at lib- erty. Expected and interpreted annuli counts from fin rays matched in 94%, 92%, 91%, and 75% of lingcod at lib- erty for one, two, three, and four years, respectively. These results validate the annual pattern of banding in fin-ray sections of fish up to 18 years of age. One of the consequences of not estab- lishing and following age determina- tion criteria is illustrated by the change in age composition due to misidentifi- cation of the first few annuli. A change in age composition has implications for stock estimates, such as mortality rates, and harvest strategies. For example, overfishing could result from underes- timated mortality rates if fishery man- agers used a strategy that set fishing mortality equal to natural mortality. It is recommended that age determination criteria be routinely assessed to provide reliable age estimates. Manuscript accepted 18 December 2000. Fish. Bull. 99:459-464 (2001). The validity of the fin-ray method of age determination for lingcod ( Ophiodon elongatus ) Gordon A. McFarlane Jacquelynne R. King Pacific Biological Station Fisheries and Oceans Canada Nanaimo, British Columbia V9R 5K6 Canada E-mail address (for G. A. McFarlane): mcfarlanes@pac.dfo-mpo.gcca Lingcod ( Ophiodon elongatus) are dis- tributed off the west coast of North America in the nearshore waters from Baja, California, to the Shumigan Islands, Alaska. They have been an important component of the commer- cial fishery off British Columbia since the mid- 1880s and a favorite target of recreational anglers since the 1950s. They inhabit nearshore waters and are commonly found along the bottom at depths ranging from 3 to 400 m; how- ever most are found in rocky areas 10 to 100 m. Growth during the first years of life is rapid and up to age 2 it is similar for males and females, both reaching an average length of 45 cm. After age 2, females grow faster than males and the growth of males tapers off at about age 8, whereas females con- tinue to grow rapidly until about age 12-14. For waters off the west coast of Canada, the maximum age recorded for lingcod has been 14 years for males and 20 years for females. Females reach lengths in excess of 100 cm, whereas males rarely exceed lengths of 90 cm. The fin-ray method for age deter- mination has been used for a wide range of freshwater and marine species such as sturgeon ( Acipensei ■ spp.), Pacif- ic salmon (Oncorhynchus spp.), brown trout ( Salmo trutta), whiter sucker (Ca- tostomus eommersoni), lake whitefish ( Coregonus clupeaformis), channel cat- fish ( Ictalurus punctatus ), common carp ( Cyprinus carpio), ide ( Leuciscus idus), tuna (Thun nus spp.), rudd ( Scardinius erythropthalmus), chub ( Squalls ceph- alus), roach ( Rutilus rutilus), bream ( Abramis brama), and yellow perch ( Per- ea fluviatilus ) (Beamish 1981). Howev- er, in only a few instances have there been attempts in recent years to val- idate age estimates with the fin-ray method ( Rien and Breamesderfer, 1994; Rossiter et al., 1995; Stevenson and Secour, 2000) despite the caution posed by Beamish and McFarlane ( 1983) that validation of age determination tech- niques is difficult, albeit critical. Beamish and Chilton (1977) devel- oped a method of aging lingcod that used thin sections of fin rays com- bined with mean annular diameter measurements to locate the position of the first and second annuli. This technique allowed estimates of mortal- ity rates, growth rates, and fecundity to be determined for each age group. These rates strongly influenced assess- ment and management approaches for lingcod. Preliminary validation of the method was accomplished by using a mark-recapture experiment in conjunc- tion with injections of oxytetracycline (OTC) (Cass and Beamish, 1983). How- ever, this preliminary validation of the method was based on only four recov- eries. A second experiment was initiat- ed in 1982 to complete the validation of the fin-ray method for lingcod. Our re- port presents the results of this second mark-recapture study combined with OTC injections. During the course of analyzing ling- cod age data in the early 1990s, it be- came apparent that the annular diam- eter measurement criteria suggested by Beamish and Chilton ( 1977) were no lon- ger being followed. Resulting parameter estimates based on these age estimates had changed. This resulted in a mis- understanding of growth, mortality, and age at maturity, which had implications for development of management strate- gies for lingcod. We therefore also report on a re-examination of the measurement 460 Fishery Bulletin 99(3) criteria proposed by Beamish and Chilton (1977) and show that accurate ages for lingcod can be produced by using the fin-ray method combined with mean annular diameter mea- surements. Methods During 14-27 July 1982, a major tag and release program was conducted on the commercial fishing grounds. La Perouse Bank (48'45°N; 125'55°W), off southwest Vancou- ver Island. The primary focus of this tagging program was for validation of age determination. Methods and survey design are outlined in Cass et al. (1983). Prior to process- ing, all fish were anesthetized with tricaine methane sul- fonate (MS222). We measured their fork length (mm) and inserted an individually numbered Floy FD68 anchor tag into the connective tissue just below the anterior base of the first dorsal fin. We injected the tagged fish with a 25 mg/kg body-weight dosage of oxytetracycline into the interperitoneal cavity (McFarlane and Beamish, 1987). Of the injected fish, a random subsample was kept in 3000-L holding tanks for 48 hours to assess short-term mortality due to the OTC injection, then released. We released a random subsample of tagged fish without OTC injections as a control group to assess long-term (6-month) mortality due to OTC injec- tion by comparing recapture rates for the control fish with those for the OTC-injected fish. Tagged lingcod were recovered by the commercial fish- ery off the southwest coast of Vancouver Island. Because the first annuli after the OTC mark would form during the winter of 1982-83, only tagged lingcod recovered after 1 January 1983 were used in the validation of age determi- nation. Returned whole fish were measured for fork length (mm), and a portion of the second dorsal fin was removed for age determination according to Beamish and Chilton (1977). Sectioned fin rays were illuminated with ultravio- let light to fluoresce the OTC mark and with regular light to identify annuli. All recovered lingcod were aged in June and August of 1987. In October 1987, a change in personnel led to an un- knowing change in the criteria for determining the first few annuli. Chilton and Beamish (1982) suggested that known average diameter of the first and second annuli could be used to estimate their location. The first and sec- ond annuli can be difficult to determine when resorption of the inner portion occurs or if the center becomes obscured by irregular-shaped deposits of opaque material. In addi- tion, the presence of checks can also make it difficult to identify the first two annuli. In all cases, an estimate of the position of the third annulus can be made by measur- ing the diameter of the first and second annuli from juve- nile fish or fish where these annuli are visible (Chilton and Beamish, 1982). By 1988, the primary reader was no lon- ger using the Chilton and Beamish (1982) criteria. The re- covered lingcod were re-aged after this change in criteria and second ages (using these criteria) were compared with original ages (not using these criteria) to assess bias. We compared age compositions of a subsample of the lingcod C i i i Figure 1 Schematic representation of section of fin ray for measure- ments of first (A) and second (B) annular diameters. Mea- surements should be made along axes perpendicular to the axis (C) of the inner groove of the fin ray. with the two age estimates to illustrate the consequences of not using the Chilton and Beamish (1982) criteria. Juvenile lingcod (less than 48 cm) were captured in 1987 in the recreational fishery in the Strait of Georgia. Fork lengths were measured (nearest mm) and dorsal fin rays were sampled during the creel survey program. These juveniles would not likely have resorbed fin-ray centers. The fin rays were sectioned (Beamish and Chilton, 1977), and illuminated with regular light. The diameter to the first and second annuli were measured with a micrometer eyepiece under 40x magnification to the nearest microm- eter (pm). Diameter measurements were made along an axis perpendicular to the inner groove of the fin ray (Fig. 1 ). If an annuli appeared as a thick zone, the range from the beginning to the end of the zone was measured and the median used as the diameter measurement. Results A total of 7429 lingcod were tagged and released during July 1982. Of these, 6946 were injected with OTC and 483 were not injected but were released as the control group for assessing long-term (6-month) mortality due to OTC injection. A total of 188 lingcod were held for 48 hours after being tagged and injected with OTC. No mortality occurred prior to release. Within the first six months (i.e. by 31 December 1982) 1442 lingcod were recovered. During this period, the return rates of OTC and non-OTC-injected lingcod were similar (19.5% and 18.4%, respectively) indicating no long-term mortality from OTC injection. McFarlane and King: The validity of the fin-ray method of age determination for Ophiodon elongcitus 461 11111 Figure 2 Annuli are indicated by dots, OTC ( oxytetracycline ) mark indicated by arrow. A, C, E were photo- graphed by using white light and B, D, F were photographed by using ultraviolet light. (A and B) Fish numbered B8223905 at liberty for 2 years, with OTC mark evident after fourth annulus. Total age was 6+. (C and D) Fish numbered B8227766 at liberty for 3 years, with OTC mark evident after twelfth annulus. Total age was 15+. (E and F) Enlargements of sections C and D after the fifth annulus. Between 1983 and 1986, 1002 tagged lingcod were recov- ered. From these, 460 lingcod fin rays were obtained for age determination validation: 348 recovered in 1983, 65 recovered in 1984, 43 recovered in 1985, and 4 recovered in 1986. The number of observed annuli after the OTC marks corresponded well to the number of expected annuli based on the number of years at liberty (Table 1). Lingcod recovered in 1983 were expected to have one annulus af- ter the OTC mark and approximately 94% of the samples were estimated to have one annulus. Lingcod recovered in 1984, 1985, and 1986 had estimated annuli after the OTC mark that corresponded to the expected number of annuli in 92%, 91%, and 75% of the samples, respectively. In most cases, the presence of the OTC mark was eas- ily discernible and appeared as a crisp distinct line. Annu- li following the mark were also clearly visible. Fish num- bered B8223905 was recovered in July 1984 and aged at 6 years. The OTC mark formed just beyond the fourth annu- lus (Fig. 2A) and was obvious under ultraviolet light (Fig. 2B). After the OTC mark, two complete annual growth zones were present corresponding to the two years at lib- erty. Each annual growth zone consisted of an opaque (summer) and a translucent (winter) zone. The OTC mark at the beginning of the summer opaque zone (Fig. 2A) in- dicated that growth had just begun prior to tagging and injection in July. Fish numbered B8227766 was recovered in July 1985 and aged at 15 years. As with fish B8223905, the OTC mark was also formed at the beginning of an opaque zone but was a broad band covering the whole summer zone 462 Fishery Bulletin 99(3) Table t The number of estimated annuli after the OTC mark versus the expected annuli based on years at large. Num- bers in bold indicate those samples in which the number of observed annuli equaled the number of expected annuli. Expected annuli Estimated annuli 12 3 4 0 1 2 3 4 3 1 — — 328 3 — — 16 60 1 1 39 1 4 3 (Fig. 20 and clearly visible with ultraviolet light (Fig. 2D). This fin ray had three annual growth zones after the OTC mark, each comprising opaque and translucent zones (Fig. 20. An annual growth zone can contain several dis- tinct translucent zones, i.e. checks. In this example, the tenth labelled annulus did not appear to be continuous and could, therefore, be interpreted as a check (Fig. 1, E and F). Although this area likely contained checks, the thickness and prominence of the labelled translucent zone around most of the fin-ray section (Fig. 20 indicated that it was the edge of the tenth annulus. Fish numbered B8224635 was recovered in July 1985 and was estimated to be 9 years old (Fig. 3). Unlike the previous two examples, the first two annuli were difficult to determine because of resorption in the central portion of the fin ray (Fig. 3). We used the diameter measurement criteria of Chilton and Beamish (1982) to locate the first and second annuli. The 46 juvenile lingcod available from the creel survey in 1987 were used for first and second annular measure- ments. The mean first annular diameter was 0.40 mm (SD=0.07) and the mean second annular diameter was 0.63 mm (SD=0.11). With one exception, there was no overlap between the measured diameter for the first an- nulus and those measured for the second annulus and the two means were significantly different (Ctest, t= 17.46, df=86, P<0.0001). The mean annular diameter did not vary across the range of fork lengths observed (Fig. 4). Ages determined by using average diameter for the lo- cation of first and second annuli measured in the late- 1980s were generally one year older than ages determined without using a measurement to locate the first two an- nuli (Fig. 5). There appeared to be no trend related to age. However, across all ages, the unmeasured estimates ranged from three years younger to three years older than estimates made with the measurement criteria (Fig. 5). This result is likely due to the exclusion of annuli that cannot be recognized or due to the inclusion of checks, mis- taken as annuli, in some fish. There was no difference in the number of annuli estimated after the OTC mark. The one-year aging bias is illustrated in the age composition Figure 3 Fish numbered B8224635 had resorbed (R) cen- tral portion of the fin ray, making identification of the first two annuli (arrows) difficult. Subse- quent annuli are indicated by dots. Total age was 9+. for the lingcod recovered in 1983 (n=341). The age com- position for the ages determined with the measurement criteria had a mode of 6 years, whereas the age composi- tion for ages determined without the measurement crite- ria had a mode of 5 years (Fig. 6). For all of the samples, the estimated ages ranged from three to 18 years. Discussion Accurate age determination is essential for proper stock assessment and fisheries management. Although aging inaccuracies of one year might not be critical in long-lived species (e.g. 80-year-old rockfish), they can have serious stock assessment implications for shorter-lived species, such as lingcod. This is likely true because shorter-lived species typically have only four or five year classes that dominant the fishery. It is therefore important to develop accurate methods for age determination and to validate those methods. The fin-ray method first described by Beamish and Chilton (1977) is an accurate method for age determination of lingcod. Annular growth is represented by an opaque (summer growth) and a translucent (winter growth) zone and the edge of the annulus is assigned to the translucent zone (Beamish and Chilton, 1977). Using McFarlane and King: The validity of the fin-ray method of age determination for Ophiodon elongotus 463 0.8 r 0.7 - E .E, 0.6 - o if) « 0.5 - D • | 0.4 - < 0.3 - □ □ □ □ □ □ □ □ □ □0 □ B □ 0.2 ■ i i » i » < 240 280 320 360 400 440 480 520 Fork length (mm) Figure 4 The mean diameter of the first (circles) and the second (squares) annuli are 0.40 mm and 0.62 mm, respectively. The mean diameters do not show a change in value across the fork lengths observed. ~G 0 if) 13 <5 O E CC -O if) D c c nJ cn > 6 weeks) stage witch flounder larvae to that of yellowtail flounder (Pleuronectes ferrugineus) and Atlantic cod ( Gadus morhua ) to examine differences in prey con- sumption rate between species. Yellowtail flounder were observed at 8000 prey/L because this prey density pro- motes good growth and survival for this species (Rabe and Brown, 2000). Similarly, the data used for Atlantic cod were taken from the behavioral observations conducted at 4000 prey/L presented in Puvanendran and Brown ( 1999). The 4000 prey/L level was the prey density that optimized growth and survival in that experiment. Results Experiment 1 : Growth and survival At hatching, the mean standard length of larvae was 5.62 mm (±0.12 mm SE). The standard length, dry weight, and body height of larvae did not differ between prey density 468 Fishery Bulletin 99(3) Table 2 Summary of ANOVA and ANCOVA results for growth and foraging response variables of witch flounder larvae at different prey densities (no. of prey per liter). An ANOVA was used for orient frequency within the size range 10.5-20.8 mm; for all other response variables an ANCOVA was used. Age (weeks after hatching) was used as the covariate for growth response variables, whereas size (mm SL) was used for behavioral response variables. (* denotes a significant difference at a=0.05). MAP Source df F P Standard length (mm) Age 1 7041.4 <0.001* Prey density 2 0.85 >0.25 Age x prey density 2 0.75 >0.25 Error 66 Dry weight (mg) Age 1 2640.4 <0.001* Prey density 2 0.06 >0.5 Age x prey density 2 0.21 >0.5 Error 18 Body height (mm) Age 1 582.1 <0.001* Prey density 2 0.03 >0.5 Age x prey density 2 0.28 >0.5 Error 18 Orientation frequency Prey density 6 4.71 <0.01* Error 42 Fixate frequency Size 1 6.94 <0.05* Prey density 6 1.29 >0.25 Size x prey density 6 0.52 >0.5 Error 84 Lunge frequency Size 1 15.8 <0.001* Prey density 6 1.35 >0.25 Size x prey density 6 0.62 >0.5 Error 84 treatments (Table 2, Figs. 1 and 2). Larvae began to increase in body height around the mean size of 15 mm (Fig. 2B). The average absolute growth rate from week 0 to week 12 for all treatments was 0.53 mm/d.. The average specific growth rate (SGR) from week 0-6 was 3.68%/d and from week 0-12 was 2.61%/d. The survival results were not corrected for fish sampled lethally (20 per tank). The survival in all treatments was similar over the course of the experi- ment (Table 3) and was unaffected by prey density at week 12 (ANOVA; F2 3= 2.75, P=0.210). Experiment 2: Behavior Larval behavior was characterized by a shift from nondirected MAPs to locomotory activities between the mean sizes of 10.5-16.2 mm (Fig. 3, A and B). The increase in locomotory activities was due to an increase in the total time spent swimming (Fig. 3A). The average duration of a swim MAP increased from 3.2 to 86.3 seconds over the study, whereas the frequency of swimming decreased sig- nificantly (Table 4). A turn MAP was used during the early stage, disappearing by the time larvae reached Rabe and Brown: Behavior, growth, and survival of Glyptocephalus cynoglossus larvae in relation to prey availably 469 10.5 mm (Fig. 3A). Pause and sink MAPs decreased when larvae reached 16.2 mm, whereas a shake MAP stopped altogether (Fig. 3B). The durations of the locomotory and nondirected MAPs changed with size (Fig. 3, A and B) but were not significantly affected by prey density (Table 4). At the end of the study there was a slight increase in pause MAPs and a concomitant decrease in the duration of swim MAPs, the result of some larvae settling during the observation periods. Throughout the study, larvae spent between 2% and 10% of the total time observed performing forag- ing activities. The variation in total time spent for- aging was largely due to variation in orientation du- ration. The duration of the fixate and lunge MAPs (<2% of total time per MAP) was relatively constant over the observation periods, whereas the duration of the orientation MAP ranged from 1% to 7% of total time (Fig. 30. The frequencies of the foraging behaviors were highly variable and many larvae did not forage dur- ing the observation periods. The frequency of the ori- entation MAP changed throughout the study period, peaking between the mean sizes of 10.5 and 20.8 mm (Fig. 4A). Within this size range, orientation fre- quency increased with increasing prey density. The effect of prey density on orientation frequency was significant (Table 2); the orientation frequency of lar- vae at 250 prey/L was significantly lower than that of larvae at 2000-16,000 prey/L within the 10.5-20.8 mm size interval (Tukey test). After 20.8 mm, the frequency of orientation was low for all sizes and treatments. The frequencies of fixate and lunge MAPs varied from 0 to 4 per two-min observation. Larvae at high- er prey densities tended to perform more fixate and lunge MAPs compared with larvae at lower prey densities (Fig. 4, B and C), although this trend was not statistically significant (Table 2). The frequen- cies of fixate and lunge MAPs increased significantly with increasing larval length (Fig. 4, B and C, Table 2). The average lunge frequency of early- and late-stage witch flounder larvae was lower than that of both yellowtail flounder and Atlantic cod larvae (Fig. 5). Discussion Witch flounder larvae grew and survived in all treatments used in our study. Our experiment is the first to examine the early growth and behavior of witch flounder larvae in relation to prey availability. Larval performance can be influenced by many factors other than prey density, including temperature (Hunter, 1981), light, (Batty, 1987; Puvanendran and Brown, 1998), prey type (Drost, 1987), and turbulence (MacKenzie and Kiprboe 1995; Brownian, 1996). We used our results 1) to describe the early growth and ontogeny of the foraging behavior in this species and 2) as a preliminary step towards understanding the behav- ioral ecology of witch flounder larvae. O) E O) 0) S Q 1000 100 10 0.1 0.01 j f 2000 p/L 4000 p/L 8000 p/L WeekO 10 20 30 40 50 60 Standard length (mm) Figure 2 Relationship between standard length (mm) and (A) dry weight (mg; note logarithmic axis) and body height (mm) for individ- ual witch flounder larvae reared at different prey densities (no. of prey per liter). Symbols represent individual larvae. Table 3 Percentage of witch flounder larvae reared at different prey densities (±SE) that survived over the experiment. Latvae sampled dead were not included in calculations. Prey density (no. of prey per liter) 2000 4000 8000 Week 2 Week 5 Week 12 36.15(1.28) 28.72 (1.54) 14.10 (2.31) 30.26 (6.15) 25.13 (5.13) 4.62 (0.51) 38.20 (4.36) 31.03 (3.85) 8.97 (4.36) Witch flounder grew and survived equally well at each of the prey densities tested. Although the range of prey den- sities used in the rearing experiment (2000-8000 prey/L) was not exhaustive, these prey densities have resulted in informative differences in growth and survival for other 470 Fishery Bulletin 99(3) c CD O CD D. 100 80 60 40 20 10 100 80 60 40 20 10 B Total locomotion Swim Turn Total nondirected Pause Shake Sink Standard length (mm) Figure 3 Mean proportion (%) of time witch flounder larvae spent performing individual (A) locomotory, (B) nondirected, and (C) foraging MAPs over standard length (mm) during two- minute observation periods. Values are means (n= 70 larvae per length) ±SE. North Atlantic marine fish larvae, such as Atlantic cod (Puvanendran and Brown, 1999) and redfish ( Sebastes sp.; Laurel et al., in press) reared under similar laboratory conditions. Therefore, we anticipated that this range of prey densities would be adequate to detect differences in growth and survival of witch flounder in relation to prey density. Furthermore, as shown in experiment 2, the lunge frequency of witch flounder was not significantly affected by prey availability, which would be expected if low prey densities (<2000 prey/L) were to reduce consumption and subsequent growth and survival. 12 10 >. o c S> 6 cr CD 4= 4 250 p/L 500 p/L 1 000 p/L 2000 p/L 4000 p/L 8000 p/L 16000 p/L >. o c CD 3 cr CD >> o c CD CD CD c 10 15 20 25 30 Standard length (mm) Figure 4 Frequency of (A) orient, (B) fixate, and (C) lunge MAPs of witch flounder larvae at different prey densities (no. of prey per liter) during two-minute observation periods over standard length (mm). Values are means (n=10) ±SE. Ontogeny of behavior Witch flounder search strategy for prey is interesting because it appeared to change from a saltatory to a cruise strategy (see O’Brien et al., 1990; Browman and O’Brien, 1992) during the study period. When larvae were less than 10 mm, foraging included many turns and brief periods of swimming that served as repositioning acts. By the time larvae reached an average size of 16.2 mm, most of their Rabe and Brown: Behavior, growth, and survival of Glyptocephclus cynoglossus larvae in relation to prey availably 471 Table 4 Summary of ANCOVA results for locomotory and nondirected MAPs of witch flounder larvae at different prey densities ( per liter). Each model was run until the larval size indicated in parentheses to satisfy model assumptions. (* denotes a difference at a=0.05). no. of prey significant MAP Source df F P Swim frequency (20.8 mm) Size 1 32.1 <0.001* Prey density 6 0.78 >0.25 Size x prey density 6 0.42 >0.5 Error 49 Swim duration (20.8 mm) Size 1 251.4 <0.001* Prey density 6 1.76 >0.25 Size x prey density 6 0.17 >0.5 Error 49 Turn duration (13.8 mm) Size 1 56.5 <0.001* Prey density 6 1.34 >0.25 Size x prey density 6 0.82 >0.5 Error 21 Pause duration (20.8 mm) Size 1 88.4 <0.001* Prey density 6 1.82 >0.25 Size x prey density 6 0.26 >0.5 Error 49 Sink duration (18.4 mm) Size 1 79.8 <0.001* Prey density 6 0.54 >0.5 Size x prey density 6 1.35 >0.5 Error 42 Shake duration (16.2 mm) Size 1 13.0 <0.001 Prey density 6 1.13 >0.25 Size x prey density 6 0.25 >0.5 Error 35 time was spent swimming, which was typically interrupted only by foraging events. During the mean size interval of 10.5-16.2 mm, the turn and shake MAPs disappeared and the frequency of pause and sink MAPs decreased. These behavioral changes were likely related to the increase in larval body height, accompanied by a substantial increase in finfold height that occurs during this time. The nature of the nondirected MAPs shake, sink, and pause is not straightforward. Sinking has been reported in other species, such as the snapper (Pagrus auratus) and, like the pause MAP, has been interpreted as a rest- ing behavior. In snapper, it occurs in yolksac larvae and in feeding larvae during night-time periods of inactivity (Pankhurst et al., 1991). Sinking is typically observed only in the early stages of other species such as the black sea bream, Acanthopagrus schlegeli (Fukuhara, 1987). How- ever, Kawamura and Ishida (1985) noted that sinking oc- curs in both yolksac larvae and larger feeding larvae of the flounder Paralichthys olivaceus immediately after at- tacking a prey item. Observations of sinking in later-stage witch flounder larvae were not related to feeding events; the persistence of these behaviors in witch flounder was likely the result of some smaller, slower-growing individu- als having been included in our observations. Witch flounder, like other species (Holling, 1965; Houde and Schekter, 1980; Werner and Blaxter, 1980; Puvanen- dran and Brown, 1999), demonstrated increased foraging behavior with prey density. However, the orient MAP was the only foraging MAP statistically affected by prey den- sity. The change in orientation frequency with size is in- teresting because this MAP was affected only by prey den- sity within a limited size range. The initial low frequency of orient MAPs followed by an increase associated with great- er larval size can be explained by changes in swimming speeds and encounter rates (Mittelbach, 1981). However, the decrease in orient frequency across treatments later in the study period is puzzling because larger larvae are gen- erally competent swimmers (Rosenthal and Hempel, 1971; Laurence, 1972; Houde and Schekter, 1980) and are expect- ed to exhibit a high prey encounter rate. This decrease in orient frequency may be due to improved foraging ability associated with greater visual acuity. Miller et al. (1993) showed that the visual angle — the smallest angle at which a stimulus may subtend the eye and remain resolvable (Neave, 1984) — decreases during the development of three species of fish larvae. Thus, the eye develops such that lar- vae can likely detect prey items in their periphery without turning the head and orienting themselves toward them. 472 Fishery Bulletin 99(3) 5 i 4 - WF WF YT YT AC AC Early Late Early Late Early Late Witch flounder, this experiment Yellowtail flounder, Rabe and Brown, 2000 Atlantic cod, Puvanendran and Brown, 1999 Figure 5 Average lunge frequency per minute for early (<2 weeks) and late stage (>6 weeks) witch flounder (WF), yellowtail flounder ( YT), and Atlantic cod larvae (AT). See text for details. Behavioral ecology The main finding of our study was that witch flounder are not affected by changes in prey availability in the same manner as other species of larvae observed under similar laboratory conditions. The typical pattern among fish lar- vae— that they increase their foraging behavior and prey consumption rate with increased prey density (Houde and Schekter, 1980) — was supported by our study. However, the results are unusual in that the effects of prey density on foraging behavior were not as strong as results that have been reported for other species. The ecological implications of these results can be il- lustrated by a comparison of the growth and behavior of witch flounder to other northwest Atlantic species ob- served under similar laboratory conditions. In rearing experiments on Atlantic cod larvae, Puvanendran and Brown (1999) found that cod have specific requirements for high prey densities. Larval survival, growth rate (0.20 mm/d), and condition were highest when larvae were reared at a prey density of 4000 prey/L. Furthermore, in the same experiment, the lunge frequency (an indicator of consumption rate) of cod larvae increased from nearly 1 to 3.5 prey items per minute over the six-week study period (Fig. 5; Puvanendran and Brown, 1999). A marked increase in lunge frequency with age is also seen in yel- lowtail flounder, another north Atlantic pleuronectiform, raised under similar conditions (Fig. 5; Rabe and Brown, 2000). In that study, the growth rate of yellowtail flounder was 0.34 mm/d. Not only is the foraging behavior of witch flounder rela- tively unaffected by variation in prey density, but its lunge frequency is much lower, suggesting that it may have a lower consumption rate and therefore lower prey require- ments compared with those of other species. The lack of a significant effect of prey density on the foraging of witch flounder is not solely a result of the larger larval size of this species. Redfish ( Sebastes ) are relatively large at ex- trusion (6-8.9 mm; Penny and Evans, 1985) and their for- aging behavior is affected by variations in prey availabil- ity (Laurel et al., in press). It is remarkable that the growth rate of witch flounder is faster than that of cod and yellowtail flounder, especial- ly given the possible lower prey consumption rate of witch flounder. Two potential mechanisms that may explain this phenomenon are high assimilation efficiency and low met- abolic requirements. Witch flounder are relatively large at hatching and for most of the study period were larger than other species of similar age. This size difference alone im- plies that its digestive system is larger, more developed, and more efficient (Govoni et al., 1986; Klumpp and von Westernhagen, 1986). Both yellowtail flounder and cod are more active and swim faster than witch flounder. Higher activity imparts a greater need for prey, which results in Rabe and Brown: Behavior, growth, and survival of Glyptocephalus cynoglossus lan/ae in relation to prey availably 473 larvae being more susceptible to starvation in the absence of prey (Hunter, 1981). The ecological significance of the lack of an effect of prey density on foraging, growth, and survival found in our study may be a reflection of the life history of witch flounder. This species has an extended larval period and is committed to being in the water column much longer than other species (Bigelow and Schroeder, 1953). During this extended larval period, witch flounder larvae will likely encounter periods of high or low plankton availability, or both. The abundance of zooplankton prey for fish larvae can vary over four orders of magnitude during the year, typically reaching a peak in the warmer months and de- creasing dramatically in the winter (Myers et al., 1994). Therefore, witch flounder larvae must be able to cope with this variation in prey availability to survive. This condi- tion requires a different strategy from that of other ma- rine fish species that have shorter larval periods and that likely rely on a “match” of spawning with plankton pro- duction to promote larval survival (Cushing, 1972). In the scope of the life history evolution of witch flounder, the long larval period and large size at metamorphosis may be a strategy to cope with intense postsettlement, size-de- pendent competition. Acknowledgments We would like to thank Deborah Bidwell for supplying the larvae used in our study. Technical assistance was kindly provided by Danny Boyce, Dena Wiseman, Olav Lyngs- tad, Donna Sommerton, Trevor Keough, Ralph Pynn, and Tracy Granter. We are grateful to Velmurugu Puvanen- dran and Pierre Pepin for insightful discussions. Funding for this research was provided by the Canadian Centre for Fisheries Innovation and a Memorial University of New- foundland Graduate Fellowship. Literature cited Altman, J. 1974. Observational study of behavior: sampling methods. Behaviour 91:449-459. Barlow, G. W. 1968. Ethological units of behavior. In The central ner- vous system and fish behavior (D. J. Ingle, ed.), p. 217-232. Univ. Chicago Press, Chicago, IL. Batty, R. S. 1987. Effect of light intensity on activity and food-search- ing of larval herring, Clupea harengus : a laboratory study. Mar. Biol. 94:323-327. Bigelow H. B., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine. Fish. Bull. 53:1-577. Braum, E. 1978. Ecological aspects of the survival offish eggs, embryos and larvae. In Ecological basis of freshwater fish produc- tion (S. D. Gerking, ed ), p. 102-136. Blackwell Scientific, Oxford. Browman, H. I. 1996. Predator-prey interactions in the sea: commentaries on the role of turbulence. Mar. Ecol. Prog. Ser. 139:301-312. Browman, H. I., and J. O’Brien. 1992. Foraging and prey search behavior of golden shiner (Notemigonus crysoleucas) larvae. Can. J. Fish. Aquat. Sci. 49:813-819. Brown, J. A., and P. W. Colgan. 1985. The ontogeny of feeding behavior in four species of centrarchid fish. Behav. Process. 9:395-411. Busacker, G. P, I. R. Adelman, and E. M. Goolish. 1990. Growth. In Methods for fish biology (C. B. Shreck and P. B. Moyle, eds.), p. 363-388. Am. Fish. Soc., Bethesda, MD. Cushing, D. H. 1972. The production cycle and the numbers of marine fish. Symp. Zool. Soc. Lond. 29:213-232. Drost, M. R. 1987. Relation between aiming and capture success in larval fishes. Can. J. Fish. Aquat. Sci. 44:304-315. Fukuhara, O. 1987. Larval development and behavior in early life stages of black sea bream reared in the laboratory. Bull. Jpn. Soc. Sci. Fish. 53:371-379. Govoni, J. J., G. W. Boehlert, and Y. Watanabe. 1986. The physiology of digestion in fish larvae. Environ. Biol. Fishes 16:59-77. Holling, C. S. 1965. The functional response of invertebrate predators to prey density and its role in mimicry and population regula- tion. Mem. Entomol. Soc. Can. 451-60. Houde, E. D. 1978. Critical food concentrations for larvae of three species of subtropical marine fishes. Bull. Mar. Sci. 28:395-411. Houde, E. D., and R. C. Schekter. 1980. Feeding by marine fish larvae: developmental and functional responses. Environ. Biol. Fishes 5:315-334. Hunter, J. R. 1981 Feeding ecology and predation on marine fish larvae. In Marine fish larvae: morphology, ecology and relation to fisheries (R. Lasker, ed.), p. 33-77. Washington Sea Grant Program, Univ. Washington Press, Seattle, WA. Kawamura, G., and K. Ishida. 1985. Changes in sense organ morphology and behavior with growth in the flounder Paralichthys olivaceus. Bull. Jpn. Soc. Sci. Fish. 51:155-165. Klumpp, D. W., and H. von Westernhagen 1986. Nitrogen balance in marine fish larvae: influence of developmental stage and prey density. Mar. Biol. 93: 189-199. Laurel, B. J., J. A. Brown, and R. Anderson In press. Behaviour, growth and survival of redfish ( Sebastes spp.) larvae in relation to prey availability. J. Fish Biol. Laurence, G. C. 1972. Comparative swimming abilities of fed and starved larval largemouth bass (Micropterus salmoides ). J. Fish Biol. 38:1-15. MacKenzie, B. R., and T. Kiorboe. 1995. Encounter rates and swimming behavior of pause- travel and cruise larval fish predators in calm and tur- bulent laboratory environments. Limnol. Oceanogr. 40: 1278-1289. Miller, J. M., J. S. Burke, and G. R. Fitzhugh. 1991. Early life history of North American flatfish: Likely (and unlikely) factors controlling recruitment. Neth. J. Sea Res. 27:261-275. Miller, T. M., L. B. Crowder, and J. A. Rice. 1993. Ontogenic changes in behavioral and histological mea- 474 Fishery Bulletin 99(3) sures of visual acuity in three species of fish. Environ. Biol. Fishes 37:1-8. Mittelbach, G. G. 1981. Foraging efficiency and body size: a study of optimal diet and habitat use by bluegills. Ecology 62: 1370-1386. Munk, P. 1995. Foraging behavior of larval cod ( Gadus morhua ) influ- enced by prey density and hunger. Mar. Biol. 122:205- 212. Myers, R. A., N. J. Barrowman, G. Mertz, J. Gamble, and H. G. Hunt. 1994. Analysis of continuous plankton recorder data in the Northwest Atlantic 1959-1992. Can. Tech. Rep. Fish. Aquat. Sci. 1966, iii + 246 p. Neave, D. A. 1984. The development of visual acuity in larval plaice ( Pleuronectes platessa L.) and turbot ( Scopthalmus inaxi- mus L.). J. Exp. Mar. Biol. Ecol. 78:167-175. O'Brien, W. J., H. I. Browman, and B. I. Evans. 1990. Search strategies in foraging animals. Am. Sci. 78: 152-160. Osse, J. W. M., and J. G. M. Van den Boogaart . 1997. Size of flatfish larvae at transformation, functional demands and historical constraints. Neth. J. Sea Res. 37:229-239. Pankhurst, P. M., J. C. Montgomery, and N. W. Pankhurst. 1991. Growth, development, and behavior of artificially reared larval Pagrus auratus (Bloch & Schneider, 1801) (Sparidae). Aust. J. Mar. Freshwater Res. 42: 391-398. Penney, R. W., and G. T. Evans. 1985. Growth histories of larval redfish ( Sebastes spp.) on an offshore Atlantic fishing bank determined by otolith incre- ment analysis. Can. J. Fish. Aquat. Sci. 42:1452-1464. Puvanendran, V., and J. A. Brown. 1998. Effect of light intensity on the foraging and growth of Atlantic cod larvae: interpopulation difference? Mar. Ecol. Prog. Ser. 167:207-214. 1999. Foraging, growth and survival of Atlantic cod larvae reared in different prey concentrations. Aquaculture 175: 77-92. Rabe, J. 1999. The behavior, growth, and survival of witch flounder and yellowtail flounder larvae in relation to prey avail- ability. M.Sc. thesis. Memorial Univ. Newfoundland, St. John’s, 135 p. Rabe, J., and J. A. Brown. 2000. A pulse feeding strategy for rearing larval fish: an exper- iment with yellowtail flounder. Aquaculture 191:289— 302. Rosenthal, H., and G. Hempel. 1971. Experimental estimates of minimum food density for herring larvae. Rapp. P.-v. Reun. Cons. int. Explor. Mer 160: 125-127. Scott, B. S., and M. G. Scott. 1988. Atlantic fishes of Canada. Can. Bull. Fish. Aquat. Sci., 219 p. Suthers, I. 2000. Significance of larval condition: comment on labora- tory experiments. Can. J. Fish. Aquat. Sci. 1534-1536. Werner R. G., and J. H. S. Blaxter. 1980. Growth and survival of larval herring ( Clupea haren- gus ) in relation to prey density. Can. J. Fish. Aquat. Sci. 37:1063-1069. Zar, J. 1999. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, NJ, 663 p. 475 Abstract— Lake Maracaibo, a large Venezuelan estuarine lagoon, is report- edly inhabited by three species of the genus Callinectes Stimpson, 1860 that are important to local fisheries: C. sapi- dus Rathbun, 1896, C. bocourti A. Milne Edwards, 1879, and C. maracaiboensis Taissoun, 1969. Callinectes maracai- boensis, originally described from Lake Maracaibo and assumed endemic to those waters, has recently been reported from other Caribbean localities and Brazil. However, because characters separating it from the morphologically similar C. bocourti are noted to be vague, we have compared these spe- cies and several congeners by molec- ular methods. Among our specimens from Lake Maracaibo and other parts of the Venezuelan coast, those assign- able to C. bocourti and C. maracaiboen- sis on the basis of putatively diagnostic characters in coloration and structural characteristics do not differ in their 16S mtDNA sequences. These molec- ular results and our re-examination of supposed morphological differences between these species suggest that C. maracaiboensis is a junior synonym of C. bocourti, which varies markedly in minor features of coloration and struc- tural characteristics. Genetic relation- ships of this species to other species of swimming crabs of the genus Calli- nectes are also explored and presented as phylogenetic trees. Manuscript accepted 21 March 2001. Fish. Bull. 99:475-481 (2001). Lack of divergence between 16S mtDNA sequences of the swimming crabs Callinectes bocourti and C. maracaiboensis (Brachyura: Portunidae) from Venezuela* Christoph D. Schubart Department of Biology University of Louisiana Lafayette, Louisiana 70504 Present address: Biologie I, Universitat Regensburg 93040 Regensburg, Germany E-mail address: christoph.schubart@biologie.uni-regensburg.de Jesus E. Conde Carlos Carmona-Suarez Centro de Ecologia Instituto Venezolano de Investigaciones Cientfficas (IVIO A P. 21827 Caracas 1020-A, Venezuela Rafael Robles Darryl L. Felder Department of Biology University of Louisiana Lafayette, Louisiana 70504 Swimming crabs of the genus Calli- nectes Stimpson, 1860 are widely dis- tributed throughout the American trop- ics and subtropics, where many species are exploited commercially (Norse, 1977; Williams, 1984). Members of this genus play key trophic roles in coastal habitats that range from sandy-mud bottoms to seagrass meadows (Arnold, 1984; Orth and van Montfrans, 1987; Wilson et al., 1987; Lin, 1991). Of the nine species of Callinectes from the tropical western Atlantic, seven have been reported from western Venezuela (Rodriguez, 1980; Williams, 1984; Carmona-Suarez and Conde, 1996). However, knowledge of these Venezuelan populations remains inadequate, despite the importance of several species in both large-scale com- mercial harvests and artisanal fish- eries of coastal villages in the area (Oesterling and Petrocci, 1995; Ferre r- Montano, 1997; Conde and Rodriguez, 1999). Persistent difficulty in identify- ing mixed samples of these species ham- pers understanding of their distribution, abundance, and population dynamics, which are essential to fishery-manage- ment decisions. In particular, consistent morphological distinction of the com- monly encountered C. bocourti from C. maracaiboensis cannot be achieved with confidence. Initially described as endemic from Lake Maracaibo, C. maracaiboensis has been reported from other Vene- zuelan waters, as well as from sites in Jamaica, Curasao, Colombia, and Brazil (Norse, 1977; Carmona-Suarez and Conde, 1996; Sankarankutty et al., 1999). Although detailed descriptions have been provided by Taissoun (1969, 1972) and Williams (1974) to diagnose C. maracaiboensis, the same authors have repeatedly emphasized its resem- blance to C. bocourti. Morphological distinction of the two species is based upon postulated defining characters of the frontal teeth, direction and form of * Contribution 75 of the University of Lou- isiana Lafayette, Laboratory for Crusta- cean Research, Lafayette, LA 70504. 476 Fishery Bulletin 99(3) the anterolateral spines, convexity of the carapace, sur- face granulation patterns, and varied morphometric mea- surements (Taissoun 1969, 1972). However, neither these characters nor additional ones proposed by other authors (Williams, 1974; Rodriguez, 1980) have proven to be of consistent use in separating the species. Carapacial color and granulation vary widely, and the outer frontal teeth of many specimens cannot be classified as a definitive tri- angular (C. maracaiboensis ) or obtuse (C. bocourti) shape. These concerns have led some authors to report only pos- sible co-occurrence of C. maracaiboensis in their samples of C. bocourti , as in records from Puerto Rico (Buchanan and Stoner, 1988). Recently, molecular markers have proven particularly valuable in decapod crustacean systematics. In addition to reconstructing phylogenetic relationships, molecular se- quences are of value in recognizing questionable species separations (e.g. Geller et ah, 1997; Sarver et al., 1998; Schneider-Broussard et ah, 1998; Schubart et ah, 1998; Schubart et ah, in press). In our study, we used sequences of a mitochondrial gene (16S rRNA) to examine genetic differentiation between the morphologically similar spe- cies C. bocourti and C. maracaiboensis and to compare these sequences to those of other swimming crabs of the genus Callinectes in a phylogenetic analysis. Materials and methods Swimming crabs were caught by hand in shallows of the Golfete de Cuare (68°37'W, 10°06'N, State of Falcon, Vene- zuela) in November-December 1998 and March 1999, and in the shallows of El Mojan, Lago de Maracaibo (71°39'W, 10°58'N, State of Zulia, Venezuela) in October 1999. Crabs were transported to the Institute Venezolano de Investig- aciones Cientificas, Caracas, in liquid nitrogen and then stored in a -70°C freezer. Two walking legs and one swim- ming leg were separated from each specimen, preserved individually in 85% ethanol, and shipped to the University of Louisiana, Lafayette (U.S.A.), for molecular analyses. Specimens were assigned to Callinectes bocourti or C. maracaiboensis according to morphological characters provided by Taissoun (1969, 1972), Williams (1974, 1978), Fischer (1978), and Rodriguez (1980). Eight specimens of C. bocoui'ti and six of C. maracaiboensis from Venezuela were used for the molecular analysis. In addition, a short fragment (-300 base pairs) of 16S mtDNA was obtained with an internal primer for one formalin-preserved speci- men of C. bocoui'ti from Colombia. These voucher speci- mens, lacking limbs (which had been removed for analy- sis), were cataloged in collections of the Laboratorio de Ecologia y Genetica de Poblaciones, Institute Venezolano de Investigaciones Cientificas (IVIC), and in the Univer- sity of Louisiana Zoological Collection (Table 1). For out- group comparisons, specimens of C. sapidus, C. oniatus Ordway, 1863, C. danae Smith, 1869, and Portunus ord- wayi (Stimpson, 1860) were sequenced (Table 1). Genomic DNA was isolated from the muscle tissue of one walking leg with a phenol-chloroform extraction (Kocher et al., 1989). Selective amplification of a 585-basepair (bp) product (547 bp excluding primer regions) from the mito- chondrial 16S rRNA gene was carried out by a polymerase- chain-reaction (PCR) (35-40 cycles; 1 min at 94°C, 1 min at 55°C, 2 min at 72°C (denaturing, annealing, and extension temperatures, respectively ) with primers 16Sar (5'-CGCCT- GTTTATCAAAAACAT-3'), 16Sbr ( 5'-C CGGTCTGAACTC A- GATCACGT-3'), 1472 ( 5 '- AGATAGAAAC C AAC CTGG-3 ' ) , and 16L15 ( 5'-GACGATAAGACCCTATAAAGCTT-3') (for references to primers see Schubart et al., 2000). 16L15 is an internal primer and, in combination with 16Sbr, was used for partial amplification of the formalin-preserved specimen. PCR products were purified with Microcon-100® filters (Millipore Corp., Bedford, MA) and sequenced with the ABI BigDye® Terminator Mix (PE Biosystems, Welles- ley, MA) in an ABI Prism 310 Genetic Analyzer® (Applied Biosystems, Foster City, CA). Sequences were manually aligned with the multisequence-editing program ESEE (Cabot and Beckenbach, 1989). New sequences were ac- cessioned to the European Molecular Biology Laboratory (EMBL) genomic library (see Table 1). Sequences avail- able online (by electronic database accession numbers that follow) were obtained for “ Callinectes sapidus ” (U75267), C. ornatus (U75268), C. similis (U75269), and C. sapidus (AJ130813). Sequence divergence was analyzed by using Kimura 2-parameter distances, UPGMA cluster analysis, and neighbor-joining (NJ) distance analysis (Saitou and Nei, 1987) with the program MEGA (Kumar et al. 1993). Sta- tistical significance of groups within inferred trees was evaluated by the interior branch method (Rzhetsky and Nei, 1992). As a second phylogenetic method, a maximum parsimony (MP) analysis was carried out with a heuristic search and random sequence addition (tree-bisection and reconnection as branch-swapping option) and by omission of gaps with the program PAUP (Swofford, 1993). Signifi- cance levels were evaluated with the same software and the bootstrap method with 2000 replicates. Results Sequencing of 15 specimens of Callinectes bocourti and C. maracaiboensis from Golfete de Cuare and El Mojan (both Venezuela) and the Rio Sinu estuary (Colombia) revealed the existence of seven different haplotypes. Two haplotypes (ht-1 and ht-5) clearly predominated (together 64.3%) and were found in both species (ht-1 in two Cal- linectes bocourti and three C. maracaiboensis from Golfete de Cuare, ht-5 in three C. bocoui'ti and one C. maracai- boensis from El Mojan). The other five haplotypes (ht-2, ht-3, ht-4, ht-6, ht-7) each differed from ht-1 or ht-5 in not more than two positions and were found in only single individuals (Fig. 1, Table 2). Consequently there is not a single diagnostic molecular character in our sequenced unit of 16S mtDNA that would discriminate between the two species of swimming crabs. On the other hand, hap- lotype distributions differed between the two sampled populations; ht-1 is found in only the Golfete de Cuare (-550 km east of Lake Maracaibo) and ht-5 is restricted to El Mojan (within Lake Maracaibo). Comparison of the Schubart et al.: DNA sequences of swimming crabs Collinectes bocourti and C. maracaiboensis 477 Table 1 Swimming crabs used in sequencing of 16S mtDNA and subsequent phylogenetic analyses. IVIC-LEGP = Laboratorio de Ecologia y Genetica de Poblaciones, Institute Venezolano de Investigaciones Cientificas, Caracas, Venezuela; ULLZ = LTniversity of Louisiana Zoological Collections, Lafayette, U.S.A.; specimens bearing numbers for both collections are archived in ULLZ; EMBL = European Molecular Biology Laboratory. Species Locality of collection Catalog number EMBL accession no. Callinectes maracaiboensis Venezuela; Falcon: Golfete de Cuare IVIC-LEGP-C-40 = ULLZ 4181 AJ298171 Callinectes maracaiboensis Venezuela: Falcon: Golfete de Cuare IVIC-LEGP-C-70 AJ298182 Callinectes maracaiboensis Venezuela: Falcon: Golfete de Cuare IVIC-LEGP-C-7 1 AJ298172 Callinectes maracaiboensis Venezuela: Falcon: Golfete de Cuare IVIC-LEGP-C-72 AJ298173 Callinectes maracaiboensis Venezuela: Zulia: El Mojan IVIC-LEGP-MZ- 1 AJ298177 Callinectes maracaiboensis Venezuela: Zulia: El Mojan IVIC-LEGP-MZ-6 AJ298183 Callinectes bocourti Venezuela: Falcon: Golfete de Cuare IVIC-LEPG-C-30 = ULLZ 4180 AJ298170 Callinectes bocourti Venezuela: Falcon: Golfete de Cuare IVIC-LEGP-C- 109 AJ298174 Callinectes bocourti Venezuela: Falcon: Golfete de Cuare IVIC-LEGP-C-112 AJ298175 Callinectes bocourti Venezuela: Falcon: Golfete de Cuare IVIC-LEGP-C-1 13 AJ298176 Callinectes bocourti Venezuela: Zulia: El Mojan IVIC-LEGP-MZ-2 AJ298181 Callinectes bocourti Venezuela: Zulia: El Mojan IVIC-LEGP-MZ-3 AJ298178 Callinectes bocourti Venezuela: Zulia: El Mojan IVIC-LEGP-MZ-4 AJ298179 Callinectes bocourti Venezuela: Zulia: El Mojan IVIC-LEGP-MZ-5 AJ298180 Callinectes bocourti Colombia: Rio Sinii ULLZ 4186 AJ298169 Callinectes sapidus Florida: Fort Pierce ULLZ 3766 AJ298189 Callinectes sapidus Venezuela: Zulia: El Mojan ULLZ 4188 AJ298190 Callinectes ornatus Venezuela: Falcon: La Vela de Coro IVIC-LEGP-LV-CO-1 AJ298187 Callinectes ornatus Venezuela: Falcon: Golfete de Cuare IVIC-LEGP-CO-3 AJ298188 Callinectes ornatus Brazil: Ensenada de Ubatuba ULLZ 4178 AJ298186 Callinectes danae Brazil: Ensenada de Ubatuba ULLZ 4179 AJ298185 Callinectes danae Venezuela: Falcon: Golfete de Cuare IVIC-LEGP-C-1 AJ298184 Portunus ordwayi Venezuela: Falcon: La Vela de Coro IVIC-LEGP-LV-9 AJ298191 Table 2 Percent (uncorrected) genetic divergence and number of differences between 547 base pairs of 16S mtDNA among 7 haplotypes (ht) corresponding to 14 specimens of Callinectes bocourti and C. maracaiboensis from Venezuela (s: transition, v: transversion, i: indel). Catalog numbers shown below correspond to IVIC-LEGP specimens listed in Table 1. ht-1 ht-2 ht-3 ht-4 ht-5 ht-6 ht-7 C-30 C-70 C-112 C-113 MZ-1 MZ-2 MZ-6 C-40 MZ-3 C-71 MZ-4 C-72 MZ-5 C-109 ht-1 — 0.18 0.18 0.18 0.18 0.55 0.18 ht-2 lv — 0.37 0.37 0.37 0.73 0.37 ht-3 Is Is, lv — 0.37 0.37 0.73 0.37 ht-4 Is Is, lv 2s — 0.37 0.73 0.37 ht-5 Is Is, lv 2s 2s — 0.37 0.37 ht-6 2s, li 2s, li, lv 3s, li 3s, li Is, li — 0.73 ht-7 Is 1 s, lv 2s 2s 2s 3s, li — 478 Fishery Bulletin 99(3) Figure 1 UPGMA ( unweighted pair-group method with arithmetic averaging) analysis of Kimura 2-param- eter genetic distances based on 547 base pairs of 16S mtDNA among seven haplotyopes of Cal- linectes maracaiboensis and C. bocourti (see Table 2 for species designation) corresponding to 14 specimens from Golfete de Cuare (haplotypes 1-4) and El Mojan (haplotypes 5-7) (both Venezu- ela) and two populations of C. sapidus. 294 base pairs obtained for one Callinectes bocourti from Colombia did not reveal any difference from ht-1, ht-3, or ht-5, suggesting that genetic homogeneity in these species can be expected in other Caribbean localities. Comparing sequences from C. danae, C. ornatus , C. sapidus, and Portunus ordwayi with those of C. bocourti and C. maracaiboensis allowed us to postulate phylogenet- ic relationships among these selected portunid taxa (Fig. 2). Marked genetic differences (Table 3) characterized all intra- and interspecific separations within the genus Cal- linectes, except in pairings of C. bocourti and C. maracai- boensis. Discussion Comparison of 15 sequences of the 16S mtDNA of Cal- linectes maracaiboensis and C. bocourti from within and outside Lake Maracaibo revealed no consistent differ- ences between these two species of swimming crabs. These results support observations from previous studies in which morphological characters have proven unreliable in separating these two species. Rather than constituting a separate species, C. maracaiboensis would appear from these findings to represent only a phenotypic extreme of C. bocouj'ti, a species of known morphological plasticity (Wil- liams, 1974). We are highly confident that our sequenced samples of C. maracaiboensis represent the population originally as- signed to that species. In addition to assuring that our samples were topotypic, we confirmed individual specimen identities by direct comparisons with deposited paratypes of Taissoun (IVIC-LEGP 425). However, we cannot make similar claims for the analyzed materials of C. bocourti. Although published records (Williams, 1984) indicate that C. bocourti is widely distributed (from Jamaica and Belize to Brazil, occasionally northward to Florida, Mississippi, and North Carolina), the type locality is southern Belize. To date, topotypic materials of C. bocourti have not been available to us for use in either our sequencing or mor- phological comparisons with Venezuelan material; such comparisons should, however, ultimately be undertaken to firmly anchor the synonymy we have proposed. Among specimens that we would presently assign to C. bocourti on the basis of structural features, one from the Rio Sinu of Colombia shared complete identity in almost 300 base pairs of mtDNA with the most common haplotypes from Venezuela. This finding at the very least suggests a large overall range for materials that we refer to as C. bocourti and a relative homogeneity among forms that we have as- signed to this name on the basis of structural characteris- tics, even over large geographic distances. The reported phylogenetic relationships and genetic dis- tances (Fig. 1) reflect a low level of genetic divergence among the seven haplotypes of C. maracaiboensis and C. bocourti, especially when compared with sequences of two populations of C. sapidus. The tree resulting from the phy- logenetic analysis (Fig. 2) also suggests that C. maracai- boensis and C. bocourti are closer to C. sapidus than to C. ornatus, C. similis, and C. danae, which is in accord with preliminary findings of Norse and Fox-Norse (1979). How- ever, our results must also be regarded as preliminary because other species of Callinectes should be sequenced before we can claim full understanding of sister-species re- lationships. Published studies to date that have attempted to separate questionable brachyuran species on the basis of 16S mtDNA have consistently reported at least a few nucleotide differences that are diagnostic for the species in question (Geller et ah, 1997; Schneider-Broussard et al., 1998; Schubart et al. 1998; Schubart et al., 2000). Only the Mediterranean species Brachynotus sexdentatus and B. gemmellari may constitute an exception to this pattern because they are identical in the sequenced fragment of 16S mtDNA (Schubart et al., in press). Comparison of our sequences with those from a previ- ous study (Geller et al., 1997) revealed important, appar- ently intraspecific differences between specimens of C. or- natus as well as between specimens of C. sapidus. Genetic Schubart et al.: DNA sequences of swimming crabs Callinectes bocourti and C. maracciiboensis 479 Table 3 Percent genetic divergence (uncorrected; excluding indels) between 553 base pairs of 16S mtDNA among species of Callinectes and the outgroup Portunus ordwayi, as used for the phylogeny in Figure 2 (1=C. similis (U75269); 2=“C. sapidus” (U75267); 3=C. sapidus (FL); 4=C. sapidus (Venezuela); 5 =C. bocourti (ht-1); 6=C. bocourti (ht-5); 7=C. danae (Venezuela); 8 =C. danae (Brazil); 9=C. ornatus (U75268); 10=C. ornatus (Venezuela CO-1); 11=C. ornatus (Venezuela CO-3); 12=C. ornatus (Brazil); 13=P. ordwayi). 2 3 4 5 6 7 8 9 10 11 12 13 1 0.18 5.61 5.79 6.33 6.15 0.54 0.54 2.53 3.07 3.07 2.89 13.02 2 5.06 5.24 5.79 5.61 0.54 0.54 2.53 3.07 3.07 2.89 12.12 3 1.27 2.53 2.35 5.79 5.79 5.97 6.87 6.87 6.69 13.92 4 2.71 2.53 6.15 6.15 6.15 7.05 7.05 7.05 13.92 5 0.18 6.51 6.51 6.33 6.51 6.51 6.33 14.10 6 6.51 6.51 6.15 6.33 6.33 6.33 14.29 7 0.00 2.71 3.62 3.62 3.44 13.92 8 2.71 3.62 3.62 3.44 13.92 9 1.63 1.63 1.45 12.66 10 0.36 0.18 12.84 11 0.18 13.02 12 13.38 96 67 r C. similis (U75269) 70\ 98 L “C.sapidus"( U75267) 88 r C danae (Venezuela) ssf C. danae (Brazil) - C. ornatus (U75268) C ornatus (Venezuela CO-1) C. ornatus (Venezuela CO-3) C. ornatus (Brazil) C. sapidus (Florida & AJ130813) C sapidus (Venezuela) C. bocourti (Venezuela ht-1) 99 L C bocourti (Venezuela ht-5) Portunus ordwayi (Venezuela) r, Figure 2 Phylogenetic relationships of five species of Callinectes based on 553 base pairs of 16S mtDNA. Tree topology is based on neighbor-joining (NJ) analy- sis with Kimura 2-parameter genetic distances. Sequences accessed online for C. similis, putative “C. sapidus and C. ornatus , and C. sapidus, are indi- cated solely by electronic database accession numbers. Analyses for Venezue- lan C. ornatus include specimens from adjacent localities (see Table 1); those for C. bocourti include haplotypes (ht) 1 and 5 (see Table 2 and Fig. 1). Con- fidence values >50 were obtained with the interior branch method for NJ distance analysis (upper values) and by maximum parsimony analysis after 2000 bootstrap replicates (lower values, in Italics). differences between Geller et al.’s (1997) specimen from North Carolina (U75268) and three specimens of C. ornatus from South America can be attributed to geo- graphical distance. In the case of C. sapi- dus, however, differences clearly exceed a level that could be interpreted as an- cestral polymorphism or biogeographical variation. Although our new sequence of C. sapidus from Florida perfectly matches one previously registered for a specimen from Louisiana (AJ130813), and both of these show limited divergence from a Ven- ezuelan specimen, another sequence pre- viously reported for C. sapidus (U75267) by Geller et al. (1997) differs markedly from the aforementioned set. The phylo- geny of all 16S sequences presently avail- able for Callinectes (Fig. 2) shows that the sequence reported for “C. sapidus ” by Geller et al. (1997) instead pairs closely with C. similis, a grouping supported by high bootstrap values. Because no mor- phological voucher specimens appear to exist for this sequenced specimen,1 we must conclude that the reported sequence was based upon a misidentified specimen of C. similis. Differences in color of some of the Bra- zilian specimens (see Sankarankutty et al., 1999) could indicate that two or more species might be involved in the C. bocour- ti complex, regardless of the synonymy that we have herein 1 Geller, J. 1999. Personal commun. Moss Landing Marine Laboratories, Moss Landing, CA 95039. proposed. However, color differences should be judged with precaution, given the known high variability of this fea- ture in C. bocourti (see Williams, 1984). Our recent color photography of specimens from Colombia, Venezuela, and 480 Fishery Bulletin 99(3) Florida has revealed marked variations; dorsal surfaces of the carapace of some specimens were predominantly olive green, those in other specimens were largely rust brown or rust patterned on green, and in yet others a rust pattern on a background of almost totally blue. As a result of our molecular analyses, especially in the absence of consistent characters to distinguish C. maracai- boensis on the basis of color or structural characteristics, we must conclude that Lake Maracaibo and adjacent Ven- ezuelan waters are inhabited by populations assignable to C. sapidus and by only one other species of the genus, C. bocourti. It appears that these are in turn the target spe- cies of a largely unregulated crab fishery which represented more than U.S. $6 million in exports to the United States in 1992 (Oesterling and Petrocci, 1995), even though it has de- clined since 1989 in probable response to overharvesting. As maximum sustainable yield has likely been achieved, these fisheries, which have been proposed for Miller’s phase II management (resource-mapping, gear related regulations), should in the near future proceed to phase III (basic bio- logical studies, long-term management) (Miller, 1999). This phase can be undertaken only with a clear understanding of the genetic entities upon which the fishery is based. Acknowledgments We thank Fernando Mantelatto for making available Bra- zilian specimens of Callinectes, Waldo Querales and Jesus Arteaga for assistance in field work, and Arnaldo Ferrer and associates at PROFAUNA (MARNR), Cuare, for pro- viding accommodations at the station. We are indebted to Joseph E. Neigel for sharing his laboratory facilities and to Gilberto Rodriguez and Hector Suarez for providing advice, specimens, and convenient access to IVIC’s crusta- cean collection. This study was supported in part by the U.S. Department of Energy (grant DE-FG02-97ER12220). Literature cited Arnold, W. S. 1984. The effects of prey size, predator size, and sediment composition on the rate of predation of the blue crab, Cal- linectes sapidus Rathbun, on the hard clam, Mercenaria mercenaria (Linne). J. Exp. Mar. Biol. Ecol. 80:207-219. Buchanan, B. A., and A. W. Stoner. 1988. Distributional patterns of blue crabs ( Callinectes sp.) in a tropical estuarine lagoon. Estuaries 11:231-239. Cabot, E.L. and A.T. Beckenbach. 1989. Simultaneous editing of multiple nucleic acid and protein sequences with ESEE. Comput. Appl. Biosci. 5:233-234. Carmona-Suarez, C. A., and J. E. Conde. 1996. Littoral brachyuran crabs (Crustacea: Decapoda) from Falcon, Venezuela, with biogeographical and ecological remarks. Rev. Brasil. Biol. 56:725-747. Conde, J. E., and G. Rodriguez. 1999. Integrated coastal zone management in Venezuela: the Maracaibo system. In Perspectives on integrated coastal zone management: principles and practice (W. Salo- mons, K. Turner, L. D. Lacerda, and R. Ramachandram, eds.), p. 297-312. Environmental Science Series. Springer- Verlag, Berlin, xviii+386 p. Ferrer-Montano, O. J. 1997. Effectiveness of two pots and other factors for har- vesting hard blue crabs Callinectes sapidus in Lake Mara- caibo, Venezuela. Ciencia 5:111-118. Geller, J. G., E. D. Walton, E. D. Grosholz, and G. M. Ruiz. 1997. Cryptic invasions of the crab Carcinus detected by molecular phylogeography. Mol. Ecol. 6:901-906. Kocher, T. D., W. K. Thomas, A. Meyer, S. V. Edwards, S. Paabo, F. X. Villablanca, and A. C. Wilson. 1989. Dynamics of mitochondrial DNA evolution in ani- mals: Amplification and sequencing with conserved prim- ers. Proc. Natl. Acad. Sci. (USA) 86:6196-6200. Kumar, S., K. Tamura, and M. Nei. 1993. MEGA: Molecular evolutionary genetics analysis, ver- sion 1.01. The Pennsylvania State Univ., University Park, PA. Lin, J. 1991. Predator-prey interactions between blue crabs and ribbed mussels living in clumps. Estuarine Coast. Shelf Sci. 32:61-69. Miller, R. J. 1999. Courage and the management of developing fisheries. Can. J. Fish. Aquat. Sci. 56:897-905. Norse, E. A. 1977. Aspects of the zoogeographic distribution of Callinectes (Brachyura: Portunidae). Bull. Mar Sci. 27:440-447. Norse, E. A., and V. Fox-Norse. 1979. Geographical ecology and evolutionary relationships in Callinectes spp. (Brachyura: Portunidae). Proceedings of the Blue Crab Colloquium, Gulf States Marine Fisheries Commission 7(1982):l-9. Oesterling, M. J., and C. Petrocci. 1995. The crab industry in Venezuela, Ecuador and Mexico. Virginia Sea Grant Marine Advisory Program, Gloucester Point, VA, and Maryland Sea Grant Extension Program, College Park, MD. Orth, R. J., and J. van Montfrans. 1987. Utilization of a seagrass meadow and tidal marsh creek by blue crabs Callinectes sapidus. 2. Seasonal and annual variations in abundance with emphasis on post-set- tlement juveniles. Mar. Ecol. Prog. Ser. 41:283-294. Rodriguez, G. 1980. Crustaceos decapodos de Venezuela. Instituto Vene- zolano de Investigaciones Cientificas, Caracas, 494 p. Rzhetsky, A., and M. Nei. 1992. A simple method for estimating and testing mini- mum-evolution trees. Mol. Biol. Evol. 9:945-967. Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method for recon- structing phylogenetic trees. Mol. Biol. Evol. 4:406-425. Sankarankutty, C., A. C. Ferreira Roman, C. S. Callado Pinto, F. E. N. Varela Barca, and M. D. A. Alencar. 1999. Callinectes maracaiboensis Taissoun (Crustacea, Decapoda, Portunidae), a species common but so far un- recorded in the northeast of Brazil. Rev. Bras. Zool. 16: 145-150. Sarver, S. K., J. D. Silberman, and P. J. Walsh. 1998. Mitochondrial DNA sequence evidence supporting the recognition of two subspecies or species of the Florida spiny lobster Panulirus argus. J. Crust. Biol. 18:177-186. Schneider-Broussard, R., D. L. Felder, C. A. Chian, and J. E. Neigel. 1998. Tests of phylogeographic models with nuclear and mitochondrial DNA sequence variation in the stone crabs. Schubart et al.: DNA sequences of swimming crabs Callinectes bocourti and C. maracaiboensis 481 Menippe adina and M. mercenaria. Evolution 52:1671- 1678. Schubart, C. D., J. A. Cuesta, and A. Rodriguez. In press. Molecular phylogeny of the crab genus Brachyno- tus (Brachyura: Varunidae) based on the 16S rRNA gene. Hydrobiologia. Schubart, C. D., J. E. Neigel and D. L. Felder. 2000. The use of the mitochondrial 16S rRNA gene for phy- logenetic and population studies of Crustacea. In The bio- diversity crisis and Crustacea. Proceedings of the fourth international crustacean congress, Amsterdam, Netherlands, 20-24 July 1998, vol. 2. Crustacean Issues 12:817-830. Schubart, C. D., J. Reimer, and R. Diesel. 1998. Morphological and molecular evidence for a new endemic freshwater crab, Sesarma ayatum sp. n.’ (Grap- sidae, Sesarminae) from eastern Jamaica. Zool. Scr. 27:373-380. Swofford, D. L. 1993. Phylogenetic analysis using parsimony (PAUP), ver- sion 3.1.1. Univ. Illinois, Champaign, IL. Taissoun, E. 1969. Las especies de cangrejos del genero Callinectes ( Brach- yura) en el Golfo de Venezuela y Lago de Maracaibo. Bol. Centro Invest. Biol. Univ. Zulia. 2:1-102. 1972. Estudio comparative, taxonomico y ecologico entre los cangrejos (Decapoda: Brachyura: Portunidae), Call- inectes maracaiboensis (nueva especie), C. bocourti (A. Milne Edwards) y C. rathbunae (Contreras) en el Golfo de Venezuela, Lago de Maracaibo y Golfo de Mexico. Bol. Centro Invest. Biol. Univ. Zulia 6:7-46. Williams, A. B. 1974. The swimming crabs of the genus Callinectes. Fish. Bull. 72:685-798. 1978. True crabs. In FAO species identification sheet for fishery purposes: western central Atlantic (fishing area 31), vol. 6 (W. Fischer, ed.), unpaginated. Food and Agricul- tural Organization of the United Nations, Rome. 1984. Shrimps, lobsters, and crabs of the Atlantic Coast of the eastern United States, Maine to Florida. Smithsonian Institution Press, Washington D.C., xviii+550 p. Wilson, K. A„ K. L. Heck Jr., and K. W. Abele. 1987. Juvenile blue crab, Callinectes sapidus, survival: An evaluation of eelgrass, Zostera marina, as refuge. Fish. Bull. 85:53-58. Abstract— Diets of young-of-the-year (YOY) walleye pollock ( Theragra chal- cogi'amma) and Pacific herring (Clupea pallasi ) were compared between sea- sons (summer and autumn), years (autumn), and allopatric and sympatric fish aggregations (autumn) in Prince William Sound (PWS), Alaska. Fish were collected principally by mid-water trawl 20 July-12 August 1995, 5-14 October 1995, and 7-13 November 1994. Prey fields were assessed from zoo- plankton samples in 1995. During the summer, the principal prey of allopatric pollock and herring was small calanoids and diet overlap was high (7?0>0.76). During the autumn, diets were composed of large calanoids, larvaceans, and euphausiids. Diet over- lap between sympatric species was greater in November 1994 (f?0<0.94) than in October 1995 (R0< 0.69). The seasonal diet shift to larger prey coin- cided with larger fish size and with decreased abundance and proportions of the principal zooplankter, small cal- anoids, and increased abundance and proportions of large calanoids and lar- vaceans in zooplankton tows. However, feeding decreased in autumn, compared with summer, especially for herring. Sympatric fish had higher rates of non- feeding than allopatric fish, and subtle differences in prey selection existed between the aggregations, but sampling variation could explain these feeding differences. The similarity in diets of YOY pollock and herring indicate the potential for competition. These species are impor- tant to commercial fisheries and as forage for marine birds and mammals. An understanding of their trophic inter- actions could help to explain shifts in fish community structure and bird pre- dation. If sympatry increases as prey resources decline, competition in au- tumn may be particularly important in regulating populations. Manuscript accepted 18 December 2000 Fish. Bull. 99:482-501 (2001). Feeding habits, prey fields, and potential competition of young-of-the-year walleye pollock ( Theragra chalcogramma) and Pacific herring (Clupea pallasi ) in Prince William Sound, Alaska, 1994-1995 Molly V. Sturdevant Auke Bay Laboratory Alaska Fisheries Science Center 11305 Glacier Highway Juneau, Alaska 99801 E-mail address: Molly.Sturdevant@noaa.gov Audra L. J. Brase Commercial Fisheries Division Alaska Department of Fish and Game PO Box 20 Douglas, Alaska 99824 Leland B. Hulbert Auke Bay Laboratory Alaska Fisheries Science Center 11305 Glacier Highway Juneau, Alaska 99801 Walleye pollock ( Theragra chalcogram- ma ) and Pacific herring (Clupea pallasi) are forage fish that inhabit the north- eastern Pacific Ocean rim. Both species are important components of marine bird, mammal, and fish diets, both sup- port important commercial fisheries in the Gulf of Alaska (GOA), and histori- cal data show dramatic variability in both their populations (Springer, 1992; Bechtol, 1997; Springer and Speckman, 1997; Anderson and Piatt, 1999). Young- of-the-year (YOY) walleye pollock and YOY Pacific herring are found at the same locations and depths during at least part of the year (Brodeur and Wilson, 1996a; Willette et ah, 1997; Stokesbury et ah, 2000) and both con- sume zooplankton as their primary prey1’2 (Willette et ah, 1997; Foy and Norcross, 1999a, 1999b). Because of these similarities and because the fre- quency and nature of their interactions may change as fish population struc- ture shifts, we investigated the poten- tial for feeding competition between these species. Several recent studies have found that the species composition of forage fish populations in the GOA and Prince William Sound (PWS) has changed dra- matically (Bechtol, 1997; Kuletz et ah, 1997; Anderson and Piatt, 1999). Short- term population changes were attribut- ed to the Exxon Valdez oil spill in March 1989 (Brown et ah, 1996a; Kuletz, 1996; Oakley and Kuletz, 1996; Kuletz et ah, 1 Jewett, S., and E. Debevee. 1999. Diet composition, diet overlap and size of 14 species of forage fish collected monthly in Prince William Sound, Alaska, 1994-1996. In Forage fish diet overlap, 1994-1996, p. 10-37. Exxon Valdez Oil Spill Restoration Project Final Report (Restoration Project 971630, Auke Bay Laboratory, National Marine Fisheries Ser- vice, 11305 Glacier Hwy., Juneau, AK 99801-8626. 2 Boldt, J. L. 1997. Condition and distri- bution of forage fish in Prince William Sound, Alaska. Unpubl. M.S. thesis, Juneau Center School of Fisheries and Ocean Science, Univ. Alaska Fairbanks, 11120 Glacier Hwy., Juneau, AK 99801, 155 p. Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragra cha/cogrammci and Clupea pallasi 483 148° W 147° W Sampling regions and stations for YOY walleye pollock and Pacific herring diet sam- ples collected in Prince William Sound, Alaska. Circles: July-August 1995; squares: October 1995; triangles: November 1994. 1997), but researchers believe that long-term changes be- gan with a climate regime shift prior to the oil spill (Bailey et al., 1995; Piatt and Anderson, 1996; Anderson and Pi- att, 1999). For example, the number of walleye pollock and other demersal fish increased at the same time that the other taxa decreased (Anderson and Piatt, 1999), the PWS spawning population of Pacific herring declined by 75% by 1993 (Brown et al., 1996b), and fish biomass of PWS de- creased by 50% (Piatt and Anderson, 1996). Concomitant decreases in populations of marine birds and mammals in PWS may be related to these shifts in the composition and abundance of forage fish prey (Oakley and Kuletz, 1996; Piatt and Anderson, 1996; Iverson et al., 1997). Apparent- ly, fewer high-quality forage fish have been available, and the species composition has changed to one in which the predominant fish taxa are less energetically valuable to marine piscivores (Piatt and Anderson, 1996; Anthony and Roby, 1997; Anderson and Piatt, 1999; Payne et al., 1999). The potential for such shifts to cascade throughout ma- rine food webs (Livingston, 1993; Springer and Speckman, 1997) is an important reason to understand the trophic in- teractions of forage fish. This reports stems from the Alaska Predator Ecosystem Experiment (APEX), a multidisciplinary study that at- tempted to link current knowledge about the forage fish of PWS with their seabird predator populations. We describe differences in the feeding habits of YOY walleye pollock and YOY Pacific herring caught in summer and autumn in PWS and compare feeding attributes of fish caught in al- lopatric (single species) and sympatric (co-occurring, mul- tispecies) aggregations in autumn to support the hypoth- esis that the presence of potential competitors may induce changes in feeding habits. We compare fish size, zooplank- ton prey fields, fish feeding habits, prey selection, food quantity, and diet overlap of these species. Materials and methods Field methods Fish stomach and zooplankton samples were collected during APEX forage fish population surveys in central, northeastern, and southwestern PWS3,4 5 (Fig. 1). In a pilot study in 1994, we sampled from 7 to 13 November aboard the Alaska Department of Fish and Game RV Medeia\ in 1995, we sampled from 20 July to 12 August aboard the charter FV Caravel le, and from 5 to 14 October aboard the RV Medeia. Surveys were conducted offshore along a grid of parallel transects spaced at two-mile 484 Fishery Bulletin 99(3) intervals and ending as near shore as possible. Bottom depths averaged approximately 120 m (range: 25-220 m). The grid was surveyed twice in summer and once, par- tially, each autumn. Hydroacoustic and hydrographic pro- file data were collected but are presented elsewhere.3 4'3 4 5 Where fish were detected with hydroacoustic equipment, we either interrupted the survey or returned after the tran- sect was completed to fish with a mid-water beam trawl. The net was generally fished 20-35 minutes each trawl.3 4 The trawl’s effective mouth opening was 50 m2, and net mesh sizes diminished from 5 cm in the wings to 1 cm in the codend. A 0.3-cm mesh liner was sewn into the codend, which terminated in a plankton bucket having 500-pm nytex mesh. In summer, beach-seine and dip-net samples occasionally supplemented the trawl catches. Subsamples of forage species (72 = 10 to 15 per species) were preserved in 10% buffered formalin-seawater solution on the vessels for later stomach analysis in the laboratory. We classified samples collected between 08:00 and 20:00 as “day” and those between 20:01 and 07:59 as “night.” In 1995, the zooplankton prey spectrum was assessed from dual vertical hauls taken at each station within two hours of fish catches by using conical nets that were 0.5 m in diameter and equipped with 303-pm mesh in summer and 243-pm mesh in autumn. We towed the nets from a standard depth of 20 m or to the depth at which fish were caught (or using a combination of both depths). Depth of hauls were categorized as “shallow” (<25 m) or “deep” ( >25 to <100 m). Samples were collected at both depths at seven stations in summer (from 95-1-53 to 95-1-62 and 95-1-112) and one station in autumn (95-2-7, Table 1). Laboratory methods After a minimum of six weeks in formalin solution, fish samples were transferred to a solution of 50% isopropa- nol for at least 10 days before stomach analysis was per- formed. Ten specimens of each species were measured (mm fork length, FL; mg wet weight), and size was used to develop age-class categories for diet samples (Smith, 1981; Paul et al., 1998a). Walleye pollock (20 to 120 mm FL) and 3 Haldorson, L. 1995. Fish net sampling. In Forage fish study in Prince William Sound, Alaska, p. 55-83. Exxon Valdez Oil Spill Restoration Project 94163A Annual Report, Juneau Center School of Fisheries and Ocean Science, Univ. Alaska Fairbanks, 11120 Glacier Hwy., Juneau, AK 99801. 4 Haldorson, L. J., T. C. Shirley, and K. O. Coyle. 1996. Bio- mass and distribution of forage species in Prince William Sound. In APEX project: Alaska predator ecosystem experi- ment in Prince William Sound and the Gulf of Alaska (D. C. Duffy, compiler), Exxon Valdez Oil Spill Restoration Project Annual Report (Restoration Project 95163 A-Q), Alaska Nat- ural Heritage Program, Department of Biology, Univ. Alaska Anchorage, 707 A Street, Anchorage, AK 99501 5 Haldorson, L. T. Shirley, K. Coyle, and R. Thorne. 1997. For- age species studies in Prince William Sound. In Alaska preda- tor ecosystem experiment in Prince William Sound and the Gulf of Alaska (D. C. Duffy, compiler). Exxon Valdez Oil Spill Resto- ration Project Annual Report (Restoration Project 96163 A-Q), Alaska Natural Heritage Program, Department of Biology, Univ. Alaska Anchorage, 707 A Street, Anchorage, AK 99501. Pacific herring (60 to 120 mm FL) were classified as YOY (age-class 0). Stomachs were excised, weighed, and their contents were removed. The weight of prey contents was recorded as the difference between full and empty stom- ach weights. Fish were considered to have been feeding if their stomachs contained more than a trace of food. Rela- tive stomach fullness was recorded as integers represent- ing empty stomachs (1), stomach containing trace contents (2), stomachs that were 25%, 50%, 75%, or 100% full (3-6), or stomachs that were distended (7). State of digestion was recorded as partially digested contents (1), mostly digested contents (2), and empty stomachs (3). Stomach contents and zooplankton samples were identi- fied with a binocular microscope to the highest taxonomic resolution possible and enumerated. The prey category of calanoid copepods was also segregated into “large” (>2.5 mm total length, TL) and “small” individuals (>2.5 mm TL). We pooled the common pelagic cyclopoid copepod Oi- thona similis with small calanoids. We subsampled all zoo- plankton samples and stomach samples when practical, using a Folsom splitter to achieve a minimum count of 200 of the predominant taxon. Counts were expanded and to- tal prey weights were determined by multiplying the ex- panded number observed by the mean weight per taxon. Weights per taxon were obtained from data on file (speci- mens from zooplankton samples or fish stomachs) collect- ed from spring through autumn of various years in south- eastern Alaska or PWS (Coyle et ah, 1990; Stark6; senior author, unpubl. data). Analytical methods Forage fish were considered to occur in allopatric aggre- gations if only one species and one age class were caught in a net haul. They were considered to be sympatric if at least two species or age classes (>10 fish each) were caught together. For this study, we restricted analyses to YOY pollock and herring that were allopatric or that co- occurred only with each other to limit the complexity of trophic interactions; we excluded pollock and herring that were caught in other types of aggregations, such as with other species or older conspecifics. We examined the size of forage fish and their feeding attributes. Size included FL and wet weight. Feeding attributes included measures of the quantity of food consumed, measures of feeding fre- quency, and measures of prey composition. Food quantity was expressed as means of the total number and weight of prey (ln-transformed), stomach fullness index (rounded to nearest 25%), and prey percent body weight (%BW; ratio of wet stomach-content weight to fish body weight). Feeding frequency was measured as the percentages of feeding fish and the percentages of fish with partially or mostly digested stomach contents. Prey composition was expressed as the percent number and percent biomass of prey categories. Zooplankton density per cubic meter and numerical percent composition was calculated for species, 6 Stark, C. 1995. Personal commun. Institute of Marine Sci- ence, Univ. Alaska Fairbanks, P. O. Box 757220, Fairbanks, AK 99775-7220. Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragra chalcogramma and Clupea pallasi 485 Table 1 Mean zooplankton density and percent density (standard errors, SE, in parentheses) from pooled vertical hauls in summer (77=37 hauls, 303-pm mesh) and autumn (n= 8 hauls, 243-pm mesh), and from shallow (< 25 m) and deep (50-100 m) hauls in Prince Wil- liam Sound, 1995. Prey category Mean density (number of zooplankton/m3 ) Summer Autumn Overall Shallow Deep Overall Shallow Deep Barnacle larvae <1 <1 <1 <1 <1 <1 Calanoids, large 29 (4) 37 (7) 28 (6) 204 (60) 104(3) 176(3) Calanoids, small 1018 (133) 1175(181) 550 (87) 828 (130) 685 (48) 426 (40) Chaetognaths 4 (<1) 3(1) 4 (1) 3 (<1) 1 (<1) 5 (<1) Cladocera 33(9) 25 (9) 6(1) 0 0 0 Cyphonautes larvae 0 0 0 205 (50) 297 (7) 333(8) Decapod larvae 1 (<1) 2 (<1) <1 0 <1 0 Euphausiid larvae 3(1) 3(1) 2 (<1) 2 (<1) 1 (<1) 2(1) Gastropods 60 (10) 99(23) 36(5) 96 (19) 141 (4) 84(5) Hyperiid amphipods 2 (<1) 2 (<1) 2 (<1) 1 (<1) 1 (<1) 1 (<1) Larvaceans 14 ( 4) 6(2) 8(3) 45(11) 50 (12) 22(7) Other 17 (2) 17 (3) 8(1) 14 (2) 18(6) 9(1) TotaP 1184(138) 1371 (191) 645 (91) 1414 (185) 1299 (64) 1064 (56) Percent composition Summer Autumn Prey category Overall Shallow Deep Overall Shallow Deep Barnacle larvae 0 0 0 0 0 0 Calanoids, large 3 3 5 13 8 17 Calanoids, small 84 84 84 58 53 40 Chaetognaths <1 <1 <1 <1 <1 <1 Cladocera 5 2 1 0 0 0 Cyphonautes larvae 0 0 0 16 23 14 Decapod larvae 0 0 0 0 0 0 Euphausiid larvae <1 <1 <1 <1 <1 <1 Gastropods 6 7 6 7 11 8 Hyperiid amphipods <1 <1 <1 <1 <1 <1 Larvaceans <1 <1 <1 2 4 2 Other 2 1 1 1 1 1 1 Numbers do not add to column totals because of rounding. principal prey taxa, and total organisms in each vertical tow by using the expanded organism count divided by the water volume of the tow; mean values of replicate tows were used to represent each station. Feeding selectivity of allopatric and sympatric aggrega- tions of pollock and herring was calculated for summer and early autumn, 1995, when zooplankton were collected at the fish sampling stations. Occasionally, in summer, zoo- plankton samples from adjacent stations were substituted for those fishing stations without prey samples (Table 1). At stations where zooplankton was collected at two depths, the selection values presented were based on zooplankton from the depth where fish diet samples were collected. We used Strauss’ linear selection index, L0 (Strauss, 1979), a measure varying from -1 to +1, where negative values in- dicate no preference for the prey taxon and positive values indicate preference for the prey taxon: Lo = r, ~Pi> where r; = percentage of 7th prey resource in the diet; and p( = percentage of 7th prey resource in the environ- ment. Prey resources for selection were defined as the species, stages, and sizes of prey pooled into principal taxa. 486 Fishery Bulletin 99(3) Feeding overlap between species and within species be- tween fish in allopatric and sympatric aggregations was described by using Horn’s overlap index (Horn, 1966; Smith and Zaret, 1982; Krebs, 1989). This index minimiz- es bias due to changing numbers of resource categories and resource evenness. Overlap was computed at two lev- els: prey resources were defined at the lowest level (spe- cies, stage, and size) and at a pooled level (principal taxa). Horn’s overlap index values, Rq, were expressed from 0 (no overlap) to 1 (total overlap) for predator species j and k: _ X(p" + p'k ) x ln< p" + Pik ] - X p" x ln p' _ X p'k x ln p"> 2 x In 2 where ptj = proportion of zth resource in total prey re- sources utilized by jth species; and P,k = proportion of zth resource in total prey re- sources utilized by kth species. We considered R0 values >0.60 to indicate similar use of resources and R0 >0.75 to indicate very similar use of resources. In this report, we examined both intraspecific and inter- specific differences in YOY pollock and herring and quali- tatively compared data between seasons, between years (in autumn), between allopatric and sympatric aggrega- tion types, and between sympatric species. We also com- pared day-night feeding frequencies and prey condition to assess principal time of day of feeding. We compared zoo- plankton densities and percent densities of principal taxa between seasons and between depths (shallow and deep) sampled each season. Because our data were limited and the sampling design unbalanced, we present means and standard errors without statistical tests. Results Zooplankton prey fields Total zooplankton densities for all samples pooled were similar in summer and autumn of 1995, averaging approximately 1200 and 1400 organisms/m3 (n= 37 and 8, mesh=303 pm and 243 pm, respectively; Table 1, Fig. 2). Approximately 3/4 of samples from each season were col- lected in daylight, between 10:00 and 20:00. Zooplankton taxa were less diverse in summer than in autumn, but small calanoids predominated in both seasons. Seasonal differences in the density and percentage contri- bution of a few taxa were apparent. Small calanoids made up a greater percentage of the total in summer compared with autumn (84% vs. 58%), but their density was similar in each season. Large calanoid density and percent density were both lower in summer than in autumn. Only a few other taxa contributed >5% to the total numbers of zoo- plankters in either season. In summer, these included cla- docera and gastropods (primarily the pteropod Limacina helicina); in autumn, they included large calanoids, bryo- zoan cyphonautes larvae, and gastropods (Table 1, Fig. 2). Calanoid species composition in the zooplankton also varied seasonally. Small calanoids in summer were pre- dominantly Pseudocalanus spp. (75%) and Acartia lon- giremis (5%); in autumn, Pseudocalanus spp. was 40%, Oithona similis, 15%, and A. longiremis, 6% of total zoo- plankton. Large calanoids in summer were principally Neocalanus and Calanus spp. (3% of total zooplankton) and in autumn, they were Metridia pacifica (10%) and Calanus spp. (1%). Among other taxa present in both sea- sons, density and percent density of gastropods were simi- lar between seasons, but larvaceans were less available in summer than in autumn. Cladocera were present only in summer and cyphonautes only in autumn. Hyperiid am- phipods, euphausiid larvae, chaetognaths, and barnacle and decapod larvae were rare in both seasons (Table 1). Zooplankton density per cubic meter varied with depth of sampling at all stations for both seasons. In summer (n= 7 stations), zooplankton density was more than twice as high in shallow hauls as in deep hauls, whereas in au- tumn (?z=l station), it was greater in the shallow tows than in the deep tows (Table 1). Depth-related differences in the abundance of principal taxa, but not in their per- centage composition, were also observed. Overall, in sum- mer, small calanoids were twice as abundant in the more shallow tows, but they comprised similar percentages at each depth. In autumn, both density and percentage of small calanoids were greater in shallow tows than in deep- er tows. Large calanoid density and percentage did not dif- fer between depths in summer, but in autumn they were lower in shallow tows compared with deep tows. Gastro- pod abundance in both seasons, and cladoceran abundance in summer, were greater in shallow tows than in deep tows, but percentages did not differ with depth. Larva- cean abundance and percentage abundance in either sea- son, and the abundance and percentage abundance of cy- phonautes in autumn, did not differ between depths. The macrozooplankters, larval euphausiids and young hyperiids, were present, but rare in our plankton tows. However, euphausiids were captured in 11% of summer trawls and in 43% of autumn trawls. In summer of 1995, 4 juvenile-adult euphausiids were present in the northeast- ern and southwestern regions of the sound but were not caught in the central region or at any stations where fish were caught; in autumn ( 19947 and 19954), they were pres- ent in trawls from all three regions but were absent from the allopatric trawl in October 1995. We have no consistent data on the presence or absence of hyperiid amphipods. Fish catches Catches of YOY walleye pollock and Pacific herring sam- pled from PWS that met our allopatric-sympatric criteria (see “Materials and methods” section) were not evenly rep- 7 Paul, A. J. 1995. Invertebrate forage species. In Forage fish study in Prince William Sound, Alaska, p. 43-54. Exxon Valdez Oil Spill Restoration Project 94163A Annual Report, Juneau Center School of Fisheries and Ocean Science, Univ. Alaska Fairbanks, 11120 Glacier Hwy., Juneau, AK 99801. Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragrc chalcogrcimma and Clupea pallasi 487 1 600 t 1400-- 0) CD Summer Autumn □ Barnacle larvae Large calanoids Small calanoids Larvaceans Euphausiids | Decapod larvae Bryozoan larvae □ M Cladocera Chaetognaths Gammarld amphipods ^ Other Malacostraca □ Cnidarians and ctenophorans □ Hyperiid amphipods [S3 Gastropods Figure 2 Seasonal zooplankton density and composition by principal taxa in Prince William Sound, Alaska 1995, Zooplankton were collected in ver- tical tows of conical nets having 303-pm mesh in summer and 243-mm mesh in autumn. resented in both seasons. Only allopatric fish qualified in the summer of 1995, but both allopatric and sympatric fish qualified in the autumns of 1994 and 1995 (Table 2). In summer, 18 of the 62 trawl hauls caught sufficient sam- ples of either species.4 Allopatric YOY pollock were col- lected at 12 stations in the central region and allopatric YOY herring were collected at one central and one north- eastern station in summer (Fig. 1). In autumn, 11 of the 25 trawl hauls (total 14 in November 1994, 3 11 in October 19954) caught sufficient samples of either species, and 36% of these caught sympatric fish in the northeastern region. Allopatric species were collected in different years: the allopatric pollock in November 1994 were collected in the southwestern region and the allopatric herring in October 1995 in the central region (Table 2). The remainder of pol- lock and herring caught did not meet our study criteria for age, number, or composition of fish per haul. The magnitude and relative species composition of the catch varied considerably, and sampling time and depth differed between seasons. In summer, the number of pol- lock caught in trawls varied by two orders of magnitude between stations, from 22 to 1689 per haul (Table 2). Her- ring catches in the alternative gear, dip net or beach seine, were of similar magnitude. In autumn, from 14 to 4156 pollock or herring were caught per trawl haul; the catches were not evenly partitioned between the species in sym- patric hauls. Similar numbers of pollock and herring were caught at some sympatric stations, whereas, at others, the number per species caught was an order of magni- tude greater. In summer, approximately half the samples were collected in the morning, half in the afternoon, and only one was collected at dusk (-22:00). The pollock were caught most often at 50-80 m trawl depths (Y=60 m) off- shore and the herring were caught at the surface and near shore. In autumn, sympatric samples were collected in darkness (approximately 22:00-23:00) at a mean depth of 30 m in bays, whereas allopatric samples were collected during daylight and in deeper water (Table 2). 488 Fishery Bulletin 99(3) Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragra chalcogramma and Clupea pallcisi 489 Table 3 Seasonal fish size and feeding attributes of allopatric and sympatric YOY walleye pollock and Pacific herring from Prince William Sound, 1994-1995. Values are mean and standard error, SE, for fish pooled from stations shown in Table 1. Species and sampling period n FL Wet weight 0.76) between allopatric species in terms of num- bers and weights of prey species or principal prey taxa. In autumn, interspecific diet overlap was also observed 490 Fishery Bulletin 99(3) allopatric allopatric sympatric sympatric Barnacle larvae Large calanoids Small calanoids Larvaceans Euphausiids □ M Decapod larvae Bryozoan larvae Cladocera Chaelognaths Gammarld amphipods □ Other Malacostraca Cnidarians and ctenophorans Hyperiid amphipods Gastropods Figure 3 Percent total number of prey consumed by YOY walleye pollock and Pacific her- ring from sympatric and allopatric aggregations in Prince William Sound, Alaska, in July-August 1995, October 1995, and November 1994. in terms of prey numbers and weights. Numeric overlap for the November 1994 sympatric species was observed at both stations, and the mean was approximately twice that for the October 1995 sympatric species (i?0= 0.97 versus 0.43). Biomass overlap between sympatric fish occurred at three out of four autumn stations when based on princi- pal prey taxa, but only at one station (in November) when based on prey species. Overall, diets of sympatric pollock and herring overlapped more in terms of biomass in No- vember (i?0=0.95) than in October (J?0=0.69). We did not have autumn samples from the same year to compare al- lopatric diet overlap between species. Prey selection We noted selection from zooplankton prey resources by pollock in summer and by pollock and herring in autumn, 1995 (Fig. 5). Selection patterns were almost identical whether calculated from shallow or deep zooplankton abundances; values were always in the same direction (selection or avoidance), and their magnitude was within 10 points. Neither species selected small calanoids, the most abundant zooplankton taxon in both seasons. In summer, even though small calanoids made up >50% of YOY pollock and herring diets (Figs. 3 and 4), this taxon was avoided by pollock and was consumed in close propor- tion to its availability by herring (Fig. 5). Summer pollock moderately selected for large calanoids, gastropods, and larvaceans, but summer herring did not strongly select for any prey category. Changes in zooplankton composition from summer to autumn were reflected in fish diets and prey selection. The percent density of small calanoids in the zooplankton de- clined by nearly 30% from summer to autumn, the per- centage consumed by fish was likewise much reduced, and selection values were more strongly negative. In contrast, both large calanoids and larvaceans were more abundant in zooplankton samples in autumn than in summer and, Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragra cho/cogramma and Clupeo pallasi 491 along with euphausiids and hyperiids, formed larger di- etary components at that time. In autumn (October 1995), pollock strongly selected for large calanoid copepods and herring strongly selected for larvaceans (Figs. 3-5). Addi- tional differences were noted in terms of prey frequency. In summer, pollock consumed hyperiids more frequently than they consumed euphausiids, and in autumn, both pollock and herring consumed euphausiids more frequent- ly than hyperiids. Seasonal feeding Several measures of feeding in autumn indicated that YOY pollock continued to feed moderately, whereas the pattern for YOY herring showed more of a decline from summer (Tables 1 and 3). Very similar high percentages of pollock and herring had fed in summer and in October 1995 (>80%), but in November 1994, the percentage of nonfeeding herring was greater than the percentage of feeding herring, and was more than twice the percentage of nonfeeding pollock (Table 3). For pollock, the November 1994 feeding measures were only lower than the high summer measures, whereas for herring, feeding measures in autumn of 1994 were lower than in either summer or autumn of 1995. Although pollock stomachs were at least half full in each season, herring stomach fullness declined from 75% full in summer to half full in October 1995, to only trace amounts of food in November 1994. Similarly, prey content %-BW declined less between sea- sons for pollock than for herring. Total prey numbers and weights were generally lower for pollock than for herring in a season, but numbers of prey were highly variable and biomass of prey was relatively stable for pollock, whereas these measures showed declining trends for herring (Fig. 6, Table 3). Digestion data and feeding-frequency data for individ- ual fish were pooled across seasons to compare day and night feeding patterns. The condition of stomach contents 492 Fishery Bulletin 99(3) Table 4 Horn’s overlap index values for total numbers and biomass of prey consumed by YOY walleye pollock and Pacific herring caught separately in summer and together in autumn in Prince William Sound, 1994-1995. No summer sympatric fish were available and autumn allopatric fish were not caught in the same year. Overlap greater than 0.60 indicates similar diets (see text). C = central; NE = northeast. Sampling period and target catch By prey Region Overlap in number By prey By prey species category Overlap in By prey species biomass category Interspecific diet overlap July 1995 allopatric fish 1995 C, NE 0.79 0.82 0.76 0.83 October 1995 sympatric fish 1995-96 NE 0.16 0.22 0.44 0.64 1995-97 NE 0.43 0.48 0.53 0.69 Average NE 0.31 0.43 0.55 0.69 Intraspecific diet overlap November 1994 sympatric fish 1994-96 NE 0.69 0.94 0.08 0.39 1994-97 NE 0.86 0.91 0.88 0.91 Average NE 0.87 0.97 0.88 0.95 Allopatric-sympatric fish October 1995 herring 0.51 0.89 0.56 0.93 November 1994 pollock 0.87 0.91 0.56 0.73 Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragra chalcogramma and Clupea pallasi 493 10 8 20 G In prey numbers I In prey biomass 10 10 20 pollock herring herring pollock pollock herring pollock herring July 1995 Oct 95 Nov 94 Oct 1995 Nov 1994 allopatric allopatric sympatric sympatric Figure 6 Total number and total biomass of prey (ln-transformed means) consumed by allopatric and sympatric YOY walleye pollock and YOY Pacific herring from Prince William Sound, Alaska, in July-August 1995, October 1995, and November 1994. Number of individuals is indicated. by time of day was different for the two species. Pollock had greater percentages of mostly digested contents dur- ing the day than during the night (before and after 22:00). Conversely, herring had greater percentages of mostly di- gested contents during the night than during the day. The day-night percentages of feeding and nonfeeding fish were similar for both, however. Comparisons between allopatric and sympatric fish aggregations The allopatric-sympatric size pattern was different for pollock and herring. We did not pool sympatric fish from October 1995 and November 1994 for comparison with allopatric groups because of the interannual differences in size and feeding measures. For pollock in November 1994, FLs of allopatric and sympatric fish were similar, but fish were approximately 1.5 g lighter in sympatric aggrega- tions than fish in allopatric aggregations. For herring in October 1995, the sizes of allopatric and sympatric fish were similar (Table 3). The allopatric-sympatric feeding pattern was somewhat different for pollock than for herring in autumn. Among November 1994 pollock, the allopatric fish consistently had the highest feeding measures and the sympatric fish consistently had the lowest feeding measures (Fig. 6, Ta- ble 3). This finding coincided with slightly increased nu- merical percentages of larvaceans, increased gravimetric percentages of euphausiids, and increased frequencies of euphausiids, amphipods, and large calanoids in the diet of allopatric pollock compared with the diet of sympatric pol- lock (Figs. 3 and 4). For October 1995 herring, allopatric fish consumed the greatest prey biomass and %BW, but their prey numbers and fullness index were lower than those of sympatric fish (Fig. 6, Table 3). This finding coin- cided with decreased numerical percentages of larvaceans, increased numerical and gravimetric percentages of large calanoids and euphausiids, and increased frequencies of occurrence of hyperiids and euphausiids in the diet of al- lopatric herring compared with diet of sympatric herring (Figs. 3 and 4). Allopatric herring were more selective of large calanoids, but selection for larvaceans was similar for both groups of herring (Fig. 5). For both species, intra- specific diet overlap between allopatric and sympatric fish was extensive at the principal taxon level. In October 1995 intraspecific overlap in terms of prey biomass was 0.93 for herring, and in November 1994, intraspecific overlap in terms of prey number was 0.91 for pollock. Discussion Zooplankton prey fields The prey fields available to planktivorous YOY walleye pollock and Pacific herring were different in summer and autumn 1995. Although we report similar total zooplank- ton densities for the two seasons, we believe that summer densities of small calanoids were underestimated with the 303-pm net. This conclusion is based on a 1995 compan- ion study that showed that a 243-pm mesh net retained small calanoids better than a 303-pm mesh net.8 The sea- Sturdevant, M. V. In review. (continued on next page ) 494 Fishery Bulletin 99(3) sonal densities presented in our study were not directly comparable because of the different gear used, and our summer mesozooplankton densities were lower than those reported from other summer collections in PWS8 9 (Cooney et al., 1981; Celewycz and Wertheimer, 1996). The food supply available to juvenile fish was likely more abun- dant in summer than in autumn, because other studies in the northeastern Pacific showed a steady decline in zoo- plankton biomass from summer to winter (Peterson and Miller, 1977; Cooney, 1986; Incze et al., 1997; Foy and Paul, 1999). Zooplankton taxonomic compositions also dif- fered seasonally, mainly in the larger percentages of alter- nate prey available in autumn. We found the surface-water feeding environment to be richer in numbers of prey than the deeper water in both seasons. Lower zooplankton density with depth has been reported by other authors in PWS,10 Shelikof Strait (Napp et al., 1996), off the Oregon Coast (Marlowe and Miller, 1975), and at Ocean Station “P” (Petersen and Miller, 1977). Large calanoids were an exception in autumn, how- ever, when their densities and percentages were greater in the deeper tows, typical of subarctic locations (e.g. Mar- lowe and Miller, 1975; Nakatani, 1988; Napp et al., 1996). High abundances of Pseudocalanus spp. and other small calanoids and high biomass of large calanoids are char- acteristic of neritic locations in subarctic Pacific waters, such as those of PWS in summer (Springer et al., 1989; Coyle et al., 1990; Celewycz and Wertheimer, 1996; Incze et al., 1997). The time of collection for zooplankton is im- portant, however, because the location of peak abundance in the water column varies with their diel vertical migra- tion. Calanoid copepods are usually most abundant at the surface at night, migrating deeper during the day (e.g. Sekiguchi, 1975). Even though most of our samples were collected during the day, it is not surprising that abun- dances were greater in shallow tows than in deeper tows because Pseudocalanus was dominant. Of the species im- portant in our study, Pseudocalanus newmani was typi- cally the shallowest in depth distribution, Calanus pacifi- cus was intermediate and Metridia pacifica, the deepest (e.g. Bollens and Frost, 1989; Frost and Bollens, 1992; Bol- lens et al., 1993; Bollens et al., 1992b). However, the verti- cal distribution patterns of these species were influenced by the presence of predators and other factors (e.g. Bol- lens and Frost, 1989; Bollens et al., 1992b; Frost and Bol- 8 ( continued ) Summer zooplankton density and composition esti- mates from 20-m vertical hauls using three net meshes. Alaska Fisheries Research Bull., Auke Bay Laboratory, National Marine Fisheries Service, 11305 Glacier Hwy., Juneau, AK 99801. 9 Sturdevant, M. V., and L. B. Hulbert. 1999. Diet overlap, prey selection, and potential food competition among allopatric and sympatric forage fish species in Prince William Sound, 1996. In Forage fish diet overlap, 1994-1996, p. 72-100. Exxon Valdez Oil Spill Restoration Project Final Report (Restoration Project 97163C), Auke Bay Laboratory, National Marine Fisheries Ser- vice, 11305 Glacier Hwy., Juneau, Alaska 99801. 10 Foy, R. J., and B. L. Norcross. In prep. Nearshore zooplank- ton community ecology in Prince William Sound, Alaska. J. Plankton Res., Institute of Marine Science, Univ. Alaska Fair- banks, P. O. Box 757220, Fairbanks, AK 99775-7220. lens, 1992; Bollens et al., 1993). The seasonal difference in depth distribution of large calanoids could have been due to seasonal vertical migrations (Sekiguchi, 1975; Mackas et al, 1993), differences in sampling time, or a response to the changing light regime in autumn. Feeding habits Summer and autumn diets of YOY allopatric and sympat- ric walleye pollock and Pacific herring in PWS were sim- ilar to those reported from other areas and to those of other pollock and herring caught during the study that did not meet our criteria (Sturdevant, unpubl. data; see “Materials and methods” section). In summer, both spe- cies consumed the abundant calanoid taxa as well as less abundant small prey. Calanoid copepods were likewise the predominant summer prey of YOY pollock in Japa- nese waters (Kamba, 1977; Nakatani, 1988), the region of the Kodiak Island-Alaska Peninsula,11 the Gulf of Alaska and eastern Bering Sea (Grover, 1990, 1991; Brodeur et al., 1997), PWS (Willette et al., 1997; Foy and Norcross, 1999a), and southeastern Alaska.12 The small calanoid Acartia clausi was particularly important in southeastern Alaska diets from August to October,12 but by late summer in other areas, euphausiids accounted for more prey bio- mass and calanoids continued to dominate numerically (Merati and Brodeur, 1996). The autumn prey composi- tion of pollock in our study was also similar to that of YOY pollock in the Gulf of Alaska (Merati and Brodeur, 1996; Brodeur et al., 2000), eastern Kamchatka (Sobo- levskii and Senchenko, 1996), and southeastern Alaska.12 In those studies, increased fish size was correlated with decreased predation on small copepods and increased pre- dation on large copepods, larvaceans, and euphausiids. By winter, large copepods virtually disappeared from diets in some areas (Sobolevskii and Senchenko, 1996); chae- tognaths and epibenthic prey such as mysids, shrimps, caprellid amphipods, and cumaceans were incorporated in the diet as vertical distributions of the fish changed and pelagic prey became scarce12 (Merati and Brodeur, 1996; Sobolevskii and Senchenko, 1996; Brodeur et al., 2000). Seasonal changes in prey have also been correlated with change in YOY pollock distribution and the use of different habitats (Nakatani, 1988). Like pollock, YOY Pacific herring depended on small calanoid prey in PWS and throughout their range. Addi- tional small prey taxa are commonly reported in Pacific herring and other species diets, including invertebrate eggs, barnacle larvae, cladocerans, oikopleurans, and ju- venile amphipods and euphausiids9 (Wailes, 1936; Sher- man and Perkins, 1971; Last, 1989; Arrhenius and Hans- nLivingston, P. A. 1985. Summer food habits of young-of-the- year walleye pollock, Theragra chalcogramma, in the Kodiak area of the Gulf of Alaska during 1985. Unpubl. manuscr., Alaska Fish. Sci. Center, National Marine Fisheries Center, 7600 Sand Point Way NE, Seattle, WA 981 15, 8 p. 12Krieger, K. J. 1985. Food habits and distribution of first- year walleye pollock, Theragra chalcogramma (Pallas), in Auke Bay, Southeastern Alaska. Unpubl. MS thesis, Univ. Alaska Juneau, 11305 Glacier Hwy., Juneau, AK 99801, 57 p. Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragra chalcogrcimma and Clupea pallasi 495 son, 1994; Arrhenius, 1996; Haegele, 1997; Willette et al., 1997). In addition to pelagic prey, epibenthic prey such as harpacticoid copepods and gammarid amphipods were im- portant in estuarine habitats of the Fraser River for Pacific herring and coastal Maine for Atlantic herring (Sherman and Perkins, 1971; Blaxter and Hunter, 1982; Levings, 1983, in Lassuy, 1989). Spatial differences were found for diets of Pacific herring from four widespread bays in PWS (Foy and Norcross, 1999a). We found seasonal differences in prey similar to reports by others: euphausiids replaced calanoids in the diets of older juvenile herring compared with younger juvenile herring (Wailes,1936; Lassuy, 1989; Haegele, 1997), and larvaceans (Foy and Norcross, 1999a, 1999b), mysids, and other malacostracans (Foy and Nor- cross, 1999a; Foy and Paul, 1999) were also more common in autumn. In Auke Bay, Alaska, juvenile herring diets varied with abundance of zooplankton prey taxa in spring and early summer and included large calanoids and eu- phausiids when they were present (Coyle and Paul, 1992). The marine distribution of the alewife, Alosa pseudoharen- gus, another herring, was correlated with the seasonal dis- tribution, availability, and abundance of its euphausiid prey (Stone and Jessop, 1994). Prey selection, feeding time, and depth Segregation by depth is one way to reduce interspecific competition among fish with overlapping distributions (e.g. Jessop, 1990; Arrhenius, 1996). Both pollock and her- ring perform diel vertical migrations (Smith, 1981; Blaxter and Hunter, 1982), the pattern of which can vary season- ally and ontogenetically (e.g., Kamba, 1977; Lassuy, 1989; Olla et al., 1996; Stokesbury et al., 2000). We observed some interspecific differences in prey selection that may relate to diel vertical distributions of predator and prey, prey preferences, ontogenetic changes in prey size, or dif- ferent feeding rhythms for both species. Neither species selected small calanoids in summer, and pollock selected taxa that herring did not (large calanoids, pteropods, and larvaceans). The summer herring were located nearshore and at the surface where densities of their main prey were twice as high as deep in the water column and where light for feeding was most intense. In contrast, summer pol- lock were located in relatively deep water with less light and lower prey densities. Even though small calanoids were the predominant prey of pollock and the predomi- nant zooplankter, they were avoided in relation to their availability. In autumn, both species avoided small cala- noids even more strongly than in summer. In autumn, sympatric pollock in shallow water selected large cala- noids, mainly Metrida pacifica, that were less available than in deeper water, but only the deeper allopatric her- ring selected them. Young Pacific herring in another study selected Calanus pacificus over Metridia lucens , perhaps in relation to fine-scale differences in prey depth distri- butions (Fortier and Leggett, 1983 in Munk et al., 1989; Bollens et al., 1993). Both sympatric herring and allo- patric herring strongly selected the larvaceans that were more evenly distributed in the water column. We primar- ily sampled fish schools that were located acoustically and assumed that the fish would be located where food was available (e.g. Arrhenius and Hansson, 1999). Compari- sons of prey selection between seasons are valid because the calculations are based on prey percentages which did not differ between the 243- and 303-pm mesh sizes used.8 The less-digested condition of herring stomach contents by day compared with night suggests that the summer fish and autumn allopatric fish were actively feeding and that the autumn sympatric fish were not. The condition of pollock prey and the fact that the fish were not located where their selected prey were in either season suggest that pollock were not actively feeding at the times we sam- pled in either season. The change in digested state of their prey with time of day also suggests that YOY pollock may seasonally or ontogenetically switch from feeding princi- pally during the day to feeding at night12 (Merati and Bro- deur, 1996; Brodeur et al., 2000). Past diel studies have reported different patterns of feeding for pollock and various herring species. For exam- ple, peak time of feeding for pollock was at midnight or just before dawn in some studies (Brodeur and Wilson, 1996a; Merati and Brodeur, 1996; Willette et al., 1997), but an- other study showed a change in feeding chronology from sunset in small fish to night in larger fish (Brodeur et al., 2000). Peak time of feeding for Atlantic, Baltic, and Pacific herring occurred in the afternoon or evening, with the lowest feeding rates in early morning (e.g. de Silva, 1973; Blaxter and Hunter, 1982; Mehner and Heerkloss, 1994; Willette et al., 1997; Arrhenius, 1998). Changes in prey composition with time of day and ontogeny have also been noted in some studies (e.g. Nakatani, 1988; Grover, 1991; Munk, 1992; Stone and Jessop, 1994) and not others (Brodeur et al., 2000). If co-occurring fish feed at different times, their diets could be highly similar without direct competition because predation on the same prey resourc- es would be temporally or spatially separated. This could be the case with small calanoid prey, which both species fed on in summer, but only herring were co-located at the depth of prey concentration. Different feeding periodicities could result in indirect competition if prey resources are limited, however. Sampling time could also affect the appearance in the diet of vertically migrating macrozooplankton, such as juvenile-adult euphausiids and hyperiid amphipods, be- cause the vertical locations of peak abundance of preda- tor and prey may not overlap continually. For example, euphausiids could be consumed at night near the surface or during the day near the bottom12 (Pearcy et al., 1979; Nakatani, 1988). Juvenile euphausiids were rare summer prey in our study and were present only in those trawls taken below the 60-m mean depth of pollock catches.2 In autumn, the euphausiid and fish distributions were more likely to overlap during the night sampling time (e.g. Bai- ley, 1989; Bollens et al., 1992a), and euphausiids were in- deed a principal prey in terms of biomass, particularly for pollock. Large calanoids and euphausiids could have been consumed at different feedings, particularly if their verti- cal distributions overlapped with those of the fish at dif- ferent times. Predation by pollock and herring on euphau- siids in areas where these macrozooplankters were not 496 Fishery Bulletin 99(3) collected suggests that the fish fed in a different area or at a different time. Comparison of allopatric and sympatric aggregations If competition occurs between sympatric species, one would expect that, given similar prey fields, the quantity or qual- ity of prey consumed would improve when fish are allo- patric. If such changes are prolonged, growth or survival differences are probable. We found intraspecific changes in diet composition and food quantity for both species from allopatric to sympatric aggregations, and weak interspe- cific differences in feeding between sympatric fish. Both species of fish avoided small calanoids whether a competi- tor was present or not, but the proportional use of ener- getically advantageous taxa (large calanoid, euphausiid, and hyperiid prey) and the selection for large calanoids were lower in the presence of competitors. Comparisons of sympatric fish showed that pollock ate proportionally more high energy-producing prey (euphausiid biomass, hyperiid frequency, large calanoid numbers) and selected Metridia spp. more strongly than herring. Overall, how- ever, high diet overlap between allopatric and sympatric fish of either species in autumn indicated little change in diet composition owing to sympatry. Because both spe- cies of sympatric fish ate less total prey than allopatric fish, quantity, rather than quality of prey decreased in the presence of competitors or perhaps as a density- dependent response. Drastic diet shifts were demonstrated in one study involving fish removal; planktivorous spe- cies became benthivorous when their competitors were removed from a lake (Persson and Hansson, 1999). In our study, however, although the intraspecific comparisons (allopatric-sympatric) were within month, they differed by depth, time, or region; the interspecific comparisons of allopatric fish were a month apart in different years. Many authors have found differences in zooplankton prey fields on these scales10 (e.g. Springer et al., 1989; Celewycz and Wertheimer, 1996), making it difficult in our study to separate the effects of sampling differences from those due to competition on diet and feeding. Seasonal feeding We observed the highest frequency of empty stomachs for both species in the November 1994 samples; empty stom- achs, however, were more frequent for herring than for pollock. Similarly, in other seasonal studies, the propor- tion of empty stomachs in YOY herring peaked in winter (Foy and Norcross, 1999a; Foy and Paul, 1999), but empty stomachs were never observed among YOY pollock, even though stomachs were least full in December.12 As zoo- plankton biomass declined between the winter months of October and February, the feeding response and whole body energy content of herring also declined (Foy and Paul, 1999). The herring relied on stored energy to overwinter (Paul et ah, 1998a; Paul and Paul, 1998), but pollock are thought to allocate energy from year-round feeding for somatic growth (Paul et al., 1998b). We also observed declines in zooplankton coincident with feeding declines in autumn; conversely, a greater food sup- ply of small calanoids supported trends toward more in- tensive feeding in summer. Feeding on small calanoids in summer could have been density dependent because this taxon was found to be most prominent in the diets when it was more abundant in zooplankton samples. Juvenile Bal- tic herring exhibited a similar functional response to great- er food densities (Arrhenius and Hansson, 1999). However, decreased feeding on small calanoids is more likely related to fish size and energy requirements. Both species (espe- cially pollock in November 1994) were larger in autumn when they ceased consuming small calanoids, and both species avoided them more often. Larger prey are more effi- cient sources of energy for larger fish, if available. The larg- er autumn fish were probably better able to prey on late stages of macrozooplankton than the summer fish (Merati and Brodeur, 1996; Haegele, 1997; Kamba, 1977), in syn- chronism with a seasonal difference in their abundance and availability13 (Bollens et al., 1992a; Stone and Jessop, 1994; Incze et al., 1997). A shift to larger prey is consis- tent with a decrease in prey numbers without a change in prey weight because fewer large prey are needed for equiv- alent weight. The biomass percentages of large prey con- sumed (large calanoids and euphausiids) and the numeric percentages of small prey consumed (larvaceans) both in- creased in autumn compared with summer (see also Foy and Norcross, 1999a). Ontogenetic changes and size-selec- tive feeding have been reported by others, as well, for juve- nile herring (e.g. Raid, 1985; Munk, 1992; Arrhenius, 1996) and juvenile pollock (Grover, 1991; Brodeur, 1998). In conjunction with the seasonal diet transition, total prey number and most other measures of quantity were lower in autumn than in summer, particularly for her- ring. Autumn declines in juvenile pollock feeding rates were also observed in a study in Southeast Alaska in con- junction with a switch from small prey to larger prey.12 Similarly, feeding declined seasonally for Atlantic herring (de Silva, 1973), Baltic herring (Arrhenius and Hansson, 1999), and Pacific herring (Foy and Norcross, 1999a). In our study, pollock consumed well above maintenance ra- tion (0.30 %BW at 7.5°C; Smith et al., 1986) in all time periods, but, the low prey %BW of herring in autumn 1994 could indicate starvation, given a ration of 1.3-3. 6 %BW at 6.2-8.7°C (de Silva and Balbontin, 1974; Arrhenius and Hansson, 1994). Continued feeding and better fish condi- tion late in the year are advantageous for survival through the extreme conditions of winter, and relate to the species different strategies for overwintering. Declines in feeding with season were also indicated by increased diet overlap between sympatric species in No- vember compared with overlap in October, which suggests a density-dependent convergence of feeding with declining prey resources in late autumn. For sympatric fish in au- 13 Mooney, J. R. 1999. Distribution, energetics, and parasites of euphausiids in Prince William Sound, Alaska. Unpubl. MS thesis, Juneau Center School of Fisheries and Ocean Sci- ence, Univ. Alaska Fairbanks, 11122 Glacier Hwy., Juneau, AK 99801, 172 p. Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragro chalcogramma and Clupea pallasi 497 tumn, lower diet overlap occurred when pollock were more selective for large calanoids and ate proportionally more euphausiids, and the herring selected larvaceans (October 1995); greater dietary overlap in autumn occurred when the species ate similar percentages of these taxa and small calanoids (November 1994). Despite larger size in autumn, energy requirements of fish are lower during this period of less abundant food due to environmental changes. In PWS, surface temper- atures declined from approximately 12°C in summer, to 10°C in October 1995, and 7-8°C in November 1994. 2 Tem- peratures below the thermocline were stable at 5-6°C, but thermocline depth varied seasonally, between approx- imately 40-60 m in summer and 50-90 m in either au- tumn.34 Additional seasonal changes in the environment have been reported by Stokesbury et al. ( 1999). From sum- mer to winter in PWS, surface waters inside bays changed from being colder to being warmer than surface waters outside bays. Such environmental temperature changes in both the vertical and horizontal planes may influence the distribution of pollock and herring, their energy budgets (Smith and Paul, 1986), and the seasonal distribution and abundance of their prey. Fish can conserve energy during times of reduced prey by altering behaviors to maximize prey searching or to de- crease metabolic costs. For example, fish can alter school cohesiveness (Brodeur and Wilson, 1996b; Stokesbury et al., 1999), restrict their movement, or shift in residence to deeper water with colder temperatures (Sogard and Olla, 1996; Ciannelli et al., 1998). However, they may do so at the risk of increased predation or decreased light for feed- ing. Herring are primarily visual feeders, requiring mini- mum light levels to feed (Blaxter and Hunter, 1982) but may also filter feed in low light (Batty et al., 1986; Stone and Jessop, 1994). They can respond to prey distributions correlated with thermocline depth (Fossum and Johannes- sen, 1979) and may shift feeding during the day to depths where light is not restrictive even if prey are concentrat- ed elsewhere (Munk et al., 1989). Similarly, the vertical distribution of juvenile pollock is affected by the relative availability of food and by numerous other factors that af- fect feeding conditions: predator presence, light, turbidity, pressure, temperature, size, and metabolic requirements (Bailey, 1989; Olla et al., 1996; Sogard and Olla, 1996; Ci- annelli et al., 1998). Juvenile pollock avoided light more and avoided cold water less with growth, especially under conditions of low zooplankton (Olla et al., 1996). Both ju- venile pollock (Smith et al., 1986) and YOY herring (de Silva and Balbontin, 1974; Arrhenius and Hansson, 1994) decreased prey consumption at colder temperatures; pol- lock also had lower maintenance rations and grew more rapidly under conditions of low food with colder temper- atures (Smith et al., 1986). Competition with pollock for food could compromise the winter survival of YOY herring by limiting the accumulation of their energy stores. Research showing major differences in the nutritional quality of forage species consumed by piscivores has sparked interest in their trophic interactions. Pollock lipid content was low compared with that of herring, but un- like herring, was not correlated with size (Anthony and Roby, 1997; Payne et al., 1999; Anthony et al., 2000). How- ever, only a few studies have examined the energetic con- sequences of feeding on different prey for juvenile pollock (e.g. Smith et al., 1986; Davis and Olla, 1992; Ciannelli et al., 1998) and herring (e.g. de Silva and Balbontin, 1974; Arrhenius and Hansson, 1999). Because prey differ in nutritional composition, caloric density (Ikeda, 1972; Lee, 1974; Deibel et al., 1992; Davis et al., 1998), size, mo- bility, and behavior, their relative abundance is not the only factor of importance. For example, larvaceans are a highly visible taxon (Bailey et al., 1975) with caloric val- ue per unit weight similar to that of crustacean zooplank- ton even though they are gelatinous (Davis et al., 1998), but many more larvaceans must be consumed to accumu- late the equivalent calories obtained from the crustaceans. Davis and Olla (1992) showed in a controlled experiment that larval pollock growth, behavior, and lipid concentra- tion were affected by the nutritional quality of prey. In a field study, herring diets had the highest energy density of all in May, when large calanoids were the most im- portant taxon (Foy and Norcross, 1999a); however, they were not examined from late fall — a period for which our study showed that euphausiids were the prominent prey and others have shown they contain higher energy den- sity compared with earlier times of the year.13 If sympatry induces feeding on prey of lesser nutritional quality for extended periods because of interference competition, fish growth and survival could be affected. For sympatry to occur, the distribution of juvenile wall- eye pollock and Pacific herring must overlap in three di- mensions: time (seasonal and diel), and both horizontal and vertical space. These species have different life his- tories (Smith, 1981; Lassuy, 1989) and patterns of move- ment change ontogenetically, suggesting that spatial over- lap is likely to vary. YOY herring generally school near the bottom along shore during the day, then move up to the surface at dusk and disperse (Blaxter and Hunter, 1982; Lassuy, 1989; Haegele, 1997). Early YOY pollock stay prin- cipally in surface water above the thermocline, perform a diel vertical migration (DVM), and disperse or move in- shore at night; depth distribution increases from summer to autumn12 and with ontogeny (e.g. Nakatani, 1988; Karp and Walters, 1994; Olla et al., 1996). Stokesbury et al. (1999) found that herring and pollock were generally depth stratified but that both were aggregated in bays in July and October. Rather than having a strong species associa- tion, they may simply have an affinity for the same habitats at some points in their life histories (Brodeur and Wilson, 1996a). Data collected monthly in PWS in 1994 (Willette et al., unpubl. data) showed that, after May, >50% of juvenile herring sets also caught juvenile pollock, and after July, >50% of juvenile pollock sets also caught juvenile herring. Willette et al. (1997) noted that diet overlap was more than twice as great for pollock and herring from sympatric sites than for these species from allopatric sites in late summer. These patterns suggest that sympatry and feeding competi- tion increase from spring through summer. Our study is the first to examine the feeding interac- tions of juvenile walleye pollock and Pacific herring. Its limited scope, because it was designed for other primary 498 Fishery Bulletin 99(3) purposes, necessitates a primarily descriptive approach. Differences that we observed in fish feeding could relate to variable sampling intensity, limited fish sample sizes, gear limitations for certain prey taxa, spatial variation in prey, or differences in fish depths. Nonetheless, the similarity of dietary requirements between YOY pollock and herring could induce competition for limited food when these fish co-occur during periods or in places of low food availabil- ity, particularly late autumn and winter. Stokesbury et al. (1999) showed spatial variation in growth of YOY her- ring among widespread bays in PWS and surmised that density-dependent interspecific competition for food was one variable that affected growth rates; likewise, Paul and Paul (1999) showed interannual and spatial variations in size and energy content of YOY pollock from PWS and speculated that prey production and delivery contributed to the differences. The densities of planktivorous predators may exceed the carrying capacity of bays where walleye pollock and Pacific herring are sympatric, adding inter- specific competition to the factors limiting growth. More studies comparing spatial and temporal patterns of distri- bution, abundance, and feeding are needed to clarify the extent and frequency of interactions between YOY pollock, herring, and other forage species and the impact of chang- es in prey and climate on these interactions. Acknowledgments We thank Lewis Haldorson, Captain Wade Loofborough, and the crew of the Alaska Department of Fish and Game’s RV Medeia, and Captain Brian Beaver and the crew of the chartered trawl vessel FV Caravelle for their dedica- tion in collecting samples and oceanographic data. We also appreciate the assistance of the biologists and technicians from the University of Alaska, Institute of Marine Science and Juneau Center for Fisheries and Ocean Science, the National Marine Fisheries Service Auke Bay Laboratory, and the U.S. Fish and Wildlife Service. We especially thank J. Boldt for fieldwork organization, and M. Auburn-Cook, R. Bailey, and S. Keegan, who analyzed most of the gut contents in the laboratory. We also appreciate the financial support of the Exxon Valdez Oil Spill Trustee Council. Literature cited Anderson, P. J., and J. F. Piatt. 1999. Community reorganization in the Gulf of Alaska fol- lowing ocean climate regime shift. Mar. Ecol. Prog. Ser. 189:117-123. Anthony, J. A., and D. D. Roby. 1997. Variation in lipid content of forage fishes and its effect on energy provisioning rates to seabird nestlings. In Forage fishes in marine ecosystems: proceedings of the international symposium on the role of forage fishes in marine ecosystems; November 13-16, 1996, Anchorage, Alaska, p. 725-729. Univ. Alaska Sea Grant College Pro- gram Report 97-01, Anchorage, AK. Anthony, J. A., D. D. Roby and K. R. Turco. 2000. Lipid content and energy density of forage fishes from the northern Gulf of Alaska. J. Exp. Mar. Biol. Ecol. 243:53-78. Arrhenius, F. 1996. Diet composition and food selectivity of 0-group her- ring ( Clupea harengus L.) and sprat ( Sprattus sprattus L.) in the northern Baltic Sea. ICES J. Mar. Sci. 53:701-712. 1998. Variable length of daily feeding period in bioenerget- ics modeling: a test with 0-group Baltic herring. J. Fish Biol. 52:855-860. Arrhenius, F., and S. Hansson. 1994. In situ food consumption by young-of-the-year Baltic Sea herring Clupea harengus : a test of predictions from a bioenergetics model. Mar. Ecol. Prog. Ser. 110:145-149. 1999. Growth of Baltic Sea young-of-the-hear herring Clu- pea harengus is resource limited. Mar. Ecol. Prog. Ser. 191: 295-299. Bailey, K. M. 1989. Interaction between the vertical distribution of juve- nile walleye pollock Theragra chalcogramma in the east- ern Bering Sea and cannibalism. Mar. Ecol. Prog. Ser. 53: 205-213. Bailey, K. M., S. A. Macklin, R. K. Reed, R. D. Brodeur, W. J. Ingraham, J. F. Piatt, M. Shima, R. C. Francis, P. J. Anderson, T. C. Royer, A. B. Hollowed, D. A. Somerton, and W. S. Wooster. 1995. ENSO events in the northern Gulf of Alaska, and effects on selected marine fisheries. CalCOFI 36:78-96. Bailey, J. E., B. L. Wing, and C. R. Mattson. 1975. Zooplankton abundance and feeding habits of fry of pink salmon, Oncorhynchus gorbuscha, and chum salmon, Oncorhynchus keta, in Traitors Cove, Alaska, with specu- lations on the carrying capacity of the area. Fish. Bull. 73(4):846-861. Batty, R. S., H. S. Blaxter, and D.A. Libby. 1986. Herring (Clupea harengus ) filter-feeding in the dark. Mar. Biol. 91:371-375. Bechtol, W. R. 1997. Changes in forage fish populations in Kachemak Bay, Alaska, 1976-1995. In Forage fishes in marine ecosystems: proceedings of the international symposium on the role of forage fishes in marine ecosystems, p. 441^54. Alaska Sea Grant College Program Report No. 97-01, Fairbanks, AK. Blaxter, J. H. S., and J. R. Hunter. 1982. The biology of Clupeoid fishes. In Advances in marine biology, vol. 20 (J. H. S. Blaxter, Sir F. S. Russell, and Sir M. Yonge, eds.), p. 1-223. Academic Press, London. Bollens, S. M., and B. W. Frost. 1989. Zooplanktivorous fish and variable diel vertical migra- tion in the marine planktonic copepod Calanus pacificus. Limnol. Oceanogr. 34(6):1072-1083. Bollens, S. M„ B. W. Frost, and T. S. Lin. 1992a. Recruitment, growth, and diel vertical migration of Euphausia pacifica in a temperate fiord. Mar. Biol. 114:219-228. Bollens, S. M., B. W. Frost, D. S. Thoreson, and S. J. Watts. 1992b. Diel vertical migration in zooplankton: field evidence in support of the predator avoidance hypothesis. Hydro- biologia 234:33-39. Bollens, S. M., K. Osgood, B. W. Frost, and S. D. Watts. 1993. Vertical distributions and susceptibilities to verte- brate predation of the marine copepods Metridia lucens and Calanus pacificus. Limnol. Oceanog. 38(8):1827-1837. Brodeur, R. D. 1998. Prey selection by age-0 Walleye pollock ( Theragra chalcogramma) in nearshore waters of the Gulf of Alaska. Environ. Biol. Fishes 51:175-186. Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragra chalcogrammci and C/upea pallasi 499 Brodeur, R., and M. Wilson. 1996a. A review of the distribution, ecology and population dynamics of age-0 walleye pollock in the Gulf of Alaska. Fish. Oceanogr. 5 (suppl. 1): 148-166. 1996b. Mesoscale acoustic patterns of juvenile walleye pol- lock ( Theragra chalcogramma ) in the western Gulf of Alaska. Can. J. Fish. Aquat. Sci. 53:1951-1963. Brodeur, R. D., M. T. Wilson, and L. Ciannelli. 2000. Spatial and temporal variability in feeding and con- dition of age-0 walleye pollock (Theragra chalcogramma ) in frontal regions of the Bering Sea. ICES J. Mar. Sci. 57:256-264. Brodeur, R. D., M. T. Wilson, J. M. Napp, P. J. Stabeno, and S. Salo. 1997. Distribution of juvenile pollock relative to frontal structure near the Pribilof Islands, Bering Sea. In Forage fishes in marine ecosystems: proceedings of the interna- tional symposium on the role of forage fishes in marine eco- systems; November 13-16, 1996, Anchorage, Alaska, p. 573- 589. Univ. Alaska Sea Grant College Program Report 97-01, Anchorage, AK. Brown, E. D., T. T. Baker, J. E. Hose, R. M. Kocan, G. D. Marty, M. D. McGurk, B. L. Norcross, and J. Short. 1996a. Injury to the early life history stages of Pacific her- ring in Prince William Sound after the Exxon Valdez Oil Spill. Am. Fish. Soc. Symp. 18:448-462 . Brown, E. D., B. L. Norcross, and J. W. Short. 1996b. An introduction to studies on the effects of the Exxon Valdez oil spill on early life history stages of Pacific herring, Clupea pallasi, in Prince William Sound, Alaska. Can. J. Fish. Aquat. Sci. 53:2337-2342. Celewycz, A. G., and A. C. Wertheimer. 1996. Prey availability to juvenile salmon after the Exxon Valdez Oil Spill. Am. Fish. Soc. Symp. 18:564—577. Ciannelli, L., R. D. Brodeur, and T. W. Buckley. 1998. Development and application of a bioenergetics model for juvenile walleye pollock. J. Fish Biol. 52:879-898. Cooney, R. T. 1986. Zooplankton. In The Gulf of Alaska: physical envi- ronment and biological resources (D. W. Hood and S. T. Zimmerman, eds.), p. 285-303. U.S. Gov. Printing Office, Wash. D.C., 654 p. Cooney, R. T., D. Urquhart, and D. Barnard. 1981. The behavior, feeding biology and growth of hatch- ery-released pink and chum salmon fry in Prince William Sound, Alaska. Alaska Sea Grant Rep. 81-5, 114 p. Coyle, K. O., and A. J. Paul. 1992. Interannual differences in prey taken by capelin, her- ring, and red salmon relative to zooplankton abundance during the spring bloom in a southeast Alaskan embay- ment. Fish. Oceanog. l(4):294-305. Coyle, K. O., A. J. Paul, and D. A. Ziemann. 1990. Copepod populations during the spring bloom in an Alaskan subarctic embayment. J. Plank. Res. 12(41:759— 797. Davis, N. D., K. W. Myers, and Y. Ishida. 1998. Caloric value of high-seas salmon prey organisms and simulated salmon ocean growth and prey consumption. North Pacific Anadr. Fish Comm. Bull. 1:146-162. Davis, M. W., and B. L. Olla. 1992. Comparison of growth, behavior and lipid con- centrations of walleye pollock Theragra chalcogramma larvae fed lipid-enriched, lipid-deficient and field-collected prey. Mar. Ecol. Prog. Ser. 90:23-30. de Silva, S. S. 1973. Food and feeding habits of the herring Clupea haren- gus and the sprat C. sprattus in inshore waters of the west coast of Scotland. Mar. Biol. 20:282-290. de Silva, S. S., and F. Balbontin. 1974. Laboratory studies on food intake, growth and food conversion of young herring, Clupea harengus ( L. ). J. Fish Biol. 6:645-658. Deibel, D., J. F. Cavaletto, M. R. Riehl, and W. S. Gardner. 1992. Lipid and lipid class content of the pelagic tunicate Oikopleura vanhoeffeni. Mar. Ecol. Prog. Ser. 88:297-302. Fossum, P, and A. Johannessen. 1979. Field and laboratory studies of herring larvae (Clupea harengus L.): proceedings of the International Council for the Exploration of the Sea (ICES) meeting at Charlotten- lund, Denmark, October, 1, 1979, 31 p. Foy, R. J., and B. L. Norcross. 1999a. Spatial and temporal variability in the diet of juve- nile Pacific herring ( Clupea pallasi) in Prince William Sound, Alaska. Can. J. Zool. 77:1-10. 1999b. Feeding behavior of herring (Clupea pallasi ) associ- ated with zooplankton availability in Prince William Sound, Alaska. In Ecosystem approaches for fisheries manage- ment, p. 1229-135. Alaska Sea Grant College Program AK-SG-99-01, Fairbanks, AK, 756 p. Foy, R. J., and A. J. Paul. 1999. Winter feeding and changes in somatic energy content for age-0 Pacific herring (Clupea pallasi) in Prince William Sound, Alaska. Trans. Am. Fish. Soc. 128:1193-2000. Frost, B. W., and S. M. Bollens. 1992. Variability of diel vertical migration in the marine planktonic copepod Pseudocalanus newmam in relation to its predators. Can. J. Fish. Aquat. Sci. 49:1137-1141. Grover, J. J. 1990. Feeding ecology of late-larval and early -juvenile wall- eye pollock Theragra chalcogramma from the Gulf of Alaska in 1987. Fish. Bull. 88:463-470. 1991. Trophic relationship of age-0 and age-1 walleye pol- lock Theragra chalcogramma collected together in the east- ern Bering Sea. Fish. Bull. 89:719-722. Haegele, C. W. 1997. The occurrence, abundance and food of juvenile her- ring and salmon in the Strait of Georgia, British Columbia in 1990 to 1994. Canadian Manuscr. Rep., Fish. Aquat. Sci. 2390, 124 p. Horn, H. S. 1966. Measurement of “overlap” in comparative ecological studies. Am. Naturalist. 100(914):419-424. Ikeda, T. 1972. Chemical composition and nutrition of zooplankton in the Bering Sea. In Biological oceanography of the north- ern North Pacific Ocean (A. Y. Takenouti, ed.), p. 433-442. Idemitsu Shoten, Tokyo, Japan. Incze, L. S., D. W. Siefert, and J. M. Napp. 1997. Mesozooplankton of Shelikof Strait, Alaska: abun- dance and community composition. Continental Shelf Res. 17(31:287-305. Iverson, S. J., K. J. Frost, and L. F. Lowry. 1997. Fatty acid signatures reveal fine scale structure of foraging distribution of harbor seals and their prey in Prince William Sound, Alaska. Mar. Ecol. Prog. Ser. 151( 1— 3):255— 27 1. Jessop, B. M. 1990. Diel variation in density, length composition, and feeding activity of juvenile alewife, Alosa pseudoharengus Wilson, and blueback herring, A. aestivalis Mitchill, at near-surface depths in a hydroelectric dam impoundment. J. Fish. Biol. 37:813-822. 500 Fishery Bulletin 99(3) Kamba, M. 1977. Feeding habits and vertical distribution of walleye pollock Theragra chalcogramma (Pallas), in early life stage in Uchiura Bay, Hokkaido. Res. Inst. North Pac. Fish., Hokkaido University, spec, vol., p 175-197. Karp, W. A., and G. E. Walters. 1994. Survey assessment of semi-pelagic Gadoids: the exam- ples of walleye pollock, Theragra chalcogramma, in the Eastern Bering Sea. Mar. Fish. Rev. 56( 1 ):8— 22. Krebs, C. J. 1989. Ecological methodology. Harper Collins Pubis., Univ. British Columbia, British Columbia, 654 p. Kuletz, K. J. 1996. Marbled murrelet abundance and breeding activity at Naked Island, Prince William Sound, and Kachemak Bay, Alaska, before and after the Exxon Valdez oil spill. Am. Fish. Soc. Symp. 18:770-784. Kuletz, K. J., D. B. Irons, B. A. Agler, J. F. Piatt, and D. C. Duffy. 1997. Long-term changes in diets and populations of pisciv- orous birds and mammals in Prince William Sound, Alaska. In Forage fishes in marine ecosystems: proceedings of the international symposium on the role of forage fishes in marine ecosystems, p. 703-706. Univ. Alaska Sea Grant College Program Report 97-01, Fairbanks, AK, 816 p. Lassuy, D. R. 1989. Species profiles: Life histories and environmental requirements of coastal fishes and invertebrates (Pacific Northwest) — Pacific Herring. U.S. Fish Wildl. Serv. Biol. Rep 82( 11.126). U.S. Army Corps of Engineers, TR-EL-82-4, 18 p. Last, J. M. 1989. The food of herring, Clupea harengus, in the North Sea, 1983-1986. J. Fish Biol. 34:489-501. Lee, R. F. 1974. Lipids of zooplankton from Bute Inlet, British Colum- bia. J. Fish Res. Board Can. 31:1577-1582. Livingston, P. A. 1993. Importance of predation by groundfish, marine mam- mals and birds on walleye pollock Theragra chalcogramma and Pacific herring Clupea pallasi in the eastern Bering Sea. Mar. Ecol. Prog. Ser. 102:205-215. Mackas, D. L., H. Sefton, C. B. Miller, and A. Raich. 1993. Vertical habitat partitioning by large calanoid cope- pods in the oceanic subarctic pacific during spring. Prog. Oceanog. 32:259-294. Marlowe, C. J., and C. B. Miller. 1975. Patterns of vertical distribution and migration of zooplankton at Ocean Station “P”. Limnol. Oceanog. 20)3):824-844. Mehner, T., and R. Heerkloss. 1994. Direct estimation of food consumption of juvenile fish in a shallow inlet of the southern Baltic. Int. Revue ges. Hydrobiol. 79(2):295-304. Merati, N., and R. D. Brodeur. 1996. Feeding habits and daily ration of juvenile walleye pollock Theragra chalcogramma, in the Western Gulf of Alaska. In Ecology of juvenile walleye pollock Theragra chalcogramma (R. D. Brodeur, P. A. Livingston, T. R. Lough- lin, and A. B. Hollowed, eds.). p. 65-79. U.S. Dep. Comm. Tech. Rep. NMFS 126. Munk, P. 1992. Foraging behaviour and prey size spectra of larval her- ring Clupea harengus. Mar. Ecol. Prog. Ser. 80:149-158. Munk, P, T. Kiorboe, and V. Christensen. 1989. Vertical migrations of herring Clupea harengus larvae in relation to light and prey distribution. Environ. Biol. Fish. 26:87-96. Nakatani, T. 1988. Studies on the early life history of walleye pollock Theragra chalcogramma in Funka Bay and vicinity, Hok- kaido. Mem. Fac. Fish. Hokkaido U. 35(l):l-46. Napp, J. M., L. S. Incze, P. B. Ortner, D. L. W. Siefert, and L. Britt. 1996. The plankton of Shelikof Strait, Alaska: standing stock, production, mesoscale variability and their relevance to larval fish survival. Fish. Oceanog. 5(suppl. l):19-38. Oakley, K. L., and K. J. Kuletz. 1996. Population, reproduction and foraging of pigeon guil- lemots at Naked Island, Alaska, before and after the Exxon Valdez Oil Spill. Am. Fish. Soc. Symp. 18:759-769. Olla, B. L., M. W. Davis, C. H. Ryer, and S. M. Sogard. 1996. Behavioural determinants of distribution and sur- vival in early stages of walleye pollock Theragra chal- cogramma\ a synthesis of experimental studies. Fish. Oceanogr. 5(suppl. 11:167-178. Paul, A. J., and J. M. Paul. 1998. Comparisons of whole body energy content of captive fasting age zero Alaskan Pacific herring (Clupea pallasi Valanciennes) and cohorts over-wintering in nature. J. Exp. Mar. Biol. Ecol. 226(11:75-86. 1999. Interannual and regional variations in body length, weight and energy content of age-0 Pacific herring from Prince William sound, Alaska. J. Fish Biol. 54:996- 1001. Paul, A. J., J. M. Paul, and E. D. Brown. 1998a. Fall and spring somatic energy content for Alaskan Pacific herring (Clupea pallasi Valenciennes 1847) relative to age, size and sex. J. Exp. Mar. Biol. Ecol. 223:133-142. Paul, A. J., J. M. Paul, and R. L. Smith. 1998b. Seasonal changes in whole-body energy content and estimated consumption rates of age-0 walleye pollock from Prince William Sound, Alaska. Estuar. Coast. Shelf Sci. 47(31:251-259. Payne, S. A., B. A. Johnson, and R. S. Otto. 1999. Proximate composition of some north-eastern Pacific forage fish species. Fish. Oceanogr. 8(31:159-177. Pearcy, W. G., C. C. E. Hopkins, S. Gronvik, and R. A. Evans. 1979. Feeding habits of cod, capelin, and herring in Bals- fjorden, northern Norway, July-August 1978: The impor- tance of euphausiids. Sarsia 64:269-277. Persson, A., and L.-A. Hansson. 1999. Diet shift in fish following competitive release. Can. J. Fish. Aquat. Sci. 56:70-78. Petersen, W. T., and C. B. Miller. 1977. Seasonal cycle of zooplankton abundance and species composition along the central Oregon coast. Fish. Bull. 75(41:717-724. Piatt, J. F., and P. Anderson. 1996. Response of common murres to the Exxon Valdez Oil Spill and long term changes in the Gulf of Alaska marine ecosystem. Am. Fish. Soc. Symp. 18:720-737. Raid, T. 1985. The reproduction areas and ecology of Baltic herring in the early stages of development found in the Soviet Zone of the Gulf of Finland. Finnish Fish. Res. 6:20-34. Sekiguchi, H. 1975. Seasonal and ontogenetic vertical migrations in some common copepods in the northern region of the north Pacific. Bull. Fac. Fish., Mie Univ. 2:29-38. Sherman, K., and H. C. Perkins. 1971. Seasonal variations in the food of juvenile herring in coastal waters of Maine. Trans. Am. Fish. Soc. 1:121- 124. Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragra cho/cogramma and C/upea pallasi 501 Smith, G. B. 1981. The biology of walleye pollock. In The eastern Bering Sea Shelf: oceanography and resources, vol. 1 (D. W. Hood and J. A. Calder, eds.), p. 527-551. U.S. Government Print- ing Office, Wash., D. C. Smith, R. L., and A. J. Paul. 1986. A theoretical energy budget for juvenile walleye pol- lock in Alaskan waters. INPFC Bull. 47:79-86. Smith, R. L., A. J. Paul, and J. M. Paul. 1986. Effect of food intake and temperature on growth and conversion efficiency of juvenile walleye pollock ( Theragra chalcogramma (Pallas)): a laboratory study. J. Cons. Int. Explor. Mer 42:241-253. Smith, E. P., and T. M. Zaret. 1982. Bias in estimating niche overlap. Ecology 63( 5 ): 1248— 1253. Sobolevskii, E. I., and I. A. Senchenko. 1996. The spatial structure and trophic connections of abun- dant pelagic fish of eastern Kamchatka in the autumn and winter. J. Ichthyol. 36( 1 ):30— 39. Sogard, S. M., and B. L. Olla. 1996. Food deprivation affects vertical distribution and activity of a marine fish in a thermal gradient: potential energy-conserving mechanisms. Mar. Ecol. Prog. Ser. 133: 43-55. Springer, A. M. 1992. A review: walleye pollock in the North Pacific Ocean — how much difference do they really make? Fish. Ocean- ogr. 1:80-96. Springer, A. M., C. P. McRoy, and K. R. Turco. 1989. The paradox of pelagic food webs in the northern Bering Sea-II: zooplankton communities. Continental Shelf Res. 9(41:359-386. Springer, A. M., and S. G. Speckman. 1997. A forage fish is what? Summary of the symposium. In Forage fishes in marine ecosystems: proceedings of the interna- tional symposium on the role of forage fishes in marine eco- systems; November 13-16, 1996, Anchorage, Alaska, p. 773- 805. Univ. of Alaska Sea Grant College FYogram Report 97-0. Stokesbury, K. D. E., R. J. Foy, and B. L. Norcross. 1999. Spatial and temporal variability in juvenile Pacific herring, Clupea pallasi, growth in Prince William Sound, Alaska. Environ. Biol. Fish. 56:409-418. Stokesbury, K. D., J. Kirsch, E. D. Brown, G. L. Thomas, and B. L. Norcross. 2000. Spatial distributions of Pacific herring, Clupea pal- lasi, and walleye pollock, Theragra chalcogramma, in Prince William Sound, Alaska. Fish. Bull. 98:400-409. Stone, H. H., and B. M. Jessop. 1994. Feeding habits of anadromous alewives, Alosa pseu- doharengus, off the Atlantic Coast of Nova Scotia. Fish. Bull. 92:157-170. Strauss, R. E. 1979. Reliability estimates for Ivlev’s Electivity Index, the Forage Ratio and a proposed linear index of food selection. Trans. Am. Fish. Soc. 108:344-352. Wailes, G. H. 1936. Food of Clupea pallasii in southern British Columbia waters. J. Biol. Board Can. l(6):477-486. Willette, M., M. Sturdevant, and S. Jewett. 1997. Prey resource partitioning among several species of forage fishes in Prince William Sound, Alaska. In Forage fishes in marine ecosystems: proceedings of the interna- tional symposium on the role of forage fishes in marine ecosystems, p. 11-29. Univ. Alaska Sea Grant College Program Report 97-01, Fairbanks, AK. 502 Age and growth of the bigeye tuna, Thunnus obesus, in the western Pacific Ocean Abstract-Bigeye tuna (Thunnus obe- sus) age and growth studies were last conducted in the western Pacific in 1967 and no study has ever attempted to age bigeye tuna from this area by using dorsal spines. The objective of our study was to estimate bigeye tuna age and growth rate in the western Pacific based on counts of growth rings on sections of the first dorsal spine. Length and weight data, and the first dorsal spine from bigeye tuna in the Tungkang (southwest of Taiwan) fish market were collected monthly from February 1997 to January 1998. In total, 1149 specimens were collected. The fork lengths of individuals ranged from 45.6 to 189.2 cm. Cross sections from dorsal spines were taken and examined under a dissecting microscope equipped with an image analysis system. The monthly percentage of specimens having a terminal translucent zone indicated that growth rings formed once a year; therefore, the age of each fish was determined from the number of visible growth rings. Von Bertalanffy growth parameters were estimated for males, females, and both sexes combined. There was no significant difference between males and females. The parameter estimates for the combined sexes were asymptotic length (LM) = 208.7 cm, growth coefficient ( K) = 0.201/yr, and age at zero length (t0) = -0.9906 yr. Manuscript accepted 12 December 2000. Fish. Bull. 99:502-509 (2001). Chi-Lu Sun Chien-Lung Huang Su-Zan Yeh Institute of Oceanography National Taiwan University Taipei, Taiwan E-mail address (for C.L. Sun): chilu@ccms.ntu.edu. tw Bigeye tuna (Thunnus obesus Lowe, 1839) are a commercially important species of tuna inhabiting the warm waters of the Atlantic, Indian, and Pacific oceans. They are found across the entire Pacific between northern Japan and North Island of New Zea- land in the west and from 40°N to 30°S in the east (Calkins, 1980; Mat- sumoto, 1998). Adult bigeye tuna are caught mainly by longlines, but sub- stantial numbers of juveniles are taken by purse seines. Taiwanese distant water tuna long- line fleets have operated throughout these three oceans since the late 1960s targeting albacore. In the early 1980s, the Taiwanese began equipping their longliners with very cold (below -55°C) freezers and deep longlines in the Indi- an and Atlantic oceans, which allowed them to target bigeye tuna for the lu- crative sashimi market in Japan. In the western Pacific, the Taiwanese off- shore longline fleets, based in domestic (Tungkang mainly) and foreign fishing ports, have landed more bigeye tuna than in the past. Growth studies of Pacific bigeye tu- na conducted in the 1950s and 1960s were based either on increments be- tween modal points in size-composition data (Iversen, 1955; Shomura and Ke- ala, 1963; Yukinawa and Yabuta, 1963; Kume and Joseph, 1966; Suda and Kume, 1967) or on the number of an- nual markings (annuli) on scales (Nose et al., 1957; Yukinawa and Yabuta, 1963). Recently, Hampton and Leroy1 and Matsumoto (1998) presented pre- liminary results from growth studies based on otolith increment counts. No previous study had aged Pacific and In- dian bigeye tuna from dorsal spines, al- though a few age determination studies existed for Atlantic bigeye tuna (Gaikov et al., 1980; Draganik and Pelczarski, 1984; Delgado de Molina and Santana, 1986; Alves et al., 1998). Accurate age structure of stocks is essential for stock assessment and fishery management. Our study provides estimates of the age and growth rate of bigeye tuna in the western Pacific from growth rings on sections of the first dorsal spine. Materials and methods Fork length (in cm), weight (in kg), and sex were determined for bigeye tuna caught by Taiwanese offshore longlin- ers in the fishing area from 23°N to 0°N and 110°E to 140°E (Fig. I2) and sold at the Tungkang fish market between February 1997 and January 1998. In addition, a total of 1149 first dorsal spines were collected. Three cross sec- tions were taken along the length of each spine above the condyle base (Fig. 2A) with a low-speed “ISOMET” saw 1 Hampton, J., and B. Leroy. 1998. Note on preliminary estimates of bigeye growth from presumed daily increments on oto- liths and tagging data. Working paper 18, eleventh meeting of the standing com- mittee on tuna and billfish, Honolulu, Hawaii, USA, 30 May-6 June 1998, 3 p. Oceanic Fisheries Programme, Secretar- iate of the Pacific Community, B.P.D5, 98848 Noumea, New Caledonia. 2 Yang, R. T., R. F. Chung, and C. L. Chang. 1982. Taiwanese offshore tuna longline fishery. Part I: fishing ground, fishing season, and fishing condition. Spec. Rep. 36, 6 p. [In Chinese with English abstract.] Insitute of Oceanography, National Taiwan University, no. 1, sec 4, Roosevelt Rd, Taipei, 106 Taiwan. Sun et al.: Age and growth of Thunnus obesus 503 (model no. 11-1280) and diamond wafering blades. Sections ranging from 0.8 to 1.0 mm thick (Fig. 2B) were examined with a dis- secting microscope (model: Olympus SZH- ILLD) with transmitted light. Images of the dorsal spine sections were captured by using an image analysis software package, a CCD (charged coupled device) camera, and a high- resolution computer monitor. Translucent rings on the section images were counted by two readers independently. When ring counts disagreed, images were read again by both readers simultaneously, and any ques- tionable spines were discarded. Spine sections as the structure to estimate age have the advantage of requiring easy sampling and easy reading (the growth rings stand out clearly), and samples are easily stored for future reexamination (Compean- Jimenez and Bard, 1983). However, early growth rings may be lost in larger specimens because of increased size of the vascularized core in the spine. Accordingly, we estimated the number of lost (obscured) rings from ob- servations of their position and number in spines from young specimens as has been done for little tunny ( Euthynnus alletteratus ) (Cayre and Di- ouf, 1983), eastern Atlantic bluefin tuna (Thunnus thynnus) (Compean-Jimenez and Bard, 1983), and Pacific blue marlin ( Makaira nigricans) (Hill et al., 1989). Age was determined from the translucent rings, assuming that two rings are formed each year — a translucent (light colored) ring formed during the slower growth period and an opaque (dark colored) ring formed during the fast growth period. This as- sumption was validated by observing a translucent or opaque edge on the dorsal-spine sections and a monthly variation in the number of translucent edg- es (Antoine et al., 1983). Distance between the center of the dorsal spine and the outer edge of each annual ring was mea- sured in microns with the software package after calibration against an optical micrometer. The cen- ter of the spine was estimated by following Cayre and Diouf (1983) (Fig. 2B). Distances (rf.) were then converted into radii (i?t-) by following Gonzalez-Gar- ces and Farina-Perez (1983). The relationship between fork length (FL) and dorsal spine radius (R) was modeled by a linear equation (Zar, 1999). Fork length was then back-cal- culated for each ring with the formula (Lee, 1920) 100 105 110 115 120 125 130 135 140 Figure 1 Fishing areas of the Taiwanese offshore tuna longline fishery in the western Pacific Ocean (Yang et al.2). FL, — a + (FL - a)Rt R where FLt = predicted fork length of the fish corre- sponding to age or ring i in cm; a = ordinate in the origin of the equation FL = a + bR-, RING Figure 2 First dorsal spine and the site of cross section (A) and the cross section showing annual rings and measurements taken (B) for age determination of the western Pacific bigeye tuna (c =width of condyle base; L1DS=length of the first dorsal spine; R=radius of spine; Rj=radius of ring i; d=diameter of spine; dpdiameter of ring i). 504 Fishery Bulletin 99(3) FL = observed fork length of the fish in cm; Rl = radius of the ring calculated as the aver- age value observed in ring i (Fig. 2B); and R = dorsal spine radius. Back-calculated fork lengths were used in Ford-Walford (Gulland, 1983) and nonlinear (Ratkowsky, 1983) meth- ods to fit the von Bertalanffy growth function (VBGF) and to obtain vital parameters by sex. Analysis of the residual sum of squares (ARSS) was employed to com- pare the VBGF between sexes (Ratkowsky, 1983; Chen et al., 1992). Weight was related to fork length by using the power function, and analysis of covariance (ANCOVA) (Steel, 1980; Zar, 1999) was conducted to examine differences between sexes. Results Spines from 1149 specimens ranging in size from 45.6 to 189.2 cm FL were examined (Table 1, Fig. 3). There was 90% agreement between the readers’ counts of growth rings and second readings improved this agreement to 95.6%, which resulted in discarding 51 specimens from analysis. The relationship between FL (cm) and weight (kg) is shown in Figure 4. The ANCOVA indicated no significant difference between males and females (P>0.05); thus the FL-W relationship with sexes combined was expressed as W = 3 x 10-5 FL2 9278 (r2=0.97, n=856). The relationship of first dorsal spine lengths (Lws) and FL was (Fig. 5) Table 1 Sample sizes, ranges of fork lengths (FL, cm), and sampling months and areas of bigeye tuna from the western Pacific Ocean. A, B, C, D, and E denote areas in Figure 1. Sampling Month area Sample size Minimum FL Maximum FL Feb 1997 A 80 70 174.5 Mar 1997 A 104 64 169.5 Apr 1997 B 70 101.3 171 May 1997 B 54 83.8 157.4 Jun 1997 B 71 75.5 162.8 Jul 1997 B, D 131 72.2 165.6 Aug 1997 E 94 78.5 187.7 Sep 1997 E 98 45.6 189.2 Oct 1997 A, C 115 86.5 176.6 Nov 1997 A, C 116 89.6 161.1 Dec 1997 C 123 104.6 162.1 Jan 1998 A 93 88.7 159.1 Total 1149 45.6 189.2 FL = 6.9367 Lws + 6.6667 (/-0.94, n= 567). The trend of the monthly percentages of terminal trans- lucent edges (Fig. 6) suggested that the period from Febru- ary to September was the long period of inhibited growth (translucent edge). From October to November, growth ap- peared to resume (opaque edge) and later, from December to January, a new translucent edge appeared; indicating the formation of one growth ring per year. Given the significant linear relationship between the dorsal spine radius and fork length (FL=26.455R + 19.916, r=0.94, n = 1098), we used spine measurements to back- Sun et al.: Age and growth of Thunnus obesus 505 Combined sexes (n= 856) l/V=0. 00003 FL2 9278 r2= 0.97 0 20 40 60 80 100 120 140 160 180 200 Fork length (cm) Figure 4 Relationships between weight and fork length of the west- ern Pacific bigeye tuna sampled at Tungkang fish market. calculate the fork lengths of previous ages. The mean back-calculated fork lengths for the first 10 years of life for the western Pacific bigeye tuna are given in Table 2. Parameters of the VBGF estimated by the Ford-Wal- ford method for males, females, and sexes combined are shown in Table 3. Growth was not significantly different between sexes (ARSS, F=1.98; df=3, 452; P> 0.05); the pooled growth curve is shown in Figure 7. VBGF param- eters computed by nonlinear regression are also shown in Table 3 and Figure 7. Length-at-age of bigeye tuna esti- mated by nonlinear regression is larger (up to age 6 years) than that estimated by the Ford-Walford method. Discussion Available genetic information supports the hypothesis of a single bigeye stock in the Pacific Ocean (Hampton et ah, Figure 6 Monthly variation in percentage of the western Pacific big- eye tuna with a terminal translucent zone in dorsal spine sections, February 1997 to January 1998. Age (year) Figure 7 Comparison of the growth curve obtained by the Ford- Walford plot with the growth curve obtained by nonlinear regression method for the western Pacific bigeye tuna. 1998; Grewe and Hampton3). Although the fishing area of the Taiwan fleet and thus the sampling area of bigeye tuna used in our study was limited to a small area of the western Pacific, our results may be representative of bigeye tuna throughout the Pacific Ocean. Monthly variation in percent terminal translucent edges in our study suggested the formation of growth rings once a year. Ehrhardt et al. ( 1996) attributed the narrow, trans- 3 Grewe, P. M., and J. Hampton. 1998. An assessment of bigeye ( Thunnus obesus ) population structure in the Pacific Ocean, based on mitochondrial DNA and DNA microsatellite analysis. University of Hawaii, Joint Institute for Marine and Atmosphere Research Contribution 98-320, 29 p. Pelagic Fish- eries Research Program, University of Hawaii at Manoa, 1000 Pope Road, Honolulu, HI 96822. 506 Fishery Bulletin 99(3) Table 2 Observed and back-calculated mean fork length (FL, cm) at age for the western Pacific bigeye tuna, Thunnus obesus. (“ — ” means there were no data owing to vascularization at core area). Numbers in normal print represent the mean back-calculated fork lengths; numbers in parentheses represent the number of specimens for which the specified ring was readable. Age (yr) n Observed mean FL (cm) Annulus number I II III IV V VI VII VIII IX X 1 8 67.9 57.0 (8) 2 79 93.7 50.0 80.8 (29) (79) 3 329 115.3 52.4 85.8 102.9 (69) (265) (329) 4 413 131.7 51.9 82.7 107.1 121.2 (19) (162) (384) (413) 5 188 145.9 — 80.9 112.1 122.8 135.8 (14) (77) (184) (188) 6 59 158.4 — — 115.5 126.6 138.2 149.5 (3) (36) (58) (59) 7 11 169.3 — 119.5 130.1 142.6 152.1 161.8 (1) (5) (9) (11) (11) 8 6 174.7 — — — 123.8 139.6 150.8 161.0 172.1 (1) (5) (6) (6) (6) 9 3 178.5 - — — 121.0 129.2 142.0 154.2 169.2 174.2 (1) (2) (2) (3) (3) (3) 10 2 188.5 - — — — 138.2 148.4 162.5 174.1 180.6 186.2 (1) (2) (2) (2) (2) (2) Total 1098 (125) (520) (794) (640) (263) (80) (22) (11) (5) (2) Weighted back-calculated mean FL (cm) 52.1 83.9 105.9 122.0 136.6 149.8 160.6 171.7 177.4 186.2 Growth increment (cm) - 31.9 21.9 16.2 14.6 13.1 10.9 11.1 5.7 8.7 Table 3 Growth parameters obtained by the Ford-Walford plot method and the nonlinear regression method for the bigeye tuna from the western Pacific Ocean. Parameter Ford-Walford plot Nonlinear regression Male Female Pooled Total7 Total7 n 278 180 458 1098 1098 K 0.1789 0.191 0.1842 0.185 0.2011 220.6 211.4 216.1 226.4 208.7 *0 -0.5566 -0.4592 -0.5266 -0.4465 -0.9906 1 Male, female, and sex-unknown combined. lucent rings to slower growth periods, whereas the wide, opaque rings were attributed to periods of fast growth. The spawning season of bigeye tuna in the western Pacific is between February and September and peaks from March to June (Sun, et al.4), the period that we found to coincide 4 Sun, C. L., S. L. Chu, and S. Z. Yeh. 1999. Note on reproduc- tion biology of bigeye tuna in the western Pacific. SCTB12/WP/ BET-4, 6 p. Twelfth meeting of the standing committee on tuna and billfish; Tahiti, French Polynesia, June 14-23, 1999. Oce- anic Fisheries Programme, Secretariate of the Pacific Commu- nity, B.P.D5, 98848 Noumea, New Caledonia. Sun et al.: Age and growth of Thunnus obesus 507 508 Fishery Bulletin 99(3) — •— Present study — O — Suda and Kume (1967), size frequency Yukinawa and Yabuta (1963), scales — is — Yukinawa and Yabuta (1963). size frequency — • — Shomura and Keala (1963), M, size frequency Figure 8 Growth curves of Pacific bigeye tuna estimated by different authors. with the slow-growth period indicated by the narrow and translucent rings. Similar findings have been reported for skipjack tuna (Antoine et ah, 1983), bigeye tuna (Gaikov et ah, 1980), and swordfish (Ehrhardt, 1992; Tserpes and Tsimenides, 1995). Our efforts only partially validate fish age; complete validation requires either mark-recapture data or the study of known-age fish in the population (Beamish and McFarlane, 1983; Prince et ah, 1995; Tser- pes and Tsimenides, 1995). We estimated the parameters of the VBGF by using the Ford-Walford and nonlinear methods and found that the nonlinear method had a better fit (r2=0.95) than the Ford- Walford method (r2=0.91). Comparisons of our VBGF pa- rameters with previous studies (Fig. 8, Table 4) showed similar results to those of Yukinawa and Yabuta (1963) who used scales and to those of Suda and Kume (1967) who also used Pacific samples of bigeye tuna. The values of t0 differed because different aging techniques were used. Following the suggestion of Gallucci and Quinn (1979), Vaughan and Kanciruk (1982), and Hanumara and Hoe- nig (1987) that Ford-Walford and other linear methods be replaced by nonlinear fitting techniques; we propose using parameters of VBGF estimated by the nonlinear method (Table 3) for description of age and growth for the western Pacific bigeye tuna. Acknowledgments We thank Pei-Ching Chiang who assisted in the field sam- pling at the fish market. We thank Sheng-Ping Wang who assisted in preparing some figures and tables. We also thank Clay E. Porch, Southeast Fisheries Sci- ence Center, National Marine Fisheries Service, and Chu-Fa Tsai, visiting professor at the Taiwan Endemic Species Research Institute, who reviewed the manu- script. We are indebted to three anonymous referees for their valuable comments and to John V. Merriner (scientific editor) and Sarah Shoffler (editorial assis- tant) for their editorial suggestions in revising the manuscript. Literature cited Alves, A., P. de Barros, and M. R. Pinho. 1998. Age and growth of bigeye tuna Thunnus obesus captured in the Madeira Archipelago. Int. Comm. Conserv. Atl. Tunas, Coll. Vol. Sci. Pap., vol. 48(2): 277-283. Antoine, L. M., J. Mendoza, and P. M. Cayre. 1983. Progress of age and growth assessment of Atlan- tic skipjack tuna, Euthynnus pelamis , from dorsal fin spines. In Proceedings of the international workshop on age determination of oceanic pelagic fishes: tunas, billfishes, and sharks (E. D. Prince, and L. M. Pulos, eds.), p. 91-97. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 8. Beamish, R. J., and G. S. McFarlane. 1983. The forgotten requirement for age validation in fisheries biology. Trans. Am. Fish. Soc. 112:735- 743. Calkins, T. P. 1980. Synopsis of biological data on the bigeye tuna, Thun- nus obesus (Lowe, 1839), in the Pacific Ocean. In Syn- opses of biological data on eight species of scombrids (W. H. Bayliff, ed.), p. 219-259. Inter-Am. Trop. Tuna Comm. Spec. Rep. 2. Cayre, P. M., and T. Diouf. 1983. Estimating age and growth of little tunny, Euthynnus alletteratus, off the coast of Senegal, using dorsal fin spine sections. In Proceedings of the international workshop on age determination of oceanic pelagic fishes: tunas, bill- fishes, and sharks (E. D. Prince, and L. M. Pulos, eds.), p. 105-110. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 8. Chen, Y., D. A. Jackson, and H. H. Harvey. 1992. A comparison of von Bertalanffy and polynomial func- tions in modelling fish growth data. Can. J. Fish. Aquat. Sci. 49:1128-1235. Compean-Jimenez, G., and F. X. Bard. 1983. Growth increments on dorsal spines of eastern Atlan- tic bluefin tuna, Thunnus thynnus , and their possible relation to migration patterns. In Proceedings of the international workshop on age determination of oceanic pelagic fishes: tunas, billfishes, and sharks (E. D. Prince, and L. M. Pulos, eds.), p. 77-86. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 8. Delgado de Molina, A., and J. C. Santana. 1986. Estimacion de edad y crescimiento del patudo (Thun- nus obesus , Lowe 1839) capturado en las Islas Canarias. Int. Comm. Conserv. Atl. Tunas, Coll. Vol. Sci. Pap., vol. 25:130-137. Draganik, B., and W. Pelczarski. 1984. Growth and age of bigeye and yellowfin tuna in the central Atlantic as far data gathered by RW “Wieczno”. Int. Comm. Conserv. Atl. Tunas, Coll. Vol. Sci. Pap., vol. 20( 1 ):96— 103. Sun et al.: Age and growth of Thunnus obesus 509 Ehrhardt, N. M. 1992. Age and growth of swordfish, Xiphias gladius, in the northwestern Atlantic. Bull. Mar. Sci. 50:292-301. Ehrhardt, N. M., R. J. Robbins, and F. Arocha. 1996. Age validation and growth of swordfish, Xiphias gla- dius, in the northwest Atlantic. Int. Comm. Conserv. Atl. Tunas, Coll. Vol. Sci. Pap., vol. 45(21:358-367. Gaikov, V. V., V. N. Chur, V. L. Zharov, and Yu P. Fedoseev. 1980. On age and growth of the Atlantic bigeye tuna. Int. Comm. Conserv. Atl. Tunas, Coll. Vol. Sci. Pap., vol. 9(21:294-302. Gallucci, V. F., and T. J. Quinn II. 1979. Reparameterizing, fitting, and testing a simple growth model. Trans. Am. Fish. Soc. 108:14-25. Gonzalez-Garces, A., and A. C. Farina-Perez. 1983. Determining age of young albacore, Thunnus ala- lunga, using dorsal spines. In Proceedings of the interna- tional workshop on age determination of oceanic pelagic fishes: tunas, billfishes, and sharks (E. D. Prince, and L. M. Pulos, eds.l, p. 117-122. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 8. Gulland, J. A. 1983. Fish stock assessment: a manual of basic methods. Wiley, Chichester, 223 p. Hampton, J., K. Bigelow, and M. Labelle. 1998. A summary of current information on the biology, fish- eries and stock assessment of bigeye tuna (Thunnus obesus 1 in the Pacific Ocean, with recommendation for data require- ments and future research. Tech. Rep. 36, 46 p. Oceanic Fisheries Programme, Secretariate of the Pacific Commu- nity, B.P.D5, 98848 Noumea, New Caledonia. Hanumara, R. C., and N. A. Hoenig. 1987. An empirical comparison of a fit of linear and non- linear models for seasonal growth in fish. Fish. Res. 5:359-381. Hill, K. T., G. M. Cailliet, and R. L. Radtke. 1989. A comparative analysis of growth zones in four calci- fied structures of Pacific blue marlin, Makaira nigricans. Fish. Bull. 87:829-843. Iversen, E. S. 1955. Size frequencies and growth of central and western Pacific bigeye tuna. Spec. Sci. Rep. U.S. Fish. Wildl. Serv. (Fish.) 162:1-40. Kume, S., and J, Joseph. 1966. Size composition, growth and sexual maturity of bigeye tuna, Thunnus obesus (Lowe), from the Japanese long-line fishery in the eastern Pacific Ocean. Bull. Inter-Am. Trop. Tuna Comm. 11(21:45-99. Lee, R. M. 1920. A review of the methods of age and growth determi- nation by means of scales. Fishery Invest., Lond., ser. II, 4(2), 32 p. Matsumoto, T. 1998. Preliminary analyses of age and growth of bigeye tuna (Thunnus obesus) in the western Pacific Ocean based on otolith increments. In Proceedings of the first world meet- ing on bigeye tuna (R. S. Deriso, W. H. Bayliff, and N. J. Webb, eds.), p. 238-242. Inter-Am. Trop. Tuna Comm. Spec. Rep. 9. Nose, Y., H. Kawatsu, and Y. Hiyama. 1957. Age and growth of Pacific tunas by scale reading. In Collection of works on fisheries science. Jubilee publication of Professor I. Amemiya, p. 701-716. Univ. Tokyo Press. Prince, E. D., D. W. Lee, J. L. Cort, G. A. McFarlane, and A. Wild. 1995. Age validation evidence for two tag-recaptured Atlan- tic albacore, Thunnus alalunga, based on dorsal, anal, and pectoral finrays, vertebrae, and otoliths. In Recent devel- opments in fish otolith research (D. H. Secor, J. M. Dean, and S. E. Campana, eds.), p. 375-398. LTniv. South Caro- lina Press, Columbia, SC. Ratkowsky, D. A. 1983. Nonlinear regression modelling. Marcel Dekker, Inc., New York, NY, 276 p. Shomura, R. S. 1966. Age and growth studies of four species of tunas in the Pacific Ocean. In Proceedings of the government’s confer- ence on central Pacific fishery resources (T. A. Manar, ed.), p. 203-219. Bureau of Commercial Fisheries, Biological Laboratory, Honolulu, Hawaii. Shomura, R. S., and B. A. Keala. 1963. Growth and sexual dimorphism in growth of bigeye tuna ( Thunnus obesus ) — a preliminary report. In Pro- ceedings of the world scientific meeting on the biology of tunas and related species. La Jolla, California, 2-14 July 1962, p. 1409-1417. FAO Fisheries Report 6. Steel, R. G. D. 1980. Principles and procedures of statistics: a biometrical approach. McGraw-Hill, Auckland, 633 p. Suda, A., and S. Kume. 1967. Survival and recruit of bigeye tuna in the Pacific Ocean, estimated by the data of tuna longline catch. Rep. Nankai. Reg. Fish. Res. Lab. ( 25 ):9 1—104. Tserpes, G., and N. Tsimenides. 1995. Determination of age and growth of swordfish, Xiphias gladius L., 1758, in the eastern Mediterranean using anal- fin spines. Fish. Bull. 93:594-602. Vaughan, D. S., and P. Kanciruk. 1982. An empirical comparison of estimation procedures for the von Bertalanffy growth equation. J. Cons. Explor. Mer 40:211-219. Yukinawa, M., and Y. Yabuta. 1963. Age and growth of bigeye tuna, Parathunnus mebachi (Kishinouye). Rep. Nankai Reg. Fish. Res. Lab. 19:103- 118. Zar, J. H. 1999. Biostatistical analysis. 4th ed., Prentice-Hall, Engle- wood Cliffs, NJ, 929 p. 510 Nocturna! occurrence of the swimming crab Ovalipes punctatus in the swash zone of a sandy beach in northeastern Japan Kazutaka Takahashi Kouichi Kawaguchi Ocean Research Institute University of Tokyo 1-15-1 Minamidai Nakano, Tokyo 164-8639, Japan Present address (for K. Takahashi): Tohoku National Fisheries Research Institute 3-27-5 Shinhama Shiogama, Miyagi 985-0001 Japan E-mail address (for K. Takahashi): issey@affrc.go jp Swimming crabs belonging to the genus Ovalipes (Crustacea: Brachyura: Por- tunidae) are distributed worldwide along sandy coastlines of subtropical and temperate waters (Stephenson and Rees, 1968). They are extremely well adapted to life on sand, with their strong swimming and burrowing abil- ities (Brown and McLachlan, 1990). Crabs of this genus are voracious carni- vores and significant predators of com- mercially important mollusks on sandy beaches (Du Preez, 1984; Brown and McLachlan, 1990). Recently, interest in their importance as a possible fisher- ies resource has increased (Sasaki and Kawasaki, 1980; Wear and Haddon, 1987). Ovalipes punctatus (De Haan) is common in the coastal waters of Ja- pan and China. It is found on sandy bottoms below the intertidal zone, but peak abundance is at 5-60 m. This species has an ontogenetic migration into deeper water (Kamei, 1976; Sasa- ki and Kawasaki, 1980). Sasaki and Kawasaki (1980) investigated the life history of O. punctatus in Sendai Bay (100 km south of the present study site), on the Pacific coast of northeast- ern Japan. Female crabs spawn from mid-September until mid-November in offshore water 40-60 m deep. The in- cubation period of the eggs is about 20 days. After a planktonic phase, the lar- vae settle on sandy bottoms in March and April. Settled crabs of both sexes grow to carapace widths (CW ) of 50-65 mm in the first year. In the second year, females and males grow to 70-90 and 80-85 mm CW, respectively. The life span of this crab is 2 to 2.5 years. The minimum size at maturity is es- timated to be 45-50 mm CW (Sasaki and Kawasaki, 1980), but the spawning population consists mainly of 2-year- old crabs of 75-80 mm CW. However, knowledge of the diet and feeding hab- its of this species has been extremely limited until now. Ovalipes species, including O. punc- tatus, generally live below the intertid- al zone. However, they are often found in the intertidal zone, especially at night (McLachlan et al., 1979; Brown and McLachlan, 1990). McLachlan et al. (1979) suggested that this occur- rence is related to feeding, but details are still unknown. During the course of an ecological study of the sandy beach in Otsuchi Bay, we frequently found O. punctatus in the swash zone at night. This note examines the significance of the swash zone in the life history of the swimming crab O. punctatus in rela- tion to feeding and molting. Materials and methods The study was conducted at Koshira- hama Beach, in Otsuchi Bay, on the Pacific coast of northern Honshu, main- land Japan, in September of 1994 and August of 1995 (Table 1). The beach is about 120 m long, bounded on both ends by rocky shores, and is catego- rized as a “sheltered beach” according to McLachlan’s (1980) rating scheme. Mean depth at 20, 50, and 100 m from the shoreline is 1.6, 3.5, and 5.5 m, respectively. The beach has a 1:13 (rise: run) slope and median particle diam- eter of the sediment is 260 pm. Other details of the study site are given in Takahashi and Kawaguchi (1995). Crabs were sampled with a beach seine 1.1 m high by 8.3 m wide with 1.0-mm mesh. The net was hauled par- allel to the shoreline at a depth of ca. 0-1 m. Samples were taken along the entire shoreline of Koshirahama Beach. All crabs were preserved in 10% formalin, and the carapace width (CW), sex, maturity and molt stage of each individual were recorded. The foregut was removed and stored in 70% ethanol for dietary analysis. Sex- ual maturity of O. punctatus was as- signed by size classes as defined by Sasaki and Kawasaki ( 1980): juvenile (<30 mm CW), immature (31-50 mm CW), or adult (>51 mm CW). Molt stages were based on the criteria of Drach and Tchernigovtzeff (1967) and Norman and Jones (1992): 1) soft, no calcification of the new exoskeleton; 2) early papershell, thin, flexible exoskel- eton, easily depressed when touched; 3) late papershell, hard exoskeleton except for the branchiostegite region which is compressible; 4) intermolt, completely hard exoskeleton; and 5) premolt, teeth on chelae well worn and having a less rounded appearance compared with those in earlier stages, complete exocuticle developed beneath the exoskeleton. The diet of O. punctatus was an- alyzed by using the points method and the percentage occurrence meth- od (Williams, 1981; Wear and Haddon, 1987). The points method assesses di- et composition in terms of both foregut fullness and estimated volume of food in the foregut. First, the relative degree of foregut fullness of each crab was es- timated visually by using six ordered classes (empty=class 0; trace=class 1; 25%=class 2; 50%=class 3; 75%=class 4; full=class 5). A visual assessment of fullness was possible because, ex- cept for the gastric mill, the foregut of O. punctatus is like a thin-walled trans- lucent bag. Manuscript accepted 18 December 2000. Fish. Bull. 99:510-515 (2001) NOTE Takahashi and Kawaguchi: Nocturnal occurrence of Ovalipes punctcitus 511 Table 1 Details of the sampling dates and times (JST) and results in the swash zone of Koshirahama Beach, Otsuchi Bay. Date Day or night Sampling time (h) No. of crabs collected 22 Aug 1995 Day 8:02 3 25 Aug 1995 Day 13:12 2 30 Aug 1995 Day 13:15 3 12 Sept 1994 Night 20:54 30 25 Aug 1995 Night 21:06 23 30 Aug 1995 Night 20:48 35 Food items were identified to the lowest possible taxon under a binocular dissecting microscope. The relative con- tribution of each prey category to the total volume of the foregut contents was assessed subjectively in the follow- ing way: a category representing 95-100% of the total con- tents was awarded 100 points; 65-95%, 75 points; 35-65%, 50 points; 5-35%, 25 points; 5% or less, 2.5 points; empty, 0 points. The points that each prey category received were weighted by multiplying by a factor that depended on the degree of foregut fullness, i.e. full = 1, 75% = 0.75, 50% = 0.5, 25% = 0.25 and trace = 0.02 (Wear and Haddon, 1987). The maximum and minimum weighted points possible for a single category in a single foregut were 100 ( 100 x 1.0) and 0.05 (2.5 x 0.02), respectively. The following percentages were calculated for each prey category (Williams, 1981): Percentage points for zth prey = V i 100; and Percentage occurrence for zth prey = (6/AD100, where atJ - the number of points for prey item i in the foregut of the yth crab; A = the total points for all the crabs and all the prey items in all the foreguts examined; N - the number of crabs examined with food in the foregut; and bt = the number of crabs with foreguts containing prey category i. Results Die! change in the swash zone occurrence pattern A total of 96 crabs (41 males, 55 females), ranging from 12 to 80 mm CW, were collected from the swash zone of Koshirahama Beach (Fig. 1). No ovigerous females were collected. Almost all specimens were collected at night (88 indi- viduals, 92% of the total catch) and only 8 crabs were col- lected during the day (Fig. 1). Juvenile (35) and immature m ro u "D > c 10 r- Day J I I L 10 20 30 40 50 60 70 80 90 Carapace width (mm) Figure 1 Histograms of carapace width for the swimming crab Ovalipes punctatus found in the swash zone of Koshira- hama Beach, Otsuchi Bay, during the day and at night. All the specimens were collected in September 1994 or August 1995. (39) crabs dominated the nocturnal catches, accounting for 40% and 44% of the total nocturnal catch, respectively (Table 2). Fourteen adult crabs were also caught at night, constituting 16% of the nocturnal samples (Table 2). Day- time catches consisted of 1 immature and 7 juvenile crabs (Table 2). On average, crabs caught at night (36 mm CW) were significantly larger than crabs caught during the day (25 mm CW) (P<0.01; f-test). All crabs were in the intermolt or the papershell stages (early and late papershell); no soft-shelled and premolt crabs were collected. During the daytime, 7 out of 8 crabs were in the papershell stages and only one individual was classified as intermolt (Table 2). At night, 5 juvenile and 5 immature papershell crabs were collected, accounting for 14% and 13% of the total catch for each category, respec- tively (Table 2). Of the adults collected at night, 9 were in the papershell stages, accounting for 64% of the total adult crabs taken. Feeding activity and diet composition Because feeding habits of early and late papershell crabs are not different from those of intermolt stages in por- tunid crabs (Norman and Jones, 1992), analysis for fore- gut fullness and diet was conducted by pooling all molt stages. The percent frequency of each class of foregut full- ness is shown separately for day and night in Figure 2. Although the daytime sample was small, 50% of the crabs had empty foreguts (Fig. 2), but at night over 70% of the 512 Fishery Bulletin 99(3) Table 2 Ovalipes punctatus caught in the swash zone of Koshirahama Beach, Otsuchi Bay, during the summers of 1994 and 1995. The numbers of each developmental stage and of papershell crabs caught during the day and at night are shown. No softshell and premolt crabs were collected. Number of papershell crabs Total number of Day or night Developmental stage of crabs (CW) crabs collected Early papershell Late papershell Day Juvenile (12-30 mm) 7 6 Immature (31-50 mm) 1 1 — Adult (51-80 mm) — — — Night Juvenile (12-30 mm) 35 4 1 Immature (31-50 mm) 39 4 1 Adult (51-80 mm) 14 3 6 Table 3 Percentage occurrence, points, and percentage points for the 12 categories of foregut content in 69 Ovalipes punctatus from Koshi- rahama Beach, Otsuchi Bay. The points express the dietary contribution in terms of foregut fullness and estimated volume of food items in the foregut (see text). Sand was excluded from the calculation of the percentage points. Food item Percentage occurrence Points Percentage points Crustaceans Haustorioides japonicus 59 1437.5 26.1 Archaeomysis kokuboi 54 1287.5 23.4 Excirolana chiltoni 30 528.7 9.6 Crangon sp. 14 540.0 9.8 Crustacean fragments 42 1115.6 20.3 Other crustaceans 4 62.5 1.1 Mollusks Bivalves 6 41.3 0.7 Gastropods 1 25.0 0.5 Fishes Unidentified fish 3 5.0 0.1 Other items Unidentified organic matter 19 456.3 8.3 Algae 3 3.8 0.1 Sand 94 970.6 — foreguts examined were more than half full (classes 3-5) and 7% had empty foreguts, which suggests that the crabs fed actively in the swash zone at night (Fig. 2). When describing the diet of portunid crabs, Williams (1981) recommended including only individuals with fore- guts more than 50% full (i.e. classes 3-5). Applying this criterion, we found 69 crabs (39 females, 30 males) for the diet analysis, 66 of which were collected at night. Because there was no significant difference in the diets by sex (chi- square test, P>0.5), the male and female data were com- bined. Forguts contained 11 prey categories plus sand; percent frequency of occurrence, total points, and the rela- tive proportions of the diet components in terms of points are listed in Table 3. Although sand is not considered a part of the diet, it contributed more than 14% to the total points of the foregut contents. Small crustaceans were frequent and predominant in the foreguts of O. punctatus (Table 3). The sand-burrowing am- phipod Haustorioides japonicus was the most frequent prey, contributing 26.1% of the total points. Second most frequent prey was the sand-burrowing mysid Archaeomysis kokuboi (23.3% of the total points), followed by the sand-burrowing isopod Excirolana chiltoni (9.6%) and sand shrimps Cran- gon sp. (9.8%). Fragments that probably came from these crustaceans produced 20.3% of the total points (Table 3). The results show 90.2% of the total diet volume of O. punc- tatus consisted of crustacean prey. Mollusks and fish were not important in the diet (1.3% collectively). NOTE Takahashi and Kawaguchi: Nocturnal occurrence of Ovalipes punctatus 513 Rank of foregut fullness index □ class 0 □ class 1 □ class 2 EB class 3 □ class 4 nclass 5 Figure 2 Difference in the foregut fullness index of the swimming crab Ovalipes punc- tatus in the swash zone of Koshirahama Beach, Otsuchi Bay between day and night. Class 0 = empty foregut, class 1 = trace items in foregut, class 2 = foregut 25% full, class 3 = 50%> full , class 4 = 75% full, class 5 = full. Discussion Ovalipes punctatus occurred nocturnally in the swash zone of Koshirahama Beach, Otsuchi Bay. The nocturnal occurrence in the swash zone for congeners such as O. tri- maculatus from South Africa (reported as O. punctatus, see Schoeman and Cockcroft, 1993) is often observed else- where (McLachlan et ah, 1979; Brown and McLachlan, 1990). McLachlan et al. (1979) suggested that occurrence in the swash zone is related to feeding. Generally, Ovali- pes species are active nocturnally (Caine, 1974; Du Preez, 1983). Our observations of foregut fullness and abundance support these findings. Furthermore, because peracarid crustaceans that live close to the sandy shoreline domi- nated the foregut contents (Kamihira, 1979; Takahashi and Kawaguchi, 1995; authors’ pers. obs.), we concluded that part of the crab population migrates from deeper waters into the swash zone at night to exploit the crusta- cean fauna of sandy beaches. Feeding in the swash zone is advantageous for the crabs because important prey, such as the amphipod H. japoni- cus, the mysid A. kokuboi, and the isopod E. chiltoni, are abundant in the swash zone of Koshirahama Beach. These species are considerably more abundant near the shore- line than other macrobenthic organisms that live in the offshore area (authors’ pers. obs.). The peracarids, especial- ly mysids and isopods, burrow into the sand during day- time, then emerge into water column in the swash zone at night (Takahashi and Kawaguchi, 1997; authors’ pers. obs.) which makes them more vulnerable to predation by O. punctatus and fishes (Takahashi et al., 1999). The predomi- nance of peracarid crustaceans in the foreguts of the crabs suggests that O. punctatus capture the peracarids effective- ly in the swash zone of Koshirahama Beach at night. Species of Ovalipes are known as opportunistic, broad- spectrum predators. They are extremely versatile, feed- ing on both less active prey, such as mollusks, and on mobile animals, such as amphipods, mysids, isopods, and fish (Caine, 1974; Haefner, 1985; Wear and Haddon, 1987; Ropes, 1989; Stehlik, 1993). The swash zone also serves as a refuge from predation. Species of Ovalipes (including congeneric species) are oc- casionally found in the guts of fish and decapod crusta- ceans (McDermott, 1983; Du Preez and McLachlan, 1984; Mitchell, 1984; Wear and Haddon, 1987; Stehlik, 1993). Cannibalism is a major cause of mortality in many por- tunid crabs during their early life stages; the importance of refuges has been emphasized as a factor regulating crab recruitment in natural populations (Hines and Ruiz, 1995). In estuarine ecosystems, habitats with structural complexity, i.e. seagrass beds, mussel beds, filamentous algae, are effective as refuges for juvenile crabs, whereas risk of predation is higher in open sand, the main habitat of Ovalipes (Wilson et al., 1987; Ryer et al., 1997, Moksnes et al., 1998). In sand habitat, small Ovalipes burrow deep by using reverse gill current to hide from predators during the day (Barshaw and Able, 1990). Nocturnal emergence of Ovalipes makes them more vul- nerable to cannibalism. In the swash zone of Koshiraha- ma Beach at night, almost all the O. punctatus were ju- venile and immature crabs. This finding suggests that small individuals avoid the deeper zone, where the risk of cannibalism is high. For blue crabs, Callinectes sapidus, shallow water areas function as refuges from predation 514 Fishery Bulletin 99(3) when habitats have little structural complexity (Dittel et al., 1995; Hines and Ruiz, 1995). The swash zone is prob- ably difficult for larger predators to penetrate and they risk being stranded. In addition, birds that exploit the swash zone are less active at night (Brown and McLach- lan, 1990). Softshell and premolt crabs were not collected in the swash zone; it appears that this zone may not be a molt- ing ground owing to its dynamic characteristics. However, most diurnal and adult crabs in the samples were paper- shell crabs, which are also vulnerable to predation; the swash zone, therefore, may be important as a refuge for postmolt crabs. In our study, juvenile and immature crabs dominated the population of O. punctatus in the swash zone during late summer. From the size of the crabs, these were first- year crabs born in the previous winter or spring, and therefore their occurrence in the swash zone may be lim- ited to the summer and fall, when these juvenile and im- mature crabs appear (Sasaki and Kawasaki, 1980). A simi- lar seasonal occurrence of juvenile O. ocellatus from July to October on exposed sandy beaches of the mid-Atlantic Coast of the United States has been reported in an inshore area (McDermott, 1983). The occurrence and behavior of O. punctatus in the swash zone suggest that this habitat is important as a feeding ground and a refuge from preda- tion. Intertidal sand-burrowing peracarids, the major diet items of O. punctatus, also consume particulate organic matter and organisms in the swash zone (Kamihira, 1992; Takahashi and Kawaguchi, 1998). Thus by feeding in the swash zone, young O. puncta tus transport some of the sec- ondary production from the swash zone to the offshore ar- ea through their ontogenetic migration into deeper water. Acknowledgments We are grateful to Captain K. Morita and T. Kawamura, and K. Hirano of the Otsuchi Marine Research Center, Ocean Research Institute (ORI), for their help in field sampling. We would like to thank C. Vallet of Soka Uni- versity who translated the French articles. We also thank T. R Hirose of the Japan Sea National Fisheries Research Institute and the staff of the Plankton Laboratory of ORI for their help with fieldwork and for fruitful discussions during the course of our study. Literature cited Barshaw, D. E., and K. W. Able. 1990. Deep burial as a refuge for lady crabs Ovalipes ocellatus : comparisons with blue crabs Callinectes sapi- dus. Mar. Ecol. Prog. Ser. 66:75-79. Brown, A. C., and A. McLachlan. 1990. Ecology of sandy shores. Elsevier Science Publish- ing Company Inc., Amsterdam, 328 p. Caine, E. A. 1974. Feeding of Ovalipes guadulpensis (Saussure) (Decap- oda: Brachyura: Portunidae), and morphological adapta- tions to a burrowing existence. Biol. Bull. 147:550-559. Dittel, A. I., A. H. Hines, G. M. Ruiz, and K. Ruffin. 1995. Effects of shallow water refuge on behavior and den- sity-dependent mortality of juvenile blue crabs in Chesa- peake Bay. Bull. Mar.Sci. 57:902-916. Drach, P, and C. Tchernigovtzeff. 1967. Sur la methode de determination des estades d’inter- mue et son application generale aux crustaces. Vie Milieu 18:595-609. Du Preez, H. H. 1983. The effects of temperature, season and activity on the respiration of the three spot swimming crab, Ovalipes punctatus. Comp. Biochem. Physiol. 75A:353-362. 1984. Molluscan predation by Ovalipes punctatus (De Haan) (Crustacea: Brachyura: Portunidae). J. Exp. Mar. Biol. Ecol. 84:55-71. Du Preez, H. H., and A. McLachlan. 1984. Biology of the three-spot swimming crab, Ovalipes punctatus (De Haan) II. Growth and molting. Crusta- ceana 47:113-120. Haefner, P. A. Jr. 1985. Morphometry, respiration, diet and epizoites of Ovali- pes stephensoni Williams, 1976 (Decapoda, Brachyura). J. Crust. Biol. 5:658-672. Hines, A. H., and G. M. Ruiz. 1995. Temporal variation in juvenile blue crab mortality: nearshore shallows and cannibalism in Chesapeake Bay. Bull. Mar. Sci. 57:884-901. Kamei, M. 1976. Biology of Hiratsume-gani, Ovalipes punctatus (DE HAAN), in Sagami Bay. Jpn J Ecol 26:65-69. [In Jpn., Engl, abstr.] Kamihira, Y. 1979. Ecological studies of macrofauna on a sandy beach of Hakodate, Japan II. On the distribution of peracarids and the factors influencing their distribution. Bull. Fac. Fish. Hokkaido Univ. 30:133-143. [In Jpn., Engl, abstr.] 1992. Ecological studies of sand-burrowing amphipod Haus- torioides japonicus Dogielinotidae), on the south-western Hokkaido, Japan. Rev. Hakodate Univ., spec. no. 1:1-106. [In Jpn., Engl, abstr.] McDermott, J. J. 1983. Food web in the surf zone of an exposed sandy beach along the mid-Atlantic Coast of the United States. In Sandy beaches as ecosystems (A. McLachlan and T. Eras- mus, eds.), p. 529-538. Junk Publishers, The Hague. McLachlan, A. 1980. The definition of sandy beaches in relation to expo- sure: a simple rating system. S. Afr. J. Sci. 76:137-138. McLachlan. A., T. Wooldridge, and G. Van der Horst. 1979. Tidal movements of the macrofauna on an exposed sandy beach in South Africa. J. Zool. Lond. 187:433-442. Mitchell, S. J. 1984. Feeding of ling Genypterus blacodes (Bloch and Sch- neider) from New Zealand offshore fishing grounds. N. Z. J. Mar. Freshwater Res. 18:265-274. Moksnes, P-O, L. Pihl, and J. van Montfrans. 1998. Predation on post larvae and juveniles of the shore crab Carcinus tnaenas: importance of shelter, size and can- nibalism. Mar. Ecol. Prog. Ser. 166:211-225. Norman, C. P, and M. B. Jones. 1992. Influence of depth, season, and moult stage on the diet analysis of the velvet swimming crab Necora puber ( Branchy - ura, Portunidae). Estuar. Coast. Shelf Sci. 34:17-83. Ropes, J. W. 1989. The food habits of five crab species at Pettaquamscutt River, Rhode Island. Fish. Bull. 87: 197-204. NOTE Takahashi and Kawaguchi: Nocturnal occurrence of Ovalipes punctatus 515 Ryer, C. H., J. van Montfrans, and K. E. Moody. 1997. Cannibalism, refugia and the molting blue crab. Mar. Ecol. Prog. Ser. 147:77-85. Sasaki, K., and T. Kawasaki. 1980. Some aspects of the reproductive biology of the swimming crab, Ovalipes punctatus (De Haan), in Sendai Bay and its adjacent waters. Tohoku J. Agric. Res. 30:183-194. Schoeman, D. S., and A. C. Cockcroft. 1993. On the misidentification of a common sandy crab belonging to the genus Ovalipes Rathbun, 1898. S. Afr. J. Zool. 28:124-125. Stehlik, L. L . 1993. Diets of the brachyuran crabs Cancer irroratus, C. borealis , and Ovalipes ocellatus in the New York Bight. J. Crust. Biol. 13:723-735. Stephenson, W., and M. Rees. 1968. A revision of the genus Ovalipes Rathbun, 1898 (Crus- tacea, Decapoda, Portunidae). Rec. Aust. Mus. 27:213- 261. Takahashi, K., T. Hirose, and K. Kawaguchi. 1999. The importance of the intertidal sand-burrowing per- acarid crustaceans as prey organisms of fish in the surf zone of sandy beach at Otsuchi Bay, northeastern Japan. Fisheries Sci. 65:856-864. Takahashi, K., and K. Kawaguchi. 1995. Inter- and intraspecific zonation in three species of sand-burrowing mysids, Archaeomysis kokuboi, A. greb- nitzkii and Iiella ohshimai , in Otsuchi Bay, northeastern Japan. Mar. Ecol. Prog. Ser. 116:75-84. 1997. Diel and tidal migrations of the sand-burrowing mysids, Archaeomysis kokuboi , A. japonica and Iiella ohshi- mai, in Otsuchi Bay, northeastern Japan. Mar. Ecol. Prog. Ser. 148:95-107. 1998. Diet and feeding rhythm of the sand-burrowing mysids, Archaeomysis kokuboi and A. japonica in Otsuchi Bay northeastern Japan. Mar. Ecol. Prog. Ser. 162:191- 199. Wear, R. G., and M. Haddon. 1987. Natural diet of the crab Ovalipes catharus (Crusta- cea, Portunidae) around central and northern New Zea- land. Mar. Ecol. Prog. Ser. 35:39-49. Wilson, K. A., K. L. Heck Jr., and K. W. Able. 1987. Juvenile blue crab, Callinectes sapidus, survival: an evaluation of eelgrass, Zostera marina, as refuge. Fish. Bull. 85:53-58. Williams, M. J. 1981. Methods for analysis of natural diet in portumd crabs (Crustacea: Decapoda: Portunidae). J. Exp. Mar. Biol. Ecol. 52:103-113. 516 Fishery Bulletin 99(3) Errata Fisheiy Bulletin 98(41:759-766 Lindley, Steven T., Michael S. Mohr, and Michael H. Prager Monitoring protocol for Sacramento River winter chinook salmon, Oncorhynchus tshawytscha: application of statistical analysis to recovery of an endangered species Erratum 1: page 759, right column, line 10. The equation within the text should read R, =N,/N,_3. Erratum 2: page 760. Equation 1 should read ^ _ 1 ~ Pgoal s / yfil Erratum 3: page 761. Equation 2 should read ? P - Pgoal 0 = 7= =-• a / yn Fishery Bulletin 99(1 ):49-62 Gharrett, Anthony J., Andrew K. Gray, and Jonathan Heifetz Identification of rockfish (Sebastes spp.) by restriction site analysis of the mitochondrial ND-3/ND-4 and 12S/16S rRNA gene regions Erratum: page 56, line 5 Read “S. melanops” instead of “S. maliger” so that the sen- tence begins “Similarly, the haplotypes of S. melanops and S. flavidus (subgenus Sebastosomus ) were tightly clustered . . .”. Fishery Bulletin 99(3) 517 Superintendent of Documents Publications Order Form *5178 1 I YES, please send me the following publications: Subscriptions to Fishery Bulletin for $55.00 per year ($68.75 foreign) The total cost of my order is $ . Prices include regular domestic postage and handling and are subject to change. (Company or Personal Name) (Please type or print) (Additional address/attention line) (Street address) (City, State, ZIP Code) (Daytime phone including area code) (Purchase Order No.) Charge your order. IT’S EASY! Please Choose Method of Payment: ] Check Payable to the Superintendent of Documents | GPO Deposit Account [ | j j ! j { j ] VISA or MasterCard Account To fax your orders (202) 512-2250 (Credit card expiration date) (Authorizing Signature) Mail To: Superintendent of Documents P.O. Box 371954, Pittsburgh, PA 15250-7954 Thank you for your order! Fishery Bulletin Guidelines for contributors Content of papers Articles Articles are reports of 10 to 30 pages (double spaced) that describe original research in one or a combination of the following fields of marine science: taxonomy, biology, genetics, mathematics (including modeling), statistics, engineering, eco- nomics, and ecology. Notes Notes are reports of 5 to 10 pages without an abstract that describe methods and results not supported by a large body of data. Although all contributions are subject to peer review, responsi- bility for the contents of articles and notes rests upon the authors and not upon the editor or the publisher. It is therefore important that authors consider the contents of their manuscripts care- fully. Submission of an article is un-derstood to imply that the article is original and is not being considered for publication elsewhere. Manuscripts must be written in English. Authors whose native language is not English are strongly advised to have their manuscripts checked for fluency by English-speaking colleagues prior to submission. Preparation of papers Text Title page should include authors’ full names and mailing addresses (street address required) and the senior author’s telephone, fax number, e-mail address, as well as a list of key words to describe the contents of the manuscript. Abstract must be less than one typed page (double spaced) and must not contain any citations. It should state the main scope of the research but emphasize the author’s con- clusions and relevant findings. Because abstracts are circulated by abstracting agencies, it is impor- tant that they represent the research clearly and concisely. General text must be typed in double- spaced format. A brief introduction should state the broad significance of the paper; the remainder of the paper should be divided into the following sec- tions: 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 subdi- vision (i.e. there can be no subdivision without at least two subheadings). The entire text should be intelligible to interdisciplinary readers; therefore, all acronyms and abbreviations should be written out and all lesser-known technical terms should be defined 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 Sci- entific style and format: CBE manual for authors, editors, and publishers (6th ed.) for editorial style and the most current issue of the American Fish- eries Society’s common and scientific names of fishes from the United States and Canada for fish nomenclature. Dates should be written as fol- lows: 11 November 1991. Measurements should be expressed in metric units, e.g. metric tons (t). The numeral one (1) should be typed as a one, not as a lower-case el (1). Footnotes Use footnotes to add editorial comments regarding claims made in the text and to document unpub- lished works or works with local circulation. Foot- notes should be numbered with Arabic numerals and inserted in 10-point font at the bottom of the first page on which they are cited. Footnotes should be formatted in the same manner as citations. If a manuscript is unpublished, in the process of review, or if the information provided in the footnote has been conveyed verbally, please state this information as “unpubl. data,” “manuscript in review,” and “personal commun.,” respectively. Authors are advised wherever possible to avoid ref- erences to nonstandard literature (unpublished lit- erature that is difficult to obtain, such as internal reports, processed reports, administrative reports, ICES council minutes, IWC minutes or working papers, any “research” or “working” documents, laboratory reports, contract reports, and manu- scripts in review). If these references are used, please indicate whether they are available from NTIS (National Technical Information Service) or from some other public depository. Footnote format: author (last name, followed by first-name initials); year; title of report or manuscript; type of report and its administrative or serial number; name and address of agency or institution where the report is filed. Literature cited The literature cited section comprises works that have been published and those accepted for pub- lication (works in press) in peer-reviewed jour- nals and books. Follow the name and year system for citation format. In the text, write “Smith and Jones (1977) reported” but if the citation takes the form of parenthetical matter, write “(Smith and Jones, 1977).” In the literature cited section, list citations alphabetically by last name of senior author: For example, Alston, 1952; Mannly, 1988; Smith, 1932; Smith, 1947; Stalinsky and Jones, 1985. Abbreviations of journals should conform to the abbreviations given in the Serial sources for the BIOSIS previews database. Authors are responsible for the accuracy and completeness of all citations. Literature citation format: author (last name, followed by first-name initials); year; title of report or article; abbreviated title of the journal in which the article was published, volume number, page numbers. For books, please provide publisher, city, and state. Tables Tables should not be excessive in size and must be cited in numerical order in the text. Headings in tables 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 legend. Other incidental comments may be footnoted (use italic arabic numerals for footnote markers). Use asterisks only to indicate probability in statistical data. Place table legends on the same page as the table data. We accept tables saved in most spread- sheet software programs (e.g. Microsoft Excel). Please note the following: • Use a comma in numbers of five digits or more (e.g. 13,000 but 3000). • Use zeros before all decimal points for values less than one (e.g. 0.31). Figures Figures include line illustrations, computer-gener- ated line graphs, and photographs (or slides). They must be cited in numerical order in the text. Line illustrations are best submitted as original draw- ings. Computer-generated line graphs should be printed on laser-quality paper. Photographs should be submitted on glossy paper with good contrast. All figures are to be labeled with senior author’s name and the number of the figure (e.g. Smith, Fig. 4). Use Helvetica or Arial font to label anatomical parts (line drawings) or variables (graphs) within figures; use Times Roman bold font to label the dif- ferent sections of a figure (e.g. A, B, C). Figure leg- ends should explain all symbols and abbreviations seen within the figure and should be typed in dou- ble-spaced format on a separate page at the end of the manuscript. We advise authors to peruse a recent issue of Fishery Bulletin for standard for- mats. Please note the following: • Capitalize the first letter of the first word of axis labels. • Do not use overly large font sizes to label axes or parts within figures. • Do not use boldface fonts within figures. • Do not create outline rules around graphs. • Do not use horizontal lines through graphs. • Do not use large font sizes to label degrees of longitude and latitude on maps. • Indicate direction of degrees longitude and latitude on maps (e.g. 170°E). • Avoid placing labels on a vertical plane (except ony axis).' •Avoid odd (nonstandard) patterns to mark sections of bar graphs and pie charts. Copyright law Fishery Bulletin, a U,S. government publication, is not subject to copyright law. If an author wishes to reproduce any part of Fishery Bulletin in his or her work, he or she is obliged, however, to acknowledge the source of the extracted literature. Submission of papers Send four printed copies (one original plus three copies) — clipped, not stapled — to the Scientific Edi- tor, at the address shown below. Send photocopies of figures with initial submission of manuscript. Original figures will be requested later when the manuscript has been accepted for publication. Do not send your manuscript on diskette until requested to do so. Dr. John V. Merriner Scientific Editor, Fishery Bulletin Center for Coastal Fisheries and Habitat Research, NOS 101 Pivers Island Road Beaufort, NC 28516 Once the manuscript has been accepted for pub- lication, you will be asked to submit a software copy of your manuscript. The software copy should be submitted in WordPerfect or Word format (in Word, save as Rich Text Format). Please note that we do not accept ASCII text files. Reprints Copies of published articles and notes are avail- able free of charge to the senior author (50 copies) and to his or her laboratory (50 copies). Additional copies may be purchased in lots of 100 when the author receives page proofs. sff 1 1 S3 ?ISK U.S. Department of Commerce Volume 99 Number 4 October 2001 Fishery Bulletin U.S. Department of Commerce Donald L. Evans Secretary National Oceanic and Atmospheric Administration Scott B. Gudes Acting Under Secretary for Oceans and Atmosphere National Marine Fisheries Service William T. Hogarth Acting Assistant Administrator for Fisheries Scientific Editor Dr. John V. Merriner Editorial Assistant Sarah Shoffler Center for Coastal Fisheries and Habitat Research, NOS 101 Pivers Island Road Beaufort, NC 28516 The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fish- eries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115-0070. Periodicals postage is paid at Seattle, WA, and at additional mailing offices. POST- MASTER: Send address changes for sub- scriptions to Fishery Bulletin, Superin- tendent of Documents, Attn.: Chief, Mail List Branch, Mail Stop SSOM, Washing- ton, DC 20402-9373. Although the contents of this publica- tion have not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this peri- odical is necessary according to law for the transaction of public business 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: $45.00 domestic and $56.25 foreign. Cost per single issue: $28.00 domestic and $35.00 foreign. See back for order form. Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C 15700 Seattle, Washington 98115-0070 Editorial Committee Dr. Dr. Dr. Dr. Dr. Dr. Dr. Dr. Andrew E. Dizon Harlyn O. Halvorson Ronald W. Hardy Richard D. Methot Theodore W. Pietsch Joseph E. Powers Harald Rosenthal Fredric M. Serchuk National Marine Fisheries Service University of Massachusetts, Boston University of Idaho, Hagerman National Marine Fisheries Service University of Washington, Seattle National Marine Fisheries Service Universitat Kiel, Germany National Marine Fisheries Service Fishery Bulletin web site: fishbull.noaa.gov The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. It began as the Bulletin of the United States Fish Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. 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 99 Number 4 October 2001 Fishery Bulletin Contents The conclusions and opinions expressed in Fishery Bulletin are solely those of the authors and do not represent the official position of the National Marine Fisher- ies Service (NOAA) or any other agency or institution. The National Marine Fisheries Service (NMFS) does not approve, recommend, or endorse any proprietary product or pro- prietary material mentioned in this pub- lication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales pro- motion which would indicate or imply that NMFS approves, recommends, or endorses any proprietary product or pro- prietary 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. Articles 519-527 Arendt, Michael D., Jon A. Lucy, and Thomas A. Munroe Seasonal occurrence and site-utilization patterns of adult tautog, Tautoga onitis (Labridae), at manmade and natural structures in lower Chesapeake Bay 528-544 Gharrett, Anthony J., Andrew K. Gray, and Vladimir Brykov Phylogeographic analysis of mitochondrial BNA variation in Alaskan coho salmon, Oncorhynchus kisutch 545-553 Hurst, Thomas P., and David O. Conover Diet and consumption rates of overwintering YOY striped bass, Morone saxatilis, in the Hudson River 554-571 Lo, Nancy C. H., John R. Hunter, and Richard Charter Use of a continuous egg sampler for ichthyoplankton surveys: application to the estimation of daily egg production of Pacific sardine (Sard inops sagax ) off California 572-587 Loher, Timothy, David A. Armstrong, and Bradley G. Stevens Growth of juvenile red king crab ( Paralithodes camtschaticus) in Bristol Bay (Alaska) elucidated from field sampling and analysis of trawl-survey data 588-600 MacNair, Leslie S., Michael L. Domeier, and Calvin S. Y. Chun Age, growth, and mortality of California halibut, Paralichthys californicus, along southern and central California 601 -616 Miyashita, Shigeru, Yoshifumi Sawada, Tokihiko Okada, Osamu Murata, and Hidemi Kumai Morphological development and growth of laboratory-reared larval and juvenile Thunnus thynnus (Pisces: Scombridae) Fishery Bulletin 99(4) 617-627 Patterson III, William F, James H. Cowan Jr, Charles A. Wilson, and Robert L. Shipp Age and growth of red snapper, Lutjcinus campechanus, from an artificial reef area off Alabama in the northern Gulf of Mexico 628-640 Sipe, Ann M., and Mark E. Chittenden Jr. A comparison of calcified structures for aging summer flounder, Paralichthys dentatus 641 -652 Somerton, David A., and Peter Munro Bridle efficiency of a survey trawl for flatfish 653-664 Wilson, Charles, A., and David L. Nieland Age and growth of red snapper, Lutjanus campechanus, from the northern Gulf of Mexico off Louisiana 665-670 Notes Arendt, Michael D., John E. Olney, and Jon A. Lucy Stomach content analysis of cobia, Rachycentron canadum, from lower Chesapeake Bay 671-678 Blandon, Ivonne R., Rocky Ward, Tim L. King, William J. Karel, and James P. Monaghan Jr. Preliminary genetic population structure of southern flounder, Paralichthys lethostigma, along the Atlantic Coast and Gulf of Mexico 679-684 Hernandez-Lopez, Jose L., Jose J. Castro-Hernandez, and Vicente Hernandez-Garcia Age determined from the daily deposition of concentric rings on common octopus (Octopus vulgaris ) beaks 685-690 Jones, Thomas S., and Karl 1. Ugland Reproduction of female spiny dogfish, Squalus acanthias, in the Oslofjord 691-696 Porter, Steven M., Annette L. Brown, and Kevin M. Bailey Estimating live standard length of net-caught walleye pollock ( Theragra chalcogramma) larvae using measurements in addition to standard length 697-701 Takagi, Motohiro, Tetsuro Okamura, Seinen Chow, and Nobuhiko Taniguchi Preliminary study of albacore ( Thunnus alalunga) stock differentiation inferred from microsatellite DNA analysis 702 Acknowledgment of reviewers 703 Index 711 Erratum 712 Subscription form 519 Abstract — Ultrasonic transmitters were surgically implanted into adult tautog ( n -27 , 400-5 14 mm TL ) to document sea- sonal occurrence and site utilization at four sites situated within known tautog habitat near Cape Charles, Virginia, in lower Chesapeake Bay. Tagged tautog were released at the same sites where originally caught within 2 h of capture. Sites were continuously monitored with automated acoustic receivers between 9 November 1998 and 13 October 1999. Two sites consisted of natural bedform materials and two sites consisted of manmade materials. Ninety-four per- cent of tautog (n= 15) released in fall 1998 remained inshore during winter at sustained water temperatures of 5-8°C, rather than moved offshore during winter as documented for tautog off New York, Rhode Island, and Mas- sachusetts. Ninety-one percent ( /z = 1 0 ) of tautog released in spring 1999 re- mained inshore during summer when water temperature was 27°C and in the absence of an important food item, blue mussels (Mytilus edulis). These find- ings conflict with assertions that tautog move to cooler water in summer when water temperatures reach 20°C. Tautog released at natural bedform sites were detected only at these sites throughout the study. Tautog released at manmade structures also displayed high site-utili- zation patterns, but several tautog peri- odically moved 2-10.2 km away from these sites over featureless bottom, a known deterrent to emigration for large temperate labrids in other waters. Ben- thic communities were similar at man- made sites and natural bedform sites, and movement away from manmade sites may have been influenced by hab- itat size as well as habitat structure. Understanding temporal and spatial utilization of habitats is an important first step to identifying essential fish habitat and to evaluating and protect- ing fishery resources within Chesa- peake Bay and elsewhere. Manuscript accepted 11 April 2001. Fish. Bull. 99:519-527 (2001). Seasonal occurrence and site-utilization patterns of adult tautog, Tautoga onitis (Labridae), at manmade and natural structures in lower Chesapeake Bay* Michael D. Arendt School of Marine Science College of William and Mary Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Present address: Marine Resources Division, Department of Natural Resources Marine Resources Research Institute 217 Fort Johnson Road Charleston, South Carolina 29422-2559 E-mail address: arendtm@mrd.dnr.state.sc. Jon A. Lucy Sea Grant Marine Advisory Program Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Thomas A. Munroe National Marine Fisheries Service National Systematics Laboratory National Museum of Natural History Washington, D C. 20560-0153 The labrid Tautoga onitis (tautog) is a highly prized game fish targeted by anglers fishing at natural and man- made structure (Briggs, 1977; Lucy and Barr, 1994). Tautog are distrib- uted between Georgia (Parker, 1990) and Nova Scotia (Bigelow and Schro- eder, 1953); peak abundance is found between Massachusetts and the Dela- ware Capes.* 1 Slow growth rate, late age at maturity, predictable distribution, and localized population structure sug- gest high vulnerability to overexploi- tation (Hostetter and Munroe, 1993). Extended residence at accessible fish- ing sites may increase the potential for overexploitation; thus, residence and site-utilization patterns of tautog throughout this species’ distribution range must be well understood for effec- tive management of this resource. Tag-recapture studies in New York, Rhode Island, and Massachusetts sug- gest that adult tautog spend spring and fall months inshore, may move offshore during the warmest summer months (Cooper, 1966; Briggs, 1969), and over- winter offshore (Cooper, 1966; Briggs, 1977). Tautog leave inshore waters at varying rates between July and Oc- tober (Cooper, 1966) and are recap- tured in coastal waters in fall (Cooper, 1966; Briggs, 1977), consistent with in- direct observations on seasonal abun- dance (Stolgitis, 1970; Olla et ah, 1974). In contrast, tag-recapture studies re- port limited evidence of a seasonal in- shore-offshore migration for tautog in the Chesapeake Bay and coastal Virgin- ia waters.2 Seasonal abundance data also suggest that tautog remain inshore in Chesapeake Bay ( Hostetter and Mun- roe, 1993) and in Delaware Bay (Eklund and Targett, 1991) during winter. * Contribution 2391 of the Virginia Institute of Marine Science, Gloucester Point, Vir- ginia 23062. 1 Atlantic States Marine Fisheries Commis- sion (ASMFC). 1996. Fishery manage- ment plan for tautog, rep. 25, 56 p. [Avail- able from ASMFC, 1444 Eye Street NW, Washington, DC 20005.1 2 Virginia Game Fish Tagging Program. 1995-1999. Marine Resources Commis- sion, 968 Oriole Dr. South, Suite 102, Vir- ginia Beach, VA 23451. 520 Fishery Bulletin 99(4) The tag-recapture method is not a suitable method for evaluating site-utilization patterns because this technique does not provide information on the loca- tion of tagged animals between times of release and recapture. Furthermore, the tag-recapture method requires that tagged animals be recaptured, and be reported as recaptured, before any information is available. Ultrasonic telemetry, however, enables continuous observations on all tagged animals, in their natural environment, without requiring that tagged animals be recaptured (Winter, 1996). Pre- viously, only in one other study (Olla et al., 1974) was ultrasonic telemetry used to monitor adult tau- tog. Olla et al. (1974) tagged and “tracked” 10 adult tautog ultrasonically in Great South Bay, NY, for up to 80 h after their release. Although an important study, their sample size was small and total observa- tions too limited (<400 h, single season) to document seasonal occurrence and site-utilization patterns for this species. Ultrasonic telemetry was selected to address sea- sonal occurrence and site-utilization patterns of adult tautog in lower Chesapeake Bay, given that tag-recapture methods can be applied only within limitations and given that the poor visibility and strong currents preclude direct underwater observa- tions of this species in this turbid estuary. Rather than collect detailed positional data over short pe- riods of time (days) for a few tautog, we chose to collect seasonal occurrence and site-utilization data for two large groups of tautog (n=16 and 11) at four specific sites located within known tautog habitat. Sites were monitored by using a fixed, submerged hydrophone array between November 1998 and Sep- tember 1999. The first objective of our study was to determine if tautog remained inshore at natural and manmade structures in lower Chesapeake Bay dur- ing winter and summer. The second objective was to docu- ment and describe site-utilization patterns within inshore study sites. Data for daily activity patterns are presented elsewhere (Arendt et al., in press). Materials and methods Tautog were caught, tagged, and released at four sites sit- uated within a 1.5 km x 6 km area near Cape Charles, Virginia (Fig. 1). Side-scan sonar (Sea Scan Technology, Ltd., White Marsh, VA) was used to measure dimensions of the four study sites and to map the surrounding sea- floor. The Texeco Wreck, a 30 m x 100 m shipwreck, was located in 18 m of water west of the Susquehanna Channel (30-40 m deep) in an area characterized by flat, relatively featureless bottom topography (Wright et al., 1987). The three remaining sites (Airplane Wreck, Coral Lump, and Ridged Bottom) were located in 8-15 m of water east of the Susquehanna Channel in an area characterized by sand flats and deep, mud-bottomed channels (Wright et al., 1987). The Airplane Wreck (40 m x 20 m) consisted of concrete rubble. The Ridged Figure 1 Location of study sites for telemetric study of tautog released in lower Chesapeake Bay near Cape Charles, VA. Texeco Wreck (TX) is located in 18 m of water on a plain west of Susquehanna Channel (30-40 m deep). Coral Lump (CL), Ridged Bottom (RB), and Airplane Wreck (AW) are located in 8-15 m of water on a flat east of the Susquehanna Channel. Bottom (30 m x 100 m) and Coral Lump (100 m x 300 m) sites consisted of natural bedforms. Otter trawl, oyster dredge, and underwater video surveys indicated that all sites were densely populated by several species of sponges, colonial bryozoans, mollusks, and crustaceans. Tautog were caught with standard two-hook bottom rigs baited with pieces of blue crab or clam and were brought aboard with a nylon landing net. Tautog were observed in an aerated live well up to 2 h before transmitters were implanted. Total length (mm) and sex (White, 1996) were recorded. Only tautog >400 mm TL were tagged ultrasoni- cally. This minimum size increased the odds of transmit- ters weighing less than 1.25% of fish body weight in wa- ter (Winter, 1996), based on size-weight relationships for tautog in Virginia (Hostetter and Munroe, 1993; White, 1996). Tautog >400 mm were also reproductively mature (Hostetter and Munroe, 1993; White, 1996). Surgical procedures were similar to those used in Nemetz and Macmillan (1988), Mortensen (1990), Holland et al. (1993), Szedlmayer (1997), and Thoreau and Baras (1997). In preparation for surgical implantation of trans- mitters, level-four anesthesia (Mattson and Ripple, 1989; Prince et al., 1995) was induced by immersing tautog in a 325-mg/L solution of MS-222. Once anesthetized, a small Arendt et al.: Seasonal occurrence of site-utilization patterns of Tautoga onitis 521 (25-mm) incision was made on each fish immediately dor- sal to the ventral midline between pelvic fins and anus with a sterilized, disposable razor blade. The transmit- ters was coated in sterile mineral oil and placed into the visceral cavities of tautog with the transducer end of the transmitter facing forward. Braided, polyglycolic acid sutures (Dexon®, I-III), surgical staples (Proximate Plus MD 35W®), and acrylic adhesive (Krazy glue®) were used to close the incision. Betadine was used periodically throughout the surgical procedure and antibiotics (Nu- Flor®) were injected intramuscularly to increase postsur- gical survival (Schramm and Black, 1984; Bart and Dun- ham, 1990; Poppe et al., 1996). Tautog were revived in the aerated live well and released within 0.5 h after surgery. Preliminary evaluation of surgical procedures with “dum- my” transmitters indicated 100% transmitter retention, 86% survival, and normal swimming, feeding, reproduc- tive behavior, and physiology for tautog >400 mm TL held up to 418 days in captivity (Arendt, 1999). Tautog were fully recovered from surgery <1 to 6 days after release (Arendt, 1999) according to detection patterns recorded by automated acoustic receivers (Arendt and Lucy, 2000). V-16-1H-R256 coded transmitters ( 16 mm x 48 mm, 9 g in water; Vemco, Ltd. , Shad Bay, Nova Scotia, Canada ) were used in our study. Signal repeat intervals for coded trans- mitters (69 kHz) varied randomly between 45 and 75 s, which extended battery life to 111 d. Transmitters were primarily detected with a submerged array of automated acoustic receivers (VR1, Vemco, Ltd.); however, transmit- ters were also detected from a research vessel with acous- tic hydrophones (V10, VH65, Vemco, Ltd.) and an electron- ic receiver (VR60, Vemco, Ltd.). Omnidirectional VR1 receivers were deployed 100-150 m to the west and east of the perimeter of each of the four sites. Detection radius for each receiver was approximately 400 m. Detection areas for both receivers were overlapped to create three distinct reception zones: a central reception zone common to both receivers and two peripheral recep- tion zones unique to either receiver. VR1 receivers were moored 1. 5-3.0 m above the seafloor to provide an unob- structed line-of-sight for transmitter signal reception (i.e. positioned above the “structure” associated with each site) and to reduce acoustic interference from suspended mate- rial associated with strong bottom currents. Mooring units consisted of a railroad wheel (227 kg), stainless steel air- craft cable (0.64 cm, 7 x 19 strand), and subsurface and surface floats. Receivers were retrieved every three to six weeks and detection data (transmitter ID, date and time of detection) were downloaded directly to a shipboard com- puter by means of a VR1-PC cable interface (Vemco, Ltd.). Data for tautog released in fall 1998 were collected for the duration of the transmitter battery life. Data for tau- tog released in spring 1999 were collected until all VR1 receivers were permanently removed from each site. VR1 detections for each tautog were sorted into hourly bins and examined graphically. Because tautog are diurnally active and nocturnally quiescent (Olla et al., 1974; Arendt et al., in press), only daytime detection was used to determine site-utilization patterns. Tautog were considered resident at a particular site each day if they were detected at the same site throughout the day (morning, mid-day, evening). Total fish-days (sum of all days between date of first and last detection for all fish) for each calendar season were classified as 1) days when tautog were resident at the site of initial release, 2) days when tautog were detected at an alternative site, or 3) days when they were not detected at all. Site-utilization patterns of tautog were examined for each site in fall (from 9 Nov 98 to 20 Dec 98), winter (from 21 Dec 98 to 20 Mar 99), spring (from 21 Mar 99 to 20 Jun 99), and summer (from 21 Jun 99 to 9 Sep 99). To increase the probability that tautog would be re- ported as recaptured should recapture occur, transmitters were labeled with the specimen’s ID number, a $50 “re- ward” notice, and a phone number. Tautog were tagged externally with a small, orange t-bar anchor tag (TBA2; Hallprint, Holden Hill, South Australia) used by the Vir- ginia Game Fish Tagging Program and with a larger, green t-bar tag (SHD-95; Floy Mfg., Seattle, WA) containing the specimen’s ID number, a $50 reward” notice, and a phone number. Internal and external reward notices were includ- ed because Szedlymayer ( 1997) had observed that internal reward notices persisted longer than external reward no- tices for red snapper ( Lutjanus eampeehanus ) in the Gulf of Mexcio, and that internal reward notices were noticed accidentally. In addition to distinct marking of each tau- tog, colorful reward posters were posted at over 40 mari- nas, boat ramps, and tackle shops in lower Chesapeake Bay and literature describing the project was mailed to over 5000 homes and businesses. Results Twenty-seven adult tautog (400-514 mm TL) were tagged with ultrasonic transmitters and released (16 in fall 1998, 11 in spring 1999) near Cape Charles, VA (Table 1). Four tautog were released at each of the four sites in fall 1998 and at the Coral Lump and Texeco Wreck in spring 1999. Two tautog were released at the Ridged Bottom and one tautog was released at the Airplane Wreck in spring 1999. Similar numbers of tautog were tagged and released at manmade (77 = 13) and natural bedform (77 = 14) sites. Eighty-one percent (77=22) of all tautog released were males; 19% were females. Thirteen percent (/?=2) of tau- tog released in fall 1998 were female; 27% (77=3) of all tau- tog released in spring 1999 were female. Both female tau- tog released in fall 1998 (ID19, ID28) were released at the Texeco Wreck. One female tautog was released at the Tex- eco Wreck (ID37) and two female tautog were released at the Coral Lump (ID39, ID40) in spring 1999. Ninety-four percent (77=15 of 16) of tautog released in fall 1998 remained inshore within lower Chesapeake Bay during winter. Ninety-one percent <77=10 of 11) of tautog released in spring 1999 remained inshore within the Bay during summer. All tautog (77=14) released at natural sites remained in- shore and were detected only at their respective release sites. Tautog released at the Ridged Bottom and Coral Lump sites were detected 99% of fish-days in fall, 71-91% of fish -days in winter, 64-100% of days in spring, and 522 Fishery Bulletin 99(4) Table 1 Summary of fish-days for data from 27 adult tautog (400-514 mm TL) tagged and released with ultrasonic transmitters at four sites in lower Chesapeake Bay near Cape Charles, Virginia. Recaptured tautog are noted with an asterisk (*). Abbreviations for sites: CL = Coral Lump; TX = Texeco Wreck; RB = Ridged Bottom; AW = Airplane Wreck. Fish days ID Sex TL Site Date released At site Not detected At alternate site Total 1 M 432 CL 9 Nov 1998 174 8 0 182 18 M 406 CL 9 Nov 1998 155 19 0 174 19 F 495 TX 10 Nov 1998 58 105 2 165 20* M 470 TX 10 Nov 1998 0 167 1 168 21 M 406 RB 10 Nov 1998 88 10 0 98 22 M 400 RB 10 Nov 1998 178 0 0 178 23 M 483 AW 13 Nov 1998 152 13 0 165 24 M 432 AW 13 Nov 1998 112 45 0 157 25 M 432 CL 3 Dec 1998 183 3 0 186 26 M 400 CL 3 Dec 1998 177 4 0 181 27 M 514 TX 4 Dec 1998 177 0 0 177 28 F 413 TX 4 Dec 1998 42 130 13 185 29* M 400 AW 7 Dec 1998 147 14 1 162 30 M 419 AW 7 Dec 1998 46 21 0 67 31 M 445 RB 8 Dec 1998 144 24 0 168 32 M 419 RB 8 Dec 1998 88 39 0 127 33 M 406 TX 21 Apr 1999 0 136 6 141 34* M 432 CL 28 May 1999 69 0 0 69 35 M 445 TX 28 May 1999 97 7 0 104 36 M — TX 28 May 1999 103 1 0 104 37* F 445 TX 28 May 1999 99 5 0 104 CO oo M 483 CL 7 Jun 1999 59 0 0 59 39* F 483 C,L 7 Jun 1999 59 0 0 59 40* F 432 CL 7 Jun 1999 59 0 0 59 41 M 445 AW 7 Jun 1999 10 2 2 14 42* M 406 RB 9 Jun 1999 58 0 0 58 43 M 406 RB 9 Jun 1999 56 2 0 58 98-100% of fish-days in summer (Fig. 2). Reduced detec- tion in winter and spring was partially attributed to two tautog (ID21, ID32) that were detected 46 d and 75 d less than the mean (175 d) for other tautog (72=12) released at the same time. To provide a more conservative estimate of site utilization, these tautog were listed as “not detected” at these sites for these days. Eighty-five percent (/2 = 11) of tautog released at man- made structures remained inshore and all but three of these were detected only at their respective release sites. Site utilization by tautog at the Texeco Wreck was low (34-71% of fish-days) in all seasons (Fig. 2). One tautog (ID20) released at the Texeco Wreck was only detected at the Texeco Wreck for three hours after being tagged and released in fall 1998. Two additional tautog (ID 19, ID28) spent 64-70% of fish-days away from the Texeco Wreck in fall, winter, and spring, and a fourth tautog (ID33) spent 100% of fish-days away from the Texeco Wreck in spring and summer. One of these tautog (ID28) was detected at the Coral Lump for 6 days in January 1999 (VR1 receivers) and all three tautog were regularly detected with the VR60 receiver at a site 2 km south of the Texeco Wreck. Tautog released at the Airplane Wreck were resident 99% of fish- days in fall, 66% of fish-days in winter, and 58% of fish-days in spring (Fig. 2). Reduced detection in winter resulted partially from one tautog (ID30) being detected 108 days less than the mean (175 days) for other tautog 0z = 12) re- leased at the same time. To provide a more conservative estimate of site utilization, this tautog was listed as “not detected” at the Airplane Wreck for these days. No sum- mer data were available for tautog at the Airplane Wreck because the single tautog (ID41) released there in spring 1999 was detected at this site for two days following re- lease, was later detected at the Texeco Wreck, and then was never detected again at any site. Tautog 41 was listed as “not detected” at the Airplane Wreck from the time of Arendt et al.: Seasonal occurrence of site-utilization patterns of Tautogo onitis 523 Airplane Wreck Texeco Wreck Ridged Bottom Coral Lump Manmade structures Natural bedforms B fall □ winter ■ spring □ summer Figure 2 Seasonal site-utilization patterns of tautog released at manmade (Airplane Wreck and Texeco Wreck) and natural bedform (Ridged Bottom and Coral Lump) sites near Cape Charles, Virginia, in fall (9 Nov 98 to 20 Dec 98), winter (21 Dec 98 to 20 Mar 99), spring (21 Mar 99 to 20 Jun 99), and summer (21 Jun 99 to 9 Sep 99). Site-utilization patterns were greatest for tautog released at natural bedform sites. leaving this site until VR1 receivers were removed from this site, 112 d later. Eight tautog (27%) released in our study were subse- quently recaptured (Table 1). Two tautog (ID20, ID29) released in fall 1998 were recaptured (13%) away from release sites by commercial fishermen in spring 1999. Tau- tog 20 was released at the Texeco Wreck on 10 November 1998 and either left this site the same day or its transmit- ter failed to transmit data. This tautog was subsequently recaptured 10.2 km northeast of the Texeco Wreck in a crab pot on 27 April 1999, 169 days later. Tautog 29 was first caught at the Airplane Wreck on 13 November 1998 and held in a wire cage with several other tautog at the Airplane Wreck for five days as part of a catch-release mortality study on tautog.3 After being released at the Air- plane Wreck on 18 November 1998, this tautog was re- captured at the Airplane Wreck on 7 December 1998 (19 d later) and ultrasonically tagged and released. Tautog 29 remained at the Airplane Wreck until 12 May 1999, then was recaptured in a gill net 2 km east of the Air- 3 Lucy, J. A., and M. D. Arendt. 1999. Exploratory field evalua- tion of hook-release mortality in tautog ( Ta u toga onitis) in Lower Chesapeake Bay, Virginia. Rep. VMRC-99-10, 11 p. [Available from Marine Advisory Program, Virginia Institute of Marine Science, Gloucester Point, VA 23062.] plane Wreck on 19 May 1999. Recreational fishermen re- captured six tautog released in spring 1999 (55%) at the same sites where these fish were released 114—211 d earli- er and 8-13 weeks after VR1 receivers were removed from sites. Discussion Tautog remained inshore within lower Chesapeake Bay during winter, at sustained water temperatures of 5-8°C (Arendt et al., in press). Although detected at these water temperatures, tautog overall were detected less than during ot her times of the year, most likely because tautog remained inactive and within structures for several days at a time (Arendt et al., in press). Inshore occurrence of tautog in winter has been observed in eastern Long Island Sound,4 Delaware Bay (Eklund andTargett, 1991), and lower Chesa- 4 Auster, P. J. 1989. Species profiles: life histories and environ- mental requirements of coastal fishes and invertebrates (North and Mid-Atlantic) — tautog and cunner. U.S. Fish and Wildlife Service Biological Report 82 (11.105). U.S. Army Corps of Engi- neers Rep. TR EL-82-4, 13 p. NOAA’s National Undersea Research Program, Univ. Connecticut at Avery Point, Groton, CT 06430. 524 Fishery Bulletin 99(4) peake Bay (Hostetter and Munroe, 1993). When water tem- peratures remains above 9-10°C,5 a viable winter inshore fishery exists for tautog within lower Chesapeake Bay. Within the winter fishery, inshore catches occur predomi- nantly in December and March, whereas offshore catches occur predominantly in January and February. Occurrence of an inshore winter fishery for tautog in Virginia is unique within this species’ geographic distribution. Tautog remained inshore during the summer at a maxi- mum sustained water temperature of 27°C (Arendt et ah, in press). Summer residence data have been supported by direct underwater observations.6 Infrequent recreational catches of tautog in lower Chesapeake Bay during sum- mer have also been reported.2 Inshore, summer residence of tautog has been documented for Great South Bay, NY, when water temperatures were 19-24°C (Olla et al., 1974) and Narragansett Bay, RI, at maximum sustained water temperatures of 22°C.7 These findings contradict reports from Virginia (Adams, 1993), New York (Briggs, 1969), and Rhode Island (Cooper, 1966) that adult tautog may move offshore to cooler water during summer. Tautog remained inshore during summer in the absence of blue mussels ( Mytilus edulis), a primary food item of tautog in northern areas (Bigelow and Schroeder, 1953; Ol- la et ah, 1974). In June 1998-99, using underwater video, otter trawl, and oyster dredge, we documented large clus- ters of live blue mussels at study sites and noted growth of mussels on VR1 mooring units. By July 1998-99, blue mussels were not present at any of these sites. Absence of mussels in July in both years was most likely due to le- thal effects of water temperatures >27°C (Wells and Gray, 1960). Because blue mussels are not present in lower Chesa- peake Bay year round, the diet of tautog inhabiting lower Chesapeake Bay throughout the year may be more diverse than that of tautog in northern areas. Stomach contents from an ultrasonically tagged tautog recaptured in Octo- ber 1999 at the Ridged Bottom site consisted primarily of the bryozoan Alcyinidium verilli. At an artificial fishing reef near Cape Charles, VA, tautog consumed a variety of crustaceans, shellfish, bryozoans, and hydroids.8 Similar temporal distributions of blue mussels have been reported 5 White, G. G., J. E. Kirkley, and J. A. Lucy. 1997. Quantitative assessment of fishing mortality for tautog ( Tautoga onitis ) in Virginia. Preliminary report, 54 p. Department of Fisheries Science and Marine Advisory Program, Virginia Institute of Marine Science, College of William and Mary, Gloucester Point, VA 23062. 6 Hager, C. 1999 (July). Personal communication of direct un- derwater observation of tautog at Plantation Light (3-8 m depth), 2 km southeast of Texeco Wreck study site. School of Marine Science, College of William and Mary, Virginia Institute of Marine Science, Gloucester Point, VA 23062. 7 Castro, K. 1999 (October). Personal communication of water temperature observation for Narragansett Bay. East Farm- Fisheries Center, Univ. Rhode Island, Kingston, RI 02881. 8 Feigenbaum, D., C. Blair, and A. J. Provenzano. 1985. Artifi- cial reef study — year II report. Virginia Marine Resources Com- mission rep. VMRC-83-1 185-616, 57 p. VA Institute of Marine Science, Gloucester Point, VA 23062. in Delaware Bay, where diets of tautog subsequently shift towards alternative, less nutritious food items at times when blue mussels are unavailable.9 Year-round occurrence of adult tautog in Chesapeake Bay differs from seasonal (spring and fall) inshore occur- rence of adult tautog in Great South Bay, NY (Olla et al., 1974; Briggs, 1977), Narragansett Bay, RI (Cooper, 1966), and the Weweantic River Estuary, MA (Stolgitis, 1970). Year-round occurrence of ultrasonically tagged tautog was consistent with large-scale patterns of occurrence of tau- tog from conventionally tagged tautog (127-584 mm TL) in the Virginia Game Fish Tagging Program between 1995 and 1999. 2 5 Of 563 recaptured tautog that were originally tagged in lower Chesapeake Bay, excluding Cape Charles, and adjacent coastal waters, 85% (n=476) were recaptured at the same sites where released 0-1214 d earlier, includ- ing 20 multiple recaptures of individuals at the same sites where originally released. High incidence of recaptures of tautog at the same sites where they were released oc- curred during all seasons. Only 5% of all recapture events involved movement (8-97 km) of tautog between Chesa- peake Bay and adjacent coastal waters. Daily detection patterns were almost always similar for both VR1 receivers at sites, indicating that tautog re- mained within the central signal reception area of both VR1 receivers and in the general vicinity of sites through- out the entire day. Tautog were generally not detected by VR1 receivers at night; however, tautog likely remained at sites throughout the night (Arendt et al., in press). Tau- tog were likely detected less often (or not at all) at night because of nocturnal quiescence in or near structure (Olla et al., 1974) and therefore were effectively out of range of VR1 receivers because of the presence of an acoustic barrier (Matthews, 1992; Pearcy, 1992; Bradbury et al., 1995,1997; Zeller, 1997). Featureless bottom topography, a known deterrent to emigration for large, temperate labrids ( Notolabrus tetri- cus, N. fucicola, Pictilabrus laticlavius, Pseudolabrus psit- taculus) in Tasmania (Barrett, 1995), did not act as a de- terrent to emigration for tautog in lower Chesapeake Bay. Two tautog (ID20, ID28) released at the Texeco Wreck in fall 1998 traversed a wide (2-km), deep (37-40 m) mud- bottomed channel (Wright et al., 1987) on at least three occasions, and one of these tautog (ID20) was subsequent- ly recaptured. Two tautog (ID19, ID28) traveled between the Texeco Wreck and a site 2 km south of the Texeco Wreck on at least 12 occasions. A third tautog (ID33) left the Texeco Wreck almost immediately after being released and was subsequently detected (VR60 receiver) at this site throughout the remainder of our study. Movement by tautog was assumed to represent actual movement by tagged tautog as opposed to movement of a 9 Steimle, F., K. Foster, W. Muir, and B. Conlin. 1999. The diet of tautog collected on an artificial reef in Delaware Bay and interannual effects of prey availability (and notes on other tautog diet studies in the middle Atlantic Bight). First Biennial Conference on the biology of tautog and cunner, Mystic, CT, 30 November-1 December 1999. National Marine Fisheries Ser- vice, Highlands, NJ 07732. Arendt et al.: Seasonal occurrence of site-utilization patterns of Tautoga onitis 525 predator that may have preyed upon tagged tautog, even though these three tautog were never recaptured nor vi- sually observed after release. Adult tautog are very in- frequently preyed upon by sharks in Virginia (Gelsleich- ter et al., 1999); however, sharks are not likely present in Chesapeake Bay when water temperatures are 5-8°C. At these water temperatures, large striped bass (Morone saxatilis ) pose the only possible predatory threat to tau- tog. A recent study on feeding habits of adult striped bass in Chesapeake Bay found no tautog in the stomachs of more than 2000 striped bass, many of which were collected from the Chesapeake Bay Bridge-Tunnel complex, a well- known fishing area for adult tautog (Walters, 1999). All tautog detected or recaptured away from release sites were released at manmade sites, which also hap- pened to be the smallest sites. No information was avail- able regarding the origin of these manmade sites, but both have been in place for at least 20 years.1011 Stone et al. (1979) concluded that artificial reefs reach a stable state after five years. Benthic macrofauna collected at manmade sites during our study were similar to benthic macrofau- na collected at natural sites, suggesting that food may have been similar between manmade and natural sites. Given these observations, habitat size may be an impor- tant factor for adult tautog in determining the scale of lo- cal movements between adjacent habitats. Understanding the relationship between habitat size and site utilization warrants further investigation, especially with recent in- creased interest in the construction of artificial habitats for purposes of stock enhancement and enhanced fishing opportunities for tautog. Sex ratio of female to male tautog in our study was heavily skewed (1:3.5) towards male tautog due to oppor- tunistic sampling, and the preponderance of male tautog likely contributed to the high levels of site utilization ob- served in our study. In laboratory settings, adult male tau- tog aggressively defend territories throughout most of the year (Olla et al., 1978, 1980) and only during the spawn- ing season are female tautog permitted to enter territo- ries (Olla and Samet, 1977; Olla et al., 1981). Overall ac- tivity, including male agonistic behavior, also decreases as water temperatures approach annual minimum and maxi- mum values (Olla et al., 1978, 1980). Although sample size in our study was too small to distinguish site-utilization patterns by sex, it is worth noting that both tautog that left the Texeco Wreck in Nov-Dec and that periodically re- turned to this site throughout the winter and spring were females. In contrast, during the spring spawning season, three females (one at Texeco Wreck, two at Coral Lump) remained at release sites throughout the spring-summer monitoring period and all three were subsequently recap- tured by recreational fishermen at these same sites in the fall. More sex-specific data are needed to fully comprehend 10 Verry, S. 1998. Automated wreck and obstruction informa- tion system, special area report (37°00N-37°30N; 076°00W- 07 6°30W). NOAA/NOS, Hydrographic Services Branch/N/CS3 1, Silver Spring, MD 20910. 11 Jenrette, J. 1998. Personal commun. Captain, FV Bucca- neer, P.O. Box 149, Route 1108, Cape Charles, VA 23310. what role reproductive biology and social structure have on seasonal site-utilization patterns. Site-utilization patterns exhibited by ultrasonically tagged tautog were consistent with patterns reported for tautog released at these same sites from the Virginia Game Fish Tagging Program (VGFTP). Between April 1998 and October 1999, 40 tautog, tagged and released at these sites, were recaptured, including one tautog recap- tured twice at the same site. Six of eight (75%) tautog orig- inally released at the Texeco and Airplane Wrecks were re- captured away from these sites. Of these six tautog, three moved to the Coral Lump and Ridged Bottom and three moved to sites located 26.9 to 43.2 km away in lower Ches- apeake Bay. In contrast, 32 tautog tagged and released at the Coral Lump and Ridged Bottom sites were recaptured, all but two (which moved from the Ridged Bottom to the Coral Lump) were recaptured where released. One addi- tional fish moved to the Coral Lump from an artificial reef located 4 km to the northeast and within 2 km of where both tautog were recaptured by commercial fishermen in spring 1999. Ultrasonically and conventionally tagged tautog re- leased near Cape Charles, VA, in lower Chesapeake Bay demonstrated high site utilization at and high site affin- ity (returned to release sites after short emigration) for release sites. Extended residence by tautog at familiar sites during annual environmental extremes is considered more beneficial than emigration to more optimal environ- mental conditions because residence at familiar sites re- duces the risk of not finding suitable shelter, food, mates, or of encountering predators (Olla et al., 1978). Although directed seasonal offshore movements were not observed in our study, movements between adjacent inshore loca- tions occurred several times, including movement to adja- cent locations during the periods of seasonal thermal ex- tremes. Understanding temporal and spatial utilization of habitats is an important first step to identifying essential fish habitat and critical to evaluating and protecting fish- ery resources within Chesapeake Bay and elsewhere. Acknowledgments This research represents part of M. D. Arendt’s Master’s thesis; earlier drafts of this manuscript were reviewed by J. Hoenig, J. Musick, D. Evans, and W. DuPaul. J. Jenrette of the FV Bucaneer , C. Machen of the RV Langley, G. Pon- gonis, and J. Olney Jr. and VIMS divers (B. Gammisch, W. Reisner, T. Chisholm, and W. Stockhausen) provided valu- able field assistance. We thank Vemco, Ltd., and Physical Sciences Department, VIMS, for assistance with equip- ment operation and deployment. We also thank J. Jen- rette and S. Speirs (for collectively tagging and releasing 300+ tautog at study sites and recapturing 5 of 8 tautog tagged ultrasonically), volunteer anglers, and the com- mercial fishing industry for tagging, releasing, and report- ing recaptured tautog. This project was funded by the Recreational Fishing Development Fund of the Virginia Marine Resources Commission and by General Funds from VIMS. 526 Fishery Bulletin 99(4) Literature cited Adams, A. J. 1993. Dynamics of fish assemblages associated with an offshore artificial reef in the southern Mid-Atlantic Bight. M.A. thesis. School of Marine Science, College of William and Mary, Gloucester Point, VA, 97 p. Arendt, M. D. 1999. Seasonal residence, movement, and activity of adult tautog ( Tautoga onitis ) in lower Chesapeake Bay. M.S. thesis, School of Marine Science, College of William and Mary, Gloucester Point, VA, 104 p. Arendt, M. D., and J. A. Lucy. 2000. Recovery period and survival of ultrasonically tagged adult tautog in the lower Chesapeake Bay using auto- mated receivers. In Biotelemetry 15: proceeding of the 15th international dymposium on biotelemetry, Juneau, Alaska (J. H. Eiler, D. J. Alcorn, and M. R. Neuman, eds.), p. 117-125. International Society on Biotelemetry, Wag- eningen. The Netherlands. Arendt, M. D., J. A. Lucy and D. A. Evans. In press. Diel and seasonal activity patterns of adult tautog, Tautoga onitis, in lower Chesapeake Bay, inferred from ultrasonic telemetry. Environ. Biol. Fishes Barrett, N. S. 1995. Short- and long-term movement patterns of six temperate reef fishes (Families Labridae and Monocanthi- dae). Mar. Fresh. Res. 46(5):853-860. Bart, A. N., and R. A. Dunham. 1990. Factors affecting survival of channel cat fish after sur- gical removal of testes. Prog. Fish. Cult. 52(4):241-246. Bigelow, H. B., and W. C. Schroeder. 1953. Bony fishes, class Osteichthyes. Cunner tribe or wrasses. Family Labridae, tautog Tautoga onitis (Linnaeus). In Fishes of the Gulf of Maine, p. 478—484. Fish. Bull. 53. Bradbury, C., J. M. Green, and M. Bruce-Lockhart. 1995. Home ranges of female cunner, Tautogalabrus adsper- sus (Labridae), as determined by ultrasonic telemetry. Can. J. Zool. 73(7): 1268-1279. 1997. Daily and seasonal activity patterns of female cunner, Tautogalabrus adspersus (Labridae), in Newfoundland. Fish. Bull. 95(4): 646-652. Briggs, P. T. 1969. The sport fisheries for tautog in the inshore waters of eastern Long Island. N.Y. Fish Game J. 16( 2 ):238 — 254. 1977. Status of tautog populations at artificial reefs in New York waters and effect of fishing. N.Y. Fish Game J. 24(2):154-167 Cooper, R. A. 1966. Migration and population estimation of the tautog, Tautoga onitis (Linnaeus), from Rhode Island. Trans. Am. Fish. Soc. 95(3):239-247. Eklund, A. M., and T. E. Targett. 1991. Seasonality of fish catch rates and species composi- tion from the hard bottom trap fishery in the middle Atlan- tic Bight (US East Coast). Fish. Res.l2(l):l-22. Gelsleichter, J., J. A. Musick and S. Nichols. 1999. Food habits of the smooth dogfish, Mustelus canis, dusky shark, Carcharhinus obscurus, Atlantic sharpnose shark, Rhizoprionodon terraenovae, and the sand tiger, Car- charias taurus, from the northwest Atlantic Ocean. Envi- ron. Biol. Fish. 54:205-217. Holland, K. N., J. D. Peterson, C. G. Lowe, and B. M. Wetherbee. 1993. Movements, distribution, and growth rates of the white goatfish Mulloides flavolineatus in a fisheries conser- vation zone. Bull. Mar. Sci. 52(3):982-992. Hostetter, E. B., and T. A. Munroe. 1993. Age, growth, and reproduction of tautog Tautoga onitis (Labridae: Perciformes) from coastal waters of Vir- ginia. Fish. Bull. 91(l):45-64. Lucy, J. A., and C. G. Barr. 1994. Experimental monitoring of Virginia’s artificial reefs using fishermen’s catch data. Bull. Mar. Sci., 55(2-31524— 537. Matthews, K. R. 1992. A telemetric study of the home ranges and homing routes of lingcod Ophidion elongatus on shallow rocky reefs off Vancouver Island, British Columbia. Fish. Bull. 90(4):784-790. Mattson, N. S., and T. H. Ripple. 1989. Metomidate, a better anesthetic for cod ( Gadus morh ua ) in comparison with benzo-caine, MS-222, chlorobu- tanol, and phenoxyethanol. Aquaculture 83(19891:89-94. Mortensen, D. G. 1990. Use of staple sutures to close surgical incisions for transmitter implants. Am. Fish. Soc. Symp. 7:380-383. Nemetz, T. G., and J. R. MacMillan. 1988. Wound healing of incisions closed with a cynaoacry- late adhesive. Trans. Am. Fish. Soc. 117(2 ): 190— 195. Olla, B. L., A. J. Bejda, and A. D. Martin. 1974. Daily activity, movements, feeding, and seasonal occurrence in the tautog, Tautoga onitis. Fish. Bull. 72(1): 27-35. Olla, B. L., and C. Samet. 1977. Courtship and spawning behavior of the tautog, Tautoga onitis (Pisces: Labridae), under laboratory condi- tions. Fish Bull. 75(31585-599. Olla, B. L., C. Samet, and A. L. Studholme. 1981. Correlates between number of mates, shelter avail- ability and reproductive behavior in the tautog Tautoga onitis. Mar. Biol. 62(41239-248. Olla, B.L., A. L. Studholme, A. J. Bejda, and C. Samet. 1978. Effect of temperature on activity and social behavior of the adult tautog Tautoga onitis under laboratory condi- tions. Mar. Biol. 45(41369-378. 1980. Role of temperature in triggering migratory behavior of the adult tautog Tautoga onitis under laboratory condi- tions. Mar. Biol. 59(1123-30. Parker, R. O., Jr. 1990. Tagging studies and diver observations of fish pop- ulations on live-bottom reefs of the U.S. southeastern coast. Bull. Mar. Sci. 46(31749-760. Pearcy, W. G. 1992. Movements of acoustically-tagged yellowtail rockfish Sebastes flavidus on Heceta Bank, Oregon. Fish. Bull. 90(41726-735. Poppe, T. T., T. A. Mo, and K. Langdal. 1996. Surgical implantation of radio tansmitters into the abdominal cavity of brook trout and brown trout. Norsk Veterinaertidsskr. 198(4124 1-244. Prince, A. M. J., S. E. Low, T. J. Lissimore, R. E. Diewert, and S. G. Hinch. 1995. Sodium bicarbonate and acetic acid: an effective ane- thestic for field use. N. Am. J. Fish. Manage. 15(11170- 172. Schramm, H. L., Jr., and D. J. Black. 1984. Anesthesia and surgical procedures for implanting radio transmitters into grass carp. Prog. Fish. Cult. 46(3): 185-190. Stolgitis, J. A. 1970. Some aspects of the biology of the tautog, Tautoga onitis (Linnaeus), from the Weweantic River Estuary, Arendt et al.: Seasonal occurrence of site-utilization patterns of Tautoga onitis 527 Massachussetts, 1966. M.S. thesis, Univ. Massachusetts, Amherst, MA, 48 p. Stone, R. B., H. L. Pratt, R. O. Parker Jr., and G. E. Davis. 1979. A comparison of fish populations on an artificial and nat- ural reef in the Florida Keys. Mar. Fish. Rev. 41(9):1— 11. Szedlmayer, S. 1997. Ultrasonic telemetry of red snapper, Lutjanus eampe- chanus, at artificial reef sites in the northeast Gulf of Mexico. Copeia 1997(4):846-850. Thoreau, X., and E. Baras. 1997. Evaluation of surgery procedures for implanting telemetry transmitters into the body cavity of tilapia Oreo- chromis aureus. Aquat. Living Resour. 10( 4 ):207— 2 1 1 . Walters, J. F„ III. 1999. Diet composition and feeding habits of large striped bass (Morone saxatilis) in Chesapeake Bay. M.S. thesis, School of Marine Science, College of William and Mary, Gloucester Point, VA, 124 p. Wells, H. W„ and I. E. Gray. 1960. The seasonal occurrences of Mytilus edulis on the Carolina coast as a result of transport around Cape Hat- teras. Biol. Bull. (Woods Hole) 119:550-559. White, G. G. 1996. Reproductive biology of tautog, Tautoga onitis, in the lower Chesapeake Bay and coastal waters of Virignia. M.A. thesis, School of Marine Science, College of William and Mary, Gloucester Point, VA, 100 p. Winter, J. 1996. Advances in underwater biotelemetry. In Fisheries techniques, 2nd ed. (B. R. Murphy and D. W. Willis, eds.), p. 555-590. Am. Fish. Soc., Bethesda, MD. Wright, L. D., D. B. Prior, C. H. Hobbs, R. J. Byrne, J. D. Boon, L. C. Schaffner and M. O. Green. 1987. Spatial variability of bottom types in the lower Chesa- peake Bay and adjoining estuaries and inner shelf. Estuar. Coast. Shelf Sci. 24(6):765-784. Zeller, D. C. 1997. Home ranges and activity patterns of coral trout Plec- tropomus leopardus (Serranidae). Mar. Ecol. Prog. Ser. 154(19971:65-77. 528 Abstract — Mitochondria] DNA(mtDNA) haplotypes of coho salmon ( Oncorhyn - chus kisutch) sampled from northern Pacific Ocean and Bering Sea drain- ages formed two monophyletic clades between which nucleotide divergences averaged 2.95 substitutions per 1000 nucleotides. These data were obtained from restriction endonuclease diges- tions of PCR products that included over 97% of the mtDNA genome and resolved 16 different haplotypes in 258 fish from 13 locations. Comparisons of haplotype compositions of populations indicated that the Bering Sea drain- ages and one Kodiak Island population clustered separately from nine other Gulf of Alaska populations, including one from Asia. Rates of gene flow among populations estimated from haplotype frequencies (assuming an equilibrium between gene flow and random drift) were low (about one female per genera- tion between drainages within regions) in relation to allozyme-based estimates of gene flow for other Pacific salmon spe- cies. Much of the haplotype frequency variation was within-region variation. Haplotypes from both clades occur in many extant populations, suggesting that gene flow, population movements, or recolonization followed divergence of refugial isolates. Nested clade analysis of the geographic distribution of mtDNA haplotypes indicated that coho salmon demographic history has been influ- enced by recent isolation by distance and that historic population fragmen- tation was preceded by range expan- sion. These observations are consistent with effects expected from Pleistocene glacial advances and retreats. Manuscript accepted 19 March 2001. Fish. Bull. 99:528-544 (2001). Phylogeographic analysis of mitochondrial DNA variation in Alaskan coho salmon, Oncorhynchus kisutch Anthony J. Gharrett Andrew K. Gray Fisheries Division, School of Fisheries and Ocean Sciences University of Alaska Fairbanks 1 1 120 Glacier Highway Juneau, Alaska 99801 E-mail address (for Anthony J. Gharrett): ffajg@uaf.edu Vladimir Brykov Institute of Marine Biology Russian Academy of Science, Far East Branch 690041 Vladivostok, Russia In drainages flowing into the Gulf of Alaska and Bering Sea, coho salmon ( Oncorhyjichus kisutch) are the least numerous and their population struc- ture the least understood of Pacific salmon species (Oncorhynchus spp.). Many populations spawn in late fall or winter in remote drainages that are difficult to access. Spawning popula- tions are often small and separated widely (Sandercock, 1991). In larger rivers spawning adults may return to small, often transient, headwater streams. After emergence, fry and juve- niles may move to rearing areas, usually much lower in the river (Sandercock, 1991), forming complex admixtures from spawning populations in large drainages. Studies of allozyme variation have provided insight into the population structure of Pacific salmon species (e.g. Zhivotovsky et al., 1994, and references therein). Patterns of genetic variability often provide evidence of relationships between populations resulting from coancestry or gene flow, and genetic divergence among populations may be used for stock identification (Shaklee et al., 1999). In coho salmon, the low level of allozyme variation resolves relatively little population genetic structure (Rei- senbichler and Phelps, 1987; Wehrhahn and Powell, 1987; Bartley et al., 1992; Pustavoit, 1995). This low level of al- lozyme variation is consistent with nu- merous spawning populations that have small effective sizes and low levels of gene flow, such as that of coho salmon. Analysis of DNA variation adds di- mensions of interpretation not possible with allozyme data (Avise et al., 1988; Avise, 1989). “Gene trees” for mitochon- drial DNA (mtDNA) haplotypes are es- pecially informative. The mitochondrial genome is transmitted (primarily) ma- ternally (Gyllensten et al., 1991), and mtDNA is haploid and clonally inherit- ed with no recombination. Consequently, mutations accumulate over time within a clonal mtDNA line or haplotype. Com- parison of different haplotypes provides a basis for reconstruction of matriarchal genealogies. There is also evidence that mtDNA sequences may diverge (evolve) faster than many nuclear sequences (Brown et al., 1982). The extent of nu- cleotide divergence provides a temporal basis for comparing haplotype lineages. The rates of divergence of mtDNA have been roughly estimated for a number of species pairs by comparing observed nu- cleotide sequence divergence with fossil records that document the emergence of those species (Brown et al., 1979; Shields and Wilson, 1987 and references there- in). Although applications of these mo- lecular “clocks” are questionable when extended to species for which the fossil record is poor or missing, deductions about relative (as opposed to absolute) divergence times can be made from the extent of nucleotide change within a spe- cies. Furthermore, the geographic distri- Gharrett et al.: Phylogeographic analysis of mitochondrial DNA variation in Oncorhynchus kisutch 529 170°W 160°W 150° W 140° W 130°W Figure 1 Sites from which coho salmon were sampled for mtDNA analysis. Squares denote samples used in both the preliminary and secondary analyses; circles denote collections used only for the secondary analysis. Locations are Hugh Smith River (1), Fish Creek, a Taku River tribu- tary (2), Berner’s River (3), Indian River (4), Ford Arm River (5), Crooked Creek (6), Little Susitna River (7), Buskin River (8), Karluk River (9), Eek River (10), Kanektok River (11), Delta Clearwater River, a Yukon River tributary (12), and Kamchatka River (13). bution of haplotypes and their genealogical relationships can provide information about the historic demography and gene flow of a species. Templeton and colleagues (e.g. Tem- pleton and Sing, 1993, Castelloe and Templeton, 1994; Tem- pleton, 1998) have developed methods to examine both the shape of the “gene tree” and the geographic distribution of haplotypes, which they term “nested clade analysis of geo- graphic distances.” The objectives of our study were to survey the geograph- ic distribution of mtDNA variation in Alaskan coho salm- on populations along the Gulf of Alaska and Bering Sea and to use that information and the mtDNA haplotype “gene tree” to deduce the nature of the historic demo- graphic processes that influenced the contemporary geo- graphic distribution of coho salmon. Materials and methods Coho salmon were sampled from 12 drainages in Alaska and one in Asia (Fig. 1). Samples of heart tissue from each specimen were preserved in 95% ethanol or a solution of 20% dimethyl sulfoxide (DMSO) and 0.25M ethylene- diaminetetraacetic acid (EDTA) at pH 8, saturated with NaCl (Seutin et al., 1991). Total genomic DNA was isolated by phenol-chloroform extraction (Wallace, 1987) or with Puregene DNA™ iso- lation kits (Gentra Systems Inc., Minneapolis, MN). Se- quences were PCR-amplified using primers that targeted seven regions of the mtDNA genome in pieces that range from about 2115 to 2689 base pairs (bp) (Fig. 2, Table 1). The regions were designated ND3/ND4 (including genes for the NADH dehydrogenase-3 subunit and NADH de- hydrogenase-4L and -4 subunit genes), ND5/ND6 (includ- ing genes for the NADH dehydrogenase-5 and -6 sub- units), Cytb/D-loop (including the cytochrome b gene and the control region), 12S/16S (including 12S rRNA gene and most of the 16S rRNA gene), ND1/ND2 (including the NADH dehydrogenase- 1 and NADH dehydrogenase-2 subunit genes), COI/COII (including most of the cyto- chrome oxidase I subunit gene and the cytochrome oxidase II subunit gene), and A8/COIII (including genes for the ATPase-8 and -6 subunits and the cytochrome oxidase III subunit gene). The seven mtDNA regions were amplified by denaturation at 94°C for 5 min, followed by 30 cycles of 1 min at 94°C, 1 min at 55°C, and 3 min at 72°C |0.2 mM of each dNTP, 0.2p M of each primer, 2 mM MgCl2, 50 mM KC1, and 10 mM Tris-HCl, pH 8.3 with 1 unit of Taq polymerase (Perkin Elmer, Norwalk, CT) in a 50-pL reac- tion], except that amplifications of regions A8/COIII and ND3/ND4 required 3 mM instead of 2 mM MgCl.,. Subsamples of PCR products of each mtDNA region were digested with each of 12 restriction enzymes. The endonu- cleases recognized six bases ( Ase I ), multiple six-base sites ( Ava I, Hind II, Sty I), multiple 5 base sites (BstN I), and four bases (RstU I, Cfo I, Dde I, Hint I, Mho I, Msp I, Rsa I). Digestion reactions were carried out under conditions recommended by the manufacturers. The resulting frag- 530 Fishery Bulletin 99(4) Table 1 Primers for PCR amplification of coho salmon mitochondrial DNA (mtDNA), location in relation to O. mykiss (Zardoya et al., 1995), PCR fragment sizes, and fragment overlaps (bp). 0. mykiss Fragment Region Sequence locations size Overlap Source 12S/16S 5' AATTCAGCAGTGATAAACATT 3' 1234-1254 1 5' AGATAGAAACTGACCTGGATT 3' 3615-3635 2402 121 1 ND1/ND2 5' ACCTCGATGTTGGATCAGG 3' 3515-3533 1 5' ATTAAAGTGNTTGA(T/G)TTGCATTC 3' 6181-6203 2689 -430 1,5 COI/COII 5TAATCGTCACAGCCCATGCCTTCGT 3' 6634-6658 2 5' GGTCAGTTTCAGGGTTCAGGTTTAGC 3' 9079-9104 2471 166 2 A8/COIII 5' CTAGTGACATGCCCCAACTCAACC 3' 8939-8962 2,3 5' TCATAAGGCGGTCATGGACTTAAACC 3' 11028-11053 2115 480 2,3 ND3/ND4 5' TTAC GCGTAT AAGTGACTTC C AA 3' 10574-10596 2,3 5' TTTTGGTTCCTAAGACCAATGGAT 3' 12881-12904 2331 32 2,3 ND5/ND6 5' AAC AGCTC ATC C ATTGGTCTTAGG 3' 12873-12896 2, 4, 5 5' TTAC AAC G ATGGTTTTTC ATGTC A 3' 15319-15342 2470 19 2, 4, 5 Cytb/D-loop 5' TGAA( G/A 1ACCACCGTTGTTATTC AA 3' 15324-15347 2, 4, 5 5' TAGGGCCTCTCGTATAACCG 3' 1321-1340 2659 107 2, 4,5 12S/16S 1 Consensus from Anderson et al. (1981); Anderson et al. (1982); Bibb et al. (1981); Roe et al. (1985); Chang et al. (1994). 2 Unpublished mtDNA sequence data from our lab. 3 Thomas and Beckenbach (1989). 4 Cronin et al. (1993). 5 Carney et al. (1997). Gharrett et al.: Phylogeographic analysis of mitochondrial DNA variation in Oncorhynchus kisutch 531 ments were separated by electrophoresis through a 1.5% agarose gel (a mixture composed of one part Ultra Pure™ agarose [BRL Gibco, Grand, NY] and two parts Synergel™ [Diversified Biotech Inc., Boston, MA]) in 0.5xTBE buffer (TBE is 90 mM Tris-boric acid, and 2 mM EDTA, pH 7.5). DNA in the gel was stained with ethidium bromide and photographed on an ultraviolet light transillumina- tor. Digests that produced fragments too small for detec- tion in agarose/Synergel™ gels were resolved in 12% poly- acrylamide (29:1 acrylamide:bisacrylamide; lxTBE) gels. DNA fragments separated in polyacrylamide gels were stained with SYBR Green 1 Nucleic Acid Stain™ (Molec- ular probes, Eugene, OR), which is more sensitive than ethidium bromide. Either a 1-kilobase (kb) ladder or Hae III digested (f>x 174 RF phage DNA (BRL Gibco, Grand, NY) was used as a molecular weight reference for estimating restriction fragment sizes. Restriction sites were inferred from fragment patterns that could be related to each other by the gain or loss of a single site. Composite haplotypes were constructed from restriction fragment patterns of all restriction enzymes across all mtDNA PCR regions. Using the rules of Castel- loe and Templeton (1994) to resolve ambiguities, we con- structed the single most probable parsimonious tree de- picting restriction site changes between haplotypes. Using REAP (McElroy et al., 1990), we estimated haplotype (nu- cleon) and nucleotide diversities within populations (Nei, 1987) as well as average nucleotide divergences between populations. Nucleotide divergence between populations takes into account both the haplotype frequency differ- ences between populations and the nucleotide divergences between haplotypes (Nei and Tajima, 1983; Nei, 1987; Nei and Miller, 1990). Homogeneity of haplotype diversities among populations was tested by using the Monte-Carlo simulation in REAP (McElroy et al., 1990X10,000 iter- ations; Hedges, 1992) to establish probability levels for goodness-of-fit statistics (Roff and Bentzen, 1989). Populations were clustered from pair-wise nucleotide di- vergences by using the Fitch and Margoliash (1967) least- squares method (FITCH in PHYLIP; Felsenstein, 1995). For comparisons between populations, the precision of es- timates of nucleotide divergence depends on sample size. Therefore, stability of the topology was examined by boot- strapping (2000 iterations; Hedges, 1992) over individuals within each collection. A consensus tree (CONSENSE in PHYLIP; Felsenstein, 1995) that shows the stability of the topology, but not the branch lengths, was generated from the set of bootstrapped trees. The hierarchical structure of the expanded set of coho samples was analyzed by analysis of molecular variance (AM OVA; Excoffier et al., 1992) with Arlequin (Schneider et al., 1997). Collections were grouped geographically into four regions: Southeast Alaska, Southcentral Alaska, Ber- ing Sea, and Asia. With appropriate choices of divergence matrices, the analysis can examine the structure from haplotype frequencies (e.g. Weir, 1996) or from nucleotide diversities based on paths between haplotypes traced through a haplotype tree (Excoffier et al., 1992). Signifi- cance (PMC) of d>-statistics (Excoffier et al., 1992) was es- timated from distributions of the statistics generated by 17,000 permutations (Hedges, 1992) at the appropriate level of hierarchy. Nested clade analysis of geographical distributions of haplotypes and subclades (e.g. Templeton and Sing, 1993; Castelloe and Templeton, 1994; Templeton, 1998) was con- ducted with GEODIS 2.0 (Posada et al., 2000). Results Our general approach was to conduct a broad preliminary survey to obtain genome-wide information for mtDNA restriction site variation. Subsequently, we focused on the variable restriction sites and examined larger sample sizes and additional populations. From those results we con- structed a fine-scale mtDNA “gene tree” and analyzed the geographic distributions of mtDNA lineages to deduce the nature of the historical demographic changes that influ- enced present-day population structure. The approach also allowed us to determine the effects on estimates of molec- ular parameters that occur when analyses focus on vari- able restriction sites. Diversities of coho salmon mtDNA Ten coho salmon from each of seven drainages (Fig. 1) were analyzed to survey broadly the species’ mtDNA vari- ability using 12 restriction endonucleases (Appendix 1). The total number of restriction sites inferred from restric- tion fragment patterns was 298 (an average of 291.28 per haplotype), which corresponds to 1284 nucleotides (an average of 1254.80), or a maximum of 7.73% (an average of 7.56%) of the coho salmon mtDNA genome (Table 2). Sixteen sites (1.2% of the total) were variable. Individu- ally, the regions averaged between 29 and 57 restriction sites, which correspond to a maximum of between 5.58% and 9.57% of the nucleotides in a region. Although the amplified regions had some overlaps (Table 1), no vari- able sites were observed in the areas of overlap; and no invariant sites were shared between regions, except possi- bly Dde I sites in the 408-bp overlap between A8/COIII and ND3/ND4 (Table 1). Because of the large total number of sites examined, a few overlapping Dde I sites would cause only a slight decrease in nucleotide diversity estimates and have little effect on nucleotide divergence estimates. Restriction site variation was observed in five of the sev- en PCR-amplified mtDNA regions for the 12 restriction endonucleases used. Between zero and five variable sites were observed per region. No variation was detected in the A8/COIII and ND3/ND4 regions. The largest number of variable sites (5) and the highest level of nucleotide diver- sity (5.99 substitutions per 1000 base pairs) were observed in the ND5/ND6 region (Table 2). Because there is no recombination between heterolo- gous mtDNA molecules, the composite haplotype is the ap- propriate unit to consider in genetic analysis (e.g. Avise, 1989). Our preliminary survey discovered 11 haplotypes (Table 3). As a whole, the sample of 70 fish had a haplotype diversity of 0.803 and a nucleotide diversity of 1.70 substi- tutions per 1000 base pairs. Haplotype diversities within 532 Fishery Bulletin 99(4) Table 2 Number and variability of restriction sites observed in each of the seven mtDNA regions we examined using 12 restriction endo- nucleases in our preliminary survey. Region Fragment size Mean number of sites Mean number of nucleotides % coverage Number of variable sites Number of haplotypes Haplotype diversity USE) Nucleotide diversity (per 1000) 12S/16S 2402 53.5 226.7 9.44 3 4 0.485 ±0.042 2.15 ND1/ND2 2689 41.5 184.7 6.87 3 4 0.485 ±0.042 2.65 COI/COII 2471 39.7 168 6.8 3 3 0.057 ±0.038 0.25 A8/COIII 2115 29 126.7 5.99 0 1 0 0 ND3/ND4 2331 31 130 5.58 0 1 0 0 ND5/ND6 2470 36.5 157.2 6.36 5 4 0.470 ±0.040 5.99 Cytb/D-loop 2659 57 248.7 9.35 2 3 0.670 ±0.014 1.87 Total 16,642 291.28 1254.8 7.54 16 11 0.803 ±0.024 1.70 Table 3 Observed numbers of each mtDNA haplotype, haplotype diversity, and nucleotide diversity (substitutions per 1000 bp) within col- lections of coho salmon examined in a preliminary survey of North Pacific Ocean coho salmon (Nei and Tajima, 1983; Nei ,1987). Standard errors are in parentheses. Homogeneity of haplotype frequencies ( PMC < 10~ 4 ) was tested with using Monte-Carlo simula- tion based on 10,000 resampling iterations to estimate probability (Roff and Bentzen, 1989). Collection n A-D cluster Haplotype E-H cluster Haplotype Haplotype diversity Nucleotide diversity A A' B C C' D E E' F G H Hugh Smith River 10 3 0 0 4 1 0 2 0 0 0 0 0.78 1.31 Fish Creek (Taku River) 10 7 1 0 0 0 1 0 0 1 0 0 0.53 0.88 Ford Arm River 10 2 0 4 1 0 0 2 0 0 1 0 0.82 1.65 Crooked Creek 10 5 0 0 4 0 0 1 0 0 0 0 0.64 0.78 Eek River 10 0 0 0 4 0 0 4 0 0 0 2 0.71 1.87 Delta Clearwater River 10 0 0 0 0 0 0 0 0 0 0 10 0.00 0.00 Kamchatka River 10 5 0 0 4 0 0 0 1 0 0 0 0.06 0.94 Total 70 22 1 4 17 1 1 9 1 1 1 12 0.80 1.70 (±0.024) (±0.000) Average 0.59 1.06 (±0.11) (±0.24) collections from individual drainages ranged from 0.00 to 0.82 and nucleotide diversities ranged from 0.00 to 1.87 substitutions per 1000 bp. The distribution of haplotypes among collections was highly heterogeneous, which is note- worthy given the small sample sizes (n=10) (Table 3). Phylogenetically, there were two clusters of haplotypes (haplotypes A-D and haplotypes E-H) (Fig. 3). Haplotypes of the two clusters differed by five or more restriction sites, and the average nucleotide divergence between individual haplotypes in the two clusters averaged 2.72 nucleotide substitutions per 1000 nucleotides, as compared to an av- erage nucleotide diversity within each cluster of 0.87. The cluster of A-D accounted for the majority of fish, but hap- lotypes of both clusters were observed in collections from all drainages, except for fish from the Delta Clearwater River, which had only haplotype H. Bootstrap estimates of nucleotide divergence between clusters and average nu- cleotide diversity within clusters, estimated from the en- tire sample of 70 fish, were 2.98 ±0.06 and 0.35 ±0.05 (sub- stitutions per 1000 nucleotides), respectively. Expanded coho mtDNA survey based on variable sites We increased the number of populations surveyed to 13 and increased sample sizes to 20, except for the Kam- Gharrett et al.: Phylogeographic analysis of mitochondrial DNA variation in Oncorhynchus kisutch 533 Figure 3 Haplotype tree for coho salmon mtDNA restriction site variation. Haplotypes A-H, including A', C', and E', were observed in the preliminary survey (70 fish); haplotypes 1-0 were observed in the subsequent survey ( 188 additional fish); and haplotype P was detected in digests conducted to resolve positions on the tree of the latter haplotypes. Arrows indicate the direction of site gain. Ambiguities in the topology were resolved (Castelloe and Templeton, 1994) and the nested structure was constructed (Templeton and Sing, 1993). Hap- lotypes A', C', and E' were pooled with A, C, and E, respectively, for the nested clade analysis. The results of the analysis are provided in Table 5. chatka River (n=17) and Delta Clearwater River (n= 21) populations. To make analysis of larger numbers of sam- ples practical, we focused on restriction sites in five mtDNA regions that defined the most abundant eight (A-H) of the 11 haplotypes. Haplotypes A', C', and E', each of which was represented by a single individual, were eliminated. In this survey, we analyzed site variation for the following PCR region by restriction endonuclease com- binations: 12S/16S rRNA Cyt b/D-loop ND5/ND6 ND1/ND2 ND3/ND4 Cfo I BstN I Ava I Dde I Cfo I Dde I Dde I Sty I This survey detected 62 restriction sites (56.47 on aver- age in each haplotype), which correspond to 262.67 nucleo- tides (238.22 on average). The proportion of the genome screened was a maximum of 1.58% (1.43% on average). To completely resolve the placement of haplotypes I, J, K, and L in the “gene tree,” we digested their ND5/ND6 regions with Mbo I and their COI/COII regions with Dde I. An ad- ditional haplotype (P) was resolved. In total, the expanded survey resolved three additional haplotypes (I, J, and P) within the A-D clade and five additional haplotypes (K, L, M, N, and O) within the E-H clade (Table 4, Appendix 2, and Fig. 3). Haplotypes of both clusters appeared in most drainages. It is notable that four drainages included hap- lotypes of only a single clade: Delta Clearwater had 20 of haplotype H and one of the related haplotype O; Indian River had 19 of haplotype A and one of haplotype C; Bern- ers River had five of hapotype A, 11 of haplotype C, one of haplotype I, two of haplotype J, and one of haplotype P; and the Little Susitna collection had four A haplotypes, 15 C haplotypes, and one I haplotype. Within drainages, haplotype diversities ranged from 0.10 to 0.73 and nucleo- tide diversities ranged from 0.22 to 9.07 substitutions per 1000 bp. The 13 collections exhibited strong heterogeneity (P<10^, Table 4). In a Fitch-Margoliash phenogram that estimates rela- tionships among drainages based on haplotype frequen- cies (Fig. 4), Delta Clearwater River differed strongly from the other collections, and the collections from systems that drain into the Bering Sea and from Karluk Lake on Kodiak Island, clustered separately. The remaining col- 534 Fishery Bulletin 99(4) Table 4 Observed numbers of each mtDNA haplotype, haplotype diversity, and nucleotide diversity (substitutions per 1000 bp) within col- lections of North Pacific coho salmon screened for variable sites detected in a preliminary survey (Table 3)(Nei and Tajima, 1983; Nei, 1987). Standard errors are in parentheses. Homogeneity of haplotype frequencies ( PMC < 10-4) was tested by using Monte-Carlo simulation based on 10,000 resampling iterations to estimate probability (Roff and Bentzen, 1989). Collection n Haplotype Haplotype diversity Nucleotide diversity A B C D E F G H I J K L M N 0 p Hugh Smith River 20 6 0 12 0 2 0 0 0 0 0 0 0 0 0 0 0 0.57 3.64 Fish Creek (Taku River) 20 11 0 3 2 0 2 0 0 0 1 0 1 0 0 0 0 0.68 5.68 Berners River 20 5 0 11 0 0 0 0 0 1 2 0 0 0 0 0 1 0.64 2.40 Indian River 20 19 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0.10 0.22 Ford Arm River 20 10 6 1 0 2 0 1 0 0 0 0 0 0 0 0 0 0.68 4.51 Crooked Creek 20 9 0 10 0 1 0 0 0 0 0 0 0 0 0 0 0 0.57 2.50 Little Susitna River 20 4 0 15 0 0 0 0 0 1 0 0 0 0 0 0 0 0.42 1.13 Buskin River 20 5 0 12 0 3 0 0 0 0 0 0 0 0 0 0 0 0.58 4.69 Karluk River 20 1 0 9 0 5 0 0 0 0 0 1 0 0 4 0 0 0.73 9.07 Eek River 20 1 0 7 0 9 0 0 3 0 0 0 0 0 0 0 0 0.68 8.38 Kanektok River 20 5 0 8 0 5 0 0 1 0 0 0 0 1 0 0 0 0.75 7.91 Delta Clearwater River 21 0 0 0 0 0 0 0 20 0 0 0 0 0 0 1 0 0.1 0.2 Kamchatka River 17 9 0 7 0 1 0 0 0 0 0 0 0 0 0 0 0 0.58 2.7 Total 258 85 6 96 2 28 2 1 24 2 3 1 1 1 4 1 1 0.734 6.73 (±0.016) (±0.00) Average 0.54 4.08 (±0.06) (±0.83) lections spanned the southern portion of the geographic range from southern Southeast Alaska to the Kamchatka Peninsula. Within that set of collections, the two coastal Southeast Alaskan collections (Ford Arm and Indian Riv- er) appeared to form a weak cluster, but there was no obvi- ous structure among the remaining collections. We conducted AMOVA analyses reflecting a geograph- ical hierarchy: Southeast Alaska, Southcentral Alaska, western and interior Alaska (Bering Sea), and Asia to examine the geographic basis of variation. Although the Karluk River collection resembled Bering Sea collections more than other northern Gulf of Alaska collections, we in- cluded it with the Southcentral Alaska group to maintain the geographic basis of the analysis. Analyzing the data based on haplotype frequencies (analogous to allelic dif- ferences in analysis of variance described by Weir [1996]) revealed highly significant divergence among collections ((PST=0.291, PMC<0.0001), most of which is attributable to average divergence among drainages within a region ( d>gc=0.227, PMC<0.0001 ), rather than differences between regions (rr=0.083, PMC=0.094). Incorporating relation- ships between the haplotypes into the analysis increased the proportion of the total divergence observed among drainages ( &ST= 0.449, PMC<0.0001) and among drainages within regions (d>sc=0.273 , PMC<0.0001). The estimate of the proportion of divergence among regions also increased (0ct=O.‘242, Pmc= 0.083). Estimates of long-term gene flow [AIe(pm=(d>Yy_1_1^2] from (c0+ cfT), where T = temperature; t = time in hours since ingestion; b0 and 6X = the coefficients of the exponential relation- ship between evacuation rate and tempera- ture; and Hurst and Conover: Diet and consumption rates of Morone saxatilis in the Hudson River 547 Table 1 Summary of collections and diets of overwintering YOY striped bass collected in the Hudson River Estuary, 1993-97. %F = mean frequency of occurrence, %W = mean percentage of wet weight; n = number of fish sampled per date. Winter 1993 1994 1995 1996 1997 No. of dates sampled 1 8 7 6 4 L Temperature (°C) 2.0- -5.1 4.2- -10.4 1.0-9. 4 3. 3-6. 2 5.0- -6.1 n 20- -34 15- -53 8-37 16-29 18- -52 Percentage empty stomachs 71.0 52.5 72.3 57.6 73.4 Mean gut fullness ±SE 0.44 ±0.103 0.51 ± 0.056 0.34 ±0.066 0.18 ±0.037 0.26 ± o o a o Prey type Scientific name %F %W %F %W %F %W %F %W %F %w Amphipods Gammarus sp. 66.7 59.1 92.1 90.2 58.2 39.5 86.3 52.4 34.3 15.7 Sand shrimp Crangon septimspinosa 37.0 27.1 4.2 4.2 7.1 8.1 0 0 3.1 5.3 Grass shrimp Palaemonetes pugio, Palaemonetes vulgaris 38.9 6.4 0 0 4.3 4.1 0 0 3.1 7.5 Mysid shrimp 0 0 0 0 0 0 3.0 7.1 18.8 4.3 Unidentifiable C. septimspinosa, 0 0 0.7 0.4 27.5 22.8 16.4 32.5 21.4 26.4 shrimp Crabs P. pugio, P. vulgaris 33.3 5.7 0 0 0 0 0 0 0 0 Polychaetes 0 0 0 0 17.9 23.3 0 0 0 0 Oligochaetes 0 0 0 0 0 0 0 0 23.9 38.8 Fish Anchoa mitchilli, Menidia menidia, Ammodytes americanus 1.9 1.7 5.4 5.3 2.9 2.2 7.1 8.0 3.1 1.9 UIR7 40.7 24.4 11.9 8.5 28.2 6.0 31.8 29.2 42.2 25.1 1 UIR = unidentified invertebrate remains. These were not included in the calculation of %W of other prey items. c0 and cx = the intercept and slope of the linear relation- ship between temperature and the lag prior to the beginning of evacuation. If no lag is present in the data, the parameters c0 and Cj approach 0. The effect of body size on evacuation was as- sessed by examining the relationship between fish length and deviation from the best fitting evacuation model. Diet of YOY striped bass Overwintering YOY striped bass were collected with a 9-m bottom trawl (38-mm stretch mesh codend) from the lower Hudson River estuary, the only known wintering aggre- gation for the Hudson River population (Dovel, 1992). Sampling occurred throughout the overwintering period of five consecutive years. Fish were captured during day- light hours between river mile 0 and 9 in conjunction with the New York Power Authority’s hatchery evaluation and tagging program. Sampling in each year began in mid- December and ended in late March or early April. Bottom water temperatures were measured during sampling. The number of fish analyzed for diet per date ranged from 8 to 53 depending on catch rates (Table 1). Captured fish were individually wrapped and immediately frozen for preservation. In the laboratory, fish were thawed, mea- sured (mm TL), weighed (g wet weight), and stomach con- tents were removed. Prey were identified to the lowest possible taxon and weighed. All prey in each category in each stomach were weighed as a group because the abun- dance and small size of the dominant prey item made individual measurements unfeasible (gammarid amphi- pods; often >50/stomach and <0.005 g each). Diets of YOY striped bass were described by the contribution of items expressed on the basis of both weight (%W —weight of prey - /weight of all identifiable prey ) and frequency (%F = 100 x no. of stomachs with prey J no. of stomachs with prey). Unidentifiable items were measured separately and are expressed as a percentage of the total weight of all stom- ach contents. Estimation of consumption rates Daily consumption rate estimates were generated by com- bining gut fullness values (S=total weight of stomach contents/fish weight) of overwintering YOY striped bass with laboratory-determined gastric evacuation rates with the method of Eggers (1979): C = 24 R, x S, 548 Fishery Bulletin 99(4) where S = average gut fullness of field-caught fish; and Re= exponential evacuation rate at measured field temperatures determined from labora- tory experiments described above. This simplified version of the Elliott and Persson (1978) model is appropriate when sampling is not conducted in discrete time intervals. We used gut-fullness measures from fish collected throughout the day, as described above. Although fish were sampled only during daylight hours, the sampling interval (24 hours) is substantially shorter than the evacuation time (>72 hours), even at the warmest temperatures observed; hence any diel feeding patterns will have little effect on our consumption estimates. Standard deviations of consumption estimates were gen- erated from a Taylor expansion of the consumption model dC ' Z 2 aR + ' dC 1 Z <4+2 ' dC " ‘ dC ' [dS l dRe J s l dRe ) dS ) The variance in the evacuation rate parameter (Re) was estimated through a bootstrap procedure by fitting the evacuation model to 1000 sets of 152 observations sampled with replacement from the evacuation rate data. Since Rc is a function of temperature, the variance in Re at a given temperature was determined by inserting the_tempera- ture into the 1000 model fits. The variance in S was esti- mated for each date as var(S-)/n. The standard deviation of the consumption estimate de- pends upon the covariance between evacuation rate (Re), measured in laboratory experiments, and gut fullness ( S ), observed in wild fish. Because we have no way of measur- ing the covariance between these parameters, we evaluat- ed the importance of this term by calculating the variance under three assumptions: 1) Sand Re are not correlated, 2) they covary perfectly, and 3) they display perfect nega- tive covariance. Analysis of feeding patterns We investigated feeding patterns ofYOY striped bass by examining the relationship between several factors (body size, time of year, water temperature, and energy storage) and gut fullness, at the individual and population level. Gut fullness was chosen over consumption rate for three rea- sons. First, gut fullnesses were measured directly, whereas consumption rates were estimated from a model by using gut fullnesses. Second, consumption estimates are directly dependent on temperature, an independent variable in these analyses. Finally, we were interested in determining the conditions that stimulate feeding, which we believe are more immediately reflected in the gut fullness measures. Feeding patterns at the individual level were investi- gated by examining the relationships between gut fullness and the independent variables of body size, lipid level, wa- ter temperature, and time of year. Gut fullness values of individual fish were dominated by zeros (empty stomachs) and could not be transformed appropriately, preventing the use of parametric statistics. Because empty stomachs may reflect either a lack of prey availability or reduced appetite, we concentrated analyses at the individual level on the slope of the 95% quantile of gut fullness regressed against the independent variables (Scharf et al., 1998). This technique has been used to examine scatter-plots, when boundaries of a relationship between two factors are of interest, rather than the mean. In our case we were in- terested in the factors that stimulate feeding as opposed to developing a predictive relationship between indepen- dent variables and gut fullness. The 95% quantile line de- scribes the relationship between maximum observed gut fullness and the independent variables. Mean gut fullness of all fish captured on a given date was used to investigate feeding patterns at the population level and was compared to average lipid level of fish captured on that date, time of year, and water temperature, by using Kendall’s coeffi- cient of rank correlation (Sokal and Rohlf, 1981). Lipid levels were determined for a subsample of fish from the diet analysis. This involved determination of the percentage of dry body weight comprising non polar lipids (those used primarily for energy storage). Details of the compositional analysis procedure can be found in Hurst et al. (2000) and Schultz and Conover (1997). Compositional analysis was performed on 15 to 40 fish from each sam- pling date (except 13 December 1995). The relationship between lipid level and gut fullness at the individual level was examined by using all fish for which both pieces of in- formation were available (zz =587), whereas analyses at the population level used the average lipid level observed on a given date [n= 28). Results Gastric evacuation rate Evacuation rates of overwintering YOY striped bass declined greatly with temperature. At mid-winter temper- atures, evacuation rates were among the lowest reported for any fish species. Time to 50% evacuation ranged from 31 hours at 11°C to 101 hours at 2°C (Fig. 1). The best fit parameters in the evacuation model were 60=-0. 00685, 6j= 0.126, c0= 25.757, and c1=-1.8033. We observed a lag between feeding and a measurable loss of material from the stomach. The length of the ob- served lag decreased as temperature increased from 18.2 hours at 2°C to 6.1 hours at 11°C. The exponential mod- els used here to describe digestion fitted the experimental data as well as, or significantly better than, other common models (linear and square root; Bromley 1994). The amount of variability among individuals in evacua- tion rate increased as temperature decreased, leading to a poorer fit of the evacuation models at the lower temper- atures (Fig. 1). Deviations from the model increased sig- nificantly as temperature decreased (P=0.018; ANOVA of absolute value of sample deviations from evacuation mod- el). Body size had no effect on evacuation rate among ju- venile striped bass. We found no correlation between re- sidual values from the evacuation model and fish length (r=-0.03 P=0.685). Hurst and Conover: Diet and consumption rates of Morone saxatilis in the Hudson River 549 tr Hours since ingestion Figure I Gastric evacuation patterns of YOY striped bass fed Crangon septemspinosa at four tem- peratures. Points are observations of percentage of initial meal remaining in the stomach after a predetermined time. Lines represent evacuation model fitted by using all data. Diet composition Benthic invertebrates were the dominant prey of YOY striped bass overwintering in the lower Hudson River estu- ary, making up 95.0% of the diet by weight; the remaining 5.0% were various fish prey (Table 1). Most striped bass captured during winter had empty stomachs (64.8%), and on two dates, all fish sampled had empty stomachs (30 March 1994 and 6 December 1994). Unidentifiable stomach contents ranged from 6.0% weight (%W) in 1995 to 29.2%W in 1996. The most common item in the diet was gammarid amphipods (those identified to genus were Gammarus sp., but most were not identified) making up 51.3%W and occurring in 81% of stomachs with food (%F). Several spe- cies of shrimp were also common dietary items, including sand shrimp (C. septemspinosa ) and grass shrimps (Palae- monetes spp.). Other invertebrates that were important items on several dates included mysid shrimp, polychaetes, and oligochaetes. Fish were rarely found in the stomachs of YOY striped bass (3.8%F) but included bay anchovy (Anchoa mitchilli), Atlantic silversides (Menidia menidia ), and American sand lance (Ammodytes americanus ). The importance of individual prey taxa varied among years and among dates within each year, although low sample sizes of fish with prey prevent drawing strong conclusions from these data. In 1994, striped bass fed al- most exclusively on gammarid amphipods (90.2%W and 92.1%F), but these prey were less important in other years. Shrimp were important in 1993 (33.5%W, mostly C. septemspinosa) and 1995-97 (34.9 to 43.6%W all species). Polychaetes were observed in striped bass diets only in 1995 (23.3%W and 17.9%F). Oligochaetes were found com- monly in 1997, averaging 18.5%F and 23.9%W of the diet but did not occur in other years. Consumption rates Assuming an exponential evacuation pattern, consumption rate estimates for overwintering YOY striped bass ranged from 0 to 0.29% body weight per day (%bw/day, Fig. 2). Stan- dard deviations of estimated consumption averaged 46% of the estimate (range: 20.2-100.9%). The standard devi- ation of the consumption estimate was relatively insensi- tive to assumptions of the covariance between S and Re. The ratio of estimates under the least conservative assump- tion (perfect covariance) to that under the most conserva- tive assumption (perfect negative covariance) was less than 1.08, with one exception (1.17 on 20 December 1993). The standard deviations presented in Figure 2 are the interme- diate values, based on the assumption of no covariance. Feeding patterns Among individuals, we observed a strong negative relation- ship between maximum gut fullness and the level of stor- age lipids. The 95th quantile of gut fullness was 3% for fish with lipid reserves of 2% dry weight and decreased to under 1% for fish with lipid levels in excess of 19%< dry weight (Fig. 3). No significant patterns were observed between individual gut fullness levels and water temperature, time 550 Fishery Bulletin 99(4) of year, or body size. At the population level, only date was significantly correlated with mean gut fullness (Kendall’s rank correlation: P- 0.032; Fig. 4). Addition of temperature and average lipid level did not provide significant improve- ments in the fit of the model (P>0.20 for both). Discussion Evacuation rates Laboratory-measured evacuation rates of YOY striped bass at representative winter temperatures were among the lowest reported for any fish species. None of the estimates compiled from multiple studies by He and Wurtsbaugh (1993) included rates as low as the 0.027 we measured at 11°C. The reason for this observation is not that striped bass have exceedingly low digestion rates compared with other fishes, but rather that evacuation rate is strongly related to temperature and is rarely measured at the lower end of temperatures encountered by each species. For exam- ple, the lowest temperature at which evacuation rates in white perch (M or one americana) have been measured is 12.6°C (Parrish and Margraf 1990), despite the fact that this fish often occupies waters near 0°C. The only evacua- tion rates we found in the literature comparable to those we measured were those for largemouth bass at 4°C (f?e=0.006; Cochran and Adelman, 1982, from data in Markus, 1932) and for brown trout at 1.8°C (7^=0.026; Jensen and Berg, 1993), both measured near the species’ thermal minima. Factors other than temperature have been found to in- fluence evacuation rates such as prey type, method of feed- ing (voluntary vs. force-feeding), and body size. All of our experiments were conducted with a single prey type (C. septemspinosa) at one ration level (2% of body weight). Meal size has been found to affect evacuation rate in lab- oratory studies (Smith et al., 1989; Andersen, 1998), but this effect is substantially smaller than that of tempera- ture (He and Wurtsbaugh, 1993). Our meal size was slight- ly higher than the mean, but well within the range ob- served among individual fish. We found that body size did not affect evacuation rate over the size range of over- wintering YOY striped bass (88-150 mm). In several stud- ies comparing evacuation rates of various prey, large bod- ied shrimp, including C. septemspinosa, were found to be evacuated at lower rates than soft bodied prey or smaller shrimp species (Nelson and Ross, 1995; Singh-Renton and Bromley, 1996; Lankford and Targett, 1997), whereas oth- er studies found no differences among prey types (Juanes and Conover, 1994; dos Santos and Jobling, 1995). Al- though C. septemspinosa and other shrimp species are common diet items of YOY striped bass, they were not the dominant item. If other prey items, such as amphi- pods, are digested and evacuated more rapidly, consump- tion rates of wild fish will be underestimated when non- shrimp prey dominate the diet. We allowed a lag in our description of meal evacuation based on laboratory observations. Such a lag has been ob- served in several other studies, including both those where fish fed voluntarily (Gerald, 1973; Grove et al., 1985) and were force-fed (Vondracek, 1987). The length of the lag ob- served in juvenile striped bass decreased as temperature increased, as seen in juvenile turbot ( Seopthalmus maxi- mus\ Grove et al., 1985). The lag prior to beginning of evac- uation could be an artifact of experimental conditions or a natural delay in the passing of food from the stomach to the intestine following ingestion of shelled prey. Our mod- el estimated the evacuation rate parameter ( Re ) after the lag because this parameter represents the rate of passage of food from the stomach. Consumption rates based on gut fullness levels may be overestimated if there is a consid- erable lag prior to the beginning of digestion. However, the digestive lag we observed was short compared with Hurst and Conover: Diet and consumption rates of Morone saxotilis in the Hudson River 551 co 05 05 CD (3 2 - 0 - / * * * JS. 4 *. t . 10 15 20 25 Lipid level (% dry weight) 30 35 Figure 3 Relationship between gut fullness and lipid level of overwintering YOY striped bass through five winters. Points are gut fullness and lipid levels of individual fish. Q co +i 1 Jan 1 Feb Date 1 Apr Figure 4 Relationship between mean gut fullness (±1 SE) observed on a given date and date for collections of YOY striped bass through five winters. the total evacuation time, averaging only 11% of the time required to reach 75% digestion. Sim- ulations that varied the length of the digestive lag time and the evacuation rate showed that for lag times in the range we observed, consump- tion rates were overestimated by less than 5%. If the lag observed in the laboratory experiments does not occur in the wild, estimated consump- tion rates are unbiased. Diet Diets of overwintering YOY striped bass in the lower Hudson River estuary were dominated by benthic invertebrates such as gammarid amphi- pods and several shrimp species. Juvenile striped bass do not appear to undergo a major diet shift from summer to winter but appear to focus more heavily on amphipods in winter. Studies of summer diets of striped bass in the Hudson River from the 1970s and 1990s found a slightly more diverse diet than we observed in winter. Summer diets included copepods, chironomids, and isopods, and more commonly incorporated fish and polychaetes (Gardinier and Hoff, 1982; Hurst, unpubl. data). Data from other estuaries suggest that the diets of overwintering juvenile striped bass may vary regionally. In Chesapeake Bay, Hartman and Brandt (1995a) found that fish prey accounted for 20-25%W of YOY striped bass diets in winter, substantially more than the 2-9%W we observed in the Hudson River. In the Miramichi River estuary, overwintering fish fed primarily on shrimp (mysids and C. septemspi- nosa) and ceased feeding when temperatures fell below 3°C in late November (Robichaud-LeBlanc et al., 1997). The studies from Chesapeake Bay and the Miramichi River estuary did not examine interannual variability in diets. Overwintering juvenile striped bass appear to be opportunistic feeders, their diets reflective of the epibenthic invertebrate community in the low- er Hudson River and similar to published informa- tion on the diets of co-occurring species. Although there have been no surveys of the benthic com- munity in the lower Hudson River in winter, data available from summer surveys in the Hudson River estuary and the adjacent Raritan estuary suggest dominance of the epibenthic community by gammarid amphipods, shrimp, and annelids (Ristich et al., 1977; Steimle and Caracciolo-Ward, 1989). The winter fish community in the lower Hudson River is composed primarily of striped bass, white perch, Atlantic tomcod (Microgadus tomcod), and winter flounder (Pseudo- pleuronectes americanus). Winter diets of Atlantic tomcod were dominated by gammarid amphipods and copepods and were significantly less diverse in winter than in spring and autumn (Grabe, 1977, 1980). Diets of age-1 and older striped bass overwintering in the lower Hudson River estu- ary are similar to those of YOY fish, although the incidence of juvenile fish prey (including YOY striped bass) increased with body size (Dunning et al., 1997). Consumption rates Consumption estimates of overwintering YOY striped bass in the Hudson River were consistently below l%bw/day, with only 34% of captured fish containing prey items. Our field estimates of the consumption rates of over- wintering striped bass are significantly lower than pub- 552 Fishery Bulletin 99(4) lished estimates of maximum consumption measured in laboratory experiments (Hartman and Brandt, 1995b). Their estimates of maximum consumption (for a 10-g fish) increased from 0.45%bw/day at 1°C to 7.10%bw/day at 10°C. The highest percentages of predicted maximum con- sumption achieved by YOY striped bass in the Hudson River occurred on 6 January 1994 (25.8%), and on three dates in early winter 1995 (15.5-18.0%). On other dates consumption was generally below 10% of maximum esti- mated from the Hartman and Brandt (1995b) model. Bull et al. (1996) developed a model of consumption for overwintering juvenile Atlantic salmon in which appe- tite was related to anticipated metabolic requirements. To minimize the risks of starvation and predation associated with foraging, they predicted that appetite of overwinter- ing fish should be highest in early winter (when future metabolic needs are greatest) or when internal energy re- serves are low. Our results suggest that a similar model might be appropriate for overwintering striped bass. Gut fullness levels of YOY striped bass were related to level of lipid reserves at the individual level (Fig. 3). Although some fish had empty stomachs at all energy levels, the highest observed gut fullnesses were found in fish with low lipid levels. This finding suggests that overwintering striped bass increase feeding activity when energy re- serves become depleted. Such feeding patterns have been documented in laboratory experiments with Atlantic salm- on (Metcalfe and Thorpe, 1992) and striped bass (Hurst and Conover, 2001) but have not previously been observed among fish feeding in the wild. At the population level, we observed a negative relation- ship between mean gut fullness and date; gut fullnesses were higher in early winter than late winter (Fig. 4). This pattern was predicted in the Atlantic salmon model (Bull et al., 1996) but could also be due to external factors such as depleted food resources at the end of winter. Benthic prey production is likely reduced by low winter temperatures, and standing stocks may become depleted as winter pro- gresses. Further work is required to determine fully the causes and implications of reduced gut fullnesses observed in later winter. Reduced feeding in late winter may reflect the availability of sufficient energetic reserves and suggests that starvation is unlikely. Conversely, reduced feeding due to depressed prey availability in late winter would indicate a strong potential for winter starvation. The role of starva- tion in winter mortality will depend greatly on determining if variable feeding patterns among years are due to inter- nally controlled variations in feeding motivation or environ- mentally imposed constraints on prey availability. Acknowledgments We thank J. Powers for assistance with stomach content and lipid analyses. Significant assistance with field collec- tions was provided by the New York Power Authority and the crews of the RV Pannaway and the RV Heather M II. S. Munch provided advice on data analysis. R. Cowen, R. Cer- rato, D. Lonsdale, E. Houde, and two anonymous reviewers provided valuable comments on this manuscript. This work was funded by a graduate fellowship in population biology from the Electric Power Research Institute (T.P. Hurst) and by research grants from the Hudson River Foundation for Science and Environmental Research and the National Science Foundation (OCE9530209 to D.O. Conover). Literature cited Andersen, N. G. 1998. The effect of meal size on gastric evacuation in whit- ing. J. Fish Biol. 52:743-755. Boreman, J., and H. M. Austin. 1985. Production and harvest of anadromous striped bass stocks along the Atlantic coast. Trans. Am. Fish. Soc. 114:3-7. Boynton, W. R., T. T. Polgar, and H. H. Zion. 1981. Importance of juvenile striped bass food habits in the Potomac estuary. Trans. Am. Fish. Soc. 110:56-63. Bromley, P. J. 1994. The role of gastric evacuation experiments in quanti- fying the feeding rates of predatory fish. Rev. Fish Biol. Fish. 4:33-66. Bull, C. D., N. B. Metcalfe, and M. Mangel. 1996. Seasonal matching of foraging to anticipated energy requirements in anorexic juvenile salmon. Proc. Royal Soc. London Series B, 263:13-18. Cochran, P. A., and I. R. Adelman. 1982. Seasonal aspects of daily ration and diet of largemouth bass, Micropterus salrnoides, with an evaluation of gastric evacuation rates. Environ. Biol. Fishes 7:265-275. Cunjak, R. A., and G. Power. 1987. The feeding and energetics of stream-resident trout in winter. J. Fish Biol. 31:493-511. Daiber, F. 1956. A comparative analysis of the winter feeding habits of two benthic stream fishes. Copeia 1956:141-151. Diana, J. S. 1979. The feeding pattern and daily ration of a top carni- vore, the northern pike ( Esox lucius). Can. J. Zool. 57: 2121-2127. dos Santos, J., and M. Jobling. 1995. Test of a food consumption model for the Atlantic cod. ICES J. Mar. Sci. 52:209-219. Dovel, W. L. 1992. Movements of immature striped bass in the Hudson estuary: estuarine research in the 1980’s. State Univ. New York Press, Albany, NY, p. 276-300. Dunning, D. J., J. R. Waldman, Q. E. Ross, and M. T. Mattson. 1997. Use of Atlantic tomcod and other prey by striped bass in the lower Hudson River estuary during winter. Trans. Am. Fish. Soc. 126:857-861. Eggers, D. M. 1979. Comments on some recent methods for estimating food consumption by fish. J. Fish. Res. Board Can. 36: 1018- 1020. Elliott, J. M., and L. Persson. 1978. The estimation of daily rates of food consumption for fish. J. Anim. Ecol. 1978:977-991. Foltz, J. W., and C. R. Norden. 1977. Seasonal changes in food consumption and energy content of smelt (Osmerus mordax) in Lake Michigan. Trans. Am. Fish. Soc. 106:230-234. Foy, R. J., and A. J. Paul. 1999. Winter feeding and changes in somatic energy content Hurst and Conover: Diet and consumption rates of Morone saxatilis in the Hudson River 553 of age-0 Pacific herring in Prince William Sound, Alaska. Trans. Am. Fish. Soc. 128:1193-1200. Gardinier, M. N., and T. B. Hoff. 1982. Diet of striped bass in the Hudson River estuary. N.Y. Fish Game J. 29:152-165. Gerald, V. M. 1973. Rate of digestion in Ophiocephalus punctatus Bloch. Comp. Biochem. Physiol. 46A: 195-205. Grabe, S. A. 1977. Food and feeding habits of juvenile Atlantic tomcod, Microgadus tomcod, from Haverstraw Bay, Hudson River. Fish. Bull. 76:89-94. 1980. Food of age 1 and 2 Atlantic tomcod, Microgadus tomcod, from Haverstraw Bay, Hudson River, New York. Fish. Bull. 77:1003-1006. Grove, D. J., M. A. Moctezuma, H. R. J. Flett, J. S. Foott, T. Watson, and M. W. Flowedew. 1985. Gastric emptying and the return of appetite in juve- nile turbot, Scopthalmus maximus L., fed on artificial diets. J. Fish Biol. 26:339-354. Hartman, K. J., and S. B. Brandt. 1995a. Trophic resource partitioning, diets, and growth of sympatric estuarine predators. Trans. Am. Fish. Soc. 124:520-537. 1995b. Comparative energetics and the development of bio- energetics models for sympatric estuarine piscivores. Can. J. Fish. Aquat. Sci. 52:1647-1666. He, E., and W. A. Wurtsbaugh. 1993. An empirical model of gastric evacuation rates for fish and an analysis of digestion in piscivorous brown trout. Trans. Am. Fish. Soc. 122:717-730. Hurst, T. P, and D. O. Conover. 1998. Winter mortality of young-of-the-year Hudson River striped bass (Morone saxatilis): size-dependent patterns and effects on recruitment. Can. J. Fish. Aquat. Sci. 55:1122-1130. 2001. Activity-related constraints on overwintering young- of-the-year striped bass (Morone saxatilis). Can. J. Zool. 79:129-136. Hurst, T. P, E. T. Schultz, and D. O. Conover. 2000. Seasonal energy dynamics of young-of-the-year Hudson River striped bass. Trans. Am. Fish. Soc. 129:145-157. Jensen, J. W., and T. Berg. 1993. Food rations and rate of gastric emptying in brown trout fed pellets. Prog. Fish-Cult. 55:244-249. Juanes, F., and D O. Conover. 1994. Rapid growth, high feeding rates, and early piscivory in young-of-the-year bluefish (Pomatomus saltatrix). Can. J. Fish. Aquat. Sci. 51:1752-1761. Keast, A. 1968. Feeding of some Great Lakes fishes at low tempera- tures. J. Fish. Res. Board Can. 25:1199-1218. Lankford, T. E., and T. E. Targett. 1997. Selective predation by juvenile weakfish: post-con- sumptive constraints on energy maximization and growth. Ecology 75:1049-1061. Markle, D. F.,G. C. Grant. 1970. The summer food habits of young-of-the-year striped bass in three Virginia rivers. Chesapeake Sci. 11:50-54. Markus, H. C. 1932. The extent to which temperature changes influence food consumption in largemouth bass. Trans. Am. Fish. Soc. 62:202-210. Metcalfe, N. B., and J. E. Thorpe. 1992. Anorexia and defended energy levels in over-winter- ing juvenile salmon. J. Anim. Ecol. 61:175-181. Minton, J. W., and R. B. McLean. 1982. Measurement of growth and consumption of sauger (Stizostedion canadense): implications for fish energetics studies. Can. J. Fish. Aquat. Sci. 39:1396-1403. Nelson, G. A., and M. R. Ross. 1995. Gastric evacuation in little skate. J. Fish Biol. 46: 977-986. Parrish, D. L., and F. J. Margraf. 1990. Gastric evacuation estimates of white perch, Morone americana, determined from laboratory and field data. Environ. Biol. Fishes 29:155-158. Post, J. R., and D. O. Evans. 1989. Size-dependent overwinter mortality of young-of-the- year yellow perch (Perea flavescens): laboratory, in situ, and field experiments. Can. J. Fish. Aquat. Sci. 46:1958- 1968. Ristich, S. S., M. Crandall, and J. Fortier. 1977. Benthic and epibenthic macroinvertebrates of the Hudson River. I. Distribution, natural history and commu- nity structure. Estuar. Coast. Mar. Sci. 5:255-266. Robichaud-LeBlanc, K. A., S. C. Courtenay, and J. M. Hanson. 1997. Ontogentetic diet shifts in age-0 striped bass, Morone saxatilis, from the Miramichi River estuary, Gulf of St. Lawrence. Can. J. Zool. 75:1300-1309. Rulifson, R. A., and S. A. McKenna. 1987. Food of striped bass in the upper Bay of Fundy, Canada. Trans. Am. Fish. Soc. 116:119-122. Scharf, F. S., F. Juanes, and M. Sutherland. 1998. Inferring ecological relationships from the edges of scatter diagrams: comparisons of regression techniques. Ecology 79:448-460. Schultz, E. T., and D. O. Conover. 1997. Latitudinal differences in somatic energy storage: adaptive responses to seasonality in an estuarine fish (Atherinidae: Menidia menidia). Oecologia 109:516-529. Singh-Renton, S., and P. J. Bromley. 1996. Effects of temperature, prey type and prey size on gastric evacuation in small cod and whiting. J. Fish Biol. 49:702-713. Smith, R. L., J. M. Paul, and A. J. Paul. 1989. Gastric evacuation in walleye pollock, Theragra chal- cogramma. Can. J. Fish. Aquat. Sci. 46:489-493. Sogard, S. M. 1997. Size-selective mortality in the juvenile stage of tele- ost fishes: a review. Bull. Mar. Sci. 60:1129-1157. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, second ed. W. H. Freeman and Co., New York, NY, 859 p. Steimle, F. W., and J. Caracciolo-Ward. 1989. A reassessment of the status of the benthic macro- fauna of the Raritan Estuary. Estuaries 12:145-156. Thompson, J. T., E. P. Bergersen, C. A. Carlson, and L. R. Kaeding. 1991. Role of size, condition, and lipid content in the over- winter survival of age-0 Colorado squawfish. Trans. Am. Fish. Soc. 120:346-353. Toneys, M. L., and D. W. Coble. 1980. Mortality, hematocrit, osmolality, electrolyte regula- tion, and fat depletion of young-of-the-year freshwater fishes under simulated winter conditions. Can. J. Fish. Aquat. Sci. 37:225-232. Vondracek, B. 1987. Digestion rates and gastric evacuation times in rela- tion to temperature of the Sacramento squawfish, Ptycho- cheilus grandis. Fish. Bull. 85:159-163. 554 Abstract— In 1996, using the contin- uous underway fish egg sampler (CU- FES), we carried out a preliminary study with the daily egg production method (DEPM) for estimating fish biomass of Pacific sardine, Sardinops sagax. Full- water-column abundance of sardine eggs was correlated with the abundance of eggs taken at 3-m depth with a CUFES; however direct conversion of CUFES samples to the full-water-column abun- dance, required by the DEPM, would add considerable variance to an esti- mate of daily egg production. Our pre- liminary study also indicated that the average size of an egg patch for Pacific sardine was 22 km diameter and that all stages of sardine eggs were not equally vulnerable to a CUFES at 3-m depth. Using these findings as a guide, in 1997 we carried out an adaptive alloca- tion DEPM survey in which the num- bers of sardine eggs collected with the CUFES were used to determine sub- sequent locations of vertical net tows. All the vertical net tows were taken in a high-density stratum where the CUFES collected at least two eggs per minute; egg density in this stratum was computed by using only vertical tows. The remaining survey area, where the CUFES collected fewer than two eggs per minute, constituted a low-density stratum where egg density was esti- mated by using the ratio of CUFES egg abundance in these two strata multi- plied by the egg density in the high-den- sity stratum. Conventional statistics were used because the sampling units were survey lines spaced at 22-km or greater intervals. An acceptable level of precision (CV=21%) for a daily pro- duction of 2.57 eggs 0.05 m2/d (51.4 eggs m2/d), was achieved by using only 141 vertical net tows. Therefore, the CUFES will enhance the DEPM by increasing precision of the estimates, or by reducing costs in relation to a survey composed of only vertical water- column tows, when the CUFES is used adaptively to establish sampling strata for vertical water column tows. Manuscript accepted 19 March 2001. Fish. Bull. 99:554-571 (2001). Use of a continuous egg sampler for ichthyoplankton surveys: application to the estimation of daily egg production of Pacific sardine ( Sardinops sagax ) off California Nancy C. H. Lo John R, Hunter Richard Charter Southwest Fisheries Science Center National Marine Fisheries Service, NOAA 8604 La Jolla Shores La Jolla, California 92037 E-mail address (for N C. H. Lo): Nancy.Lo@noaa.gov The continuous underway fish egg sam- pler (CUFES) (Checkley et ah, 1997; van der Lingen et ah, 1998; Checkley et ah, 1999; Watson et ah, 1999) is a new device that provides high-resolu- tion spatial maps of fish eggs by siev- ing the eggs from water pumped from a fixed depth while a survey ship is underway. These data may be used as an index of fish abundance, or to study spawning habitats, and if converted to the numbers of eggs in the full-water- column, they may be used to estimate fish biomass by using one of the egg production methods (Hunter and Lo, 1997). The objective of our study was to evaluate the use of the CUFES in the daily egg production method (DEPM) of estimating the spawning biomass of pelagic fishes (Lasker et ah, 1985; Parker, 1985). In the DEPM, biomass is calculated from the number of staged eggs taken in plankton samples and the daily fecundity of the parents. In a standai'd DEPM estimate, eggs are sampled in vertical net tows, starting from a point below the maximum depth of the eggs (typically 70 m) and end- ing at the surface. These vertical sam- ples are taken within a grid of sta- tions located 4 nmi apart, a sampling interval known to produce uncorrelat- ed samples of anchovy eggs (Smith and Hewitt, 1985). Use of the CUFES in the DEPM has several potential advantag- es over the use of a standard fixed-grid survey. Continuous sampling may in- crease the precision of the biomass es- timate because it provides an increased spatial resolution of egg patches (Hunt- er and Lo, 1997). Continuous sampling may save ship time and thereby reduce the cost of sampling per transect mile, and the increased spatial resolution of egg patches provides new knowledge regarding the spawning behavior of the species. Several potential disadvantag- es of using a CUFES in the DEPM also exist; the gains in precision pro- vided by continuous sampling may be diminished because of the increases in variance due to converting numbers of eggs taken in a CUFES to a full-wa- ter-column abundance; a much more complicated formula for variance of the estimate may be needed because of correlated CUFES samples; and the numbers of staged eggs taken in a CUFES may be biased, either because a CUFES damages eggs, making the staging of them more subject to error or because all stages may not be equal- ly vulnerable to sampling at the 3-m depth. Clearly these issues need to be resolved if a CUFES is to be used in a DEPM survey. We used the following approach to evaluate the use of a CUFES in the estimation of daily egg production. We conducted a pilot CUFES survey in 1996 with the objective of examining the spatial properties of Pacific sardine (Sardinops sagax ) eggs sampled by a CUFES. We determined how the num- bers of eggs taken in a CUFES were related to their abundance in the full- water-column, the spatial correlation of egg samples, and the condition of eggs after passing through a CUFES. Our Lo et al.: Application of the continuous egg sampler to estimation of the daily egg production of Sardinops sogax 555 Table 1 Total number of samples and positive samples of sardine eggs sampled in CUFES and CalVET surveys. The range and mean of duration (minutes) for CUFES collections for 1994, 1996, and 1997 surveys. 1994 (18 Apr— 11 May) 1996 pilot CUFES survey 1997 DEPM survey leg 1 (15-21 Mar) leg 2 (21 Mar-6 Apr) leg 1 (11 Mar-27 Mar) leg 2 (28 Mar-7 Apr) CUFES7 (positive) 1396 905 896 331 (889) (568) (550) (137) Duration mean range (min) — 3.5 18.13 30 26.9 2-5 3-35 1-54 1-34 CalVET1 2 (positive) 684 91 — 141 — (74) (66) — (102) — Survey area (km2) 380,175 157,000 174,196 (253,850 in USA) 1 Total collection in leg 1, 1996, was 1437. The first 41 tows were experimental. 2 CalVET surveys in 1996 and 1997 were taken in the high-density stratum; in 1994 they were taken in a fixed grid over the whole survey area. analysis of the 1996 cruise was used to develop an adap- tive allocation survey design, similar to that proposed by Thompson et al. ( 1992), to estimate the biomass of sardine that incorporates a CUFES into the DEPM. Using the adaptive allocation survey design, we conducted a DEPM survey for Pacific sardine with a CUFES in 1997, and the precision of our estimate of daily production of eggs was compared with the precision of the estimate provided by a DEPM survey carried out in 1994 under standard meth- ods (Lo et al., 1996). We also considered the accuracy of shipboard counts of eggs collected in a CUFES in relation to counts of preserved samples made after cruise end. We present our results in chronological order, beginning with those from the 1996 pilot survey which provided the ratio- nal for the new survey design in 1997; we next describe the new survey design, and end with a comparison of the 1997 CUFES and DEPM survey with the conventional DEPM survey in 1994. Survey data and spatial models Survey data The data used in our study were taken from three ichthyo- plankton surveys (Table 1); 1 A pilot CUFES cruise in 1996. This cruise consisted of leg 1 (during which both a CUFES was used and full- water-column tows were taken ) and leg 2 (during which only a CUFES was used to survey the large geographic area of spawning sardine (Fig. 1); 2 A DEPM survey in 1997. This survey for Pacific sardine employed a new survey design with the CUFES and the California vertical tow (CalVET, see below) (Smith et al., 1985) (Fig. 2). The allocation of CalVETs was deter- mined by the egg density observed from the CUFES. 3 Results of the 1994 DEPM survey off California, U.S., and Baja California, Mexico (Lo et al., 1996). This was a conventional fixed-grid DEPM survey employing only CalVETs. This cruise was used as a standard for com- paring DEPM surveys with and without the CUFES. The CalVET net consisted of 150-pm nylon netting. The diameter of the CalVET net frame was 25 cm; the tow was lowered to a depth of 70 m and was retrieved vertically. The CUFES was installed midship on the NOAA vessel Dcwid Starr Jordan onto the intake pipe over the side of the vessel; it extended 3 m below the water surface (see il- lustration in Checkley et al., 1997). Eggs were sieved from the water flow with the 500-pm nylon mesh of the CUFES concentrator. The density of eggs taken in the CalVET net was ex- pressed as the number of eggs/0.05 m2 of sea surface wa- ter, a standard procedure in the DEPM, where 0.05 m2 is the area of the mouth opening of the CalVET net. All eggs taken in the CalVET samples, regardless of survey, were counted and staged in the laboratory. The density of eggs taken in the CUFES was expressed as the number of eggs taken per minute. The interval over which eggs accumu- lated in CUFES samples varied depending on their abun- dance. When abundance was low, samples were collected over intervals of 0.5 h equivalent to 7.4 km or 4 nmi on the transect line; when abundance was high, they were col- lected over intervals of at least 1 minute (0.24 km or 0.13 nmi). All CUFES samples were counted at sea, preserved, and recounted in the laboratory. All eggs taken by the CUFES in the pilot survey, but not the subsequent DEPM survey, were staged in the laboratory. We used the system detailed in Lo et al. ( 1996) and grouped eggs by their ages into half-day age classes for egg mortality computation and one-day age classes for spatial statistical analysis. In our study, we considered the estimates of only the daily production of eggs, P0, one of the key parameters in 556 Fishery Bulletin 99(4) 118° 00' 117° 30‘ 117° 00' Figure 1 Sardine eggs/minute from CUFES samples and survey pattern in cruise 9603. Leg 1: 13-21 March 1996 (A) and leg 2 (B): 6-21 March 1996. The star on the lower right corner of section B is the reference point (0,0) to compute vario- grams (see text). Lo et al.: Application of the continuous egg sampler to estimation of the daily egg production of Sardinops sagax 557 124“ 122“ 120“ 118“ 116“ 38" N 36“ 34“ 32“ ■ i i 124“ 122“ 120” 118“ 116“W Figure 2 Sardine eggs/minute from CUFES samples and survey pattern in cruise 9703, 11 March-6 April 1997, with two strata: a high-density stratum (open area) and a low-density stratum (shaded area). the DEPM. P0 is only one of six parameters used in esti- mating biomass with the DEPM. In the DEPM, biomass is related to egg production using the model Bs = P^A /(R/ Wf) SF, where Bs - biomass for area A, P0- the daily egg pro- duction at age 0 day per unit sea surface area; Wf - the average female weight; S = the fraction of females spawn- ing per day; F =batch fecundity; and R - the fraction of the biomass that is female (Alheit, 1993; Hunter and Lo, 1997). The denominator, (R/W^)SF, is the number of eggs/ biomass in grams and is also called the daily specific fe- cundity. In the 1994 survey all parameters were estimated but in the 1997 survey we estimated only P0 and A and used historical data for the other parameters (Hill et al.1). Variograms of sardine density (eggs/minute) As the first step in developing a DEPM survey design for the CUFES, we used geostatistical techniques (Cressie, 1 Hill, Kevin, T. M. Yaremko, L. D. Jacobson, N. C. H. Lo, and D. A. Hannan. 1998. Stock assessment and management recommendations for Pacific sardine ( Sardinops sagax ) in 1997. Marine region. Admin. Rep. 98-5, Cal. Dep. of Fish and Game, 8604 La Jolla Shores Dr., La Jolla, CA 92037. 1991; Barange and Hampton, 1997; and Fletcher and Sumner, 1999) to describe the spatial structure of the egg distribution and estimated the major diameters of sardines egg patches from the 1996 survey. We applied variogram models (Cressie, 1991; Petitgas, 1993) to CUFE samples (in units of eggs/minute) grouped into three age groups: 1-day (4-27 h), 2-day (28-51 h), and 3-day (52-75 h). Variogram (y(h)) is defined as the variance of difference between values that are h units apart and is a function of variance and covariance: 2y(h) = var[w(x) - u(x + h )] = 2(var(u) - co v(/i)) if var[w(x)] = var[«(x + h)], where u(x) = the eggs/min at location x; u(x+h) = the eggs/min at the location h (nmi) away from x; var(u) = the variance; and co v(/i) = the covariance of eggs/min that are h (nmi) apart. y(h) is the semivariogram. For simplicity, we refer to y(h) as the variogram. We used S+SpatialStats (Kaluzny et al., 558 Fishery Bulletin 99(4) 1996) and EVA software (Petitgas and Prampant2) to ana- lyze and visualize spatial distributions of sardine eggs/ min and to compute a variogram for each of the three age groups of sardine eggs. Two basic assumptions of the variogram (Eq. 1) are intrinsic stationarity and isotropy. Intrinsic stationarity means that a constant mean exists and the variance of egg density is defined by the mag- nitude of h. Isotropy means that spatial correlation and the range of correlation do not change with direction. The variogram is normally expressed as a function of three parameters: range, sill, and nugget effect. The range is the distance beyond which the observations are not corre- lated. The sill is the variance of the random field and is the asymptotic value of y(h) . The nugget effect measures the micromeasurement error and the white noise for h close to 0 (Cressie, 1991). Ideally, for h close to 0, the variogram ( y(h )) will be close to zero because the observations tend to be similar. As h increases, the observations become h units apart and tend to be different, or co v(h) decreases, and the variogram in- creases. At a certain distance, h*, cov(h) approaches zero, and for h >h*, the variogram approaches its asymptote (sill). The distance, h*, is the range. The range (h*) was estimated from a model that best fits the data and was used to estimate the diameter of the patch of sardine eggs because eggs whose distances are less than h* nmi are correlated and thus are likely to be in the same patch. Conversely, eggs that are more than h* nmi apart are no longer correlated and thus are assumed to be in different patches. We chose the robust (or stable) estimator of the variogram (Cressie and Howkins, 1980; Cressie, 1991): gram: range, sill and nugget effect. The range was then used as the estimate of the diameter of the patch for each of the three age groups and total number of eggs. Results of the 1996 pilot CUFES survey Spatial correlation and patch size of sardine eggs For each of the three age groups and the total number of eggs, the four-directional variograms of the residuals of ln(eggs/min+l) from LOESS (Fig. 3) indicated that the variograms for transects at 0°, 45°, and 135° were clearer than the cross-transect (90°), particularly for the 1-day-old eggs, 2-day-old eggs, and total egg category. The variogram in the direction of within-transect (0 degree) had the clear- est signal because intervals between adjacent collections were the shortest. Because the variograms for each direction looked similar, we used the variogram in the within-transect direction to assess the spatial correlation of sardine eggs. The spherical model was chosen to fit the variogram (Cressie, 1991): y(h;6) -c0+cs (3 / 2)(||/r||/as) - (1/ 2)(||/z||/as)' c0 + cs, h = 0 0 < INI < as (3) No- where c0 = the nugget effect; cs = the practical sill (variance - nugget); as = the range; and II h || = Euclidean distance. T l4 | u(x) - u(x + h)\V~ 2 ~(/l) — _N(h) * ~ [0.457 + 0.494/|Afi/i)|]Afi/i)4 ’ where |M/j)| = the number of distinctive pairs; and h = the distance (nmi) between any two locations. To avoid possible trends, we first ran a local regression model (LOESS) of ln(eggs/min+l) against line distance (the distance computed from the survey lines, y-axis) and sta- tion distance fic-axis) (Chamber and Hastie, 1992) where the reference point (0,0) is a pseudo station in Mexico (sta- tion 260 on CalCOFI line 980; Fig. IB). We then constructed a variogram for the residuals from the local regression model (LOESS) in four directions clockwise from the tran- sect (0° ,45°, 90°, and 135°) to examine possible aniso- trophy. We chose natural logarithm (In) transformed data because the distribution of eggs was skewed. Finally, we used an interactive S+ function (model. variogram function) to determine the estimates of parameters for each vario- 2 Petitgas, P., and A. Prampant. 1993. EVA (estimation vari- ance), a geostatistical software on IBM-PC for structure charac- terization and variance computation. CM1993/D:65, 81st meeting of ICES. IFREMER laboratoire d’Ecologie Halieutique, BP 21105,44311 Nantes (France), Cedex 03, 33 p. For total number of eggs (all eggs combined), the range (2) of the residuals of ln(eggs/min+l) was 22.2 km (12 nmi), sill (variance in the random field) was 0.4, and nugget was 0.05. For 1-day-old eggs, the range was 14.8 km (8 nmi) and the sill was 0.3. For 2-day-old eggs, the range was 18.5 km (10 nmi), the sill was 0.15, and nugget was 0.005, and for 3-day-old eggs, range was 22.2 km ( 12 nmi), the sill was 0.065, and nugget was 0.005 (Fig. 4). Because the maxi- mum range was 22.2 km (12 nmi), eggs collected more than 22.2 km (12 nmi) apart were considered uncorrelat- ed; therefore, transect lines spaced intervals of 12 nmi or greater were considered independent. The data also indi- cated the gradual dispersion of egg sardine patches with time as described by Smith (1973) because the diameter of sardine egg patches increased from 14.8 km for 4-27 h old eggs to 22.2 km for eggs 52-75 h old. Conversion of CUFES egg density to full-water- column abundance and distribution of egg stages Egg counts from 91 paired samples collected with the CUFES and CalVETs during leg 1 of the 1996 survey (Table 1) were used to derive a conversion factor from eggs/minute of CUFES sample to CalVET catch (R). We used a regression estimator to compute the ratio of eggs/ minute from the CUFES to eggs/tows from CalVETs, R = pv/px, where y = eggs/minute; x = eggs/tow; and R = the catch ratio. The estimator of R is R = Z(x x y )/Z(x2). Lo et al.: Application of the continuous egg sampler to estimation of the daily egg production of Sardinops sagax 559 Distance (nmi, km) Distance (nmi, km) 0.25 0.20 0.15 o.io 0.05 0.0 0 6 0.4 0.2 0.0 Figure 3 Variogram of residuals of In (sardine e.ggs/minute+1) for age groups: 4-27 h (A), 28-51 h (B), and 52-75 h (C), and total eggs for four directions (D). Degree 0 is the direction perpendicular to the coastal line, degree 90 is the direction along the coastal line during leg 2 of cruise 9603. On the x-axis, the inner ticks are in km and the outer ticks are in nmi. Paired samples were taken wherever high abundance of eggs appeared in samples collected with the CUFES. We obtained egg/minute = 0.73 eggs/0. 05m2 (CV=0.16) (Fig. 5). This means that for one egg observed from a Cal- VET, one would expect to see, on average, 0.73 eggs/min. Or for one egg/min from a CUFES, one would expect to see 1.5 eggs/tow. A striking difference existed between the data from the 1996 pilot survey and the full survey car- ried out in 1997. In 1997, the catch ratio of eggs/minute to eggs/tows was 0.25 (CV=0.08) from 110 pairs of CalVET and CUFES of which at least one sample was positive (Fig. 5). This means that one egg/tow from a CalVET tow was equivalent to approximately 0.25 egg/min from a CUFES, or one egg/minute from the CUFES was equiva- lent to 4 eggs/tow from a CalVET sample. The ephemeral nature of such conversion coefficients was not known to us when we developed the design for the 1997 survey. How- ever, the variance associated with the direct 1996 conver- sions was a strong incentive to reduce the effects of direct conversions in the design of the 1997 survey and in the calculation of biomass. To determine if the CUFES provides an unbiased sam- ple of all sardine eggs stages, we compared the distribu- tions of developmental stages between the two samplers taken in 91 paired CUFES and CalVET samples during leg 1 of the 1996 survey (Fig. 1). A 2 x 11 contingency ta- ble was constructed for total counts of each of 11 stages of eggs collected with the CalVET and the CUFES. A chi- square statistic was computed to test the null hypothesis that the distribution of stages was independent of the samplers. The chi-square (%2) analysis showed that the distribu- tion of stages was not the same between the two samplers (%2=188.47, df =10, P-value <0.01) (Table 2). The differ- ence was primarily due to eggs of stages I, III, V, and VI. The CUFES caught only two stage-I eggs; therefore we de- cided to run a %2 test with stage-I and stage-II eggs col- lapsed into one group. A similar conclusion was reached for the later case i = 1,2, and A, - the area size. ' Aj + Aj Simulation Bootstrap simulations were conducted to provide the pos- sible biases and another estimate of the standard error of daily egg production (P0) and the instantaneous mortality rate (z) for each stratum and the entire survey area under the adaptive allocation sampling scheme. As mentioned ear- lier, CalVETs were taken on nine transect lines and not on line 8 and line 11. In the simulation, nine transects with CalVETs were sampled with replacement and estimation procedures described in previous section were followed. To evaluate the effect of weighting, we included weighted and unweighted nonlinear regression where the weight was the inverse of the standard error of egg production of each age group and yolksac larvae. One thousand iterations were run, and the standard deviation of 1000 estimates was the bootstrapped standard error of the estimates. Bias was the difference between the average of 1000 estimates and the estimate from the original data. The bias-corrected esti- mate was the original estimate minus the bias. ■^0,2 - ^0,1 9’ (6) Results of the 1997 CUFES and DEPM survey Zjm' where ml = the total CUFES time (minutes) in the zth transect; and Xjx - was eggs/min in the jth stratum and ith tran- sect. The variance of q was computed according to that of a ratio estimator (Eq. 4). Daily egg production for the total survey area ( P0 ) Pn was computed as a weighted average of P0 1 and P0 2, where D _ P0,A + ^0,2^2 A, + A-, (8) — P0 jlfj + P02w2 = P01[i<-\+qw.2] Daily egg production The daily egg production for each half-day category and yolksac larval production and their ages (d) were used to construct an embryonic mortality curve for the high-den- sity stratum (Eq. 5, Fig. 8, Table 3). The daily egg produc- tion in the high-density stratum (P0 j) based on unweighted nonlinear regression was 5.04 eggs/0.05 m2/d (100.8 eggs/ m2/d ,CV=0.25) and egg mortality was z=0.21 (CV=0.73) for an area (Aj) of 66,841 km2 (19,530 nmi2) (Eq. 8, Fig. 9). The ratio (q) of egg density between the low- density stratum and high-density stratum from CUFES samples was 0.211 (CV=0.43) (Eq. 7). Therefore, in the low-density stratum, the egg production (P02) was 1.064 eggs/0.05 m2/d (21.28 eggs/m2/d, CV=0.49) for an area (A2) of 107,255 km2 (31,338 nmi2). The estimate of the daily egg production for the entire survey area was 2.57/0.05 m2 (51.4/m2, CV=0.27) (Eq. 8, Table 4). The weighted nonlinear regression produced estimates slightly different from those with unweighted nonlinear regression: P0 x =4.76/0.05 m2 (59.2/m2, CV=0.18), z = 0.35(CV=0.14), and the P0 for the entire survey area was 2.43/0.05 m2 (48.6/m2, CV=0.21). The bootstrap estimate of P0 (5.10) was similar to the original estimate P0 (5.04) in the high-density stratum. The standard error ofP0 (1.6) from the bootstrap analysis was higher than the estimate from the original data (1.28). The bias of P0 (0.06) was negligible because the ratio 564 Fishery Bulletin 99(4) of bias to standard error was less than 0.25 (Efron and Tibshirani, 1993). However, the daily instantaneous mor- tality rate (2=0.29) from the bootstrap was higher than 0.21 from the original data. The bias-corrected 2 was 0.13 (SE=0.12). For the entire survey area, the bootstrapped P0 was 2.60/0.05 m2 (SE=0.71; CV=0.27), similar to the origi- nal estimate (2.57/0.05 m2). The bootstrap results indicated that the weighted non- linear regression produced a downward biased estimate ofP0 (bias=-1.26=3.50-4.76) in the high-density stratum. The standard error (1.60) of P0 from bootstrap was also higher than that computed from the original data (0.86). The fact that the ratio of bias to standard error (1.26:1.60) was greater than 0.25 may indicate that the weighted non- Lo et al.: Application of the continuous egg sampler to estimation of the daily egg production of Sordinops sagax 565 Table 4 Estimates of daily egg production (P0\ number of eggs/0.05 m'2/day at age 0), daily instantaneous mortality rate (z) for the high- density stratum (66,841 km2) and low-density stratum ( 107,255 km2), PQ for the entire survey area ( 174,096 km2), the catch ratio of eggs/min in the low-density stratum to eggs/min in the high-density stratum ( q ), their standard errors (SE), estimated bias, and the bias-corrected estimates for the unweighted and weighted nonlinear regression from 1000 iterations of bootstrap simulation. High-density stratum Low-density stratum Entire survey area P0,l SE (P01 ) 2 SE (z) P -*0,2 SE (PQ 2) P0 SE (P0) Q SE ( q ) Unweighted nonlinear regression survey 5.04 1.28 -0.21 0.16 1.06 0.52 2.57 0.72 0.211 0.09 bootstrap mean 5.10 1.60 -0.29 0.28 1.10 0.53 2.60 0.85 0.22 0.08 SE 1.60 0.72 0.12 0.11 0.48 0.28 0.71 0.34 0.09 0.03 CV 0.31 0.46 -0.40 0.38 0.45 0.52 0.27 0.41 0.41 0.40 bias 0.06 0.32 -0.08 0.12 0.03 0.01 0.03 0.13 0.009 -0.01 bias corrected 4.98 0.96 -0.13 0.04 1.04 0.51 2.54 0.58 0.202 0.10 Weighted nonlinear regression survey 4.76 0.86 -0.35 0.050 1.004 0.45 2.43 0.51 0.211 0.09 bootstrap mean 3.50 0.72 -0.27 0.056 0.730 0.33 1.80 0.41 0.22 0.08 SE 1.60 0.35 0.10 0.025 0.400 0.21 0.76 0.20 0.09 0.03 CV 0.46 0.49 -0.36 0.440 0.550 0.63 0.42 0.48 0.42 0.40 bias -1.26 -0.14 0.08 0.006 -0.274 -0.12 -0.63 -0.10 0.009 0.01 bias corrected 6.02 1.00 -0.43 0.044 1.278 0.57 3.06 0.61 0.202 0.10 linear regression was unwarranted in our case. One pos- sible reason is that the standard errors (SE) of egg produc- tion in each age group were between 1.3 and 2.8, whereas the SE for yolksac larvae was 0.16 (Table 3). Although the weighted regression should be used when the varianc- es of data are unequal, we believe that there is too large a disparity between variance of eggs and yolksac larvae, and too much weight was assigned to yolksac larvae for a weighted nonlinear regression. The spawning biomass was computed from the daily egg production from the 1997 survey, and the historical daily specific fecundity (number of eggs/gram of biomass) was 23.55 (Macewicz et al., 1996) assuming the daily spe- cific fecundity was the same as that for 1994-96. Sardine spawning biomass in 1997 would be 379,940 t3 for an area of 174,096 km2 (50,868 nmi2) from San Diego to San Fran- cisco. No variance of Bs was computed because no variance of the number of eggs per population weight (g)/day was available. Comparison of results from the 1997 CUFES and DEPM survey with results from a conventional DEPM survey We compared the results of the 1997 CUFES and DEPM design with those of a conventional DEPM ( 1994), in which only CalVET samples were taken, to illustrate how the 3 The spawning biomass would be 359,280 metric tons if P0 of 2.43/0.05 m2 from weighted nonlinear regression was used. CUFES allocation design may affect the performance of a DEPM survey of Pacific sardine. We believe the compari- son is instructive, even though the two surveys differed somewhat in area and population size; the conventional 1994 survey covered a larger area than the 1997 survey (380,175 km2 vs. 174,096 km2), and, the total biomass of sardine was smaller in 1994 than in 1997 (111,493 t (Lo et. al., 1996) vs. 379,940 t, respectively). In both sur- veys staged eggs and yolksac larvae from CalVET samples were used in the calculation of P0. An obvious difference in the results of the two surveys was that only 11% (74/684) of CalVET samples were pos- itive for sardine eggs in the 1994 conventional survey, whereas in the 1997 survey 72% (102/141) were positive (Table 5). These results indicate that CUFES was effec- tive in allocating CalVET samples and thereby reducing ship-time costs per survey mile. The coefficients of varia- tion (CV) for the estimates of P0 were similar: 0.22 for the conventional survey compared with 0.27 for the CUFES- based DEPM. Thus, variance penalty for using the ratio estimator q did not greatly diminish the benefit in using the CUFES in the DEPM. This simple statistical compari- son, however, does not reveal the greatest potential bene- fits in using a CUFES. The allocation of CalVETs would be most useful when the population is at a lower level, as it was in 1994, because at such levels one must cover a large survey area to assure an unbiased estimate, but the popu- lation is probably concentrated in a very small fraction of the area where CalVET samples will be allocated. In addi- tion, the high-resolution maps of the spatial distribution 566 Fishery Bulletin 99(4) Table 5 Sardine daily egg production (PQ) from a conventional survey (1994), compared with a CUFES and DEPM survey (1997). 1994 Conventional DEPM survey 1997 CUFES and DEPM survey Area: 380,175 km2 Area: 174,096 km2 CalVET samples CalVET samples total 684 total 141 positive for eggs 72 positive for eggs 102 positive percent 11% percent positive 72% CUFES samples None CUFES samples total 1227 total positive 687 percent positive 56% high-density stratum 84% low-density stratum 40% Daily egg production Daily egg production 0.169/0.05 m2 Po 2.57/0.05 m2 CV 0.22 CV 0.27 spawning biomass 111,493 t spawning biomass 379,940 t of eggs provided by the CUFES have not as yet been incor- porated into the DEPM design, but we believe in the long- term it will be possible to improve the accuracy of surveys by doing so, as well as possible to develope new insights into the processes involved in selection of spawning habi- tats by parent fish. Comparison of shipboard egg counts with preserved egg counts A key element of the allocation design was that the allo- cation of CalVET samples was based on near real-time counts and on the identification of eggs by CUFES opera- tors. After having been counted by CUFES operators at sea, the eggs were preserved in vials, and later recounted and identified in the laboratory by experts. We compared egg counts in the laboratory with those taken at sea from the 1997 survey to determine the reliability of shipboard counting and identification and to indicate the extent to which a difference affected the final estimate of the daily egg production. Within each stratum, we first computed an overall mean eggs/min for the laboratory and ship count by using a ra- tio estimator (where y=total number of eggs and x=the to- tal minutes for each transect summed over all transects) (Table 6). We also computed a ratio of eggs/min from the laboratory to that from the ship for each transect line and obtained a weighted ratio for each stratum, where weight was the duration for each transect within a stratum (Eq. 7). The ratio for the entire survey area was a weighted av- erage of two ratios, one from each stratum, and weight was the area of each stratum, as in Equation 5. In the high-density stratum, the ratio from the labora- tory counts to the ship counts was 1.19:1 (CV=0.03). In the low-density stratum, the ratio from the laboratory to the ship was 1.22:1 (CV=0.05) (Table 6). The overall ratio for the entire survey area was 1.20:1 (CV=0.10); therefore the laboratory count was higher than the ship count by 20% (Table 6). Of a total 1227 collections, 687 were positive according to laboratory counts. There were 16 collections where ship counts were positive but counts in the laboratory were zero. Eggs of other species in those 16 CUFES collections were obviously misidentified as sardine eggs. (Table 6). A total of 658 pairs had positive counts, out of which 130 pairs had equal positive counts. A total of 524 pairs had zero counts. Therefore there were 654 equal counts between laboratory and ship (524 zeros and 130 positive counts) and 573 pairs ( = 1227-654) with a mismatch. The relationship between the total counts from the laboratory and the ship confirmed the undercount from the ship (Fig. 9). The variance of under- counts increased with the total counts from the laboratory. One outlier was a collection where the ship count was zero and the laboratory count was 341. Although the absolute undercounts increased with the total number of eggs, the percent of undercount decreased with the total egg count. Conversion coefficient between two strata (g) In the 1997 CUFES and DEPM survey, one of the primary functions of the CUFES collections was to provide a con- version coefficient ( q ) of egg density between two strata to convert the egg production in the high-density stratum (P0 j) to the low-density stratum (P02), be. P02 =P01 x q (Eq. 6). The conversion coefficients ( q ) computed from the laboratory and ship counts were similar: 0.213 (CV=0.44) from the laboratory and 0.211 (CV=0.43) from the ship counts. Therefore, the bias of using egg counts at sea to calibrate egg production per unit area in the low-density stratum would be small: if the ratio of 0.213:1 were used, the egg production for the entire survey area would be 2.60/0.05 m2 instead of 2.57/0.05 m.2 Lo et al.: Application of the continuous egg sampler to estimation of the daily egg production of Sarclinops sagax 567 Table 6 Comparison of egg density and distribution of zero and positive egg counts from CUFES samples with those taken in the laboratory and at sea (ship), 1997. CVs are in shown in parentheses. Egg density (eggs/min) High-density stratum Low-density stratum Overall q ship 4.16 (0.41) 0.47 (0.45) 0.213 (0.44) laboratory 4.91 (0.4) 0.57 (0.45) 0.211 (0.43) ratio 1.19(0.03) 1.22(0.05) 1.20 (0.10) Distribution of zero and positive egg counts Laboratory Ship zero positive total zero 524 29 553 positive 16 6581 674 total 540 687 1,227 1 658 = 130 with exact match + 428 without match. Discussion Presently the most productive use of the CUFES in the DEPM is to use the CUFES as an efficient auxiliary infor- mation provider and the CalVET, or another full-water- column tow, as the primary sampler. Given our present level of knowledge of egg distributions and accuracy of predicting them, it would be folly to depend only on the CUFES for an estimate of P0. Predictive egg-distribution models (Sundby, 1983; Westgard, 1989) are promising and sometime in the future might make exclusive use of the CUFES practical in a DEPM. Exclusive use of the CUFES as a sampler in the DEPM also would require detailed knowledge of the stage-specific vulnerability of eggs to the sampler because a bias seemed to exist for sardine. In short, full-water-column tows are essential for accu- rately estimating egg production today, but the CUFES can greatly facilitate the allocation of such tows. The CUFES enhancement of the basic DEPM design (Lasker, 1985) may have broad application because the DEPM is used world-wide for estimating the spawning biomass of sardines, anchovies, and other species of fishes (Priede and Watson, 1993;Zeldis, 1993; Lo, 1997). Thus, we feel it is useful to discuss some of the new features of the DEPM survey that we developed for estimating P0 with the CUFES. Critical value for allocation sampling To use the CUFES in our DEPM survey design requires selecting an egg density that triggers full-water-column sampling (CalVET sampling). We used a critical value of 2 eggs/min, which was equivalent to 3 or 8 eggs/tow, depending on the conversion factor. If the 1997 conversion factor (eggs/min=0.25 egg/tow) was correct, we could lower the critical value to 1 egg/min, equivalent to 4 eggs/tow. This would create a larger high-density stratum and more CalVET tows would be allocated. This range of critical values (3-8 eggs/tow) was similar to the value (5 eggs/tow) used in a stratified sampling design for an anchovy survey in Biscay Bay in Spain (Petitgas, 1997). As a result, the precision of P0 and z would likely be improved. Increasing the area for the high-density sampling also reduces the potential for bias from the assumption of a constant egg mortality between strata. On the other hand, lowering the critical value diminishes the gain from using the CUFES. Clearly, an optimal critical value exists for each species and survey area. The critical value can be determined prior to the survey or during the survey by using order statistics (Thompson and Seber, 1996; Quinn et al., 1999). The extent that the critical value can be fine tuned to deliver an optimum balance between the CUFES and Cal- VET samples for a particular region, species, and season is unknown. The large difference in catch ratios between our 1996 (0.73) and 1997 (0.25) surveys certainly does not sup- port the idea of fine tuning. These differences may over- state the expected variability for sardine because the ar- eas were different; the 1996 samples were taken over a very limited portion of the survey area , whereas in 1997, the sample pairs were taken where high-density spawning occurred (Figs. 1A, 2, and 6). Interestingly, our 1997 esti- mate (0.25) is similar to that computed by us for sardine off South Africa (van der Lingen et al., 1998). Variability in catch ratios The extent of vertical mixing of eggs is probably the main factor affecting the variation of the catch ratio between surveys. Because the selection of the optimal critical value for CalVET sample allocation depends upon on the rela- tionship between catches in the CUFES and those in the CalVET, or the catch ratio, it seems useful to consider what 568 Fishery Bulletin 99(4) the ratio would be if water were perfectly mixed in the whole-water-column. If the water were perfectly mixed, fish eggs would be distributed randomly and the catch ratio (between CalVET and CUFES) would be a constant, because egg count would be proportional to the volume of water filtered through the two samplers. A CUFES filters on the average 0.64 m3 of water/min and the CalVET fil- ters 3.5 m3 of water. Therefore under perfect mixing, the ratio of eggs/min to eggs/tow = 0.64 m3/3.5 m3 = 0.18 for the daytime when fish schools are in deeper water, assum- ing the vertical distribution of sardine eggs was similar to that of anchovy eggs (Moser and Pommeranz, 1999). At night, when fish schools are in the upper 50 meters,4 the ratio would be 0.64 m3/2.5m3 = 0.26. This means that on the average, for four to six eggs seen in a CalVET catch, we would expect one egg/min in the CUFES. The fact that the estimated ratio for two years were 0.76 in 1996 and 0.25 in 1997, indicates that more sardine eggs appeared in the upper 3 m than would be expected under perfect mixing and less mixing occurred in 1996 than in 1997. Clearly, if all the eggs were in the upper 3 m of water column, the ratio would be 1:1 for the same surface area. As the extent of vertical mixing in the sea is highly variable, we believe that calibration tows are always needed, even if the CUFES is used only as an index of egg abundance. Ver- tical egg mixing models might eventually help to reduce calibration requirements. Identification and counting of eggs at sea Identification and counting of fish eggs while the ship is underway was an essential ingredient of our adaptive allo- cation design. The eggs of some commercially important clupeid species are often difficult to distinguish from those of co-occurring species (Ahlstrom and Moser, 1980; Mata- rese and Sandknop, 1984; Watson and Sandknop, 1996). Further, many melanostomiin species have eggs with char- acteristics (e.g. large diameter, wide perivitelline space, segmented yolk) that are similar to those of co-occurring sardines and other clupeid eggs. During a DEPM survey off Oregon, Bentley et al. (1996) encountered a type of melanostomiin egg similar to the egg of Pacific sardine at 24 of 46 stations. Our experience has shown that the risk of misidentification increases when identifications are made at sea; during our CUFES and DEPM survey in 1997, about 1% of the eggs initially identified as sardines turned out to be melanostomiin eggs after examination in the laboratory. Fortunately, melanostomiin eggs were in such low abundance that they had no effect on the criti- cal values of density. The possibility of misidentification differs depending on season, location, and target species. Clearly all shipboard positive records for areas that lie outside of the known spawning range and season of the 4 Castillo Valderrama, P. R. 1995. Distribucion de los princi- pales recursos pelagicos durante los veranlos de 1992 a 1994. Instituo Del Mar Del Peru, Informe No 114. Instituto del Mar delPeru (IMARPE), Esquina Gamarra Y General Valle Chu- cuito, Callao Peru, 24 p. target species should be checked in the laboratory after the cruise. Owing to the great abundance of sardines and anchovy eggs, the effect of misidentification on biomass estimation is probably trivial if the survey is conducted during peak spawning months. Egg counts on shipboard were somewhat lower than shoreside counts. Even though the effect of the difference is negligible from the standpoint of biomass estimation, we recommend that shoreside measurement be maintained. In particular, for collections that contain a large amount of other organisms (e.g. salps) which makes it difficult to count the eggs on aboard the ship, shoreside counting would be necessary. Stratified design An important feature of the survey design was the strati- fication of sampling by egg density and sampler type. In the high-density stratum, we used only staged eggs from CalVET samples to estimate 2 and P0, whereas in the low- density stratum we collected only CUFES samples. We believe it would not be useful to use staged eggs from CUFES samples in the low-density stratum to estimate z and P0 directly because of the low egg abundance, possibil- ity of stage-specific bias, and lack of yolksac larvae (the CUFES does not sample yolksac larvae well). It seemed preferable to use, for the low-density stratum, the esti- mate of P0 for the high-density stratum, adjusted by the ratio of egg densities taken in CUFES at high and low strata. This, of course, requires the assumption that egg mortality did not differ between strata. A direct test of this assumption is impractical because of the large sam- pling effort needed at low density to obtain a sufficient number of positive samples to detect a difference in mor- tality rates. Fortunately, the effect of this potential bias is diminished because the low-density stratum contributes fewer eggs. In our example, the low-density stratum con- tributed about 25%5 of the daily production. An alterna- tive approach would be to allocate CalVET sampling to the low-density stratum. We believe this approach would not be cost effective because the number of positive CalVETs would be so low. Another important element of the stratified design was the use of transect lines as the sampling unit. Model-based geostatistics are needed (Fletcher and Sumner, 1999) for data from continuous samplers such as echo-sounders and CUFESs, unless sampling units are defined such that da- ta are uncorrelated among sampling units in the survey design (Armstrong et al., 1988). Because conventional de- sign-based statistical procedures are easier to apply, we preferred using a transect line as our sampling unit, which requires that the minimum sample size allows a between- transect distance greater than the diameter of the egg patch. Fortunately the within-transect CUFES collections provided the information needed on the spatial structure to determine the distance of CalVET lines to insure samples are uncorrelated. In our case, tows, a minimum of 22.2 km 5 25% =100 x (1.064x107, 255)/(2.57xl74, 096) (Table 4). Lo et al.: Application of the continuous egg sampler to estimation of the daily egg production of Sardinops sagax 569 (12 nmi) apart, would ensure uncorrelated samples and provide the unbiased estimates of abundance of eggs in each development stage of sardine eggs. Use of yolksac larvae Although yolksac larvae have been used to estimate P0 with the DEPM method (Lo, 1985; 1986; Lo et al., 1996; Hunter and Lo, 1997), but are not a requirement for using the CUFES, we included yolksac larvae to estimate P0 because the mortality rate of eggs and yolksac larvae are similar for northern anchovy (Lo, 1985, 1986) and because the development of early stages of anchovy and sardine is similar (Ahlstrom, 1943; Zweifel and Lasker, 1976; Moser and Alhstrom, 1985; Lo et al., 1996). Our threshold for taking CalVET samples, 2 eggs per minute, generated 102 positive CalVETs (72% were positive). This number is far fewer positive tows than the number needed to be assured a significant slope for the regression of numbers of eggs on their ages. In anchovy, where the eggs are less patchy than sardine, about 500 positive tows were required to assure a CV of the estimate of mortality rate close to 0.6 (Lo, 1997). To obtain a significant slope in the 1997 survey, we used the number and average age of yolksac larvae (adjusted for observed mean temperature) as well as staged eggs. This introduced a potential bias, because by doing this we assumed that yolksac larvae have the same mortality rate as eggs. To avoid this bias, the number of staged egg sam- ples should be increased. One approach would be to stage the eggs taken with a CUFES in the high-density stra- tum and combine them with the CalVET samples from the same stratum. This would greatly increase the number of staged eggs available but requires the assumption that CUFES staged eggs are an unbiased sample of the full- water-column, a condition not met in our 1996 pilot study. Another approach would be to increase the number of CalVET samples per mile in the high-density stratum from the present one sample per 4 nmi to a higher fre- quency. A doubling of the CalVET sampling rate would significantly increase survey costs and may not increase the number of positive samples sufficiently to be able to use only eggs for estimation of P0. Considering the relative risks and costs of these approaches, we feel that the use of the yolksac larvae in the estimation of P0 was prefer- able. In other species, the distribution of eggs may be less concentrated than they are for sardine and sampling at the 4 nmi sampling rate may be adequate. Clearly, other solutions may exist, and we recommend considering these issues with each new application. Conclusions We conclude that the CUFES is a useful tool with the DEPM when it is used adaptively to establish sampling strata for CalVETs. In this mode, the CUFES increases precision and reduces cost per transect mile. Perhaps one of the major benefits of using a CUFES in the DEPM is that one can better afford to expand survey boundaries and thereby reduce the potential bias of not enclosing the entire population, a very common bias in field surveys in general (Gunderson, 1993). Clearly, the more contiguous the distribution of spawned eggs within the surveyed hab- itat, the greater the benefits in using the CUFES. Acknowledgments We thank three unanimous reviewers for their comments; David Griffith, Ron Dotson, and Amy Hays for operating the CUFES for 1996 and 1997; David Ambrose, William Watson, and Elaine Acuna for counting and staging eggs of CUFES and CalVET samples; Geoffrey Moser for provid- ing the information on misidentification of sardine eggs; and crew members of NOAA RV David Starr Jordan for their cooperation. Literature cited Alheit, J. 1993. Use of the daily egg production method for estimating biomass of clupeoid fishes, a review and evaluation. Bull. Mar. Sci. 53(2)750-767. Ahlstrom, E. H. 1943. Studies on the Pacific pilchard or sardine ( Sardinpos caerulea ). 4. Influence of temperature on the rate of devel- opment of pilchard eggs in nature. U.S. Fish. Wild], Serv., Spec. Sci. Rep. 23, 26 p. All 1 strom, E. H., and H. G. Moser. 1980. Characters useful in identification of pelagic marine fish eggs. Calif. Coop. Oceanic Fish. Investig. Rep. 21:121- 131. Armstrong, M., P. Shelton, I. Hampton, G. Jolly, and Y. Melo. 1988. Egg production estimates of anchovy biomass in the southern Benguela. Calif. Coop. Oceanic. Fish. Investig. Rep. 29:137-157. Barange, M., and I. Hampton. 1997. Spatial structure of co-occuring anchovy and sardine pop- ulations from acoustic data: implications for survey design. Fish. Oceangr. 6:294-108. Bentley. P. J., R. L. Emmett, N. C. H. Lo, and H. G. Moser. 1996. Egg production of Pacific sardine ( Sardinops sagax ) off Oregon in 1994. Calif. Coop. Oceanic. Fish. Investig. Rep. 37:193-200. Chambers, J. M., and T. J. Hastie (ed). 1992. Statistical models in S. Wadsworth and Brooks/Cole Advanced Books and Software, Pacific Grove, CA, 608 p. Checkley, D. M., Jr, R. C. Dotson, and D. A. Griffith. 1999. Continuous, Underway sampling of eggs of northern anchovy (Engraulis mordax ) and Pacific sardin e (Sardinops sagax ) in spring 1996 and 1997 off southern and central California. Deep-Sea Res. (special rep.) II 47(2000) 1139- 1155. Checkley, D. M., Jr., P. B. Ortner, L. R. Settle, L.R., and S. R. Cummings. 1997. A continuous, underway fish egg sampler. Fish. Oceanogr. 6:58-73. Cressie, N. 1991. Statistics for spatial data. John Wiley and Sons, Inc. New York, Chichester, Toronto, Brisbane, Singapore, 900 p. Cressie, N., and D. M. Howkins. 1980. Robust estimation of the variogram. I. J. Int. Assoc. Math. Geology 12:115-125. 570 Fishery Bulletin 99(4) Efron, B., and R, J, Tibshirani. 1993. An introduction to the bootstrap. Monographs on sta- tistics and applied probability 57. Chapman and Hall, New York, London, 436 p. Fletcher, W. J., and N. R. Sumner. 1999. Spatial distribution of sardine ( Sardinops sagax) eggs and larvae: an application of geostatistics and resampling to survey data. Can. J. Aquat. Sci. 56:907-914. Gunderson. D. R. 1993. Surveys of fisheries resources. John Wiley and Sons, Inc., New York, Chichester, Brisbane, Toronto, Singapore, 248 p. Goodman, L. A. 1960. On the exact variance of products. J. Am. Stat. Assoc. 55(N292):708-713. Hunter, J. R., and N. C. H. Lo. 1997. The daily egg production methods of biomass esti- mation: some problems and potential improvements. OZEANOGRAFIKA, Boletin De La Sociedad De oceano- grafica De Gipuzkoa, Medalla De Oro De La Ciudad De San Sebastian 2:9-40. Kaluzny, S. P., C. Vega, T. P. Cardoso, and A. A. Shelly. 1996. S+ spatialstats. User’s manual, version 1. Mathsoft, Inc., Seattle, WA, 226 p. Lasker, R (ed). 1985. An egg production method for estimating spawning biomass of pelagic fish: applicat ion to the northern anchovy (Engraulis mordax). U.S. Dep. Commer., NOAATech. Rep. NMFS 36, 99 p. Lo, N. C. H. 1985. Egg production of the central stock off northern anchovy, Engraulis mordax , 1951-82. Fish. Bull. 83(2)137- 150. 1986. Modeling life-stage-specific instantaneous mortality rates, an application to northern anchovy, Engraulis mor- dax, eggs and larvae. Fish. Bull. 84(2):395-406. 1997. Empirical analyses of the precision of regression estimates of daily egg production and egg mortality of pelagic fish for daily egg production method. OZEANO- GRAFIKA, Boletin De La Sociedad De Oceanografica De Gipuzkoa, Medalla De Oro De La Ciudad De San Sebastian 2:71-89. Lo, N. C. H., Y. A. Geen Ruiz, M. Jaacob Cervantes, H. G. Moser, and R. J. Lynn. 1996. Egg production and spawning biomass of Pacific sar- dine ( Sardinops sagax ) in 1994, determined by the daily egg production method. Calif. Coop. Oceanic. Fish. Inves- tig. Rep. 37:160-174. Macewicz, B. J., J. J. Castro Gonzalez, C. E. Cotero Altamirano, and J. R. Hunter, 1996. Adult reproductive parameters of Pacific sardine, sar- dinops sagax, during 1994. Calif. Coop. Oceanic. Fish. Investig. Rep. 37:140-151. Matarese, A. C., and E. M. Sandknop. 1984. Identification of fish eggs. In Ontogeny and system- atics of fishes (H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L. Richardson, eds.), p. 27-31. Am. Soc. Ichthyol. Herpetol. Spec. Publ. 1. Moser, H. G., and E. H. Ahlstrom. 1985. Staging anchovy eggs. In An egg production method for estimating spawning biomass of pelagic fish: applica- tion to the northern anchovy, Engraulis mordax (R. Lasker, ed.), p. 37-42. U.S. Dep. Commer., NOAA Technical Report NMFS 36. Moser, H. G., and T. Pommeranz. 1999. Vertical distribution of eggs and larvae of northern anchovy, Engraulis mordax, and of the larvae of associated fishes at two sites in the Southern California Bight. Fish. Bull. 97:920-943. Parker, K. 1985. Biomass model for the egg production method. In An egg production method for estimating spawning biomass of pelagic fish: application to the northern anchovy, Engraulis mordax (R. Lasker, ed.), p. 5-6. U.S. Dep. Commer., NOAA Technical Rep. NMFS 36. Petitgas, P. 1993. Geostatistics for fish stock assessments: a review and an acoustic application. ICES J.Mar. Sci. 50:285-298. 1997. Use of disjunctive Kriging to analyse an adaptive survey design for anchovy, Engraulis encraicolus, eggs in Biscay Bay. OZEANOGRAFIKA, Boletin De La Sociedad De Oceanografica De Gipuzkoa, Medalla De Oro De La Ciudad De San Sebastian 2:121-132. Priede, I. G., and J. J. Watson. 1993. An evaluation of the daily egg production method for estimating biomass of Atlantic mackerel ( Scomber scom- brus). Bull. Mar. Sci. 53(2)891-911. Quinn, T .J., II, D. Hanselman, D. Clausen, C. Lunsford, J. Heifetz. 1999. Adaptive cluster sampling of rockfish populations: proceedings of the American Statistical Association, 1999 Joint Statistical Meeting, Biometrics section. Am. Stat. Assoc., Alexandria, VA. Smith, P. E. 1973. The mortality and dispersal of sardine eggs and larvae. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 164:282- 292. Smith, P. E., and R. P. Hewitt. 1985. Anchovy egg dispersal and mortality as interred from close interval observations. Calif. Coop. Oceanic. Fish. Investig. Rep. 26:97-112. Smith, P. E., W. Flerx, and R. P. Hewitt. 1985. The CalCOFI vertical egg tow (CalVET). In An egg production method for estimating spawning spawning bio- mass of pelagic fish: application to the northern anchovy (Engraulis mordax ) (R. Lasker, ed.), p. 27-32. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 36. Sundby, B. 1983. A one-dimensional model for the vertical distribution of pelagic fish eggs in the mixed layer. Deep-Sea Res., 30: 645-661. Thompson, S. K., F. L. Ramsey, and G. A. F. Seber. 1992. An adaptive procedure for sampling animal popula- tions. Biometrics 48:1195-1199. Thompson, S. K., and G. A. F. Seber. 1996. Adaptive sampling. John Wiley and Sons, Inc., New York, Chichester, Brisbane, Toronto, Singapore, 265 p. van der Lingen, C. D., D. Checkley Jr, M. Barange, L. Hutchings, and K. Osgood. 1998. Assessing the abundance and distribution of eggs of sardine, Sardinops sagas, and round herring, Etru- menus whiteheadi, on the western Agulhas Bank, South Africa, using a continuous, underway fish egg sampler. Fish. Oceanogr. 7( 1 ):35— 47. Watson, W., and E. M. Sandknop. 1996. Clupeidae. In The early stages of fishes in the Cali- fornia Current region (H. G. Moser, ed.), 159-171 p. Calif. Coop. Oceanic Fish. Invest. Atlas 33. Watson, W., R. L. Charter, H. G. Moser, R. D. Vetter, D. A. Ambrose, S. R. Charter, L. I. Robertson, E. M. Sandknop, E. A. Lynn, and J. Stannard. 1999. Fine-scale distributions of planktonic fish eggs in the Lo et al.: Application of the continuous egg sampler to estimation of the daily egg production of Sardinops sagax 571 vicinities of Big Sycamore Canyon and Vandenberg Ecolog- ical Reserves, and Anacapa and San Miguel Island, Califor- nia. Calif. Coop. Oceanic. Fish. Investig. Rep. 40:128-153. Westgard, T. 1989. Two models of the vertical distribution of pelagic fish egg in the turbulent layer of the ocean. Rapp. R-V. Reun. Cons. Int. Explor. Mer 191:195-200. Zeldis, J. R. 1993. Applicability to egg surveys for spawning-stock bio- mass estimation of snapper, orange roughy, and hoki in New Zealand. Bull. Mar. Sci.:53(2) 864-890. Zweifel, J. R., and R. Lasker. 1976. Prehatch and posthatch growth of fishes — a general model. Fish. Bull. 74(31:609-621. 572 Growth of juvenile red king crab ( Paralithodes camtschaticus ) in Bristol Bay (Alaska) elucidated from field sampling and analysis of trawl-survey data Timothy Loher David A. Armstrong School of Fisheries and Aquatic Sciences Box 355020 University of Washington Seattle, Washington 98195 Present address (for T Loher): International Pacific Halibut Commission P.O. Box 95009 Seattle, Washington 98145-2009 E-mail address (for T. Loher): loher@u washington.edu Bradley G. Stevens AFSC Kodiak Laboratory Kodiak Fisheries Research Center National Marine Fisheries Service 301 Research Court Kodiak, Alaska 99615 Abstract— An analysis of in situ growth rate was conducted for juvenile red king crab ( Paralithodes camtschat- icus) in Bristol Bay, Alaska. Growth of early juveniles (-2-40 mm CL; age 0-3 yr) was determined by fitting sea- sonalized Gompertz growth models to length-frequency data. The parame- ters of the growth model and resulting size-at-age estimates were compared with those from studies conducted at Unalaska and Kodiak Islands by fitting the same growth model to published length-frequency data from separate sources. Growth of late juvenile and early reproductive crabs, -30-100 mm carapace length (CL), was examined by analyzing length-frequency data from the National Marine Fisheries Service annual Bering Sea trawl survey from 1975 through 1999. Mean CL associ- ated with strong size modes of crabs in Bristol Bay length-frequency distribu- tions was resolved by using the FiSAT software package (FAO-ICLARM Stock Assessment Tools) to track the modal size progression of strong year classes and assign mean size-at-age character- istics to the stock. Growth of early juvenile crabs was slower in Bristol Bay than that observed by other researchers at Unalaska or Kodiak. Sizes at 1, 2, and 3 years after settlement were estimated to be -9 mm, 23 mm, and 47 mm CL in Bristol Bay compared with 16 mm, 38 mm, and 66 mm CL at Unalaska; at Kodiak, estimated sizes of 12 mm and 42 mm were obtained for age-1 and age-2 crabs, respectively. Within the Bristol Bay trawl survey data, a total of 24 modes were identified for both males and females <-100 mm CL, which included the modal progression of two year classes that presumably settled in 1976 and 1990. The 1976 year class grew slowly and would not have recruited to the reproductive stock until -9 years after settlement, whereas the 1990 year class appeared to recruit at -8 years after settlement. Both esti- mates indicate that Bristol Bay red king crabs are older at reproductive maturity than the -6 years after settle- ment presently assumed. An attempt to resolve discrete mean size-at-age from the length-frequency data met with little success because variability in growth between year classes mark- edly obscured size-at-age characteris- tics in the stock. Manuscript accepted 20 March 2001. Fish. Bull. 99:572-587 (2001). Population abundance of red king crab (Paralithodes camtschaticus ) in Bristol Bay, Alaska (Fig. 1) is typified by great variability. The maximum abundance of harvestable male crabs in the stock over the last 25 years has fluctuated by over an order of magnitude, peak- ing at nearly 60 million individuals in 1977, and has fallen to less than 3 mil- lion from 1983 to 1985 (Loher et al., 1998; Zheng and Kruse1). Strong fish- eries in the late 1970s and early 1980s were followed by substantially reduced abundance in recent years, and fishery closures in 1981 and 1994, leading to concern over the status of the popula- tion by both management agencies and fishermen. Such concern prompted the Alaska Department of Fish and Game to develop a detailed harvest-based recovery plan for the fishery (Zheng et al.2) that considers the stock to be fully recovered once it reaches an effective spawning biomass of 55 million pounds. This biomass has not occurred in Bris- tol Bay since 1981 (Zheng and Kruse1; Zheng et al.2), but it was chosen as an appropriate rebuilding level primarily on the basis of length-based recruit- ment models that, in turn, rely on the inferred underlying stock-recruitment relationship for the population (Zheng et al., 1995a, 1995b). Present assumptions regarding red king crab growth rates suggest that Bristol Bay crabs recruit to the re- productive stock at an age of approx- imately 6 years after settlement (i.e. -seven years following egg fertilization; Zheng et al., 1995a, 1995b), which cor- responds to mean sizes of 105 mm car- apace length (CL) and 97 mm CL for males and females, respectively. These size-at-age values are based on Ste- vens and Munk (1990), Weber (1967), and Balsiger (1974). However, none of these sources represents a comprehen- sive treatment of growth from settle- ment through maturity based entirely on field-collected data from the Bristol Bay region. Balsiger’s (1974) growth 1 Zheng, J., and G. H. Kruse. 1999. Status of king crab stocks in the eastern Bering Sea in 1999. Alaska Department of Fish and Game, Reg. Inf. Rep. 5J99-09. Division of Commercial Fisheries, Alaska Depart- ment of Fish and Game, RO. Box 25526, Juneau, Alaska, 99801. 2 Zheng, J., M. C. Murphy, and G. H. Kruse. 996. Overview of population estimation methods and recommended harvest strat- egy for red king crabs in Bristol Bay. Alaska Department of Fish and Game, Reg. Inf. Rep. 5J96-04. Division of Com- mercial Fisheries, Alaska Department of Fish and Game, P.O. Box 25526, Juneau, Alaska, 99801. Loher et al.: Growth of Paralithodes ccimtschaticus 573 model was derived from data from Bristol Bay, but it con- sidered only individuals >81 mm carapace length, well above the size at which multiple yearly molts are expect- ed. Weber’s (1967) work was based on juvenile crabs col- lected more than 40 years ago at Unalaska Island (Fig. 1) in the eastern Aleutian Islands, which is located west of Unimak Pass and the coastal shelf break and which is oceanographically separated from Bristol Bay by the southern origins of the Bering Slope current. Stevens and Munk’s (1990) growth model was based on data from the Kodiak region (Fig. 1); a separate growth model was de- veloped by Stevens (1990) to consider the eastern Bering Sea, but it relies on Weber’s (1967) findings regarding the growth of early juvenile crabs. Accurate growth rate information is important to prop- erly calibrate the length-based recruitment model and also to determine appropriate time lags between spawning and subsequent recruitment. Given the likelihood of temporal and geographic variability in growth rates, it is likely that the Bristol Bay stock exhibits growth rates different from those observed at different locations by the aforemen- tioned researchers. Environmentally induced changes in molt schedule resulting in considerable variability in size- at-age is a common feature in Crustacea (e.g. Hartnoll, 1982; Hill et al., 1989; Huner and Romaire, 1990; Wain- right and Armstrong, 1993; Tremblay and Eagles, 1997) and probably also in red king crab (Stevens, 1990; Stevens and Munk, 1990). Thus, further analysis of the growth of prerecruit crabs within Bristol Bay is needed. In our study, we analyzed growth of early juvenile red king crabs in Bristol Bay by fitting growth equations to length-fre- quency data collected between 1983 and 1991. We com- pared the inferred growth rate for Bristol Bay with that determined from identical models fitted to Weber’s (1967) data, and with two sets of length-frequency data avail- able for the Kodiak region (Dew, 1990; Donaldson et al., 1992) to assess whether application of these growth rates to Bristol Bay crabs is appropriate. We then expanded our analyses to include older prerecruit crabs in Bristol Bay by analyzing length-frequency distributions from 25 years of southeast Bering Sea trawl survey data in order to iden- tify growth patterns associated with strong year classes and in order to elucidate size-at-age characteristics dis- played by the population. Materials and methods Figure 1 Map of western Alaska, showing the locations of Bristol Bay, Unalaska Island, and Kodiak Island. results of Weber (1967), who collected early juvenile red king crabs at Unalaska Island in 1958 and 1959, as well as data collected in the Kodiak region from 1987 to 1989, originally published in Dew ( 1990), and from 1990 to 1991, originally published in Donaldson et al. (1992). Note that, in the present paper, all ages and year-class designations are referenced to approximate settlement date. That is, they refer to the postsettlement age of benthic crabs and do not include the larval phase or the egg incubation period. This postsettlement age should be taken into con- sideration for applications in which the year when eggs were extruded is important. Bristol Bay Waters of the Bering Sea east of 163. 5°W longitude and south of 59°N latitude will be considered “Bristol Bay.” This area is larger than the area that is often referred to as Bristol Bay; it is more typically con- sidered the “southeast Bering Sea.” However, because we also report information from Unalaska Island, also in the southeastern Bering Sea, we choose to make a distinction between Bristol Bay and Unalaska Island in order to avoid confusion. In 1983, surveys of juvenile red king crab abun- dance were conducted throughout the Bristol Bay region (Fig. 2A) with surface-deployed try-net otter trawls and a rock dredge, during three sampling periods: 18 April-7 May, 2—17 June, and 9-23 September (McMurray et al.3). Growth of early juvenile red king crab Growth of early juvenile crabs, from settlement through approximately 3 years after settlement, was examined by using catch data from targeted sampling in Bristol Bay obtained from three sources: 1) work conducted in 1983 under the auspices of the Outer Continental Shelf Assessment Program (OCSEAP) (documented in McMur- ray et al.3), 2) OCSEAP work conducted in 1985 (previ- ously unpublished), and, 3) work conducted in 1991 by the National Marine Fisheries Service (NMFS; documented in Stevens and Macintosh4). In addition we reviewed the 3 McMurray, G., A. H. Vogel, P. A. Fishman, D. A. Armstrong, and S. C. Jewett. 1984. Distribution of larval and juvenile red king crab ( Paralithodes camtschatica) in Bristol Bay. U.S. Dep. Commer., NOAA, OCSEAP Final Report 53( 19861:267-477, Anchorage, Alaska. [Available from D.A. Armstrong at: School of Fisheries and Aquatic Sciences, Univ. Washington, Box 355020, Seattle, WA 98195.] 4 Stevens, B. G., and R. A. Macintosh. 1991. Cruise 91-1 Ocean Hope 3: 1991 eastern Bering Sea juvenile red king crab survey, May 24-June 3, 1991. U.S. Dep. Commer., NOAA, NMFS, AFSC, RACE. Seattle, Washington. [Available from B. G. Ste- vens at AFSC Kodiak Laboratory, National Marine Fisheries Service, 301 Research Court, Kodiak, Alaska, 99615.] 574 Fishery Bulletin 99(4) Location of survey stations in Bristol Bay (east of 163.5° west longitude) visited during 1983 OCSEAP studies (A), 1985 OCSEAP studies (B, open squares), and by Stevens and Macintosh in 1991 (B, closed circles). Try nets had a headrope length of 5.4 m and either wooden or aluminum doors measuring 0.4 m wide x 0.9 m tall. The rock dredge was constructed of a rigid steel frame with a mouth opening 0.9 m wide x 0.4 m tall. The 1985 survey was conducted from 19 July to 2 August, nearshore (gen- erally inside the 50-m isobath), along the north Aleutian Shelf from Unimak Pass through the Port Moller region (Fig. 2B) also with try net and rock dredge. The 1991 survey (Stevens and Macintosh4) was conducted from 26 May to 1 June along the North Aleutian Shelf from Port Moller to Kvichak Bay (Fig. 2B) with a 3.1-m wide beam trawl. The carapace length (CL) of all red king crab cap- tured in the above surveys was measured and recorded onboard the vessels. For each sampling period, length-frequency histograms were constructed to identify individual age classes (co- horts) within the data. Male and female crabs were pooled because sex-specific growth rates are not apparent until reproductive age (Weber, 1967; Dew, 1990), and the com- bined length-frequency data were analyzed by using the FiSAT software package (Gayanilo and Pauly, 1997) to de- termine the mean size (±1 SD) for each identifiable cohort. FiSAT employs a combination of Bhattacharya’s method (Bhattacharya, 1967) and NORMSEP (Hasselblad, 1966; Pauly and Caddy5) to decompose complex size-frequency distributions into a series of best-fit normal curves that represent each cohort within the data set. In our study, the “mean size” of a cohort refers to the mean (±1 SD) of its associated best-fit normal curve, as determined by FiSAT size-frequency decomposition. 5 Pauly, D., and J. F. Caddy. 1985. A modification of Bhat- tacharya’s method for the analysis of mixtures of normal dis- tributions. FAO Fisheries Circular 781, Sales and Marketing Group, Information Division, FAO, Viale delle Terme di Cara- calla, 00100 Rome. Postsettlement age, in Julian days, was then calculated for each cohort. For each year’s length-frequency histo- gram, the cohort with the smallest mean CL was assigned age 0, and the subsequent sizes assigned age 1 and age 2 (Fig. 3). The age of each cohort (in days) was calculated as the time from settlement in the cohort’s settlement year until the median sampling date for the survey period. Set- tlement was estimated to be 15 July of each year because numerous 2-mm-CL individuals occurred in late July dur- ing the 1983 surveys, and a carapace length of 2 mm is typical of the first benthic instar (Kurata, 1961; Donaldson et al., 1992; Loher and Armstrong, 2000). Postsettlement age was plotted against the associated mean CL and two growth curves were fitted to the data. The first curve was a seasonalized version of the von Bertalanffy growth mod- el obtained from Anastacio and Marques (1995), where growth in carapace length is expressed as -[irxZ?xU-qM,()+Cx(iCxZ)/2n)xsin2mxU-ti)]T'^ ^ (-jq The second curve was a seasonalized Gompertz model, where growth in carapace length is expressed as Lt=Lmax[e-e~{K"' where (in both models) Lt = ^ max ~ t = ^ min ~ carapace length at time t; maximum carapace length; given time; time at which the carapace length is the minimum size for the life-stage of interest (in this case, minimum size for benthic red king crab=2 mm); Loher et al.: Growth of Paralithodes camtschaticus 575 Carapace length (mm) Figure 3 Length-frequency histograms of early juvenile red king crabs captured in Bristol Bay from 1983 to 1991. Each age class is denoted by shading: open bars = age 0+; shaded bars = age 1+; closed bars = age 2+. For each sampling date an aver- age time after settlement (ATAS) is reported. The ATAS is the number of days estimated to have elapsed between 15 July of the sampling year (the most recent settlement event) and the median sampling date. For example, the ATAS for crabs sampled on 15 September was 62 days. This value corresponds to the estimated age of age 0+ crabs in that sample, whereas age 1+ crab would have been -365 days older than the ATAS, and age 2+ crabs -730 days older than the ATAS. Note that the June samples fell at nearly one calendar year following the previous year’s settlement; hence, age 0+ crab in those samples were nearly age 1. Arrows indicate the mean CL of each size cohort, as determined with FiSAT; mean values were used to fit the growth curve. ts - lag-time between the start of growth and the first seasonal growth oscillation; oscillations are sinusoi- dal with a one-year period; K = intrinsic growth rate; C = parameter ranging from 0 to 1 that controls the strength of the seasonal growth oscillation; 0 = no seasonal signature; 1 = strong seasonality with a brief period each year during which growth ceases; and D = parameter expressing metabolic deviation from the von Bertalanffy 2/3 metabolic rule (in our study, D = 1 [no deviation]). Unalaska and Kodiak islands Growth rate of early juve- nile red king crab at Unalaska Island was assessed by using data published in Weber ( 1967). In that study imma- ture crabs were collected, primarily with SCUBA, during four sampling periods in 1958 (22 April-17 May; 30 May-6 June; 13 July; 17 September-1 October) and two periods in 1959 (11-24 February; 24 May-2 June). Two data sources were used that originated in Kodiak Island: 1) Donaldson et al. (1992), who documented growth of red king crab in artificial habitat collectors for approximately one year after settlement between June 1990 and May 1991, and; 2) Dew ( 1990) who collected data on podding age 1+ to 2+ red king crab using SCUBA observations between 20 Novem- ber 1987 and 3 June 1989. The data in Weber (1967) and Dew (1990) were used to construct length-frequency his- tograms (males and females pooled) and the characteris- 576 Fishery Bulletin 99(4) Table 1 Points used to construct the growth curve for early juvenile red king crab in Bristol Bay. Data originally presented Macintosh4 are indicated as such; previously unpublished data are listed as “new.” in Stevens and Collection date Source Estimated age (days after settlement) n Size range (mm) Mean size ±1 SD 19 Jul-2 Aug 1985 new 11.5 6 2 2.00 ±0.000 10 Sep-19 Sep 1983 new 62 134 3. 0-6.0 3.97 ±0.581 21 Apr-6 Jun 1983 new 283.5 10 4. 0-6.0 4.27 ±0.730 26 May-1 Jun 1991 Stevens and Macintosh 319.5 35 4-14 7.70 ±1.645 5 Jun-16 Jun 1983 new 332 13 3. 0-7. 2 5.19 ±0.982 18 Jul-2 Aug 1985 new 376.5 3 8. 0-8. 4 8.13 ±0.231 10 Sep-19 Sep 1983 new 427 45 10-28 14.63 ±3.447 21 Apr-6 May 1983 new 648.5 30 10-32 16.70 ±5.389 26 May-1 Jun 1991 Stevens and Macintosh 684.5 16 18-29 22.53 ±3.562 5 Jun-16 Jun 1983 new 697 25 12-26 17.05 ±4.102 26 May-1 Jun 1991 Stevens and Macintosh 1049.5 7 34-47 40.36 ±3.681 tics of individual cohorts were resolved by using FiSAT. Donaldson et al. (1992) did not provide their original data; therefore we simply used the mean size-at-age values that they reported. Mean size-at-age values for each region were then fitted with seasonalized von Bertalanffy and Gompertz curves, as described previously. Age of crabs at Unalaska Island was calculated by using 1 July as the approximate settlement date, as suggested by Weber (1967); 14 June was used for Kodiak Island because Don- aldson et al. (1992) first collected benthic instars on this date. Mean size-at-age of late juvenile through early reproductive-age crabs Growth of late juvenile (-age 2+) through early repro- ductive-age crabs was examined by using data from the annual NMFS groundfish trawl survey from 1975 to 1999 (please refer to Otto [1986] for a description of the annual trawl survey protocol and its spatial coverage). These data were used to construct annual length-frequency his- tograms for the Bristol Bay stock; length-frequency dis- tributions were examined to identify strong size modes, and the mean CL (±1 SD) associated with each mode was determined by using FiSAT. Two approaches were then employed to assign discrete mean size-at-age categories to the population, focusing on size modes with mean CL <100 mm. First, each strong year class that recruited to the population was identified, and the growth of individu- als in these year classes was determined by examination of progression of their length-frequency modes, from the first appearance of the year classes until they could no longer be resolved. Second, the mean CL associated with every strong size mode that appeared in all the annual length-frequency distributions, over the entire time series, was plotted to identify commonly occurring mean CLs that might represent consecutive size-at-age categories. Bottom temperature Area-averaged near-bottom temperature was calculated for each year from 1975 to 1999, to more fully interpret the growth rates observed in Bristol Bay region with respect to variation over time. The area-averaged near-bot- tom temperature represented the mean temperature over all trawl-survey stations located within the Bristol Bay region. Because the survey data were sometimes incom- plete and sampling stations were not located in precisely the same location each year, the temperature at the center of each trawl survey station was statistically interpolated with known values at surrounding points (i.e. “kriged”; see Cressie, 1993). Kriging was performed with the Surfer 6.04 software package (Keckler, 1994) and a linear var- iogram model (Cressie, 1993). Area-averaged near-bottom temperatures for the Bristol Bay region were then cal- culated by using the kriged estimates obtained from the center of trawl survey stations. Results Early juvenile red king crab Resolution of age classes from length-frequency data col- lected in Bristol Bay in 1983, 1985, and 1991 provided 1 1 estimates of mean size-at-age representing individuals ranging from approximately 12 to 1050 days after set- tlement (refer to Fig. 3, Table 1). Weber (1967) provided information resulting in 24 mean size-at-age estimates (Table 2) at Unalaska, for crabs estimated to be -86-964 days after settlement. Donaldson et al. (1992) and Dew (1990) provided 22 estimates of mean size-at-age for crabs -0-720 days after settlement (Table 3) at Kodiak Island. Plots of early juvenile mean size-at-age are pre- sented in Figure 4 for Bristol Bay, Unalaska Island, and Loher et al.: Growth of Parahthodes camtschciticus 577 Table 2 Points used to construct the Unalaska early juvenile red king crab growth curve from data of Weber (1967) (his Appendix Table II). The original data come from crabs that were collected at a number of study sites around Unalaska Island; the “sample area” column in this table refers to different study sites as designated by Weber ( 1967). Refer to the original paper for descriptions of each specific sample area. Sample Estimated age Molt Size range Mean size (mm) Collection date area (days after settlement) status (mm) ±1 SD 17 Sep-1 Oct 1958 2 86 not reported 4-6 4.81 ±0.634 11-24 Feb 1959 2 232.5 postmolt 7-10 8.32 ±0.750 22 Apr-17 May 1958 2 309 not reported 8-17 12.62 ±1.799 24 May-2 Jun 1959 2 333.5 pre- and postmolt 8-16 11.26 ±1.406 30 May-6 Jun 1958 2 338.5 not reported 9-18 12.73 ±1.433 13 Jul 1958 2 379 not reported 13-18 16.06 ±0.869 17 Sep-1 Oct 1958 2 452 not reported 15-34 24.08 ±3.994 17 Sep-1 Oct 1958 4 452 not reported 17-29 20.22 ±1.500 11 -24 Feb 1959 2 598.5 premolt 23-35 29.62 ±2.376 11-24 Feb 1959 2 598.5 postmolt 25-44 35.12 ±3.798 11-24 Feb 1959 4 598.5 pre- and postmolt 22-40 30.56 ±4.200 22 Apr-17 May 1958 1 674 not reported 29-35 31.85 ±1.682 22 Apr-17 May 1958 2 674 not reported 23-48 36.68 ±4.392 24 May-2 Jun 1959 4 698.5 premolt 28-41 29.10 ±2.783 24 May-2 Jun 1959 4 698.5 postmolt 31-41 33.73 ±2.919 24 May-2 Jun 1959 2 698.5 premolt 25-41 35.00 ±2.388 24 May-2 Jun 1959 2 698.5 postmolt 28-51 39.65 ±4.238 30 May-6 Jun 1958 2 703.5 not reported 29-50 39.75 ±4.373 17 Sep-1 Oct 1958 4 817 not reported 41-53 45.86 ±3.308 11-24 Feb 1959 1 963.5 premolt 33-51 40.85 ±3.162 11-24 Feb 1959 1 963.5 postmolt 41-69 49.17 ±4.925 22-24 Feb 1959 3 963.5 premolt 42-74 52.70 ±6.000 11-24 Feb 1959 3 963.5 postmolt 51-82 66.23 ±8.030 11-24 Feb 1959 4 963.5 premolt 43-78 57.19 ±7.659 Kodiak Island data. Seasonalized Gompertz growth functions were fitted to the data; von Bertalanffy growth curves are not presented because their fits were poorer by comparison (Table 4). Model fits and parameter estimates for Gompertz growth functions are reported in Table 5, along with model estimates of CL at ages 1.0 and 2.0 at all sites, and at age 3.0 for Bristol Bay and Unalaska Island; predicted sizes at ages 0.9, 1.9, and 2.9 are also reported because the Bering Sea trawl survey occurs ~0.1 years prior to the settlement season. No data were available for Kodiak Island crabs beyond age ~2; therefore model predictions are not reported beyond this age. Intrinsic growth rate was lowest in Bristol Bay and was coupled with a very strong sea- sonal signature, with no growth expected dur- ing mid-winter. The growth rate observed at Kodiak Island was higher than in Bristol Bay, with a moderate seasonal signature, whereas Estimated age (years after settlement) Figure 4 Seasonalized Gompertz growth curves fitted to mean size-at-age data from Bristol Bay, Unalaska Island, and Kodiak Island. Note that no data were available for Kodiak crabs > ~2 years after settlement. 578 Fishery Bulletin 99(4) Table 3 Points used to construct the Kodiak early juvenile red king crab growth curve. Data from Donaldson et al. (1992) come directly from their Table 1; data from Dew (1990) are compiled from his Figures 7 and 8. Collection date Estimated age (days after settlement) Size range (mm) Mean size (mm: ±SD Donaldson et al. (1992) 14 Jun 1990 0 1.9-2. 4 2.18 ±0.118 28 Jun 1990 14 1.8-2. 6 2.18 ±0.171 13 Jul 1990 29 2. 0-3. 3 2.61 ±0.363 26 Jul 1990 42 2. 5-3. 4 2.84 ±0.152 10 Aug 1990 57 2. 4-4. 2 3.64 ±0.408 23 Aug 1990 70 2. 6-5. 4 3.78 ±0.362 7 Sep 1990 85 3.4-5. 1 4.47 ±0.421 21 Sep 1990 109 4. 2-6. 4 4.98 ±0.605 16 Oct 1990 124 3.4-6. 1 5.29 ±0.491 4 Dec 1990 173 4. 6-7. 9 6.58 ±0.703 11 Feb 1991 212 4. 6-9.0 7.71 ±1.201 27 Mar 1991 257 6.1-10.3 7.95 ±0.769 14 May 1991 335 8.7-13.0 10.46 ±1.079 29 May 1991 350 8. 6-9. 9 9.33 ±0.474 Dew (1990) 6 Oct 1988 480 17.5-34.0 25.67 ±2.446 18 Nov 1988 523 19.0-35.5 27.87 ±2.795 20 Nov 1987 525 22.0-32.5 26.93 ±2.572 18 Dec 1987 553 22.0-27.4 28.10 ±2.726 24 Feb 1989 621 25.0-43.0 35.27 ±3.693 8 Mar 1988 641 22.0-41.5 34.09 ±3.688 19 Apr 1988 675 26.5-44.5 36.49 ±3.176 3 Jun 1988 720 31.0-49.0 40.18 ±3.361 growth rate observed at Unalaska Island was most similar to that at Kodiak Island, and very little seasonal signature could be detected. Estimated mean size at age 2 at Un- alaska Island (37.6 mm CL) was much closer to the value estimated for Kodiak Island (42.2 mm CL) than for Bris- tol Bay (22.7 mm CL, Table 5); at age 3, Bristol Bay crabs were expected to average ~20 mm smaller than crabs ob- served at Unalaska Island. It is important to note that the choice of settlement dates for the models had negli- gible effect on size-at-age estimates, except where such changes caused cohorts to be re-assigned to younger age classes than seemed reasonable (i.e. if early spring settle- ment dates had been chosen, very small crabs would have been assigned to late age 0, instead of being considered immediately postsettlement crabs). Changes in settlement date primarily affected estimated size-at-age over the first few months but had little impact on estimation of size at 1, 2, or 3 years after settlement. Late juvenile through early reproductive-age crabs Male red king crabs collected in NMFS trawl surveys from 1975 to 1999 ranged from 6 to 201 mm CL, and all Table 4 Comparison of coefficient of determination (r2) goodness- of-fit values associated with seasonalized Gompertz and von Bertalanffy growth curves fitted to early juvenile red king crab size-at-age data. Region Gompertz von Bertalanffy Bristol Bay 0.968 0.842 Unalaska Island 0.922 0.863 Kodiak Island 0.995 0.919 sizes between 23 and 198 mm CL were observed. Female red king crabs ranged from 7 to 192 mm CL; all sizes between 24 and 165 mm CL. Over the entire time series, two particularly strong year classes recruited to the Bris- tol Bay population whose growth and size-at-age char- acteristics could be tracked by modal size progression. The first year class was evident from 1979 to 1984, first appearing as size modes with mean CL = ~32 mm and Loher et al.: Growth of Pcrolithodes camtschoticus 579 ~30 mm for males and females, respectively (Fig. 5, Table 6); these crabs were most likely ~2.9 years of age and rep- resented the year class that settled in 1976. Males of this year class grew to a mean CL = ~88 mm by 1983 and the mode became indistinguishable in 1984; females grew to a mean CL = ~84 mm by 1983, and -92 mm by 1984 (Fig. 5, Table 6). The second strong year class was evident from 1994 to 1999 (Fig. 6, Table 6) and first appeared at larger sizes than did the 1976 year class. Males first appeared at a mean CL = -54 mm and grew to a mean CL = -137 mm in 1999; females first appeared at -53 mm CL in 1994 and grew to -109 mm in 1999. This year class first appeared at sizes that were essentially equivalent to the sizes that individuals from the 1976 year class had attained at age 3.9. Thus, this second year class most likely represented crabs that settled in 1990. For both year classes, growth of males and females was similar up to -85 mm CL, after which females grew more slowly than males. However, the growth rates of the two year classes were not equivalent: the 1976 year class grew slower than the 1990 year class (Fig. 7) be- cause the 1976 year class required two years (from 1980 to 1982) to progress from a mean CL = -50 mm to mean CL = -70 mm, whereas the 1990 year class achieved this level of growth within a single year (from 1994 to 1995). Mean sizes during the following two years of growth (from 1982 to 1984 for the 1976 year class; 1995 to 1997 for the 1990 year class, Table 6) were similar between year classes. Within all of the annual length-frequency distributions, considering only size modes with mean CL <100 mm, 24 modes were identified for both males and for females. These included the modes presented previously for the 1976 and 1990 year classes, and an additional 16 modes for males, and 14 modes for females. The additional modes represent other year classes whose modal progression Table 5 Characteristics of seasonalized Gompertz growth curves fitted to size-at-age data from Bristol Bay, Unalaska Island, and Kodiak Island. Values are reported for ages 0.9, 1.9, and 2.9 because these ages roughly correspond to trawl survey data: the trawl survey typically occurs in late May, -0.9 years following the previous year’s settlement. For all curves, Lmax = 200 mm CL. See “Materials and methods” section for definitions of model parameters. Model Bristol Unalaska Kodiak parameters Bay Island Island r2 (fit) 0.968 0.922 0.995 K 0.415 0.421 0.634 C 1.000 0.275 0.599 ts (year) 0.553 0.535 0.680 Kun -3.922 -3.286 -2.659 Length estimates from the models age 0.0 3.1 3.6 2.7 age 0.9 7.0 14.5 9.9 age 1.0 8.5 16.4 11.7 age 1.9 19.5 34.5 38.0 age 2.0 22.7 37.6 42.2 age 2.9 41.9 62.7 — age 3.0 46.7 66.4 — Slowest growth in January January February could not be tracked for a substantial period of time. The mean CL (±1 SD) of all 24 male and female inodes is presented in Figure 8, plotted sequentially by increas- Summary of mean FiSAT. size (±1 SD) of the cohorts depicted in Table 6 size-frequency progression plots (Figs. 5 and 6), as determined with Year class Year Male mean size (mm) ±1 SD Female mean size (mm) ±1 SD Age (years after settlement) 1976 1979 31.9 ±2.79 29.6 ±2.04 2.9 1980 50.7 ±4.10 50.2 ±3.40 3.9 1981 63.2 ±5.06 64.0 ±5.73 4.9 1982 70.1 ±7.50 71.4 ±8.20 5.9 1983 88.4 ±14.74 83.5 ±11.23 6.9 1984 cohort indistinct 92.3 ±4.84 7.9 1990 1994 54.1 ±4.67 52.7 ±3.75 3.9 1995 73.4 ±4.90 71.9 ±7.28 4.9 1996 86.5 ±8.01 83.5 ±5.89 5.9 1997 104.2 ±8.44 97.4 ±3.62 6.9 1998 117.8 ±11.40 105.3 ±8.01 7.9 1999 136.5 ±11.92 109.3 ±7.79 8.9 580 Fishery Bulletin 99(4) Males Females CD n E Carapace length (mm) Figure 5 Yearly length-frequency histograms for male (left) and female (right) Bristol Bay region red king crabs <130 mm CL, 1979-1984, revealing mean size-at-age of the 1976 year class by modal progression. Arrows indicate the cohort’s mean CL each year, as determined with FiSAT; dark shading indicates CLs within 1 SD of the mean. The vertical lines located at 105 mm CL and 97 mm CL indicate size at reproductive recruitment for males and females, respectively, as defined by Zheng et al. (1995a, 1995b). Note that the 1979 histograms have been truncated at n = 80 observations in order to make the relevant size modes more visible. The 1984 male size mode was not resolved with FiSAT; we considered this size mode to be too indistinguishable from the surrounding data to yield an accurate result. ing mean CL. This figure represents an attempt to iden- tify commonly occurring sizes that may represent the ex- pected CLs of consecutive age classes. For male crabs, a cluster of three observations occurred at ~35 mm CL, separated from the remaining observations, and likely representing a single age class; for females, two observa- tions occurred at the same approximate size, separated from the remaining modes. Considering the estimated age Loher et al.: Growth of Paralithodes camtschaticus 581 Males Females Carapace length (mm) Figure 6 Yearly length-frequency histograms for male (left) and female (right) Bristol Bay region red king crabs <155 mm CL, 1994-1999, revealing mean size-at-age of the 1990 year class with modal progression. Arrows indicate the cohort’s mean CL each year, as determined with FiSAT, and dark shading indicates CLs within 1 SD of that mean. The vertical lines located at 105 mm CL and 97 mm CL indicate size-at-reproductive recruitment for males and females, respectively, as defined by Zheng et al. (1995a, 1995b). at size 2.9 of ~40 mm obtained from the growth curve pre- sented earlier (Fig. 4, Table 5), these small crab are likely age 2.9. At larger sizes, there was considerable overlap between modes, suggesting that the mean CL of older age classes is quite variable and difficult to distinguish among years. Bottom temperature From 1975 to 1999, area-averaged near-bottom tempera- tures in Bristol Bay ranged between a low of 0.7°C in 1976 and a high of 5.2 °C in 1981 (Fig. 9). In general, the coldest temperatures were observed in the 1970s, warmest tern- 582 Fishery Bulletin 99(4) Males Females 1976 year class 1976 year class 1990 year class 1990 year class 40 60 80 100 120 40 60 80 100 120 Carapace length (mm) Figure 7 Best-fit normal curves, as determined with FiSAT, of the length-frequency modes associated with male and female red king crabs of the 1976 (solid curves) and 1990 (dashed curves) year classes, from age 3.9 to 6.9, demonstrating the divergence in size-at-age between the two year classes. In particular, note the large difference in apparent molt increment dis- played by the year classes between ages 3.9 and 4.9. All curves have been standardized to a uniform height in order to facilitate their comparison, because the number of crabs that were observed differed between years. peratures during the early 1980s, and moderate tempera- tures from the mid-1980s through the late 1990s. Discussion Our analyses demonstrate that Bristol Bay red king crab grow slower than previously assumed. The present stock- recruitment relationship used to manage this population is based on growth models (Weber, 1967; Balsiger, 1974; Ste- vens and Munk, 1990) that suggest a time lag of six years between settlement and subsequent recruitment (i.e. seven years after fertilization), where reproductive recruitment is defined to occur at ~97 mm and -105 mm CL for females and males, respectively (Zheng et al., 1995a, 1995b). Our results indicate that mean age at full reproductive recruit- ment is likely 8-9 years after settlement. Crabs that settled in 1990 began to reach reproductive size in 1997, at 7 years after settlement, and at this age only -50% of the individu- als were at or above the reproductive size cut-off; the mode did not become fully recruited for an additional 1 to 2 years. Individuals of the 1976 year class grew even slower and, on average, were still slightly smaller than reproductive size at 8 years after settlement. This year class would not have begun to contribute substantially to the reproductive stock until at least 9 years after settlement. The discrepancy between our estimates of mean age- at-recruitment and presently accepted values is due in part to incorrect assumptions of the latter regarding the growth of early juveniles in Bristol Bay. Weber (1967) con- ducted one of the few comprehensive in situ studies of ear- ly juvenile size-at-age in Alaskan waters and clearly dem- onstrated that red king crab should be expected to reach a mean CL = -66 mm CL three years after settlement. This conclusion contrasted sharply with earlier data from Bristol Bay that showed strong size modes with means of 4 mm, 9 mm, and 17 mm CL in early summer samples in 1956, 1957, and 1958, respectively (Fisheries Agency of Japan6), apparently representing the modal progression of the 1956 year class, and suggesting a much slower growth rate. More recent studies conducted in Kodiak (Dew, 1990; Donaldson et al., 1992) have supported Weber’s (1967) conclusions; thus, the observations made by the Fisheries Agency of Japan6 have been largely ignored. Our results suggest that the data from Kodiak is in good accord with Weber’s (1967) conclusions, but that mean CLs of 4 mm, Fisheries Agency of Japan. 1959. Report of research on king crab in the eastern Bering Sea. Int. N. Pac. Fish. Comm. Annu. Rep., p. 71-78. [Available from Secretariat, North Pacific Anad- romous Fish Commission, Suite 502, 889 West Pender Street, Vancouver, British Columbia, V6C 3BC.] Loher et al.: Growth of Paralithodes camtschaticus 583 Modal mean carapace length Figure 8 Mean CL ±1 SD of all strong size modes of Bristol Bay red king crabs appearing in the annual length- frequency data from NMFS Bering Sea trawl surveys, 1975-1999; only modes with mean CL <100 mm are plotted. Size modes are plotted in length-sequential order along the x-axis, starting with the smallest observed mean CL. The numbers adjacent to each mode indicate the year during which that mode was observed. A grouping of modes separate from the remaining data, indicated by the dashed line and the arrow, likely represents a discrete age class; these small crabs are most likely age 2.9 (refer to Fig. 4). The remaining observations are continuous; substantial size overlap between consecutive modes suggest that discrete size-at-age categories are lacking at >~40 mm mean CL. Year Figure 9 Area-averaged June bottom temperatures in the Bristol Bay region from 1975 to 1999. The five-year running averages represent a five-year period that includes the plotted date and the four years preceding it; for example, the five-year average temperature plotted at 1981 (3.0°C) represents mean area-averaged temperature from 1977 to 1981. 584 Fishery Bulletin 99(4) 9 mm, and 17 mm are consistent with sizes at age 0, 1, and 2, respectively, within Bristol Bay. This may be due, in part, to reduced molt frequency in Bristol Bay: the da- ta obtained from Weber (1967) indicate that mid-winter molting was not uncommon during his study, whereas our data suggest that this may not be typical in Bristol Bay. It is difficult to determine whether the discrepancies repre- sent actual regional differences, or simply differences be- tween different studies conducted at different sites and at different times, but our results indicate that it is probably inappropriate to apply the early juvenile growth rate ob- tained by Weber (1967) to the Bristol Bay stock; use of the faster growth rate likely results in lower estimates of age- at-recruitment than the population displays. Variable growth of older juveniles (age 3+) may further delay age-at-recruitment. This variability is evident when comparing growth of the 1976 year class to that of the 1990 year class; age-at-recruitment was 1-2 years greater in the former, due to slow growth of prerecruits. In par- ticular, note that females of the 1976 year class, averaging 83.5 mm CL in 1983, displayed a mean increase in cara- pace length during molting (molt increment [MI]) of ~9 mm during the 1984 spring molt, whereas females of the 1990 year class, also averaging 83.5 mm CL in 1996, ex- hibited a mean MI of ~14 mm during the 1997 spring molt. As a result, females from the 1976 year class required four years to grow from a mean CL = ~50 mm to a mean CL = ~92 mm CL and at ~8 years after settlement were still slightly smaller than the estimated size for full repro- ductive recruitment (Zheng et ah, 1995a, 1995b). Females from the 1990 year class were able to accomplish slightly more mean growth, from ~53 mm to ~97 mm CL, in only three years. The large difference in growth rate between the 1976 and 1990 year classes may have been caused by water temperature differences during the two time periods. Molt schedules and growth rates can be strongly influenced by ambient temperature (Kurata, 1960, 1961; Nakanishi, 1985), and considerable variability in size at maturity has been observed over the species’ geographic range in both males (Paul et al., 1991) and females (see review in Blau, 1990; Otto et al., 1990). Though a number of factors may contribute to the observed variability, reduced growth as- sociated with colder bottom temperatures has been in- voked to explain the smaller size-at-maturity observed in the Norton Sound population as compared with other Ber- ing Sea stocks (Blau, 1990; Otto et al., 1990), and modeling suggests that regional and temporal variation in temper- ature can have broad effects on age-at-recruitment (Ste- vens, 1990; Stevens and Munk, 1991). June bottom tem- perature profiles in Bristol Bay suggest that the 1976 year class was subjected to lower temperatures than the 1990 year class, primarily at early juvenile ages. Although a detailed analysis of temperature-dependent growth would require year-round temperature records, which are not available for this region, June temperatures in Bristol Bay may serve as a proxy for thermal conditions throughout the year. Bottom temperatures in Bristol Bay are linked to seasonal sea ice, that in some years covers much of Bristol Bay (NIC, 1994; Wyllie-Echeverria, 1995; Neibauer, 1998), and the development of sea ice can have strong effects on bottom temperature conditions throughout the year. “Cold pool” bottom waters (<1°C) produced in the winter during ice formation may persist well into the summer, and po- tentially into the following winter, once insulated from surface heating by the development of the summer ther- mocline (Azumaya and Ohtani, 1995). In many Crustacea, temperature primarily affects the molt schedule and has little influence on the magnitude of the MI (Hartnoll, 1982; Wainright and Armstrong, 1993), but laboratory studies conducted with red king crab indi- cate substantial variability in Ml-at-age, across ranges of temperatures, as well as under stable environmental con- ditions. Rearing crabs ~6.3 mm CL under constant tem- peratures of ~10°C, Molyneaux and Shirley (1988) report- ed changes in CL at molt that ranged from -4.4% to 52.2%; similarly, for juvenile premolt crabs 33-36 mm CL, reared at ~5.0°C, Gharrett (1986) observed Mis ranging from 3 to 8 mm. At reproductive age, Weber and Miyahara ( 1962) observed that MI varied between 5 and 23 mm CL per molt in adult males, and large variability in MI associated with water temperatures between 0° and 12°C has been demonstrated for ovigerous females (Shirley et al., 1990). Because the changes in mean CL the we observed for the Bristol Bay stock were determined through modal analysis of the entire population, it is reasonable to suspect that ap- parent differences in growth between years and cohorts do not represent MI variability but may be explained as vari- ability in the number of molted versus unmolted crab with- in particular survey years. This is reasonable to assume, considering that red king crab may skip molting so that the annual molt schedule is replaced by a biennial or triennial cycle (Weber and Miyahara, 1962; McCaughran and Pow- ell, 1977; Balsiger, 1974). However, closer examination of the trawl survey data indicates that, of the 23 size modes of crab observed, none comprised less than 94% new-shelled crabs that had recently molted (senior author, unpubl. da- ta). A high proportion of newly molted crabs was character- istic of nearly all the identifiable size modes (senior author, unpubl. data): of the 53 modes identified, 35 comprised en- tirely new-shelled crabs, 17 comprised 94-99% new-shelled crabs, and only one comprised >10% old-shell individuals (the female cohort with mean CL=~94mm CL in 1982; pop- ulation^ 1.4% old-shell). Thus, the difference in growth rates observed between the two year classes of late juve- niles cannot be explained by variations in molt frequency; the trawl survey data support the conclusion that variabil- ity in MI is a characteristic shared by both sexes across a range of ages. Substantial variability in MI is an important life his- tory characteristic that confounds attempts to assign dis- crete size-at-age categories to a population. The greater the variability in MI among individuals and over time, the greater will be the range of sizes associated with crabs in a given year class, making modes more difficult to resolve from one another. Size-at-age values presently used for Bristol Bay red king crabs were derived from studies consisting of 1-4 years of data (Weber and Miyahara, 1962; Weber, 1967; McCaughran and Powell, 1977; Incze et al., 1986), but our analyses show that a strong tenden- Loher et at: Growth of Paralithodes camtschaticus 585 cy toward specific mean size-at-age is not apparent if lon- ger time scales are considered. Size-at-age characteristics may be different depending on which year class is consid- ered, and among crabs >~40 mm CL, mean CLs of iden- tifiable cohorts displayed a fairly continuous distribution with considerable overlap between adjacent size modes; we could not identify specific mean CLs that could be con- sistently assigned to various age classes. In addition to confounding age estimates, variable MI may cause differ- ent year classes to recruit at different ages, over different time spans, and increase the number of year classes that constitute each year’s new recruitment. These issues have been treated elsewhere with respect to variable intermolt period (Stevens, 1990); variability in MI will produce the same effects. We explored the possibility that overlap between size modes might be attributable to changes in growth rate within the population over time. That is, because bottom temperatures in Bristol Bay were colder during the 1970s than they have been more recently, growth was expected to be slower in the 1970s than later in the time series (Ste- vens, 1990). Thus, one might expect size modes to fall clos- er together early in the time series and to be spaced farther apart later. However, we were unable to resolve a clear tem- poral component in the data; even consecutive year classes sometimes had different mean size-at-age and growth increment characteristics. For example, the 1976 year class displayed a mean CL = ~51 mm (range: 44-57 mm) and ~63 mm CL (range: 56-71mm) in 1980 and 1981, respectively. These represented clear and well-separated size modes at ages 3.9 and 4.9, respectively. However, note the occurrence of a strong mode in the 1979 data, with a mean CL = ~58mm CL (range=50-65 mm); this mode probably represents a single year class, settled in 1975, which would be expected to have mean size-at-age charac- teristics similar to those of the 1976 year class. Yet, this mode of the 1975 year class fell almost precisely between, and its range encompassed the mean sizes of both age 3.9 and age 4.9 crabs from the 1976 year class. Such features were not uncommon in the length frequencies that were presented. For applications requiring accurate size-at-age information, the onerous task of year-by-year and year- class-by-year-class assessments may be necessary. In summary, our results demonstrate that both male and female red king crab in Bristol Bay reach maturity at least one year later than presently assumed (i.e. at ~7 years after settlement) due to slower growth from settle- ment through age 3. Furthermore, variability in MI in late juveniles can result in further reduction in growth rate such that reproductive recruitment is delayed by an ad- ditional 1-2 years (i.e. reproductive recruitment at ~8-9 years after settlement). From a management perspective, variable MI is a life-history characteristic that should be considered in growth- and length-based models of recruit- ment. Present models attempt to simulate MI variability (Zheng et al., 1995a), but little information exists on the magnitude of that variability and its changes over time and space. Such information will be valuable to managers to calibrate recruitment models with respect to lag times between spawning and subsequent recruitment, as well as to predict how year classes enter the spawning population and the fishery; more research in this area is warranted. Inappropriate growth rate and lag-time assumptions have resulted in assigning the wrong year’s spawning stock bio- mass to subsequent recruitment levels in the Bristol Bay stock-recruitment curve; spawning stock abundances are presently offset -1-2 years from the recruitment levels that they generated. This offset may affect the precise shape of the stock-recruitment curve and alter some of the models associated with it. Such changes may prove negli- gible with respect to actual harvest strategies, but it may be prudent to make the appropriate adjustments given knowledge of greater age at reproductive recruitment. The Bristol Bay red king crab stock has been typified by large fluctuations in fishable abundance and by rela- tively rare, strong recruitment pulses generating the bulk of the fishery. Most recently, relatively strong catches from 1997-99, yielding a combined landed catch of -35 million pounds of crabs with an exvessel value estimated at over $137 million (Morrison et al.7), were supported almost en- tirely by the 1990 year class (i.e. were spawned by the 1989 reproductive stock). Former assumptions would have led us to assign this pulse to the 1990 spawning stock and assume that the planktonic larval phase and settlement occurred during 1991. From a management standpoint, the ramifications of such an error may be negligible be- cause estimated effective spawning biomass was similar in both 1989 and 1990 (Zheng and Kruse1). However, as we try to elucidate the mechanisms that generated this strong year class, it is crucial that we accurately determine when those crabs were larvae, early benthic individuals, and later stage crabs. Physical forcing, for example, has been shown to play a large role in determining recruitment vari- ability in a number of commercially important crustacean species worldwide (e.g. Polovina and Mitchum, 1992; Po- lovina et al., 1993; McConnaughey et al., 1994; Rothlisberg et al., 1994; Jones and Epifanio, 1995; McConnaughey and Armstrong, 1995; Rozenkranz et al., 1998). Similar correla- tions between recruitment and physical parameters have been attempted for king crabs (Zheng and Kruse, 2000), but our ability to identify causes of recruitment variabil- ity relies upon associating the correct life history stages with physical forcing events. Even seemingly minor errors in growth-rate assumptions can have serious impacts on our understanding of population dynamics. Acknowledgments Funding for preparation of this manuscript was provided by the University of Washington, Seattle, through a Victor and Tamara Loosanoff Fellowship. We wish to thank Julie 7 Morrison, R., F. Bowers, R. Gish, E. Wilson, W. Jones, and B. Palach. 2000. Annual management report for shellfish fisher- ies of the Bering Sea. In Annual management report for shell- fish fisheries of the westward region, 1999, p. 147-261 Regional Information Report 4KOO-55, Kodiak. Division of Commercial Fisheries, Alaska Department of Fish and Game, RO. Box 25526, Juneau, Alaska, 99801. 586 Fishery Bulletin 99(4) Keister, P. Sean MacDonald, and our anonymous review- ers for thoughtful suggestions that improved the quality of this manuscript, and Steven Hare (International Pacific Halibut Commission) and Gary Walters (NMFS, AFSC) for providing near-bottom temperature data. Literature cited Anastacio, P. M., and J. C. Marques. 1995. Population biology and production of the red swamp crayfish Procambarus clarkii (Girard) in the lower Mendego River Valley, Portugal. J. Crustacean Biol. 15(1):156-168. Azumaya, T., and K. Ohtani. 1995. Effect of winter meteorological conditions on the for- mation of the cold bottom water in the eastern Bering Sea Shelf. J. Oceanogr. 51:665-680. Balsiger, J. W. 1974. A computer simulation model for the eastern Bering Sea king crab population. Ph.D. diss.. School of Fisheries, Univ. Washington, Seattle, WA, 197 p. Bhattacharya, C. G. 1967. A simple method of resolution of a distribution into Gaussian components. Biometrics 23:115-135. Blau, S. F. 1990. Size at maturity of female red king crabs (Paralithodes carntschatica) in the Adak management area, Alaska. In Proceedings of the international symposium on king and Tanner crabs, p. 105-116. Univ. Alaska Sea Grant Report 90-04, Anchorage, AK. Cressie, N. A. C. 1993. Statistics for spatial data. John Wiley and Sons, New York, NY, 900 p. Dew, B. 1990. Behavioral ecology of podding red king crab, Paralith- odes carntschatica. Can. J. Fish. Aquat. Sci. 47:1944-1958. Donaldson, W. E., S. C. Beyersdorfer, D. Pengilly, and S. F. Blau. 1992. Growth of red king crab, Paralithodes camtschaticus (Tilesius, 1815), in artificial habitat collectors at Kodiak, Alaska. J. Shellfish Res. 11(1 ):85— 89. Gayanilo, F. C., Jr., and D. E. Pauly. 1997. FAO-ICLARM stock assessment tools (FiSAT): ref- erence manual. FAO Computerized Information Series (Fisheries) 8, FAO, Rome, 262 p. Gharrett, J. A. 1986. Effects of a hydrocarbon-contaminated diet on sur- vival, feeding, growth, and molting of juvenile red king crab, Paralithodes carntschatica (Tilesius). M.S. thesis. Univ. Alaska, Juneau, AK, 178 p. Hartnoll, R. G. 1982. Growth. In The biology of Crustacea, volume 2: embryology, morphology, and genetics (L. B. Ebele, ed.), p. 1 1 1-196. Academic Press. New York, NY. Hasselblad, V. 1966. Estimation of parameters for a mixture of normal dis- tributions. Technometrics 8:431-444. Hill, J., D. L. Fowler, and M. J. VanDenAvyle 1989. Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (mid- Atlantic): blue crab. Biol. Rep. US Fish Wildl. Serv., 27 p. Huner, J. V., and R. P. Romaire. 1990. Crawfish culture in the southeastern USA. World Aquaculture 21:58-66. Incze, L. S., R. S. Otto, and M. K. McDowell. 1986. Recruitment variability of juvenile red king crab. Paralithodes carntschatica , in the southeast Bering Sea. In North Pacific workshop on stock assessment and manage- ment of invertebrates (G. S. Jamieson, and N. Bourne, eds.), p. 370-378. Can. Spec. Pub. Fish. Aquat. Sci. 92. Jones, M. B., and C. E. Epifanio. 1995. Settlement of brachyuran megalopae in Delaware Bay: an analysis of time series data. Mar. Ecol. Prog. Ser. 125:67-76. Keckler, D. 1994. Surfer for Windows: user’s manual. Golden Soft- ware, Golden, CO, 428 p. Kurata, H. 1960. Studies on the larvae and postlarvae of Paralithodes carntschatica . III. The influence of temperature and salin- ity on growth of the larvae. Bull. Hokkaido Reg. Fish. Res. Lab. 17:9-14. 1961. Studies on the larvae and postlarvae of Paralithodes carntschatica. IV. Growth of the postlarvae. Bull. Hok- kaido Reg. Fish. Res. Lab. 18:1-9. Loher, T., and D. Armstrong. 2000. Effects of habitat complexity and relative larval supply on the establishment of early benthic phase red king crab ( Paralithodes camtschaticus Tilesius, 1815) populations in Auke Bay, Alaska. J. Exp. Mar. Biol. Ecol. 245(1 ):83-109. Loher, T., P. S. Hill, G. A. Harrington, and E. Cassano. 1998. Management of Bristol Bay red king crab: a critical intersections approach to fisheries management. Rev. Fish. Sci. 6(3): 169— 251 . McCaughran, D. A., and G. C. Powell. 1977. Growth model for Alaskan king crab (Paralithodes carntschatica). J. Fish. Res. Board Can. 34: 989-995. McConnaughey, R. A., and D. A. Armstrong. 1995. Potential effects of global climate change on Dunge- ness crab ( Cancer magister) populations in the northeast Pacific Ocean. In Climate change and northern fish popu- lations (R. J. Beamish, ed.), p. 291-306. Can. Spec. Pub. Fish. Aquat. Sci. 121. McConnaughey, R. A., D. A. Armstrong, B. M. Hickey, and D. R. Gunderson. 1994. Interannual variability in coastal Washington Dunge- ness ( Cancer magister) populations, larval advection and the coastal landing strip. Fish. Oceanogr. 3:22-38. Molyneaux, D. B., and T. C. Shirley. 1988. Molting and growth of eyestalk-ablated juvenile red king crabs, Paralithodes carntschatica (Crustacea: Lithodi- dae). Comp. Biochem. Physiol. 91:245-251. Nakanishi, T. 1985. The effect of the environment on the survival rate, growth and respiration of eggs, larvae and post-larvae of king crab (Paralithodes carntschatica). In Proceedings of the international king crab symposium, p. 167-185. Univ. Alaska Sea Grant Report 85-12, Anchorage, AK. Neibauer, H. J. 1998. Variability on Bering Sea ice cover as affected by a regime shift in the north Pacific in the period 1947-1996. J. Geophys. Res. 103(C12):27717-27737. NIC (National Ice Center). 1994. Navy/NOAA National Ice Center (NIC) weekly sea ice concentrations and extents, 1972-1994. National Ice Center, Washington, D.C. [Compact disc.] Otto, R. S. 1986. Management and assessment of Bering sea king crab stocks. In North Pacific workshop on stock assessment and management of invertebrates (G. S. Jamieson and N. Bourne, eds.), p. 195-227. Can. Spec. Pub. Fish. Aquat. Sci. 92. Loher et al.: Growth of Paralithodes camtschciticus 587 Otto, R. S., R, A. Macintosh, and R A. Cumminskey. 1990. Fecundity and other reproductive parameters of female red king crab (Paralithodes camtschatica ) in Bris- tol Bay and Norton Sound. In Proceedings of the inter- national symposium on king and Tanner crabs, p. 65-90. Univ. Alaska Sea Grant Report 90-04, Anchorage, AK. Paul, J. M., A. J. Paul, R. S. Otto, and R. A. Macintosh. 1991. Spermatophore presence in relation to carapace length for eastern Bering Sea blue king crab ( Paralithodes platy- pus, Brandt, 1850) and red king crab (Paralithodes camts- chaticus (Tilesius, 1815)). J. Shellfish Res. 10:157-163. Polovina, J. J., W. R. Haight, R. B. Moffit, and F. A. Parrish. 1993. The role of benthic habitat, oceanography, and fish- ing on the population dynamics of the spiny lobster, Panu- lirus marginatus (Decapoda, Palinuridae), in the Hawaiian Achipelago. In Proceedings of the fourth annual work- shop on Lobster Biology and Management. Crustaceana 68:203-212. Polovina, J. J., and G. T. Mitchum 1992. Variability in spiny lobster recruitment and sea level in the northwestern Hawaiian Islands. Fish. Bull. 90:483-493. Rothlisberg, P. C., J. A. Church, and C. B. Fandry. 1994. A mechanism for near-shore concentration and estua- rine recruitment of post-larval Penaeus plebejus Hess (Decap- oda, Penaeidae). Estuarine Coastal Shelf Sci. 40:115-138. Rozenkranz, G. E., A. V. Tyler, G. H. Kruse, and H. J. Neibauer. 1998. Relationship between wind and year class strength of Tanner crabs in the southeast Bering Sea. Alaska Fish. Res. Bull. 5:18-24. Shirley, T. C., S. M. Shirley, and S. Korn. 1990. Incubation period, molting and growth of female red king crabs: effects of temperature. In Proceedings of the international symposium on king and Tanner crabs, p. 51-63. Univ. Alaska Sea Grant Report 90-04, Anchorage, AK. Stevens, B. G. 1990. Temperature-dependent growth of juvenile red king crab (Paralithodes camtschatica ) and its effect on size-at- age and subsequent recruitment in the eastern Bering Sea. Can. J. Fish. Aquat. Sci. 47:1307-1317. Stevens, B. G., and J. E. Munk. 1990. A temperature-dependent growth model for juvenile red king crab , Paralithodes camtschatica , in Kodiak, Alaska. In Proceedings of the international symposium on king and Tanner crabs, p. 293-304. Univ. Alaska Sea Grant Report 90-04, Anchorage, AK. Tremblay, M. J., and M. D. Eagles. 1997. Molt timing and growth of the lobster, Homarus amer- icanus, off northeastern Cape Breton Island, Nova Scotia. J. Shellfish Res. 16:383-394. Wainwright, T. C., and D. A. Armstrong. 1993. Growth pattern in Dungeness crab (Cancer magister Dana): synthesis of data and comparison of models. J. Crustacean Biol. 13:36-50. Weber, D. D. 1967. Growth of the immature king crab Paralithodes camts- chatica (Tilesius). Bull. Int. N. Pac. Fish. Comm. 21:21-53. Weber, D. D., and T. Miyahara. 1962. Growth of the adult male king crab Paralithodes camtschatica (Tilesius). Fish. Bull. 62:53-75. Wyllie-Echeverria, T. 1995. Sea ice conditions and the distribution of walleye pol- lock (Theragra chalcogramma) on the Bering and Chukchi Sea Shelf. In Climate change and northern fish popula- tions (R. J. Beamish, ed.), p. 131-136. Can. Spec. Publ. Fish. Aquat. Sci. 121. Zheng, J. M., and G. H. Kruse. 2000. Recruitment patterns of Alaskan crabs in relation to decadal shifts in climate and physical oceanography. ICES J. Mar. Sci. 57:438-451. Zheng, J., M.C. Murphy, and G. H. Kruse. 1995a. A length-based population model and stock-recruit- ment relationships for red king crab, Paralithodes camtschati- cus, in Bristol Bay. Can. J. Fish. Aquat. Sci. 52:1229-1246. 1995b. Llpdated length-based population model and stock- recruitment relationship for red king crab in Bristol Bay, Alaska. Alsk. Fish. Res. Bull. 2:114-124. 588 Age, growth, and mortality of California halibut, Paralichthys californicus, along southern and central California Leslie S. MacNair Michael L. Domeier Calvin S. Y. Chun California Department of Fish and Game 4665 Lampson Avenue, Suite C Los Alamitos, California 90720 E-mail address (for C. S Y Chun, contact author): cchun@dfg.ca gov Abstract— California halibut, Paralich- thys californicus, collected by a 400- mesh eastern trawl in southern (Mex- ican border to Point Conception) and central (Point Conception to Tomales Bay) California were aged by using whole and sectioned otoliths (sagittae) to determine age, growth, and mor- tality. Males represented 69% of the sample from southern California and 53% of the sample from central Califor- nia. A higher proportion of California halibut were older in southern Califor- nia than in central California. Although California halibut can live as long as 30 years, the oldest fish found in our study was 13 years old. For southern Califor- nia, the von Bertalanffy growth func- tion (VBGF) was L,=925.3(l-e-°08u+2-2>) for males and L(=1367.7(l-e_008u+1'2)) for females. For central California, the VBGF was L,=956.7(l-e-°10“+21>) for males and L(=1477.1( l-e_010('+0-2)) for females. The VBGF showed that at the same age, females on average were larger than males in both southern and central California. The VGBF also showed that both male and female hal- ibut in central areas on average were larger than halibut from southern Cal- ifornia. Instantaneous mortality rates of halibut in southern California were estimated at 0.91 for males and 0.68 for females. Mortality estimates for cen- tral California could not be calculated because of small sample sizes. Manuscript accepted 9 March 2001. Fish. Bull. 99:588-600 (2001). California halibut, Paralichthys califor- nicus, is an important flatfish species for sport and commercial fisheries in nearshore waters off central and south- ern California (Frey, 1971). The gear most commonly used commercially to harvest California halibut are otter trawls, gill nets, and trammel nets, whereas hook-and-line or spear are used mostly in the recreational fishery (Kramer and Sunada, 1992). Information on age, growth, and mor- tality of California halibut are impor- tant to the management of both the sport and commercial halibut fisheries. Previous studies have addressed these issues to a limited extent. Pattison and McAllister (1990) studied techniques to age halibut and determined length-at- age of halibut along the California coast, but most of the fish were taken south of Point Conception and they did not do separate analyses for southern and cen- tral California. Sunada et al. (1990) an- alyzed age, size, and sex composition of commercial landings in southern Cal- ifornia. Haaker (1975) studied the bi- ology of California halibut in Anaheim Bay, including age and growth, but the fish he used were generally less than three years of age, immature, and in- cluded fish only to 510 mm total length. In contrast to other studies, our study provides information on age and growth for central and southern California sep- arately. Mortality for southern Califor- nia was also calculated. Methods Southern California was defined as the area between the U.S. -Mexico border and Point Conception, and central Cali- fornia as the area between Point Con- ception and Tomales Bay. Bottom trawl surveys were conducted off central Cal- ifornia from 8 July through 3 August 1993 and off southern California from 14 February through 18 March 1994. California halibut were caught with a 400-mesh eastern trawl (15 m wide x 1.5 m high; 9.8 cm mesh body, 8.5 cm mesh codend). For both areas, sampling effort was stratified by depth: 0-20 fathoms (fm), 21-40 fm, and 41-60 fm. Fifty stations within each depth stra- tum were randomly selected for sam- pling by trawling parallel to isobaths at each station. The sample sizes were not proportional to the estimated area of the strata (Table 1). An additional ten California halibut were collected with the same 400-mesh eastern trawl gear in southern Cali- fornia from July 1994 to June 1995. These were included in the age and growth determinations. However, they were not used for mortality estimates because they were collected at a differ- ent time period from the other halibut collected off southern California. Each halibut was measured to the nearest millimeter for total length (TL) and standard length (SL). Gonads were examined visually to determine the sex of the fish. Sagittal otoliths were re- moved and stored dry in vials for age determination. Pattison and McAllis- ter (1990) determined that when com- pared with other fish structures, oto- liths provided the most reliable ages for California halibut. They found that in an otolith, an opaque band was formed during spring and summer (from April to October) and a translucent band was MacNair et at: Age, growth, and mortality of Paralichthys californicus 589 Table 1 Number of trawls and number of California halibut used in age and growth analyses by depth strata and region. Asterisk (*) indicates that these fish were not used in the mortality estimates because they were caught at a later date. Stratum Southern California Central California Depth (fm) No. of trawls No. of halibut used Area (nmi2) No. of trawls No. of halibut used Area (nmi2) 1 0-20 58 913 +3* 575 47 254 587 2 21-40 46 165 +2* 495 39 22 905 3 41-60 45 13 +1* 425 40 0 1124 Total 1091 +6* 276 deposited in winter (from November to March). A combination of an inner opaque and adjacent out- er translucent zone represents an annular growth ring or annulus (Casselman, 1983). This growth ring was validated by Pattison and McAllister (1990) with oxytetracycline-marked otoliths; they found that only one complete ring (opaque and translucent) was formed each year. Pattison and McAllister suggested that age es- timates of older fish (age 11 or older) may be underestimated when whole otoliths, rather than sectioned otoliths, are read. In the preliminary analysis in our study, to test if a significant dif- ference existed between estimating age with whole otoliths and estimating age with sectioned otoliths, an otolith from each of 80 randomly selected fish (10 fish from each 100-mm size class from 200 to 900 mm TL) was viewed both whole and sectioned to estimate age. A mutually determined age (an age agreed upon by the two readers) was established for the whole otolith, as well as for the sectioned version of the otolith. If the age readings were identical for each reader, then that age was consid- ered the correct age. Otherwise, the two readers discussed the readings and agreed upon an age. The non- parametric, paired Wilcoxon test was applied to the pairs of whole and sectioned otolith ages for each fish. In addi- tion, a paired Wilcoxon test was applied to otoliths for fish restricted to 8 years and older (n=32). In addition to determining whether to use whole or sec- tioned otoliths, the preliminary analysis also involved de- termining how halibut age was to be estimated: either by using the mutually determined age, as already described, or an averaged age. The averaged age was a single value calculated as the mean of the four independent readings (whole or sectioned version of an otolith read by two read- ers on two different occasions) and rounded to the nearest integer. Sample means generally have the same type of sampling distribution as individual observations, but with a smaller variance. Hence, we applied the nonparametric paired Wilcoxon test to the pairs of averaged and mutually determined ages for each otolith (n= 80). The results from the preliminary analysis determined how we conducted the process of age determination in the main part of the study. We acknowledge that using sectioned otoliths is the rec- ommended procedure when analyzing long-lived fish. Al- though individual California halibut may live as long as 30 years (Frey, 1971; Pattison and McAllister, 1990), all the fish in our study were relatively young (less than 14 years old). Because of our preliminary analysis and because sec- tioning a large number of otoliths (potentially over 1300) would have entailed considerable costs, all otoliths were initially read whole. We did use sectioned otoliths when the variability in the age readings for that otolith seemed to be unacceptably high and when the whole otolith was judged unreadable because of the anise oil residue. Before recording ages, two readers consulted with each other to standardize their technique for counting annuli. Each otolith was initially read whole, that is, the surface of the uncut otolith was read for the number of annuli pres- ent. The whole otolith was submerged in water, illuminat- ed with fiber optic light transmitted from the sides, and viewed against a dark background with a dissecting mi- croscope (16x to 25x). Figure 1 is an example of an otolith 590 Fishery Bulletin 99(4) Figure 2 An example of a whole otolith that was unreadable (A) according to our criteria, but readable sectioned (B). Age was determined to be 9 years. viewed whole. Some whole otoliths from central California were soaked in anise oil during initial handling to facilitate viewing annuli. The anise oil, however, left a residue that later caused problems in reading the otoliths. Therefore, these otoliths had to be sectioned before they could be read. For each otolith (either whole or sectioned), the two read- ers counted the annuli twice at different times, for a total of four independent readings. The birth date for a halibut was assigned to 1 January (Williams and Bedford, 1974). The mean and standard deviation of the four age deter- minations for each whole otolith were calculated. Because of our preliminary analysis, if the standard deviation was less than 1.5 years, the mean, rounded to the nearest in- teger, was the estimated age for that halibut. A standard deviation of 1.5 years was arbitrarily chosen as the upper limit for acceptance of an age determination. This devia- tion represented a difference in ages of 3 years or more among readings. If the standard deviation was 1.5 years or larger the whole otolith was judged unreadable. These otoliths were cut laterally through the nucleus with a di- amond blade on a Buehler low-speed Isomet saw. Three sections, approximately 0.38 mm thick, were cut from each otolith and mounted on a glass slide with Eukitt clear mounting medium. These sections were viewed with a compound microscope (25x to lOOx) with transmitted light. The best of three sections on a slide was chosen for age determination. Figure 2 is an example of a whole oto- lith that was unreadable, but readable when sectioned. The mean and standard deviation of the four age deter- minations for each sectioned otolith were calculated. In ac- cordance with our preliminary analysis, if the standard deviation was less than 1.5 years, the rounded mean was considered the estimated age. If the standard deviation was 1.5 years or larger, the otolith was not used in the analysis because it was felt to be unreadable. The index of average percent error was calculated to compare with- in-reader precision (Beamish and Fournier, 1981) for both whole and sectioned otoliths. Several analyses were made on total lengths and ages by sex and region. Comparisons of both length and age distributions between sexes for each region were made by using the two-sample Kolmogorov-Smirnov test (Holland- er and Wolfe, 1973). Length-at-age data for females were compared with those for males sampled from the same re- gion by using the nonparametric Mann-Whitney test. The Mann-Whitney test was also used to compare the length- at-age data between regions for each sex separately. The von Bertalanffy growth function was fitted to the length-at-age data for individual fish (Ricker, 1975); pa- rameters were estimated by nonlinear least squares by us- MacNair et at: Age, growth, and mortality of Paralichthys californicus 591 ing the Gauss-Newton method in the program NLIN from SAS (1990). Outliers were removed when the von Ber- talanffy growth function was fitted to the data. Outliers were data points that clearly stood apart in scatter plots of individual total length versus age. Means and standard deviations of observed length-at-age, and the theoretical von Bertalanffy growth function (VBGF) were plotted for females and males by region. Hotelling’s T2 test (Bernard, 1981) compared growth pa- rameters between male and female halibut in each region. This test was also used to determine if growth in southern California was different from growth in central California, for each sex separately. The estimated annual survival rate, S, was calculated for males and females separately by region by using Rob- son and Chapman’s method (1961). Total annual mortal- ity, A, was calculated by using the complement of annual survival, A = 1 - S. The instantaneous mortality rate, Z, was estimated by using -ln(S) (Ricker, 1975). All statistical analyses were performed at the 0.05 level of significance. Results In our preliminary analysis, a subsample of whole otoliths from 80 fish were read and then sectioned to determine the best method for reading otoliths for age determina- tion. In Figure 3 is a plot of the mutually determined ages from sectioned and whole otoliths, along with a 45 degree line and a linear regression. No significant difference was found between ages from whole and sectioned otoliths when using all 80 fish (paired Wilcoxon test, P=0.21), and when using only older fish (8 years and older) (paired Wil- coxon test, P=0.11). Also, as part of the preliminary analysis, in the compari- son of averaged ages with mutually determined ages for the estimation of halibut age, we found no significant dif- ferences for both whole (paired Wilcoxon test, mean dif- ferenced. 15, P=0.23) and sectioned ototliths (paired Wil- coxon test, mean differenced. 16, Pd. 06). Therefore, ages were estimated for each individual by calculating an aver- age of four independent readings from whole or sectioned otoliths because an average of several readings is consid- ered a more reliable measurement than a single reading (Fleiss, 1986). The total number of trawls conducted by strata and re- gion are listed in Table 1. The original intent was to per- form 50 trawls at each depth stratum for both southern and central California. However, substrate obstructions and other problems limited the number of successful tows at most strata. A total of 58 trawls were successfully com- pleted in stratum 1 in southern California. Because most of the halibut were collected in a single stratum (stratum 1), comparisons among strata were not made (Table 1). In southern California, a total of 916 indi- viduals were collected from stratum 1 (84% of total), 167 individuals from stratum 2, and only 14 individuals from stratum 3. In central California, a total of 254 individuals were collected from stratum 1 (92% of total), 22 individu- als from stratum 2, and none from stratum 3. Figure 3 A scatter plot of mutually determined ages for sectioned otoliths and whole otoliths. The dotted line is the linear regression through the origin (slope is 0.947); solid line is the 45 degree line. For southern California, pairs of whole otoliths from 1109 individuals were initially examined to estimate age. Twenty-two pairs of whole otoliths were unreadable, six pairs had no discernible opaque and translucent zones, and sixteen pairs had large variations in the four age readings (SD >1.5). These 22 pairs of whole otoliths were sectioned to determine age. Of these, two were eliminated because of large variations in the four age readings (SD >1.5). An additional ten were eliminated as outliers be- cause lengths were anomalously small or large for the es- timated age when the von Bertalanffy growth function was fitted to the data. Thus, a total of 1097 individuals were used to evaluate age and growth for halibut in south- ern California. The sample consisted of 69% males (761 in- dividuals) and 31% females (336 individuals). Tables 2 and 3 provide counts of total length versus age for males and females, respectively, in southern California. In central California, pairs of whole otoliths from 292 individuals were initially examined to estimate age. Oto- liths from 28 halibut had no discernible opaque and trans- lucent zones and 49 had large variations in the four age readings (SD >1.5), resulting in a total of 77 otoliths that were sectioned to estimate age. Of these, fourteen otolith pairs were eliminated because of large variations in read- ings (SD >1.5). An additional two became outliers that were removed when the von Bertalanffy growth function was fitted to the data. Thus, a total of 276 individuals were used to evaluate age and growth for halibut in central California. The sample consisted of 53% (147 individuals) males and 47% (129 individuals) females. Counts of total length versus age for males and females, respectively, in central California are shown in Tables 4 and 5. Within-reader index of average percent error for whole otoliths ranged from 8.1%< to 13.2%. For sectioned otoliths, 592 Fishery Bulletin 99(4) Table 2 Male California halibut total length versus age (mostly whole otoliths, some sectioned) in southern California. TL (mm) Age (yr) Mean age 1 2 3 4 5 6 7 8 9 10 11 12 201-225 2 1 2.3 226-250 2 5 2 2.0 251-275 1 20 14 2.4 276-300 1 51 59 4 1 2.6 301-325 14 33 4 2 2.9 326-350 7 21 18 5 3.4 351-375 4 32 34 17 3 3.8 376-400 11 25 16 6 1 4.3 401-425 5 10 25 8 12 2 5.3 426-450 8 15 15 11 2 5.7 451-500 6 14 26 21 6 6 6.3 501-525 1 7 14 17 5 2 6.5 526-550 1 5 9 8 4 2 7.5 551-575 2 9 9 9 3 1 7.2 576-600 1 4 7 1 3 0 1 1 7.6 601-625 1 6 5 2 3 8.0 626-650 1 4 1 2 7.5 651-675 0.0 676-700 0.0 701-725 1 1 8.5 Mean length 259 289 319 377 417 482 506 538 552 578 577 586 Median length of distribution: 387 mm TL within-reader average percent error ranged from 4.7% to 8.9%. In southern California, significant differences between males and females were found in comparisons of length distributions (Kolmogorov-Smirnov: D=0.42, P=0.0001). A higher proportion of males were smaller than females (Fig. 4). The median length for males was 387 mm TL (range: 210-707 mm TL); the median length for females was 544 mm TL (range: 241-987 mm TL) (Tables 2, 3, and 6). Fe- males on average were significantly larger than males for ages 3 through 10 (Mann- Whitney tests, Table 6). Some differences in lengths were also found between males and females in central California. Most of the larger fish were females (Fig. 4). However, the length distributions of males and females in central California were not signifi- cantly different (Kolmogorov-Smirnov: D= 0. 14, P= 0. 12). The median length for males was 437 mm TL (range: 285-781 mm TL); the median length for females was 444 mm TL (range: 262-1039 mm TL) (Tables 4, 5, and 7). Females on average were significantly larger than males for ages 3 and 5 through 8 (Mann- Whitney tests, Table 7). In addition, comparisons were made between regions for each sex separately. In comparisons of mean length at ag- es, the males in central California on average were signifi- cantly larger than those in southern California for ages 2 through 9, except age 8 (Mann- Whitney tests, Table 8). The females in central California on average were signifi- cantly larger than those in southern California for only ages 5, 7, and 8 (Mann-Whitney tests, Table 8). However, the median length of females in southern California (544 mm) was greater than the median length for those in cen- tral California (444 mm). In southern California, age distributions were signifi- cantly different for males and females (Kolmogorov- Smirnov: D=0.18, P=0.0001). A higher proportion of fe- males were older fish compared with males. The primary age mode for females was age 6 and for males it was age 3 (Fig. 5). In central California, age distributions were al- so different for males and females (Kolmogorov: D=0.20, P=0.008). Males had a higher percentage in the 6 to 8 year range and a lower percentage in the 3 to 5 year range com- pared with females (Fig. 5). In central California, the pri- mary mode for both males and females was age 3. With both regions combined, females were found up to age 13 and males up to age 12. In comparisons of age distribu- tions between regions, we found that a higher proportion of halibut in southern California was older than halibut in central California (Fig. 5). Because differences in length-at-age were found be- tween sexes, von Bertalanffy growth parameters were cal- MacNair et at: Age, growth, and mortality of Paralichthys californicus 593 Table 3 Female California halibut total length versus age (mostly whole otoliths, some sectioned) in southern California. TL (mm) Age (yr) Mean age 1 2 3 4 5 6 7 8 9 10 11 12 13 201-225 0.0 226-250 3 2.0 251-275 1 5 4 2.3 276-300 7 7 2.5 301-325 1 2 2.7 326-350 3 6 5 3.1 351-375 2 4 2 3.0 376-400 4 5 2 3.8 401-425 6 12 3 1 4.0 426-450 5 5 3 1 4.0 451-500 4 9 10 4 3 1 4.9 501-525 1 17 5 1 5.3 526-550 3 9 8 1 1 1 5.6 551-575 4 3 10 6 2 6.0 576-600 4 3 8 4 4 1 6.2 601-625 2 4 8 6 2 6.1 626-650 3 10 9 4 1 6.6 651-675 2 1 10 6 3 2 6.5 676-700 3 1 1 0 0 1 7.3 701-725 3 1 1 1 8.0 726-750 2 0 2 7.0 751-775 1 7.0 776-800 1 1 3 9.4 801-825 1 1 1 9.0 826-850 1 1 1 9.0 851-875 2 0 1 8.7 876-900 0.0 901-925 1 1 9.5 926-950 1 1 9.5 951-975 1 13.0 976-1000 1 11.0 Mean length 252 292 364 463 520 591 616 666 736 828 836 000 964 Median length of distribution: 544 mm TL culated separately for males and females for both regions (Table 9). Outliers were removed when the von Berta- lanffy growth function was fitted to the data. With outli- ers removed, the standard errors of the asymptotic mean length, Ltc, became considerably smaller. Graphs of the von Bertalanffy growth function are shown in Figure 6 (where females and males are compared by region) and Figure 7 (where southern and central California are com- pared by sex). Significant differences between males and females in southern California were found for all three growth pa- rameters (Hotelling’s T2 tests, P<0.0001). For central Cali- fornia, two of the three growth parameters were signifi- cantly different between males and females (Hotelling’s T2 tests, P<0.0001). Females grew faster and on average were larger than males at the same age in both regions (Fig. 6). Similar comparisons were made between regions for each sex separately (Fig. 7). For females, all three growth parameters were significantly different between southern and central California (Hotelling’s T2 tests, P<0.003). In contrast, the only significant difference between regions for males was in K, the Brody growth coefficient (Hotell- ing’s T2 tests, P<0.0001). In southern California, survival rates were based on ages 7 to 12 years for males and ages 6 to 11 years for females. The estimated annual survival rates, S, for halibut from south- 594 Fishery Bulletin 99(4) Table 4 Male California halibut total length versus age (mostly whole otoliths, some sectioned) in central California. TL (mm) Age (yr) 1 2 3 4 5 6 7 8 9 10 11 12 Mean age 201-225 0.0 226-250 0.0 251-275 0.0 276-300 4 1 2.2 301-325 4 3 2.4 326-350 5 4 2.4 351-375 9 3.0 376-400 2 15 8 3.2 401-425 9 1 1 3.3 426-450 2 7 2 4.0 451-500 3 6 2 4 1 1 4.8 501-525 2 1 5 1 3 6.2 526-550 1 1 7 1 1 6.0 551-575 1 2 4 1 6.6 576-600 1 0 0 2 7.0 601-625 1 1 1 8.0 626-650 1 2 2 1 7.5 651-675 1 0 0 2 6.0 676-700 1 0 1 1 7.7 701-725 1 8.0 726-750 1 0 0 0 1 9.0 776-800 1 10.0 Mean length 000 326 385 448 493 539 596 582 649 781 737 000 Median length of distribution: 437 mm TL Figure 4 Length frequencies for male and female California halibut sampled off southern California (So. CA) and central California (Central CA). MacNair et al.: Age, growth, and mortality of Paralichthys californicus 595 Table 5 Female California halibut total length versus age (mostly whole otoliths, some sectioned) in central California. TL (mm) Age (yr) Mean age 1 2 3 4 5 6 7 8 9 10 11 12 13 201-225 0.0 226-250 0.0 251-275 1 2.0 276-300 3 4 2.6 301-325 5 7 2.6 326-350 2 10 2 3.0 351-375 1 7 2.9 376-400 8 5 3.4 401-425 3 2 3.4 426-450 2 6 3.8 451-500 1 3 2 4.2 501-525 7 2 4.2 526-550 2 1 3 4.2 551-575 1 3 1 4.0 576-600 1 2 2 1 4.5 601-625 1 4.0 626-650 1 1 0 0 1 6.7 651-675 1 1 4.5 676-700 1 0 1 1 1 6.3 701-725 1 0 0 1 6.5 726-750 3 5.0 751-775 1 0 1 6.0 776-800 1 6.0 801-825 1 4.0 826-850 1 8.0 851-875 1 9.0 876-900 1 1 8.5 901-925 1 2 8.7 926-950 0.0 951-975 1 10.0 976-1000 0.0 1001-1025 1 12.0 1026-1050 1 12.0 Mean length 000 313 373 493 606 672 726 810 842 970 000 1027 000 Median length of distribution 444 mm TL em California were 0.40 (95% confidence interval=0.34, 0.46) for males and 0.51 (95% confidence interval=0.45, 0.57) for females. Estimated annual mortality rates, A, were 0.60 for males and 0.49 for females. Instantaneous mortalities, Z, were estimated at 0.91 for males and 0.68 for females. We were unable to calculate survival rates and mortalities for central California halibut because the sample sizes were too small. Discussion The California halibut can be a long-lived species, living as long as 30 years (Frey, 1971; Pattison and McAllister, 1990). However, none of the fish in our study was older than 13 years, and most of the fish were less than 11 years old (Fig. 5). We feel that the low average percent error in our study showed that both whole and sectioned otoliths were reliable methods for aging California halibut for the age range of our study. Pattison and McAllister (1990) suggested that because of the asymmetrical growth of the sagittae in larger fish, ages assigned to older fish (age 11 or older) may be un- derestimated if whole otoliths, rather than sectioned oto- liths, were read. Manooch and Potts (1997) in a growth study of greater amberjack also did not find whole oto- liths useful in aging fish. In contrast, we found that age estimates from whole and sectioned otoliths were not sig- 596 Fishery Bulletin 99(4) nificantly different, even when the analysis was limited to ages 8 to 13 years. However, our conclusions are lim- ited to primarily younger fish because most of the fish in our study were less than 11 years old. Pattison and McAllister recommended sectioned otoliths for fish aged 11 years or older. Nonetheless, we concur that use of sec- tioned otoliths in older individuals may reduce the varia- tion in assigned ages. We feel that the larger percentage of unreadable whole otoliths in central California versus southern California was due primarily to the preparation techniques used for age determination. The anise oil used in the initial handling of some central California otoliths created a white film that pre- cluded the otoliths from being read at a later time. Because of problems encountered with anise oil, we do not recommend applying it to otoliths until one is ready to age them. Table 6 Mean total length (mm) and standard deviation for male and female California halibut by age group in southern California. Mann- Whitney comparisons of male and female halibut for mean length at age. Asterisk (*) indicates statistical significance. Male Female Age Mann-Whitney (yr) n Mean SD Min Max n Mean SD Min Max P-value 1 4 259.3 25.0 232 290 i 252.0 — 252 252 — 2 103 289.4 28.4 210 369 21 291.9 38.4 241 368 0.77 3 178 319.3 39.3 220 411 42 364.4 63.9 254 477 0.0001* 4 no 377.3 42.2 291 501 54 462.6 86.7 342 660 0.0001* 5 106 417.2 55.0 294 561 58 520.5 64.3 389 671 0.0001* 6 92 481.7 61.9 362 646 70 591.2 67.7 401 735 0.0001* 7 97 505.6 65.1 399 645 41 616.1 60.5 493 758 0.0001* 8 40 538.3 62.6 403 707 26 666.0 101.4 492 859 0.0001* 9 23 551.5 68.0 451 707 11 736.2 131.9 534 934 0.0002* 10 6 578.2 34.5 542 619 9 828.3 67.4 720 937 0.002* 11 1 577.0 — 577 577 2 836.0 213.5 685 987 — 12 1 586.0 - 586 586 — — — — — — 13 — — — — — 1 964.0 — 964 964 — Total 761 336 Table 7 Mean total length (mm) and standard deviation for male and female California halibut by age-group in central California. Mann- Whitney comparisons of male and female halibut for mean length at age. Asterisk (*) indicates statistical significance. Age (yr) Male Female Mann-Whitney P-value n Mean SD Min Max n Mean SD Min Max 2 15 325.9 29.8 285 381 12 312.9 28.7 262 367 0.29 3 46 384.8 38.1 293 472 46 373.4 70.0 279 593 0.02* 4 26 447.8 60.5 380 660 35 493.5 102.3 328 820 0.07 5 9 492.8 59.0 424 582 17 606.2 101.4 474 758 0.007* 6 20 538.6 49.4 479 680 4 672.0 87.0 582 782 0.006* 7 13 595.6 72.7 490 748 2 725.5 43.1 695 756 0.05* 8 13 582.5 74.4 475 719 5 810.4 99.3 694 925 0.002* 9 3 648.7 29.5 619 678 5 842.0 112.2 547 923 — 10 1 781.0 — 781 781 1 970.0 — 970 970 — 11 1 737.0 — 737 737 — — — — — — 12 — — — — — 2 1026.5 17.7 1014 1039 — Total 147 129 MacNair et al.: Age, growth, and mortality of Pciralichthys californicus 597 The data in our study confirmed that females grow fast- er than males, as found in previous studies (Pattison and McAllister, 1990; Sunada et al., 1990), and that growth was different between males and females for each region. Fe- males in both regions had larger asymptotic mean lengths, , than males (Table 9). Growth curves for males and females crossed between ages 1 and 2 for southern California and at age 3 for cen- tral California (Fig. 6). This finding suggests that young male and female halibut are of similar size until about ages 2 and 3 in southern and central California, respectively. However, Haaker (1975) found that juvenile female hali- but in Anaheim Bay grew faster than males. The crossover in growth curves may be due to the small sample sizes for ages 1 and 2 in our study. The growth curves also involved some extrapolation for very young ages. Our study showed lower mean length-at-age than that obtained by Pattison and McAllister (1990); however, the differences for females were slight. These differences may partly be attributed to different environmental conditions in the years prior to each study. Pattison and McAllister used otoliths collected from 1955-66 and 1984-88, where- as we used otoliths collected from 1993 to 1995. Manooch and Potts (1997) in a study of greater amberjack in the Gulf of Mexico also cited temporal changes as one of the factors that may have contributed to differences in growth between greater amberjack in southern Florida and northern Gulf of Mexico. The geographic range of the samples may also have affected the results because we found growth differences between southern and cen- tral California. Pattison and McAllister’s study encom- passed samples from central and southern California, but they did not separate the two regions in their analysis. In addition, gear selectivity may have contributed to the differences found between our two studies. Pattison and McAllister used halibut that were collected with several gear types (trawl, gill net, beach seine, hook and line, and spear), whereas our study sampled halibut with only trawl gear of a specific mesh size. Potts et al. ( 1998) found while studying Vermillion snapper from the southeastern Unit- ed States that the different gear types used in sample col- lection may bias growth results. Sunada et al. (1990) found that for commercial halibut landings in southern California the asymptotic mean length, L^, was 909 mm TL for males and 1445 mm TL for females. In comparison, our study showed for southern Table 8 Mann-Whitney comparisons of California halibut from southern and central California for mean length at age. Asterisk (*) indicates statistical significance. Age (yr) Mean observed total length (mm) Sample n size Mann- Whitney P-value Southern Central Southern Central Males 2 289.4 325.9 103 15 0.0001* 3 319.3 384.8 178 46 0.0001* 4 377.3 447.8 110 26 0.0001* 5 417.2 492.8 106 9 0.0006* 6 481.7 538.6 92 20 0.0002* 7 505.6 595.6 97 13 0.0002* 8 538.3 582.5 40 13 0.08 9 551.5 648.7 23 3 0.03* Females 2 291.9 312.9 21 12 0.10 3 364.4 373.4 42 46 0.92 4 462.6 493.5 54 35 0.21 5 520.5 606.2 58 17 0.002* 6 591.2 672.0 70 4 0.07 7 616.1 725.5 41 2 0.04* 8 666.0 810.4 26 5 0.01* 9 736.2 842.0 11 5 0.17 California a slightly higher for males at 925 mm TL, and a lower Lx for females at 1368 mm TL. The asymp- totic mean length for females may be lower because fewer large females were collected in our study. Although our study had different growth parameters than those of Su- nada et al. (1990), the differences were only slight. The differences may have been due to gear vulnerability. Su- nada et al. (1990) collected halibut from both trawl nets and gill nets and found that fewer females were collected in trawl nets, which may be attributed to trawl gear being restricted to offshore areas during spawning season. Also Table 9 Von Bertalanffy growth parameters with asymptotic standard errors (SE) for California halibut by sex in southern and central California. Length is given in mm TL. Region Sex n u SE K SE *0 SE Southern male 761 925.3 121.4 0.08 0.02 -2.2 0.41 female 336 1367.7 273.4 0.08 0.03 -1.2 0.48 Central male 147 956.7 211.9 0.10 0.05 -2.1 0.91 female 129 1477.1 308.1 0.10 0.04 -0.2 0.43 598 Fishery Bulletin 99(4) Age (years) Figure 6 Von Bertalanffy growth curves for comparing growth between sexes by region. MacNair et al.: Age, growth, and mortality of Paralichthys californicus 599 females, which are generally larger, may be more capable of escaping a trawl net than males. Other factors that may have affected the growth parameters include the follow- ing: method of collection, time periods of sampling, and dif- ferent environmental conditions in the years prior to each study. Many factors can affect growth rates of fish including differences in the seasonality of spawning, environmental factors, amount and size of food, and genetics (Weatherly and Gill, 1987; Moyle and Cech, 1988). Southern and cen- tral California are biogeographically different. Differences between central and southern California coastal waters include temperature, water circulation patterns, bottom topography, and substrate. In particular, Jow (1990) noted differences in bottom topography and substrate between the trawl areas for California halibut in southern and northern (San Francisco Bay area) California. The bio- geographical differences among these regions could possi- bly cause growth rates of the same species to differ. Stud- ies of other fish species have shown differences in growth between geographical regions. For example, Parrish et al. (1985) found latitudinal differences in the growth of northern anchovy, Engj-aulis mordax. They found that the growth rate of juvenile northern anchovy in central Cali- fornia was greater than that of juvenile northern anchovy in southern California. Butler et al. (1996) found differ- ences in maturation and length at age of Pacific sardine between latitudinal regions. Sardine appeared to mature at a younger age off both southern and Baja California than off Monterey; one-, two-, and three-year old fish were smaller at age off Baja and larger off Monterey. Deriso et al. (1996) confirmed the findings of sardine growth differ- ences by Butler et al. ( 1996). The exact mechanism for lati- tudinal differences in growth rates of California halibut is still unclear and further research is needed to clarify this phenomenon. Acknowledgments We wish to express our appreciation to Larry Jacobson with the National Marine Fisheries Service, for his detailed review of an earlier draft of the paper. We appreciate the efforts of Steve Wertz, Sandra Owen, and Paul Gregory who helped with various aspects of this study. We also thank the reviewers for their comments and suggestions that improved the paper. 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. 600 Fishery Bulletin 99(4) Bernard, D. R. 1981. Multivariate analysis as a means of comparing growth in fish. Can. J. Fish. Aquat. Sci. 38:233-236. Butler, J. L., M. L. Granados G., J. T. Barnes, M. Yaremko, and B. J. Macewicz.. 1996. Age composition, growth, and maturation of the Pacific sardine ( Sardinops sagax ) during 1994. Calif. Coop. Oce- anic Fish Invest. Rep. 37:152-159. Casselman, J. M. 1983. Age and growth assessment of fish from their cal- cified structures-techniques and tools. In Proceedings of the international workshop on age determination of oce- anic pelagic fishes: tunas, billfishes, sharks (E. Prince and L. Pulos, eds.) p. 1-17. U.S. Dep. Commer, NOAA Tech. Report 8. Deriso, R. B., J. T. Barnes, L. D. Jacobson, and P. R. Arenas. 1996. Catch-at-age analysis for Pacific sardine (Sardinops sagax ) 1983-1995. Calif. Coop. Oceanic Fish Invest. Rep. 37:175-187. Fleiss, J. L. 1986. The design and analysis of clinical experiments. John Wiley and Sons, New York, NY, 432 p. Frey, H.W„'ed. 1971. California’s living marine resources and their utiliza- tion. Calif. Dep. Fish Game, Sacramento, CA, 148 p. Haaker, P. L. 1975. The biology of the California halibut, Paralichthys cali- fornicus (Ayres), in Anaheim Bay, California. In The marine resources of Anaheim Bay (E. D. Lane and C. W. Hill, eds), p. 137-151. Calif. Dep. Fish Game, Fish Bull. 165. Hollander, M., and D. A. Wolfe. 1973. Nonparametric statistical methods. John Wiley and Sons, New York, NY, 503 p. Jow, T. 1990. The California halibut trawl fishery. In The Califor- nia halibut, Paralichthys californicus , resources and fisher- ies (C. W. Haugen, ed. ), p. 229-241. Calif. Dep. Fish Game, Fish Bull. 174. Kramer, S. H., and J. S. Sunada. 1992. California halibut. In California’s living marine re- sources and their utilization (W. S. Leet, C. M. Dewess, and C. W. Haugen, eds.), p. 94-97. California Sea Grant, Univ. Calif., Davis, CA. Manooch, C. S. Ill, and J. C. Potts. 1997. Age, growth, and mortality of greater amberjack, Seri- ola dumerili, from the U.S. Gulf of Mexico headboat fish- ery. Bull. Mar. Sci. 61:671-683. Moyle, P. B., and J. J. Cech Jr. 1988. Fishes: an introduction to ichthyology, 2nd ed. Pren- tice-Hall, Inc., Englewood Cliffs, NJ, p. 593 Parrish, R. H., D. L. Mallicoate, and K. F. Mais. 1985. Regional variations in the growth and age composi- tion of northern anchovy, Engraulis mordax. Fish. Bull. 83:483-496. Pattison, C. A. and R. D. McAllister. 1990. Age determination of California halibut, Paralich- thys californicus. In The California halibut, Paralichthys californicus, resources and fisheries (C. W. Haugen, ed.), p. 207-216. Calif. Dep. Fish Game, Fish Bull. 174. Potts, J. C., C. S. Manooch III, and D. S. Vaughan. 1998. Age and growth of vermillion snapper from the south- eastern United States. Trans. Am. Fish. Soc. 127:787- 795. Ricker, W. E. 1975. Computation and interpretation of biological statis- tics of fish populations. Bull. Fish. Res. Board Can. 191, 382 p. Robson, D. S., and D. G. Chapman. 1961. Catch curves and mortality rates. Trans. Am. Fish. Soc. 90:181-189. SAS. 1990. SAS/STAT user’s guide, version 6, fourth ed., vol. 2. SAS Institute, Inc., Cary, NC, 846 p. Sunada, J. S., P. V. Velez, and C. A. Pattison. 1990. Age, size, and sex composition of California halibut from southern California commercial fishery landings, 1983-1988. In The California halibut, Paralichthys cal- ifornicus, resources and fisheries (C.W. Haugen, ed.), p. 303-319. Calif. Dep. Fish Game, Fish Bull. 174. Weatherly, A. H„ and H. S. Gill. 1987. The biology of fish growth. Academic Press, Orlando, FL, 443 p. Williams, T., and B. C. Bedford. 1974. The use of otoliths for age determination. In Aging of fish: proceedings of an international symposium, p. 114- 123. Unwin Brothers Ltd., Surrey, England. 601 Abstract— Morphological development and growth of larval and juvenile Pacific bluefin tuna, Thunnus thynnus, were studied from laboratory-reared spec- imens. Average body length (BL) of newly hatched larvae was 2.83 mm and larvae grew on average 5.80 mm by 10 days, 10.62 mm by 20 days, and 35.74 mm by 30 days after hatching. Growth was especially accelerated after 20 days from hatching. Newly hatched larvae had small melanophores scattered over their bodies except for the finfold. On day 1 after hatching (3.35-3.74 mm BL), a characteristic melanophore pat- tern appeared and it was partially maintained until day 3 after hatching. At approximately 4 mm BL, larvae had developed melanophore patterns simi- lar to those of preflexion Thunnus spp. larvae, such as melanophores on the dorsum of the gut, midlateral trunk, and tail, at the dorsal and ventral mid- lines of trunk and tail, and on the lower jaw. Erythrophores appeared at 4.63 mm BL in the caudal area. Jaw teeth appeared at 5-6 mm BL. The preoper- cular angle spine, anterior preoper- cular spine, and posttemporal spine, developed at approximately 7 mm BL, when erythrophores appeared on the trunk and tail. The notochord flexion occurred between 6 and 8 mm BL. At approximately 8 mm BL, erythro- phores disappeared and juvenile color- ation appeared on the trunk and tail, consisting of dense patches of melano- phores at the dorsal and anal-fin bases, embedded melanophores, and melano- phores at the periphery of the eye. The adult complement of fin-ray counts was attained at 10 mm BL, when the juve- nile melanophore pattern was attained, although the pattern was not fully developed. Specimens larger than 20 mm BL did not have erythrophores. Squamation began at 27 mm BL and head spines disappeared by 38 mm BL. Manuscript accepted 10 April 2001. Fish. Bull. 99:601-616 (2001). Morphological development and growth of laboratory-reared larval and juvenile Thunnus thynnus (Pisces: Scombridae) Shigeru Miyashita Yoshifumi Sawada Tokihiko Okada Osamu Murata Hidemi Kumai Fisheries Laboratory of Kinki University, 3153, Shirahama, Wakayama 649-2211, Japan E-mail address (for S. Miyashita): miyasita@cypress ne.|p The Pacific bluefin tuna is a large scom- brid that migrates transoceanically between the western and eastern Pacific Ocean (Orange and Fink, 1963; Fujita, 1998). The Pacific bluefin tuna, Thun- nus thynnus orientalis (Tennninck and Schlegel), is considered a separate sub- species from the Atlantic bluefin tuna, T. thynnus thynnus Linnaeus (Gibbs and Collette, 1967; Collette and Nauen, 1983). Recently, Cho and Inoue (1993) showed reproductive isolation between the two subspecies. Most recently, Col- lette (1999) considered these subspe- cies to be separate species on the basis of morphology and molecular data. Our knowledge of the early life his- tory of oceanic species, such as tunas, is incomplete. Morphological character- istics common to scombrid larvae are large head, gape, and eyes; development of head spination; and posterior migra- tion of the anus (Collette et al., 1984). In studies of early-stage bluefin tuna (e.g. Yabe et al. , 1966; Matsumoto et al. , 1972; Richards and Potthoff, 1974; Kohno et al., 1982), melanophore pat- terns were used to identify tuna larvae. Erythrophore patterns were proposed as another characteristic by Ueyanagi (1966) and Matsumoto et al. (1972). However, ontogenetic changes of mela- nophore and erythrophore patterns and morphological characteristics have not sufficiently been investigated because of the difficulty in obtaining a complete series of wild specimens in their early developmental stages. Morphological, physiological, and be- havioral information has recently been collected from laboratory-reared Pacific bluefin tuna specimens (Harada et al., 1971a; Kaji et al., 1996; Takii et al., 1997; Kumai, 1998; Miyashita et al., 1998; Miyashita et al., 2000, in press; Sawada et al., 2000). The ability to pro- duce a complete series of early-stage specimens provides an opportunity to enhance our understanding of the ear- ly development of bluefin tuna. In this study, we describe the morphological development and pigment patterns of reared larval and juvenile bluefin tuna. Materials and methods Eggs were obtained on 10 July 1994 from bluefin tuna that had been caught off Ohshima when 20-40 cm in total length 10-year class] and that had been raised for seven years naturally in a net cage at the Ohshima Experiment Sta- tion of the Kinki University Fisheries Laboratory. The net cage was equipped with a polyvinyl chloride sheet sur- rounding the top 2 m of the net side wall to prevent eggs from flowing out of the net. Fertilized eggs were collected at the surface with a 330-pm mesh net. Eggs were transferred to a tank in a land-based nursery center and kept in a fine mesh net at 24.5°C for approxi- mately 12 hours. The following morn- ing, eggs were transferred to a 20-m3 rearing tank and larvae hatched at 602 Fishery Bulletin 99(4) approximately 2200 h on 1 1 July. The first 24 hours after hatching were counted as day 0 (day-0) after hatching. Water temperature ranged from 24.5 to 27.7°C during the rearing period (x=25.5°C, Fig. 1). The feeding scheme for larvae and juveniles was as follows: rotifers, Brachionus rotundiformis, from day 2 to day 22 after hatching, Arte- mia nauplii from day 10 to day 25, and other live fish larvae ( Oplegnatlius fasciatus) from day 12 to day 30. At 1000 h daily from day 0 (12 July) to day 30 (11 Au- gust), 20 tuna were sampled. Before preservation in 5% formalin, individual computer-captured video images were made with a video camera attached to a binocular micro- scope while all specimens were under MS-222 anesthesia. Ten body parts were measured from these images: total length (TL), body length (BL), preanal length, body height at the base of the first dorsal spine, head length, head height, snout length, eye diameter, caudal peduncle depth, and upper jaw length. Video image measurements to 1/102 mm with an accuracy of ±1% were obtained from NIH Im- age (version 1.61) image analysis software (NIH, 1997). Before and during notochord flexion, BL was measured from the tip of the upper jaw to the end of the notochord. After notochord flexion, BL was measured from the tip of the upper jaw to the posterior margin of the hypurals. All specimens (/? =620, 3.49-37.78 mm BL) were used to de- termine growth estimates and to record pigmentation pat- terns and spine and fin development. Sixty-nine specimens (5.26-33.68 mm BL) were stained with alizarin red-S to describe spine development, and squamation. Pigment patterns, especially the position of body melanophores in relation to myomeres, were deter- mined for specimens from 3 to 16 days after hatching (n=180, 3.63-9.96 mm BL). At irregular intervals from day 0 to day 22, one to five fish (a total of 53 specimens, 3.38-26.76 mm BL) were examined while anesthetized with MS-222 to determine erythrophore distribution. Ad- ditional samples were examined at irregular intervals from hatching to day 2. A representative series of specimens from this study was deposited in the Aquatic Natural History Museum of the Fisheries Research Station, Kyoto University (FAKU 129041-129075). Results Structural characteristics and growth Eggs, having a mean diameter of 1.02 mm (SD=0.02; n= 20), were spherical and pelagic and had a single oil globule of 0.26 ±0.05 (mean ±SD) mm in diameter. Newly hatched larvae measured 2.83 ±0.16 mm in BL (mean ±SD, n= 20; Fig. 2A). At 3 days after hatching (3.81 ±0.14 mm BL), the mouth was developed and larvae began feeding. Yolk was present until day 5 (4.32 ±0.13 mm BL). Larvae grew to 5.58 ±0.35 mm by day 10, 9.34 ±1.36 mm by day 20 and 30.36 ±4.34 mm by day 30 after hatching (Fig.l). From hatching to day 20, growth rate was 0.33 mm/day, then increased to 2.10 mm/day from day 20 to day 30. Notochord flexion began as early as 5.76 mm BL, and the largest preflexion larva was 5.83 mm BL. The small- est postflexion larva was 7.46 mm BL; the largest flexion larva, 8.65 mm BL. Thus, the flexion stage occurred from 6 to 8 mm BL (Fig. 2E). The relative values of the nine body parts, except preanal length, increased during larval development, then subse- quently declined (Fig. 3). The ratio of head length, head height, eye diameter, snout length, and upper jaw length to BL reached a maximum at 11-13 mm BL. Head length in- creased rapidly from 10% BL at hatching to 40% at 13 mm BL and then decreased slowly to 34 % BL at 40 mm BL. Head height also increased rapidly from 10% BL at hatch- ing to 27% at 11 mm BL and decreased slowly to 23% BL at 40 mm BL. Eye diameter increased from 10% BL at hatch- ing to 15% at 11 mm BL and then decreased slowly to 9% BL at 40 mm BL. Snout length rapidly increased from 2% BL at hatching to 14% at 12 mm BL and then decreased slowly to 9% BL at 40 mm BL. Upper jaw length increased rapidly from 5% BL at hatching to 20-30% at 12 mm BL and then decreased slowly to 14% BL at 40 mm BL. The Miyashita et at: Morphological development and growth of Thunnus thynnus 603 Figure 2 Development of Thunnus thynnus (A) newly hatched larva, 2.91 mm BL; ( B ) 1-day-old yolksac larva, 3.54 mm BL; ( C ) 4-day-old preflexion larva, 4.02 mm BL; ( D ) 7-day-old flexion larva, 5.34 mm BL; (E) 15-day-old postflexion larva, 7.23 mm BL; (F) 20-day-old postflexion larva, 8.64 mm BL; ( G ) 25-day-old juvenile, 13.55 mm BL. relative depth of the caudal peduncle attained a maximum (ca. 7% BL) at 8 mm BL. In contrast, body height and total length were maximum in relation to BL at approximately 16 and 20 mm BL, respectively. Preanal length increased from 35-44% BL at 3-7 mm BL to about 65% BL at ap- proximately 15 mm BL. 604 Fishery Bulletin 99(4) Table 1 Numbers of fin spines and rays of Thunnus thynnus larvae and juveniles. Roman and Arabic numerals show numbers of spines and soft rays, respectively. Size range (BL; mm) First dorsal fin Second dorsal fin Dorsal finlet Anal fin Pectoral fin Pelvic fin Number of specimens examined BL <5.99 0 0 0 0 0 0 10 6.00-6.99 I-X 0 0 0-6 0-13 0-1+ 1 17 7.00-7.99 0-XIV 0-13 0-8 0-12 0-31 0-1+5 13 8.00-8.99 IX-XIV 12-14 6-9 7-14 11-31 1+4— 1+5 11 9.00-9.99 XI-XV 14 8-9 13-14 16-31 1+5 8 10.00-19.99 XIV-XV 14 8-9 14 31-34 1+5 14 20.00-29.99 XV 14 8-9 13-14 31-36 1+5 19 30.00-37.78 XV 14 8-9 14 31-35 1+5 8 Miyashita et al.: Morphological development and growth of Thunnus thynnus 605 Fin development The number of fin rays increased between 6 and 10 mm BL (Table 1). The smallest specimen possessing dorsal spines was 6.32 mm BL. Caudal-fin rays, second dorsal-, pelvic-, anal-, and pectoral-fin rays appeared at 7.10, 7.76, 8.00, 8.47, and 8.67 mm BL, respectively. The smallest specimen with an adult complement of fin-ray counts (Iwai et al., 1965; Collette and Nauen, 1983) was 9.68 mm BL, whereas the largest specimen with an incomplete fin-ray complement was 10.02 mm BL. The number of pectoral-fin rays for specimens >10 mm BL varied among individuals. Pigmentation Newly hatched larvae (2.62-3.05 mm BL) had small mela- nophores scattered over the body, head, notochord, yolk, and oil globule, but not on the finfold (Fig. 2A). On day 1 after hatching (3.35-3.74 mm BL) (Fig. 2B), den- dritic melanophores were visible from snout to fore- and mid- brain and on the dorsum of the gut; three clusters of mela- nophores occurred in the dorsal midline along with punctate melanophores; melanophores occurred along the ventral mid- line (punctate melanophores formed a line at the anterior caudal area and a cluster of melanophores occurred in the posterior caudal area); and melanophores appeared along the lateral midline of the trunk. The eyes were pigmented. On day 2 after hatching (3.40-4.18 mm BL), the pattern of melanophores was transitional between those of 1-day and older preflexion larvae. Melanophores on the snout disappeared, whereas clusters of melanophores on the an- teriodorsum of the trunk (second myomere), anteriordor- sum of the caudal area (17th to 21st myomere) and pos- teriodorsum of the caudal area (30th to 39th myomere), as well as on the anterio- and midventral edge (14th to 26th myomere), posteroventral edge (28th to 39th myo- mere), and lateral midline (posterior to the 13th myomere) of the caudal area began to shrink (although they were still large [Tables 2-5] ). On day 3 after hatching, forebrain melanophores disap- peared in all specimens >3.76 mm BL, whereas the small- est specimen lacking these melanophores was 3.63 mm BL. All specimens >3.82 mm BL had melanophores scat- tered on the dorsum of the gut, which formed a cap of me- lanophores over the gut in later stages. Clusters of mela- nophores on the dorsal and ventral midline, and lateral midline of the body now appeared punctate and additional punctate melanophores appeared in these areas (Tables 2-5). Melanophores were present on the midbrain, on the anteriodorsal side and on the base of the hindbrain; and several melanophores appeared on the lower side of the lower jaw. The smallest specimen with melanophores on the lower jaw was 3.41 mm BL, whereas the largest lack- ing them was 9.46 mm BL (day 16). Most specimens >5.5 mm BL had these lower jaw melanophores. On and after day 4 (3.92-4.43 mm BL), all specimens had the melano- phore pattern of preflexion larvae (Fig. 2C). Melanophores began to develop on several other areas in larvae larger than about 4.5 mm BL (Fig. 2, D-F). Em- bedded melanophores appeared in the lateral muscle at 4.49 mm BL (Table 6), upper jaw tip at 5.47 mm BL, oper- culum and preoperculum at 6.18 mm BL, the membrane of the first dorsal fin at 6.32 mm BL, forebrain at 6.83 mm BL, and the cleithral symphysis at 10.79 mm BL. The larg- est specimens lacking melanophores in these areas were those that were 6.44 mm BL (for upper jaw tip), 6.72 mm BL (for lower jaw tip), 6.94 mm BL (for operculum), 9.69 mm BL (for forebrain), 10.28 mm BL (for cleithral sym- physis), and 10.49 mm BL (for preoperculum). Melanophores forming a dorsal cap of the gut enlarged at about 6.5 mm BL and the rim reached the ventral sur- face of the gut in some specimens >7.5 mm BL. The body melanophore pattern changed with growth. Dorsal mid- line melanophores disappeared from the trunk from 4.5 mm BL to 7 mm BL, and ventral and lateral midline me- lanophores disappeared from the trunk (Tables 2-5). Larvae showed partial juvenile pigment patterns from about 8 mm BL (Fig. 2F). Dorsolateral and dorsal midline melanophores appeared at the trunk and increased in fre- quency in the caudal area from 7.5 mm BL (Table 2). Em- bedded melanophores frequently appeared at the posterior trunk and caudal area (Table 6). Dense patches of mela- nophores appeared at the fin base of the first and second dorsal, and anal fin bases (Tables 7 and 8; these melano- phores occurred on the myosepta), as well as at the periph- ery of the eye, especially below and behind the eye from about 8.0 mm BL. Embedded melanophores appeared near the notochord and neural and haemal spines, and melano- phores at the lateral midline of body extended internally from about 8.2 mm BL. Generally, melanophores on the dorsal and ventral midline, lateral midline, and those em- bedded increased with BL (Tables 2-6). The smallest specimen having juvenile pigmentation (i.e. densely pigmented patches appearing on the body) was 13.55 mm BL (Fig. 2G). More distinct body patches appeared and increased in number as juveniles grew. In many larvae, melanophores were found in the area of the hypural bones and on the caudal finfold or fin rays. The smallest specimens having melanophores in the area of hypural bones and on the caudal finfold were 6.42 mm BL and 6.81 mm BL, respectively. Erythrophores first appeared at 4.63 mm BL (6 days af- ter hatching) on the tail: thirty-four erythrophores at the posteroventral edge, five at the posterolateral midline, and two at the posterodorsal edge (Table 9). As larvae grew, erythrophores appeared at the caudal finfold at 4.75 mm BL, on the lower jaw at 6.10 mm BL, and at the hypural plate at 6.96 mm BL; at the hypural plate and caudal finfold, erythrophores were observed only on specimens 6.96-7.20 and 4.75-7.20 mm BL, respectively. At the dor- sal and ventral edge of the trunk and tail, erythrophores became larger and decreased in number in larvae >7 mm. At the ventral edge, adjacent erythrophores were united. At the lateral midline of the body, the number decreased at 8.5 mm BL. Erythrophores disappeared from the lower jaw at 11.26 mm BL, from the posterolateral midline and dorsal edge of the tail at 15.85 mm BL, and from the ventral edge of the tail at 19.72 mm BL. Erythrophores were not observed on the upper jaw. The number of eryth- rophores, especially the ventral edge erythrophores, de- 606 Fishery Bulletin 99(4) Table 2 Incidence of dorsal midline melanophores at the trunk and tail (%) in Thunnus thynnus. Each melanophore is Myomere Size range (BL; mm) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 3.50-3.99 14 7 14 7 7 7 7 36 29 36 4.00-4.49 4 4 4 15 4 4 7 7 15 41 22 4.50-4.99 4 4 4 4 7 15 7 5.00-5.49 8 8 8 8 17 5.50-5.99 10 10 10 10 10 6.00-6.49 20 30 30 6.50-6.99 17 7.00-7.49 10 10 10 10 20 20 30 7.50-7.99 10 10 10 20 40 20 40 20 10 20 8.00-8.49 8 15 31 23 62 77 77 85 85 85 69 69 62 8.50-8.99 23 54 38 46 46 54 54 54 62 85 85 85 85 92 85 92 92 92 92 9.00-9.49 25 67 17 25 33 33 83 58 50 58 58 67 67 67 58 75 83 67 67 9.50-9.99 30 80 20 30 40 40 100 70 60 70 70 80 80 80 70 90 100 80 80 I- Table 3 1 Incidence of the ventral midline and ventrolateral melanophores of the trunk and tail (%) in Thunn us thyn nus. Each melanophore not observed on the first to sixth myomeres. Myomere Size range (BL; mm) 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 3.50-3.99 7 7 7 7 7 14 14 29 7 7 50 7 21 21 4.00-4.49 4 4 7 7 22 22 11 15 7 19 4.50-4.99 10 5 5 0 10 10 29 24 19 19 24 33 t 5.00-5.49 8 17 17 17 5.50-5.99 10 10 30 | i 6.00-6.49 10 10 30 10 6.50-6.99 6 11 6 7.00-7.49 20 10 7.50-7.99 8.00-8.49 8.50-8.99 8 8 8 9.00-9.49 9.50-9.99 creased as the fish grew; no erythrophores were observed in specimens >20 mm BL. Head spination Major head spines first appeared in specimens at 4—7 mm BL. The posterior preopercular angle spine and anterior preopercular spine appeared at 4.23 mm BL (Fig. 2D), and the posttemporal spine at 6.80 mm BL. The largest speci- men lacking anterior and posterior preopercular spines was 4.48 mm BL, and the largest specimen lacking posttempo- ral spines, 7.21 mm BL (Fig. 2E). No other spines or spine- lets of the head appeared in later developmental stages. The number of anterior and posterior preopercular spines increased 2-3 (nine maximum) and 5-7 (nine maximum), respectively, in the size range of 5-16 mm BL (Fig. 2, F Miyashita et al.: Morphological development and growth of Thunnus thynnus 607 Table 2 recorded for the myomere on which it occurs, n indicates the number of specimens examined at each size range. Myomere 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 n 29 29 7 7 7 7 14 14 21 21 36 43 14 14 7 14 14 7 22 7 7 19 15 4 11 19 4 19 19 33 30 26 30 15 19 27 7 11 4 4 4 7 11 7 11 15 4 11 15 19 15 19 22 7 7 21 8 8 17 25 8 17 8 17 8 8 8 17 17 17 25 25 8 12 10 10 10 10 10 30 10 20 10 10 10 20 10 20 20 30 10 20 10 10 50 40 50 10 10 10 10 22 11 6 6 6 6 6 11 6 22 22 6 11 22 6 18 10 10 10 20 10 20 10 30 70 30 20 10 20 20 20 10 20 10 20 0 10 30 20 20 10 30 40 10 40 10 10 54 38 38 31 23 38 31 23 23 15 31 8 31 15 23 8 15 15 8 38 13 92 77 77 77 69 69 69 69 69 62 54 69 69 69 54 54 62 46 38 23 13 83 67 58 75 58 83 58 58 67 58 75 50 50 33 33 25 58 67 33 17 12 100 80 70 90 70 100 70 70 80 70 90 60 60 40 40 30 70 80 40 20 10 Table 3 is recorded for the myomere on which it occurs, n indicates the number of specimens examined at each size range. Melanophores were Myomere 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 n 29 14 7 57 7 64 29 14 21 79 50 36 57 50 36 43 36 14 15 15 37 11 26 37 41 41 30 44 44 26 26 48 30 33 52 27 19 33 29 29 24 43 38 38 52 24 57 43 43 19 33 29 57 21 17 8 33 17 17 42 42 17 25 83 50 50 50 33 58 33 25 12 20 30 40 60 10 10 60 60 40 70 60 60 30 20 30 20 20 10 30 10 20 60 30 30 10 30 30 40 50 20 0 10 30 20 0 10 6 11 22 28 39 39 33 56 28 44 28 28 22 50 22 17 39 18 10 30 50 20 50 60 60 40 60 80 50 40 60 30 30 30 10 10 10 20 30 20 40 40 50 50 50 90 70 60 50 70 40 60 10 8 15 15 23 54 46 62 54 69 62 62 54 62 46 46 13 8 15 8 23 31 54 46 62 46 85 77 54 77 77 85 31 13 8 25 33 42 58 33 58 67 67 67 75 50 42 12 10 40 40 40 60 70 80 70 70 70 70 60 40 10 and G). The number of anterior and posterior preopercular spines decreased at 17 nun BL. At >19.23 mm BL, no ante- rior preopercular spines were observed, except in one spec- imen (24.00 mm BL) that had two anterior preopercular spines. Juveniles had three posttemporal spines when these spines were fully developed. The largest specimen with a posttemporal spine was 24.23 mm BL, and the largest spec- imen examined, 37.78 mm BL, had no spines on its head. Teeth and squamation Upper and lower jaw teeth were first observed at 5.35 and 6.32 mm BL, respectively. Palatine teeth appeared at 7.20 mm BL and most specimens >9 mm BL had fully developed palatine teeth. At >18.00 mm BL, specimens did not have these teeth. Squamation was incomplete in the size range examined in this study. The smallest scaled specimen (27.37 mm BL) 608 Fishery Bulletin 99(4) Table 4 Incidence of melanophores on the left lateral side of the trunk and tail (%) in Thunnus thynnus. Each melanophore is recorded for the the first to eighth myomeres. Myomere oize range (BL; mm) 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 3.50-3.99 7 7 7 14 7 29 14 21 7 29 14 7 14 4.00-4.49 4 11 4 7 11 11 11 11 11 4.50-4.99 5 5 5 5 10 5 10 5 29 10 5.00-5.49 8 8 8 8 17 8 5.50-5.99 20 20 20 20 10 40 10 6.00-6.49 20 10 20 10 6.50-6.99 6 0 6 6 6 28 11 17 22 7.00-7.49 7.50-7.99 20 20 10 20 20 20 20 8.00-8.49 15 8 15 8 8 8 15 8.50-8.99 8 15 8 8 8 9.00-9.49 8 17 17 9.50-9.99 10 10 10 20 Table 5 Incidence of melanophores on the right lateral side of the trunk and tail (%) in on the first to nineth myomeres. Thunnus thynnus . Each melanophore is recorded for Myomere £>ize range (BL; mm) 10 11 12 13 14 15 16 17 18 19 20 21 22 23 3.50-3.99 0 7 7 7 14 7 21 14 21 14 29 14 7 14 4.00-4.49 4 11 4 4 11 4 15 11 11 11 4.50-4.99 5 5 10 5 10 5 19 10 5.00-5.49 17 8 17 8 33 8 5.50-5.99 10 20 10 20 10 10 40 10 6.00-6.49 20 10 10 30 10 6.50-6.99 6 6 6 6 6 28 11 17 22 7.00-7.49 20 10 20 10 10 7.50-7.99 10 10 10 20 20 10 8.00-8.49 31 15 8.50-8.99 15 15 31 31 31 15 9.00-9.49 8 8 17 17 9.50-9.99 10 10 30 had anterior lateral line scales. Scales near the anterior part of the lateral line appeared at 30.31 mm BL, and postorbital scales at 30.81 mm BL. Discussion Development and growth Thunnus thynnus eggs were reported to have a smooth chorion, narrow perivitelline space, homogeneous yolk. and a single oil globule (Miyashita et ah, 2000), and this was reconfirmed in our study. These characteristics are similar to those found in other Thunnus eggs: T. thynnus from the Mediterranean (Podoa, 1956), T. obesus (Kikawa, 1953), T. albacares (Harada et al., 1971b; Mori et ah, 1971), and T. alalunga (Yoshida and Otsu, 1963). The growth strategy of bluefin tuna apparently is to de- velop foraging structures before other organs and at an early stage to enable feeding on larger organisms. Parts of the head were large from early development (Fig. 3) as mentioned by Collette et al. (1984) and reached maximum Miyashita et al.: Morphological development and growth of Thunnus thynnus 609 Table 4 myomere on which it occurs, n indicates the number of specimens examined at each size range. Melanophores were not observed on Myomere 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 n 14 36 7 21 7 7 14 14 7 14 7 14 7 14 30 11 7 15 4 4 11 11 4 7 11 7 7 7 27 14 10 19 14 5 5 5 5 19 10 14 5 10 5 21 17 8 8 8 33 8 17 25 8 8 12 20 10 10 20 20 40 10 10 30 10 10 10 10 20 10 20 10 10 11 11 6 11 17 6 6 11 6 11 6 17 11 6 18 20 10 10 10 10 20 10 10 10 10 10 20 20 10 20 20 30 10 10 40 20 10 10 8 8 15 8 15 31 31 8 8 8 13 15 8 8 31 15 23 15 23 15 15 8 8 8 8 8 13 8 8 17 17 25 25 42 58 33 50 33 50 33 33 50 42 12 10 10 20 20 20 50 60 40 50 40 30 20 50 10 Table 5 the myomere on which it occurs, n indicates the number of specimens examined at each size range. Melanophores were not observed Myomere 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 n 14 36 7 21 7 7 14 14 14 7 7 14 7 7 14 4 30 19 7 15 7 4 15 4 7 15 19 7 7 27 14 10 14 5 5 19 5 5 5 21 17 25 17 17 42 25 8 17 8 17 8 12 10 30 20 20 10 10 50 30 10 10 10 10 10 10 20 10 10 10 10 10 6 11 11 11 17 6 6 11 6 11 6 17 11 6 18 10 10 30 10 10 10 10 20 20 20 10 10 10 30 20 10 10 10 20 30 30 20 10 31 8 8 8 15 15 8 8 8 13 8 15 15 54 23 31 46 46 31 31 31 8 8 23 15 13 8 17 8 8 25 25 42 50 50 42 25 42 25 25 42 25 12 30 20 20 20 10 10 30 30 40 50 50 30 20 50 10 ratios in relation to BL at sizes smaller than other body parts, such as total length and body height, except for the caudal peduncle depth (Fig. 4). Another ontogenetic characteristic of growth in T. thyn- nus, the posterior migration of the anus (Collette et al., 1984), was also confirmed: the anus initially positioned in the anterior part of the body, was located at the center of the body at late flexion, and thereafter in the posterior part of the body (Fig. 3). Kaji et al. (1996) examined the relative growth of lar- vae of the Pacific bluefin tuna at 3-14 mm BL and report- ed that body proportions showed constant relative growth values from 10 mm BL, except for preanal length. Our re- sults showed that at 10 mm BL, constant ratios had not been reached in relation to BL. This difference may be due to an insufficient number of larger specimens in Kaji et al.’s ( 1996) study. Even at the juvenile stage, ratios in rela- tive BL decreased gradually, except for a few cases (Fig. 3), suggesting that body proportions of juvenile bluefin tuna are different from those of adult fish. Early growth is rapid in T. thynnus compared with oth- er marine fishes cultured in Japan (e.g. red sea bream. 610 Fishery Bulletin 99(4) Table 6 Incidence of melanophores embedded in the lateral muscle and on the myosepta (%) in Thunnus thynnus. Each melanophore is recorded on the first to nineth myomeres. Myomere Size range (BL; mm) 10 11 12 13 14 15 16 17 18 19 20 21 22 23 3.50-3.99 4.00-4.49 4 4.50-4.99 4 4 4 7 4 4 4 4 4 7 5.00-5.49 8 8 8 5.50-5.99 10 10 6.00-6.49 10 10 10 10 10 6.50-6.99 7.00-7.49 10 10 10 10 20 10 10 10 7.50-7.99 10 10 10 10 8.00-8.49 8 8 8 8 8 8 8 8 15 23 31 15 8.50-8.99 8 0 8 8 0 0 15 31 38 46 54 62 9.00-9.49 8 8 8 8 25 33 33 42 50 50 50 58 9.50-9.99 10 10 10 10 30 40 40 50 60 60 60 70 Table 7 Incidence of melanophores on the dorsal fin base (%) in Thunnus thynnus. Each melanophore is recorded for the nearest myomere, for larvae <7.50 mm BL. Myomere Size range (BL; mm) 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 7.50-7.99 10 10 10 10 10 20 20 20 30 20 30 20 30 20 20 20 20 10 8.00-8.49 8 8 8 8 8 15 15 15 23 15 23 15 23 15 15 15 15 8 8.50-8.99 8 15 31 46 46 46 46 46 54 38 54 31 31 31 31 46 46 9.00-9.49 8 17 25 50 58 83 83 83 83 83 83 83 83 83 83 83 83 83 9.50-9.99 20 40 70 80 100 60 60 60 60 100 100 100 100 100 90 100 100 Table 8 Incidence of melanophores on the anal fin base (%) in Thunnus thynnus. Each melanophore is recorded for the nearest myomere. n indicates the number of specimens examined at each size range. Melanophores were not observed on the first to twenty-first myomeres and for larvae <7.0 mm BL. Myomere Size range (BL; mm) 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 n 7.00-7.49 10 10 20 10 20 20 20 10 10 10 10 10 10 7.50-7.99 10 20 30 30 50 40 60 40 40 40 40 20 30 20 20 20 10 10 8.00-8.49 15 8 15 31 38 31 62 69 54 62 62 46 31 15 13 8.50-8.99 8 8 8 8 31 31 38 62 69 69 69 69 69 62 62 38 23 13 9.00-9.49 17 33 42 58 75 83 83 75 67 67 50 42 33 25 12 9.50-9.99 20 40 50 70 90 100 100 90 80 80 60 50 40 30 10 Miyashita et al.: Morphological development and growth of Thunnus thynnus 611 Table 6 for the myomere in which it occurs, n indicates the number of specimens examined at each size range. Melanophores were not observed Myomere 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 10 10 14 27 21 12 10 10 6 6 11 11 18 10 10 10 10 10 10 10 10 10 10 10 30 20 30 20 10 30 20 20 20 30 10 23 23 31 23 38 31 54 38 23 38 15 38 38 31 15 15 13 69 62 69 77 77 77 77 77 77 69 62 62 62 62 46 23 13 75 75 83 83 75 83 75 75 75 75 75 75 58 58 67 33 12 90 90 100 100 90 100 90 90 90 90 90 90 70 70 80 40 10 Table 7 n indicates the number of specimens examined at each size range. Melanophore was not observed on the first to fourth myomeres and Myomere 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 n 10 10 20 20 30 20 30 20 20 30 20 10 10 10 8 8 15 15 23 15 23 15 15 23 15 8 8 13 31 46 38 46 46 38 46 46 38 38 31 46 23 8 0 15 13 83 83 83 83 83 83 83 83 83 83 58 58 58 42 42 8 12 90 90 100 100 100 100 80 80 90 80 80 70 60 20 30 10 20 10 Pagrus major, and white trevally, Pseudocaranx dentex [Sawada unpubl. data; Fig. 5] ). Rapid growth in bluefin tu- na is pronounced after day 20 (if =9.34 mm BL on day 20) when larvae begin to metamorphose into juveniles. This type of growth phase seems common to scombrids, for ex- ample, T. albacares (Harada et al, 1971b; Kaji et al., 1999), Scomber japonicus (Watanabe, 1970), Auxis tapeinosoma (=A. rocheri , Harada et al., 1973), and Sarda orientalis (Harada et al., 1974) and reaches its height by the time of development of external physical features and the diges- tive system (Kohno et al., 1984; Miyashita et al., 1998), and by the time of development of their selective feeding characteristics (Sawada et al., 2000). On day 19 (x =10.17 mm TL), the mouth size of T. thyn- nus averaged 2.54 mm according to Shirota’s (1970) cal- culation (mouth size index=upper jaw length x 20-5). This size is comparable to that of other scombrids, such as T. albacares (3.3 mm) and Katsuwonus pelamis (3.0 mm) at 10 mm TL, although much larger than that of other spe- cies (see Table 2 of Shirota, 1970). The digestive system of 71 thynnus develops earlier than that of other fishes (Rich- ards and Dove, 1971; Miyashita et al., 1998) and attains an adultlike structure by the juvenile stage, thus allowing bluefin tuna to use food efficiently early in development. The development of pigment, spines, jaw teeth, and squamation in 71 thynnus is summarized in Figure 6. The jaw teeth of 7! thynnus first appeared at 5.35 mm BL (day 10), the palatine teeth at 7 mm BL (flexion stage), and were fully developed at 9 mm BL (postflexion stage). At 7 mm BL, 71 thynnus larvae fed on other fish larvae and at 8 mm BL (postflexion stage), they were cannibalistic. Teeth appearance and development corresponded to the piscivo- rous stage during the fast-growth phase after the postflex- ion stage. Future improved laboratory-rearing techniques 612 Fishery Bulletin 99(4) Table 9 Number of erythrophores on the body of Thunnus thynnus larvae and juveniles. Body length (mm) Dorsal edge Ventral edge Lateral midline Lower jaw Hypural plate Caudal fin finfold 3.38 3.41 4.63 2 34 5 4.75 8 32 4 1 5.02 4 2 5.09 4 36 5.11 3 54 4 5.35 24 4 5.78 1 2 4 6.10 1 24 5 4 6.72 3 1 6 6.91 1 1 4 6.96 2 17 6 2 2 6.96 1 8 3 7.03 8 6 1 7.06 8 12 8 1 5 2 7.12 8 10 5 7.20 6 8 6 1 3 1 7.92 1 8 6 1 8.16 6 8 3 8.41 1 11 8 3 8.74 1 4 5 8.81 4 7 34 2 8.92 3 3 1 9.02 1 9.35 3 12 5 3 9.38 3 2 3 9.56 2 2 1 9.60 9 3 1 9.70 5 1 6 9.93 2 2 2 10.03 1 23 10.56 4 1 2 7 10.97 4 1 3 11.26 1 15 12 2 11.71 12.20 4 13.71 5 14.00 14.79 3 1 14.83 1 15.33 2 4 1 15.49 15.85 1 3 2 16.86 18.73 1 19.71 19.72 2 20.40 21.87 24.85 25.89 Miyashita et al.: Morphological development and growth of Thunnus thynnus 613 FLS Mpc CZJ c Relative growth Preanal Caudal peduncle depth _ □ Snout _ □ Eye diameter J □ Head height □ Upper jaw □ Head □ Body height □ Total T 1 1 1 — 1 — 1 — 1 — 1 — 1 1 1 T — ^ 1 1 1 1 t" ■ ■ i ' ■ ■ ■ i ■ ' ■ . i ■ . i ■ ~r ■ ■ -t ■ | ■ , i i | ■ i i | 0 5 10 15 20 25 30 35 Body length (mm, BL) Figure 4 Schematic representation of the relative growth (length of body parts in relation to body length) in hatchery-reared bluefin tuna, Thunnus thynnus. Flexion larva subdivision (FLS) and numerical com- plement of fin rays (NCF) are also shown. In relative growth, □ = the peak value of body part proportion (in relation to BL) and ■ = the attainment of a constant value of body part proportion (in relation to BL). will allow further study of the structure and development of feeding-related bony elements for T. thynnus, as has been the case for S. japonicus (Kohno et al., 1984) and Lates calearifer (Kohno et al., 1996). Pigmentation Accurate identification of Thunnus larvae requires an extensive knowledge of individual, growth-associated, and geographic variations in melanophore and erythrophore patterns. We obtained information on the individual and growth-associated variations of these patterns from labo- ratory-reared specimens. The melanophore distribution pattern of T. thynnus lar- vae and juveniles showed four distinct characteristic peri- ods of development: newly hatched, 1-3 days after hatch- ing, preflexion to postflexion, and juvenile. T. thynnus thynnus larvae <3.99 mm BL from the Med- iterranean had embedded melanophores (Kohno et al., 1982); but we did not observe these in our specimens until 4.49 mm BL. This difference may be a subspecific differ- ence between T. thynnus orientalis and T. thynnus thyn- nus, or it may be due to insufficient numbers of specimens larger than 6.0 mm BL in Kohno et al.’s study, or to the experimental culture product as mentioned later. Wild-caught T. thynnus orientalis larvae (3.15-8.20 mm BL) most frequently had two melanophores both on the dorsal and ventral edges of the trunk and tail (Nishikawa, 1985). This pattern resembles that of T. thynnus thynnus larvae from the Mediterranean (2.53-5.25 mm BL, Kohno et al., 1982). But our cultured specimens in the same size range had greater numbers of these melanophores (Tables 2 and 3). Nishikawa (1985) reported the appearance of melanophores at the dorsal fin somewhat earlier than we observed their appearance, but their appearance on the lower and upper jaws occurred in the same size range in both studies where specimens were examined from the same population. Laboratory-reared fish larvae have been 40 r Figure 5 Growth of bluefin tuna (Thunnus thynnus ), red sea bream (Pagrus major), and white trevally ( Pseudocaranx dentex). Red sea bream and white trevally were reared at the same temperature range (24-27°C) as bluefin tuna (Sawada, unpubl. data). shown to have greater pigmentation than wild-caught lar- vae (e.g. Pagrus major [Fukuhara and Kuniyuki, 1978], Sparus sarba [Kinoshita, 1986], Parapristipotna trilinea- tum [Kimura and Aritaki, 1985], and Nibea mitsukurii [Kinoshita and Fujita, 1988]). Thus, the source of speci- mens, wild-caught or laboratory-reared, may account for observed differences in larval pigmentation between our study and Nishikawa’s study ( 1985). A slight difference in the occurrence of lower jaw, trunk, and tail melanophores has been reported among T. thyn- nus from the Mediterranean (Kohno et al., 1982), the At- lantic (Richards and Potthoff, 1974), and the Pacific (Mat- sumoto et al., 1972). Lower jaw melanophores appeared in all specimens larger than 3.0 mm BL from the Atlantic 614 Fishery Bulletin 99(4) Erythrophores Ventral notochord Dorsal notochord Lateral notochord Lower jaw Melanophores Gut Ventral tail Midbrain Forebrain Low. jaw tip Internal Upp. jaw tip First dorsal fin Anal fin base Dorsal fin base pectoral fin Spines Inner preopercular spine Preopercular spine Posttemporal spine Jaw teeth Upper jaw teeth Lower jaw teeth Squamation Lateral line scale Postorbital scale Corselet scale 0 5 10 15 20 25 30 35 Body length (mm, BL) Figure 6 Schematic representation of the development of pigment, spines, jaw teeth, and squamation in hatch- ery-reared bluefin tuna (Thunnus thynnus). □ = appearance of pigment and spines was subject to indi- vidual variation; ■ = pigment and spines were present in all specimens examined. and larger than 4.0 mm BL from the Pacific. In our speci- mens, the incidence of these melanophores was 7.1% at 3.50- 3.99 mm BL, 29.6% at 4.00-4.49 mm BL, 57.1% at 4.50 — 4.99 mm BL, 83.3% at 5.00-5.49 mm BL, and more than 90% in specimens >5.50 mm BL. The largest speci- men we observed lacking these melanophores was 9.46 mm BL. Our data are similar to those for specimens from the Mediterranean. The difference in lower jaw melano- phore distribution between our specimens and those of Matsumoto et al. (1972) might be explained by geographic variation of the melanophore pattern in the Pacific bluefin tuna. Further study is needed to confirm this hypothesis. Ueyanagi (1966) reported species-specific characteris- tics of the erythrophore distribution pattern of Thunni- nae larvae from the Pacific Ocean: T. alalunga consis- tently had more erythrophores on the dorsal edge of the trunk and tail in front of the caudal peduncle than did T. albacares , which had only one or two erythrophores at the caudal peduncle; T. thynnus and T. obesus had pat- terns transitional between those of T. alalunga and T. al- bacares. Our data for T. thynnus generally agreed with those of Ueyanagi (1966); however, the number of dorsal erythrophores in one individual ranged more (0—8) than in Ueyanagi’s study (1-5). In addition, ventral erythro- phores appeared in large numbers (more than 20) at the preflexion stage (4.63-6.10 mm BL). From examinations of tuna larvae taken in Hawaiian waters, Matsumoto et al. (1972) considered the erythrophore pattern useful as a morphological characteristic for identifying tuna larvae. At present, however, information is limited on erythro- phore patterns of other Thunninae larvae and juveniles and on the difference of these patterns in wild-caught and laboratory-reared specimens. Thus collection of such in- formation is needed to establish the erythrophore pattern as a species-identifying characteristic of Thunninae lar- vae and juveniles. Thunnus thynnus did not have xanthophores from hatching to the preflexion stage. However, T. albacares (Harada et al., 1971b; Mori et al., 1971) and T. obesus (Ya- sutake et al., 1973) have clusters of xanthophores in the Unfolds of both the dorsal and ventral fins, and on the dor- sal body, respectively. Thus, the xanthophore pattern can be used for distinguishing these Thunnus species at the preflexion stage. Much additional work on development of tunas under controlled conditions, as well as the study on wild-caught materials, is needed to understand their early life history. We believe recent progress in the technology of rearing tu- nas will yield important information on the early life his- tory of Thunnus spp. Miyashita et al.: Morphological development and growth of Thunnus thynnus 615 Acknowledgments We would like to express our thanks to I. Nakamura (Kyoto University), I. Kinoshita (Kochi University), K. L. Main (Mote Marine Laboratory), anonymous reviewers, and the editorial office of Fishery Bulletin for their helpful suggestions and advice. This paper is dedicated to the late T. Harada, who was the previous professor and director of the Fisheries Laboratory of Kinki University and who was one of the leaders of our tuna aquaculture study. Literature cited Cho, S., and S. Inoue. 1970. Intra- and interspecific restriction fragment length polymorphism in mitochondrial genes of Thunnus tuna species. Bull. Nat. Res. Inst. Far. Seas Fish. 30:207-225. Collette, B. B. 1999. Mackerels, molecules, and morphology. Soc. Fr. Ich- tyol. 25:149-164. Collette, B. B., and C. E. Nauen. 1983. FAO species catalogue. Vol. 2: Scombrids of the world: an annotated and illustrated catalogue of tunas, macker- els, bonitos and related species known to date, 137 p. FAO Fish. Synop. 125, FAO, Rome. Collette, B. B., T. Potthoff, W. J. Richards, S. Ueyanagi, J. L. Russo, and Y. Nishikawa. 1984. Scombroidei: development and relationships. In Ontogeny and systematics of fishes (H. G. Moser, W. J. Rich- ards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr. and S. L. Richardson, eds.), p. 591-620. Am. Soc. Ichthyol. Herpe- tol., Spec. Publ. 1. Fujita, K. 1998. Species and ecology of tunas. In From the produc- tion to the consumption of tunas ( S. Ono, ed. ), p. 1-49. Sei- zando, Tokyo. Fukuhara, O., and K. Kuniyuki. 1978. Morphological development in the wild larvae of Madai, Chrysophyrys major , as compared to that in the lab- oratory-reared larvae. Bull. Nansei Reg. Fish. Res. Lab. 11:19-25. Gibbs, R. H., Jr. and B. B. Collette. 1967. Comparative anatomy and systematics of the tunas, genus Thunnus. Fish. Bull. 66:65-130. Harada, T., H. Kumai, K. Mizuno, O. Murata, M. Nakamura, S. Miyashita, and H. Furutani. 1971a. On the rearing of young bluefin tuna. Mem. Fac. Agric. Kinki Univ. 4:153-157. [In Japanese with English summary.] Harada, T., K. Mizuno, O. Murata, S. Miyashita, and H. Furutani. 1971b. On the artificial rearing of larvae in yellowfin tuna. Mem. Fac. Agri. Kinki Univ., p. 145-152. [In Japanese with English summary.] Harada, T., O. Murata, and H. Furutani. 1973. On the artificial fertilization and rearing of larvae in marusoda, Auxis tapeinosoma. Kinki Univ. Fac. Agric. Bull. 6:113-116. [In Japanese with English summary.] Harada, T., O. Murata, and S. Miyashita. 1974. On the artificial fertilization and rearing of larvae in bonito. Kinki Univ. Fac. Agric. Bull. 7:1-4. [In Japanese with English summary.] Iwai, T., I. Nakamura and K. Matsubara. 1965. Taxonomic study of the tunas. Misaki Mar. Biol. Inst. Special Report 2:1-51. [In Japanese with English summary.] Kaji, T., M. Tanaka, Y. Takahashi, M. Oka, and N. Ishibashi. 1996. Preliminary observations on development of Pacific bluefin tuna Thunnus thynnus (Scombridae) larvae reared in the laboratory, with special reference to the digestive system. Mar. Freshwater Res. 47:261-269. Kaji, T., M. Tanaka, M. Oka, H. Takeuchi, S. Ohsumi, K. Teruya, and J. Hirokawa. 1999. Growth and morphological development of laboratory- reared yellowfin tuna Thunnus alhacares larvae and juve- niles, with special emphasis on the digestive system. Fish. Sci. 65:700-707. Kikawa, S. 1953. Spawning of bigeye tuna in the sea near the Marshall Islands. Rep. Southern Sea National Research Institute 24:156-158. [In Japanese.] Kimura, S., and M. Aritaki. 1985. Studies on rearing and development of larval and juvenile threeline grunt (Pisces: Haemulidae). Bull. Fac. Fish. Mie Univ. 12:193-205. Kinoshita, I. 1986. Postlarve and juveniles of silver sea bream, Sparus asrba, occurring in the surf zones of Tosa Bay, Japan. Japan. J. Ichthyol. 33:7-30. Kinoshita, I., and S. Fujita. 1988. Larvae and juvenile of blue drum, Nibea mitsukurii , occurring in the surf zones of Tosa Bay, Japan. Japan. J. Ichthyol. 35:25-30. Kohno, H., T. Hoshino, F. Yasuda, and Y. Taki. 1982. Larval melanophore patterns of Thunnus alalunga and T. thynnus from the Mediterranean. Japan. J. Ich- thyol. 28:461-465. Kohno, H., R. Ordonio-Aguilar, A. Ohno, and Y. Taki. 1996. Osteological development of the feeding apparatus in early stage larvae of the seabass, Lates calcarifer. Ich- thyol. Res. 43:1-9. Kohno, H., M. Shimizu, and Y. Nose. 1984. Morphological aspects of the development of swim- ming and feeding functions in larval Scomber japonicus. Bull. Japan. Soc. Sci. Fish. 50:1125-1137. Kumai, H. 1998. Studies on bluefin tuna artificial hatching, rearing and reproduction. Nippon Suisan Gakkaishi 64:601-605. Matsumoto, W. M., E. H. Ahlstorm, S. Jones, W. L. Klawe, W. J. Richards, and S. Ueyanagi. 1972. On the clarification of larval tuna identification par- ticularly in the genus Thunnus. Fish. Bull. 70:1-12. Miyashita, S., K. Kato, Y. Sawada, O. Murata, Y. Ishitami, K. Shimizu, S. Yamamoto, and H. Kumai. 1998. Development of digestive system and digestive enzyme activities of larval and juvenile bluefin tuna, Thun- nus thynnus, reared in the laboratory. Suisanzoshoku 46:111-120. Miyashita, S., Y. Sawada, N. Hattori, H. Nakatsukasa, T. Okada, O. Murata, and H. Kumai. In press. Mortality of northern bluefin tuna (Thunnus thyn- nus) due to trauma caused by collision during growout cul- ture. J. World Aqua. Soc. 31. Miyashita, S., Y. Tanaka, Y. Sawada, O. Murata, N. Hattori, K. Takii, Y. Mukai, and H. Kumai. 2000. Embryonic development and effects of water temper- ature on hatching of the bluefin tuna, Thunnus thynnus. Suisanzoshoku 48:199-207. Mori, K., S. Ueyanagi, and Y. Nishikawa. 1971. The development of artificially fertilized and reared 616 Fishery Bulletin 99(4) larvae of yellowfin tuna, Thunnus albacares. Bull. Far Seas Fish. Res. Lab. 5:219-232. [In Japanese with English summary.] NIH (National Institutes of Health). 1997. NIH image analysis software. NIH, Bethesda, MD, 104 p. Nishikawa, Y. 1985. Identification for larvae of three species of genus Thun- nus by melanophore patterns. Bull. Far Seas Fish. Res. Lab. 22:119-129. [In Japanese with English summary.) Orange, C. J., and B. D. Fink. 1963. Migration of a tagged bluefin tuna across the Pacific Ocean. Calif. Fish and Game 49:307-308. Podoa, E. 1956. Uova, larve e stadi giovanili di Teleostei; Scombrifor- mes. Fauna e Flora del Golfo di Napoli, Monogr. 38:471- 507. Richards, W. J., and G. R. Dove. 1971. Internal development of young tunas of the genera Kat- suwonus, Euthynnus,Auxis, and Thunnus. Copeia 1:72-78. Richards, W. J., and T. Potthoff. 1974. Analysis of the taxonomic characters of young scom- brid fishes, genus Thunnus. In The early life history of fish (J. H. S. Blaxter, ed.), p. 623-648. Springer- Verlag, Berlin, Heidelberg, New York. Sawada, Y., S. Miyashita, M. Aoyama, M. Kurata, Y. Mukai, T. Okada, O. Murata, and H. Kumai. 2000. Rotifer size selectivity and optimal feeding density of bluefin tuna, Thunnus thunnus, larvae. Suisanzoshoku 48:169-177. Shirota, A. 1970. Studies on the mouth size of fish larvae. Bull. Jap. Soc. Sci. Fish. 36:353-368. [In Japanese with English summary.) Takii, K., S. Miyashita, M. Seoka, Y. Tanaka, Y. Kubo, and H. Kumai. 1997. Changes in chemical contents and enzyme activities during embryonic development of bluefin tuna. Fisheries Sci. 63:1014-1018. Ueyanagi, S. 1966. On the red pigmentation of larval tuna and its use- fulness in species identification. Rep. Nankai. Reg. Fish. Res. Lab. 24:41-48. [In Japanese with English summary.) Watanabe, T. 1970. Morphology and ecology of early stages of life in Japa- nese common mackerel, Scomber japonicus Houttuyn, with special reference to fluctuation of population. Bull. Tokai Reg. Fish. Res. Lab. 62:1-259. [In Japanese with English summary.) Yabe, H., S. Ueyanagi, and H. Watanabe. 1966. Studies on the early life history of bluefin tuna Thun- nus thynnus and on the larva of southern bluefin tuna T. maccoyii. Rep. Nankai Reg. Fish. Res. Lab. 23:95-129. [In Japanese with English summary.) Yasutake H., G. Nishi, and K. Mori. 1973. Artificial fertilization and rearing of bigeye tuna ( Thunnus obesus ) on board, with morphological observa- tions on embryonic through to early post-larval stage. Bull. Far Seas Fish. Res. Lab. 8, June:71-78. [In Japanese with English summary.] Yoshida, H. O., and T. Otsu. 1963. Synopsis of biological data on albacore Thunnus germo (Lacepede) 1880 (Pacific and Indian Oceans). FAO Fish. Rep. 2(6):274-318. 617 Abstract— Age and growth were exam- ined of red snapper, Lutjan us campecha- nus, captured in an extensive (3100 km2) artificial reef area off Alabama in the northern Gulf of Mexico. Sagittal otoliths were removed from individu- als (n=1755) sampled from recreational catches and tournament landings. Mar- ginal increment analysis of sectioned otoliths revealed that a single opaque zone formed annually in sagittae from January through May. Fish ages were estimated from the number of opaque zones in otoliths, timing of opaque zone formation, sampling date, and a presumed birthdate of 1 July. Esti- mated growth of recaptured red snap- per (n=288) from a tagging experiment was similar to growth estimated from otolith-aged fish and corroborated oto- lith aging methods. The von Bertalanffy growth function fitted to length-at-age data was TL = 969 mm ( l-e-0192*'-0020*) (P<0.001; r2=0.99), which was similar to reported growth functions for west- ern Gulf of Mexico and Atlantic red snapper. Results of our study are con- sistent with the single stock hypothesis for Gulf of Mexico red snapper. Manuscript accepted 15 March 2001. Fish. Bull. 99:617-627 (2001). Age and growth of red snapper, Lutjanus campechanus, from an artificial reef area off Alabama in the northern Gulf of Mexico William F. Patterson III Coastal Fisheries Institute 204 Wetlands Resources Building Louisiana State University Baton Rouge, Louisiana 70803 E-mail address: wpatte2@LSU.edu James H. Cowan Jr Department of Marine Sciences University of South Alabama Mobile, Alabama 36688 Charles A. Wilson Department of Oceanography and Coastal Studies Louisiana State University Baton Rouge, Louisiana 70803-7503 Robert L. Shipp Department of Marine Sciences University of South Alabama Mobile, Alabama 36688, Red snapper, Lutjanus campechanus, are large, predatory reef fish belonging to the family Lutjanidae. They gen- erally are found from Cape Hatteras, North Carolina, to the Yucatan Penin- sula, including the waters of the Gulf of Mexico (GOM) but not the Caribbean Sea (Hoese and Moore, 1998). Through- out their range red snapper are dis- tributed along the continental shelf out to the shelf’s edge and demonstrate affinity for vertical structure. Adults aggregate on or near coral reefs, gravel bottoms, or rock outcrops, as well as on artificial reefs, oil rigs, and wrecks (Moseley, 1966; Szedlmayer and Shipp, 1994; Stanley and Wilson, 1997). Red snapper support economically valuable recreational and commercial fisheries in U.S. waters of the GOM (GMFMC, 1989). Federal management of GOM red snapper is based on the as- sumption that fish from Florida to Tex- as constitute a single stock. Although genetic evidence supports this assump- tion (Camper et al., 1993; Gold et ah, 1997 ; Heist and Gold, 2000 ), fish are not distributed uniformly across the north- ern GOM. Fisheries-dependent data suggest there is a center of red snapper abundance off southwest Louisiana and a second, smaller center of abundance off Alabama (Goodyear, 1995a; Schir- ripa and Legault, 1999). For example, from 1981 to 1998 estimated Louisiana landings of red snapper (commercial and recreational catch from state and federal waters) accounted for 32.6% (mean) ±1.5% (SE) of the total GOM harvest, whereas Alabama landings ac- counted for 11.4% ±0.9% of the total catch (Schirripa and Legault, 1999). Although fewer red snapper are har- vested from waters off Alabama than from the northwestern GOM, the red snapper fishery off Alabama is unique in several ways. Given that Alabama’s GOM coastline represents only about 3.0% of the coastline from Tampa, Flor- ida, to Brownsville, Texas, a dispropor- tionately high percentage of the GOM red snapper harvest is caught and land- 618 Fishery Bulletin 99(4) Map of artificial reef permit areas off Alabama. Date of creation is given for each area. ed there. The productivity of the red snapper fishery off Alabama occurs despite the fact that few high-relief (>1 m) natural reefs exist on the continental shelf in the north central GOM (Parker et ah, 1983; Shultz et ah, 1987; Schroeder et ah, 1989); however, off the coast of Alabama exists a 3100-km2 area designated for artificial reef de- ployment (Fig. 1). The correlation between high catch rates and the creation of artificial reefs off Alabama has caused some to speculate that artificial reefs have increased the productivity of the GOM red snapper stock in this area (Szedlmayer and Shipp, 1994; Minton and Heath, 1998). Despite the implied effect of artificial reefs on the red snapper fishery off Alabama, few studies have focused on red snapper in this area. The objective of our study was to estimate growth rates of adult red snapper captured off Alabama with otolith-aging and mark-recapture meth- ods and to compare growth estimates with growth of adult red snapper from the western GOM and the southeastern United States. As a corollary to our primary objective, we also attempted to validate presumed annual growth rings in otoliths with marginal increment analysis and present a comparison of estimated growth of otolith-aged red snapper with estimated growth of tagged individuals. Methods Otolith aging Red snapper were sampled from July 1995 to September 1999. All fish were caught over artificial reef sites off Ala- bama. Red snapper shorter than the legal size limit (380 mm total length [TL] for most of the study) were ran- domly sampled from undersize fish caught during research cruises to tag red snapper. Fish longer than the legal size limit were either randomly sampled from recreational catches or sampled opportunistically at spearfishing or hook-and-line fishing tournaments. Total length and fork length (FL) were measured to the nearest mm for all fish, and whole weight was measured to the nearest 0.1 g. Sex was determined for most fish by macroscopic examination of the gonads. Both sagittae were removed from each sam- pled individual, rinsed of any adhering tissue, and stored in paper coin envelopes until processing. Otoliths were sectioned in a transverse plane following the methods of Cowan et al. (1995) and were read under transmitted light with either a Micro Design® model 925 microfiche projector or an Optimas® image analysis sys- tem (Media Cybernetics, 1999). Otoliths were read inde- pendently by two readers. Blind counts of opaque zones of each sectioned otolith were made along the ventral margin of the sulcus acousticus from the core to the proximal sur- face; marginal increments were scored following Beckman et al. (1991) (Table 1, Fig. 2). Otoliths for which counts of opaque zones differed between readers were read a second time. Precision among readers was evaluated with the co- efficient of variation (CV) (Chang, 1982), index of preci- sion (D) (Chang, 1982), and average percent error (APE) (Beamish and Fournier, 1981). Age was estimated from the number of opaque zones in otolith sections, timing of opaque zone formation, assumed birthdate, and sampling date. It was assumed that opaque Patterson et al.: Age and growth of Lut/anus campechanus 619 Table 1 Scale used in determining condition of red snapper otolith marginal increments. Margin score Margin description 1 Opaque zone begins to form at edge; zone is <1/3 the thickness of the previ- ous opaque zone. 2 Opaque zone at edge is between 1/3 and 2/3 the thickness of the previous opaque zone. 3 Opaque zone at edge is >2/3 the thick- ness of the previous opaque zone. 4 Translucent zone begins to form at edge; zone is <1/3 the thickness of the previous translucent zone. 5 Translucent zone at edge is between 1/3 and 2/3 the thickness of the previ- ous translucent zone. 6 Translucent zone at edge is >2/3 the thickness of the previous translucent zone. zones constitute annuli and that annulus formation be- gins 1 January (see “Results” section). The birthdate for red snapper in the north central GOM was assumed to be 1 July, which follows the convention of Goodyear (1995a) and is based on the peak in red snapper spawning in the north central and northeastern GOM (Collins et al., 1996; Szedlmayer and Conti, 1999). According to these assump- tions, young-of-the-year northern GOM red snapper form their first annulus in sagittae beginning in January when fish are approximately 0.5 yr old. Therefore, the opaque zone closest to the otolith core represents only 0.5 yr of life, which was accounted for in the aging algorithm. Age (in d) was estimated (for most fish) by first sub- tracting one opaque zone from the total number of opaque zones in a given otolith and multiplying the difference by 365 d. Next, 182 d was added to the product to account for the first 0.5 yr of life. Finally, the day of year (number of days since 1 January) the fish was sampled was added to account for the number of days in the sampling year that the fish was alive. The result was divided by 365 d to es- timate age in years. Age was estimated similarly for fish that were sampled in November and December and had already begun forming opaque margins, except two was subtracted from the total number of opaque zones before multiplying by 365 d in order to assign fish to the correct year class. For fish that were sampled in January that did not have an opaque margin, zero was subtracted from the total number of opaque zones. Von Bertalanffy (1938, 1957) growth functions (VBGFs) were fitted to TL-at-age data with Proc NLIN in SAS (SAS Institute, Inc., 1996). Von Bertalanffy growth functions were fitted separately for males and females and a likeli- hood ratio test was used to test for difference in growth Figure 2 Digital images of (A) the proximal view of the left sagitta from a 408-mm-TL female red snapper sampled in August 1996 and (B) a thin section of a sagitta from a 683-mm-TL female red snapper sampled in January 1996. White squares on the thin section designate opaque zones (n=6); edge score is one. between sexes (Kimura, 1980; Cerrato, 1990). Von Berta- lanffy growth functions also were fitted for the complete data set and for the complete set excluding tournament sampled fish. Weight-TL relationships were modeled with non-linear regression for females and males following Ricker (1975). The functions were computed with Proc NLIN in SAS (SAS Institute, Inc., 1996). Sex-specific weight-TL rela- tionships were made linear by taking the log of weight and TL, and difference between sexes was tested with an anal- ysis of covariance (ANCOVA) on the log-transformed data (SAS Institute, Inc., 1996). Lastly, a weight-TL nonlinear regression was computed for females and males combined with Pi'oc NLIN in SAS (SAS Institute, Inc., 1996). Mark-Recapture A tagging study of adult red snapper was conducted off Alabama from March 1995 to August 1999. Fish were caught with hook and line over nine artificial reef tagging sites in the Hugh Swingle general permit area for arti- 620 Fishery Bulletin 99(4) 100 200 300 400 500 600 700 800 900 1000 1100 TL (mm) Figure 3 Distribution of total length (TL) of red snapper sampled for age and growth estimation. ficial reef deployment (labeled 1986 in Fig. 1). Each fish was measured to TL and FL, tagged with a yellow Floy® (Seattle, WA) internal anchor tag, and released. Individu- als were recaptured on subsequent tagging trips or were recovered by recreational and commercial fishermen. Red snapper recaptured on tagging trips were measured to TL and FL and recovered fish were measured when fishermen retained carcasses of tagged fish. To corroborate red snapper age estimates, VBGF para- meters estimated from otolith-aged fish were incorporated into Fabens’ (1965) length increment model to predict TL at recapture of tagged fish (Labelle et al., 1993; Thompson et al., 1999). Solving for TL at recapture, Fabens’ model is computed as TLrt = 77, + (L„ - *,-) ( 1 - e~kAt.), ( 1 ) where TLr, = predicted TL at recapture of individual i; TLt = TL at release of individual i; Loo= TL asymptote from VBGF estimated from otolith-aged fish; K = growth coefficient from VBGF estimated from otolith-aged fish; and t{ = time at liberty of individual i. Predicted TL at recapture from Fabens’ method was plot- ted against observed TL at recapture to compare growth model predictions to observed values. Growth of tagged red snapper was estimated by regress- ing their change in TL on days at liberty (SAS Institute, Inc., 1996). A linear regression also was computed on TL- at-age data over the size range of recaptured individuals from the tagging study (SAS Institute, Inc., 1996). Slopes of resultant regressions were compared to assess if esti- mated growth rates were different between tagged and otolith aged fish. Results Age and growth Sagittae were collected from 1755 red snapper, including 360 fish shorter than the minimum size limit (380 mm TL) caught on tagging cruises, 289 fish from tournaments, and 1106 fish randomly sampled from recreational catches (Fig. 3). Sex was not determined for 279 fish, of which 61 individuals were immature. Mean TL (±SE) was 518.3 (±5.04) mm for males and 529.5 (±5.96) mm for females. Total length of immature individuals ranged from 208 to 309 mm. Of the 1755 otoliths sectioned for age determination, reader 1 and reader 2 agreed on the number of opaque zones for 1610 (91.7%) fish after the first reading; opaque zones in 23 otoliths were deemed not interpretable owing to sample preparation. Count disagreement between read- ers was the following: one opaque zone for 123 fish, two zones for 18 fish, and three zones for four fish. Otoliths were read a second time if reader counts were not in agree- ment and second otoliths were sectioned and read of the 23 fish for which otoliths were rejected after the first read- ing. After the second reading, agreement was reached for 1676 (95.6%) otoliths, including all 23 otoliths that were second sections. Disagreement between readers after the second reading was as follows: one opaque zone for 71 fish, two zones for 7 fish, and three zones for one fish; fish for Patterson et al.: Age and growth of Lutjcinus compechanus 621 which agreement was not reached after the sec- ond reading were not assigned ages or included in the growth estimation. Precision estimates were 1.25% for APE, 0.90% for CV, and 0.64% for D after the second reading. A clear pattern exists in the marginal in- crement scores, demonstrating that one opaque zone is formed annually in winter (Fig. 4). Most otoliths had opaque margins by January and had translucent margins by June, thus, timing of opaque zone formation appears to be from January through May for most fish. Eight fish (of 118) sampled in November (1997) and six fish (of 66) sampled in December (1996) had opaque margins. Four fish (of 47) sampled in January (1997) had translucent margins. The oldest female sampled was 34.1 yr old and the oldest male sampled was 33.2 yr old. The female to male ratio was 1.1:1 overall but was 1.5:1 for fish between 10 and 20 yr old and 3.4:1 for fish greater than 20 yr old (Fig 5, A and B). Von Bertalanffy growth functions for females and males mod- eled separately were Females: TL = 976( i_e-°-19102df=1 =0.2879), therefore, sexes were modeled together. Von Bertalanffy growth functions computed for all fish and excluding tour- nament sampled fish were All fish: TL = 969( 1-^-° 192(M)020>) <^3;1.672=47’690; P<0.001; r2=0.99) (Fig. 6) Excluding tournament fish: TL = 1181(l-e_0 120(;+a652)) n= 97 82 - O O 66 47 O * translucent opaque 31 31 78 77 57 97 23 118 O O 17 41 o O O • • • • SONDJ FMAMJ J ASONDJ FM 1996 1997 1998 Month Figure 4 Plot of monthly margin edge scores of red snapper sagittae. Symbol size is relative to the percentage of fish with the corresponding margin edge score. Monthly sample size is given. (F. =43,550; P<0.001; r2=0.99). 3;1,461 The growth model including all fish was similar to other VBGFs estimated for GOM and southeast U.S. Atlantic red snapper (Fig. 7, Table 2). Weight-TL nonlinear regression models for females and males modeled separately were Females: Weight = (4.46 x 10_9)TL3 18 (P1;73= 42,577; P<0.001; r2=0.98) Males: Weight = (5.18 x 10~9)TL3 16 (F1;728= 43,956; P<0.001; r2=0.99). Log-transformed weight-TL relationships were not signifi- cantly different between males and females (ANCOVA test for equal slopes; F1 1 465=1.54, P=0.2145; ANCOVA test for equal intercepts; P1 1 465=1.49, P= 0.2226), therefore, sexes were modeled together. The resultant nonlinear regres- sion model was 400 300 a> t 200 100 k\\\i Unknown ■■■ Females Males Figure 5 Distributions of age for (Al all sampled red snapper and ( B ) all fish greater than 10 yr old. Legend in section A is the same for both distributions. 622 Fishery Bulletin 99(4) Figure 6 Scatterplot of red snapper TL-at-age data for all fish assigned ages. Plotted line is the von Bertalanffy growth function fitted to the data. Weight = (4.68 x 10 ~9)7X317 (FV1 467=85,961; PcO.OOl; r2=0.98). Mark-Recapture A total of 2932 adult red snapper were tagged. Total length at recapture was measured for 288 of 519 recaptured individuals. Mean TL (±SE) of recaptures was 344 (±4.1) mm (range: 183-660 mm) at release and 423 (±5.6) mm (range: 253-726 mm) at recapture, and mean time at lib- erty (±SE) was 334 (±16.3) d (range: 12-1501 d). Predicted TL from Faben’s method corresponded well to observed TL at recapture for tagged fish (Fig. 8). The linear regression of change in TL on days at liberty was statistically sig- nificant (F1286=631.0, P<0.001, r2=0.76) and had a slope of 0.238 mm/d (Fig. 9A). The range in estimated ages of otolith-aged fish for the linear regression of TL on age was 356-2599 d (approximately one to seven years). The model was statistically significant (F11 541=6309, PcO.OOl, r2=0.80) and had a slope of 0.240 mm/d (Fig. 9B). Discussion Validating the periodicity of opaque zone formation in oto- liths as annual is imperative for age and growth studies where otoliths are used as aging structures (Beamish and McFarlane, 1983, 1987). Annual formation of opaque zones in otoliths has been validated for several tropical and su- tropical lutjanids (Manooch, 1987; Fowler, 1995; Cappo et al., 2000), and annual formation of opaque zones in red snap- per sagittae has been reported from the northwestern GOM and the southeastern U.S. Atlantic (Render, 1995; Manooch and Potts, 1997; Wilson and Nieland, 2001). Our marginal increment analysis of red snapper otoliths demonstrates that opaque zones in adult red snapper sagittae also are formed annually in the north central GOM. The pattern of monthly marginal increment scores reveals that some fish begin opaque zone formation as early as November and some do not have translucent margins until midsummer; however, the general pattern of opaque zone formation is from Jan- uary through May. Render (1995) and Wilson and Nieland (2001) have reported a similar pattern for adult red snapper in the northwestern GOM, and opaque zone formation in winter-spring has been shown for sagittae of several other teleosts in the northern GOM (Maceina et al., 1987; Beck- man et al., 1989, 1990, 1991; Thompson et al., 1999). Because relatively few old red snapper were sampled in our study, and samples of old fish came mostly from tour- naments in summer months, we were unable to validate ages beyond 8 years. Baker ( 1999) compared age estimates of otolith-aged GOM red snapper with age estimates from radiometric dating of their otolith cores and reported that age estimates from the two methods were highly correlat- ed for fish up to 37 yr old, thus corroborating otolith-based estimates of age for older fish. Therefore, despite lack of annulus validation for older red snapper ages in our study, we are confident that otolith-based age estimates of older individuals are accurate. Age and growth Red snapper are long-lived reef fish; we observed a maxi- mum age of 34 yr for females and 33 yr for males. Other Patterson et al.: Age and growth of Lutjanus campechcinus 623 Table 2 Results from previous studies of the age and growth of Gulf of Mexico and southeastern U.S. Atlantic red snapper. Study and location Aging structure Maximum age (yr) L ^ mm (TL) K ^0 Szedlmayer and Shipp (1994); north central GOM sectioned otoliths 42 1025 0.150 not reported Wilson and Nieland, 2001; northwest GOM sectioned otoliths 53 938 0.175 -0.530 Nelson and Manooch (1982); GOM scales and sectioned otoliths 15 941 0.170 -0.10 Nelson and Manooch (1982); southeast U.S. Atlantic scales and sectioned otoliths 15 975 0.160 0.00 Manooch and Potts ( 1997); southeast U.S. Atlantic sectioned otoliths 25 955 0.146 0.182 Figure 7 Von Bertalanffy growth functions estimated from TL-at-age data for Gulf of Mexico red snapper. Legend indicates source of each function. authors have reported fish over 40 yr old (Szedlmayer and Shipp, 1994; Render, 1995; Wilson and Nieland, 2001) and the oldest fish aged to date is 53 yr old (Render, 1995; Wilson and Nieland, 2001). Among western Atlantic lutjanids for which maximum age has been reported, GOM red snapper has the greatest longevity (Acosta and Appledoorn, 1992; Manickchand-Heileman and Phillip, 1996; Potts et al., 1998; Hood and Johnson, 1999). Maximum ages of 30+ and 40+ yr have been reported for several species of Pacific lutjanids (reviewed in Rocha-Oliva- res, 1998). Overall, the numbers of sampled males and females were nearly equal, but females were predominant in samples greater than 10 yr old. A similar pattern was observed in Pacific red snapper, Lutjanus peru , where the female-to-male ratio was essentially equal (1.1:1) for fish less than 10 yr old and 2.4:1 for fish older than 10 yr (Ro- cha-Olivares, 1998, Fig. 6). Rocha-Olivares (1998) concluded from catch curve analyses that differences in sex-specific numbers at age for L. peru resulted from males experi- encing a higher mortality rate. Differential mortality between sexes is not unusual in lutjanids (Grimes, 1987) and the predominance of female GOM red snapper in older age classes may result from a higher mortality rate for males. Fish sampled from tournaments were included in growth estimation because few large, old individuals were sam- pled randomly from the recreational catch. The inclusion of tournament sampled fish could bias growth estimates because tournament anglers target large fish; thus the po- tential exists for them to catch or spear the fastest grow- ing individuals at a given age (Ottera, 1992; Vaughan and Burton, 1993; Goodyear, 1995b). Without the fish sampled from tournaments, however, the VBGF did not reach an asymptote; therefore growth parameters were poorly esti- mated (Hirschhorn, 1974). We feel that excluding the tour- nament fish from growth function estimation introduced far greater bias than including them. Mark-Recapture Comparisons between estimated growth of tagged red snapper and otolith-aged fish corroborate otolith aging methods. Predicted TL of tagged individuals obtained with Fabens’ (1965) method and VBGF parameters estimated from otolith-aged fish are coincident with observed TL at recapture because predicted and observed values corre- sponded well to the line of 1:1 agreement. Although ours 624 Fishery Bulletin 99(4) was a graphical rather than a statistical exercise (Labelle et al., 1993; Thompson et ah, 1999), results corroborate otolith-based estimates of growth. Direct comparison of linear growth functions computed for tagged fish and otolith-aged fish provides further sup- port for otolith-based estimates of growth (slopes from the two equations differed by only 0.002 mm/d). Francis (1988) cautioned against direct comparison of age-based growth estimates with growth estimates from tagging data because different information results from the dif- ferent data types. The linear model fitted to length-at-age data predicts TL of fish for a given age (in d), whereas the linear model fitted to tagging data predicts the increment of growth expected of a fish at liberty for a given number of days. However, because the relationship between TL and age is linear over the range of TL from the tagging data (Szedlmayer and Shipp, 1994), we submit that the slopes can be compared as estimates of mean growth. The slope of the age-based linear model is an estimate of the pop- ulation growth rate of young fish, whereas the slope of the length-based linear model is an estimate of the mean growth of tagged individuals. That growth estimates for otolith-aged and tagged red snapper are very similar cor- roborates the aging method with otoliths and indicates that on average tagging did not affect fish growth. Red snapper off Alabama The red snapper fishery off Alabama is unique in several ways. Despite the reported lack of high-relief natural reef structures on the continental shelf off Alabama (Parker et al., 1983; Shultz et al., 1987; Schroeder et al., 1989), fish- ermen land a disproportionately high percentage of the annual GOM red snapper harvest from Alabama coastal waters (Schirripa and Legault, 1999). Although Alabama red snapper landings have accounted for approximately 11% of total GOM landings from 1981 to 1998, the recre- ational fishery off Alabama accounted for mean (±SE) 21.1 (±2.25)% of GOM recreational landings from 1981 to 1998 and 26.5 (±1.83)% of recreational landings from 1990 to 1998 (Schirripa and Legault, 1999). Total landings of red snapper from the U.S. GOM averaged 3.5 x 103 metric tons from 1981 to 1998, with commercial and recreational land- ings essentially equal; however, 94% of Alabama red snap- per landings from 1990 to 1998 came from the recreational sector of the fishery (Schirripa and Legault, 1999). The red snapper fishery off Alabama is also unique be- cause it is prosecuted almost entirely over artificial reefs (Minton and Heath, 1998; Shipp, 1999). Artificial reef con- struction began in the 1950s when charter boat operators gained permission from the Marine Resources Division of the Alabama Department of Conservation to place 250 car bodies on the sea floor off Alabama (Hosking and Swingle, 1989; Minton and Health, 1998). Since the 1950s, artifi- cial reefs have been constructed of a variety of materials including car bodies, liberty ships, army tanks, kitchen appliances, newspaper bins, and most recently, prefabri- cated concrete structures. Reef building increased dra- matically in the 1980s with the creation (by permit) of ar- tificial reef areas that now encompass a total of 3100 km2, Patterson et al.: Age and growth of Lutjanus campechcinus 625 and it is estimated that over 15,000 artificial reefs cur- rently exist off Alabama (Hosking and Swingle, 1989; Szedlmayer and Shipp, 1994; Minton and Health, 1998; Havard1). The correlation of catch rates with artificial reef con- struction has caused some to conclude that artificial reefs off Alabama have increased the production of red snapper, as opposed to merely aggregating fish from surrounding areas (Szedlmayer and Shipp, 1994; Min- ton and Health, 1998). Among the evidence cited in sup- port of this conclusion was that red snapper grew fast- er, attained larger sizes, and lived longer over artificial reefs off Alabama than other places throughout their range (Szedlmayer and Shipp, 1994). More recent data, including those from the current study, indicate that GOM red snapper off Alabama grow similarly to and reach similar sizes as fish from the northwestern GOM and the Atlantic (Render, 1995; Manooch and Potts, 1997; Wilson and Nieland, 2001). It is also apparent that the maximum longevity of fish off Alabama is no greater than that of fish of other locations in the GOM (Szedlmayer and Shipp, 1994; Render, 1995; Wilson and Nieland, 2001). These results have important implications for man- agement of GOM red snapper. That growth of fish off Alabama is similar to growth of fish in the north- western GOM is consistent with the management par- adigm that northern GOM red snapper constitute a single stock (Goodyear, 1995a) and contrary to the hy- pothesis that fish off Alabama are unique (Szedlmayer and Shipp, 1994). Moreover, if fishing mortality rates are higher off Alabama than other places in the north- ern GOM but growth is the same, production of red snapper may be lower off Alabama than in other areas. Furthermore, if red snapper recruit to artificial reefs off Alabama from other areas in the northern GOM, Alabama’s red snapper fishery may serve as a net sink for stock-specific production. Future research is needed to compare mortality as well as growth of fish from dif- ferent areas in the northern GOM and to estimate the source of recruits to Alabama’s artificial reefs. Acknowledgments Funding for this project was provided by NOAA through MaRFIN (grant numbers NA57FF0054 and NA87FF0424). We thank Scott Baker, Forrest Davis, Andy Fischer, Walter Ingram, Jessica McCawley, Dave Nieland, and Melissa Woods for help in sampling red snapper and sectioning and reading otoliths. We thank Mike and Ann Thierry for allowing us to tag red snapper onboard their charter fishing vessel Lady Ann. We thank volunteer fish- ermen who helped on tagging cruises and commercial 1 Havard, R. 1999. Personal commun. Alabama Department of Conservation, Marine Resources Division, P.O. Box 189, Dau- phin Island, AL 36528. and recreational fishermen who reported recaptured fish. Lastly, we thank Amada Gonzales for collecting data from fishermen who reported recaptured fish. Literature cited Acosta, A., and R. S. Appeldoorn. 1992. Estimation of growth, mortality, and yield per recruit for Lutjanus synagris (Linnaeus) in Puerto Rico. Bull. Mar. Sci.. 50:282-291. Baker, M. S„ Jr. 1999. Radiometric age validation of red snapper, Lutjanus campechanus, and red drum, Scianenops ocellatus, from 626 Fishery Bulletin 99(4) the northern Gulf of Mexico. Master’s thesis, Louisiana State Univ., Baton Rouge, LA, 59 p. Beamish, R. J., and D. A. Fournier. 1981. A method for comparing the precision of a set of age determinations. Can. J. Fish. Aquat. Sci. 38:982-983. Beamish, R. J., and G. A. McFarlane. 1983. The forgotten requirement for age validation in fish- eries biology. Trans. Am. Fish. Soc. 112:735-743. 1987. Current trends in age determination methodology. In Age and growth of fish (R. C. Summerfelt and G. E. Hall, eds.), p. 15-42. Iowa State Univ. Press, Ames, IA. 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. 1991. Age and growth of sheepshead Archosargus probatoceph- alus in Louisiana waters using otoliths. Fish. Bull. 89:1-8. 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. 87:17-28. Camper, J. D., R. C. Barber, L. R. Richardson, and J. R. Gold. 1993. Mitochondrial DNA variation among red snapper (Lutjanus compechanus ) from the Gulf of Mexico. Mol. Mar. Biol. Biotech. 2:154-161. Cappo, M., P. Eden, S. J. Newman, and S. Robertson. 2000. A new approach to validation of periodicity and timing of otolith opaque zone formation in the otoliths of eleven species of Lutjanus from the central Great Barrier Reef. Fish. Bull. 98:474-488. Cerrato, R. M. 1990. Interpretable statistical tests for growth comparisons using parameters in the von Bertalanffy equation. Can. J. Fish. Aquat. Sci. 47:1416-1426. Chang, W. B. 1982. A statistical method for evaluating the reproducibility of age determination. Can. J. Fish. Aquat. Sci. 39:1208-1210. Collins, L. A., A. G. Johnson, and C. P. Kiem. 1996. Spawning and annual fecundity of the red snapper ( Lutjanus ca/npechanus) from the northeastern Gulf of Mexico. ICLARM Conf. Proc. 48:174-188. Cowan, J. H., Jr, R. L. Shipp, H. K. Bailey IV, and D. W. Haywick. 1995. Procedure for rapid processing of large otoliths. Trans. Am. Fish. Soc. 124:280-282. Fabens, A. J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. Fowler, A. J. 1995. Annulus formation in otoliths of coral reef fish — a review. In Recent developments in fish otolith research (D. H. Secor, J. M. Dean, and S. E. Campana, eds.), p. 45-63. Univ. South Carolina Press, Columbia, SC. Francis, R. I. C. C. 1988. Are growth parameters estimated from tagging and age-length data comparable? Can. J. Fish. Aquat. Sci. 45:936-942. Gold, J. R., F. Sun, and L. R. Richardson. 1997. Population structure of red snapper from the Gulf of Mexico as inferred from analysis of mitochondrial DNA. Trans. Am. Fish. Soc. 123:386-396. Goodyear, C .P. 1995a. Red snapper in U.S. waters of the Gulf of Mexico. Stock assessment report MIA-95/96-05. Miami Labora- tory, Southeast Fisheries Science Center, Miami, FL. 1995b. Mean size at age: an evaluation of sampling strate- gies using simulated red grouper data. Trans. Am. Fish. Soc. 124:746-755. Grimes, C. B. 1987. Reproductive biology of Lutjanidae: a review. In Tropical snappers and groupers: biology and fisheries man- agement (J. J. Polivina and S. Ralston, eds.), p. 239-294. Westview Press, Inc., Boulder, CO. GMFMC (Gulf of Mexico Fishery Management Council). 1989. Amendment number 1 to the reef fish fishery man- agement plan. GMFMC, Tampa, FL. 356 p. Heist, E. J., and J. R. Gold. 2000. DNA microsatellite loci and genetic structure of red snapper in the Gulf of Mexico. Trans. Am. Fish. Soc. 129:469-475. Hirschhorn, G. 1974. Effect of different age ranges on von Bertalanffy parameters in three fishes and one mollusk of the north- eastern Pacific Ocean. In The aging of fish (T. Bagenal, ed.), p.192-199. Unwin Brothers, Ltd., London. Hoese, H. D., and R. H. Moore. 1998. Fishes of the Gulf of Mexico, 2nd ed. Texas A&M Press, College Station, TX, 422 p. Hood, P. B., and A. K. Johnson. 1999. Age, growth, mortality, and reproduction of vermillion snapper, Rhomboplites aurorubens, from the eastern Gulf of Mexico. Fish. Bull. 97:828-841. Hosking, W., and H. A. Swingle. 1989. Alabama’s artificial reefs. MS-AL Sea Grant Publ. 9-019, Mobile, AL, 6 p. Kimura, D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull. 77:765-776. Labelle, M., J. Hampton, K. Bailey, T. Murray, D. A. Fournier, and J. R. Sibert. 1993. Determination of age and growth of south Pacific albacore ( Thunnus alalunga ) using three methodologies. Fish. Bull. 91:649-663. 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. Manickchand-Heileman, S. C., and D. A. T. Phillip. 1996. Reproduction, age, and growth of Caribbean red snapper (Lutjanus purpureus) in waters off Trinidad and Tobago. ICLARM Conf. Proc. 48:137-149 Manooch, C. S., III. 1987. Age and growth of snappers and groupers. In Trop- ical snappers and groupers: biology and fisheries man- agement (J. J. Polivina and S. Ralston, eds.), p. 329-373. Westview Press, Inc., Boulder, CO. Manooch, C. S., Ill, and J. C. Potts. 1997. Age and growth of red snapper, Lutjanus campecha- nus, Lutjanidae, collected along the southeastern United States from North Carolina through the east coast of Flor- ida. J. Elisha Mitchell Sci. Soc. 113:111-122. Media Cybernetics. 1999. Optimas, version 6.5. Media Cybernetics, Silver Spring, MD. Minton, R. V., and S. R. Health. 1998. Alabama’s artificial reef program: building oases in the desert. Gulf of Mex. Sci. 16:105-106. Moseley, F. N. 1966. Biology of the red snapper, Lutjanus aya Bloch, of the northwestern Gulf of Mexico. Publ. Inst., Mar. Sci. Univ. Texas 11:90-101. Nelson, R. S., and C. S. Manooch. 1982. Growth and mortality of red snappers in the west-cen- tral Atlantic ocean and northern Gulf of Mexico. Trans. Am. Fish. Soc. 111:465-475. Patterson et al.: Age and growth of Lutjanus campechcinus 627 Ottera, H. 1992. Bias in calculating growth rates in cod (Gadus morhua L.) due to size selective growth and mortality. J. Fish Biol. 40:465-467. Parker, R. O., D. R. Colby, and T. D. Willis. 1983. Estimated amount of reef habitat on a portion of the U.S. south Atlantic and Gulf of Mexico continental shelf. Bull. Mar. Sci. 33:67-86. Potts, J. C., C. S. Manooch III, and D. S. Vaughan. 1998. Age and growth of vermillion snapper from the south- eastern United States. Trans. Am. Fish. Soc. 127:787-795. Render, J. H. 1995. The life history (age, growth, and reproduction) of red snapper (Lutjanus campechanus) and its affinity for oil and gas platforms. Ph. D. diss., Louisiana State Univ., Baton Rouge, LA, 76 p. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can., Ottawa, 382 p. Rocha-Olivares, A. 1998. Age, growth, mortality, and population characteristics of the Pacific red snapper, Lutjanus peru, off the southeast coast of Baja California, Mexico. Fish. Bull. 96:562-574. SAS Institute, Inc. 1996. Statistics, version 6.11. SAS Institute Inc., Box 8000, Cary, NC. Schirripa, M. J., and C. M. Legault. 1999. Status of red snapper in U.S. waters of the Gulf of Mexico updated through 1998. Report SFD-99/00-75, Sus- tainable Fisheries Division, Miami Laboratory, National Marine Fisheries Service, Miami, FL. Schroeder, W. W., A. W. Shultz, and J. J. Dindo. 1989. Inner-shelf hardbottom areas, northeastern Gulf of Mexico. Trans. Gulf Coast Asso. Geol. Soc. 38:535-541. Schultz, A. W., W. W. Schroeder, and J. R. Abston. 1987. Along-shore and offshore variations in Alabama inner- shelf sediments. In Coastal sediments and processes, 9th symposium of coastal sedimentology, p. 141—152. Florida Sate Univ., Tallahassee, FL. Shipp, R. L. 1999. The artificial reef debate: are we asking the right ques- tions? Gulf of Mex. Sci. 17:51-55. Stanley, D. R., and C. A. Wilson. 1997. Seasonal and spatial variations in the abundance and size distribution of fishes associated with a petroleum plat- form in the northern Gulf of Mexico. Can. J. Fish. Aquat. Sci. 54:1166-1176. Szedlmayer, S. T., and J. Conti. 1999. Nursery habitats, growth rates, and seasonality of age-0 red snapper, Lutjanus campechanus , in the northeast Gulf of Mexico. Fish. Bull. 97:626-635. Szedlmayer, S. T., and R. L. Shipp. 1994. Movement and growth of red snapper, Lutjanus campechanus, from an artificial reef area in the northeast- ern Gulf of Mexico. Bull. Mar. Sci. 55:887-896. Thompson, B. A., M. Beasley, and C. A. Wilson. 1999. Age distribution and growth of greater amberjack, Seriola dumerili, from the north-central Gulf of Mexico. Fish. Bull. 97:362-371. Vaughan, D. S., and M. L. Burton. 1993. Estimation of von Bertalanffy growth parameters in the presence of size-selective mortality: a simulated exam- ple with red grouper. Trans. Am. Fish. Soc. 123:1-8. von Bertalanffy, L. 1938. A quantitative theory of organic growth. II. Inquiries on growth laws. Human Biol. 10:181-213. 1957. Quantitative laws in metabolism and growth. Quart. Rev. Biol. 32:217-231. Wilson, C. A., and D. L. Nieland. 2001. Age and growth of red snapper, Lutjanus campecha- nus, from the northern Gulf of Mexico off Louisiana. Fish Bull. 99:653-664. 628 Abstract— Calcified structures of sum- mer flounder, Paralichthys dentatus, were evaluated to identify the best age determination method. Scales, the cur- rently preferred structure, were com- pared with opercular bones and to right and left whole and sectioned otoliths for ages 0 to 10. All structures showed concentric rings that were interpreted as annual; however structures differed greatly in the clarity of their presumed annual marks. Right and left otoliths generally gave the same age, although they differed in the clarity of marks. Sec- tioned otoliths, particularly right ones, were the best aging structure. Right sec- tioned otoliths consistently showed the clearest marks and had the highest confi- dence scores, lowest reading times, and highest agreement within and between readers, 97% and 96%, respectively. Left sectioned otoliths took twice as long to prepare and were more difficult to interpret than right sectioned otoliths. Whole otoliths were the second best structure and were adequate to age 4 or 5, after which sectioning greatly improved the clarity of marks. Scales were inferior to, and often did not give the same age readings as, whole and sectioned otoliths. Compared with otoliths, scales tended to overage at younger ages and to underage at older ages. Opercular bones were undesir- able for aging summer flounder. They were often unclear and inconsistent, and they had the lowest confidence scores, the highest reading times, and only 46% within-reader agreement. A major source of disagreement in scale and otolith age readings was the pres- ence of an early, presumably false, mark on some structures. We compare the formation of this early mark in summer flounder with early mark formation on otoliths of Atlantic croaker, a species with similar life history traits. Manuscript accepted 21 March 2001. Fish. Bull. 99:628-640 (2001). A comparison of calcified structures for aging summer flounder, Paralichthys dentatus* Ann M. Sipe Virginia Institute of Marine Science College of William and Mary Gloucester Point, Virginia 23062 E-mail address: amsipe@vims.edu Mark E. Chittenden Jr. Virginia Institute of Marine Science College of William and Mary Gloucester Point, Virginia 23062 The summer flounder, Paralichthys den- tatus, ranges from Nova Scotia to Flor- ida, although it is most abundant from Massachusetts to North Carolina (Gins- burg, 1952; Leim and Scott, 1966; Gutherz, 1967). In regions of high abun- dance, it is one of the most important commercial and recreational fishes on the Atlantic coast (MAFMC, 1987). In the Chesapeake Bay region, for example, summer flounder support an extensive recreational fishery from about March to November, when they are present in the lower portions of the Chesapeake Bay and in coastal waters (Hildebrand and Schroeder, 1928; MAFMC, 1987; Desfosse, 1995). They then support a strong commercial fishery during the fall and winter, when they move offshore to the continental shelf (Ginsburg, 1952; Bigelow and Schroeder, 1953; Poole, 1962; MAFMC, 1987). Many studies have reported difficul- ties with the structures used for age de- termination of summer flounder. Prior to about 1980, whole left otoliths were the most commonly used structure (Poole, 1961; Eldridge, 1962; Smith and Daiber, 1977; Powell, 1982). However, there were disagreements over the location and in- terpretation of the first presumed an- nual mark (Poole, 1961; Eldridge, 1962; Smith and Daiber, 1977), largely a result of uncertainties about first year growth rates. This and other problems with whole otoliths (summarized in Smith et al., 1981) prompted a comparison of age determination structures by Shepherd (1980), who reported that presumed an- nual marks were more distinct on scales than on whole otoliths. Consequently, scales became the preferred structure for aging summer flounder (Smith et al., 1981; Dery, 1988; Almeida et al., 1992). More recently, Szedlmayer et al. (1992) examined first year growth rates to resolve the location and interpreta- tion of the first mark on whole otoliths, but scales have remained the preferred structure (Bolz et al., 2000). Difficulties have also been reported in using summer flounder scales (Dery, 1988; Desfosse, 1995; Bolz et al., 2000). Desfosse (1995) used marginal incre- ment analysis to validate scales for ages 1 to 3. He reported only 46% within-read- er agreement past age 4, however, indi- cating that marks on scales are not very distinct at older ages. He attributed dis- agreements to false or indistinct annuli and to crowding of annuli at the scale edge in older fish. Most recently, Bolz et al. (2000) reported only 53% agree- ment for ages 1 to 5 in a between-agency exchange of scales, with agreement in- creasing to only 83% after they resolved as many disagreements as possible. They attributed most of the remaining dis- agreements to the choice of a first an- nual mark and to differing opinions on what constituted a false mark on scales. A reexamination of calcified struc- tures for aging summer flounder is needed, given their economic impor- tance and the reported difficulties in * Contribution 2380 from the Virginia Insti- tute of Marine Science, College of William and Mary, Gloucester Point, VA 23062 Sipe and Chittenden: A comparison of calcified structures for aging Paralichthys dentatus 629 age determination with whole otoliths and scales. Pre- vious studies have never evaluated sectioned otoliths in summer flounder, even though sectioned otoliths have of- ten proven a superior structure in other species, especially at older ages when scales and other structures can under- age fish (Beamish and McFarlane, 1983). Further study is especially needed because the location of the first mark on otoliths has recently been determined (Szedlmayer et al., 1992). In addition, no work has been done to determine if right-left differences in the location of the focus result in differences in age determination. The main objective of our study was therefore to evalu- ate and compare whole otoliths, sectioned otoliths, scales, and opercular bones for aging summer flounder. We in- cluded opercular bones because many studies, on a variety of species, have found them to be superior to other struc- tures and to have very distinct and easy to read marks (for examples, see LeCren, 1947; Donald et al., 1992; Hostet- ter and Munroe, 1993). A second objective was to compare right and left otoliths for potential differences in age based on differences in the location of the focus. Calcified struc- tures were evaluated in terms of preparation and reading times, confidence in presumed annual mark clarity, agree- ment between repeated age readings, structure growth with fish growth, age agreement between different struc- tures of the same fish, and increases in the number of pre- sumed annual marks with structure size and fish size. Fi- nally, we discuss the formation of early, presumably false, marks on summer flounder otoliths and scales that result- ed in difficulties in age interpretation Methods Sample collection To minimize difficulties interpreting marks on the edge of the structures, collections of summer flounder were made far from the time of presumed annual mark formation, which occurs in May and June on the scales of Chesa- peake Bay summer flounder (Desfosse, 1995). Summer flounder were collected from commercial fisheries in the Chesapeake Bay region from September through Novem- ber of 1998 (n=165). Additional juvenile fish (n- 11) were collected by the Virginia Institute of Marine Science juve- nile bottom trawl survey in October of 1998 in the lower Chesapeake Bay and James River. Fish were processed for total length (TL), total weight (TW), and sex, and the calcified structures were removed as follows. Both saggital otoliths were removed, wiped clean, and stored dry in tissue culture cell wells. Scales were removed from just above the lateral line anterior to the caudal peduncle (Shepherd, 1980; Dery, 1988) and stored in coin envelopes. Both opercular bones were re- moved according to the methods of LeCren (1947), stored in coin envelopes, and frozen. The collection of summer flounder was stratified into six length-based categories of 100 mm each to include as many age groups as possible in the final study sample. A ran- dom sample of 15 fish was then chosen from the first five categories. The last category included the six largest fish, all of which were used in the comparison, for a total of 81 fish. All calcified structures in the final study sample were assigned random numbers before preparation and aging. Summer flounder in the final study sample ranged in size from 209 to 758 mm TL and from 80.8 to 7304.6 g TW and in age from 0 to 10 years (determined from sectioned otoliths, as reported in this study). Preparation of calcified structures for age determination Whole otoliths were examined in water on a dark back- ground with reflected light at 120 to 240x magnification. Thin opaque bands, which appeared white under reflected light, were presumed to represent annual marks (Fig. 1A). Two counting paths were used for mark enumeration. The primary counting path was from the focus to the anterior margin of the otolith. The secondary counting path, used to verify the primary counting path reading, was from the focus to the posterior margin of the otolith. With calipers to 0.05 mm, whole otolith total length (WOTL) was mea- sured as the largest distance from the anterior to the pos- terior edge and whole otolith radial length (WORD was measured from the center of the focus to the tip of the anterior edge. A paired sample t-test was used to test for right-left differences in WORL. After all whole otolith readings were made, right and left otoliths were mounted sulcal groove down onto card- board with crystal bond adhesive and sectioned trans- versely through the focus with a variable speed Beuhler Isomet saw. The resulting sections, about 0.5 mm thick, were mounted on clear glass slides and immersed in crys- tal bond. Sections were viewed with transmitted light and bright field at 240x magnification. Thin opaque bands, which appeared dark with transmitted light, were pre- sumed to represent annual marks (Fig. IB) and were counted along the ventral side of the sulcal groove. Sec- tioned otolith radial length (SORL) was measured to 0.001 mm along the ventral arm of the sulcal groove from the center of the focus to the otolith edge by using a compound video microscope with the Optimas image analysis sys- tem (Media Cybernetics, 1999). Broken otoliths were not measured if they were fractured along the focus. A paired sample t-test was used to test for right-left differences in SORL. Opercular bones were prepared according to the meth- ods of LeCren (1947). Briefly, they were soaked in cold tap water for several minutes to thaw and to partially loosen surrounding skin, then soaked for 1 minute in simmering water, after which the skin was easily removed with a toothbrush. The opercular bones were then rinsed with cold tap water and air-dried. Opercular bones were exam- ined dry with transmitted light and in water with reflected light on a dark background. Presumed annual marks (Fig. 1C) were defined as sharp transitions from relatively nar- row translucent zones to relatively wide opaque zones that were continuous from the anterior to the posterior mar- gin of the bone (Bagenal and Tesch, 1978; Hostetter and Munroe, 1993). Translucent zones appeared white under 630 Fishery Bulletin 99(4) Figure 1 Marks on calcified structures taken from a 5-year-old (determined from sectioned otoliths) female summer flounder, TL = 687 mm, collected in mid-January. Arrows indicate individual marks counted (as described in the “Methods” section). (A) Whole otoliths, viewed with reflected light on a black background. Arrows indicate presumed annual marks along the primary counting path; dots indicate presumed annual marks along the secondary path. Whole otolith radial lengths were measured along the pri- mary counting path. ( B ) Transverse otolith sections, viewed in transmitted light. Sectioned otolith radial lengths were measured along the counting path (indicated by arrows) from the center of the focus to the edge along the ventral arm of the sulcal groove. (C) Right opercular bone, viewed with reflected light on a black background. AA = articular apex. (D) Scale impressions, viewed in transmitted light. White arrows indicate marks that appear on only one of the scales. Asterisks indicate probable false marks. Both scales have a probable false mark prior to the first mark counted. transmitted light and dark under reflected light, whereas opaque zones appeared dark under transmitted light and white under reflected light. The first presumed annual mark was defined as the first opaque zone after the first translucent zone, where the first translucent zone occu- pied the central focal area of the opercular bone. Both bones were examined, and the one with the clearest marks was used for aging. Opercular bone radial length (OpRL) was measured to 0.05 mm from the center of the articular apex to the anterior margin edge with calipers. Scales were soaked in water until flexible and brushed gently with a soft bristle toothbrush. Then 5 or 6 clean, symmetrical, unregenerated scales were dried, taped to an acetate sheet, inserted between two new acetate sheets, and pressed in a Carver laboratory scale press for 2 minutes at 15,000 pounds of pressure and 60°C. Scale impressions were read with a Bell-Howell R753 micro- fiche reader at 20x and 32x. Presumed annual marks were identified with standard scale reading criteria as de- scribed in Smith et al. (1981), Dery (1988), and Almeida Sipe and Chittenden: A comparison of calcified structures for aging Paralichthys dentatus 631 et al. (1992). Briefly, readers enumerated marks (Fig. ID) that exhibited “cutting over” in both lateral fields of the scale that was accompanied by a clear narrow zone in the anterior portion of the scale. Scale radial length (ScRL) was measured to 0.001 mm from the center of the focus to the anterior edge of the scale by using a compound video microscope with the Optimas image analysis system (Me- dia Cybernetics, 1999). Evaluation of calcified structures Each structure was examined for age by two readers — twice by reader 1 and once by reader 2. Structures were read in a randomly selected order with no knowledge of fish size or collection date. Ages were assigned on the basis of presumed annual mark counts. Different struc- tures from the same fish were read independently, includ- ing right and left otoliths, and at least one week separated the first and second readings of the same structure. Preliminary evaluations of structures included prepara- tion times, reading times, confidence in the clarity of pre- sumed annual marks, growth of the structures with size of the fish, and agreement in repeated age readings of the same structure (precision). Structures judged accept- able based on those criteria were then evaluated further for agreement in age readings between different struc- tures from the same fish and to see if the number of pre- sumed annual marks increased with structure size and fish size. Our preliminary evaluation indicated otoliths and scales to be superior to opercular bones; therefore opercular bones were not evaluated further. Preparation time, a measure of the processing efficien- cy of a structure, was evaluated as the time taken to pre- pare structures for reading. Clarity of presumed annual marks on a structure was evaluated using both reading times and confidence scores. Reading time was measured as the time taken to read a given structure in an indi- vidual fish. Confidence scores, expressed on a scale of 1 (low) to 5 (high), were assigned by the reader to each read- ing based on the clarity of the marks. Differences in confi- dence scores between structures were tested at a = 0.05 by using the normal approximation to the Mann-Whitney test for ordinal data (Zar, 1996). The assumption that structure growth is directly relat- ed to fish growth was evaluated using regression analysis (Zar, 1996). Structure sizes (ScRL, OpRL, WOTL, WORL, SORL) were regressed on fish TL to determine if the re- lationships were significant and increasing. Sample sizes varied in these regressions, and in regressions of the num- ber of presumed annual marks on structure size described below because some structures were broken in prepara- tion and could not be measured. Precision in age determinations for a given structure was evaluated using simple percent agreement in repeat- ed readings within and between readers. Within-reader agreement compared the first and second readings by reader one, and between-reader agreement compared the first readings of each of the two readers. Reader comments on structure features were evaluated to determine the proximal causes of disagreements. Scales that disagreed in the initial two readings by reader 1 were reread independently a third time by reader 1 to reach a consensus for use in between-structure compar- isons. Likewise, right and left otoliths that disagreed in the initial two readings by reader 1 were read a third time to reach a consensus. Structures that showed no agree- ment in three readings (1 of 81 for scales, 1 of 81 for sec- tioned otoliths) were not included in between-structure comparisons. Agreement in presumed annual mark counts between different structures of an individual fish was evaluated by using simple percent agreement between structures and simple linear regression procedures. For the regressions, ages determined by one structure were regressed on ages determined by another structure, and the slope of the re- gression line was tested to see if it differed significantly from one. A slope of one implies that y = x and that the two structures give the same age. For each regression, we used as the x-variable the structure judged to be superior in the preliminary evaluations. The assumption that the number of presumed annual marks on a structure is directly related to structure size and to fish size was evaluated using regression analysis (Zar, 1996). The number of presumed annual marks on a structure was regressed on structure size (ScRL, WOTL, WORL, SORL) and on fish TL to determine if the relation- ships were significant and increasing. Results Comparative appearance of calcified structures All four calcified structures showed concentric marks that were interpreted as annual (Fig. 1). However, these struc- tures differed greatly in the clarity of presumed annual marks. Presumed annual marks on both whole and sectioned otoliths (Fig. 1, A and B) were typically clear, consistent, and easy to interpret, especially for sectioned otoliths. The right-left difference in the location of the focus had moder- ate effects on mark clarity for both whole and sectioned otoliths, as described below. Whole otolith marks were most easily read at younger ages, but age had little effect on sectioned otolith mark clarity. The few disagreements in otolith ages were primarily caused by an early, presum- ably false, mark that often occurred prior to the first pre- sumed annual mark (Fig. 2). This early mark appeared as a thin opaque band close to, but distinct from, the focus and was found on both young (Fig. 2A) and older (Fig. 2B) fish. We tried not to count this early mark in our age read- ings, because it did not occur consistently in all fish. Fi- nally, only one otolith of 81 pairs was poorly calcified and unable to be read whole, although its age was easily deter- mined upon sectioning. Presumed annual marks on opercular bones (Fig. 1C) were fairly clear in some fish, but they were more often poorly defined, inconsistent, and difficult to follow across the structure, making age interpretation difficult and high- ly subjective. Opercular bones commonly exhibited un- 632 Fishery Bulletin 99(4) Figure 2 Right whole otoliths showing an early, presumably false, mark. (A) is from a 299-mm-TL age-1 fish collected in September, and ( B ) is from a 442-mm age-4 fish collected in October. White arrows point to the early marks. Black arrows indicate primary counting path (anterior field), dots indicate secondary counting path (pos- terior field). clear transitions from translucent to opaque zones; the first one or two marks were particularly difficult to distinguish, even on young fish. Zone transitions were often easier to interpret towards the edge of the structure in older fish, although this too varied greatly from fish to fish. The example in Figure 1C is unusually clear and easy to read. Presumed annual marks on scales (Fig. ID) were clearer than those on opercular bones, but they still required much subjective interpretation. Figure ID shows some of the common problems encountered with scales, including presumably false marks (as- terisks) and marks that were present on only some scales from the same fish (white arrows). In addi- tion, many fish had regenerated, asymmetrical, or otherwise damaged scales, making it difficult and time-consuming to choose acceptable scales to press. For example, about 20 scales were pressed in order to obtain two scales that were adequate to show in Figure ID. Interpretation of age from scales of older fish was extremely difficult because marks at the scale edges were often obscured or crowded together, particularly in the narrow lateral fields. Finally, a major source of disagreement in age determination from scales resulted from an early, presumably false, mark that often occurred prior to the first presumed annual mark (Fig. ID, asterisk). Because this early mark did not appear consistently in all fish or even on several scales from the same fish, we tried not to count it in our age readings. Preparation times, reading times, and confidence in clarity of marks Preparation times were short and reasonable for all structures, at less than 15 minutes per fish. Whole oto- liths took by far the shortest time because no preparation was required before reading (Table 1). Sectioned right otoliths and opercular bones required 4 to 6 minutes to prepare, whereas scales and sectioned left otoliths took much longer to prepare, about 11 and 14 minutes, respec- tively. Left sectioned otoliths took much longer to prepare than right sectioned otoliths primarily because they broke much more frequently during sectioning. Reading times were short and reasonable for all struc- tures, at less than three minutes per fish. Sectioned right otoliths had by far the shortest reading time, at only 0.27 minutes per fish (Table 1). Whole otoliths and sectioned left otoliths had the next shortest reading time, at only about 0.4 to 0.6 minutes per fish. Scales (1.2 min) and opercular bones (2.4 min) both required much more reading time than otoliths, indicating that otoliths could be aged more easily. Reader confidence scores varied greatly between struc- tures. Sectioned otoliths had by far the highest confidence scores, with values of 4.9 and 4.8 for the right and left, re- spectively (Table 1). Whole otoliths had somewhat lower confidence scores, with values of 4.1 and 3.8 for the right and left, respectively. Confidence scores were much lower for scales (3.2) and especially for opercular bones (2.3), indi- cating that these structures were not as easily interpreted. All confidence scores were significantly different from one another (Z=2.10 to 4.18; P<0.0001 to 0.013; individual val- ues not reported). Regression of structure size on fish size All calcified structures grew in size as summer flounder body length grew, indicating that each structure could be useful for back-calculation studies. All regressions of structure size on total length were significant at P < 0.001, and all slopes were positive (Table 2). All regressions were strong and explained much of the variation in structure size, generally 90% or more, with coefficient of determina- tion values (100 r2) ranging from 72% to 98%. Values for 100 r2 were less than 91% only for right and left sectioned otoliths, which were 72% and 85%, respectively. Agreement in age determinations for the same structure Agreement ( precision) between repeated age readings varied greatly between calcified structures. Precision by the same reader was highest by far (95% to 97%) for sectioned right and left otoliths and left whole otoliths (Table 3). Precision was somewhat lower in right whole otoliths (89%) than in Sipe and Chittenden: A comparison of calcified structures for aging Parcilichthys dentatus 633 Table I Average preparation times (min), reading times (min) ± standard error (SE ), and confidence scores ( ±SE ) for summer flounder calcified structures. Preparation Reading Confidence Structure time time score Opercular bones 4.63 2.43 ±0.20 2.31 ±0.16 Scales 10.50 1.20 ±0.13 3.21 ±0.15 Sectioned otoliths Right 5.86 0.27 ±0.04 4.91 ±0.04 Left 13.93 0.57 ±0.09 4.75 ±0.05 Whole otoliths Right 0.00 0.45 ±0.06 4.10 ±0.11 Left 0.00 0.41 ±0.04 3.84 ±0.10 left whole otoliths; however this could be attributed to the reader learning to use reflected lighting more effectively during the second reading, because 7 of the 9 consensus readings for right otoliths agreed with the second reading. Within-reader agreement was lower with the use of scales (80%), but precision varied with age. Agreement in repeated scale readings was actually high for ages 0 to 4 (92%, n= 52), but it decreased to only 59% for fish over age 4 (/?= 29). Preci- sion was lowest by far in opercular bones (46%), where there were no patterns in agreement by age. Because opercular bones showed the lowest precision and the poorest mark clarity, we did not include them in further evaluations. Agreement in age determinations between readers also varied greatly among calcified structures. Precision be- tween readers was highest by far (96%) for right sectioned otoliths (Table 3). Agreement was somewhat lower (86% to 88%) for left sectioned otoliths and whole otoliths. Agree- ment was lowest by far for scales (58%), reflecting the over- all poor clarity of marks and the resulting subjectiveness in scale age readings compared with otolith age readings. Comparison of right and left otoliths Differences in right and left radial lengths were observed for both whole and sectioned otoliths. The right radial length was significantly shorter than the left in whole otoliths (paired 7=17.59, df=73, P<0.0001; Fig. 1A). How- ever, for sectioned otoliths, the right radial length was sig- nificantly longer than the left (paired t =-11.72, df=43, P<0.0001; Fig. IB) because the right otolith is thicker at the focus, where the transverse cross section was taken. Right and left whole otoliths generally gave the same age readings. Reader one had high age agreement between right and left whole otolith readings (96%), and the null hypothesis that the slope of the line equals one was not rejected (P=0.077, Fig. 3A). Although right and left whole otoliths generally indicat- ed the same age, they differed in mark clarity. When the posterior field (secondary counting path) was used to verify Table 2 Regression statistics for relationships between structure size and summer flounder total length (TL). Structure abbreviations are defined in the “Methods” section of the text, n = sample size. All regressions were significant at P< 0.001. Structure Equation n 100 r2 Opercular bones OpRL = -2.280 + 0.0772 TL 66 98 Scales ScRL = -0.348 + 0.0126 TL 81 93 Sectioned otoliths Right SORL = -0.015 + 0.0027 TL 66 85 Left SORL = 0.015 + 0.0018 TL 47 72 Whole otoliths Right WORL = 0.642 + 0.0089 TL 76 91 Left WORL = 0.601 + 0.0111 TL 77 93 Right WOTL = 1.280 + 0.0164 TL 76 94 Left WOTL = 1.530 + 0.0156 TL 77 91 Table 3 Average percent agreement, within and between readers, for presumed annual mark counts on summer flounder cal- cified structures. Structure Within reader Between reader Opercular bones 46 — Scales 80 58 Sectioned otoliths Right 97 96 Left 95 88 Whole otoliths Right 89 86 Left 97 87 or determine the number of presumed annual marks, the right otolith was generally much easier to read than the left because of the greater distance between the focus and the posterior margin on the right otolith (Fig. 1A). This greater distance made the marks further apart and more easily dis- tinguishable on the right than on the left otolith. The dif- ference in mark clarity was greatest for older fish and was also reflected in significantly higher confidence scores for the right whole otolith than for the left (Table 1). Right and left sectioned otoliths also generally gave the same age readings. Reader one had high age agreement between right and left sectioned otolith readings (94%), and the null hypothesis that the slope of the line equals one was not rejected (P=0.393, Fig. 3B). 634 Fishery Bulletin 99(4) Although right and left sectioned otoliths generally gave the same age, presumed annual marks were usually clear- er and easier to interpret on the right otolith. Right sec- tioned otoliths had a much longer counting path and were therefore easier to age than left sectioned otoliths, where the marks were more crowded and less clearly defined (Fig. IB). This difference was also reflected in higher con- fidence scores and lower reading times for the right sec- tioned otolith than for the left (Table 1). Comparison of different calcified structures from the same fish Whole and sectioned otoliths generally gave the same age readings. The number of presumed annual marks on whole and sectioned otoliths showed high agreement (95%), with 100% agreement for fish under age 4 (Fig. 4A). In addition, the null hypothesis that the slope of the line equals one was not rejected (P=0.901). Although whole and sectioned otoliths generally provid- ed the same age, presumed annual marks were often clear- er on sectioned otoliths than on whole ones, especially in older fish, where crowding of marks at the edge of whole otoliths became a problem. This observation is supported by the much higher confidence scores for sectioned otoliths (Table 1). As a specific example, the oldest fish in the com- parison showed very clear marks and was aged 10 in ev- ery reading using both right and left sectioned otoliths (Fig. 5A), and all confidence scores were 5. Marks were less clear on the whole otolith (Fig. 5B), however, with between 8 and 10 marks counted in different readings, and an average confidence score of only 2.5. In general, the use of sectioned otoliths appeared to greatly in- crease mark clarity in fish over age 4 or 5. Scales and sectioned otoliths often did not give the same age readings. Agreement in the number of presumed annual marks on scales and sectioned otoliths was undesirably low, at only 80% (Fig. 4B). In addition, the null hypothesis that the slope of the line equals one was rejected (P=0.047). Scales tended to overage compared with sectioned otoliths in fish age 4 and younger, but to underage in fish older than age 4. Agreement between scales and sectioned otoliths was fairly high for ages 0 to 4 (86%, u=56) but decreased to only 65% in fish over age 4 (ti=23). Scales and whole otoliths often did not give the same age readings. Agreement in the number of presumed annual marks on scales and whole oto- liths was also undesirably low, at only 76% (Fig. 6). In addition, the null hypothesis that the slope of the line equals one was again rejected (P=0.039). As with sectioned otoliths, scales tended to over- age compared with whole otoliths in fish age 4 and younger and to underage in fish older than age 4. Agreement between whole otoliths and scales was fairly high for ages 0 to 4 (85%, 77=53) but decreased to only 56% in fish over age 4 (77=25). Increase in number of marks with structure size and fish size Mark counts on calcified structures increased as structure size and fish size increased, indicating that each structure tested could be useful in age determination. All regressions of mark counts on structure size were significant at P<0.001, and all slopes were positive (Table 4). Regressions were generally strong and explained much of the vari- ation in mark counts because 100 r2 values were high, generally from 80% to 86%. Values for 100 r2 were lowest for left sectioned otolith radius and scale radius, at 67% and 73%, respectively. Likewise, all regressions of mark counts on fish size were significant at P < 0.001, and all slopes Sipe and Chittenden: A comparison of calcified structures for aging Paralichthys dentatus 635 were positive (Table 5). All regressions were again strong, with 100 r2 values from 83% to 86%. Discussion Comparative evaluation of sectioned otoliths Our findings indicate that sectioned otoliths are the best structure for aging summer flounder over the age range 0 to 10 years. Sectioned otoliths had the shortest reading times, the highest confidence scores, the highest within- and between-reader agreement, and they were consistently clearer and easier to read than whole otoliths, scales, and oper- cular bones. These findings are new for summer flounder because no published studies have used sectioned otoliths to age this species. These findings generally agree, however, with many studies on other species that have found sectioned otoliths to be the best aging structure (for examples, Beamish, 1979; Chilton and Beamish, 1982; Beamish and McFarlane, 1983; Lowerre-Barbieri et al., 1994). Right sectioned otoliths were generally superi- or to left sectioned otoliths. Although we found high agreement in age between right and left sec- tioned otoliths, right otoliths were much easier to prepare, and they had a larger counting path, which made it easier to identify the marks, result- ing in shorter reading times, higher confidence scores, and higher reader agreement. Although we have found sectioned otoliths to be the best structure for determining the age of summer flounder, our studies have not proven their accuracy. To do so would require known-age methods or at least marginal increment meth- ods. However, until validation is done, we feel there is sufficient evidence to recommend that sectioned otoliths replace the current practice of using scales for aging summer flounder. Comparative evaluation of whole otoliths Our findings indicate that whole otoliths are the second best structure for aging summer flounder over the age range of 0 to 10 years. Whole oto- liths had no preparation time and had the second shortest reading times, the second highest confidence scores, the second highest within- and between-reader agreement, and the highest agreement with sectioned oto- liths. Whole otoliths were generally easy to read in fish less than age 4 or 5, and we feel they are adequate for these younger ages, especially in large-scale production aging where preparation time is important. We found that the right whole otolith was often easier to read than the left when the secondary counting path was used. Therefore, although former studies have used the left whole otolith only (Poole, 1961; Eldridge, 1962; Smith and Daiber, 1977; Powell, 1982), we suggest that the right should be included in future work. Our findings on preparation and reading times, confi- dence scores, within- and between-reader agreement and agreement with sectioned otoliths are generally new be- cause the literature has not reported detailed evaluations of whole otoliths in summer flounder. Given our findings, we do not agree with the current preference for using scales rather than whole otoliths in summer flounder. In- deed, we disagree with the original reasons for rejecting otoliths, which included 1) poor calcification and poor con- trast between opaque and translucent zones (Shepherd, 1980; Smith et al., 1981; Dery, 1988), 2) obscurement of the first mark as the fish ages (Powell, 1982), 3) deviation from the generalized pattern of opaque and translucent zone for- 636 Fishery Bulletin 99(4) Figure 5 Right sectioned (A) and whole (B) otolith from a female summer flounder, 10 years old (determined from sectioned otoliths) and 758 mm TL, collected in November. Arrows on the sectioned otolith indicate presumed annual marks. On the whole otolith, arrows indi- cate primary counting path (anterior field), dots indicate secondary counting path (posterior field). Ten marks are visible in the posterior field of the whole otolith, but only eight marks are visible in the anterior field. Table 4 Regression statistics for relationships between the number of marks (Marks) and calcified structure size for summer flounder. Structure abbreviations are defined in the “Meth- ods” section of the text, n = sample size. All regressions were significant at P < 0.001. Structure Equation n 100) Scales Marks = -2.64 + 1.080 ScRL 80 73 Sectioned otoliths Right Marks = -3.39 + 5.424 SORL 65 80 Left Marks = -3.36 + 6.996 SORL 46 67 Whole otoliths Right Marks = -4.56 + 1.664 RWOR 75 85 Left Marks = -4.47 + 1.367 LWOR 76 86 Right Marks = -4.80 + 0.919 RWOT 75 86 Left Marks = -4.80 + 0.934 LWOT 76 82 mation in temperate fishes (Smith et al., 1981), and 4) a narrow opaque zone as compared to the translucent zone (Smith et al., 1981). We address these issues in turn below. Table 5 Regression statistics for relationships between the number of marks (Marks) on calcified structures and summer flounder total length (TL). All regressions were significant at P < 0.001, and sample sizes were 80 fish. Structure Equation 100 r2 Scales Marks = -3.69 + 0.0151 TL 83 Sectioned otoliths Marks = -3.86 + 0.0155 TL 85 Whole otoliths Marks = -3.90 + 0.0157 TL 86 We rarely observed poor calcification or poor contrast between opaque and translucent zones of whole otoliths. Rather, our procedures gave good contrast between opaque and translucent zones, so that we had high confidence in our age readings. In addition, we found only one otolith of 81 pairs to be poorly calcified. This otolith was easily aged once it was sectioned, and its pair was not poorly calcified and was aged with high confidence. We saw little evidence that the first mark becomes ob- scured at older ages on whole otoliths, as indicated by our high agreement between whole and sectioned otoliths. The hypothesis that the first mark becomes obscured was Sipe and Chittenden: A comparison of calcified structures for aging Paralichthys dentatus 637 based on overlap in back-calculated sizes at the second and third marks on whole otoliths (Powell, 1982). However, size in any year class can vary greatly because summer flounder spawn over a protracted season (Smith, 1973; Morse, 1981; Able et al., 1990). Therefore, fish in adjacent year class- es can be expected to overlap in size, and Pow- ell’s results do not necessarily mean that the first mark becomes obscured with age. Smith et al. ( 1981) reported that summer floun- der otoliths deviated from the general pattern of opaque and translucent zone formation seen in other temperate fishes and suggested that opaque zones formed in fall-winter, the reverse of the usual spring-summer formation in other temper- ate species. We saw no evidence of this reversal. Our fish were collected from October through De- cember, so we should have observed opaque edg- es on the otolith if the timing of mark formation were reversed from other temperate fishes. In- stead, we observed relatively wide translucent zones on the otolith edges. In addition, other stud- ies have not found a reversal in the time of mark formation (Poole, 1961; Powell, 1982; Wenner et al., 1990), and Desfosse (1995) found that opaque zones appeared to form on whole otoliths at approximately the same time as scale marks (May through July). Finally, Smith et al. (1981) presented no data to support their hy- pothesis that opaque zones formed in the fall and winter. Indeed, their Figure 5 shows an opaque edge on a whole otolith from a summer flounder captured in June. In agreement with studies in other species (see referenc- es below), we found the translucent zone to be wider than the opaque zone on summer flounder otoliths. Smith et al. (1981) felt this was an anomalous occurrence and used it to reject whole otoliths. We disagree with their analy- sis, however, because many other fishes in our study area, including Atlantic croaker (Barbieri et al., 1994a), weak- fish (Lowerre-Barbieri et al., 1994), and Spanish macker- el (Gaichas, 1997) have otoliths with a wide translucent zone and a narrow opaque zone. Such a pattern reflects the fact that opaque zones form over a short time period in these species: April-May in Atlantic croaker and weakfish (Barbieri et al., 1994a; Lowerre-Barbieri et al., 1994) and May-June in Spanish mackerel (Gaichas, 1997). In addi- tion, although the sample size was limited (n=93), Des- fosse (1995) found evidence, using marginal increments, that opaque zone formation on summer flounder otoliths occurs over a similarly short time period (May to July). Fi- nally, regardless of whether opaque zones are narrower or wider than translucent zones, otoliths can be used for age determination if the mark can be proven annual. Comparative evaluation of scales Our findings indicate that scales are inferior to, and much less desirable than, both sectioned and whole otoliths for aging summer flounder. Scales had significantly lower con- fidence scores and much higher reading times than sec- tioned and whole otoliths because marks on scales were often difficult to interpret using objective aging criteria. False marks were common, and different scales from the same fish often indicated different ages. As a result, both within- and between-reader percent agreement and agree- ment with whole and sectioned otolith age were undesir- ably low in scales, especially in fish over age 4. We feel that scales should not be used for aging summer flounder if oto- liths, especially sectioned otoliths, are available. The difficulties we found with summer flounder scales generally agree with reports in the literature. Dery (1988), Desfosse (1995), and Bolz et al. (2000), for examples, have reported similar problems interpreting scale marks. Like us, Desfosse (1995) found low within-reader scale agree- ment (only 46%) in fish over age 4. Desfosse (1995) re- ported high agreement between scales and whole otoliths (98%) for ages 0 to 5, much higher than the 85% agree- ment we found for ages 0 to 4. However, 90% of his fish (rc=170) were ages 0 to 2 and only one was age 5, a likely ex- planation for his high percent agreement. Shepherd ( 1980) reported high agreement (91%) between scales and whole otoliths for moderately old fish (ages 4 to 6), but his sample size was only 21 fish, only one of which was age 6. Our study reported lower overall agreement between whole oto- liths and scales (76%), but we examined fish over a much wider age range (ages 0 to 10) than previously reported. Comparative evaluation of opercular bones Our comparative studies have found opercular bones to be inferior to both sectioned and whole otoliths in summer flounder, and even to scales. Opercular bones had the lowest confidence scores, the highest reading times, only 46% within reader agreement, and they often exhibited unclear transitions from translucent to opaque zones, par- ticularly at early ages. For these reasons, we feel that oper- cular bones should not be used for aging summer flounder. 638 Fishery Bulletin 99(4) These findings are new for summer flounder, because no previous studies have used opercular bones to age this species. We were disappointed at the poor performance of opercular bones because they have been reported useful in many other species, including perch (LeCren, 1947), carp (McConnell, 1952), yellow perch (Bardach, 1955), north- ern pike (Frost and Kipling, 1957), tautog (Cooper, 1967; Hostetter and Munroe, 1993), and goldeye (Donald et al., 1992). Many of these studies show photographs of oper- cular bones with clear, easily recognized marks that have been interpreted as being formed annually. These studies, however, generally have not validated age determination in opercular bones; therefore it is unclear whether they give accurate ages in these other species. Formation of early marks on otoliths and scales We sometimes observed an early, presumably false, mark prior to the first presumed annual mark on both otoliths and scales of summer flounder. Although we attempted not to count this early mark, it appeared to be the pri- mary cause for disagreements between our readers in aging both otoliths and scales. This problem has not been reported in summer flounder otoliths, although there is evidence of this early false mark on scales (Dery, 1988; Bolz et al., 2000). Indeed, a primary problem cited by Bolz et al. (2000) for differences in interpretation of summer flounder scales was the choice of a first annual mark. The early, presumably false, mark that sometimes oc- curred on summer flounder otoliths and scales appears similar to the first mark reported for Atlantic croaker otoliths (Barbieri et al., 1994a) and might be explained by similarities in certain life history traits of these two species. Both species have a protracted spawning season and spawn over a similar time frame in the Chesapeake Bay region: Atlantic croaker from mid-summer to late fall (Wallace, 1940; Haven, 1957; Barbieri et al., 1994b), and summer flounder from early fall to early winter (Smith, 1973; Morse, 1981; Able et al., 1990). Barbieri et al. ( 1994a) reported the formation of a first mark on Atlantic croaker otoliths in the first spring following hatching, at 5 to 10 months, with two patterns of early mark formation: 1) the first mark close to, but distinct from, the focus in early hatched fish, and 2) the first mark nearly continuous with the focus in late hatched fish. As with Atlantic croaker, we suggest that the first mark on summer flounder otoliths and scales, which we have referred to as an “early, presum- ably false, mark,” might actually be laid down in the first spring following hatching, at 5 to 8 months, with the same two patterns of early mark formation. Previous summer flounder aging studies interpret the first annual mark to be laid down on scales and otoliths in the second spring following hatching (Smith et al., 1981; Szedlmayer et al., 1992), at 17 to 20 months, one year af- ter the first annual mark is laid down on Atlantic croaker otoliths. Despite this difference, fish from these two species that are hatched at the same time are currently placed in the same year class. It thus appears that the current age determination methods differ between these two species. For example, according to current conventions (Bolz et al.. 2000), a summer flounder hatched in October 2000 would be called age 1 on 1 January 2002, at a biological age of 15 months. This age is several months before the first pre- sumed annual mark is laid down on the structures in the second spring following hatching (2002), even though an “early” mark might have been laid down in the first spring following hatching (2001). Similarly, an Atlantic croaker hatched in October 2000 would be called age 1 on 1 Janu- ary 2002 (Barbieri et al., 1994a), at a biological age of 15 months. However, this age is 8 months after the first an- nual mark is laid down on the otolith, which occurs in the first spring following hatching (2001). Therefore, the two species differ in the way the first annual mark is assigned. To resolve the issue of early mark formation in summer flounder, we suggest that calcified structures of young-of- the-year fish be examined to determine when the early mark is formed, as Barbieri et al. (1994a) did for Atlantic croaker. Barbieri et al.’s (1994a) validated method automat- ically assigns an early first mark, formed at 5 to 10 months, to all Atlantic croaker otoliths, whether the mark is dis- tinct or not. If the “early, presumably false, mark” in sum- mer flounder is similar to the first annual mark in Atlantic croaker, an early first mark could likewise be assigned to summer flounder otoliths. If this were done, disagreements on the first mark on summer flounder structures would be fewer, and summer flounder and Atlantic croaker would be aged in exactly the same way. That is, both fish would al- ready have a first annual mark on the structure when ages are advanced to 1 on the 1 January arbitrary birthdate. Acknowledgments We would like to thank Chesapeake Bay commercial fish- ermen for their cooperation and James Owens (VIMS) for helping us to obtain samples from them. T. Ihde (VIMS) helped develop some of the aging methods and was the second reader in our study. J. Foster (VIMS) helped develop ideas for the early mark portion of the discussion. Finan- cial support for this study was provided by a Wallop/ Breaux Program Grant for Sport Fish Restoration from the U.S. Fish and Wildlife Service through the Virginia Marine Resources Commission, Project F-88-R-2. Literature cited Able, K. W., R. E. Matheson, W. W. Morse, M. P. Fahay, and G. Shepherd. 1990. Patterns of summer flounder Paralichthys dentatus early life history in the Mid-Atlantic Bight and New Jersey estuaries. Fish. Bull. 88:1-12. Almeida, F. P, R. E. Castaneda, R. Jesien, R. E. Greenfield, and J. M. Burnett. 1992. Proceedings of the NEFC/ASMFC summer flounder, Paralichthys dentatus, aging workshop, 11-13 June 1990, Northeast Fisheries Center, Woods Hole, Massachusetts. U.S. Dep. Commer, NOAA Tech. Memo. NMFS-F/NEC-89, 7p. Bagenal, T. B., and F. W. Tesch. 1978. Age and growth. In Methods for assessment of fish Sipe and Chittenden: A comparison of calcified structures for aging Paralichthys dentatus 639 production in fresh waters, 3rd ed. (T. B. Bagenal, ed.), p. 101-136. Blackwell Scientific Publications, Oxford. Barbieri, L. R., M. E. Chittenden Jr., and C. M. Jones. 1994a. Age, growth, and mortality of Atlantic croaker, Micro- pogonias undulatus , in the Chesapeake Bay region, with a discussion of apparent geographic changes in population dynamics. Fish. Bull. 92:1-12. Barbieri, L. R., M. E. Chittenden Jr., and S. K. Lowerre-Barbieri. 1994b. Maturity, spawning, and ovarian cycle of Atlantic croaker, Micropogonias undulatus, in the Chesapeake Bay and adjacent coastal waters. Fish. Bull. 92:671-685. Bardach, J. E. 1955. The opercular bone of the yellow perch (Perea flavescens), as a tool for age and growth studies. Copeia 2:107-109. Beamish, R. J. 1979. Differences in the age of Pacific hake ( Merluccius pro- ductus) using whole otoliths and sections of otoliths. J. Fish. Res. Board Can. 36:141-151. Beamish, R. J., and G. A. McFarlane. 1983. The forgotten requirement for age validation in fish- eries biology. Trans. Am. Fish. Soc. 112:735-743. Bigelow, H. B., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine. U.S. Fish Wild], Serv., Fish. Bull. 53:1-577. Bolz, G. R., J. P. Monaghan Jr., K. L. Lang, R. W. Gregory, and J. M. Burnett. 2000. Proceedings of the summer flounder aging workshop, 1-2 February 1999, Woods Hole, Massachusetts. U.S. Dep. Commer., NOAATech. Memo. NMFS-NE-156, 15 p. Chilton, D. E., and R. J. Beamish. 1982. Age determination methods for fishes studied by the Groundfish Program at the Pacific Biological Station. Can. Spec. Publ. Fish. Aquat. Sci. 60:1-102. Cooper, R. A. 1967. Age and growth of the tautog, Tautoga onitis (Linnaeus), from Rhode Island. Trans. Am. Fish. Soc. 96:134—142. Dery, L. M. 1988. Summer flounder, Paralichthys dentatus. In Age determination methods for Northwest Atlantic species (J. Penttila and L. M. Dery, eds.), p. 97-102. U.S. Dep. Commer., NOAATech. Rep. NMFS 72. Desfosse, J. C. 1995. Movements and ecology of summer flounder, Paralich- thys dentatus, tagged in the southern Mid-Atlantic Bight. Ph.D. diss.. College of William and Mary, Williamsburg, VA, 187 p. Donald, D. B., J. A. Babaluk, J. F. Craig, and W. A. Musker. 1992. Evaluation of scales and operculum methods to deter- mine age of adult goldeyes with special reference to a domi- nant year-class. Trans. Am. Fish. Soc. 121:792-796. Eldridge, P. J. 1962. Observations on the winter trawl fishery for summer flounder, Paralichthys dentatus. M.S. thesis, College of William and Mary, Williamsburg, VA, 58 p. Frost, W. E., and C. Kipling. 1957. The determination of the age and growth of pike (Esox lucius L.) from scales and opercular bones. J. Cons. 24:314-341. Gaichas, S. K. 1997. Age and growth of Spanish mackerel, Scomberomorus maculatus, in the Chesapeake Bay region. M.S. thesis, College of William and Mary, VA, 111 p. Ginsburg, I. 1952. Flounders of the genus Paralichthys and related genera in American waters. U.S. Fish Wild], Serv., Fish. Bull. 52:267-351. Gutherz, E. J. 1967. Field guide to the flatfishes of the Family Bothidae in the western North Atlantic. U.S. Fish Wildl. Serv. Circ. 263,47 p. Haven, D. S. 1957. Distribution, growth, and availability of juvenile croak- er, Micropogon undulatus , in Virginia. Ecology 38:88-97. Hildebrand, S. F., and W. C. Schroeder. 1928. Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish 43:1— 366. Hostetter, E. B., and T. A. Munroe. 1993. Age, growth, and reproduction of tautog Tautoga onitis (Labridae:Perciformes) from coastal waters of Vir- ginia. Fish. Bull. 91:45-64. LeCren, E. D. 1947. The determination of the age and growth of the perch (Perea fluviatilis) from the opercular bone. J. Anim. Ecol. 16:188-204. Leim, A. H. and W. B. Scott. 1966. Fishes of the Atlantic Coast of Canada. Fish. Res. Board Can., Bull. 155:1-485. Lowerre-Barbieri, S. K., M. E. Chittenden Jr., and C. M. Jones. 1994. A comparison of a validated otolith method to age weakfish, Cynoscion regalis, with the traditional scale method. Fish. Bull. 92:555-568. MAFMC (Mid-Atlantic Fishery Management Council). 1987. Fishery management plan for the summer flounder fishery, 159 p. MAFMC. Dover, DE. McConnell, W. J. 1952. The opercular bone as an indicator of age and growth of the carp, Cyprinus carpio Linnaeus. Trans. Am. Fish. Soc. 81:138-149. Media Cybernetics. 1999. Optimas, version 6.5. Media Cybernetics, Silver Spring, MD. Morse, W. W. 1981. Reproduction of the summer flounder, Paralichthys dentatus ( L. >. J. Fish. Biol. 19:189-203. Poole, J. C. 1961. Age and growth of the fluke in Great South Bay and their significance to the sport fishery. N.Y. Fish Game J. 8:1-18. 1962. The fluke population of Great South Bay in relation to the sport fishery. N.Y. Fish Game J. 9:93-117. Powell, A. B. 1982. Annulus formation on otoliths and growth of young summer flounder from Pamlico Sound, North Carolina. Trans. Am. Fish. Soc. 11 1:688-693. Shepherd, G. 1980. A comparative study of ageing methods for summer flounder (Paralichthys dentatus). Lab. Ref. Doc. 80-13, Northeast Fisheries Science Center, NMFS, Woods Hole, MA, 26 p. Smith, R. W., and F. C. Daiber. 1977. Biology of the summer flounder, Paralichthys denta- tus, in Delaware Bay. Fish. Bull. 75:823-830. Smith, R. W., L. M. Dery, P. G. Scarlett, and A. Jearld Jr. 1981. Proceedings of the summer flounder (Paralichthys dentatus) age and growth workshop, 20-21 May 1980, NEFC, Woods Hole, Massachusetts. U.S. Dep Commer., NOAATech. Memo. NMFS-F/NEC-1 1, 30 p. Smith, W. G. 1973. The distribution of summer flounder, Paralichthys dentatus, eggs and larvae on the continental shelf between Cape Cod and Cape Lookout. Fish. Bull. 71:527-535. 640 Fishery Bulletin 99(4) Szedlmayer, S. T., K. W. Able, and R. A. Rountree. 1992. Growth and temperature induced mortality of young- of-the-year summer flounder ( Paralichthys dentatus) in southern New Jersey. Copeia 1992:120-128. Wallace, D. H. 1940. Sexual development of the croaker, Micropogon utidu- latus , and distribution of the early stages in Chesapeake Bay. Trans. Am. Fish. Soc. 70:475-482. Wenner, C. A., W. A. Roumillat, J. E. Moran Jr., M. B. Maddox, L. B. Daniel, and J. W. Smith. 1990. Investigations of the life history and population dynam- ics of marine recreational fishes in South Carolina: part 1. South Carolina Wildl. Mar. Resourc. Dep., Mar. Resourc. Res. Instit., 26 p. Zar, J. H. 1996. Biostatistical analysis. Prentice Hall, Upper Saddle River, NJ, 662 p. 641 Abstract— A bottom trawl catch of flat- fish is composed of fish that were ini- tially in the path of the trawl net and fish that were initially in the path of the bridles and were subsequently herded into the net path. Bridle effi- ciency (i.e. the proportion of fish in the bridle path that are herded into the net path) for seven species of flatfish was estimated by fitting a model of the herding process to data collected during a field experiment. The exper- iment consisted of repeatedly making trawl hauls with three different bridle lengths. The model was then fitted to the catch of each species, by length class, as a function of the widths of the net and door spreads at the three bridle lengths. Bridle efficiency was indepen- dent of body length for five species (yellowfin sole, Limanda aspera\ flat- head sole, Hippoglossoides elassodon; rock sole, Lepidopsetta bilineata ; Dover sole, Microstomus pacificus; rex sole, Glyptocephalus zachirus) and ranged from 0.22 for rex sole to 0.40 for rock sole. Bridle efficiency declined with increasing body length for two species (English sole, Parophrys vetu- lus\ Pacific sanddab, Citharichthys sor- didus ), ranging from about 0.35 to 0.10 over the lengths sampled. Manuscript accepted 15 March 2001. Fish. Bull. 99:641-652 (2001). Bridle efficiency of a survey trawl for flatfish David A. Somerton Peter Munro Alaska Fisheries Science Center 7600 Sand Point Way NE Seattle, Washington 98125 E-mail address (for D A. Somerton): david.somerton@noaa.gov Swept-area estimates of biomass pro- duced by bottom trawl surveys are typ- ically used as relative indices of stock size rather than absolute measures because the sampling efficiency of the trawls1 is rarely known. As a means to estimate trawl efficiency, Dickson (1993a) developed a model describing the trawling process and the associated fish behavior from an undisturbed state before arrival of a fishing vessel until passage of a trawl codend. For benthic species that remain closer to the sea bottom than the head-rope height, this model specifies that trawl efficiency is primarily determined by two processes: escapement of fish out of the net path by passing either under the footrope or through the mesh (net efficiency) and herding of fish into the net path by the doors, mudclouds, and bridles (bridle efficiency). In our study, we focused on fish herding, considering the problem of quantitatively estimating bridle effi- ciency or the proportion of fish encoun- tering the bridles that are herded into the net path. Underwater observations of trawls in operation have revealed that the pro- cess of herding differs between semi- pelagic species, such as Atlantic cod (Gadus morhua) and haddock ( Melano - grammus aeglefinus ) that tend to swim near the bottom, and benthic species, such as flatfish that often remain in direct contact with the bottom (Main and Sangster, 1981b). Atlantic cod are apparently stimulated to herd by the sight of the trawl doors and the mud- clouds that form in the wake of the doors, whereas flatfish are stimulated to herd by the close proximity or ac- tual touch of the doors and lower bri- dles as they sweep along the bottom (Hemmings, 1969; High, 1969; Marty- shevskii and Korotkov, 1969; Main and Sangster, 1981a, 1981b). Flatfish, once stimulated to move, tend to swim a relatively short distance in a direction perpendicular to the bridles where they slow down or settle on the bottom un- til the bridle again overtakes them. This pattern is usually repeated, pro- ducing what appears to an observer on the trawl headrope as a zigzag path to- ward the center of the trawl footrope (Main and Sangster, 1981b). While this is occurring, however, flatfish may ei- ther swim over the bridle or become ex- hausted and be overtaken by the bridle and thereby escape the herding pro- cess. Bridle efficiency is determined by the relative frequency of such events. Perhaps because the mechanism of herding seems more obvious for flatfish than other types of fish, the earliest mathematical models of fish herding were based on the contact-swim-con- tact pattern of flatfish herding (Hem- mings, 1969; Foster et al., 1981; Fuwa et al,. 1988; Fuwa, 1989). These models were focused on individual fish and typ- ically considered both the physical as- pects of the trawl, such as shape and speed, as well as the biological aspects of the fish, such as size and endurance. Although these models provided a more structured way of examining fish herd- ing and allowed the quantitative eval- uation of the relative importance of the various aspects of the process, they were of limited use in estimating bri- dle efficiency of specific trawls because some of the required parameters were unknown or difficult to estimate. The approach to modeling fish herd- ing changed markedly after Engas and Godp (1989a) conducted trawling ex- 1 Throughout this paper, we will refer to the trawl as an entire fishing gear comprising the net, bridles, and doors. 642 Fishery Bulletin 99(4) periments to quantitatively examine the effects of varia- tions in bridle length on catch for each area and fish size. Reasoning that the changes in catch that accompanied the changes in bridle length provided information on the ef- ficiency of herding, Dickson (1993b) estimated the bridle efficiency coefficients for Atlantic cod and haddock by fit- ting a simple linear model of the trawl herding process to data from Engas Godp’s (1989a) experiments. Ramm and Xiao (1995) conducted a similar trawl herding experiment and developed a new type of herding model appropriate for cases in which escapement at the center of the trawl is zero. In our study, we extended Dickson’s (1993a) herding model so that it is more specific to the peculiarities of flat- fish herding and additionally extended Dickson’s (1993b) approach to fitting the model to data by providing a more rigorous statistical foundation. The model was then ap- plied to herding data for the 83-112 Eastern bottom trawl (Armistead and Nichol, 1993) used by the Alaska Fisher- ies Science Center (AFSC) to conduct its annual bottom trawl survey of the Eastern Bering Sea. These data were collected during two herding experiments, patterned af- ter those of Engas and Godo (1989a), in which emphasis was placed on seven species of flatfish: yellowfin sole ( Li - manda aspera ), flathead sole ( Hippoglossoides elassodon), rock sole ( Lepidopsetta bilineata ), English sole ( Paroph - rys vetulus ), Dover sole ( Microstomus pad ficus), rex sole ( Glyptocephalus zachirus), and Pacific sanddab ( Cithar - ichthys sordidus). Materials and methods Nn = DLWn Nb = DL(Wd-Wn), where L = tow length; Wn = the width of the net path; and Wd = the width of the door path (Fig. 1). (2) Combining all terms, the total catch then can be expressed as N = knDLWn + knkbDL (Wd-Wn). ( 3 ) At this point the model is quite similar to the one proposed by Dickson (1993a) to describe the trawl catching process for Atlantic cod and haddock. We now modify the model to make it more specific to flatfish by considering that herd- ing is restricted to the portion of the bridle path in which the lower bridle is in contact with the bottom. This can be expressed as N = knDLWn + knhDL ( Wd - Wn - Woff), ( 4 ) where W0„ = the width of the path in which the bridle is not in contact with the bottom; and h - the herding coefficient or the proportion of fish in the bridle contact path that is herded into the net path. Thus, kb is the average efficiency over the entire bridle path and li is the average efficiency in the portion of the bridle path where herding actually occurs. Development of the herding model Consider that the area of the bottom swept by a trawl con- sists of two components, the area between the wingtips of the net (net path), and the area between the wing tips and the doors (bridle path; Fig. 1). The catch (AO obtained in a trawl can be represented as some proportion, kn, of the number of fish in the net path ( Nn ) plus some propor- tion, P, of the number of fish in the bridle path (Nb). If the probability of capt ure for a fish herded into the net path is identical to that of a fish initially in the net path, then P can be considered equal to knkb , where kh is the bridle effi- ciency2 or the proportion of fish within the bridle path that is herded into the net path. Algebraically this is expressed as N = knNn + knkbNb. (1) The numbers of fish within the net and bridle paths are equal to the fish density (D) multiplied by the path areas, that is, 2 In Dickson (1993a) this same quantity is referred to as sweep efficiency (ks). We use the term “bridle efficiency” because the 83-112 Eastern trawl does not have sweeps connecting the bri- dles to the doors as does the trawl considered by Dickson. Description of the herding experiments The strategy that we used in our herding experiments was to repeatedly trawl in such a way that the width of the net path was held approximately constant, whereas the width of the door path was varied among three dis- tances by changing the length of the bridles. The first of two herding experiments was conducted 25 July-2 August 1994 aboard the FV Arcturus in the eastern Bering Sea and focused on yellowfin sole, flathead sole, and rock sole. The second experiment was conducted 14-25 September 1994 aboard the RV Alaska off the coast of Washington State and focused on English sole, Dover sole, rex sole, and Pacific sanddab. For both experiments, we used a blocked sampling design to minimize the effects on catch of the spatial variation in fish density. In each geographic block, three nearby, but nonoverlapping, 30-min trawl hauls at a speed of 1.5 m/s were made with each of three bridle lengths chosen in random order. Bridles measured 27 m, 55 m (the standard length used on AFSC surveys), and 82 m in length and were constructed of 16-mm diameter steel cable. Tailchains connecting the doors to the bridles (Fig. 1) were always constructed of 13-mm diameter long link chain but differed in length between vessels (Table 1). Trawl doors were always a “V” style measuring 1.8 m x 2.7 m and weighing 910 kg. On all hauls, spreads of the doors were measured at the tops of the wings simultane- Somerton and Munro: Bridle efficiency of a survey trawl for flatfish 643 ously and continuously with an acoustic trawl mensura- tion system. Tow length was measured as the straight line distance between the GPS positions of first and last foot- rope contact with the bottom determined with a time-depth recorder attached to the headrope. Bottom temperature was recorded at two second intervals with a bathyther- mograph and averaged 3.0°C during the eastern Bering sea experiment and 7.0°C in the West Coast experiment. Bottom depth averaged 76 m during the eastern Bering sea experiment and 67 m in the West Coast experiment. The catch from each haul was first sorted to species; then the catch of each species was weighed in the aggregate and fish of each species were individually measured for total length in centimeters. Both experiments were conducted during daylight hours to best approximate the sampling protocol used on most AFSC bottom trawl surveys. Estimating Woff The herding experiments provided data on N, L, Wd, and Wn, but obtaining data to estimate Woft required a differ- ent type of experiment. We obtained the necessary data by directly observing the contact of the lower bridle with the bottom on an experimental cruise conducted 9-12 Septem- 644 Fishery Bulletin 99(4) Table 1 Trawl configuration parameters for each herding experiment. Included are the bridle and tail chain lengths that were used, the number of sampling blocks occupied, and the means and standard deviations (in parentheses) of the wing spread, door spread, and bridle angle. Also included are the mean values for the bridle off-bottom path width ( Wo^). WA = Washington State. Experiment Bridle length (m) Tailchain length (m) Number of blocks Wing spread (m) (Wn) Door spread (m) (Wd) Bridle angle (deg) (a) Bridle off-bottom path width (m) Woff) Bering Sea 27.3 13.9 15 18.0(0.6) 44.6(2.1) 18.9(1.2) 23.3 54.6 13.9 15 17.3 (0.6) 58.7(2.9) 17.6 (1.0) 21.8 81.6 13.9 15 17.1 (0.5) 71.6(3.7) 16.6(1.0) 20.5 West Coast (WA) 27.3 8.0 19 17.7 (0.3) 41.9(0.7) 20.0 (0.5) 24.1 54.6 8.0 19 17.1 (0.6) 57.5 (1.5) 18.8(0.7) 23.1 81.6 8.0 19 16.8 (0.3) 70.0 (1.5) 17.3(0.5) 21.3 ber 1995 aboard the RV Alaska off the coast of Washington State using the 83-112 Eastern trawl equipped with stan- dard 55-m bridles. The experiment consisted of recording views of the lower bridle with a silicon-intensified target (SIT) video camera while the trawl was in operation. The camera was mounted in a positively buoyant case attached to the upper bridle with a 1-m tether line and aligned so that the lower bridle directly below the attachment point could be viewed. The tether was positioned at 5-m inter- vals between 23 m and 43 m behind the doors. One 15-min tow at each of the five bridle positions was taken at both depths of 20 m and 35 m. The video tapes were subsequently analyzed to deter- mine the degree of bottom contact at each bridle position. This was done by viewing each tow at ten randomly chosen 10-sec sampling intervals. Bottom contact was recognized by the presence of sediment that is mixed into the water at the point of contact. Because the bottom is irregular, when the bridle has weak contact, just the tops of the irregulari- ties are touched, whereas when the bridle has strong con- tact, the entire bottom in the field of view is touched. The average percentage of the bridle in the field of view that was in contact with the bottom was scored according to the following four-level scale: 1) no contact, 2) <25% contact, 3) >25% and <75% contact, 4) >75% contact. The degree of bot- tom contact at each bridle position and depth was then es- timated as the mean of the ten evaluations. The length of the bridle not in contact with the bottom (Loff, Fig. 1) was defined as the distance between the at- tachment of the tailchain to the door and the point along the bridle where bottom contact was 50%. To estimate Lof-f> a logistic function was fitted to the bottom-contact and bri- dle-position data with generalized linear modeling (Ven- ables and Ripley, 1994), then the fitted logistic equation was evaluated at a bottom contact score of 3.0 (i.e. bottom contact >25% and <75%). Variance of Lo^ was estimated with bootstrapping (Efron and Tibshirani, 1993), where the bootstrap samples were obtained by randomly choos- ing, with replacement, from the haul mean scores at each bridle position. In subsequent analysis, Loff was treated as a constant for all three bridle lengths and for both vessels. However the variable actually used in the herding model, Woff , was calculated as L0^ multiplied by the sine of the bridle an- gle (Fig. 1) and therefore varied slightly between bridle lengths and experiments because the bridle angle varied. In addition to the camera placements for measuring we also made placements at bridle positions both clos- er to the doors to examine for any evidence of herding in the area where the lower bridles were not in contact with the bottom and closer to the wing tips to determine if bot- tom contact was maintained continuously near the junc- tion of the bridle and wing. Fitting the model to the herding data The herding model (Eq. 4) was modified in several ways to clarify the way it was fitted to the experimental data and to better define its underlying statistical structure. First, because of the block design of the experiment, fish density < D ) was considered to be a constant within each block but to vary between blocks. Furthermore, net efficiency (kj was also considered to be a constant for all tows within a block because depth and other bottom conditions are nearly constant but vary between blocks. Because D and kn are both block-specific constants and are confounded in the model, they were combined into a new constant k. Second, herding can vary with fish length (Engas and Godp, 1989a; Dickson, 1993b), therefore Equation 4 was modified to allow length dependency. The modified equa- tion is Nljk = kl[LWn\. +kthk[L(.Wd-Wn -W^)].. +£ (5) where the subscript i = block number; j = bridle length within a block; k - fish length class and e = ~ MO, 1.0, which is not feasible. In our herding experiments, however, R exceeded door spread ratios for all species except English sole and Pacific sanddab; therefore, in most cases the unmod- ified herding model failed to provide feasible estimates of kb. By comparison, estimates of kb from the modified model are feasible over the entire observed range of catch ratios. In our initial attempts to fit the herding model to experi- mental data, we treated Wo^ as an additional parameter to be estimated in the fitting process. However, it soon became clear that W0^and h are confounded in the model and that convergent solutions required an independent estimate of Woff. We were able to obtain such an estimate and ob- serve fish behavior near the bridle by using video methods, because in the standard configuration of the 83-112 East- ern trawl the bridles are not obscured by mudclouds over much of their lengths. However, in perhaps more typical cases where the bridles are obscured by the mudclouds, es- timation of Waff with video methods would be problematic. A further complication in such cases is that the mudclouds themselves might provide a herding stimulus and confuse the interpretation of W^, 650 Fishery Bulletin 99(4) of assumptions. First, we assumed herding efficiency was independent of bridle length. If, however, exhaustion was a primary determinant of bridle efficiency, then it is pos- sible that individual endurance would be exceeded more Mode! assumptions The adequacy of the herding model and the appropriate- ness of our herding efficiency estimates rest on a variety Somerton and Munro: Bridle efficiency of a survey trawl for flatfish 651 frequently at longer bridle lengths and would thereby diminish herding efficiency. This assumption was tested by comparing the fit of a herding model with a bridle-length dependent h to that of an unmodified model. The modi- fied model provided a better fit for only one species, Pacific sanddab. Because this species is the smallest of the seven (Fig. 3), it is likely that its swimming endurance might have been exceeded at the longest bridle length. However, for the remaining species the assumption is valid. Second, we assumed that none of the fish in the area spanned by Wo^ are herded into the net path. Although we recognize that the doors and the mudclouds immediately behind the doors must herd some fish toward the net path, such fish would be herded into the section of the bridle un- obscured by the mudcloud and likely escape beneath the lower bridle. We attempted to confirm this possibility by examining video observations from this area of the bridle to see if flatfish escaped or showed any herding behavior, but the results were equivocal because few fish, and prob- ably only moving fish, were seen. Third, we assumed that in changing the length of the bridles we did not alter trawl geometry in any way other than by increasing the area exposed to herding. As in the studies of Engas and Godp (1989a) and Ramm and Xiao ( 1995), increased bridle length in our study produced a small, but significant, reduction in wing spread (Table 1). If this change produced a change in footrope contact and net efficiency, then it would lead to a biased estimate of herding efficiency (net efficiency is explicitly assumed to be independent of bridle length in Eq. 4). Other pos- sible changes in trawl geometry are that 1) at the longest length the upper bridle may sag sufficiently to touch the bottom within La^ and 2) at the shortest length the bridle tension may be sufficient to lift the wingtip off the bot- tom. With such changes, and perhaps even without, Lo^ may not be constant at all bridle lengths. Although we ex- amined the performance of the trawl along the entire dis- tance from the wing tips to the doors at the standard bri- dle length, we had insufficient ship-time to examine the performance at all bridle lengths to verify that the bridles performed as intended. We recommend that further stud- ies of herding efficiency include additional research to ver- ify that standard trawl performance is maintained for all experimental configurations. One approach for doing this is to simultaneously estimate escapement under the foot- rope, perhaps by using an auxiliary net attached beneath the trawl net as in Engas and Godp (1989b), so that the assumed independence of net efficiency and bridle length could be tested. Fourth, we have conducted these experiments in rela- tively small areas that do not encompass all of the bot- tom conditions found over the areas covered by the bottom trawl surveys; therefore the results are potentially biased and certainly less variable than if they were conducted over such a range of conditions. Bridle efficiency is only one component needed to deter- mine trawl efficiency (i.e. the proportion of fish within the door path that are caught). The remaining component, for flatfish at least, is net efficiency (i.e. the proportion of fish within the net path that are caught). If estimates of net ef- ficiency are also available, perhaps obtained with the use of an auxiliary net to capture escaping fish (Engas and Godp, 1989b), they can be included with estimates of bri- dle efficiency in Dickson’s (1993a) model of the trawl fish- ing process to produce estimates of trawl efficiency. Such estimates could then be used to convert relative indices of fish abundance from trawl survey to absolute estimates of abundance. Acknowledgments We thank Asgeir Aglen, Martin Dorn, Olav Rune Godp, Jack Turnock, and Gary Stauffer for reviewing the manu- script and for making helpful suggestions. Literature cited Armistead, C. E., and D. G. Nichol. 1993. 1990 bottom trawl survey of the eastern Bering Sea continental shelf. U.S. Dep. Commer., NOAATech. Memo. NMFS-AFSC-7, 190 p. Bridger, J. P. 1969. The behaviour of demersal fish in the path of a trawl. FAO Fish. Rep. 62(3):695-715. Dickson, W. 1993a. Estimation of the capture efficiency of trawl gear. I: Development of a theoretical model. Fish. Res. 16: 239-253. 1993b. Estimation of the capture efficiency of trawl gear. II: Testing a theoretical model. Fish. Res. 16:255-272. Efron, B., and R. Tibshirani. 1993. An introduction to the bootstrap. Chapman and Hall, New York, NY, 436 p. Engas, A., and O. R. Godp. 1989a. The effect of different sweep lengths on the length composition of bottom-sampling trawl catches. J. Cons. Int. Explor. Mer 45:263-268. 1989b. Escape of fish under the fishing line of a Norwegian sampling trawl and its influence on survey results. J. Cons. Int. Explor. Mer 45:269-276. Foster, J. J., C. M. Campbell, and G. C. W. Sabin. 1981. The fish catching process relevant to trawls. Can. Spec. Publ. Fish. Aquat. Sci. 58:229-246. Fuwa, S. 1989. Fish herding model by ground ropes considering reac- tion of fish. Nippon Suisan Gakkaishi 55:1767-1771. Fuwa, S., O. Sato, K. Nashimoto, and N. Higo. 1988. Fish herding model by ground rope. Nippon Suisan Gakkaishi 54:1155-1159. Hemmings, C. C. 1969. Observations on the behaviour of fish during capture by the Danish seine net and their relation to herding by trawl bridles. FAO Fish. Rep. 62:645-655. High, W. L. 1969. Scuba diving, a valuable tool for investigating the behavior of fish within the influence of fishing gear. FAO Fish. Rep. 62(3):323 — 331. Hilborn, R., and M. Mangel. 1997. The ecological detective: confronting models with data. Princeton Univ. Press, Princeton, NJ, 315 p. Main, J., and G. I. Sangster. 1981a. A study of the sand clouds produced by trawl boards 652 Fishery Bulletin 99(4) and their possible effect on fish capture. Scott. Fish. Res. Rep. 20:1-20. 1981b. A study of the fish capture process in a bottom trawl by direct observations from a towed underwater vehicle. Scott. Fish. Res. Rep. 23:1-23. Martyshevskii, V. N., and V. N. Korotkov. 1969. Fish behaviour in the area of the trawl as studied by bathyplane. FAO Fish. Rep. 62(3):781-791. Ramm, D. C., and Y. Xiao. 1995. Herding in groundfish and effective pathwidth of trawls. Fish. Res. 24:243-259. Seber, G. A. F., and C. J. Wild. 1989. Nonlinear regression. Wiley, New York, NY, 768 p. Venables, W. N., and B. D. Ripley. 1994. Modern applied statistics with S-plus. Springer- Verlag, New York, NY, 462 p. 653 Abstract— The red snapper (Lutjanus campechanus) is currently under rig- orous federal and state management in the Gulf of Mexico because of appar- ent overfishing. Management strategies implemented to promote recovery of the species are dependent upon knowl- edge of various demographic variables, including the ages of individuals, the dis- tribution of these ages (cohort strength) within the population, and maximum longevity. Thus a dependable and pre- cise aging method is of principal impor- tance. Counts of annuli in otolith thin sections have been used to age many species of fish, including red snapper. However, the utility of this method for aging red snapper has been questioned by those who dispute both the appar- ent longevity (over 50 yr) of red snap- per and the position of the first annulus within the red snapper otolith. We counted annuli and assessed edge condition in sagittal otoliths of 3791 red snapper collected from the north- ern Gulf of Mexico off Louisiana during the periods from 1989 to 1992 and from 1995 to 1998. Opaque annuli were vali- dated by marginal increment analysis to form once per year from December through June. Estimated ages ranged from 0.5 to 52.6 yr for individuals from 104 mm to 1039 mm total length and from 0.02 kg to 22.79 kg total weight. Among the 2546 specimens of known sex, both sexes evidenced rapid growth of individuals until about age 8-10 yr, after which growth slows considerably. Von Bertalanffy growth models for total length at age were significantly differ- ent for males and females, indicating differential growth between the sexes, with females typically obtaining larger sizes at older ages than do males. Manuscript accepted 15 March 2001. Fish. Bull. 99:653-664 (2001). Age and growth of red snapper, Lutjanus campechanus, from the northern Gulf of Mexico off Louisiana* Charles A. Wilson Coastal Fisheries Institute and Department of Oceanography and Coastal Sciences School of the Coast and Environment Louisiana State University Baton Rouge, Louisiana 70803-7503 E-mail address: cwilson@lsu.edu David L. Nieland Coastal Fisheries Institute School of the Coast and Environment Louisiana State University Baton Rouge, Louisiana 70803-7503 The red snapper , Lutjanus campechanus (Poey) (family: Lutjanidae), is resident on the continental shelves of the Gulf of Mexico (GOM) and northwest Atlan- tic Ocean from the Bay of Campeche, Mexico, to Massachusetts; however, it is found only occasionally north of Cape Hatteras, North Carolina (Rivas, 1966; Robins and Ray, 1986; Hoese and Moore, 1998). Although Rivas ( 1966) suggested that red snapper may also occur off Bermuda, the Bahamas, and northern Cuba, it has been reported neither from Bermuda by Smith-Vaniz et al. (1999) nor from the Bahamas by Bohlke and Chaplin (1993), nor from Cuba by Allen (1985). The species is replaced in the Caribbean Sea and southward by the Caribbean red snapper, L. pur- pureus (Rivas, 1966; Robins and Ray, 1986; Hoese and Moore, 1998 ). Although the appellation “red snapper” has been widely used to identify as many as 12 commercially marketed lutjanids (Camber, 1955; Carpenter, 1965), it is correctly applied only to L. campecha- nus (Robins et al., 1991). The binomina L. ay a, L. blackfordi, and Neomaenis ay a, all of which appear in the litera- ture (e.g. Moseley, 1966) are synonyms of L. campechanus and refer to the red snapper sensu Robins et al. (1991). The red snapper has been and re- mains a significant component of both the commercial and recreational fish- eries in the GOM. However, document- ed commercial landings from United States territorial waters declined pre- cipitously from historic highs of about 6389 metric tons (t) in 1965 to 1015 t in 1991; estimated recreational land- ings similarly waned from 4734 t in 1979 to 581 1 in 1990 (Schirripa and Le- gault* 1). Since 1991 both fisheries have been constrained by size limits, creel or trip limits, and quotas as established by the Gulf of Mexico Fishery Man- agement Council (GMFMC). The best efforts of the GMFMC and the com- mercial and recreational sectors not- withstanding, overfishing of red snap- per in the GOM may persist (Schirripa and Legault2). * Contribution LSU-CFI-00-01 of the Coastal Fisheries Institute, Louisiana State Uni- versity, Baton Rouge, LA 70803-7503. 1 Schirripa, M. J., and C. M. Legault. 1997. Status of the red snapper in U. S. waters of the Gulf of Mexico: Updated through 1996. Contribution rep. MLA-97/98-05 from Sustainable Fisheries Division, Miami Lab- oratory, Southeast Fisheries Science Cen- ter, National Marine Fisheries Service, 75 Virginia Beach Drive, Miami, FL 33149-1099. [Not available from NTIS.] 2 Schirripa, M. J., and C. M. Legault. 1999. Status of the red snapper in U. S. waters of the Gulf of Mexico: Updated through 1998. Contribution rep. SFD-99/00-75 from Sustainable Fisheries Division, Miami Lab- continued 654 Fishery Bulletin 99(4) Accurate information on the age structure of the red snapper populations in the GOM is essential for monitor- ing year-class strength, for conducting stock assessments, and for documenting population recovery. Previous efforts at estimating red snapper age have employed a variety of aging methods. Bradley and Bryan (1975) cited the long spawning season and a constant recruitment into the pop- ulation as reasons for the difficulty in assigning red snap- per ages from length-frequency data. Moseley (1966) used scale annuli to age red snapper to age 4 yr and advanced spawning as the causal factor in check formation. Among 240 red snapper taken off the west coast of Florida, Futch and Bruger (1976) estimated red snapper ages of 1 to 5 yr from 200 readable whole otoliths; however, they postu- lated ages up to 20 yr for larger individuals whose oto- liths were unreadable. Comparisons of ages derived from whole otoliths, scales, and vertebrae by Bortone and Hol- lingsworth (1980) revealed all three hard parts to be of equal utility in aging red snapper at age 1 and 2 yr. Wade (1981) also used scales to age red snapper to 9 yr. Nelson and Manooch (1982) reported red snapper ages from 1 to 16 yr based on both scales and sectioned otoliths and dem- onstrated once yearly scale annulus formation in June and July from monthly mean marginal growth. A recent study of red snapper otoliths significantly extended the hypothe- sized longevity of red snapper in the GOM to 42 yr (Szedl- mayer and Shipp, 1994). Render (1995) provided a pre- liminary validation of yearly opaque annulus formation in sagittal otoliths and reported ages from 0 to 53 yr for red snapper in Louisiana waters. Examinations of otolith sec- tions from 537 red snapper captured in the northwestern Atlantic Ocean from Beaufort, North Carolina, south to the Florida Keys manifested a maximum longevity of 25 yr (Manooch and Potts, 1997). Among 907 red snapper from the GOM off Alabama, Patterson (1999) reported opaque annulus formation from January through June and maxi- mum ages of 30 yr for females and 31 yr for males. Despite these efforts, the longevity of red snapper remains contro- versial. Small sample sizes, a paucity of older specimens, and the failure to present legitimate validations of ages from hard parts (Beamish and McFarlane, 1983) have var- iously hampered the above studies. It has further been speculated that larger and presumably older red snapper form numerous false annuli within otoliths (Rothschild et al.3) . And both the timing of deposition and the position of the putative first annulus remain in question. Otolith analyses have proven consistent in estimating ages of many fish species, including several from the tem- perate waters of the northern GOM (Johnson et al., 1983; Barger, 1985; Beckman et al., 1988; Beckman et al., 1990, 1991; Crabtree et al., 1992; Murphy and Taylor, 1994; Franks et al., 1998; Thompson et al., 1998). Herein we pres- 2 (continued) oratory, Southeast Fisheries Science Center, National Marine Fisheries Service, 75 Virginia Beacli Drive, Miami, FL 33149-1099. [Not available from NTIS.] 3 Rothschild, B. J., A. F. Sharov, and A. Y. Bobyrev. 1997. Red snapper stock assessment and management for the Gulf of Mexico. Report submitted to the National Marine Fisheries Service, Office of Science and Technology, 1315 East-West High- way, Silver Spring, MD 20910 USA. [Not available from NTIS.] ent our interpretations of the use of sagittal otoliths to es- timate ages of red snapper from the GOM off Louisiana. Specifically we address the timing of formation and posi- tion of the first annulus, validation of the once yearly accre- tion of opaque annuli, longevity, and reader reproducibility. We further describe the growth of red snapper with von Bertalanffy growth models for both males and females. Methods and materials Red snapper from recreational and commercial catches were sampled from 1989 to 1992 and from 1995 to 1998 by personnel of the Louisiana State University (LSU) Coastal Fisheries Institute and the Louisiana Department of Wild- life and Fisheries (LDWF). Although the vast majority of our sampling efforts were targeted at both wholesale facil- ities and charter boat docks located in Grand Isle and Port Fourchon, LA, the area of coverage in the northern GOM extended from off the Mississippi River Delta in the east to off Galveston, TX, in the west. Morphometric mea- surements (total length [TL] or fork length [FL] in mm, total weight [TW] or eviscerated body weight [BW], i.e. mass with liver, digestive tract, and reproductive organs removed, in g) were recorded, both sagittal otoliths were removed, and sex was determined, when possible, for each specimen. Eviscerated body weight was converted to TW, when necessary, with the equation TW-1.10KBW) - 26.32 (linear regression, df=418; P<0.001; ^=0.996) and TL was estimated from FL with the equation TL=1.073 (FL) + 3.56 (linear regression, df= 1015; P<0. 001; r2=0.999) . All undamaged, intact otoliths were weighed to the near- er 0.1 g. The left sagittal otolith from each individual was embedded in an epoxy resin and subsequently sectioned with a low-speed saw equipped with a watering blade as described in Beckman et al. (1988). In those instances where the left sagitta was damaged or unavailable, the right sagitta was substituted. Examinations of otolith sec- tions were made with a compound microscope and trans- mitted light at 40x to lOOx magnification. Counts of an- nuli (opaque zones) were accomplished by reading along the medial surface of the transverse section dorsal or ven- tral to the sulcus; annuli were often inconsistent in other regions of the otolith section. Annuli were counted by two readers without knowledge of date of capture or morpho- metric data. The appearance of the otolith margin was al- so coded as either opaque or translucent (Beckman et al., 1988). Sections were recounted a second time by both read- ers when initial counts disagreed. Rather than excluding the small number of individuals for which a consensus could not be reached after a second reading (n=27), the as- signed annulus count for these was that of the more experi- enced reader (reader 1). Reader 2’s annulus count and edge condition were used in those circumstances where reader l’s were missing (n= 2). Annulus counting error between the two readers was evaluated after both the initial and second readings of the otolith sections. Reproducibility of the resultant age estimates was evaluated with the coeffi- cient of variation, the index of precision (Chang, 1982), and average percent error (Beamish and Fournier, 1981). Wilson and Nieland: Age and growth of Lut/cinus ccimpechanus 655 The periodicity of opaque annulus formation was deter- mined by marginal increment analysis and by plotting the proportion of otoliths with opaque margins by month of capture (Beckman et ah, 1988). To assess the possi- bility of false annulus formation among either younger or older red snapper, those individuals of age <5 yr and those >5 yr were also analyzed as above. Fork lengths at 100% maturity, 420 mm for males and 440 mm for females (Render, 1995), are achieved at about age 5. If one opaque and one translucent zone are shown to be formed each year, validation of annuli as being accreted once yearly is accomplished. Ages of red snapper were estimated from opaque annu- lus counts and date of capture with the equation Age (days) = -182 + ( annulus count x 365) + ((m - 1) x 30) + d , where m = the ordinal number ( 1-12) of the month of cap- ture; and d = the ordinal number (1-31) of the day of the month of capture. The 182 days that were subtracted from each age estimate are an accommodation for the uniform 1 July hatching date which was assigned for all specimens (Render, 1995; Collins et al., 1996). Age in yr was derived by dividing the age in days by 365. Thus a red snapper captured on 1 January which exhibits five opaque annuli (including an opaque margin [see below]) in its otoliths would have an estimated age of 1644 days or 4.5 yr. Our age esti- mation method also assumed that opaque annulus for- mation at the otolith margin uniformly commenced in January. The small number of individuals captured in Sep- tember, October, November, and December that evidenced early formation of opaque annuli had their ages adjusted by subtracting 365 days from their age estimates. Con- versely, a larger number of individuals captured in Janu- ary, February, and March had otoliths with translucent margins — evidence of an assumed delay in opaque annu- lus formation. The age estimates of these were augmented by the addition of 365 days. Total length-TW regressions were fitted with linear re- gression to the model TL = a TWh from log10-transformed data. Male and female regressions were compared with analysis of covariance (SAS, 1985). Only those red snap- per for which sex could be determined were used to fit growth models. Von Bertalanffy growth models of TL at age were fitted with nonlinear regression by least squares (SAS, 1985) in the form TLt=L„( where TLt = TL at age t\ Lx = the TL asymptote; k = a growth coefficient; t = age in yr; and t0 = a hypothetical age when TL is zero. Growth models were generated for three groups of red snapper within which the age and TL of all individuals were extant: 1) all specimens, 2) all specimens of known sex, and 3) specimens of known sex for which growth models were fitted independently for each sex. Likelihood ratio tests (Cerrato, 1990) were used to test for differences between males and females, both in growth models and in growth parameter estimates. Significance level for statis- tical analyses was 0.05 unless indicated otherwise. Results During eight years of variable collection effort, 3791 red snapper from recreational ( n =274 ) and commercial ( n =35 1 7 ) catches were sampled for morphometric data and sagittal otoliths. Among the 1438 male and 1542 female specimens for which sex could be determined, females ranged from 242 to 1039 mm TL and from 0.16 to 22.79 kg TW; males were 245-946 mm TL and 0.19-13.70 kg TW. Composite ranges for all specimens of either known or unknown sex were 104-1039 mm TL and 0.02-22.79 kg TW; however, 67.6% of 3787 available TL were between 325 and 525 mm and 80.0% of 3718 available TW were less then 2.5 kg (Fig. 1). Neither the slopes (df=2,932; F=3.41; P<0.065) nor the intercepts (df=2,932; F- 3.16; P<0.075) of the TL-TW regres- sions were found to differ significantly between males and females; thus data for the two sexes were combined and a single predictive equation was generated TW=l.n x 10-8(PL)304 Sagittae of red snapper are ovate, laterally compressed, and have an indented sulcus acousticus on the proximal surface (Fig. 2A). Although one can count purported an- nuli in relatively small whole otoliths of red snapper less than age 5 (Futch and Bruger, 1976), it is difficult to dis- cern annuli in the larger otoliths of older individuals. Thin transverse sections of these older otoliths showed semidi- stinct translucent and opaque annuli that alternated from the core to the growing edge (Fig. 2, B and C).The presump- tive first annulus posed the most consistent problem for the readers. This annulus appeared as a diffuse “smudge” of opaque material variously located from totally isolated and somewhat distant from the core (Fig. 2B) to contig- uous to and continuous with the otolith core (Fig. 2C). Annulus counts ranging from 0 to 53 and edge condi- tions were determined by at least one reader for all 3791 individuals sampled. Reader 2 considered all the otolith sections to be of sufficient quality to produce annulus counts; reader 1 provided annulus counts from all but two sections. After the initial counts, consensus between read- ers was achieved for 2804 individuals A second reading of the 987 sections for which annulus counts differed pro- duced consensus for 3762 individuals. The degree of agree- ment in red snapper opaque annulus counts between the two readers in each of the two readings was assessed. Av- erage percent error (APE), coefficient of variation (CV), in- dex of precision (D), and percentages of absolute differenc- es in counts are given in Table 1. 656 Fishery Bulletin 99(4) Proportions of otoliths with opaque margins were plot- ted by month of capture for all individuals (n=3791), for those individuals presumed to be sexually immature (ages less than or equal to 5, n=2143), and for those from indi- viduals of presumptive sexual maturity (ages greater than 5, n=948). Each of the three plots (Fig. 3) features a sin- gle broad peak and a single broad valley and conclusively demonstrates opaque annulus formation from December through June and translucent annulus formation from Ju- ly through. November. Thus, the assumption of one to one correspondence between opaque annulus counts and esti- mated red snapper age in years is validated. Furthermore, this correspondence is validated for immature and mature individuals of all ages. Having demonstrated once yearly accretion of opaque annuli, we estimated ages from 0.5 to 52.6 yr from the annulus counts of the red snapper in our study. The vast majority of specimens examined were ages 2-5 and only 1.2% of the total number were greater than age 15 yr (Fig. 4). The few ages greater than 15 yr, which were not repre- sented in our sample, were 24, 28, 31, 34, 39, 40, 42-46, and 49-50. The otolith section from the oldest specimen examined is shown in Figure 5. Among 3787 individuals, the single von Bertalanffy growth model which best describes red snapper TL at age was TL (mm) = 941 [1 _e -o.i8u + o.55)]; (r2=0.72). Wilson and Nieland: Age and growth of Lut/anus campechanus 657 Figure 2 Medial view (A) and transverse sections (B and C) of red snapper (Lat- janus campechanus) left sagittal otoliths. Abbreviations for all are Do = dorsal, V = ventral, P = posterior, A = anterior, SA = sulcus acousticus, Di = distal, and Pr = proximal. Arrows with numbers indicate positions of opaque annuli and their enumeration. Not surprisingly, the von Bertalanffy growth model gen- erated for all individuals of known sex ( rz =2979 ) was quite similar to the above: TL fmm) = 935 [1 -e -o-istt + O-M^ (r2=0.72). However, a likelihood ratio test revealed that von Ber- talanffy growth models for males (/t = 1438 ) and females (/2=1541) were significantly different from one another ( X2=7 5 .09; df=l,2979; P<0.0001). The resultant models for TL at age were Female TL (mm) = 977 [l _ e -o.ie u + o.63)]; (r2=0.71) Male TL (mm) = 904 [l _ e “° 19u + 048)], (r2=0.73) 658 Fishery Bulletin 99(4) Predicted TL at age for males and females (Figs. 6 and 7) generated with the above equations illustrated rapid and roughly equivalent growth to an age of approximately 8-10 yr after which the growth curves diverge. Differential growth between males and females was further demon- strated with likelihood ratio tests which indicated signifi- cant differences in (x2=13.05; df=l,2979; P<0.0003) and k (%2=7. 16; df=l, 2979; P<0. 0075). No significant difference was detected in t0 ( x2= 1-34; df=l,2979; P<0.247). Owing to the large variability in observed TL at age (Fig. 6), TL was a poor estimator of red snapper age within the range of TL encountered in our sampling efforts. Discussion Otolith annuli as indicators of age in years have been validated for many freshwater and marine fish species, including the Australian lutjanids L. adetii and L. quin- quelineatus (Newman et al., 1996) and L. argentimacula- tus, L. bohar, L.carponotatus, L. erythropterus, L. gibbus, L. johnii, L. malabaricus , L. monostigina , L. rivulatus, L. sebae, and L. vitta (Cappo et al., 2000). Previous studies of red snapper have used scales (Moseley, 1966; Wade, 1981), otoliths (Futch and Bruger, 1976; Szedlmayer and Shipp, 1994; Render, 1995; Manooch and Potts, 1997; Patterson, Table 1 Average percent error (APE), coefficient of variation (CV), and index of precision (D) differences in red snapper (Lut- janus campechanus ) otolith annulus counts for two read- ers on first and second readings, n = 3791. lsl reading 2nd reading APE 3.736 0.0934 CV 0.045 0.0011 D 0.032 0.0008 0 73.96% 99.29% ±1 23.27% 0.61% ±2 2.24% 0.08% > ±3 0.53% 0.02% 1999), scales and otoliths (Nelson and Manooch, 1982) and scales, otoliths, and vertebrae (Bortone and Hollingsworth, 1980) to estimate ages. Among these, early attempts to val- idate age estimation from circuli of scales and annuli of otoliths have suffered from two shortcomings: 1) a small sample size and 2) a paucity of individuals over age 10 yr. Nevertheless, they have produced a general consensus Wilson and Nieland: Age and growth of Lutjanus ccimpechanus 659 40 35 30 „ 25 ' 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 >15 Age (years) Figure 4 Age-frequency histogram («=3791) for red snapper, Lutjanus campechanus, from the northern Gulf of Mexico. Specimens were collected from the recreational and commercial fisheries of the Gulf of Mexico off Louisiana from 1989 to 1992 and from 1995 to 1998. that transparent annuli (Nelson and Manooch, 1982) are formed during the spawning season. May to September in the GOM (Collins et ah, 1996). Our validation of opaque annulus formation in otoliths of red snapper during the winter and spring seasons is in substantial agreement with previous efforts. Given that yearly formation of opaque annuli has been validated for substantial numbers of red snapper from the Atlantic waters off North Carolina south to Florida (Manooch and Potts, 1997) and the GOM waters off Alabama (Patterson, 1999) and Louisiana (Render, 1995; our study), and the validation among Australian conge- neric species cited above, the one-to-one correspondence between annuli and age in years should be indisputable. Certainly, the reproducibility statistics indicate that the annuli of red snapper otoliths are more difficult to count than those in otoliths of some other species. Comparisons of between-reader age estimates in several species of the family Sciaenidae have yielded almost 100% agreement (Beckman et ah, 1988; Beckman et ah, 1990; Barbieri et ah, 1993; Lowerre-Barbieri et ah, 1995). Sciaenid oto- liths are comparatively massive and annuli are especially well defined. Conversely, red snapper otoliths are relative- ly thin and fragile and the annuli become less well de- fined with increasing age. But, even given the above, a first reading followed by a second reading produced con- sensus in age estimates for 99.29% of those red snapper considered in our study. Patterson (1999) reported 93.8% between-reader consensus of red snapper annulus counts after two readings. Quite unlike the situation in sciaenids, training and experience are critical to achieving high be- tween-reader consensus on red snapper annulus counts. The variable position and the diffuse appearance of the first annulus formed during the first winter following hatching (age approximately 6 months) are presumed to be functions of both the protracted red snapper spawn- ing season and the rapid growth rate of juvenile red snapper. Those individuals that are spawned early in the season will experience proportionally more growth (and presumably more translucent zone accretion adjacent to the otolith core) than will a late spawned individual be- fore opaque annulus accretion begins during the following winter; thus the first opaque annulus will be more distant from the otolith core in the former instance than under the latter circumstance. Also with the first opaque annu- lus accreting at a rate theoretically corresponding to the rapid growth rate experienced during the juvenile stage, the resulting first annulus is broader and more diffuse in appearance than annuli produced during times of reduced growth rates in later life. A more complete understanding of first annulus formation in red snapper could improve insights into both recruitment patterns and growth rates of individuals within a spawning season. The distribution of ages among our sample population (Fig. 4) is certainly not reflective of the age distribution of red snapper in the GOM off Louisiana. Age-0 and age-1 specimens have been largely unavailable to our sampling efforts owing to minimum size limits applied to the rec- reational and commercial fisheries during the 1990s. Also 660 Fishery Bulletin 99(4) Figure 5 Cross section of sagittal otolith from the oldest red snapper, Lutjanus campechanus, from among 3791 specimens sampled from the recreational and commercial fisheries of the Gulf of Mexico off Louisiana from 1989 to 1992 and from 1995 to 1998. Abbreviations are Do = dorsal, V = ventral, SA = sulcus acousticus, Di = distal, and Pr = proximal. Arrows with numbers indicate positions of opaque annuli and their enumeration. the dominance of ages 2-5 may reflect the practices of the fishermen who target red snapper and a migratory aspect of the species’ life history. Age-0, and to a lesser extent age-1, red snapper are known to inhabit shallow water areas devoid of complex habitats or vertical relief where some are vulnerable to capture in trawls. This behavior is illustrated in fishery-independent trawl data from the GOM, specifically the Fall Groundfish Survey and the Summer SEAMAP Survey, in which the great majority of red snapper captured are age 0 and 1 (Schirripa and Le- gault2). It has been hypothesized that the disappearance of red snapper from the trawl data at age 1 represents their migration to structures such as oil and gas platforms that presumably provide refuge from large predators (Render, 1995). It is during this residence at the numerous oil and gas platforms off Louisiana that red snapper become vul- nerable to fishing gear. Because the platforms are easily located and potentially harbor large populations of red snapper and other fish species (Stanley and Wilson, 1996, 1998), they are the preferred destinations for both com- mercial and recreational fishermen. The very low numbers of individuals of age >6 in our sample population likely re- sult from both removal from the population through fish- ing and natural mortalities and emigration away from the oil and gas platforms to alternative habitats where they are less susceptible to capture. It is difficult to compare the maximum observed red snapper longevity reported in our study with those re- ported in earlier studies (Moseley, 1966; Futch and Bru- ger, 1976; Wade, 1981; Nelson and Manooch, 1982; Render, 1995; Szedlmayer and Shipp, 1994; Patterson, 1999) be- cause of the assortment of aging techniques (scales, whole or sectioned otoliths, length frequencies) and the variety of sources (commercial, recreational, or both) used. All show a predominance of relatively young individuals (< 10 yr). Flowever, recent advances and refinements in otolith prep- aration technology have allowed red snapper to be aged reliably up to the following ages: 42 yr (Szedlmayer and Shipp, 1994), 53 yr (Render, 1995), 31 yr (Patterson, 1999), and 52 yr (our study). Despite the sparsity of old red snap- per among these research efforts, there can be little doubt that red snapper at least have the potential to achieve ag- es of 40-50 yr and more. The red snapper growth models that we present are similar to those of earlier studies (Nelson and Manooch, 1982; Szedlmayer and Shipp, 1994; Manooch and Potts, 1997; Patterson, 1999) which did not produce separate models for the two sexes and variously applied weighted Wilson and Nieland: Age and growth of Lutjanus campechcinus 661 Figure 6 Observed total length (mm) at age and relationship of age to total length predicted from von Bertalanffy growth models for male (n=1438) and female (n = 1542) red snapper, Lutjanus campechanus, from the northern Gulf of Mexico. Closed circles and narrow line represent males; open squares and thick line represent females. or unweighted analyses (Fig. 7). All models predict rapid, and very much similar, growth during the first 8-10 years of life and slower growth thereafter. Asymptotic lengths among the above varied from 936 to 1025 mm. Our von Bertalanffy growth models also predicted a greater as- ymptotic TL and slightly faster growth for female red snapper. Among marine teleosts from the GOM, a similar pattern of growth has been shown for red drum, Sciaenops ocellatus (Beckman et al., 1988), sheepshead, Archosargus probatocephalus (Beckman et al., 1991), and cobia, Rachy- centron canadum (Franks et al., 1998). However, this phe- nomenon in red snapper may be the result of a prepon- derance of data from the commercial fishery included in our analyses. Owing to minimum size limits enforced in both the commercial and recreational fisheries, we had ac- cess to few red snapper less than age 2 and few less than 250 mm TL. Conversely, the preference of commercial red snapper fishermen and wholesalers for smaller, plate-size individuals afforded us little opportunity to sample larger, and presumably older, red snapper; these are the individu- als that can influence estimation of k and which ultimately drive estimation of Lx. The addition of another 20-30 old specimens of both genders could have profound effects on the estimations of Lx for both sexes. Furthermore, growth studies of lutjanids in Australian waters (Davis and West, 1992; McPherson and Squire, 1992; Newman et al., 1996) report faster growth and larger size at age among males. Proportionally greater expenditures of energy in the pro- duction of gametes by females is advanced to explain this observation (Newman et al., 1996). Thus, although the growth rates and asymptotic lengths for male and female red snapper are shown to differ statistically in our study, questions of the biological veracity and the biological sig- nificance of these differences remain unresolved. In addition to the differences in red snapper estimated growth rates between the sexes, there is an obvious high degree of diversity in individual growth rates. Owing to the large variability in age at a given TL (Fig. 6), this vari- able is a poor estimator of red snapper age. Our data in- dicate that red snappers of 400 mm, 600 mm, and 800 mm TL could be ages 2-7 yr, 3-9 yr, and 5-35+ yr, respec- tively. As a more concrete example, consider the Interna- tional Game Fishing Association world-record red snapper caught by rod and reel, the otoliths of which were given to us for age analysis. This individual was caught off the coast of Louisiana by Doc Kennedy of Grand Isle, LA, on 23 June 1996; it was 22.79 kg (50 lb, 4 oz) TW, 1039 mm (40.9 in) TL, and 965 mm (38 in) FL. Given the immense size of this specimen, one would reasonably expect it to be ancient by red snapper standards. However, our analysis 662 Fishery Bulletin 99(4) Figure 7 Comparative von Bertalanffy growth models for red snapper, Lutjanus campechanus, from the western north Atlantic Ocean. S&S 94 = Szedlmayer and Shipp (1994), P 99 = Patterson (1999), N&M 82 = Nelson and Manooch (1982), and M&P 97 = Manooch and Potts (1997). Males and Females are from our study. revealed it to be only 19.98 yr. Conversely, the two oldest red snapper we encountered, age 52.63 and 51.73 yr, were a comparatively small 851 mm TL and 862 mm TL, respec- tively, and 7.886 kg TW and 9.188 kg TW, respectively. A similar pattern was noted by Patterson (1999) among the red snapper that he sampled from the GOM off Alabama. Personnel at the LSU Coastal Fisheries Institute con- tinue to investigate the nuances of deriving red snapper ages from sagittal otoliths. Although our marginal incre- ment analysis demonstrates that a single opaque incre- ment is formed each year, our sample size among older individuals, albeit larger than any previous investigation, is probably inadequate for absolute validation of this phe- nomenon. Thus, some have and will continue to question once yearly annulus accretion among red snapper older than 20 yr. A solution for this problem may lie in radio- metric aging techniques with protocols that analyze vari- ous radionuclides in the otoliths. Also, core-to-first-annu- lus measurements made on otolith sections from age-0 and age-1 individuals would contribute to a better under- standing of when and how the first annulus is accreted. Acknowledgments We appreciate the able assistance of Ed Moss (LDWF), Bruce Thompson, Scott Baker, and many others in our red snapper sampling efforts. To Louise Stanley (reader 1) and Andrew Fischer (reader 2), we offer our gratitude not only for assistance in sampling, but also for the many long hours at the microscope counting annuli. Jane and Lonnie Black of USA Fish, Golden Meadow and Cameron, LA, graciously allowed us access to their facilities. We also thank Joey Trosclair and the employees of Trosclair Canning Com- pany, Cameron, LA, for their help and the many commer- cial and recreational fishermen who allowed us to sample their catches. Michael Schirripa, National Marine Fisher- ies Service, is gratefully recognized for his many tangible contributions to our red snapper research. The comments of Richard Shaw, Brigitte Nieland, and three anonymous reviewers contributed greatly to the manuscript. Funding for this research was provided by the U. S. Department of Commerce Marine Fisheries Initiative (MARFIN) Pro- gram (grant numbers NA90AA-H-MF762, NA57FF0287, and NA77FF0544). Finally, we take great pride in dedi- cating this paper to the memory of Jeffery H. Render, our friend and colleague, for his pioneering efforts in the study of red snapper life history in the northern GOM. Literature cited Allen, G. R. 1985. FAO species catalogue. Vol. 6: Snappers of the world: an annotated and illustrated catalogue of lutjanid species known to date, 208 p. FAO, Rome. Wilson and Nieland: Age and growth of Lutjanus campechanus 663 Barbieri, L. R., M. E. Chittenden Jr, and C. M. Jones. 1993. Age, growth, and mortality of Atlantic croaker, Micro- pogonias undulatus, in the Chesapeake Bay region, with a discussion of apparent geographic changes in population dynamics. Fish. Bull. 92:1-12. Barger, L. E. 1985. Age and growth of Atlantic croakers in the northern Gulf of Mexico, based on otolith sections. Trans. Am. Fish. Soc. 111:847-850. Bradley, E., and C. E. Bryan. 1975. Life history and fishery of the red snapper (Lutjanus campechanus) in the northwestern Gulf of Mexico. Proc. Gulf Caribb. Fish. Inst. 27:77-106 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. 1983. The forgotten requirement for age validation in fish- eries biology. Trans. Am. Fish. Soc. 112:735-743. 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. 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. 1991. Age and growth-rate estimation of sheepshead Archo- sargus probatocephalus in Louisiana waters using otoliths. Fish. Bull. 89:1-8. Bohlke, J. E„ and C. C. G. Chaplin. 1993. Fishes of the Bahamas and adjacent tropical waters. Univ. Texas Press, Austin, TX, 771 p. Bortone, S. A., and C. L. Hollingsworth. 1980. Aging red snapper, Lutjanus campechanus , with oto- liths, scales, and vertebrae. Northeast Gulf Sci. 4:60-63. Camber, C. I. 1955. A survey of the red snapper fishery of the Gulf of Mexico, with special reference to the Campeche Banks. Fla. Board Conserv. Tech. Ser. No. 12:1-64. Cappo, M., P. Eden, S. J. Newman, and S. Robertson. 2000. A new approach to validation of periodicity and timing of opaque zone formation in the otoliths of eleven species of Lutjanus from the central Great Barrier Reef. Fish. Bull. 98:474-488. Carpenter, J. S. 1965. A review of the Gulf of Mexico red snapper fishery. U.S. Dep. Interior, Fish and Wildl. Serv., Bur. Comm. Fish. Circ. 208:1-35. Cerrato, R. M. 1990. Interpretable statistical tests for growth comparisons using parameters in the von Bertalanffy equation. Can. J. Fish. Aquat. Sci. 47:1416-1426. Chang, W. B. 1982. A statistical method for evaluating the reproduci- bility of age determination. Can. J. Fish. Aquat. Sci. 39: 1208-1210. Collins, L. A., A. G. Johnson, and C. P. Keim. 1996. Spawning and annual fecundity of the red snapper (Lutjanus campechanus) from the northeastern Gulf of Mexico. In Biology, fisheries and culture of tropical grou- pers and snappers (F. Arreguin-Sanchez, J. L. Munro, M. C. Balgos, and D. Pauly, eds.), p. 174-188. ICLARM Conf. Proc. 48. Crabtree, R. E., E. C. Cyr, R. E. Bishop, L. M. Falkenstein, and J. M. Dean. 1992. Age and growth of tarpon, Megalops atlanticus, larvae in the eastern Gulf of Mexico, with notes on relative abun- dance and probable spawning areas. Env. Biol. Fishes 35: 361-370. Davis, T. L. O., and G. J. West. 1992. Growth and mortality of Lutjanus uittus (Quoy and Gaimard) from the North West Shelf of Australia. Fish. Bull. 90:395-404. Franks, J. S., J. R. Warren, and M. V. Buchanan. 1998. Age and growth of cobia, Rachycentron Canadian , from the northeastern Gulf of Mexico. Fish. Bull. 97:459-471. Futch, R. B., and G. E. Bruger. 1976. Age, growth, and reproduction of red snapper in Flor- ida waters. In Proceedings: colloquium on snapper-grou- per fishery resources of the western central Atlantic Ocean (H. R. Bullis Jr. and A. C. Jones, eds.), p. 165-183. Florida Sea Grant College Program Report 17. Hoese, H. D., and R. H. Moore. 1998. Fishes of the Gulf of Mexico. Texas A&M Univ. Press, College Station, TX, 422 p. Johnson, A. G., W. A. Fable Jr., M. L. Williams, and L. E. Barger. 1983. Age, growth, and mortality of king mackerel, Scomb- eromorus caualla, from the southeastern United States. Fish. Bull. 81:97-106. Lowerre-Barbieri, S. K., M. E. Chittenden Jr., and L. R. Barbieri. 1995. Age and growth of weakfish, Cynoscion regalis, in the Chesapeake Bay region with a discussion of historical changes in maximum size. Fish. Bull. 93:643-656. Manooch, C. S., Ill, and J. C. Potts. 1997. Age and growth of red snapper, Lutjanus campecha- nus, Lutjanidae, collected along the southeastern United States from North Carolina through the east coast of Flor- ida. J. Elisha Mitchell Sci. Soc. 113:111-122. McPherson, G. R., and L. Squire. 1992. Age and growth of three dominant Lutjanus species of the Great Barrier Reef inter-reef fishery. Asian Fish. Sci. 5:25-36. Moseley, F. N. 1966. Biology of red snapper, Lutjanus aya Bloch, of the northwestern Gulf of Mexico. Publ. Inst. Mar. Sci., Univ. Texas 11:90-101. Murphy, M. D., and R. G. Taylor. 1994. Age, growth, and mortality of spotted seatrout in Flor- ida waters. Trans. Am. Fish. Soc. 123:482-497. Nelson, R. S., and C. S. Manooch III. 1982. Growth and mortality of red snappers in the west-cen- tral Atlantic Ocean and northern Gulf of Mexico. Trans. Am. Fish. Soc. 111:465-475. Newman, S. J., D. McB. Williams, and G. R. Russ. 1996. Age validation, growth and mortality rates of the trop- ical snappers (Pisces : Lutjanidae) Lutjanus adetii (Castel- nau, 1873) and L. quinquelineatus (Bloch, 1790) from the central Great Barrier Reef, Australia. Mar. Freshwater Res. 47:575-584. Patterson, W. F., III. 1999. Aspects of the population ecology of red snapper Lut- janus campechanus in an artificial reef area off Alabama. Ph.D. diss., Univ. South Alabama, Mobile, AL, 164 p. Render, J. H. 1995. The life history (age, growth, and reproduction) of red snapper ( Lutjanus campechanus ) and its affinity for oil and gas platforms. Ph.D. diss., Louisiana State Univ., Baton Rouge, LA, x + 76 p. Rivas, L. R. 1966. Snappers of the western Atlantic. Commer. Fish. Rev. 32( 1 ):41 — 44 . 664 Fishery Bulletin 99(4) Robins, C. R., and G. C. Ray. 1986. A field guide to Atlantic Coast fishes of North America. Houghton Mifflin Company, Boston, MA, 354 p. + 64 pi. 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 edition. Am. Fish. Soc. Spec. Publ. 20, Bethesda, MD, 183 p. SAS (Statistical Analysis System). 1985. SAS user’s guide: statistics, version 5 ed. SAS Insti- tute, Cary, NC, 956 p. Smith- Vaniz, W. F., B. B. Collette, and B. E. Luckhurst. 1999. Fishes of Bermuda: history, zoogeography, annotated checklist, and identification keys. Allen Press Inc., Law- rence, KS, 424 p. Stanley, D. R., and C. A. Wilson. 1996. Abundance of fishes associated with a petroleum plat- form as measured with dual-beam hydroacoustics. ICES J. Mar. Sci. 53:743-475. 1998. Spatial variation in fish density at three petroleum platforms as measured with dual-beam hydroacoustics. Gulf Mexico Sci. 16:73-82. Szedlmayer, S. T., and R. L. Shipp. 1994. Movement and growth of red snapper, Lutjanus campechanus, from an artificial reef area in the northeast- ern Gulf of Mexico. Bull. Mar. Sci. 55(2-31:887-896. Thompson, B. A., M. Beasley, and C. A. Wilson. 1998. Age distribution and growth of greater amberjack, Seriola dumerili, from the north-central Gulf of Mexico. Fish. Bull. 97:362-371. Wade, C. W. 1981. Age and growth of spotted seatrout and red snapper in Alabama. Proc. Ann. Conf. S.E. Assoc. Fish & Wildl. Agencies 35:345-354. 665 Stomach content analysis of cobia, Rachycentron canadum, from lower Chesapeake Bay* Michael D. Arendt School of Marine Science College of William and Mary Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Present address: Marine Resources Research Institute South Carolina Department of Natural Resources Division 217 Fort Johnson Road Charleston, South Carolina 29422-2559 E-mail address: arendtm@mrd.dnr.state.scus John E. Olney Department of Fisheries Science School of Marine Science College of William and Mary Virginia Institute of Marine science Gloucester Point, Virginia 23062 Jon A. Lucy Sea Grant Marine Advisory Program Virginia Institute of Marine Science Glooucester Point, Virginia 23062 Cobia ( Rachycentron canadum) is a migratory pelagic species that is found in tropical and subtropical seas of the world, except in the central and eastern Pacific Ocean. The United States ranks third in total commercial production of cobia, however, recreational land- ings generally exceed commercial land- ings by an order of magnitude (Shaffer and Nakamura, 1989). In the western Atlantic Ocean, cobia migrate to Ches- apeake Bay in spring and summer to spawn, and the productive waters of the Bay are believed to constitute important foraging grounds (Joseph et ah, 1964; Richards, 1967). Cobia are known to move to areas of high food abundance, particularly crusta- cean abundance (Darracott, 1977). Recent feeding studies in the north- ern Gulf of Mexico and off North Car- olina have reported geographic differ- ences in cobia diet and have indicated that the relative importance of fishes versus crustaceans is variable and that cephalopods constitute the least signif- icant prey items. In the northern Gulf of Mexico, Franks et al. ( 1996) reported that fish (primarily anchovies, Anchoa sp. ) dominated (% index of relative im- portance [IRI | , Pinkas et ah, 1971) the diet of juvenile cobia (236-440 mm FL). Meyer and Franks (1996) reported crustaceans (primarily portunid crabs) occurred in 79.1% of stomachs and rep- resented 77.6 % of total prey items con- sumed by cobia (373-1,530 mm FL) in the northern Gulf of Mexico. Fish (primarily hardhead catfish, A/'ius fe- lis, and American eel, Anguilla rostra- ta) increased in importance with in- creasing cobia size. Fish were found in 58.5% of all cobia stomachs (20.3% of total prey) but occurred in 84.4% of stomachs of cobia 1150-1530 mm FL (Meyer and Franks, 1996). In con- trast, Smith ( 1995) observed decreased importance of teleosts in the diet of cobia (39-142 cm FL) in North Caroli- na. Elasmobranch fishes and portunid crabs dominated the diet of cobia >9 kg (Smith, 1995). Tag-recapture data collected between 1995 and 1999 document localized movement of cobia within lower Chesa- peake Bay during summer, as well as the return of individual cobia to spe- cific locations or general regions of the lower Bay in subsequent summers.* 1 Al- though Chesapeake Bay is an impor- tant destination for migrating cobia, feeding habits of cobia in the Bay have never been thoroughly examined. Our study documents cobia feeding habits in Chesapeake Bay and compares find- ings with similar cobia studies from North Carolina and the northern Gulf of Mexico. Methods Cobia were sampled opportunistically at marinas and fishing tournaments in lower Chesapeake Bay between June and July 1997. Intact stomachs were removed by cutting above the car- diac sphincter (esophagus) and below the pyloric sphincter (large intestine). Stomachs were labeled, bagged, trans- ported on ice to the VA Institute of Marine Science, and examined in rela- tively fresh condition. An incision was made along the longitudinal axis and the contents of stomachs were emp- tied onto a 500-pm mesh sieve for rins- ing and sorting. Contents were blotted dry on paper towels before counts, dis- placed volumes ( 1-L graduated cylin- der), and identifications to the lowest possible taxon were made. When pos- sible, carapace widths (mm) of crabs were measured with calipers. An index of relative importance (IRI) for all prey items combined was calculated with the formula (% Number + % Volume) x (% Frequency ), as described by Pinkas et al. (1971) and subsequently used by Smith (1995), Meyer and Franks * Contribution 2390 of the Virginia Insti- tute of Marine Science, Gloucester Point, VA 23062. 1 Annual reports. Virginia Game Fish Tag- ging Program, Marine Resources Com- mission, 968 Oriole Dr. South, Suite 102, Virginia Beach, VA 23451. Manuscript accepted 13 April 2001. Fish. Bull. 99:665-670 (2001). 666 Fishery Bulletin 99(4) 700 600 ♦ C. sapidus * O. ocellatus ♦ *♦ 500 400 300 200 100 ♦ 50 60 70 80 90 100 110 120 130 140 150 Cobia fork length (cm) Figure 1 Volume (mL) of portunid crab consumed in lower Chesapeake Bay as a function of cobia size. (1996), and Franks et al. (1996). Frequency (%F) was the percent of all stomachs that contained food. Results and discussion Stomach contents from 114 adult cobia (37-141 cm FL) were examined. Seventy-eight stomachs (68%) contained at least one identifiable, nonbait prey item. One species of bivalve, one species of hydroid, six species of crustacean, one elasmobranch, and 16 species of teleost were observed (Table 1). Mean volume of prey per stomach was 150.6 mL (range: 2-680 mL). Mean number of different prey per stomach was 1.9 species (range, 1-5 species). Blue crab ( Callinectes sapidus) and lady crab ( Ovalipes ocellatus ) dominated the diet of cobia. Index of relative importance (IRI) values for blue crab (5257) and lady crab (3665) were two orders of magnitude higher than IRI values for other prey items. Larger cobia consumed greater volumes of crab (Fig. 1). Additionally, larger cobia consumed larger (sublegal, <13 cm) blue crabs, although similar-size lady crabs were consumed by all siz- es of cobia (Table 2). Atlantic croaker ( Micropogonias un- dulatus), hogchoker ( Trinectus maculatus ), and fish and crab remains constituted the top food items volumetrical- ly after portunid crabs. Although relatively high volumes were observed, IRI values for Atlantic croaker and hog- choker were low owing to infrequent occurrence of these items. High volume of crab remains was consistent with high volumes, counts, and frequency of occurrence of por- tunid crabs in cobia stomachs. Fish remains likely result- ed from finfish bait (predominantly menhaden, Brevoortia tyrannus, and Atlantic croaker, Micropogonias undulatus). Crab and fish remains not identifiable to family and fin- fish bait were excluded from IRI calculations. Consumption of blue crab was greatest along the west- ern shore of the Bay and least at the mouth of the Bay. Conversely, consumption of lady crab was greatest at the mouth of the Bay and least along the eastern shore of the Bay (Fig. 2A). A greater percentage of lady crabs were consumed by male cobia than by female cobia, whereas a greater percentage of blue crabs were consumed by fe- male cobia (Fig. 2B). IRIs for blue crab and lady crab were similar in June, but dramatically different in July. In July, IRI for blue crab was twice as high as that for lady crab. Prey availability and habitat utilization were likely re- sponsible for determining location, sex, and within-season differences in composition of portunid crabs consumed by cobia. June samples were predominantly male cobia col- lected from the Bay mouth and eastern shore of the Bay, whereas July samples were predominantly female cobia collected from the western Bay shore. Cobia feeding habits in lower Chesapeake Bay were more similar to feeding habits reported for cobia from North Carolina (Smith, 1995) than to feeding habits of cobia from the northern Gulf of Mexico (Franks et al., 1996; Meyer and Franks, 1996). Portunid crabs dominated the diet of cobia in Chesapeake Bay and North Carolina (Smith, 1995). Elasmobranch fishes were consumed exclu- sively by large cobia in Chesapeake Bay and North Caro- lina (Smith, 1995). In North Carolina (Smith, 1995), cobia fed on stingrays ( Dasyatis sp.) and smooth dogfish ( Muste - lus canis). Cownose ray (Rhinoptera bonasus ), previously unreported as a food item of cobia, was the only elasmo- branch consumed by cobia in Chesapeake Bay. Cownose ray was observed from eight cobia stomachs; however, on- NOTE Arendt et al.: Stomach content analysis of Rachycentron cancidum 667 Table 1 Index of relative importance (IRI) of prey items consumed by cobia (n= 78) in lower Chesapeake Bay, June-July 1997. IRI was calculated as (% Number + % Volume ) x (% Frequency), as described in Smith ( 1995) and Meyer and Franks (1996). Frequency ( %F ) is the percent of stomachs that contained food. UnID = unidentified. Frequency (stomachs) %F Number (counts) %N Volume (mL) %V IRI Phylum Cnidaria Class Hydrozoa Family Sertulariidae Sertularia sp. 1 1.28 1 0.14 8 0.07 0 Phylum Mollusca Class Bivalva Family Mytilidea Mytilus edulis 2 2.56 47 6.60 17 0.14 17 Phylum Arthropoda Class Crustacea Family Squillidae Squilla empusa 2 2.56 2 0.28 11 0.09 1 Family Pagruidae 3 3.85 3 0.42 5 0.04 2 Family Portunidae Callinectes sapidus 46 58.97 226 31.74 6736 57.31 5252 Ovalipes ocellatus 43 55.13 321 45.08 2505 21.31 3660 Family Canceridae Cancer sp. 2 2.56 2 0.28 26 0.22 1 Cancer irroratus 1 1.28 1 0.14 5 0.04 0 Family Xanthidae Neopanope sp. 1 1.28 3 0.42 5 0.04 1 Phylum Chordata Class Chonrichthyes Family Myliobatidae Rhinoptera bonasus 7 8.97 8 1.12 81 0.69 16 Class Osteichthyes Family Clupeidae Opisthonema oglinum 1 1.28 3 0.42 160 1.36 2 Brevoortia tyrannus 1 1.28 1 0.14 180 1.53 2 UnID Clupeidae 2 2.56 2 0.28 275 2.34 7 Family Batrachoididae Opsanus tau 1 1.28 1 0.14 185 1.57 2 Family Ophidiidae Ophidion marginatum 1 1.28 1 0.14 42 0.36 1 Ophidion sp. 3 3.85 4 0.56 47 0.40 4 Family Sygnathidae Syngnathus floridae 1 1.28 1 0.14 2 0.02 0 Syngnathus sp. 2 2.56 2 0.28 9 0.08 1 Hippocampus erectus 1 1.28 1 0.14 3 0.03 0 Hippocampus sp. 5 6.41 49 6.88 52 0.44 47 Family Pomatomidae Pomatomus saltatrix 1 1.28 1 0.14 180 1.53 2 Family Sciaenidae Micropogonias undulatus 3 3.85 3 0.42 663 5.64 23 Family Uranoscopidae continued 668 Fishery Bulletin 99(4) Table 1 (continued) Frequency (stomachs) %F Number (counts) %N Volume (mL) %V IRI Family Uranoscopidae (continued) Astroscopus sp. 1 1.28 1 0.14 40 0.34 1 UnID Pleuronectiformes 3 3.85 4 0.56 5 0.04 2 Family Bothidae Paralichthyes dentatus 2 2.56 2 0.28 25 0.21 1 Scopthalamus aquosus 1 1.28 1 0.14 15 0.13 0 UnID Bothidae 1 1.28 1 0.14 2 0.02 0 Family Soleidae Trinectes maculatus 9 11.54 14 1.97 440 3.74 66 UnID Soleidae 3 3.85 3 0.42 15 0.13 2 Family Diodontidae Chilomycterus schoepfi 3 3.85 3 0.42 15 0.13 2 Totals 712 11754 Table 2 Consumption of portunid crabs as a function of five cobia size classes in lower Chesapeake Bay. Cobia size class (cm FL) <100 100-109.9 110-119.9 120-129.9 >130 Callinectes sapidus n 9 16 12 23 42 Mean carapace width (mm) 55.2 72.3 75.0 116.7 94.7 Range of carapace width (mm) 33-86 42-120 45-120 42-120 35-120 Ovalipes ocellatus n 42 37 38 32 27 Mean carapace width (mm) 39.1 41.2 37.7 45.3 37.1 Range of carapace width ( mm ) 26-50 30-56 20-51 20-51 27-55 ly the jaw plates (22-49 mm width) remained; thus, the IRI for cownose ray is likely underestimated. Flatfishes and syngnathids represented the teleosts most frequently consumed by cobia in Chesapeake Bay and North Caroli- na. Hogchoker ( Trinectes maculatus ) was more frequently consumed by cobia in Chesapeake Bay (our study) than in North Carolina (Smith, 1995). Blackcheek tonguefish ( Symphurus plaigusa) was regularly consumed by cobia in North Carolina (Smith, 1995) but was absent from cobia stomachs in Chesapeake Bay. In Chesapeake Bay, Syngna- thus sp. (pipefish) and Hippocampus sp. (seahorse) were only important in the diet of smaller cobia, whereas in North Carolina, these fishes were important in the diet of all cobia (Smith, 1995). Cobia in our study predominantly consumed benthic and epibenthic prey items, most notably portunid crabs. Feeding studies on other large predatory fishes in Chesa- peake Bay during the summer months, such as bluefish, Pomatomous saltatj'ix, and weakfish, Cynoscion regalis, reveal that these species consume prey items associated predominantly with pelagic food webs (Hartman and Brandt, 1995a, 1995b). Striped bass, Morone saxatalis, a large predator present in Chesapeake Bay year-round, is also reported to feed predominantly on pelagic fishes (Hartman and Brandt, 1995a, 1995b; Walters, 1999). During the summer, portunid crabs were consumed by bluefish, weakfish, and striped bass in Chesapeake Bay; however, portunid crab consumption was typically less than 5% of total prey items present in the stomachs of these fishes (Hartman and Brandt, 1995a, 1995b; Wal- ters, 1999). Red drum, Sciaenops ocellatus, a large pred- ator that uses Chesapeake Bay between spring and fall is also reported to selectively consume portunid crabs in estuarine environments in the Gulf of Mexico (Scharf and Schlict, 2000); however, feeding habits of red drum in Chesapeake Bay have not been documented. Although Percent of stomachs with crabs NOTE Arendt et al.: Stomach content analysis of Rachycentron ccinadum 669 Q C. sapidus ■ O. ocellatus □ Both crab species Western Shore 0=51) Eastern Shore Bay mouth, coastal 0=7) 0=6) Unknown 0=14) B C. sapidus 1 0. ocellatus O Both crab species B Male 0= 21) Female 0=57) Figure 2 Consumption of blue crab ( Callinectes sapidus) and lady crab (Ovalipes ocel- latus) by cobia in lower Chesapeake Bay by location (A), sex < B ), and month (June and July, C). 670 Fishery Bulletin 99(4) other large predators in Chesapeake Bay occasionally consume portunid crabs, cobia is the only large predator in Chesapeake Bay for which selective consumption of portunid crabs is documented. Acknowledgments Stomachs were collected in conjunction with a study on cobia reproduction in Chesapeake Bay. We thank Susan Crute, Jason Romine, and Holly Simpkins for technical assistance. Donnie Wallace and Harry Johnson Jr. pro- vided access to their facilities at Wallace’s Bait and Tackle, Fox Hill, VA. We thank the recreational and commercial fishermen who allowed their catches to be sampled. This project was supported by grants (RF96-8, RF97-9) from the Virginia Marine Resources Commission upon recom- mendation by the Recreational Fishing Advisory Board. Literature cited Darracott, A. 1977. Availability, morphometries, feeding, and breeding activity of a multi-species, demersal fish stock of the west- ern Indian Ocean. J. Fish. Biol. 10( 1 ): 1—16. Franks, J. S., N. K. Garber, and J. R. Warren. 1996. Stomach contents of juvenile cobia, Rachycentron canadum , from the northern Gulf of Mexico. Fish Bull. 94(2):374-380. Hartman, K. J., and S. B. Brandt. 1995a. Trophic resource partitioning, diets, and growth of sympatric estuarine predators. Trans. Am. Fish. Soc. 124:520-537. 1995b. Predatory demand and impact of striped bass, bluefish, and weakfish in the Chesapeake Bay: applica- tions of bioenergetics models. Can. J. Fish. Aquat. Sci. 52:1667-1687. Joseph, E. B., J. J. Norcross, and W. H. Massmann. 1964. Spawning of cobia, Rachycentron canadum , in the Chesapeake Bay area, with observations of juvenile speci- mens. Ches. Sci. 5:67-71. Meyer, G. H., and J. S. Franks. 1996. Food of cobia, Rachycentron canadum , from the north- central Gulf of Mexico. Gulf Res. Rep. 9(3):161-167. Pinkas, L., M. S. Oliphant, and I. L. K. Iverson. 1971. Food habits of albacore, bluefin tuna, and bonito in California waters. Calif. Dep. Fish Game Fish Bull. 152, 105 p. Richards, C. E. 1967. Age, growth, and fecundity of the cobia, Rachycentron canadum, from Chesapeake Bay and adjacent mid-Atlantic waters. Trans. Am. Fish. Soc. 96(3):343-350. Scharf, F. S„ and K. K. Schlicht. 2000. Feeding habits of red drum ( Sciaenops ocellatus) in Galveston Bay, Texas: Seasonal diet variation and preda- tor-prey size relationships. Estuaries 23( 1):128-139. Shaffer, R. V., and E. L. Nakamura. 1989. Synopsis of biological data on the cobia, Rachycentron canadum (Pisces: Rachycentridae). U.S. Dep. Commer., NOAA Tech Rep NMFS 82 [FAO Fisheries Synopsis 153], 21 p. Smith, J. W. 1995. Life history of cobia, Rachycentron canadum ( Osteich- thyes: Rachycentridae), in North Carolina waters. Brim- leyana 23:1-23. Walters, J. F., III. 1999. Diet composition and feeding habits of large striped bass ( Morone saxatilis ) in Chesapeake Bay. M.S. thesis, School of Marine Science, College of William and Mary, Gloucester Point, VA, 124 p. 671 Preliminary genetic population structure of southern flounder, Paralichthys lethostigma, along the Atlantic Coast and Gulf of Mexico Ivonne R. Blandon Rocky Ward Perry R. Bass Marine Fisheries Research Station Coastal Fisheries Division Texas Parks and Wildlife Department Palacios, Texas 77465 E-mail address (for I R Blandon): ivonne.blandon@tpwd.state.tx.us Tim L. King U.S. Geological Survey-Biological Resource Division Leetown Science Center Aquatic Ecology Laboratory 1700 Leetown Road Kearneysville, West Virginia 25430 William J. Karel Perry R. Bass Marine Fisheries Research Station Coastal Fisheries Division Texas Parks and Wildlife Department Palacios, Texas 77465 James P. Monaghan Jr. North Carolina Division of Marine Fisheries 3441 Arendell Street Morehead City, North Carolina 28557 Southern flounder, Paralichthys letho- stigma, inhabit coastal waters from Albemarle Sound, North Carolina to the Baja Laguna Madre del Sur in northern Mexico, but they are appar- ently absent from southern Florida (Ginsburg, 1952). This species inhab- its coastal bays, sounds, and lagoons from spring to fall and migrates off- shore to spawn in late fall and winter (Stokes, 1977). Valuable sport and com- mercial fisheries for southern flounder exist in both the northern Gulf of Mexico (Warren et al.1; Robinson et al.2) and the western North Atlantic (Monaghan3). Declines in southern flounder ab- solute abundance in some regions (e.g. Texas during the 1980s; Fuls and McEachron4) have prompted some management agencies to institute re- strictions on recreational and commer- cial fisheries including reductions in bag limits and minimum size. Should these measures fail to recover this fish- ery, other measures may be considered by managers, including further restric- tions on harvest, or artificial propaga- tion and stocking, or both. Implemen- tation of such enhancement programs requires that genetic surveys be con- ducted to determine genetic variabili- ty and stock structure of managed fish populations (King et al., 1995). Fail- ure to understand underlying genetic structure prior to implementing stock- ing programs places the genetic re- sources of target species at risk (Al- lendorf et al., 1987) and may result in the reduction or loss of among-popula- tion variability and changes in within- population genetic characteristics. Ge- netic analyses of population structure may also provide insight into manage- ment options that do not require stock- ing (Nelson and Soule, 1987). The objective of our study was to characterize population structure of southern flounder in coastal regions of the northern Gulf of Mexico and north- western Atlantic Ocean and to test the null hypothesis of no genetic differen- tiation within the region surveyed. If this hypothesis was rejected, a number of processes would operate to structure southern flounder population(s). Ge- netic differentiation in some nearshore organisms in the northern Gulf of Mex- ico (e.g. Sciaenops ocellatus\ Gold and Richardson, 1999) has been explained as isolation by distance (Wright, 1943). This model describes a population structured by isolation caused by lim- ited individual migration potential in relation to the size of the species’ dis- tribution (Kimura and Weis, 1964). The hypothesis of isolation by distance is supported when geographic distance and genetic distance are positively cor- related. Alternatively, differentiation may arise as an adaptive response to localized environmental conditions (King and Zimmerman, 1993) or from the operation of physical barriers, such 1 Warren, T. A., L. M. Green, and K. W. Spiller. 1994. Trends in finfish landings of sport- boat anglers in Texas marine waters May 1974-May 1992, 259 p. Manage. Data Ser. 109, Texas Parks and Wildlife (TPW), Coastal Fish. Div., Austin, TX 78744. 2 Robinson, L., P. Campbell, and L. Butler. 1995. Trends in Texas commercial fish- ery landings, 1972-1994, 133 p. Manage. Data Ser. 117, Texas Parks and Wildlife (TPW), Coastal Fish. Div., Austin, TX 78744. 3 Monaghan, J. P, Jr. 1996. Migration of paralichthid flounders tagged in North Carolina. Study 2 in Life history aspects of selected marine recreational fishes in North Carolina, 44 p. Completion Rep. Grant F-43, Segments 1-5, North Carolina Division of Marine Fisheries, 3441 Aren- dell Street, Morehead City, NC 28557. 4 Fuls, B., and L. W. McEachron. 1997. Trends in relative abundance and size of selected finfishes and shellfishes along the Texas coast: November 1975-Decem- ber 1995, 108 p. Manage. Data Ser. 137, Texas Parks and Wildlife (TPW), Coastal Fish. Div., Austin, TX 78744. Manuscript accepted 9 February 2001. Fish. Bull. 99:671-678 (2001). 672 Fishery Bulletin 99(4) Figure 1 Collection sites for southern flounder in North American waters. LLM = Lower Laguna Madre; MAT = Matagorda Bay; GAL = Galveston Bay; SAB = Sabine Lake; MS = Mississippi; AL = Alabama; STA = St. Augustine, Florida; NC = North Carolina. as current patterns (King et al., 1994). Support for com- peting models comes from examination of specific patterns observed in structured populations. Such patterns, if ob- served, may have important management implications. Materials and methods Southern flounder were collected during the summers of 1996 and 1997 by rod and reel, flounder gigs, or Texas Parks and Wildlife (TPW) gill nets in four Texas bays (Sabine Lake, Galveston Bay, Matagorda Bay, and lower Laguna Madre), by pound nets in Core Sound, North Caro- lina, from gill nets in estuarine waters near Biloxi, Missis- sippi, and from commercial fish houses in Dauphin Island, Alabama, and St. Augustine, Florida (Fig. 1). Southern flounder from commercial fish houses were reported by the house operator to be caught locally. A majority of individu- als collected were adult and were not reliably assignable to year classes. Samples were screened by using isoelec- tric focusing (IEF) of sarcoplasmic proteins (methods of Ward et al., 1995) to insure that individuals belonging to other Paralichthys species were not included in our analy- ses (necessary because of accidental inclusion of congener- ics both in samples obtained from commercial sources and in samples of juvenile flounder obtained during routine resource sampling in Texas). Skeletal muscle, liver, heart, and kidney tissues were excised from fresh or frozen fish. Sample preparation and electrophoretic techniques and conditions followed those of King and Pate (1992). Gel and electrode buffers used were tris-borate-EDTA, pH 8.0 (Selander et al., 1971), tris- citrate, pH 8.0. (Selander et al., 1971), lithium hydroxide, pH 8.0 (Selander et al., 1971), borate buffer, pH 9.0 (modi- fied from Sackler, 1966), and Poulik’s discontinuous sys- tem, pH 8.7 (Selander et al., 1971). Histochemical meth- ods were primarily taken from Manchenko ( 1994). Genetic nomenclature followed the recommendations of Shaklee et al. (1990). BIOSYS-1 (Swofford and Selander, 1981) was used to generate an allele frequency table, to estimate the pro- portion of loci heterozygous ( H ) in the average individual, the proportion of polymorphic loci in individuals from each bay population, and genetic divergence using the E-statistics of Wright (1978). Exact tests calculated by GENEPOP (v. 3.1; Raymond and Rousset, 1995) were used to test for conformance of genotypic frequencies at each locus within a sample to Hardy- Weinberg expectations, genotypic linkage disequilibrium, and allelic and genotypic heterogeneity. Pairwise differences between samples were tested by using the genic differentiation randomization test in GENEPOP. Results were combined over loci with Fisher’s method (Sokal and Rohlf, 1994). Differences be- tween each pair of populations were summarized by us- ing the chord distance of Cavalli-Sforza and Edwards (1967). An unrooted phylogenetic tree was fitted to the chord distance matrix by using the neighbor-joining (NJ) algorithm. TreeView (Page, 1996) was used to visualize NOTE Blandon et al.: Genetic population structure of Parcilichthys lethostigma 673 the tree. The strength of support for each node in the tree was tested by bootstrapping over loci with NJBPOP (Cornuet et ah, 1999). To further quantify spatial hetero- geneity, the fixation index (FST ) was calculated for each lo- cus to provide measures of interpopulation differentiation and estimates of reduction in heterozygosity of a subpopu- lation due to population subdivision. A %2 test was used to test the null hypothesis, FST= 0 (Workman and Niswan- der, 1970). All /2 probability values from tests for con- formance to Hardy- Weinberg expectations, heterogeneity, linkage disequilibrium, and FST were adjusted for multiple simultaneous tablewide tests by using sequential Bonfer- roni adjustments to minimize type-I statistical inference errors (Rice, 1989). Partitioning of variance components among geographic regions and within samples was accomplished by using a hierarchical analysis of molecular variance (AMOVA, Cockerham, 1969, 1973) with the package ARLEQUIN (version 2, Schneider et ah, 1999). Sample sites were nest- ed into regional groupings for separate analyses (Atlantic versus Gulf of Mexico, and Atlantic combined with east- ern Gulf sites versus western Gulf). A phenogram was generated from the chord distance matrix with the neigh- bor-joining (N-J) algorithm. The N-J phenogram, with bootstrap estimates (as percentage of 10,000 replications) obtained by resampling loci within samples, was gener- ated with NJBPOP (Cornuet, et al., 1999). The signifi- cance of the relationship between genetic (i.e. chord) and geographic (bay to bay shoreline distance) distance matri- ces was determined by sampling the randomization dis- tribution generated from 1000 replications with the MX- COMP (matrix comparison) routine in NTSYS-PC 2.0 by Rohlf (1997) to allow a Mantel test (Mantel, 1967). An assignment test (WHICHRUN 4.1, Banks and Eichert, 2000) tested the ability to discriminate population of ori- gin based on an individual’s multilocus genotypic profile. Clinal trends in heterozygosity and allele frequency were examined by using nonparametric correlation analy- ses (SAS Institute, 1989). Significance of Spearman’s cor- relation coefficients was determined as the probability a correlation differed from zero. Probabilities less than 0.05 were considered statistically significant (Snedecor and Co- chran, 1980). Results Misidentifications detected by IEF resulted in a reduced sample size in some samples especially from Alabama and Florida. Examination of 46 enzymes and structural pro- tein systems in southern flounder produced scorable phe- notypes for 68 putative gene loci. Two dimeric esterase loci (ESTD-V* and ESTD 2*. 3.1.1.. IIUBMBNC, 19921 ) a tripeptide aminopeptidase locus ( PEPB-2* , 3.4.13..), glyc- erol-3-phosphate dehydrogenase (G3PDH *, 1.1. 1.8), two glucose-6-phosphate isomerase loci ( GPPA and GPPB \ 5.3. 1.9), phosphoglucomutase (PGM*, 5. 4. 2. 2) and two glucose-6-phosphate dehydrogenase loci (G6PDH-1* and G6PDH-2* , 1.1. 1.49) were resolved, scored as variable, and included in analyses. The remaining 59 loci were mono- morphic or could not be scored consistently and were omit- ted. All polymorphic loci bad the same common allele across all localities (Table 1). ESTD-1*, ESTD-2 *, GPPB*, and G6PD-1* each expressed an allele unique to a single locality. The percentage of polymorphic loci ( P0 99) averaged 7.5% and ranged from 4.41% to 10.29% (Table 1). Mean individual heterozygosity ranged from 0.03 (SE=0.02) in North Carolina and Florida to 0.12 (SE=0.12) in Matago- rda Bay. Statistically significant clinal relationships were found in the Gulf of Mexico between the frequency of the common allele of the G6PDH-2* locus and degrees west longitude (rs=-0.829, P<0.05) and degrees north lat- itude (rs=0.829, P<0.05). Further differentiation of popu- lations in the western Gulf of Mexico was observed at the G3PDH* locus. The G3PDH*1 allele occurred at a fre- quency of 12.0%) in the Laguna Madre, declined to a fre- quency of 1.4% in Galveston Bay, and was absent in all collections east of Galveston Bay. Four loci had alleles con- fined to a single locality in this same region of Gulf of Mexico (Table 1), including G6PDH-1*2 which was limited to the Laguna Madre. Samples were in Hardy-Weinberg equilibrium at the nine allozyme loci surveyed for each of the 72 possible comparisons of variable loci for each locality, except ESTD-2 * in Galveston Bay. No statistically significant genotypic linkage disequilibrium was observed between any loci for any population. Tests for homogeneity of allelic frequencies at variable loci across all localities were statistically significant at six of the nine loci: ESTD-2 * (PcO.Ol), G3PDH * (P<0.01), GPPA (P=0.01), GPPB* (PcO.Ol), PGM* (PcO.Ol), and G6PDH*-2 (P<0.01). Tests for homogeneity of genotypic frequencies across all localities were significant at four loci: G3PD1P (P<0.01), GPPB* (P<0.01), PGM* (PcO.Ol), and G6PDH-2* (PcO.Ol). FST averaged 0.088 and varied from 0.181 for G6PDH-2* to 0.001 for G6PDH-1*. All esti- mates of Fst were found to be statistically different from 0, except for loci G6PDH-1 *, PEPB-2 *, and GPPA*. Regional differentiation among southern flounder was demonstrated (Fig. 2) by using chord distance and N-J clustering. The greatest discontinuity was between all samples from Galveston Bay eastward and a cluster com- posed of Matagorda Bay and the Laguna Madre. No ap- parent differentiation exists between Atlantic Coast pop- ulations and populations collected from the eastern Gulf of Mexico. This pattern is supported by the outcome of the assignment test (Table 2). Percentage correct assign- ment ranged from 0% for Florida to 78% for Matagorda Bay. However, when assignments between groups identi- fied in the cluster analysis were examined, the overall correct percentage increased to 81% . The assignment test supports the interpretation of the cluster analysis, sug- gesting within-region differentiation was minimal and not geographically consistent, but between-region differences were considerable. A different pattern was discerned by using AMOVA (Table 3). The majority of the variance (>99%) was distributed within samples and the among- sites-within-coasts component was nonsignificant. Com- parison of Atlantic versus Gulf of Mexico sample sites was statistically significant (P=0.04) and comparison of west- ern Gulf of Mexico sample sites with those from the east- 674 Fishery Bulletin 99(4) Table t Allele frequencies, sample sizes (n), percentage of polymorphic loci per population, and mean heterozygosity per locus. LM = Laguna Madre, Texas; MAT = Matagorda Bay, Texas; GAL = Galveston Bay, Texas; SAB = Sabine Lake, Texas; MIS = off Biloxi, Mississippi; ALA = Mobile Bay, Alabama; STA = off St. Augustine, Florida; NC = Core Sound, North Carolina. Locus Allele LM MAT GAL SAB MIS ALA STA NC ESTD-1 * I’3 *1 0.972 0.918 0.889 0.917 0.885 1.000 1.000 0.959 * 2 0.028 0.082 0.111 0.083 0.115 0.000 0.000 0.027 *3 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.014 n 54 61 36 24 13 21 20 37 ESTD-2* ]-3 * i 0.009 0.000 0.014 0.000 0.000 0.000 0.000 0.054 *2 0.981 1.000 0.917 1.000 1.000 1.000 1.000 0.946 *, 3 0.009 0.000 0.069 0.000 0.000 0.000 0.000 0.000 n 54 61 36 24 13 21 20 37 PEPB-2 * l-3 *i 0.074 0.016 0.014 0.000 0.038 0.024 0.025 0.000 *2 0.926 0.984 0.986 0.958 0.962 0.976 0.975 1.000 * 3 0.000 0.000 0.000 0.042 0.000 0.000 0.000 0.000 n 54 61 36 24 13 21 20 37 G3PDH* 23 * i 0.120 0.098 0.014 0.000 0.000 0.000 0.000 0.000 *2 0.778 0.779 0.972 0.848 0.808 0.833 0.875 0.946 *3 0.102 0.123 0.014 0.152 0.192 0.167 0.125 0.054 n 54 61 36 23 13 21 20 37 GPI-A* 1-3 * i 1.000 0.992 1.000 1.000 0.962 0.929 0.975 1.000 *2 0.000 0.008 0.000 0.000 0.038 0.071 0.025 0.000 n 54 61 36 24 13 21 20 36 GPI-B* lA ■'7 0.000 0.000 0.000 0.125 0.000 0.000 0.000 0.000 *2 1.000 1.000 0.986 0.875 1.000 1.000 1.000 1.000 *3 0.000 0.000 0.014 0.000 0.000 0.000 0.000 0.000 n 54 61 36 24 13 21 20 37 PGM * 1A * i 0.009 0.008 0.086 0.000 0.000 0.000 0.000 0.000 * 2 0.991 0.992 0.829 1.000 1.000 1.000 1.000 1.000 *3 0.000 0.000 0.086 0.000 0.000 0.000 0.000 0.000 n 54 61 35 24 13 21 20 37 G6PDH-1* 13 *i 0.991 1.000 1.000 1.000 1.000 1.000 1.000 1.000 * 2 0.009 0.000 0.000 0.000 0.000 0.000 0.000 0.000 n 54 61 36 24 13 21 20 37 G6PDH-2* I-3 *i 0.000 0.000 0.114 0.000 0.000 0.095 0.025 0.000 *2 0.639 0.648 0.871 0.896 0.962 0.881 0.975 1.000 ■:3 0.361 0.352 0.014 0.104 0.038 0.024 0.000 0.000 n 36 61 35 24 13 21 20 37 % polymorphic loci 10.29 8.82 10.29 7.35 7.35 5.88 5.88 4.41 Heterozygosity 0.10 0.12 0.09 0.08 0.08 0.09 0.03 0.03 Direct count (SE) (0.05) (0.06) (0.04) (0.03) (0.03) (0.04) (0.02) (0.02) ' Buffer = TVB (Tris-borate-EDTA, pH 8.0). - Buffer = borate 9. 3 Tissue = muscle. 3 Tissue = liver. ern Gulf and the Atlantic coast approached significance (P=0.07). The results of chi-square tests for population dif- ferentiation (Table 4) allowed a pairwise examination of differences among the eight sampling sites. Eastern sam- ple sites were not significantly different (P>0.01) from their nearest geographic neighbors. Western sites, from Sabine Lake to the lower Laguna Madre, were statis- tically different from at least one neighbor, suggesting more pronounced genetic differentiation in the western Gulf of Mexico. Among Gull' of Mexico samples, no sta- NOTE Blandon et al.: Genetic population structure of Paralichthys lethostigmci 675 Table 2 Assignment test outcomes. Values indicate frequency of assignment of individuals from a collection locality (rows) to the locality to which it is most similar (columns). Numbers on the diagonal indicate correct assignments. The “East” category combines collections from North Carolina to Sabine Lake, Texas, and the “West” category combines Texas sites from Galveston Bay to Laguna Madre. NC = Core Sound, North Carolina; STA = off St. Augustine, Florida; ALA = Mobile Bay, Alabama; MIS = off Biloxi, Mississippi; SAB = Sabine Lake, Texas; GAL = Galveston Bay, Texas; MAT = Matagorda Bay, Texas; LM = Laguna Madre, Texas. Assigned to From NC STA ALA MIS SAB GAL MAT LM NC 29 0 4 0 2 0 0 0 STA 13 0 5 1 0 1 0 0 ALA 8 0 6 3 0 0 0 0 MIS 4 1 0 4 3 1 0 0 SAB 15 1 4 2 10 5 0 0 GAL 9 0 2 1 2 22 1 0 MAT 8 0 6 6 1 0 14 25 LM 7 0 4 0 0 1 3 19 East 158 1 West 33 61 MAT m Figure 2 Unrooted neighbor-joining tree depicting the underlying structure found in the pairwise Cavalh-Sforza and Edwards chord distance matrix for eight collections of southern flounder. Numbers indicate bootstrap support in 1000 replications and white ellipses indicate nodes with bootstrap support below 50%. LLM = Lower Laguna Madre; MAT = Matagorda Bay; GAL = Galveston Bay; SAB = Sabine Lake; MS = Mississippi; AL = Alabama; STA = St. Augus- tine, Florida; NC = North Carolina. tistically significant correlations were observed be- tween chord distances and geographic distances h-0.416, £=1.549, P=0.06). Discussion Genetic differentiation was not extensive over most of the range of P. lethostigmci examined in this study. Samples collected from Core Sound in North Carolina to Sabine Lake on the upper Texas coast were genetically similar. However, a discontinuity in allele frequencies was identified on the Texas coast between Galveston and Matagorda Bays. In addition, statistically significant dines in allele fre- quencies at the G6PDH-2 * locus and in average individual heterozygosity were observed across the Gulf of Mexico. These observations do not suggest the occurrence of independent stocks of southern flounder in the Gulf of Mexico but do support the hypothesis that genetic structuring is present. Southern flounder samples collected off St. Augus- tine, FL, and off North Carolina cluster with sam- ples in the Gulf of Mexico from Galveston Bay, Texas, eastward, despite a modern-era distribu- tional gap that encompasses the southern reaches of Florida from the Loxahatchee River on the Atlan- tic Coast to the Caloosatchee Estuary on the Gulf Coast (NOAA5). This apparent gap may not repre- sent an effective barrier to gene flow or may be of such recent origin that differentiation has been minimal. It is also possible that differences existed that were undetected by techniques used in our study. Significant correlation between genetic and geographic distance was not found, lending no support to application of the isolation by distance model (Wright, 1943) to pop- 5 NOAA. 1985. Gulf of Mexico coastal and ocean zones stra- tegic assessment:data atlas, 188 p. Strat. Assess. Branch, Ocean Assess. Div., Off. Oceanography Mar. Assess., Nat. Ocean Serv. and the Southeast Fish. Center, NMFS, NOAA, U.S. Dep. Commer., 75 Virginia Beach Dr, Miami, FL 33149. 676 Fishery Bulletin 99(4) ulation structuring in southern flounder. The estimated ^st value of 0.088 for southern flounder was greater than that found for Sciaenops oeellatus (FST=0.022, Gold et ah, 1994), Cynoscion nebulosus (FST= 0.009, King and Pate, 1992), or Pogonius chromis (FST=0.013, Karel, unpubl. data) in this region. Physical or biotic factors may have caused greater isolation for southern flounder than for other nearshore species. Cluster analysis suggested that Table 3 Hierarchical analyses of molecular variation (AMOVA) among mtDNA composite haplotypes of southern flounder ( Paralichthys lethostigma ) from the U.S. Atlantic coast and Gulf of Mexico. Source of variation Variance % variation P' All sites Among coasts 0.00026 0.06 0.04 Among sites within coasts -0.00020 -0.04 0.92 Within sites 0.44985 99.99 0.74 Eastern versus western cluster Among clusters 0.00042 0.09 0.07 Among sites within regions -0.00034 -0.08 0.10 Within sites 0.44985 99.98 0.76 ' Probability of finding a more extreme variance component by chance alone (1000 permutations). the region of greatest differentiation occurs along the mid- dle Texas coast. Similar genetic structure in Crassostrea virginica may be explained by seasonal current patterns in Corpus Christi Bay, Texas (King et al., 1994). A com- parable mechanism may operate in southern flounder; off- shore currents may have resulted in reduced dispersion of eggs and larvae between the upper and middle Texas coasts. Currents off the Texas coast are seasonally vari- able and complex (Cochrane and Kelly, 1986) and may aid in reducing egg and larval dispersion between regions of the western Gulf of Mexico. Many marine species spawn in the open ocean, have eggs and larvae with an extensive planktonic stage, or are highly mobile as adults. It is not surprising that such or- ganisms are often panmictic, or exhibit subdivision on on- ly broad levels (e.g. Sciaenops oeellatus-, Gold et al., 1994). When genetic differentiation is found in marine organ- isms (e.g. Cynoscion nebulosus', King and Pate, 1992), ex- tensive regions of clinal change in allele frequency may be seen and may be an adaptive feature (King and Zimmer- man, 1993). A critical question for fishery management is how much differentiation is necessary to indicate bi- ologically significant population structuring. Gold et al. (1994), for instance, found that red drum were subdivided (albeit weakly) between Atlantic and Gulf of Mexico sub- populations despite relatively high levels of gene flow be- tween the populations and failure of cluster analyses to consistently segregate localities into proper geographic re- gions. However, significant differences in allele frequen- cies for two loci were found and the researchers were able to demonstrate, through hierarchical gene-diversity anal- ysis, that 20% of the variation in gene diversity was re- lated to within-region differences. Table 4 Tests Inf. (<0.01) Inf. (<0.01) Inf. Inf. (<0.01) (<0.01) Laguna Inf. Inf. Inf. 42.178 Inf. Inf. 26.956 Madre (<0.01) (<0.01 ) (<0.01) (<0.01) (<0.01) (<0.01) (0.07) NOTE Blandon et al.: Genetic population structure of Parcilichthys lethostigmci 677 In our study, allele frequency discontinuities and clinal variation in genetic characters were identified among southern flounder inhabiting the western portion of the species’ range. Short-term goals (e.g. supplementing ex- ploited populations) that fail to account for this structur- ing could result in management programs that undermine the long-term resource management objective: maintain- ing the evolutionary potential of this species. The results of this survey suggest that southern flounder populations should be considered potentially distinct pending further resolution of population differentiation and clinal varia- tion. Stocking efforts, especially those involving interbay transfers, should be undertaken only after careful consid- eration of all pertinent information, and be contingent up- on careful studies of genetic variation at the appropriate local level. Observed discontinuities and clinal variations in allele frequencies may indicate adaptation to localized conditions and should be incorporated into comprehensive management strategies developed for southern flounder. Acknowledgments Financial support for this research was received from the Sport Fish Restoration Program of the U. S. Fish and Wild- life Service. Field personnel of the TPW Coastal Fisheries Division provided invaluable assistance in sample collec- tion. Additional samples were provided by Greg Stunz of Texas A&M University, and Bradley Randall and his colleagues in the Fishery Department of the Gulf Coast Research Laboratory of Ocean Springs, Mississippi. We thank Larry McEachron and Bob Colura for administra- tive support and assistance with sample collection, and Roberta Vickers, David Westbrook, and Eric Young for their laboratory assistance, as well as Andrew Shaw for the preparation of Figure 1. Literature cited Allendorf, F. W., N. Ryman, and F. M. Utter. 1987. Genetics and fishery management: past, present, and future. In Population genetics and fishery management (N. Ryman and F. Utter, eds.), p. 1-19. Univ. Washington Press, Seattle, WA. Banks, M. A., and W. Eichert. 2000. WHICHRUN (version 3.2) a computer program for population assignment of individuals based on multilocus genotype data. J. Hered. 91:87-89. Cavalli-Sforza, L. L., and A. W. F. Edwards. 1967. Phylogenetic analysis: models and estimation proce- dures. Am. J. Hum. Gen. 19:233-257. Cochrane, J. D., and F. J. Kelly. 1986. Low-frequency circulation on the Texas-Louisiana continental shelf. J. Geophys. Res. 91:10645-10659. Cockerham, C. C. 1969. Variance of gene frequencies. Evolution 23:72-83. 1973. Analysis of gene frequencies. Genetics 74:679-700. Cornuet, J.-M., S. Piry, G. Luikart, A. Estoup, and M. Solignac. 1999. New methods employing multilocus genotypes to se- lect or exclude populations as origins of individuals. Genet- ics 153:1989-2000. Ginsburg, I. 1952. Flounders of the genus Paralichthys and related genera in American waters. Fish. Bull. 52:267-351. Gold, J. R., T. L. King, L. R. Richardson, D. H. Bohlmeyer, and G. C. Matlock. 1994. Genetic studies in marine fishes. VII: allozyme differ- entiation within and between red drum ( Sciaenops ocella- tus) from the Gulf of Mexico and Atlantic Ocean. J. Fish Biol. 116:517-185. Gold, J. R., and L. R. Richardson. 1999. Population structure of two species targeted for marine stock enhancement in the Gulf of Mexico. Bull. Natl. Res. Inst. Aquacult. (suppl. 1 ):79-87. IUBMBNC (International Union of Biochemistry and Molecular Biology, Nomenclature Committee). 1992. Enzyme nomenclature. Academic Press, Orlando, Florida, 862 p. Kimura, M., and G. H. Weis. 1964. The stepping stone model of population structure and the decrease of genetic correlation with distance. Genet- ics 49:561-576. King, T. L., and H. O. Pate. 1992. Population structure of spotted seatrout inhabiting the Texas Gulf coast: an allozymic perspective. Trans. Am. Fish. Soc. 121:746-756. King, T. L., R. Ward, I. R. Blandon, R. L. Colura, and J. R. Gold. 1995. Using genetics in the design of red drum and spotted seatrout stocking programs in Texas: a review. In Uses and effects of cultured fishes in aquatic ecosystems (H. L. Schramm Jr. and R. G. Piper, eds.), p. 499-502. Am. Fish. Soc. Sym. 15, Am. Fish. Soc., Bethesda, MD. King, T. L., R. Ward, and E. G. Zimmerman. 1994. Population structure of eastern oysters (Crassostrea virginica) inhabiting the Laguna Madre, Texas and adja- cent bay systems. Can. J. Fish. Aquat. Sci. 51:215-222. King, T. L., and E. G. Zimmerman. 1993. Clinal variation at aspartate aminotransferase-2 in spotted seatrout, Cynoscion nebulosus (Cuvier), inhabiting the north-western Gulf of Mexico. Anim. Gen. 24:59-61. Manchenko, G. P. 1994. Handbook of detection of enzymes on electrophoretic gels. CRC Press, Boca Raton, FL, 341 p. Mantel, N. A. 1967. The detection of disease clustering and a generalized regression approach. Cancer Res. 27:209-220. Nelson, K., and M. Soule. 1987. Genetical conservation of exploited fishes. In Popu- lation genetics and fishery management (N. Ryman and F. Utter, eds.), p. 345-368. LTniv. Washington Press, Seattle, WA. Page. D. D. M. 1996. TREEVIEW: an application to display phylogenetic trees on personal computers. Comput. Appl. Biosci. 12: 357-358. Raymond, M., and F. Rousset. 1995. GENEPOP (version 1.2): population genetics software for exact tests and ecumenism. J. Heredity 86:248-249. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43: 223-225. Rohlf, F. J. 1997. NTSYS-pc: numerical taxonomy and multivariate analysis system, version 2.0. Exeter Software, Setauket, NY, 32 p. Sackler, M. L. 1966. Xanthine oxidase from liver and duodenum of the rat: 678 Fishery Bulletin 99(4) histochemical localization and electrophoretic heterogene- ity. J. Histochem. Cytochem. 14:326-333. SAS Institute Inc. 1989. SAS/STAT user’s guide, version 6, 4th ed., vol. 1. SAS Institute, Cary, NC, 890 p. Schneider, S., D. Roessli, and L. Excoffier. 1999. ARLEQUIN ver. 2.0: a software for population genetic data analysis. Genetics and Biometry Laboratory, LTniv. Geneva, Geneva, Switzerland. Selander, R. K., M. H. Smith., S. Y. Yang, W. E. Johnson, and J. B. Gentry. 1971. Biochemical polymorphism and systematics in the genus Peromyscus. Variation in the old field mouse ( Pero - inyscus polionotus). Studies in Genetics IV, Univ. Texas Publ. 7103:49-90. Shaklee, J. B„ F. W. Allendord, D. C. Morizot, and G. S. Whitt. 1990. Gene nomenclature for protein-coding loci in fish. Trans. Am. Fish. Soc. 119:2-15. Snedecor, G. W., and W. G. Cochran. 1980. Statistical methods. Iowa State Univ. Press, Ames, IA, 524 p. Sokal, R. R., and F. J. Rohlf. 1994. Biometry: the principles and practice of statistics in biological research, 3rd ed. W. H. Freeman and Co., New York, NY, 880 p. Stokes, G. M. 1977. Life history studies of southern flounder (Paralich- thys lethostigma ) and Gulf flounder (P. albigutta) in the Aransas Bay areas of Texas. Texas Parks and Wildlife Dep. Tech. Ser. 25:1-37. 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. Heredity 72:282-283. Ward, R., I. R. Blandon, and B. W. Bumguardner. 1995. Hybridization among members of the genus Morone (Pisces: Percichthyidae) in Galveston Bay, Texas. Texas J. Sci. 47:155-158. Workman, P. L., and J. D. Niswander. 1970. Population studies on southwestern Indian tribes II. Local genetic differentiation in the Papago. Am. J. Hum. Gen. 22:24-49. Wright, S. 1943. Isolation by distance. Genetics 28:114-138. 1978. Evolution and the genetics of populations, vol. 4. Vari- ability within and among natural populations. Univ. Chi- cago Press, Chicago, IL, 580 p. 679 Age determined from the daily deposition of concentric rings on common octopus (Octopus vulgaris ) beaks Jose L. Hernandez-Lopez Jose J. Castro-Hernandez Departamento de Biologia Universidad de Las Palmas de Gran Canaria Apdo. 550, Las Palmas de Gran Canaria Canary Islands, Spain E-mail address (for J. J Castro-Hernandez, contact author): |ose|uan.castro@biologia ulpgc.es Vicente Hernandez-Garcia Sea Fisheries Institute, Research Station 72-600 Swinoujscie, Poland The common octopus ( Octopus vulga- ris Cuvier, 1797) is an Atlantic and Mediterranean species (Guerra, 1992; Mangold, 1998). It is one of the most important target species of the North- west African fisheries (Hernandez and Bas, 1993; Foucher et al., 1998). The common octopus catch reported for this area in 1994 was 137,844 t, represent- ing 47.17% of the total world octopus catch. In 1996 it was 156,300 t, repre- senting 50.03% of the total world octo- pus catch (FAO, 1998). Octopus age and growth have been determined by laboratory rearing stud- ies (Itami et ah, 1963; Nixon, 1969; Mangold and Boletzky, 1973; Smale and Buchan, 1981; Villanueva, 1995) and by field studies (Guerra, 1979; Hat- anaka 1979; Pereiro and Bravo de La- guna, 1979). Growth rates can be cal- culated for animals maintained in the laboratory, but comparison with growth under natural conditions is question- able (Mangold, 1983). In field studies, growth and age can be correlated when there is clear evidence that a single year class from a stable population is under consideration, but where the spawning season is very long as in the common octopus (Mangold, 1983; Guerra, 1992), identifying year classes is difficult (Guerra, 1979, Hatanaka, 1979). Cephalopod age has been determined by several methods: Guerra (1979), Pereiro and Bravo de Laguna (1979), and others have reported growth and age correlations by following a single year class from a stable population. Concentric rings on statoliths (Young, 1960), the internal shell, and eye lenses (Gongalves, 1993) of Octopus vulgaris have been reported. Raya and Hernan- dez-Gonzalez (1998) observed marks on the internal rostral area of beaks from common octopus, possibly related to daily growth. None of these methods has been val- idated for known-age Octopus vulgaris. Furthermore, all require fairly complex methods for preparing structures pri- or to observation under the microscope (polishing, embedding in resin, and sec- tioning with a diamond, etc.) which hinders their application to field stud- ies. This paper provides an easy new method of determining Octopus vulgar- is age based on the upper beak micro- structure, validated for the paralarval period. Material and methods The study was carried out on 275 common octopus ( 164 males and 111 females) collected between January 1998 and May 1999, from catches of the small-scale fishery off the island of Gran Canaria (The Canary Islands, central-east Atlantic). An additional sample of 27 Octopus vulgaris paralarvae was obtained from spawning females that deposited and incubated egg bunches in plastic bur- rows inside a 12,000-L tank. The embryonic development took between 25—30 days at a temperature range of 19-22°C. Once hatched, the paralarvae were transferred to transparent 12-L containers with open seawater flow, in July 1997 and June and July 1999, and reared in the laboratory at 19-22°C water temperature and natural pho- toperiod. The dates of hatching and death of each paralarva were record- ed. The bottom was siphoned daily to remove dead individuals. During rear- ing, paralarvae were fed with recently hatched crab zoeae (see Hernandez- Garcia et al., 2000). Ventral mantle length (VML) was measured in both benthic octopus and paralarvae to the nearest 0.1 mm. We used VML as the body measurement because we consider it to be more ac- curate than dorsal mantle length. To- tal body weight (TW) of benthic octo- pus was recorded to the nearest 0.01 g (to the nearest 0.0001 g in paralarvae). With the exception of the paralarvae, all the specimens were sexed. The beak of each animal, including paralarvae, was removed and stored in 70% ethyl alcohol. Lower and upper beaks were sagittally sectioned with scissors to obtain two symmetrical half beaks (Fig. 1). The half beaks were cleaned with water and the mucus cov- ering the inner part of the lateral walls was removed by rubbing it softly with the fingers (obviously, this operation was not necessary in the case of the paralarvae beaks). By using a stereoscopic microscope, the concentric rings in the lateral wall of each beak were counted from the rostral tip area to the opposite end of the lateral wall. Because of the lack of pigmentation in paralarvae beaks, the concentric rings in their lateral wall were more easily counted with a microscope. Rings of each beak were Manuscript accepted 10 April 2001. Fish. Bull. 99:679-684 (2001). 680 Fishery Bulletin 99(4) Upjter beak Half upper beak (3D view) (inner view) Figure 1 Diagram of upper beak of Octopus vulgaris and lateral wall showing growth bands, rostral tip area, the distal posterior beak end, and counting line. counted at least three times by the same person, and those with less than two identical counts were rejected from analysis. The number of rings in beaks of paralarvae was com- pared with the number of days each one lived. Results Octopus obtained from the small-scale fishery ranged in size from 4.8 to 165 mm VML, and weighed between 0.38 and 3926 g. Females ranged from 60 to 165 mm VML and weighed from 215 to 3926 g. Males ranged from 58 to 160 mm VML and from 200 to 3167 g in weight. We were unable to sex two individuals (4.8 mm VML, 0.38 g and 8.1 mm VML, 0.60 g). Paralarvae ranged from 1.0 to 2.7 mm VML and from 0.001 to 0.005 g in weight. The internal lateral walls of upper beaks from 302 in- dividuals (27 paralarvae and 275 individuals in benthic stages of Octopus vulgaris) revealed a pattern of concen- tric bands deposited from the rostral tip of the beak to the opposite margin of the lateral wall, parallel to the beak edges. Both halves of the upper beaks showed similar spa- tial and density patterns of microstructures. Lower beaks showed no regularity in the pattern of bands along the lateral walls and were discarded. On the upper beaks the distance from the rostral tip to the distal end of the later- al wall showed a positive correlation with the VML (Pear- son’s correlation: r=0.825; PcO.OOl) (Fig. 2). Ring counts were more difficult near the rostral tip where rings were frequently discontinuous. Counts were easier to make near the edges of the lateral wall which was less highly pigmented. Paralarvae survived in tanks from 3 to 26 days. For 48.1% of paralarvae, the concentric ring count in the later- al wall of the upper beak equaled the number of days that they lived. Otherwise, in 22.2% and 29.6% of paralarvae the number of rings counted were one more or one less, respectively, than the number of days of age. These data (Fig. 3) indicate that daily deposition of a growth incre- ment in the lateral wall of the upper beak begins on day one after hatching. The weight-age and VML-age relation- ships of paralarvae (Fig. 4, A and B) were similar to those found by Villanueva (1995), with differences attributable to rearing conditions. Given the correlation between incre- ment counts and age of paralarvae, we applied the upper beak ring count method for age determination of 272 com- mon octopus, ranging from 4.8 to 165 nun VML. Results should be taken with caution for the benthic stages of oc- topus pending the validation of growth of adults and the frequency of rings deposition. Increments counted on beaks from octopus collected in the wild ranged from 53 to 398 corresponding to indi- viduals of 0.38 and 3926 g body weight, respectively (4.8 and 165 mm VML). Males and females had no difference in the number of rings counted in the lateral walls of up- per beaks (ANOVA, F=0.0006, P=0.98). The age of males ranged between 3.2 and 12.3 months (95-369 rings), and females ranged between 3.1 and 13.3 months (93-398 rings), see Figure 5 (A and B) for weight-age and MVL-age relationships for benthic octopus. Discussion In Octopus vulgaris and other shallow-water cephalo- pods, regular patterns of activity and evidence of endog- enous rhythms induced by the light-dark cycles have been reported in both field and laboratory animals (Cobb et ah, 1995). These endogenous rhythms may be reflected in a NOTE Hernandez-Lopez et at: Age of Octopus vulgaris determined from concentric rings on beaks 681 Beak lateral wall length = -64. 166+ 1 9.522 In ( VML) Pearson's Correlation: r = 0.825; r2 - 0.6808 F = 580. 1 5; P < 0.001 ; n = 275 Figure 2 Relationship of beak lateral wall length to ventral mantle length (VML) of ben- thic Octopus vulgaris. chitinous structure such as the heaks (Raya and Hernan- dez-Gonzalez, 1998) or in calcium deposits in statoliths. Statoliths are the hard structures most commonly used for cephalopod age estimation (Lipinski, 1986, 1993; Arkh- ipkin, 1993), although the presence of concentric rings in the internal shell, beaks, and eye lenses have also been used (Clarke, 1965; Gonpalves, 1993; Raya and Hernan- dez-Gonzalez, 1998). When beaks are used, erosion of the rostral area during the life of the animal may bias age determination toward underestimation (some of the first rings may be eroded and therefore not counted). We found evidence of incomplete increments on the edge of the lat- eral wall, near the rostral tip area; therefore, ages we provide for benthic adults are to be considered minimum estimates. If rings on the lateral walls of the upper beaks are laid down daily and can be accurately counted even in the old- est specimens (as indicated by the pattern in paralarvae), then our results are consistent with a lifespan of 12-13 months in the Canary Island waters. Rava and Hernan- dez-Gonzalez (1998) gave a lifespan of 10-12 months for octopus caught off the coast of northwest Africa (21-26°N ) 682 Fishery Bulletin 99(4) although they reported some heavier but younger speci- mens than we found. This difference could be due to dis- crepancies in the aging methods or, as in the case of Man- gold ( 1983), areas off the coast may have different growth patterns and lifespans. Thus, Smale and Buchan (1981) proposed a lifespan of 9-12 months in females and 12-15 in males Octopus vulgaris from the South African coast. Several authors have noted that size (and probably weight) may not reliably indicate age in field-caught ceph- alopods (Mangold and Boletzky, 1973; Hixon, 1980) be- cause it may vary greatly depending upon factors such as food and temperature (Van Heukelem, 1979; Mangold, 1983). Cephalopods reveal great morphological variability with latitude attributed to environmental influences on development ( Hernandez-Garcia and Castro, 1998), and probably on lifespan. The length and weight ranges of oc- topus caught off the Canary Islands are within the ranges reported for this species off East Africa, are the limits of range (upper and lower) off South Africa (Smale and Bu- chan, 1981) and in the western Mediterranean Sea (Man- gold, 1983), although the range recorded in our study (Canary Islands) is closer to that reported for the Mediter- ranean Sea. The smallest octopi that we examined from fishery catches were 4.8 and 8.1 mm VML (0.38 and 0.60 g TW, re- spectively), well outside the minimum commercial length (90-100 mm VML). Their estimated ages were 51 and 91 days old. In the English Channel, the planktonic phase for common octopus has been estimated at 3 months (Rees, 1950; Rees and Lumby, 1954); and average weight of 0.2 g at settling may be normal regardless of temperature (Man- gold, 1983). Octopus typically spend the first 5-12 weeks of life as an active predator on plankton (Mangold, 1983); they change gradually from a planktonic to benthic life style (Boletzky, 1977) in some way dependent upon tem- perature (Mangold 1983). We did not observe marked dif- ferences in ring pattern spacing indicative of the transi- tion between planktonic and benthic life styles. However, the distance between rings does change during the ben- thic phase of life — a feature that seems related to water temperature — the rings being larger than average during winter and smaller during summer. NOTE Hemandez-Lopez et al.: Age of Octopus vulgaris determined from concentric rings on beaks 683 A Weight = -735 1.6 +1 607.7 In (no. of rings) Pearson’s Correlation: r = 0.8 1 7; r2 - 0.67 1 F = 547.09; P < 0.001 ; n = 275 0 50 100 150 200 250 300 350 400 450 B VML = -167.52 + 49.566 In (no. of rings) Pearson's Correlation: r = 0.775; r2 - 0.604 200 1 F = 409.61; P < 0.001; n = 275 oJ f *6 r- t t t r t t 0 50 100 150 200 250 300 350 400 450 Number of rings Figure 5 Plot of the number of rings counted on the lateral walls of the upper beak against (A) body weight and (B) ventral mantle length (VML) of benthic Octopus vulgaris. Acknowledgments We thank Ana Y. Martfn-Gutierrez for her assistance in data collection. This study was funded by the Secretariat for Fishing of the Autonomous Government of the Canary Islands (Spain). Literature cited Arkhipkin, A. 1993. Age, growth, stock structure and migratory rate of pre-spawning short-finned squid Illex argentinus based on statolith ageing investigations. Fish. Res. 16(4):313-338. Boletzky, S. v. 1977. Post-hatching behaviour and mode of life in cephalo- pods. Symp. Zool. Soc., Lond. 38:557-567. Clarke, M. R. 1965. “Growth rings” in the beaks of the squid Moroteutlus m,gens(Oegopsida: Onychoteuthidae). Malacologia3(2):287- 307. Cobb, C. S., S. K. Pope, and R. Williamson. 1995 . Circadian rhythms to 1 ight-dark cycles in the lesser octo- pus, Eledonecirrhosa. Mar. Freshwater Behav. Physiol. 26: 47-57. FAO (Food and Agriculture Organization of the LTnited Nations). 1998. Fishery statistics. Capture production 1996. FAO yearbook, vol. 82. FAO, Rome. Foucher, E., M. Thiam, and M. Barry. 1998. A GIS for the management of fisheries in West Africa: preliminary application to the octopus stock in Senegal. In Cephalopod biodiversity, ecology and evolution (A. I. L. Payne, M. R. Lipinski, M. R. Clarke, and M. A. C. Roeleveld, eds.), p. 337-346. S. Afr. J. Mar. Sci. 20. Gongalves, J. M. A. 1993. Octopus vulgaris Cuvier, 1797 (polvo-comun): sinopse da biologia e exploragao. Ms.C. thesis, Universidade dos Azores, Horta, Azores 470 p. 684 Fishery Bulletin 99(4) Guerra, A. 1979. Fitting a von Bertalanffy expression to Octopus vulga- ris growth. Inv. Pesq. 43:319-327. 1992. Mollusca cephalopoda. Fauna Iberica, vol. 1. Museo Nacional de Ciencias Naturales, Consejo Superior de Inves- tigaciones Cientfficas, Madrid, 327 p. Hatanaka, H. 1979. Studies on the fisheries biology of common octopus off northwest coast of Africa. Bull. Far Seas Fish. Res. Lab. 17: 13-124. Hernandez, V., and C. Bas. 1993. Analisis de la evolucion de las tallas de los cefalopo- dos explotados en la costa del Sahara (division 34.1.3 de CECAF) entre los periodos 1967-70 y 1989-90. Bol. Inst. Esp. Oceanogr. 9(l):215-225. Hernandez-Garcia, V., and J. J. Castro. 1998. Morphological variability in Illex coindetii (Cepha- lopoda: Ommastrephidae) along the North-west coast of Africa. J. Mar. Biol. Assoc. U.K. 78:1259-1268. Hernandez-Garcia, V., A. Y. Martin, and J. J. Castro-Hernandez. 2000. Evidences of external digestion of crustaceans in Oct- opus vulgaris (Cuvier, 1797) paralarvae. J. Mar. Biol. Assoc. U.K. 80:559-560 Hixon, R. F. 1980. Growth, reproductive biology, distribution and abun- dance of three species of loliginid squid (Myopsida, Cepha- lopoda) in the northwest Gulf of Mexico. Ph.D. diss., Univ. Miami, Coral Gables, FL, 233 p. Itami, K., Y. Izawa, S. Maeda, and K. Nakai. 1963. Notes on the laboratory culture of the Octopus larvae. Bull. Jap. Soc. Sci. Fish. 29:514-520. Lipinski, M. 1986. Methods for the validation of squid age from stato- liths. J. Mar. Biol. Assoc. U.K. 66(2):505-526. 1993. The deposition of statoliths: a working hypothesis. In Recent advances in cephalopod fisheries biology (T. Oku- tani, R. K. O’Dor, and T. Kubodera, eds.), p. 241-262. Tokai University Press, Tokyo. Mangold, K. 1983. Octopus vulgaris. In Cephalopod life cycles, vol. I, species accounts (P. R. Boyle, ed.), p. 335-364. Academic Press, London. 1998. The Octopodinae from the eastern Atlantic Ocean and Mediterranean Sea. In Systematic and biogeography of cephalopods, vol. II (N. A. Voss, M. Vecchione, R. B. Toll, and M. I. Sweeney, eds.), p. 521-528. Smithsonian Contribu- tions to Zoology 586. Mangold, K., and S. Boletzky. 1973. New data on the reproductive biology and growth of Octopus vulgaris. Mar. Biol. 19:7-12. Nixon, M. 1969. The lifespan of Octopus vulgaris Lamarck. Proc. malacol. Soc. Lond. 38:529-540. Pereiro, J. A., and J. Bravo de Laguna. 1979. Dinamica de la poblacion y evaluacion de los recur- sos del pulpo del Atlantico Centro-oriental. Bol. Inst. Esp. Oceanogr. 5:69-105. Raya, C. P., and C. L. Hernandez-Gonzalez. 1998. Growth lines within the beak microstructure of the Octopus vulgaris Cuvier, 1797. In Cephalopod biodiver- sity, ecology and evolution (A. I. L. Payne, M. R. Lipinski, M. R. Clarke, and M. A. C. Roeleveld, eds.), p. 135-142. S. Afr. J. Mar. Sci. 20. Rees, W. J. 1950. The distribution of Octopus vulgaris Lamarck in Brit- ish waters. J. Mar. Biol. Assoc. U.K. 29:361-378. Rees, W. J., and J. R. Lumby. 1954. The abundance of Octopus in the English Channel. J. Mar. Biol. Assoc. 33:515-536. Smale, M. J., and P. R. Buchan. 1981. Biology of Octopus vulgaris off the east coast of South Africa. Mar. Biol. 65(0:1-12. Van Heukelen, W. F. 1979. Environmental control of reproduction and life span in octopus: an hypothesis. In Reproductive ecology of marine invertebrates (S. E. Stancyk, ed.), p. 123-133. The Belle W. Baruch Library in Marine Science 9. Univ. South Carolina Press, Columbia, SC. Villanueva, R. 1995. Experimental rearing and growth of planktonic Octo- pus vulgaris from hatching to settlement. Can. J. Fish. Aquat. Sci. 52:2639-2650. Young, J. Z. 1960. The statocyst of Octopus vulgaris. Proc. R. Soc., Ser. B 152:3-29. 685 Reproduction of female spiny dogfish, Squalus acanthias, in the Oslofjord Thomas S. Jones Hans Oeverlandsvei 10 1363 Hoevik, Norway E-mail address: tomtrussel@hotmail com Karl I. Ugland Department of Biology University of Oslo, Pb 1064 0316 Oslo, Norway The spiny dogfish (Squalus acanthias) is a relatively small shark with a char- acteristic spine in front of each dorsal fin. Its dorsal side is grayish and has sporadic white spots. Although it may reach a length of 160 cm, most individu- als in the North Sea are in the range of 80-100 cm (Ford, 1921). It is dis- tributed worldwide, absent only from tropical and polar regions (Compagno, 1984). The spiny dogfish has been harvest- ed for more than 100 years mostly for its oil-rich liver (Ketchen, 1986). At first, the oil was used for lamp fuel and as a lubricant in machines. The oil was later (during W.W.II) used as a source of vitamin A. Today the dogfish is val- ued as food in many countries (Gordon, 1986). The reproduction cycle of the spiny dogfish begins with mature females bearing several large (over 40-mm) yel- low eggs. As the eggs pass through the shell gland they are fertilized and become enclosed in a protective cap- sule (candle). The candle passes down the reproductive tract and comes to rest in the uterus. The embryos live off the large yolk sac attached under the gill region (Gilbert, 1981). As the embryos grow, they slowly absorb the yolk sack. Embryos that have com- pletely absorbed the yolk sac may still remain in the uterus for some time be- fore being born (Ford, 1921). According to Jones and Geen (1977), the embry- os also bear an internal yolk sac which nourishes them for up to 2 months af- ter birth. The dogfish reproductive cy- cle takes almost 2 years, one of the longest gestation periods of any living vertebrate (up to 24 months) (Ketchen, 1972; Nammack et al., 1985). This shark, like other sharks, is very susceptible to overfishing, not only be- cause of its long gestation period, but also because of slow growth, late ma- turity, and because it bears a small number of offspring (up to 15) (Nam- mack et al., 1985; Fahy, 1989). Exten- sive fishing since the early 1960s has led to a marked decrease in the North Sea stock. Fishing has also affected the population in the Oslofjord where the annual catch declined from 704 tons in 1979 to less than 300 tons per year during the 1990s (Official Statistics of Norway, 1996). We investigated dogfish reproduction in the Oslofjord by com- paring the reproduction parameters of dogfish caught in 1987 and 1997. The focus of our study was to evaluate the reproductive parameters of spiny dog- fish in the Oslofjord and to look for possible changes in these parameters. Materials and methods Sampling Dogfish were sampled monthly off the Hvaler Islands throughout 1987 and 1997 in gill nets and by longline at depths ranging from 50 to 460 m (Fig. 1 ). The gill nets were composed of mono- filament line (0.60 mm) with a mesh size of 285 mm. This large mesh size accounts for catches consisting mainly of larger dogfish (over 70 cm). The long- line was composed of a 5-mm line con- nected to a 7/0 dogfish hook. The nets were usually checked every 24 hours while the longline was taken up after a few hours. The samples consisted of 132 females in 1987 and 101 females in 1997. Total length and weight were measured to the nearest 0.5 cm and 5 g, respectively, as described by Saun- ders and McFarlane (1993). The same fisherman, fishing grounds and fishing gear were used in both sampling years, so that sampling bias was avoided. The fishing gear account- ed for catches of dogfish that were mainly over 70 cm in length. There was a relatively small amount of un- marketable-size fish caught and the discard rates were approximately the same in both years. Age determination The first and second dorsal spines were removed from 217 dogfish. The remaining 16 dogfish had spines that were either missing, or broken to such an extent that age determination was not possible for these individuals. The spines were air-dried at least 1 week before being cooked for approximately 3 minutes each in tap water. The flesh around the base of the spine was then removed with a scalpel and tweezers. The cleaned spines were subsequently dipped in ethyl alcohol and then dried with a soft cloth. Next, the spines were viewed under an Olympus SZH 10 zoom stereo microscope and aged according to the method described by Ketchen (1975). For large individuals with worn spines, age was modified by Ketchen’s (1975) correction curve for age Y= 0.50 9 7X2 55, where X = the diameter of the spine base in millimeters; and Y = the additional age of the spine due to it being worn. Ketchen’s correction curve is based on fish caught in the Strait of Geor- gia, British Colombia. It is quite likely that there are differences in juvenile growth in the respective dogfish popu- lations. The result may therefore be bi- Manuscript accepted 15 March 2001. Fish. Bull. 99:685-690 (2001) 686 Fishery Bulletin 99(4) Figure 1 Map of the Oslofjord. The gray shaded area represents the capture location of dogfish in the outer Osloljord. ased and in future work a correction curve will have to be made for the stock in the Oslofjord. Embryo development Both ovaries of mature females were rinsed to expose eggs. The rinsing process consisted of removing and then carefully opening the ovaries with a scalpel. The eggs were counted and registered as belonging to the right or left ovary. In pregnant females the embryos were counted and sexed. The embryos were then measured to the nearest 0.1 cm on mil- limeter paper. Three stages of embryo development were identified according to Gauld1: stage 1 (candled embryos) — eggs, apparently fertilized, are present in a protective cap- sule in the uteri; stage 2 (free-living embryos) — the candle has ruptured and embryos bearing an external yolk sac are free in the uteri; and stage 3 (full term embryos) — fully developed embryos are present in the uteri. The yolk sac is fully absorbed and the umbilical slit is more or less closed. Results Biological parameters The fish ranged in length from 54 to 1 10 cm in 1987 (mean 87 cm) and from 68 to 108 cm (mean 88 cm) in 1997. 1 Gauld, J. 1979. Reproduction and fecundity of the Scottish- Norwegian stock of spurdogs, Squalus acanthias (L. ). ICES, Council Meeting (CM) 1979/H:54. 13 p. Directory of fisheries, biblioteket, Pb. 185, 5001 Bergen. Weights ranged from 0.4 to 5.7 kg (mean 2.6 kg) in 1987 and from 1.1 to 5.5 kg (mean 2.9 kg) in 1997. Ages ranged from 9 to 35 years (mean 25 years) in 1987 and from 10 to 38 years (mean 23 years) in 1997. On average, the fish in 1987 were 1.8 cm longer than fish in 1997 at the same age (P=0.02). However, there was a large variation in the length, and age could only account for 38% of this variability. Length accounted for 76% of the variability of the weight in 1987 and 93% in 1997. Fish between 80 and 110 cm were about 500 g heavier in 1997 (P<0.01 ). This difference was reflected in the condition fac- tor, which was 3.8 in 1987 compared with 4.2 in 1997. This indicates that female dogfish caught in 1987 were longer and lighter than in 1997. Growth Growth appeared to be best represented by a linear growth equation for dogfish between the ages of 11 and 38 years, with an average annual increment of 0.7 cm per year (Fig. 2). There was no sign of reduction in growth with increasing age; that is, no asymptote was detected in these data. Age and length at maturity Maturity was defined as females bearing large ovarian eggs with a diameter of over 2 cm, candled young or free- living embryos. In both 1987 and 1997, most of the females matured between 12 and 26 years of age. PROBIT analy- sis indicated that dogfish reached 50% maturity at an age of 17.6 years in 1987 and 17.0 years in 1997 (Fig. 3). The NOTE Jones and Ugland: Reproduction of female Squcilus acanthus 687 11 ■ 1987, n=1 10 0 1997, n=79 11 14 17 20 23 26 29 Mean value of age groups (years) Figure 3 Percentage of mature fish in the different age groups 10-12, 13-15, 16-18,19-21, 22-24, 25-27 and 28-30 years in 1987 and 1997. 1997 sample also had a larger proportion of mature fish less than 15 years of age. No individual was mature under 76 cm, but all individu- als over 88 cm were mature, indicating that maturity occurs between 76 and 88 cm. The length at 50% maturity was 81 cm in both 1987 and 1997. There may be a difference in the growth rate between the two years; fish from 1997 reached maturity earlier than fish from 1987. However, because of the small sample size and difficulty with age determina- tion, caution should be used to interpret this result. Fecundity In 1987 the number of ovarian eggs with a diameter over 2 cm (indicating mature fish) varied between 5 and 14 (mean=8.2, SD=2.2). In 1997 the range was 3-17 eggs (mean=8.9, SD=3). The relationship between egg number and adult length in the 1987 and 1997 samples was not significantly different. On average, the egg production increased by one egg per 4-cm adult length. The number of free-living embryos in 1987 varied from 2 to 14 per female (mean=6.6, SD=2.7). In 1997, free- living embryos numbered between 3 and 15 (mean=7.5, SD=2.6). The increment of free-living embryos per adult length (Fig. 4) in the 1987 and 1997 sample was not sig- nificantly different. However, the level of the regression line was significantly higher in 1987 (P=0.04). On average, the females in 1987 carried 1.2 more free-living embryos. It should be emphasized that owing to the large variabil- ity, length could only explain 45% to 56% of the difference in the number of free-living embryos. Reproduction cycle Eggs on which the blastoderm was visible, were recog- nized as recently fertilized. This was part of the candle 688 Fishery Bulletin 99(4) stage (from the newly fertilized eggs up to embryos 10 cm long) that lasted from October to November the fol- lowing year. Fertilization occurred from the beginning of October to the beginning of February. Free-living embryos were observed from October to September of the follow- ing year, followed by parturation of fully mature embryos from September to December. The duration of the preg- nancy for spiny dogfish in the Oslofjord ranged from 18 to 24 months. This period of time is calculated from the time of fertilization to the time of parturation. These estimates of the reproduction phases are summarized in Figure 5. fish embryos was followed throughout their second year of development. Growth was approximately 1 cm per month for both year classes in the size range from 8 to 24 cm, which corresponds approximately to the embryos in the second year of development. However, growth was not uni- form; it could be divided into two growth phases (Fig. 6). The first phase lasted from October to May with a slow growth of 0.6 cm/month. The second phase lasted from May until December and was characterized by a more rapid growth of 1.2 cm/month. Embryonic growth Discussion The embryos ranged from 4.4 to 24.9 cm total length. The growth of the combined 1986 and 1996 year classes of dog- Free-living embryos Candled embryos Fertilization Parturation i l t i — l i t i i i i i i — i l l t t t- i i i i — i t l i i SONDJ FMAMJJ ASONDJ FMAMJ JAS Month Figure 5 Schematic time representation of the reproductive cycle of the female spiny dogfish by month (n=194). Growth The lack of an asymptote is probably due to an absence of older, larger fish. This is most likely due to over-harvesting of dogfish in the Oslofjord. According to local fishermen, heavy catches of large individuals (exceeding 110 centime- ters) were taken in the late 70s and early 80s.2 Age and length at maturity Age at 50% maturity was found to be 17.6 years in 1987 and 17.0 years in 1997. For spiny dogfish captured between the East Coast of Scotland and Norway, Aasen (1961) estimated age at 50% maturity to be in the range 12 to 14 years. In an area northwest of Scotland, Holden and Meadows (1962) reported age at 50% maturity of 9 years. However, Aasen (1961) and Holden and Meadows (1962) did not take into account worn spines, which com- bined with the difficulty of identifying all of the annual rings, means that their estimates are most likely biased downwards. Length at 50% maturity in the Oslofjord should not de- viate greatly from sizes observed elsewhere in the north- east Atlantic. Length at 50% maturity in the Oslofjord was found to be 81 cm in both 1987 and 1997. This is in general agreement with other observations in the northeast Atlantic where length at 50% maturity varied from 74 to 83 cm (Hickling, 1930; Holden and Meadows, 1964; Fahy, 1989; Gauld1). Fecundity There was no significant difference between the number of eggs observed in 1987 and 1997. How- ever, the number of eggs in the ovaries was higher than the number of embryos in the uterus (average: 8.1 eggs and 6.6 free-living embryos in 1987 and 8.9 eggs and 7.5 free-living embryos in 1997). This is probably due to the absorption of eggs in the ovaries following fertilization that was observed by Hanchet (1988), Holden and Meadows (1964), and Nammack et al. ( 1985). 2 Johansen, P. A. 1997. Personal commun. Dogfish fish- erman from the Oslofjord. Gudeberg Alle 1, 1600 Fredrik- stad, Norway. NOTE Jones and Ugland: Reproduction of female Squci/us acanthus 689 25 20 1 1 * 15 * l l * f ♦ m l 10 5 0 ^ — ' — ' — 1 — ' — 1 — ' — ’ — ' — ' — > — ' — ' — ' — ’ — ' — >— ASONDJ FMAMJ JASOND Month Figure 6 Length of the embryos during their second year of devel- opment for the combined 1987 and 1997 sample. Fecundity was within the bounds found elsewhere in the northeast Atlantic (Aasen, 1961; Gauld1). Free-living em- bryos showed the same increase in number per length in 1987 and 1997, but the level was significantly lower in 1997 (on average one less embryo per unit of length). In addition the fish were shorter and heavier in 1997. How- ever, these results must be taken with some caution be- cause of the large variability in the data. A small sample size and annual variability in both em- bryo production and food availability may have contrib- uted to the lower fecundity in 1997. With regard to the sample sizes (31 and 38 fishes), it is emphasized that the covariation between the number of free-living embryos versus length was rather low. Although the differences be- tween 1987 and 1997 were statistically significant, the overlap was so large that caution must be used in inter- pretation of the data. Reproduction cycle and embryonic growth In the Oslofjord, fertilization occurs from October to Febru- ary and parturation from October to December. Research in other areas has shown that the pregnancy lasts 22-23 months (Ford, 1921; Gauld1). Thus, the duration of preg- nancy in the Oslofjord (18 to 24 months) seems to have a greater variation than these observations. This variation may have been caused by the sample from the Oslofjord, which included dogfish that were fertilized early in the fertilization period, as well as dogfish that were fertilized towards the end of the fertilization period. This, in turn, indicates the possible minimal and maximal duration of pregnancy. Behavioral factors may influence pregnancy duration, as well. Some dogfish may inhabit cooler water for longer periods of time than other dogfish, which may increase the duration of their pregnancy. The differences in the embryo growth rate during their second year of development may be related to the water temperature. Hickling (1930) found that females migrate from deep to shallow water as pregnancy proceeds. This migration pattern exposes the embryos to different tem- perature levels that may influence embryonic growth, as found in Newfoundland by Templeman (1944). Embryos outside of Newfoundland had an average growth rate of 1. 1 cm/month and a 24-month pregnancy period (Templeman, 1944). Outside of Woods Hole, Hisaw and Albert (1947) found an average growth rate of 1.3 cm/month and a preg- nancy period of 20-22 months. Nammack et al. ( 1985) stat- ed that the lower sea temperature outside of Newfound- land could be a possible reason for this difference. In phase 1, growth is slow and corresponds to a period of low water temperature (winter and spring). As females begin to migrate towards shallower water, they encounter warmer water (summer and fall). The increased growth rate in phase 2 is therefore most likely related to an in- crease in water temperature. Acknowledgments We would like to thank P. A. Johansen for making the col- lection of the sharks possible, as well as giving us valu- able suggestions. We thank scientists and personnel of the biological station in Drpbak and H. E. Karlsen for the use of equipment associated with age determination. We are also particularly grateful to T. R. Wiley and H. Zidowitz for their comments on the manuscript. Literature cited Aasen, O. 1961 Piggha-undersokelsene. Fisken og havet 1:1-9. Compagno, L. J. V. 1984. FAO species catalogue. Vol. 4: Sharks of the world: an annoted and illustrated catalogue of shark species known to date. Part 1. Hexiformes to Lamniformes. FAO Fish. Synop. 125:1-249. Fahy, E. 1989. The spurdog (Squalus acanthias) fishery in South West Ireland. Ir. Fish. Invest, ser. B. no. 32, 22 p. Ford, E. 1921. A contribution to our knowledge of the life-histories of the dogfishes landed at Plymouth. J. Mar. Biol. Assoc. U.K. 12:468-505. Gilbert, P. W. 1981. Patterns of shark reproduction. Oceanus 24:30-39. Gordon, D. G. 1986. The spiny dogfish. Oceans 19:56-59. Hanchet, S. 1988. Reproductive biology of Squalus acanthias from the east coast. South Island, New Zealand. N.Z.. J. Mar. Fresh- water Res. 22:537-549. Hickling, C. F. 1930. A contribution towards the life history of the spurdog. J. Mar. Biol. Assoc. U.K. 16:529-576. Hisaw, F. L., and A. Albert. 1947. Observations on the reproduction of spiny dogfish, Squalus acanthias. Biol. Bull. 92:187-199. Holden, M. J., and P. S. Meadows. 1962. The structure of the spine of the spur dogfish ( Squa- lus acanthias ) and its use for age determination. J. Mar. Biol. Assoc. U.K. 42:179-197. 690 Fishery Bulletin 99(4) 1964. The fecundity of the spurdog ( Squalus acanthias). J. Cons. Perm. Int. Explor. Mer 28:418-424. Jones, B. C., and G. H. Geen. 1977. Reproduction and embryonic development of spiny dogfish (Squalus acanthias ) in the Strait of Georgia, Brit- ish Colombia. J. Fish. Res. Board Can. 34:1286-1292. Ketchen, K. S. 1972. Size at maturity, fecundity, and embryonic growth of the spiny dogfish ( Squalu sacanthias ) in British Colombia Waters. J. Fish. Res. Board Can. 29:1717-1723. 1975. Age and growth of dogfish, Squalus acanthias, in Brit- ish Colombia Waters. J. Fish. Res. Board Can. 32: 43-59. 1986. The spiny dogfish ( Squalus acanthias ) in the North- east Pacific and a history of its utilization. Can. Spec. Publ. Fish. Aquat. Sci. 88, 78 p. Nammack, M. F., J. A. Musick, and J. A. Colvocoresses. 1985. Life history of spiny dogfish off the Northeastern United States. Trans. Am. Fish. Soc. 114:367-376. Official Statistics of Norway. 1996. Fishery statistics 1961-1996. Graphic Communica- tion Systems, Inc. (GCS) A/S, Oslo, Norway, 116 p. Saunders, M. W., and G. A. McFarlane. 1993. Age and length at maturity of the female spiny dog- fish, Squalus acanthias, in the Strait of Georgia, British Colombia, Canada. Environ. Biol. Fish. 38:49-57. Templeman, W. 1944. The life-history of the spiny dogfish (Squalus acanth- ias) and the vitamin A values of dogfish liver oil. Nfld. Gov., Dep. Nat. Resour., Fish. Res. Bull. 15:1-100. 691 Estimating live standard length of net-caught walleye pollock ( Theragra chalcogramma) larvae using measurements in addition to standard length* Steven M. Porter Annette L. Brown Kevin M. Bailey Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 7600 Sand Point Way NE Seattle, Washington 98115 E-mail address (for S M Porter): steve.porter@noaa.gov Accurate measurement of larval fish lengths from ichthyoplankton samples is critical in the estimation of growth, birthdate and mortality. It is well known that larvae shrink in length 1) when caught in plankton nets (Thei- lacker, 1980; Hay, 1981), 2) between the time of collection and preserva- tion, and 3) afterwards from the effects of preservation in chemicals or from freezing (Theilacker, 1980; Fowler and Smith, 1983; Yin and Blaxter, 1986). Shrinkage caused by the plankton net (the net can damage a larva’s integu- ment; Holliday and Blaxter, 1960) and by delays in preservation can be up to 40% of the initial live length (Hay, 1981). In some cases, measurements of larval body length may be unsuitable for growth estimation (Jennings, 1991 ). In other cases, algorithms have been developed to estimate live lengths from preserved larvae (Bailey 1982, Hjorlei- fsson and Klein-MacPhee, 1992; Fey, 1999) and from net-caught and pre- served larvae (Theilacker, 1980; Thei- lacker and Porter, 1995; Fox, 1996). It has also been noted that constant shrinkage-correction factors should be applied cautiously (Pepin et al., 1998). Shrinkage of the standard length of fish larvae is species-specific ( Jennings, 1991; Fey, 1999), size-dependent (Fowl- er and Smith, 1983; Theilacker and Porter, 1995; Fey, 1999), solution-de- pendent (Hay, 1982; Tucker and Ches- ter, 1984), and may also depend to some degree on the ambient tempera- ture at the time of preservation (Hay, 1982). Other body measurements may shrink due to effects of collection and preservation. In fact, many corrections for larval shrinkage have been devel- oped with respect to estimating mor- phometric-dependent condition factors (Theilacker, 1980; McGurk, 1985; Yin and Blaxter, 1986). Measurement of otoliths has been proposed as a method to correct for shrinkage in the length of fish larvae (Leak, 1986; Radtke, 1989), but this method may not be generally appli- cable because of systematic variations in the relationship between fish size and otolith size (Neilson and Campa- na, 1990). In addition, this relationship can be nonlinear and unpredictable. Another approach to obtain shrinkage correction factors is by controlled ex- periments in which larval shrinkage is recorded from the time of death, or for the period of time larvae are in a plankton net, and again after preserva- tion. The problem with this approach is that during most sampling at sea, the time larvae enter the net to the time of their preservation is unknown and can vary from several minutes to a half hour or more, depending on their depth of capture and the duration of the tow. In our study, we used measurements of other body parts as ancillary data to estimate live lengths of preserved fish larvae caught at sea. We report on two common preservatives: ethanol and formalin. To avoid the problem of duration of time in the net, shrink- age correction equations were formu- lated by pooling data of known-length larvae that were treated for varying durations in simulated plankton tows in the laboratory and then preserved. Morphometric measurements were col- lected on preserved larvae that could be used in equations to obtain more accurate and precise estimates of live length. Rather than use alternative body part measurements as a substi- tute for length, we hypothesized that other body parts, nonshrinking or oth- erwise, could be used to correct for shrinkage, in spite of how long larvae had been in nets. Materials and methods Rearing protocol Spawning walleye pollock ( Theragra chalcogramma ) were collected by trawl in the eastern Bering Sea by the NOAA ship Miller Freeman during April 1999. Fertilized eggs were transported to the Alaska Fisheries Science Center, Seattle, Washington, where they were incubated at 6°C in the dark in 4-L glass jars filled with 3 L of filtered sea- water at a salinity of 33%0.05). Although the growth rates were not significantly different, growth rates calculated from the multivariate models were more similar to growth rates calculated from live SL (Table 2); therefore additional mor- phometric variables improve growth rate estimates over using preserved SL alone. Discussion Fish larvae shrink considerably when they are caught in nets and preserved. Shrinkage during collection is most likely caused by the plankton net, which can damage the larva’s integument (Holliday and Blaxter, 1960). Shrinkage also varies between preservatives because of the differing ionic strengths of these solutions (Parker, 1963; Hay, 1982; Tucker and Chester, 1984). We found that shrinkage of walleye pollock larvae was greater in 5% formalin than in 95%< ethanol — a finding similar to results from other spe- cies (Bailey, 1982; Fey, 1999). Often only preserved SL has been used to correct for shrinkage in length (Theilacker and Porter, 1995; Fox, 1996; Kristoffersen and Gro Vea Sal- vanes, 1998), although other measurements, such as oto- lith size, have been suggested for use to correct for larval fish shrinkage (Leak, 1986; Radtke, 1989,1990). By using measurements that are easily made and readily adapted to 694 Fishery Bulletin 99(4) Figure 1 Uncorrected (A) and corrected (B) standard length determined by using best shrinkage correction model, see Table 1) for laboratory-reared walleye pol- lock ( Theragra chalcogramma) larvae subjected to various shrinkage treat- ments and preserved in 95% ethanol (n=91). The 95% confidence limits are shown for each predicted value, and the diagonal line shows a 1:1 ratio. routine data collection, we have shown improved accuracy of shrinkage correction models by use of measurements in addition to preserved SL: for 5% formalin, 7% more of the variance was explained by adding body depth; and for 95% ethanol, 8% more of the variance was explained by includ- ing body depth and otolith diameter. For 95% ethanol, the shrinkage correction model formulated in our study may significantly underestimate SL when shrinkage is high, but because there were only two larvae in this category, results were inconclusive. Shrinkage correction models are usually formulated by using fed, laboratory-reared larvae (Jennings, 1991; Thei- lacker and Porter, 1995; Fox, 1996), but individual larvae response to handling and preservation can vary signifi- cantly (Pepin et ah, 1998). These correction models were developed to allow more accurate calculation of the growth rate of fish larvae in the sea and to apply laboratory- derived indices to the field. We have shown factors in ad- dition to preserved SL that may improve the accuracy of live-length estimates for walleye pollock. NOTE Porter et al.: Estimating live standard length of net-caught Theragra chalcogrcimma 695 E E^ JZ O) c Q) "D CD "O C E £ O) c 0) ■D CD ■D C CD O O Figure 2 Uncorrected (A) and corrected (B) standard length determined by using best shrinkage correction model, see Table 1) for laboratory-reared walleye pol- lock ( Theragra chalcogramma ) larvae subjected to various shrinkage treat- ments and preserved in 5% formalin (rc=90). The 95% confidence limits are shown for each predicted value, and the diagonal line shows a 1:1 ratio. Acknowledgments We thank Debbie Blood for care of the pollock eggs aboard ship, and for bringing them back to Seattle. Kathy Mier, Susan Picquelle, and three anonymous reviewers com- mented on drafts of the manuscript and offered improve- ments. Literature cited Bailey, K. M. 1982. The early life history of Pacific hake, Merluccius pro- ductus. Fish. Bull. 80:589-598. Fey, D. P. 1999. Effects of preservation technique on the length of larval fish: methods of correcting estimates and their impli- cation for studying growth rates. Arch. Fish. Mar. Res. 47: 17-29. 696 Fishery Bulletin 99(4) Fowler, G. M., and S. J. Smith. 1983. Length changes in silver hake (Merluccius bilinearis ) larvae: effects of formalin, ethanol, and freezing. Can. J. Fish. Aquat. Sci. 40:866-870. Fox, C. J. 1996. Length changes in herring ( Clupea harengus) larvae: effects of capture and storage in formaldehyde and alcohol. J. Plankton Res. 18:483-493. 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 178:395-400. 1982. Fixation shrinkage of herring larvae: effects of salin- ity, formalin concentration, and other factors. Can. J. Fish. Aquat. Sci. 39:1138-1143. Hjorleifsson E., and G. Klein-MacPhee. 1992. Estimation of live standard length of winter flounder Pleuronectes americanus larvae from formalin-preserved, ethanol-preserved and frozen specimens. Mar. Ecol. Prog. Ser. 82:13-19. Holliday, F. G. T., and J. H. S. Blaxter. 1960. The effects of salinity on the developing eggs and larvae of the herring. J. Mar. Biol. Assoc. U.K. 39:591- 603. Jennings, S. 1991. The effects of capture, net retention and preservation upon lengths of larval and juvenile bass, Dicentrarchus labrax (L. ). J. Fish Biol. 38:349-357. Kristoffersen, J. B., and A. Gro Vea Salvanes. 1998. Effects of formaldehyde and ethanol preservation on body and otoliths of Maurolicus muelleri and Benthosema glaciale. Sarsia 83:95-102. Leak, J. C. 1986. The relationship of standard length and otolith dia- meter in larval bay anchovy, Anchoa mitchilli (Val.). A shrinkage estimator. J. Exp. Mar. Biol. Ecol. 95:167-172. McGurk, M. D. 1985. Effects of net capture on the postpreservation mor- phometry, dry weight, and condition factor of Pacific her- ring larvae. Trans. Am. Fish. Soc. 114:348-355. Neilson, J. D., and S. E. Campana. 1990. Comment on “Larval fish age, growth, and body shrinkage: information available from otoliths.” Can. J. Fish. Aquat. Sci. 47: 2461-2463. Parker, R. R. 1963. Effects of formalin on the length and weight of fishes. J. Fish. Res. Board Can. 20:1441-1455. Pepin, R, J. F. Dower, and W. C. Leggett 1998. Changes in the probability density function of larval fish body length following preservation. Fish. Bull. 96:633-640. Radtke, R. L. 1989. Larval fish age, growth, and body shrinkage: infor- mation available from otoliths. Can. J. Fish. Aquat. Sci. 46:1884-1894. 1990. Information storage capacity of otoliths: response to Nielson and Campana. Can. J. Fish. Aquat. Sci. 47:2463-2467. Rosenthal, H., D. Kuhlmann, and O. Fukuhara 1978. Shrinkage of newly hatched larvae of the red sea bream ( Chrysophrys major Temminck and Schlegel) pre- served in formalin. Arch. Fischwiss. 29:59-63. SPSS, Inc. 1998. SYSTAT 8.0: statistics. SPSS, Inc., Chicago, IL, 1086 p. Theilacker, G. H. 1980. Changes in body measurements of larval northern anchovy, Engraulis mordax , and other fishes due to han- dling and preservation Fish. Bull. 78:685-692. Theilacker, G. H., and S. M. Porter. 1995. Condition of larval walleye pollock, Theragra chalco- gramma , in the western Gulf of Alaska assessed with his- tological and shrinkage indices. Fish. Bull. 93:333-344. Tucker, J. W., Jr., and A. J. Chester. 1984. Effects of salinity, formalin concentration and buffer on quality of preservation of southern flounder (Paralich- thys letliostigrna) larvae. Copeia 1984:981-988. Yamashita, Y., and K. M. Bailey. 1989. A laboratory study of the bioenergetics of larval wall- eye pollock, Theragra chalcogramma. Fish. Bull. 87:525- 536. Yin, M. C., and J. H. S. Blaxter. 1986. Morphological changes during growth and starvation of larval cod ( Gadus morhua L.) and flounder ( Platichthys flesus L.). J. Exp. Mar. Biol. Ecol. 104:215-228. Zar, J. H. 1996. Biostatistical analysis, third ed. Prentice Hall, Upper Saddle River, NJ, 622 p. 697 Preliminary study of albacore ( Thunnus alalunga ) stock differentiation inferred from microsatellite DNA analysis Motohiro Takagi Tetsuro Okamura Department of Cultural Fisheries Faculty of Agriculture Kochi University, Monobe Nankoku, Kochi 783-8502, Japan Seinen Chow National Research Institute of Far Seas Fisheries 5-7-1 Orido Shimizu 424-8633, Japan Nobuhiko Taniguchi Department of Cultural Fisheries Faculty of Agriculture Kochi University, Monobe Nankoku, Kochi 783-8502, Japan Present address (for N Taniguchi, contact author): Graduate school of Agriculture Tohoku University 1-1 Tsutsumidori, Amamiyamachi 1-2 Sendai 981-8555, Japan E-mail address (for N. Taniguchi, contact author): nobuhiko@bios.tohoku.ac.|p The albacore (Thunnus alalunga ) is a highly migratory large pelagic tuna, common from tropical to temperate areas of all oceans, including the Mediterranean Sea. Although alba- core populations of each ocean or each hemisphere have been managed sepa- rately, relationships between albacore populations of northern and southern hemispheres within an ocean are con- troversial. Differences in morphome- try, movement, and catch statistics of albacore between northern and south- ern hemispheres within Atlantic and Pacific Oceans have been reported (Kurogane and Hiyama, 1958; Ishii, 1965; Nakamura, 1969). Examining mtDNA variation, Chow and Ushiama (1995) detected very little genetic dif- ference between samples from north- ern and southern hemispheres. They proposed that minor gene flow may have occurred between albacore popu- lations of northern and southern hemi- spheres— enough to prevent genetic differentiation. However, it is also possible that insufficient time has elapsed since population subdivision for mtDNA genotypic rearrangement. Because all tuna species of the genus Thunnus are thought to be phylogenet- ically new (Chow and Kishino, 1995), use of highly variable genetic markers may be necessary to investigate genetic differentiation between stocks. Recently, Takagi et al. ( 1999) isolated four microsatellite loci from Pacific northern bluefin tuna (Thunnus thyn- nus oriental is) and demonstrated suc- cessful cross-species amplification of ho- mologous microsatellites in other tuna species. In our study, we used these mi- crosatellite primers to evaluate genetic variation within and between albacore samples from the Atlantic and the Pacif- ic Oceans and present genetic evidence of population structuring between and within ocean samples of albacore. Materials and methods Five albacore samples were drawn from archival stock materials in the National Research Institute of Far Seas Fisheries laboratory. One sample each came from the Northwest Pacific (NW Pacific; Japan), Southwest Pacific (SW Pacific; Australia), Southeast Pacific (SE Pacific; Chile and Peru), and two came from the Northeast Atlantic (NE Atlantic; Biscay Bay) and the Southwest Atlantic (SW Atlantic; Brazil). These samples were derived from the same sample lots used by Chow and Ushiama (1995). Nucleo- tide sequences of the four primer sets, PCR amplification conditions needed to amplify the four microsatellite loci ( Ttho-1 *, -4*, -6* and -7*) devel- oped for Pacific northern bluefin tuna (T. thynnus orientalist, and electro- phoresis procedures can be found in Takagi et al. (1999). Differentiation of allele frequencies between and among samples was estimated by a fixation index (FST) with Arlequin version 1.1 (Schneider et al., 1997). Results Alleles observed in each locus were as follows: 9 in Ttho-1 *, 29 in Ttho-4 *, 31 in Ttho-6*, and 18 in Ttho-7* (Table 1). All four sets of PCR primers success- fully amplified scorable microsatellite loci for all samples. Observed hetero- zygosity (H0) ranged from 0.391 to 0.914 at Ttho-1* , from 0.688 to 0.886 at Ttho-4 . from 0.548 to 0.884 at Ttho-6 and from 0.857 to 1 at 77/m- 7 . We found no substantial discrepancy between observed and expected number of genotypes for any locus. All samples except NE Atlantic shared the most common alleles for all loci, whereas the NE Atlantic sample shared the most common alleles only within Ttho-1 ' . The NE Atlantic sam- ple also did not share the second most common allele with the other samples for all loci. FST among all five Manuscript accepted 11 April 2001. Fish. Bull. 99:697-701 (2001). 698 Fishery Bulletin 99(4) Table 1 Allele frequency and genetic variability for four microsatelite loci surveyed for albacore used in this study. Locus Allele (base pairs) NW Pacific SW Pacific SE Pacific SW Atlantic NE Atlantic Ttho-1* 172 0.000 0.000 0.021 0.000 0.000 174 0.000 0.016 0.000 0.000 0.000 176 0.075 0.065 0.043 0.054 0.065 178 0.000 0.097 0.096 0.000 0.032 180 0.234 0.177 0.202 0.141 0.145 182 0.521 0.452 0.436 0.685 0.516 184 0.149 0.161 0.160 0.098 0.210 186 0.000 0.032 0.043 0.022 0.032 188 0.021 0.000 0.000 0.000 0.000 No. of samples 47 31 47 46 31 No. of allele 5 7 7 5 6 Effective no. of alleles2 2.82 3.62 3.70 1.99 2.96 Observed heterozygosity ( H0 ) 0.659 0.871 0.914 0.391 0.645 Expected heterozygosity (He) 0.645 0.724 0.730 0.498 0.662 Ttho-4* 134 0.000 0.000 0.000 0.000 0.032 136 0.000 0.016 0.011 0.000 0.096 138 0.011 0.016 0.033 0.011 0.065 140 0.043 0.031 0.067 0.011 0.452 142 0.032 0.031 0.167 0.000 0.032 144 0.085 0.094 0.100 0.102 0.065 146 0.383 0.313 0.178 0.261 0.048 148 0.032 0.078 0.089 0.080 0.097 150 0.011 0.094 0.056 0.023 0.032 152 0.053 0.047 0.067 0.102 0.048 154 0.096 0.031 0.044 0.046 0.032 156 0.011 0.000 0.011 0.068 0.000 158 0.032 0.047 0.000 0.057 0.000 160 0.043 0.047 0.044 0.046 0.000 162 0.021 0.031 0.033 0.034 0.000 164 0.011 0.078 0.044 0.023 0.000 166 0.053 0.000 0.000 0.023 0.000 168 0.021 0.031 0.011 0.011 0.000 170 0.000 0.000 0.022 0.011 0.000 174 0.000 0.000 0.011 0.023 0.000 176 0.000 0.000 0.000 0.023 0.000 178 0.000 0.000 0.011 0.023 0.000 180 0.032 0.000 0.000 0.000 0.000 182 0.011 0.000 0.000 0.000 0.000 184 0.011 0.000 0.000 0.000 0.000 186 0.000 0.016 0.000 0.000 0.000 190 0.011 0.000 0.000 0.000 0.000 202 0.000 0.000 0.000 0.011 0.000 232 0.000 0.000 0.000 0.011 0.000 No. of samples 47 32 45 44 31 No. of allele 20 16 18 21 11 Effective no. of alleles2 5.62 7.14 10.20 8.85 4.17 Observed heterozygosity (Hti) 0.851 0.688 0.844 0.886 0.710 Expected heterozygosity (He) 0.822 0.860 0.902 0.887 0.760 Ttho-6* 127 0.000 0.000 0.000 0.012 0.000 129 0.000 0.000 0.000 0.012 0.000 131 0.011 0.032 0.000 0.000 0.000 133 0.011 0.000 0.000 0.000 0.000 135 0.033 0.113 0.047 0.093 0.000 continued NOTE Takagi et al.: Stock differentiation of Thunnus alalunga through microsatellite DNA analysis 699 Table 1 (continued) Locus Allele base pairs) NW Pacific SW Pacific SE Pacific SW Atlantic NE Atlantic Ttho-6* (continued) 137 0.100 0.000 0.023 0.105 0.000 139 0.056 0.016 0.093 0.047 0.000 141 0.478 0.629 0.233 0.349 0.083 143 0.067 0.048 0.081 0.105 0.017 145 0.167 0.113 0.163 0.128 0.050 147 0.067 0.048 0.093 0.058 0.300 149 0.000 0.000 0.081 0.047 0.083 151 0.011 0.000 0.023 0.012 0.017 153 0.000 0.000 0.012 0.000 0.033 155 0.000 0.000 0.035 0.000 0.050 157 0.000 0.000 0.000 0.012 0.033 159 0.000 0.000 0.000 0.023 0.050 161 0.000 0.000 0.012 0.000 0.083 163 0.000 0.000 0.023 0.000 0.000 165 0.000 0.000 0.000 0.000 0.033 167 0.000 0.000 0.023 0.000 0.050 169 0.000 0.000 0.000 0.000 0.033 171 0.000 0.000 0.012 0.000 0.017 177 0.000 0.000 0.012 0.000 0.000 183 0.000 0.000 0.000 0.000 0.017 185 0.000 0.000 0.000 0.000 0.017 187 0.000 0.000 0.012 0.000 0.000 189 0.000 0.000 0.012 0.000 0.000 191 0.000 0.000 0.012 0.000 0.000 199 0.000 0.000 0.000 0.000 0.017 201 0.000 0.000 0.000 0.000 0.017 No. of samples 45 31 43 43 30 No. of allele 10 7 19 13 19 Effective no. of alleles7 3.58 2.34 8.47 5.65 7.87 Observed heterozygosity (Ha) 0.689 0.548 0.884 0.884 0.867 Expected heterozygosity (HJ 0.721 0.573 0.882 0.823 0.873 Ttho-7* 182 0.000 0.031 0.000 0.000 0.000 188 0.083 0.109 0.032 0.000 0.017 190 0.024 0.000 0.021 0.010 0.033 192 0.024 0.063 0.075 0.052 0.000 194 0.321 0.203 0.287 0.250 0.217 196 0.012 0.063 0.096 0.146 0.117 198 0.024 0.078 0.053 0.125 0.317 200 0.083 0.172 0.074 0.135 0.050 202 0.012 0.031 0.053 0.042 0.083 204 0.024 0.000 0.011 0.031 0.017 206 0.012 0.016 0.043 0.000 0.017 208 0.060 0.078 0.032 0.031 0.050 210 0.131 0.063 0.075 0.042 0.017 212 0.107 0.000 0.021 0.063 0.067 214 0.024 0.031 0.053 0.031 0.000 216 0.024 0.031 0.043 0.010 0.000 218 0.024 0.031 0.032 0.010 0.000 224 0.000 0.000 0.000 0.010 0.000 No. of samples 42 32 47 48 30 No. of allele 16 14 16 15 12 Effective no. of allele 6.49 8.93 8.06 7.63 5.59 Observed heterozygosity (Hn) 0.857 0.906 0.872 0.896 1.000 Expected heterozygosity (He) 0.846 0.888 0.876 0.869 0.821 1 Effective number of alleles was calculated with the equation l/( 1 -H l. 700 Fishery Bulletin 99(4) samples deviated considerably from zero (Fsr=0.018 to 0.070, PcO.OOl) for all loci. Significant departures of FST from zero were observed for three loci ( Ttho-1 *, Ttho-4*, and Ttho-6*) among all samples except the NE Atlantic sample ( FST =0.013 to 0.085, P<0.005) and among the three Pacific samples (FS7—0.018 to 0.074, PcO.OOl). There were also significant differences in allele frequencies be- tween some pairwise comparisons (Table 2). At Ttho-1 *, the SW Pacific sample showed significant difference from all other samples (all PcO.OOl), and there was also a sig- nificant difference between the SE Pacific and SW At- lantic samples (P=0.003). The NE Atlantic sample was significantly heterogeneous in comparison with all other samples in Ttho-4* (all PcO.OOl). Difference between the NW and SE Pacific samples was also significant in this locus (P=0.001). In Ttho-6 * the NE Atlantic sample was again significantly heterogeneous in comparison with all other samples (all PcO.OOl). Furthermore, there were also significant differences between the NW and SE Pacific, be- tween the SW and SE Pacific, and between the SW Pacific and SW Atlantic samples in this locus (all PcO.OOl). In Ttlio-7 * the NE Atlantic sample showed significant differ- ences from all three Pacific samples (PcO.OOl for the NW and SW Pacific, P=0.003 for the SE Pacific) but not from the SW Atlantic sample (P=0.017). Discussion Because the number of alleles observed in microsatellite loci is usually large and the frequency of each allele may be low, a large sample size is necessary for satisfying sub- sequent statistic analyses. Ruzzante (1997) showed that a sample size of 50