MBL U.S. Department of Commerce Ronald H. Brown Secretary National Oceanic and Atmospheric Administration D. James Baker Under Secretary for Oceans and Atmosphere National Marine Fisheries Service William W. Fox Jr. Assistant Administrator for Fisheries Scientific Editor Dr. Ronald W. Hardy Northwest Fisheries Science Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, Washington 981 12-2097 Editorial Committee Dr. Andrew E. Dizon National Marine Fisheries Service Dr. Linda L. Jones National Marine Fisheries Service Dr. Richard D. Methot National Marine Fisheries Service Dr. Theodore W. Pietsch University of Washington Dr. Joseph E. Powers National Marine Fisheries Service Dr. Tim D. Smith National Marine Fisheries Service The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115-0070. Second class postage is paid in Seattle, Wash., and additional offices. POSTMASTER send address changes for subscriptions to Fishery Bulletin, Super- intendent of Documents, Attn: Chief, Mail List Branch, Mail Stop SSOM, Washington, DC 20402-9373. Although the contents have not been copyrighted and may be reprinted entire- ly, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this period- ical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $24.00 domestic and $30.00 foreign. Cost per single issue: $12.00 domestic and $15.00 foreign. See back page for order form. Managing Editor Nancy Peacock National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C 1 5700 Seattle, Washington 981 15-0070 The Fishery Bulletin carries original research reports and technical notes on investiga- tions in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Begin- ning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 91 Number 1 January 1993 Fishery Bulletin MV3 '" Contents in Publication Awards iv List of recent NOAA Technical Reports NMFS 1 DeMartini, Edward E., Denise M. Ellis, and Victor A. Honda Comparisons of spiny lobster Panulirus marginatus fecundity, egg size, and spawning frequency before and after exploitation 8 Flores-Coto, Cesar, and Stanley M. Warlen Spawning time, growth, and recruitment of larval spot Leiostomus xanthurus into a North Carolina estuary 23 Fuiman, Lee A., and David R. Ottey Temperature effects on spontaneous behavior of larval and juvenile red drum Sciaenops ocellatus, and implications for foraging 36 Haldorson, Lewis, Marc Pritchett, David Sterritt, and John Watts Abundance patterns of marine fish larvae during spring in a southeastern Alaskan bay - 45 Hostetter, E. Brian, and Thomas A. Munroe Age, growth, and reproduction oftautog Tautoga omtis (Labridae: Perciformes) from coastal waters of Virginia 65 Jearld, Ambrose Jr., Sherry L. Sass, and Melinda F. Davis Early growth, behavior, and otolith development of the winter flounder Pleuronectes amencanus 76 Jordan, Alan R., and Barry D. Bruce Larval development of three roughy species complexes (Pisces: Trachichthyidae) from southern Australian waters, with comments on the occurrence of orange roughy Hoplostethus atlanticus 87 Krieger, Kenneth J. Distribution and abundance of rockfish determined from a. submersible and by bottom trawling Fishery Bulletin 91(1). 1993 97 Marks, Rick E., and David O. Conover Ontogenetic shift in the diet of young-of-year bluefish Pomatomus saltatrix during the oceanic phase of the early life history 1 07 McConnaughey, Robert A., and Loveday L. Conquest Trawl survey estimation using a comparative approach based on lognormal theory 1 1 9 Powell, Allyn B. A comparison of early-life-history traits in Atlantic menhaden Brevoortia tyrannus and gulf menhaden B. patronus 1 29 Renaud, Maurice, Gregg Gitschlag, Edward Klima, Arvind Shah, Dennis Koi, and James Nance Loss of shrimp by turtle excluder devices (TEDs) in coastal waters of the United States, North Carolina to Texas: March 1 988-August 1 990 1 38 Stillwell, Charles E., and Nancy E. Kohler Food habits of the sandbar shark Carcharhinus plumbeus off the U.S. northeast coast, with estimates of daily ration Notes 1 51 Chittenden, Mark E. Jr., Luiz R. Barbieri, and Cynthia M. Jones Spatial and temporal occurrence of Spanish mackerel Scomberomorus macu/atus in Chesapeake Bay 1 59 Francis, Malcolm R, Maryann W Williams, Andrea C. Pryce, Susan Pollard, and Stephen G. Scott Uncoupling of otolith and somatic growth in Pagrus auratus (Sparidae) 1 65 Lowerre-Barbieri, Susan K., and Luiz R. Barbieri A new method of oocyte separation and preservation for fish reproduction studies 171 Olson, Alan F, and Thomas R Quinn Vertical and horizontal movements of adult chmook salmon Oncorhynchus tshawytscha in the Columbia River estuary 1 79 Singhas, Lynda S., Terry L. West, and William G. Ambrose Jr. Occurrence of Echeneibothrium (Platyhelminthes, Cestoda) in the calico scallop Argopecten gibbus from North Carolina U.S. Department of Commerce Seattle, Washington Publications Awards 1990-91 National Marine Fisheries Service, NOAA The Publications Advisory Committee of the National Marine Fisheries Service is pleased to announce the awards for best publications authored by NMFS scientists and published in the Fishery Bulletin volume 89 and Marine Fisheries Review volume 52. Eligible papers are nominated by the Fisheries Science Centers and Regional Offices and are judged by the NMFS Editorial Board. Only articles which significantly contribute to the understanding and knowledge of NMFS-related studies are eligible. We offer congratulations to the following authors for their outstanding efforts. Fishery Bulletin 1 99 1 Thomas A. Munroe Western Atlantic tonguefishes of the Symphurus plagusia complex (Cynoglossidae: Pleuronectiformes), with descriptions of two new spe- cies. Fishery Bulletin 89:247-287. Dr. Munroe is with the National Systematics Laboratory, Washington DC. Honorable Mention: Donald E. Pearson, Joseph E. Hightower, and Jacqueline T.H. Chan Age, growth, and potential yield for shortbelly rockfish Sebastesjordani Fishery Bulletin 89:403-409 . Dr. Pearson is with the Tiburon Laboratory of the Southwest Fisheries Science Center, as were Dr. Chan and Dr. Hightower. Dr. Chan has recently retired, and Dr. Hightower is now with the North Carolina State University at Raleigh. Marine Fisheries Review 1990 Clyde L. MacKenzie Jr. History of the fisheries of Rantan Bay, New York and New Jersey Ma- rine Fisheries Review 52(4): 1 -45. Dr. MacKenzie is with the Sandy Hook Laboratory of the Northeast Fisheries Science Center, Highlands, New Jersey. Dr. MacKenzie received Honorable Mention for his 1 989 paper on estuanne mollusc fisheries. Honorable Mention: Wayne N. Witzell and Edwin L. Scott Blue marlin, Makaira nigricans, movements in the western North Atlan- tic Ocean: Results of a cooperative game fish tagging program, 1954- 88. Marine Fisheries Review 52(2): 12-17. Dr. Witzell is with the Miami Laboratory of the Southeast Fisheries Science Center as was his co- author, the late Dr. Scott. U.S. Department of Commerce Seattle, Washington Recent publications in the NOAA Technical Report NMFS Series 108 Marine debris survey manual Christine A. Ribic, Trevor R. Dixon, and Ivan Vimng April 92 p. 1992. 109 Seasonal climatologies and variability of eastern tropical Pacific surface waters Paul C. Fiedler April 1992, 65 p. 110 The distribution of Kemp's Ridley sea turtles (Lepidochelys kempi) along the Texas coast: An atlas Sharon A. Manzella and Jo A. Williams. May 1 992, 52 p. Ill Control of disease in aquaculture. Proceedings of the nineteenth U.S. -Japan meeting on aquacul- ture, Ise, Mie Prefecture, Japan, 29-30 October 1990 Ralph S. Svrjcek (editor) October 1992, 143 p. 112 Variability of temperature and salinity in the Middle Atlantic Bight and Gulf of Maine Robert L. Benway, Jack W Jossi, Kevin P Thomas, and Julien R. Goulet. April 1993, I08p. Some NOAA publications are avail- able by purchase from the Superin- tendent of Documents, U.S. Govern- ment Printing Office, Washington, DC 20402. Abstract. — Size-specific fecundi- ties of spiny lobster Panulirus mar- ginatus were compared for two time- periods: pre- and early exploitation or "before" (1978-81), and post- exploitation or "after" ( 1991 ). Fecun- dity was further evaluated within each time-period at two collection sites that represented the major lob- ster fishing grounds (Maro Reef and Necker Island) in the Northwestern Hawaiian Islands. Complementary data on egg size and spawning- frequency index were compared be- tween study sites and time-periods. Study sites and time-periods had no observable effects on egg size or spawning frequency, and there was no temporal effect on fecundity at Maro Reef. Fecundities at the two sites differed, however; "after" size- specific fecundity was an estimated 1619^ greater than "before" fecun- dity at Necker Island. Observations suggest that the recent increase in fecundity at Necker Island may reflect a compensatory (density- dependent) response to greater ex- ploitation at this site. Results are discussed in terms of evidence for density-dependent responses in other, exploited spiny lobster stocks. Comparisons of spiny lobster Panulirus marginatus fecundity, egg size, and spawning frequency before and after exploitation Edward E. DeMartini Denise M. Ellis Honolulu Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service. NOAA 2570 Dole Street Honolulu, Hawaii 96822-2396 Victor A. Honda Southwest Enforcement, Pacific Area Office National Marine Fisheries Service, NOAA 300 Ala Moana Boulevard Honolulu, Hawaii 96850-0001 Manuscript accepted 19 August 1992. Fishery Bulletin, U.S. 91:1-7 (1993). The spiny lobster Panulirus mar- ginatus (Quoy & Gaimard) is endemic to the Hawaiian Archipelago and Johnston Island (Brock 1973, Uchida et al. 1980). This species supported a major commercial fishery in the main Hawaiian Islands (MHI) prior to the rapid increases in demand after World War II (Uchida et al. 1980). Not until the expansion of the fishery into the Northwestern Hawaiian Is- lands (NWHI) began in 1977 did the species again support a valuable com- mercial enterprise, complemented with bycatches of slipper lobster Scyllarides squamosus (H. Milne- Edwards) and S. haanii (De Haan). A fishery management plan was cre- ated in 1983 to regulate the fishery based on minimum size limits and limited entry. Prior to 1990, annual landings av- eraged 1-2 million spiny lobster worth US$4- 6 million ex-vessel. Be- ginning in 1990 and continuing until the fishery closure in early 1991, however, landings fell heavily, equal- ing one-fifth of the long-term aver- age (Landgraf 1991). These decreases reflected real declines in abundance, as both research and commercial catch per trap-haul (CPUE) similarly declined (Landgraf 1991). The present belief is that recent declines in the spiny lobster CPUE reflect a combination of continued, heavy ex- ploitation and the occurrence of a se- ries of poor year-classes, particularly at Maro Reef, one of the two major NWHI fishing grounds (Polovina 1991). Recent research by Polovina (1989) has indicated that a density-depen- dent decrease in the size-at-onset of egg production occurred in NWHI spiny lobster from 1977 to 1986-87. Additional types of compensatory re- sponses to lower population densities may be operative and may have a major influence on the dynamics of these lobster populations, but data are lacking (Polovina 1989). Included among these compensatory mecha- nisms is an increase in size-specific fecundity, a phenomenon suggested for other species of spiny lobsters (Chittleborough 1976 and 1979, Beyers & Goosen 1987, MacDiarmid 1989). With the a priori prediction that size-specific fecundities might have increased for NWHI spiny lobster Fishery Bulletin 91(1). 1993 during the recent period of low population densities, we initiated a study of its fecundity and related repro- ductive life history. Prior to our study, little quantita- tive information existed on the fecundity of this spe- cies, and data were limited to the waters off Oahu in the MHI (Morris 1968, McGinnis 1972). Our objective is to compare the size-specific fecundities of NWHI spiny lobster between two time-periods: an early or pre-exploitation (hereafter referred to as "before") pe- riod in 1978-81, and a postexploitation ("after") period in 1991, when population densities had declined to a fraction of their pre-exploitation level. Methods and materials Specimen collection Spiny lobster were collected using baited commercial traps at Maro Reef and on the offshore bank of Necker Island, the second of the two major NWHI fishing grounds (fig. 1, Polovina 1989). Lobsters were trapped during multiple cruises aboard chartered commercial vessels and the NOAA ship Townsend Cromwell dur- ing the summertime (May- August) breeding seasons of 1978-81 (the "before" period) and on a single cruise by the Townsend Cromwell during June-July 1991 ("af- ter"). Commercial traps fished for a standard (over- night) soak period were used at each site during both time-periods. Specimens were similarly handled aboard the chartered vessels and the Townsend Cromwell. Sample processing Lobsters were sexed, carapace length (CD measured, and the egg developmental stage of egg-bearing ("ber- ried") females scored as either Stage 1 (orange = freshly extruded), Stage 2 (brown = late development), or Stage 3 (white = hatching imminent). The CL, defined as the distance along the middorsal line from the transverse ridge between the supraorbital spines to the posterior margin of the carapace, was measured to the nearest 0.1mm. Berried female specimens were either pro- cessed fresh in the ship's wet lab or flash-frozen (damp) aboard ship for processing ashore. In the laboratory, brood sizes were estimated using Stage- 1 females whenever possible so as to minimize the effect of potential egg loss (Morgan 1972, Annala & Bycroft 1987). The eight pleopods including egg clusters (setae bearing the egg masses) were separated by dis- section and placed on absorbent paper towels. Egg clus- ters were then stripped off the pleopods onto preweighed weigh boats. Each individual female's total egg comple- ment was weighed (damp weight to 0.1 mg) and then reweighed following determination of egg subsample 1 ^ i i i i i Maro Reef * "After" exploitation a ..-"' Ifi a "Before" exploitation A &.-■"' M * * ?.-■"' Stf) \ '>" A a"' ** A * * — ' .-* A >> 12 ft.''' A +J A ■3 d 3 ..-•' "Before" and "after" pooled V ..-•■'' In F = 1 73 + 2 39(ln CL) <4H ..-•■'" R Z = 60 d N = 53 P < 0.001 A 1 .i 1 i 1 1 1 1 1 y Necker Island N = 32 a/ - P < 0.001 J&* ' " O ■z. :JfeK ' — ' J^F M- >> 12 spj& • ** •3 oyrtr \ a x^* °*- 3 yf °*%D O, o y/ o o "Before" exploitation (M ' In F = 2 80 + 2 16(ln CL) d R 2 = 71 N = 35 P < 0001 1 1.1,1,1:1. 3.8 4.0 4.2 4.4 4.6 4.8 5.0 In carapace length (mm) Figure 1 Scatterplots, least-squares regressions, and regression statis- tics for In fecundity (number of eggs I versus In carapace length (in mm) for berried female spiny lobster Pan ill I run margmatus trapped at two locations in the Northwestern Hawaiian Islands. Maro Reef (top). Data for the "before" (1978-81) and "after" ( 1991 ) periods are pooled for the regression analy- sis but plotted separately; one extreme outlier was omitted (see Results). Necker Island (bottom). Data for the "be- fore" and "after" periods are plotted and analyzed separately. Arrows indicate the five most-extreme "before" data that were deleted in a re-analysis of the data (see Results). weights; these two weighings were then averaged to provide a measure of the total egg mass. Random subsamples comprising a minimum (by weight) of 1% (£=1.5%) of the female's total egg mass were weighed (0.1 mg) and later enumerated to estimate fecundity (F t =total number of eggs) by proportion: DeMartim et al Fecundity comparisons of Panuhrus margmatus F. = F. ( w. where F s = number of eggs in subsample, W s = weight of egg subsample, and W, = total weight of eggs. Some frozen-thawed egg masses were fixed in 4% formalde- hyde for 1 month to harden eggs prior to weighing and counting. A single subsample was used to characterize the fecundity of each "before" specimen. Three repli- cate subsamples were used to estimate the sampling error of "after" fecundity determinations; the three pooled subsamples provided the best estimate of "af- ter" fecundity. Total eggs were counted for one of the "after" specimens to gauge the accuracy of the weigh- ing and counting procedures. Egg sizes were estimated to complement the fecun- dity data. For a subset of both "before" and "after" Stage- 1 specimens, a minimum of 25 eggs per female were randomly chosen and measured (random axis, at 50 X ) using a dissecting microscope with calibrated eye- piece micrometer. Total egg production is the product of the number of eggs produced per spawning (brood size) and the num- ber of spawnings. For females above threshold body sizes at onset of egg production at each of the sites during the two time-periods (Polovina 1989), we in- dexed spawning frequency based on the relative fre- quencies of berried (to total) females present in his- torical catch data of the Honolulu Laboratory. We used records of catches made at Maro Reef and Necker Is- land on summertime cruises during years within pre- and postexploitation periods when sufficient data were available (1977, 1988-91). Statistical analysis Analysis of covariance (ANCOVA, SAS Proc GLM; SAS 1985) was used to compare mean fecundities between sampling periods; CL was used as a covariate to adjust for potential body-size differences between periods. As justified, central tendencies in fecundity were compared between periods ("before," "after") using least-square means (LSM) and their standard errors (SEM). Period and site (Maro Reef, Necker Island) were evaluated as class variables. Student's i-test, with degrees of freedom adjusted (as necessary) by Satterthwaite's approxima- tion for unequal variances (Bailey 1981), was used to compare indices of spawning frequency between periods. Results Size-fecundity relationships Paired CL and fecundity data were available for 54 spiny lobster from Maro Reef. At Necker Island, there were 67 analogous data pairs (Appendix A). Over 90% of the "before" specimens had Stage- 1 eggs, and Stage- 2 eggs were equally distributed among specimens from the two sites. Incidence of Stage-2 eggs appeared higher in the "after" samples from Maro Reef (8/24=33%) than in the analogous samples from Necker Island (3/32=10%). No lobsters with Stage-3 eggs were col- lected during either time-period. The coefficient of variation [CV=(SD/.r)-100] of the triplicate "after" fe- cundity estimates was about 2%. The accuracy of the mean of the two weighings of an entire egg mass was within 4% of a total count. At Maro, slopes were indistinguishable between pe- riods, regardless of whether an obvious outlier (whose residual deviated 8.5% from its predicted value) was included (In CL x period interaction: F 150 =0.06, P=0.81) or was deleted from the analysis (F 149 <0.01, P>0.99). Slopes also were indistinguishable between the two periods at Necker Island (F 163 =1.17, P=0.28). Intercepts did not differ between periods at Maro Reef (F 150 =0.22, P=0.64), but the period (intercept) ef- fect at Necker Island was significant (F 164 = 10.17, P=0.002). Greater size-specific fecundity in the "after" period at Necker Island persisted, even if the five most- extreme "before" values ( noted by the arrows in Fig. 1 ) were deleted and the analysis rerun (F 159 =4.40, P<0.05). Using all available data, the power (1 minus Type-II error) of the test for period differences at Necker Island was 84%, for a critical Type-I error of 5% (a 2 =0.05). CL significantly influenced fecundity at both Maro Reef (Fj 50=70.6, P<0.001; Fig. 1) and at Necker Island (F 164 =215.6, P<0.001; Fig. 1). After adjustment for pe- riod differences in CL, the fecundity of lobsters at Necker Island was an estimated 16±9% greater dur- ing the "after" versus the "before" period (LSM±SEM of InF = 12.224±0.034 and 12.072±0.033, respectively). Unlike the case at Necker Island, mean fecundity at Maro Reef differed only by <3% between the "before" and "after" periods (LSM±SEM = 12.680±0.041 and 12.651± 0.046, respectively). Egg size The median egg diameters of 22 females collected from Maro Reef and Necker Island during the "before" pe- riod were 0.58-0.69 mm. The analogous data for 53 "after" females were 0.61-0.73 mm, with a grand me- dian of 0.66 mm. Slopes of female body size/egg size (median diam- eter) relations were indistinguishable between sites (CL X site interaction: F 151 =0.55, P=0.46). Intercepts also were indistinguishable: egg size was uninfluenced by site (F 152 =2.75, P=0.10). Carapace length had no Fishery Bulletin 91 1 1993 effect when sites were evaluated separately (F 152 =0.11, P=0.75). However, female size significantly but weakly (P 2 =0.08) affected egg size when data for the two sites were pooled (CL effect: F I53 =2.104, P=0.04; median egg diameter = 30.2 EPU+0.039 CL mm ; N=55; EPU= 0.0197 mm). Spawning frequency The relative frequency of berried/total adult females collected at Maro Reef and Necker Island during the summer of 1988 (iV=3085 adult females), 1990 (1198), and 1991 (1165) was 0.246±3.401 (*±1SD, N=6 site- year combinations). This index of the spawning fre- quency of females did not differ (t'=0.01, P>0.9) from 0.230±0.510, the estimated frequency for 3037 females collected during the summer of 1977 (N=2 site-years). specific fecundity might be expected to co-vary with egg size and spawning frequency (Gadgil and Bossert 1970). However, in many organisms, offspring size and number often do not track one another simultaneously or to an equivalent extent (Capinera 1979, Roff 1982). Therefore, our observation that egg size did not co- vary with egg number in Panulirus marginatus should not be surprising. Perhaps strong selection for plank- tonic larvae of relatively invariant body size is typical within particular populations of spiny lobster, even though average egg sizes might differ among popula- tions of some species. This speculation is consistent with our observation that estimated egg volume var- ied only about 50% among female P. marginatus of a large range of body sizes from either site. This value is low compared with those of most marine teleosts (Bagenal 1971). Discussion Fecundity-body size relations Exponents of the curvilinear, F=aCL b , relations ob- served in this study ranged from 2.16±0.241 (statisti- cally equal to 2.0) to 2.50±0.175 (2.00.12, the subsample was 30 fish, except when more than 1000 fish were caught, at which time 40 fish were used. Lar- vae were measured to the nearest O.lmmSL. Age was determined according to the method of Warlen & Chester (1985). The estimated age of a larva was the observed number of sagittal growth increments from one reading plus the estimated number of days from hatching to first increment formation (5 d). The precision in dupli- cate readings of otoliths from 25 larvae (range 32-86 d) was estimated from the differences in paired readings. The mean (±SD) difference was 1.52±1.26 growth incre- ments, and the range was 0—4 increments. The birthdate (= spawning date) of each larva was back-calculated by subtracting its estimated age from the date of capture. Larvae spawned in a given calendar week were consid- ered in the same calendar birthweek cohort. The spawn- ing period of spot was estimated from the back-calcu- lated birthdates of larvae recruited to the estuary over all seasons. For each weekly collection, the percentage of larvae from each birthweek cohort was determined. Each percentage was multiplied by the corresponding weekly density total (larvae/100 m 3 ) to give the density of larvae from each birthweek cohort. The densities for each birthweek cohort were summed over all collections and their percentage contribution to the total density (1540 larvae/100 m ! ) for all birthweek cohorts was cal- culated. The Laird version (Laird et al. 1965) of the Gompertz growth equation was used to describe growth of spot larvae from the combined marine and estuarine collections. To stabilize the variance of length over the observed age interval, we used the log-transformed ver- sion of the Gompertz growth equation. Results Estuarine abundance and age/size distribution of larvae Abundance of larvae A total of 9760 spot larvae was collected at Pivers Island between 2 December 1987 Fishery Bulletin 91 [I). 1993 Figure 1 Location of sampling sites for spot Leiostomus xanthurus larvae at Pivers Island near Beaufort Inlet and ocean stations ("+") off the North Carolina coast. Surface-water temperatures are indicated for February 1988 stations. and 4 May 1988 (Fig. 2). The 4th of May was consid- ered the virtual end of the recruitment period, since spot densities had declined to <2 larvae/100 m 3 over the last 2 weeks of sampling. The sum of the weekly mean larval densities over all collections ( 1540 larvae/ 100 m 3 ) was used as the basis for determining the per- centage of recruited larvae from each birthweek co- hort. Larval density gradually increased during the first 10 weeks of sampling, varying from to 40 lar- vae/100m 3 (x=14.6). Approximately 9% of the larval spot recruitment occurred during this period. The pe- riod of highest density (x= 138.3 larvae/100 m 3 ) occurred during 10 February-13 April, when 88% of the larvae were recruited to the estuary. During this 10-week period, there was wide variation in relative abundance with four clear peaks, the largest occurring on 23 March when 34% of all larvae were collected. During the last 3 weeks of sampling, larval density declined to very low levels (x=3.7/100m 3 ) and represented less than 3% of the total spot larvae collected. There was varia- tion in the catch densities among net sets on any given sampling night. Excluding the first 4 weeks when no spot larvae were collected in 41% of the sets, the aver- age nightly coefficient of variation was 65.3% (range 20.3-119.5%). We assumed that larvae caught each week were newly recruited to the estuary and that they were in transit to upper reaches of the estuary past Pivers Island. These assumptions are supported by the gen- erally small standard error in the age of larvae within each collection (Fig. 3). The small observed within- sample variation in age is probably due in part to mixing of age cohorts in the ocean prior to estuarine recruitment. Since larvae were not accumulating in the lower estuary, there was no increase in standard error of mean age over time. Also, in the week follow- ing each of the four peaks (Fig. 2) densities were rela- tively low, a pattern that did not suggest substantial carryover offish from week to week. Age and length of larvae The weekly mean age of spot larvae caught at Pivers Island fluctuated between Flores-Coto and Warlen: Spawning time, growth, and recruitment of larval Leiostomus xanthurus 1 I JAN ' FEB ' MAR COLLECTION OATE Figure 2 Weekly mean density ( larvae/100 m' I of spot Leiostomus xanthurus larvae in collections at Pivers Island in the New- port River estuary, November 1987 to May 1988 (no larvae caught before 2 December). Mean weekly surface-water tem- perature calculated from hourly measurements at Pivers Island. 37 and 108 d, (x=82.4d); weekly mean SL varied from 9.2 to 22.2 mm (jr=17.2mm). The weekly mean age and SL of spot larvae (Fig. 3) increased from the beginning of the re- cruitment period, when the youngest and smallest larvae were caught, to the beginning of the period of peak recruitment density (10 February). Thereafter, average values remained high, only decreasing during the last 3 weeks. The age and size distributions of all spot larVae recruited to the estuary (Fig. 4) indicated that 68% were between 75 and 95 d, and 80% were between 15.1 and 20.0 mmSL. The mean age (±SE) and SL (±SE) of larvae corresponding to the three recruitment periods of different densities were 50.4 (±2.7), 84.6 (±1.4), and 62.6d (±3.8), and 11.8 (±0.49), 17.7(±0.31), and 13.4mm (±0.69). Spawning time Spawning, which was continuous over a 5- month period, began near mid-October and ended about mid- March (Fig. 5). Over 99% of spawning occurred from 1 Novem- ber to 24 January (22 weeks). Only about 1% was contributed by 9 weekly cohorts: 2 before and 7 after the main spawning period. Cohorts from 22 November to 10 January contributed >80% of total larvae (Fig. 5). The midpoint in spawning was the week of 20 December. E E ,6 I I- i " Ul Q LU a NOV I DEC I JAN I FEB I MAR I APR I MAY COLLECTION DATE Figure 3 Mean standard length (mmllSEl and age (dilSEl of spot Leiostomus xanthurus larvae collected weekly, November 1987 to May 1988, at Pivers Is- land in the Newport River estuary (no larvae caught before 2 December). B .cm UlU 20 22 24 STANDARD LENGTH Figure 4 Percentage distribution of larval spot Leiostomus xanthurus recruited to the Newport River estuary, No- vember 1987 to May 1988, by (A) estimated age (5d intervals) and (B) standard length ( 1 mm intervals). Fishery Bulletin 91(1), 1993 15 22 29 6 13 20 27 10 17 24 31 14 21 28 6 13 Figure 5 Percentage distribution of the number of larval spot Leiostomus xanthurus of back-calculated birthweeks recruited to the Newport River estuary, No- vember 1987 to May 1988. Age and length of birthweek cohorts Larvae from the beginning and end of the spawning period reached the estuary at younger mean ages than those from the middle period (Fig. 3). Larvae from birthweek cohorts were recruited to the estuary over periods ranging from 2 to 10 weeks, with an average of 7 weeks for the main spawning period (Table 1). In all but four cohorts (25 October, 22 November, 6 December, and 3 January), at least 50% of the larvae from the cohort reached the estuary during a single week (Table 1 ). The SL of larvae reaching the estuary (Table 2) follows a similar pattern to age. Except in one instance (3 February), the first two and the last five cohorts to reach the estuary had a mean SL <13.9mm. In the middle period, the weekly mean SL of larvae was gen- erally larger. The difference between the smallest and largest mean SL of larvae of any one birthweek cohort throughout the recruitment period varied from 0.3 to 11.7 mm. This was generally related to the total time during which a cohort recruited to the estuary. Abundance of birthweek cohorts Several birthweek cohorts contributed substantially to more than one of the recruitment peaks. Three cohorts (13, 20, 27 De- cember) contributed at least 50% of their respective total recruits to the large influx of larvae that occurred on 23 March (Table 1). The high densities of larvae (Fig. 2, Table 1) captured on 10 and 24 February, 23 March, and 13 April were collections to which some birthweek cohorts contributed >50% of their total re- cruitment (e.g., 1 and 8 November cohorts to catch of 13 April ). Oceanic abundance and age/size distribution of larvae Abundance of larvae The highest densi- ties of spot larvae collected offshore during January and February generally occurred in waters >30m (Fig. 6). Densities there ranged from 24 to 68 larvae/100 m ', but di- minished toward the coast and further off- shore toward the shelf break. In areas within 40 km of the coast at depths <30m, densities were <5 larvae/100 m ! . Except for two larvae collected off Beaufort Inlet in January (Fig. 6A), no larvae were collected within 10 km of the coast. Larval densities in the estuary at Pivers Island were always higher than at any oceanic station <30m deep within 40 km of the coast. Spot larvae were collected at two of seven stations sampled east and north of Onslow Bay during January (Fig. 6A). In February, spot larvae were collected at all but one of these stations (Fig. 6B). Densities were rela- tively low except at the station nearest Onslow Bay. Age and length distribution of larvae Ages of 351 spot larvae caught during oceanic sampling ranged from 9 to 69 d. Youngest larvae occurred farthest off- shore, over the outer continental shelf and within the Gulf Stream (Fig. 7A,B). Age of larvae varied inversely with distance from shore along the Beaufort Inlet transect in January and February and the Oregon In- let transect in February (Table 3, Fig. 7A,B). The mean age of larvae in Onslow Bay during February seems to increase toward shore from a dispersion center on the outer continental shelf south of Beaufort Inlet. Older larvae radiate to the north and west in Onslow Bay (Fig. 7B). Larvae in the transect across the continen- tal shelf off Oregon Inlet (Fig. 7B) may have a general spawning area in common with larvae collected in Onslow Bay. The length of larvae also varied inversely with dis- tance from shore (Table 3, Fig. 7C,D). Smallest larvae were found on the outer continental shelf and over the continental shelf break. The mean size of larvae was 3.4mmSL (range 2. 1-10. lmm) in January and 6.2mmSL (range 2.5-12.7 mm) in February. Spawning time Spot larvae collected off North Caro- lina were spawned over the period 8 November to 17 January (Fig. 8). As many as six cohorts were found in any one sample and the overall mean number of co- horts per sample was three. In 19 of 33 stations, >50% of the larvae were from one cohort, and in six stations >50% were from two consecutive cohorts. The remain- Flores-Coto and Warlen: Spawning time, growth, and recruitment of larval Leiostomus xanthurus 13 6 \ c o a — > M US OJ Z n 3 k, ■e re 3 h- ■C X CO c3 O T3 CM — — co CM CO as -<* CO LO OS -r LO t— lO co CO C c ^ O : i t- CO CM lO CO ^ ■^ ~ en CD CD en CO -H ri O CO CM CO CD O LO -r co CO — LO en LO LO CM X LO o lO LO CD CM CD o LO o CJS CO Lfi - CM = CO CO CO s CO o o CO o en Zl LO LO CO CO — LO 3 — 5 lO CO n o CD ir- CO CO ** o co en CO CD lo CO CM CM LC CM CO CD CO CO CO CM CO i- CO CD CD LO "^ O O •** ** t- CO CO CJi Ol *-J CD CD CM CO CO CO LO CO CD LO CD en lO CX CD — CO LO CO CO to CM CO ^ CD CM LO CD tC LO *" H LO CM CD lO CO LO CJ5 35 CM LO -r ^t CM LC CX> CM i — i f CO sc -^ o © © £ Fishery Bulletin 91(1). 1993 '3 IN tS 0J-2 c B to Q> 5 * CO ao CM 3 " CO ■^ O) cc CO CM CM 70 Tp CO GO CO CM :•" eo O CO CO CO CD CO m t~- t- o ^ ^ t~- c- CD ao o uO co cm CO tr^ CO f CD a^ co ^ cm o: co CM CD t^ GO 00 CO CO CD •■* CM i— I CO C-' co oS oS CT> O IT3 r-H O --5 CM O 0} CO iO CD CO ^ U0 CD t-h CM iri ifi cd oS © 1—5 H H H rt (M N CO O H Tf CO Ol H eouor-aoocococoo aicMT^coiOi^cDcbo cc Cft t* CO CD 05 CO -^ |Zi ~ re iO CD ir- X c- o CD ^H CM o |Z! t~- CO t- CO t- CM O »H co CI CO f- © © ■8 3 £ Flores-Coto and Warlen: Spawning time, growth, and recruitment of larval Leiostomus xanthurus Figure 6 Density (larvae/100 m') of spot Leiostomus xanthurus larvae collected in 1988 with bongo nets off the North Carolina coast during (A) 12-13 January, and (B) 2-5 February, and at Pivers Island with a neuston net. + indicates no larvae collected. 16 Fishery Bulletin 91(1), 1993 ing eight stations each had <5 lar- vae. There were eight different birthweek cohorts in January and nine in February. The principal cohorts in January were those of weeks beginning 20 and 27 De- cember, which together contrib- uted more than 88% of all larvae collected in the ocean (Fig. 8). These cohorts were the most im- portant contributors to larvae re- cruited during the week of great- est abundance, 23 March (Table 1, Fig. 2). During Febru- ary, larval cohorts were princi- pally from birthweeks beginning 20 December-23 January. Larvae from these cohorts were caught later in the estuary and contrib- uted heavily to the abundance peaks of 23 March and 13 April (Table 1). A Kolmogorov-Smirnov two- sample test (Sokal & Rohlf 1981 ) was used to compare larval birthdate distributions for a 3-week spawning period, 13 De- cember-2 January (Figs. 5&8), for larvae collected during Janu- ary-February in the ocean and February-March in the estuary. There was no significant dif- ference (max. diff. = 0.108, P>0.05) between birthdate distri- butions of 98 larvae collected in the ocean in January and 109 col- lected in February. There was also no significant difference be- tween the pooled birthdate dis- tribution data for larvae collected in the ocean in January and February (rc=207) and for 134 lar- vae collected in the estuary at Pivers Island (max. diff. = 0.180, P>0.05). Growth rate The overall growth rate of larval spot was estimated from 312 es- tuarine and 351 oceanic speci- mens combined. Larvae ranged from 9 to 108 d and 2.1 to 22.2 mmSL (Fig. 9). From the Laird-Gompertz model (Fig. 9), we predicted that spot grew from 1.2mm at Figure 7 Contour plots of the mean age (d) on (A) 12-13 January, and (B) 2-5 February, and mean standard length (mm) on (C) 12-13 January and (D) 2-5 February, of spot Leiostomus xanthurus larvae collected off the North Carolina coast in 1988. + indicates no larvae collected. hatching to 19.1 mm in 95 d, an overall average growth rate of 0.188mm/d. The size at hatching, estimated from the Laird-Gompertz model (1.2mmSL), was less Flores-Coto and Warlen. Spawning time, growth, and recruitment of larval Leiostomus xanthurus 76" (^ EMtUH S*" 3-6 / ATLANTIC OCEAN 50 100 75' 74* ^^i VJ"\ ATLANTIC OCEAN Figure 7 (continued) larvae were 9.3mmSL and 46 d old. Within-season growth was compared for similar-age larvae collected at Pivers Island be- tween early (early February) and late (early April) portions of the peak recruitment period and be- tween early (early February) and post-peak recruitment periods (late April). The mean growth rate (0.195 mm/d) of 15 larvae <17.7±1.30mmSL, 82.5±2.56 growth increments) collected 3 and 10 February ( early peak) was not significantly different (r-test, P=0.09> from that (0.204 mm/d) of 15 larvae (18.2±0.98mmSL, 81.9±3.19 growth increments) collected 6 and 13 April (late peak). However, the mean growth (0.214mm/d) of 12 larvae (14.9±1.45mmSL, 62.3±3.77 growth increments! collected on 3 February (early peak) was sig- nificantly different (r-test, P<0.01) from that (0.194mm/d) for 10 larvae (13.4±0.64mmSL, 60.7±3.43 growth increments) collected in late April (post-peak). Discussion than the 1.6-1.7 mmSL measured on laboratory-reared larvae by Powell & Gordy (1980) and estimated from wild specimens by Warlen & Chester (1985). In the log-transformed model (Fig. 9), age accounted for 98% of the variation in length. Age-specific growth rate declined from 5.8%/d at age 10 d to <0.7%/d at age 100 d. Maximum absolute growth rate occurred when The spawning period of spot in the 1987-88 season was appar- ently a continuous process occur- ring over 5 months from mid- October to mid-March. Although spawning was protracted, the greatest concentration occurred in the 2-month interval from mid-November to mid-January when 90% of the estuarine- recruited larvae were spawned. This information, while generally agreeing with earlier work on fish collected in North Carolina (Hildebrand & Cable 1930, Warlen & Chester 1985) and South Carolina (Beckman & Dean 1984), is the first to be estimated from back-calculated birthweek distributions on larvae collected weekly over the en- tire estuarine recruitment period. The contribution of birthweeks to the larval catch each week was based on Fishery Bulletin 91(1). 1993 Table 3 Linear regressions of mean es i mated age (d) and mean stan- dard length (mmSL) of spot Leiostomus xanthurus larvae on collection distance (km) from shore. Stations Transect/ along Correlation month transect Variable Intercept Slope Coefficient Beaufort Inlet 8 Age 44.875 -0.364 0.905 January SL 10.227 -0.088 0.886 Beaufort Inlet 7 Age 68.345 -0.515 0.936 February SL 13.465 -0.104 0.953 Oregon Inlet 3 Age 71.279 -1.526 0.999 February SL 15.756 -0.363 0.997 60 50 UJ 40 O < I 30 DC UJ Q. 20 10 30 UJ u < 20 t- Z o cc UJ 10 0. JANUARY 198 8 FEBRUARY 1988 15 22 29 6 13 20 27 3 1 Percen birthwe vae cai coast: 2-5 Fel NOV DEC Figure 8 ,age distribution of be eks for spot Leiostomus lght in 1988 off the N (top) 12-13 January >ruary. ck- tan ortr iiid JA :al< hu C ilx ^ ula ■us iro >tt< ted lar- ina Mill the percentage age composition and den- sity of larvae. The seasonal differences in age and size of larvae at estuarine recruitment (Fig. 3) suggest that spawning may have been nearer the coast at the beginning and end of the spawning season. As adult spot emigrate from the estuary to offshore waters in fall, at a time of decreasing photoperiod and water temperature (Mer- cer 1989), they probably seek suitable water tempera- tures for spawning ( 17.5-25°C) (Hettler & Powell 1981). This temperature range is present over much of the North Carolina continental shelf water (Stefansson et al. 1971) in the early months of spawning. During later spawning (mid-December to February), nearshore wa- ters are cooler and well mixed and water above 17.5°C is generally restricted to areas on the outer continen- tal shelf (Stefansson et al. 1971, Atkinson 1985). Also, in winter the extent of this potential spawning area may be influenced by the Gulf Stream and its occa- sional wave-like perturbations (cyclonic meanders and filaments) along its western edge which can intrude onto the continental shelf. It is known that average surface-water temperatures on the outer continental shelf are moderated by the Gulf Stream (Atkinson 1985). The winter (January and February) distribution patterns of age and size of spot larvae in Onslow Bay, with the youngest, smallest larvae occurring only furthest offshore, support the idea that spawning may occur some 90 km offshore. Spawning that occurs further offshore as the spawning season progresses has been suggested for spot (Lewis & Judy 1983, Warlen & Chester 1985) and Atlantic croaker Micropogonias undulatus (Warlen 1982). Norcross & Austin (1988) also suggested that the area of warm water encountered upon migration of Atlantic croaker from Chesapeake Bay onto the continental shelf determines spawn- ing location. Differences in age at estuarine recruitment may also be due to differences in transport rate that result from the degree to which favorable currents facilitate more rapid transport of larvae toward shore. Different transport rates and spawning distances from shore could also act in concert to produce the observed differences in age and size at recruitment. Although some of the physical processes that could affect larval transport have been discussed (Checkley et al. 1988, Miller 1988, Pietrafesa & Janowitz 1988), precise larval- fish transport mechanisms still remain unknown. The higher abundance of spot larvae in water deeper than 30 m may be a function of spawning location and subsequent larval trans- port toward shore. The reduced abundance observed inshore of the 30 m isobath may reflect fewer numbers of larvae present or their reduced vulnerability to capture by bongo nets. Mortality will also reduce larval abundance over time. Kjelson et al. (1976) and Miller et al. (1984) suggest that spot larvae offshore are more pelagic but that inshore they are more benthic-oriented. Spot larvae caught from shore to the 30 m isobath were 40-61 d old and 8.2-12.1 mm SL, and correspond to the early stages of the transformation period (Govoni 1980 and 1987, Powell & Gordy 1980) when spot begin to be more benthic. Larval spot are recruited from offshore spawning areas to estuar- ies bordering Onslow Bay. Fish of the same ages also are found along the offshore to onshore Oregon Inlet transect (Figs. 6,7). This data and the fact that larval spot were found in and near the Gulf Flores-Coto and Warlen: Spawning time, growth, and recruitment of larval Leiostomus xanthurus 19 - 0.024 ESTIMATED AGE (days) Figure 9 Growth of larval spot Leiostomus xanthurus collected from the estuary and oce- anic waters off North Carolina, November 1987 to May 1988. A Laird-Gompertz model was fit to the age/size data for 663 fish. Estimates of the parameters were obtained by fitting the log-transformed version of the model to the data. L„=length at hatching, A,i=specific growth rate at hatching, and a = exponential decay of the specific growth rate. Stream (Fig. 6) suggest that larvae are being trans- ported to areas north of Onslow Bay. The origin of these larvae is probably south of Cape Hatteras and most likely Onslow Bay or southward. These data sup- port the hypothesis of Norcross and Bodolus (1991) that spot spawned south of Cape Hatteras on the outer continental shelf contribute to recruitment in Chesa- peake Bay. Spring (March-May) spawned bluefish Pomatomus saltatrix are also thought to be transported to the Middle Atlantic Bight from spawning areas near the edge of a northerly flowing warm-water mass (Gulf Stream) in the South Atlantic Bight (McBride & Conover 1991). The extended recruitment period of 5 months ( Fig. 2 ) is a reflection of the length of spawning period (Fig. 5), although the time from spawning to recruitment var- ies throughout the season. The beginning of the maxi- mum recruitment period coincides with increasing es- tuarine water temperature. Warlen & Burke (1990) found that peak immigration into North Carolina es- tuaries of fall-winter spawned ichthyoplankton matched the period of rising water temperature. This idea agrees with the fact that spot abundances are low during cold periods. Low water temperatures (<10°C) can cause cold stress by increasing larval respiration rate and can kill spot larvae (Hoss et al. 1988). The maximum estuarine recruitment period in North Caro- lina probably varys slightly from year to year, but is basically midwinter to early spring. Our estimate of the maxi- mum recruitment period (mid-Febru- ary to mid-April) is similar to that (February-March) recorded by Hettler & Chester (1990) and that (mid^Janu- ary to mid-March ) found by Warlen & Burke ( 1990). Apparently once spot lar- vae are in the estuary they move to- ward fresher water and utilize upper reaches of estuaries as nursery areas (Weinstein et al. 1980, Allen & Barker 1990). Peak recruitment of spot to the marshes of the Cape Fear River estu- ary in North Carolina occurred during March and April (Weinstein 1979). Interannual variations may be ex- pected as a consequence of the sea- sonal changes that trigger emigration of the adults from estuaries to oceanic spawning areas and the subsequent transport rates of larvae back to the estuary. The sum of the weekly mean larval densities over all collections ( 1540 lar- vae/100 m 3 ) was almost double that of 1985-86 (estimated from Fig. 2 of Warlen & Burke 1990), but less than that of 1989-90 and about equal to 1986-87 and 1988-89 (S.M. Warlen, unpubl. data). Allen & Barker (1990) also recorded variable patterns of larval spot abundance in South Carolina estuaries during 1981-84. The four highest peaks of recruitment density (10 and 24 February, 23 March, 13 April), that contributed about 74% of all spot larvae, could be con- sequences of concentration mechanisms of larvae out- side the inlet and the subsequent facilitation of the influx of pooled larvae to the estuary. Lyczkowski- Shultz et al. (1990) suggested that larvae of spot, as well as other offshore spawners, accumulate in nearshore areas to develop and grow prior to recruit- ment. Tide may be an important mechanism which forces the larval gathering process outside and inside inlets (Pietrafesa & Janowitz 1988). The higher abun- dance of larvae just inside the inlet compared with the abundance at inshore stations seems to be a common feature for many estuarine-dependent species (Warlen 1982, Lewis & Judy 1983, Warlen & Chester 1985). Because any birthweek cohort can be widely dis- persed in the ocean, their larvae may reach the estu- ary over a period of 2-10 weeks. However, in general, >50% of the larvae of any birthweek cohort are re- cruited to the estuary in one week (Table 1). Birthweek cohorts of 25 October to 15 November had bimodal 20 Fishery Bulletin 91(1), 1993 recruitment, with small groups of younger larvae of each cohort reaching the estuary earlier, and larger groups of older (and larger) larvae recruited later (Tables 1, 2). The groups are clearly separated by a period of no recruitment, and the separation becomes less evident with later birthweek cohorts. The exist- ence of two groups of recruits from early cohorts could result from dispersion of larvae spawned over the mid- continental shelf at different spawning locations and with different rates of transport to the estuary. Birthweek cohorts after November do not appear to be recruited to the estuary as distinct early and late groups. The results show that the abundant larval cohorts of birthweeks 20 and 27 December, caught in January and February in Onslow Bay, contributed substantially to recruitment over the last four weeks in March. Birthweek cohorts of 3, 10, and 17 January were well represented in two later estuarine recruitment peaks. The comparison of birthdate distributions of larvae col- lected in the ocean in January and February and later in the estuary in February and March provided an opportunity to assess the relative survival of cohorts. The Kolmogorov-Smirnov tests showed no significant difference between ocean and estuarine birthdate dis- tributions for larvae spawned over a period (13 December-2 January) of intense spawning. Therefore, we conclude that earlier (oceanic) and later (estua- rine) larvae were from the same birthdate distribu- tion, and that survival for the daily cohorts between 13 December and 2 January was not age-specific. The lack of seasonal sampling of larvae in the ocean pre- cluded similar comparisons of birthdates throughout the spawning season. During their oceanic existence, spot larvae grew rap- idly from a hatching size of about 1.6mmSL to a mean size of 17.2 mmSL at estuarine immigration. The growth curve was sigmoidal and similar to those found by Warlen & Chester (1985) for spot larvae in North Carolina during 1978-79 and 1979-80. Parameter es- timates of the growth model for larval spot in 1987-88 in North Carolina, i.e., length at hatching (L (lll =1.156), specific growth rate at hatching (A, o ,=0.074), and the exponential decline of the specific growth rate (°==0.024), were comparable to the growth parameter estimates for 1978-79 and 1979-80 (L, ,=1.686, 1.609; A«,=0.060, 0.067; «=0.021, 0.026) found by Warlen & Chester (1985). Maximum growth rate (9.3 mm, 46 d- old larvae) was between the values that they report (8.0mm, 46d old; and 10.7mm, 45d old). There did not appear to be large differences in within-season growth of spot, although significant differences could be demonstrated. Mean growth was about 0.19- 0.21 mm/d for larvae collected during and after the peak immigration period. Acknowledgments We thank A.J. Chester for statistical assistance and J.J. Govoni, W.F. Hettler, and D.S. Peters for their helpful critical reviews of an early draft of the manu- script. Weekly mean surface-water temperatures for Pivers Island were provided by W.F. Hettler. The se- nior author gives special thanks to Consejo Nacional de Ciencia y Tecnologfa, Direccion General de Asuntos del Personal Academico de la Universidad Nacional autonoma de Mexico, and Fondo para el Desarrollo de los Recursos Humanos del Banco de Mexico, for sup- port of his sabbatical program at the Beaufort Labora- tory of the National Marine Fisheries Service where this paper was developed. Citations Allen, D.M., & D.L. Barker 1990 Interannual variations in larval fish recruitment to estuarine epibenthic habitats. Mar. Ecol. Prog. Ser. 63:113-125. Atkinson, L.P. 1985 Hydrography and nutrients of the southeastern U.S. continental shelf. In Atkinson, L.P., D.W. Menzel, & K.A. Bush (eds.), Oceanography of the southeastern U.S. coastal shelf, p. 77-92. Am. Geophys. Union, Wash. D.C. Beckman, D.W., & J.M. Dean 1984 The age and growth of young-of-the-year spot, Leiostomus xanthurus Lacepede, in South Carolina. Estuaries 7(4B):487^96. Chao, L.N., & J.A. Musick 1977 Life history, feeding habits, and functional morphology of juvenile sciaenid fishes, in the York River estuary, Virginia. Fish. Bull, U.S. 75:657- 702. Checkley, D.M. Jr., S. Raman, G.L. Maillet, & K.M. Mason 1988 Winter storm effects on spawning and larval drift of a pelagic fish. Nature ( Lond. ) 335:346-348. Fahay, M.P. 1975 An annotated list of larval and juvenile fishes captured with surface-towed meter net in the south Atlantic Bight during four RV Dolphin cruises be- tween May 1967 and February 1968. NOAA Tech. Rep. NMFS SSRF-685, 39 p. Fruge, D.J. 1977 Larval development and distribution of Micropogonias undulatus and Leiostomus xanthurus and larval distribution of Mugil cephalus and Bregmacerus atlanticus off the southeastern Louisi- ana coast. M.S. thesis, Louisiana State Univ., Baton Rouge, 75 p. Fruge, D.J., & F.M. Truesdale 1978 Comparative larval development of Micropogonias undulatus and Leiostomus xanthurus (Pisces: Flores-Coto and Warlen: Spawning time, growth, and recruitment of larval Leiostomus xanthurus 21 Sciaenidaei from the northern Gulf of Mexico. Copeia 1978:643-648. Govoni, J.J. 1980 Morphological, histological and functional aspects of alimentary canal and associated organ development in larval Leiostomus xanthurus. Rev. Can. Biol. 39:69-80. 1987 The ontogeny of dentition in Leiostomus xanthurus. Copeia 1987:1041-1046. Govoni, J.J., D.E. Hoss, & A.J. Chester 1983 Comparative feeding of three species of larval fishes in the northern Gulf of Mexico: Brevoortia patronus, Leiostomus xanthurus, and Micropogonias undulatus. Mar. Ecol. Prog. Ser. 13:189-199. Govoni, J.J., A.J. Chester, D.E. Hoss, & P.B. Ortner 1985 An observation of episodic feeding and growth of larval Leiostomus xanthurus in the Northern Gulf of Mexico. J. Plankton Res. 7:137-146. Hettler, W.F. 1979 Modified neuston net for collecting live larval and juvenile fish. Prog. Fish-Cult. 41:32-33. Hettler, W.F., & A.J. Chester 1990 Temporal distribution of ichthyoplankton near Beaufort Inlet, North Carolina. Mar. Ecol. Prog. Ser. 68:157-168. Hettler, W.F., & A.B. Powell 1981 Egg and larval fish production at the NMFS Beau- fort Laboratory, Beaufort, N.C., USA. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:501-503. Hildebrand, S.F., & L.E. Cable 1930 Development and life history of fourteen te- leostean fishes at Beaufort, N.C. Bull. U.S. Bur. Fish. 46:383-488. Hoss, D.E., L. Coston-Clements, D.S. Peters, & P.A. Tester 1988 Metabolic responses of spot, Leiostomus xanthurus and Atlantic croaker, Micropogonias undulatus, larvae to cold temperatures encountered following recruitment to estuaries. Fish. Bull., U.S. 68:483-488. Johnson, G.D. 1978 Leiostomus xanthurus. In Development of fishes of the Mid-Atlantic Bight: An atlas of egg, larval and juvenile stages. Vol. IV, Carangidae through Ephippidae, p. 203-208. U.S. Fish Wildl. Serv, FSW/ OBS-78-12. Kjelson, MA., G.N. Johnson, R.L. Garner, & J.P. John- son 1976 The horizontal-vertical distribution and sample variability of ichthyoplankton populations within the nearshore and offshore ecosystems of Onslow Bay. In Annual report to the Energy Research and Develop- ment Administration, p. 287-341. Beaufort Lab., NMFS Southeast Fish. Sci. Cent. Laird, A.K., SA. Tyler, & A.D. Barton 1965 Dynamics of normal growth. Growth 29:233- 248. Lewis, R.M., & M.H. Judy 1983 The occurrence of spot, Leiostomus xanthurus, and Atlantic croaker, Micropogonias undulatus, lar- vae in Onslow Bay and Newport River estuary, North Carolina. Fish. Bull, U.S. 81:405-412. Lyczkowski-Shultz, J., D.L. Ruple, S.L. Richardson, & J.H. Cowan Jr. 1990 Distribution of fish larvae relative to time and tide in a Gulf of Mexico barrier island pass. Bull. Mar. Sci. 46:563-577. McBride, R.S., & D.O. Conover 1991 Recruitment of young-of-the-year bluefish Pomatomus saltatrix to the New York Bight: Varia- tion in abundance and growth of spring- and sum- mer-spawned cohorts. Mar. Ecol. Prog. Ser. 78:205-216. Mercer, L.P. 1989 Fishery management plan for spot {Leiostomus xanthurus). N.C. Dep. Nat. Resour. Commun. Dev. Spec. Rep. 49, Raleigh, 81 p. Miller, J.M. 1988 Physical processes and the mechanisms of coastal migrations of immature marine fishes. In Weinstein, M.P. (ed.), Larval fish and shellfish transport through inlets, p. 68-76. Am. Fish. Soc. Symp. 3, Bethesda. Miller, J.M., J.P. Reed, & L.J. Pietrafesa 1984 Patterns, mechanisms, and approaches to the study of migrations of estuarine-dependent fish lar- vae and juveniles. In McCleave, J.D., G.P. Arnold, J.J. Dodson, & W.H. Neill (eds.). Mechanisms of mi- grations in fishes, p. 209-225. Plenum Press, NY. Norcross, B.L., & H.M. Austin 1988 Middle Atlantic Bight meridional wind compo- nent effect on bottom water temperatures and spawn- ing distribution of Atlantic croaker. Continental Shelf Res. 8:69-88. Norcross, B.L., & DA. Bodolus 1991 Hypothetical northern spawning limit and larval transport of spot. In Hoyt, R.D. (ed.). Larval fish recruitment and research in the Americas, p. 77- 88. NOAA Tech. Rep. NMFS 95. Pietrafesa, L.J., & G.S. Janowitz 1988 Physical oceanographic processes affecting lar- val transport around and through North Carolina inlets. In Weinstein, M.P. (ed.), Larval fish and shell- fish transport through inlets, p. 34-50. Am. Fish. Soc. Symp. 3, Bethesda. Powell, A.B., & H.R. Gordy 1980 Egg and larval development of the spot, Leiostomus xanthurus (Sciaenidae). Fish. Bull., U.S. 78:701-714. Siegfried, R.C. II, & M.P. Weinstein 1989 Validation of daily increment deposition in the otoliths of spot (Leiostomus xanthurus). Estuaries 12:180-185. Sogard, S.M., D.E. Hoss, & J.J. Govoni 1987 Density and depth distribution of larval gulf men- haden, Brevoortia patronus, Atlantic croaker, Micro- pogonias undulatus, and spot, Leiostomus xanthurus, in the northern Gulf of Mexico. Fish. Bull., U.S. 85:601-609. Sokal, R.R., & F.J. Rohlf 1981 Biometry, 2d ed. W.H. Freeman, San Francisco, 859 p. 22 Fishery Bulletin 91(1), 1993 Stefansson, U., L.P. Atkinson, & D.F. Bumpus 1971 Hydrographic properties and circulation of the North Carolina shelf and slope waters. Deep-Sea Res. 18:383-420. Warlen, S.M. 1982 Age and growth of larvae and spawning time of Atlantic croaker in North Carolina. Proc. Annu. Conf. S.E. Assoc. Fish. Wildl. Agencies 34:202-214. Warlen, S.M., & J.S. Burke 1990 Immigration of larvae of fall/winter spawning marine fishes into a North Carolina estuary. Es- tuaries 13:453^161. Warlen, S.M., & A.J. Chester 1985 Age, growth, and distribution of larval spot, Leiostomus xanthurus, off North Carolina. Fish. Bull, U.S. 83:587-599. Weinstein, M.P. 1979 Shallow marsh habitats as primary nurseries for fishes and shellfish, Cape Fear River, North Caro- lina. Fish. Bull., U.S. 77:339-357. Weinstein, M.P., S.L. Weiss, R.G. Hodson, & L.R. Gerry 1980 Retention of three taxa of postlarval fishes in an intensively flushed tidal estuary, Cape Fear River, North Carolina. Fish. Bull., U.S. 78:419^36. Abstract— Spontaneous behavior of young red drum Sciaenops ocella- tus was examined over a period of 8h at two acclimation temperatures (21° and 26° C) and after acute tem- perature changes between these lev- els. Three sizes of fish were used (jc=9, 23, and 34mmTL). Activity of fish acclimated to 26° C was greater than that at 21°C for fish of all sizes. Duration of pauses in spontaneous activity was generally lower at the warmer temperature. Effects of han- dling stabilized after 2-5 h. The time course for activity after an acute thermal change followed the tradi- tional model for thermal stress, with an early overshoot followed by a sta- bilized period. The overshoot was positive for upward transfers (21- 26° C) and negative for downward transfers (26-21° C). Pause duration showed a time course roughly in- verse of the trend for activity, but pause frequency was inconsistent. Effects of 5° C changes stabilized af- ter about 2 h. Results indicate that a minimum adjustment period of 2-5 h is advisable when handling young red drum for research or for stocking into natural waters. The be- havior of young red drum deprived of food at acclimation temperatures suggests they are sweep, rather than saltatory, searchers. Temperature effects on spontaneous behavior of larval and juvenile red drum Sciaenops oce/fatus, and implications for foraging* Lee A. Fuiman David R. Ottey The University of Texas at Austin. Marine Science Institute PO. Box 1267, Port Aransas. Texas 78373-1267 Searching for food is critical to sur- vival, and any factor that influences foraging behavior may have vital con- sequences, especially early in life when starvation is a serious threat. Temperature is one of the most po- tent natural factors affecting fishes, and substantial thermal variability may be experienced routinely. Such variability can span a wide range of time scales. Annual and seasonal dif- ferences in water temperature are common. Measurements of thermal effects on this scale probably reflect differences between physiologically stable (acclimated) states with re- spect to temperature. Differences in swimming performance due to accli- mation temperature are well docu- mented (Beamish 1978). Shorter- term temperature variations are also common in nature (summarized by Montgomery & MacDonald 1990). Fishes residing in shallow, lentic wa- ters can experience large amplitude, diel temperature cycles (Bamforth 1962, Smid & Priban 1978). Move- ment across a thermocline imposes an even more rapid temperature change, as does inundation of tidal marshes and pools and the act of stocking hatchery fish into surface waters. Under these circumstances, Manuscript accepted 14 September 1992. Fishery Bulletin, U.S. 91:23-35(1993). * Contribution 847 of The University of Texas at Austin Marine Science Institute. the dynamic processes of physiologi- cal adaptation to the temperature change also contribute to the overall thermal effect. Our goal was to examine the effects of temperature on spontane- ous behavior of young red drum Sciaenops ocellatus. Here, we con- strue spontaneous behavior of soli- tary young fishes deprived of food as that typically used in foraging. We designed experiments to evaluate dif- ferences in behavior at two constant temperatures and after acute in- crease or decrease in temperature. Our measures of behavior are useful for quantifying foraging effort. Materials and methods All fish were reared from eggs spawned at the Fisheries and Mari- culture Laboratory of the University of Texas Marine Science Institute. Spawning occurred in the evening at 27-28° C. Eggs were collected the following morning and placed in 150 L rearing tanks maintained at two nominal acclimation tempera- tures (21° and 26° C), where they hatched within 24 h of spawning. Lar- vae were fed rotifers {Brachionus) at 3-4 d after hatching (3mmTL); Artemia nauplii were added to the diet at 10-11 d after hatching. Roti- 23 24 Fishery Bulletin 91 (1), 1993 fers were discontinued by day 15, when larvae were approximately 4-6 mm long. Dry food supplements were provided for larger fish, and Artemia densities were diminished so that juveniles eventually subsisted entirely on dry food. These and other details of rearing followed Holt etal. (1990). Experimental design and protocol Three sizes of red drum were studied. Mean (±SD) total lengths for the small, medium, and large size- classes were 8.7 (±0.9), 22.8 (±2.3), and 34.0 (±2.6) mm. The lower acclimation temperature (21°C) is well below optimum for red drum larvae (Holt et al. 1981), so those reared at that temperature grew more slowly and exhibited higher mortality rates during the first week than did the 26° C fish. Trials were conducted at both nominal temperatures, yielding four treatments. Trials on fish placed in water of the same temperature as their rearing tank (acclimation temperature) are termed 'high' (26°C) or 'low' (21° C) controls. Trials at a temperature differing from the acclimation tempera- ture are referred to as 'upward' (21° to 26° C) or 'down- ward' (26° to 21° C) transfers. For each trial, fish were transferred individually from the rearing tank to separate transparent experimental arenas and left undisturbed for the duration of the observations. Arenas were rectangular from above, with sides in a ratio of 5:3. Arena sizes were scaled such that the longer dimension of the rectangular surface area was 6 to 8 times the average total length of the fish. Water depth was 4.6 to 7.5 times the greatest body depth of the fish (3.5 to 7.5 times depth with fins expanded). Small fish were pipetted individually, while larger fish were released from 100 mL beakers con- taining a single fish in 50 mL of water from the rear- ing tank. Behavior was recorded on videotape through a cam- era mounted above the arenas. The recorder was acti- vated from about 1 min prior to transfer until 20 min after transfer, then for 5-min periods at intervals in- creasing from 15 to 60 min. In all, behavior was quan- tified during 19 5-min observation periods beginning 5, 10, 15, 30, 45, 60, 75, 90, 105, 120, 135, 165, 195, 225, 255, 315, 375, 435, and 495 min after transfer. Each of the four treatments was applied to six fish. Temperatures in rearing tanks were controlled by balancing air temperature with submersed aquarium heaters, and were slightly above nominal levels. Mean acclimation temperatures during the 10 d preceding each trial were 21.3-21.7° C and 26.1-26.5°C. Tem- perature in the experimental arenas was maintained by controlling room air temperature. This sufficed for trials with medium and large fish; however, arenas for small fish were placed in a larger water bath to stabi- lize their temperature. During the 8h trials, the arena temperatures were 20.7-21. 5°C and 26.1-26.5°C. By design, control fish were to experience no temperature change, and upward and downward transfers were to experience ±5.0° C. Actual differences between rearing tank and trial temperatures were -0.8° to +0.2° C for controls, +4.5° to +5.0° C for upward transfers, and -4.7° to -5.7° C for downward transfers. In quantifying behaviors relevant to foraging, we recognized the dichotomy of searching techniques in fishes that travel in search of food and locate prey visually. Some species actively search while swimming and are called 'sweep' searchers (Laing 1938, O'Brien et al. 1986). Others search during brief periods when swimming is interrupted (described by Janssen 1982, and termed 'saltatory' searchers by O'Brien et al. 1989). Both types of active, visual foraging have been sug- gested for young fishes. For example, larval Atlantic herring Clupea harengus are thought to be sweep searchers (Rosenthal 1969, Rosenthal & Hempel 1970), while larval white crappie Pomoxis annularis exhibit saltatory searching behavior (Browman & O'Brien 1992). Bell (1990) suggested that the saltatory style is more generally employed by teleosts. We examined three measures of foraging behavior: activity, pause frequency, and mean pause duration. Activity, the total amount of time spent swimming during each 5-min period, is a measure of foraging effort for sweep searchers, since they search new ter- ritory while swimming. Saltatory searchers use peri- ods of inactivity (pauses) for finding food. Therefore, the frequency of pauses and their duration per obser- vation period are indices of the time spent scanning the environment. A BASIC computer program (available from the se- nior author) was written to act as an event recorder. Data were obtained by replaying the video tapes at normal speed and making the appropriate keystroke each time the subject's behavior changed (swimming, pausing). The program recorded intervals between key- strokes from which all variables were calculated. All observations from every trial ( 19 time-periods x 6 rep- licates x 4 treatments x 3 sizes = 1368 5-min obser- vation periods) were made by the junior author. When a selection of observation periods was reanalyzed, the differences in activity averaged 4.1% (extremes 0.1- 9.2%) of the combined mean. Data analysis Since observations within each temperature treatment and size-class followed the same individuals over time, repeated-measures multivariate analysis of variance (MANOVA) was used to identify effects on spontane- ous behavior due to fish size, temperature, and time Fuiman and Ottey: Temperature effects on behavior of young Saaenops ocellatus 25 since transfer. Each behavioral measure (activity, pause frequency, pause duration) was considered the 'trials' factor in a separate analysis. Size-class and tempera- ture treatment were grouping factors. Activity was re- stricted, by definition, to values between and 300 s. Data were expressed as a percentage of the total time- period, and an angular (arcsine) transformation was applied to satisfy the statistical requirement of MAN OVA for normality (Snedecor & Cochran 1967). Pause frequency and duration were not transformed. Further analyses focused on three questions: (1) How does behavior differ at the two acclimation tempera- tures? (2) What is the immediate effect of a transfer of 5°C on behavior (relative to fish maintained at accli- mation temperatures)? and (3) When does behavior stabilize after such a transfer? These questions were addressed by repeated-measures MANOVA on pairs of treatments within size-classes. Since temporal trends in variables were of greater interest than the mere presence of significant differences among the time-pe- riods, we examined first (linear)- through fourth (quar- tic)-order polynomial trends over time with univariate F statistics (Wilkinson 1990), in addition to testing for significant differences due to temperature treatments within each time-period. By design, fish transferred upward or downward ex- perienced a temperature change, but control fish did not. However, behavior of all fish could be expected to vary with time, due to effects of handling, hunger, or circadian rhythms. Effects of handling should dimin- ish with time since transfer, but hunger should in- crease over time and have a stronger influence on be- havior of smaller fish. Therefore, behavior of fish in transfers was compared with that of control fish from the same acclimation temperature to answer questions (2) and (3). Specifically, upward transfers were com- pared with the low controls, and downward transfers with the high controls, to examine changes in behav- ior relative to acclimation levels. Figures 2, 4, and 6 depict the effects of thermal transfers as differences between means of six fish in the transfer and control treatments, but statistical tests were based on results for individual fish. Results Spontaneous behavior was composed of conspicuous periods of swimming activity interspersed with pauses, which were sometimes quite long. Behavior of medium and large fish was grossly similar, their transitions from active swimming to pausing were gradual, in part because of passive coasting. At the larger sizes, swim- ming involved forward movements generated by the caudal and pectoral fins and complex maneuvers us- ing sculling motions of the pectoral fins alone. Move- ments of small fish were less fluid. Their small size and low velocities prevented them from appreciable coasting, so their motion stopped abruptly when pro- pulsive strokes of the caudal region ceased. Activity The proportion of time in active swimming was high for all sizes, usually >65%. However, there were sig- nificant differences in activity among the three size- classes, among the four temperature treatments, and within individuals over the course of the experiments. When the four temperature treatments were exam- ined separately, significant differences among size- classes were found only for downward transfers and high controls (Table 1). Activity varied significantly with time during ex- periments within most temperature treatments for small and medium fish (Table 1). Large fish showed no significant changes in activity with time in any of the treatments because variability among individuals was greater than at smaller sizes. Differences in activity over time constituted temporal trends for small fish in all treatments and medium fish in downward trans- fers. These trends were usually linear, but several higher-order polynomials were significant for down- ward transfers (Table 1). In subsequent comparisons of the different temperature treatments, we examined each size-class separately because of the highly sig- nificant differences among sizes. Controls Activity was generally greater at 26° C for all sizes offish (Fig. 1). This difference was significant for small and medium fish, but the attained signifi- cance, P, for large fish was slightly beyond the crite- rion of a=0.05 (Table 1). Mean activity followed a monotonic trend with time at both temperatures for all sizes (Fig. 1). Small fish were most active immediately after transfer, while medium fish were least active at that time. Activity levels generally became stable within approximately 4h. Mean differences (±SD) between the two control levels after 4h were 40.0 (±12.9), 49.1 (±26.6), and 45.7 (±41.5) s of swimming/5 min for small, medium, and large fish, respectively. Thus, there was remark- able similarity in the effect of acclimation tempera- ture on mean activity, but variability increased steadily with size. Significant differences between acclimation tempera- tures were common during the first 2 h of the trials, and rare thereafter (Fig. 1), suggesting that the effect of handling on activity and the rate of recovery were temperature-related. However, these effects of tempera- ture were not consistent. Activity of small fish stabi- 26 Fishery Bulletin 91(1). 1993 Significance levels (P) for F tests in repeated-measures effects significant at a=0.05. Superscripts indicate the dratic, 3=cubic, 4=quartic). Table 1 MANOVA of activity presence and order of in red drum Sciaenops ocella significant polynomial trends us. Boldface values with time ( l=linear, ndicate 2=qua- Treatment Effect Size-class All Small Medium Large All Size Treatment Time <0.001 0.001 0.001 Upward transfer Size Time 0.060 0.477 <0.001 ' 0.045 0.263 Low control Size Time 0.092 0.027 0.211 ' 0.008 0.310 Downward transfer Size Time 0.027 <0.001 0.030 234 <0.001 ' 0.856 High control Size Time 0.037 0.061 <0.001 ' 0.168 0.193 High control vs. low control Treatment Time <0.001 <0.001 0.035 0.001 0.058 0.256 Upward transfer vs. low control Treatment Time 0.031 <0.001 0.536 <0.001 0.554 0.807 Downward transfer vs. high control Treatment Time 0.016 0.001 0.043 <0.001 0.214 0.974 lized sooner at low temperatures, but activity of me- dium fish stabilized later at low temperatures. Transfers The general response to a 5° C change was an almost immediate shift in activity in the expected direction (upward transfers produced higher activity, downward transfers produced lower activity) to a level greater than the stabilized value attained after about 2.5 h (Fig. 2). In upward transfers, elevated activity persisted for 1-2 h, rising during the first 20-30 min from an implicit value of zero just prior to transfer. The response of fish to downward transfer was an almost immediate drop in activity to minimum values followed quickly by increasing activity (Fig. 2). Activity of small fish after upward transfer was significantly different from that of low controls (Table 1). Differences were concentrated in the early part of the experiments, when transferred fish were more ac- tive than controls in six of the first seven time-periods (Fig. 2). Medium and large fish did not show overall differences between upward transfers and controls (Table 11, and there were few significant differences at individual time-periods. Downward transfer also had a significant influence on activity in comparisons with high controls, but only for small and medium fish (Table 1). Transferred fish were less active than controls in the first six time- periods for small fish and in five of the first nine time- periods for medium fish (Fig. 2). Large fish did not show an overall difference between downward trans- fer and control treatments, and only a single differ- ence for individual time-periods was statistically sig- nificant (Fig. 2). Pause frequency The number of pauses during the 5-min observation periods generally decreased as fish grew (Fig. 3). Small control fish (temperatures and times combined) paused an average of 24.6 (±11.5) times in 5 min; medium and large fish paused 10.8 (±6.9) and 9.5 (±7.3) times, re- spectively. All main effects (size-class, temperature Fuiman and Ottey Temperature effects on behavior of young Sciaenops ocellatus 27 o 300 -■a -A £.__ — *- CO Time Since Transfer (h) Figure 1 Time-course for activity of young red drum Sciaenops ocellatus at two acclimation temperatures. Points represent means of six observations. Upright triangles and broken lines refer to control trials at 26° C. Inverted triangles and solid lines refer to control trials at 21°C. Filled symbols denote periods in which activity differed significantly at the two temperatures. Panels show data for small (upper), medium (middle I, and large (lower) fish. treatment, and time since transfer) had a significant influence on pause frequency. Differences among the size-classes were present within each of the treatments (Table 2). Within individual fish (all sizes combined) there were significant differences in pause frequency with time in all treatments except the downward transfers (Table 2). These differences among time-periods followed lin- ear trends in most temperature treatments for small fish. Among medium and large fish, temporal trends were less common, but curvilinear (quadratic) rela- tionships were significant in two combinations of tem- perature treatment and size-class (Table 2). Controls Pause frequency differed significantly be- tween the two control temperatures for small and large fish (Table 2). For those sizes, pause frequency was higher at 21° C (Fig. 3). Pause frequency generally increased with time since transfer, but the amount of change during the experi- ment was small for medium and large fish. Temporal trends stabilized after about 2.5 h in most treatments (Fig. 3). Trends for small fish were parallel, with a mean difference of 13.6 (±4.1) additional pauses/5 min at 21° C. Large fish had roughly parallel trends during the first 3 h, with a mean difference over that period of 7.5 (±3.0) pauses. Mean differences between control values after behavior stabilized (3h) were -12.9 (±2.0), 0.2 (±2.6), and -4.0 (±2.4) pauses/5 min for small, me- dium, and large fish, respectively. Differences between the control temperatures dur- ing individual time-periods were present throughout the experiments on small fish. For the two larger sizes, significant differences were confined to the first 2.5 h, highlighting the converging trends for those sizes. It appears that an immediate effect of handling is to depress pause frequency. The magnitude and duration of the effect is temperature-dependent. Transfers Small and large fish showed qualitatively similar responses to the 5°C temperature change, rela- tive to control values. Pause frequency was depressed by upward transfer. For the same size-classes in down- ward transfers, pause frequency increased slightly at first, then dropped to values below controls (Fig. 4). The response of medium fish to upward transfer was similar in form to the downward transfer of the other size-classes. In downward transfers their response was the same as that of low controls, producing a corrected pause frequency of zero. Variability in these apparent differences was great, and the only significant differ- ence between transferred fish and their controls was for small fish in upward transfers (Table 2). Small fish paused consistently and significantly less often after experiencing a 5°C increase (Fig. 4, Table 2). Large fish showed a similar, though smaller and non- significant, response. Mean differences between upward transfers and low controls were -6.8 (±4.3) and -4.1 (±2.8) s for small and large fish, respectively. Differ- ences at individual time-periods were rare in upward transfers, at all sizes, but the overall effect of time since transfer was always significant (Table 2). Downward transfer did not result in a significant overall effect on pause frequency, relative to controls, for small, medium, or large fish (Table 2). However, there were temporal trends in the differences between 28 Fishery Bulletin 91(1), 1993 wa 2PC -> 26°C <=> -100 Oh 200 <» * ^ 100 >» | J 1— 1 > • »— 1 1 > O -100 <3 *T3 -200 -»-^ 200 O CD i-H l—i 100 o o -X— — a— o. o o fe°_ooo o :~ — -0—s.n. — ____. o o _l I I I I I l_ --Q----C —] ' 1— —i • 1 1 ' r Op O ~0 D~D 26°C -> 21"C -100 P - 1 ' r- °a&-Q- 9 ~.~°- -O _©©- -■-?-- &--&- f o0 ,^'5° o" D " ° ° o Time Since Transfer (h) Figure 2 Time-course for the effect of an acute temperature change of 5°C on activity of young red drum Sciaenops ocellatus. Points represent differences between means for transferred fish and control fish at the acclimation temperature. Filled symbols denote periods in which activity of transferred fish differed significantly (a=0.05) from that of control fish (corrected activity*!)). Left panels show upward transfers, right panels show downward transfers. Small, medium, and large fish are presented from top to bottom. downward transfers and high controls, centered near zero, that could not be detected by an overall test for a temperature effect. Corrected pause frequency of both small and large fish in downward transfers declined throughout the experiments, relative to high controls (Fig. 4). Time had a significant effect for these sizes but not for medium fish. Pause duration This variable showed the greatest range of variation, spanning more than two orders of magnitude (note varying scales in Figs. 5&6). Small fish showed little effect of time or temperature treatment on pause du- ration, relative to effects on larger fish. At the two larger sizes, pauses were longest at the start of the experiments, stabilizing after about 2.5 h. As with the other behavioral measures, all main effects were sig- nificant (Table 3). There were significant differences in pause duration among time-periods for low controls and downward transfers (Table 3). Linear, quadratic, and cubic trends described temporal changes for small and medium fish, but no trends were found for large fish. Controls Pause duration was generally greater for fish maintained at 21° C than for those held at 26° C (Fig. 5). Largest differences occurred early in the ex- Fuiman and Ottey Temperature effects on behavior of young Sciaenops ocellatus 29 C/5 <=> o> rn t-H 40 O. 30 >> o a 20 3 cr 10 High control vs. low control Treatment Time <0.001 <0.001 0.281 0.003 0.024 0.001 Upward transfer vs. low control Treatment Time 0.021 <0.001 0.815 <0.001 0.076 0.009 Downward transfer high control vs. Treatment Time 0.979 <0.001 0.945 0.529 0.489 0.018 Activity of young red drum closely followed the tra- ditional time-course at all sizes, whether the transfer was upward or downward. The general temporal trends in activity had similar shapes and values within treat- ments. Similar experiments conducted on a single size Cunderyearlings') of Atlantic salmon also showed that spontaneous activity followed the overshoot model for adaptation (Peterson & Anderson 1969). Interestingly, activity of salmon did not decrease dramatically in downward transfers, as predicted. Rather, the time- course for activity followed trends similar to those re- sulting from temperature increases. Peak activity in the overshoot period correlated with the rate of tem- perature change, rather than the magnitude or direc- tion of the change. Also, peak activity for salmon trans- ferred to 12° C from 6°C was considerably lower than that for fish undergoing a similar downward transfer from 18° C. Our data show a similar relationship be- tween the direction of transfer and mean peak (posi- tive or negative) activity during the overshoot period. Upward transfers resulted in lesser overshoots of ac- tivity than downward transfers, and overshoot periods were later and prolonged. The differences were not as great as those found by Peterson & Anderson (1969) for salmon, probably because of the smaller differences between acclimation temperatures in our experiments (5° vs. 12°C). To our knowledge, no other investigators have ex- amined the influence of acute temperature change on pause characteristics. Pause duration followed trends that were broadly similar to the overshoot model and consistent across sizes. Variability around the trends increased with size of fish. Unlike activity, peak pause duration was essentially independent of the direction of temperature change, but there were differences among sizes. Pause frequency showed fundamentally different time-courses. These were not consistently re- lated to the direction of temperature change or fish size. Pause frequency obviously is not a good indicator of the state of thermal adaptation. Its lack of unifor- mity suggests that it is highly variable and may be influenced by numerous other factors. Acclimation differences Activity also exhibited the strongest and most consis- tent effects of different constant temperatures in young red drum. In control experiments, the pattern of activ- Fuiman and Ortey Temperature effects on behavior of young Saaenops ocellatus 31 C/3 o o cr Oh *o o o 21°C -> 26*C 3 o o L -Q--Q D o o O o -«-— ^ • I.I.I.!. ' 1 - ,^5^-^--—---°"' 1 2 3 4 S 6 7 8 25 IS 26"C -> 21"C s • *W> °o >°-£>-£L o ° o IS t°o ° O oo --0---C ^v o o 5 e -- -o—t 012345678 Time Since Transfer (h) Figure 4 Time-course for the effect of an acute temperature change of 5°C on pause frequency of young red drum Scmenops ocellatus. Annotated as in Fig. 2. ity over time followed similar trends at the two tem- peratures within each size-class, showing the same ini- tial response in activity followed by stable behavior after 2-4 h. Once stabilized, the proportion of time spent actively swimming was 21-26% greater at the higher temperature. This is equivalent to a tempera- ture coefficient (Q 10 ) of 1.5-1.6, which is similar to val- ues reported for various other whole-animal measures of fish swimming. Maximum sustainable speed of carp has a Q lft of 1.5-1.6 (Rome et al. 1984). Larval zebra danio and Atlantic herring have temperature coeffi- cients between 1.4 and 1.7 for burst distance and maxi- mum burst speed (Fuiman 1986, 1991). Larger rain- bow trout have values of 1.8 for burst distance (Webb 1978). Stabilized activity (after 4h) in the control experi- ments, combined with pause frequency, describe a tem- perature effect on the duration of active bouts between pauses. The average duration of active bouts is very nearly the ratio of total activity to pause frequency, since pause frequency is essentially equal to the num- ber of active periods. Higher values for activity at the upper temperature, accompanied by lower or equiva- lent pause frequencies, result in longer periods of ac- tivity in warmer water. This effect holds for all three size-classes, although the magnitude of the effect is greatest for small fish. All variables we evaluated relate to foraging activity, and since fish were not fed during the experiments, there should have been ample motivation for fish to 32 Fishery Bulletin 91(1). 1993 Significance levels (Pi for F tests in indicate significant effects at oc=0.05 2=quadratic, 3=cubic). Table 3 repeated-measures MANOVA of pause duration Superscripts indicate the presence and order of in red drum Scieanops ocellatus. Boldface significant polynomial trends with time ( 1= values =linear, Treatment Effect Size-class All Small Medium Large All Size Treatment Time <0.001 0.014 <0.001 Upward transfer Size Time 0.027 0.080 <0.001 0.174 0.410 Low control Size Time 0.013 <0.001 0.071 <0.001 '*> 0.420 Downward transfer Size Time 0.039 0.022 0.379 : 0.024 ' 0.729 High control Size Time 0.183 0.203 0.001 0.391 0.437 High control vs. low control Treatment Time 0.062 0.001 0.033 <0.001 0.202 0.474 Upward transfer vs. low control Treatment Time 0.134 0.009 0.224 <0.001 0.500 0.476 Downward transfer vs high control Treatment Time 0.035 0.310 0.036 0.011 0.026 0.781 search for food once the effects of handling subsided. Fish in warmer water would be expected to have a higher demand for food at any time in the experiment due to a higher metabolic rate. Further, small fish should have a higher demand than larger fish. In- creased demand for food should be met by an increase in the volume searched per unit time. The means by which a fish increases the volume searched per unit time depends on the type of search- ing employed (i.e., sweep vs. saltatory). Search vol- ume for sweep searchers is proportional to the dis- tance traveled. A sweep searcher in warmer water should swim faster, for longer periods, and/or pause less often. For saltatory searchers, search volume is directly proportional to pause frequency. Saltatory searchers can scan greater volumes by increasing both pause frequency and swimming speed between pauses. In the absence of suitable prey, their pause duration should be constant and only as long as nec- essary to scan each field completely. Our results for total activity and mean bout duration from the sta- bilized levels of control experiments follow the pre- dictions for a sweep searcher: In warmer water, ac- tivity is higher, pause frequency is shorter, and the duration of active bouts is longer. Predictions for a saltatory searcher are contradicted, since pause fre- quency is not higher in warmer water. In fact, the opposite is true for small fish, which should be most strongly affected by hunger. Conclusion Experimental conditions cannot mimic the various natural scenarios of transient temperature fluctuation. Diel temperature changes are usually gradual and pre- dictable. Yet, even when acclimated to regular circa- dian cycles, swimming behavior may be influenced by ambient temperature (Fuiman 1986). Traversing a ther- mocline is more abrupt, but it is predictable and at least partly voluntary, allowing for some degree of physiological preparation. Such vertical migrations can be beneficial to a fish's daily energy budget (Brett 1971, Wurtsbaugh & Neverman 1988), but there may Fuiman and Ottey: Temperature effects on behavior of young Saaenops ocellatus 33 ._ 1000 Q a 12 3 4 5 6 7 8 Time Since Transfer (h) Figure 5 Time-course for pause duration of young red drum Sciaeiwps ocellatus at two acclimation temperatures. Annotated as in Fig. 1. be an immediate cost in terms of locomotor efficiency. Perhaps the most hostile type of natural temperature fluctuation is exemplified by inundation of tidal pools. It is both unpredictable and abrupt, and concomitant physical changes (e.g., sound and pressure) probably add to the thermal effects on behavior. Our results show that acute temperature changes of 5°C alone engender behavioral changes that persist for about 2h. These changes may act directly on a fish's ability to forage normally, or they may be merely indicators of a generally stressed condition in which a fish could be more susceptible to predators or disease. In addition to the thermal impacts, our fish exhib- ited a handling effect which was overcome in 3-5 h. The most similar circumstance experienced by young red drum in the field occurs during stocking from hatcheries into natural waters. Young red drum, 3-40 mm in length, have been stocked into bays and estuaries routinely since 1975 (Dailey 1991). Stocked fish experience the combined effects of handling and temperature change (often more than 5°C). Simi- larly, fish used in laboratory experiments often in- cur handling and thermal stress. Our results sug- gest that even careful handling affects behavior for at least as long as a 5°C temperature change. Mini- mal handling and an acclimation period of 2-5 h would benefit young red drum during both labora- tory experiments and stocking. Acknowledgments This research was supported by a grant from the Sid W. Richardson Foundation. We are grateful to Dr. Connie R. Arnold, Janie Munoz, and Dana Allen for providing red drum eggs. Gerald R. Hoff and Dennis M. Higgs assisted in caring for the fish. Citations Bamforth, S.S. 1962 Diurnal changes in shallow aquatic habi- tats. Limnol. Oceanogr. 7:343-353. Beamish, F.W.H. 1978 Swimming capacity. In Hoar, W.S., & D.J. Randall (eds.), Fish physiology, vol. 7, p. 101- 187. Academic Press, NY. Bell, W.J. 1990 Searching behaviour: The behavioural ecology of finding resources. Chapman & Hall, NY, 400 p. Brett, J.R. 1971 Energetic responses of salmon to temperature. A study of some thermal relations in the physiology and freshwater ecology of sockeye salmon (Oncorhynchus nerka). Am. Zool. 11:99-113. Browman, H.I., & J.W. O'Brien 1992 The ontogeny of search behavior in the white crappie, Pomoxis annularis. Environ. Biol. Fish. 34:181-195. Dailey, J. 1991 Fish stocking in Texas bays: 1975-1990. Man- age. Data Ser. 54, Texas Parks Wildl. Dep., Fish. Wildl. Div., Coastal Fish Br, Austin, 38 p. Fuiman, L.A. 1986 Burst-swimming performance of larval zebra da- nios and the effects of diel temperature fluctua- tions. Trans. Am. Fish. Soc. 115:143-148. 1991 Influence of temperature on evasive responses of Atlantic herring larvae attacked by yearling her- ring. J. Fish Biol. 39:93-102. 34 Fishery Bulletin 9 1 1 1 ), 1993 C/3 a o -10 -t— » cd 150 t-i =3 Q SO a> 75 »-i V-( O U 25 21 °C -> 26'C o„ e _2_ Q ___Q„_ 3 _l i I , 1_ i i i i o - O n ,0-°-0'°" a' _ o~ - ~o- -■=■■£, c o o i ...,.,., . 1 . 1 1 1 1 ' 1 ' T o - 3 O ojoo-e — e-o-Q_o__ c o 1 1 1 1 1 o 1 1 - 12 3 4 5 6 7 8 26°C -> 21°C •n (S • 9 £> o o V~5- ..Oo"Xr:--o-nQ-__ ,_o._^.„...-Q----!: -J . 1 . L o 1 1 1 • 1 — o o ' o o o 1 . 1 . 1 1 12 3 4 5 6 7 8 Time Since Transfer (h) Figure 6 Time-course for the effect of an acute temperature change of 5°C on pause duration of young red drum Sciaenops ocellatus. Annotated as in Fig. 2. Holt, G.J., C.R. Arnold, & CM. Riley 1990 Intensive culture of larval and post larval red drum. In Chamberlain, G.W., R.J. Miget, M.G. Haby (eds.), Red drum aquaculture, p. 53-56. Texas A&M Univ. Sea Grant Coll. Prog., Galveston. Holt, J., R. Godbout, & C.R. Arnold 1981 Effects of temperature and salinity on egg hatch- ing and larval survival of red drum, Sciaenops oeellata. Fish. Bull., U.S. 79:569-573. Janssen, J. 1982 Comparison of searching behavior for zooplank- ton in an obligate planktivore, blueback herring iAlosa aestivalis) and a facultative planktivore, bluegill (Lepomis machrochirus). Can. J. Fish. Aquat. Sci. 39:1649-1654. Kinne, O. 1963 The effects of temperature and salinity on ma- rine and brackish water animals. I. Tempera- ture. Oceanogr. Mar. Biol. Annu. Rev. 1:301-340. Laing, J. 1938 Host-finding by insect parasites. I. Observations on the finding of hosts by Alysia manducator, Mor- moniella vitripennis, and Trichogramma evan- escens. J. Anim. Ecol. 6:298-317. Montgomery, J.C., & JA. MacDonald 1990 Effects of temperature on nervous system: Impli- cations for behavioral performance. Am. J. Physiol. 259:R191-R196. O'Brien, W.J., B.I. Evans, & G.L. Howick 1986 A new view of the predation cycle of a plank- tivorous fish, white crappie (Pomoxis annularis). Can. J. Fish. Aquat. Sci. 43:1894-1899. O'Brien, W.J., B.I. Evans, & H.I. Browman 1989 Flexible search tactics and efficient foraging in saltatory searching animals. Oecologia 80:100-110. Peterson, R.H., & J.M. Anderson 1969 Influence of temperature change on spontaneous locomotor activity and oxygen consumption of Atlan- Fuiman and Ottey Temperature effects on behavior of young Saaenops ocellatus 35 tic salmon, Salmo salar, acclimated to two tempera- tures. J. Fish. Res. Board Can. 26:93-109. Precht, H., H. Laudien, & B. Havteen 1973 The normal temperature range. In Precht, H., J. Christophersen, H.,Hensel, & W. Larcher (eds.), Temperature and life, p. 302-399. Springer-Verlag, Berlin. Prosser, C.L. 1964 Perspectives of adaptation: Theoretical as- pects. In Dill, D.B., E.F. Adolph, & C. G. Wilber (eds.), Handbook of physiology. Sec. 4: Adaptation to the en- vironment, p. 11-25. Am. Physiol. Soc, Wash. D.C. Rome, L.C., P.T. Loughna, & G. Goldspink 1984 Muscle fiber recruitment as a function of swim speed and muscle temperature in carp. Am. J. Physiol. 247:R272-R279. Rosenthal, H. 1969 Untersuchungen uber das Beutefangverhalten bei Larven des Herings Clupea harengus. Mar. Biol. (Berl.) 3:208-221. Rosenthal, H., & G. Hempel 1970 Experimental studies in feeding and food require- ments of herring larvae (Clupea harengus L. ). In Steele, J.H. (ed.l, Marine food chains, p. 344- 364. Univ. Calif, Berkeley. Smid, P., & K. Priban 1978 Microclimate in fishpond littoral ecosystems. In Dykyjova, D., & J. Kvet (eds.), Pond littoral ecosys- tems, structure and functioning, p. 104-112. Springer- Verlag, Berlin. Snedecor, G.W., & W.G. Cochran 1967 Statistical methods. Iowa State Univ. Press, Ames, 693 p. Webb, P.W. 1978 Effects of temperature on fast-start performance of rainbow trout {Salmo gairdneri). J. Fish. Res. Board Can. 35:1417-1422. Wilkinson, L. 1990 SYSTAT: The system for statistics. SYSTATInc, Evanston IL, 677 p. Wurtsbaugh, WA, & D. Neverman 1988 Post-feeding thermotaxis and daily vertical mi- gration in a larval fish. Nature (Lond.) 333:846-848. AbStraCt.-Ichthyoplankton were sampled weekly in Auke Bay, south- eastern Alaska, from March or early April through June, 1986-89. The spring primary production bloom oc- curred in April, and was followed in May by the annual maximum in herbivorous copepods. Each year, the five most-abundant fish larvae were osmerids. Pacific sandlance Ammodytes hexapterus, walleye pollock Theragra chalcogramma, fiathead sole Hippoglossoid.es elasso- don, and rock sole Pleuronectes bilineatus. Each species tended to occur at the same time every year, and could be categorized either as synchronous species that were present at the time copepod abun- dance was maximized, or early spe- cies that were most abundant be- fore the spring phytoplankton bloom. Pacific sandlance and rock sole lar- vae always reached maximum abun- dance prior to the spring bloom, whereas larvae of walleye pollock, fiathead sole, and osmerids were most abundant at the time of the copepod maximum. Physical and bi- otic conditions experienced by early and synchronous larvae differ mark- edly, suggesting that survival through early life history is deter- mined by different processes in the two groups. Abundance patterns of marine fish larvae during spring in a southeastern Alaskan bay Lewis Haldorson Marc Pritchett David Sterritt John Watts School of Fisheries and Ocean Sciences, University of Alaska 1 1 1 20 Glacier Highway. Juneau, Alaska 99801 Manuscript accepted 20 August 1992. Fishery Bulletin. U.S. 91:36-44 ( 1993 1. Fluctuation in recruitment to ex- ploited fish populations remains a central problem in marine fish man- agement. There are indications that much of the variation in year-class abundance in marine fish populations results from processes and events in planktonic early-life-history stages (Houde 1987, Pepin & Myers 1991). Interannual variation in survival through egg and larval life stages is undoubtedly determined by mul- tiple and interacting mechanisms; however, timing of reproduction has often been implicated as a factor contributing to the success or fail- ure of year-classes. For example, Hjort's (1914) critical-period hypoth- esis and Cushing's (1975) match- mismatch hypothesis describe the importance of synchrony between production of larval fishes and their planktonic prey. In subarctic regions, nearshore marine ecosystems display marked seasonal changes in physical and bi- otic conditions (Smetacek et al. 1984). In such systems, timing of reproduction may be extremely im- portant, as conditions that result in high survival through planktonic life-history stages may be transi- tory. A dominant feature in the an- nual subarctic nearshore production cycle is the spring phytoplankton bloom, an event that contributes much of the annual production (Smetacek et al. 1984). The phyto- plankton bloom is followed by the herbivorous copepod maximum (Smetecek et al. 1984), a period of 1-2 months that produces an an- nual optimum in foraging conditions for those larval fishes that feed on copepod nauplii. Water temperature and predator density may also de- termine survival of fish eggs and larvae (Houde 1987) and could con- stitute important constraints on timing of reproduction. In this paper we report the re- sults of a 4-year investigation of lar- val fishes in a coastal subarctic ma- rine ecosystem. Our observations describe when larvae of some north- east Pacific Ocean fish species oc- cur relative to the spring produc- tion cycle. The study was part of an interdisciplinary project (AP- PRISE, Association of Primary Pro- duction and Recruitment in a Sub- arctic Ecosystem) that provided a detailed description of the physical and biotic environment present dur- ing the period from late winter through early summer. Study area The study was conducted in Auke Bay (lat. 58 22' N, long. 134 40' W), southeast Alaska. (Fig. 1). The 16knr Bay varies in depth from 40 to 60 m. Physical conditions in Auke Bay are typical of nearshore subarctic marine 36 Haldorson et al.: Spring abundance patterns of marine fish larvae 37 Figure 1 Auke Bay study area in southeast Alaska, and location of the Auke Bay Monitor (ABM I station. Materials and methods Fish larvae were collected in Auke Bay at a station designated ABM (Fig. 1) from mid-March or early April through mid-June, 1986-89. The ABM station was selected be- cause it had been used in previ- ous studies (summarized in Coyle & Shirley 1990). Samples were collected on the same day each week between 0800 and 1300, with the exception on the second week of April 1986 (Fig. 4). Each week five replicate samples were collected with a 1 nr Tucker trawl constructed of 505 |x mesh and fitted with a digital flowmeter in the middle of the net opening. Each replicate was collected at the ABM station by towing the net in a double-oblique trajectory to a depth of 30-35 m. The vessel systems. The water column is isothermal until April, when surface warming and increased freshwater run- off contribute to formation of a pycnocline (Bruce et al. 1977, Ziemann et al. 1991). In the 4 years of this study water temperature prior to stratification var- ied from 3 to 5 C, was colder in 1986 and 1989, and warmer in 1988 (Fig. 2; data from Ziemann et al. 1990). Stratification, indicated by diverging tempera- tures at 5 and 20m, began in April (Fig. 2; data from Ziemann et al. 1990). Auke Bay exhibits a typical subarctic annual pro- duction cycle (Williamson 1978, Ziemann et al. 1991). The spring phytoplankton bloom began in early April, 1986-89 (Fig. 3; data from Ziemann et al. 1990), in response to several consecutive days of relatively high light levels (Ziemann et al. 1991). Chlorophyll biomass peaked in late April or early May each year, with a subsequent decline resulting from nutrient limitation (Ziemann et al. 1991). The herbivorous copepod maximum began 2-4 weeks after the spring phytoplankton bloom (Fig. 3; data from Coyle & Paul 1990), with Pseudocalanus spp. copepods dominant in every year (Coyle et al. 1990). Copepod nauplii in the size-ranges consumed by larval fishes were typi- cally in low density prior to the herbivorous copepod maximum and reached maximum density in May, although there was considerable interannual varia- tion in nauplii density during the period of peak abundance (Paul et al. 1991). 12 - 1986 10 - 8- 6- ^ m 0m 4- -•-e 3 *^ ^^ r* 1987 10 - J knv 8- / "^"•"-rj 6- 1 1 -§■> _ j i' v- 4 - »=*= ^r^^^* 1988 o ,0 " o """ 8- 0. E 6- Ui 1- 4- — B - * 1989 10 - A. J\ k/ 8- 1 JS sf V 6- J M^^~ I-"* - * 4- •-*— &&Z-+ v^ 60 9° 120 150 18 MARCH APRIL MAY JUNE Figure 2 Spring and early-summer water temperatures in Auke Bay, Alaska, at 5 and 20 m ( 1986-89; data from Ziemann et al. 1990). 38 Fishery Bulletin 91 [1), 1993 800 - 600- 400 - 1986 -5000 -4000 -3000 ■2000 — - o — Chi oroohvll udocslanui .....o---o-' COPEPOD DENSITY (no./m 3 ) 1988 .o\/« "■o---o* o- A -o 1989 600 - r\ •4000 400 - 200- , o-o~.« X? r*-* -3000 -2000 -1000 -0 6 ' 90 120 150 It MARCH APHIL MAY j UNE Figure 3 Chlorophyll concentrations and densities of Pseudocalanus spp. in Auke Bay during spring and early summer 1986-89 (chlorophyll data from Ziemannn et al. 1990, Pseudocalanus data from Coyle & Paul 1990). LU o 40 - 1986 30 - 20- 10 - n - , — • \f W V^J 1987 V 20 -| 1988 15- 10 - 5 - Figure 4 Mean densities of all fish larvae, excluding osmerids, in Auke Bay, Alaska during spring and early summer, 1986-89. Error bars are 1SE; where no bars are visible, they are obscured by the point symbol. speed was about 1.5 kn. Each tow lasted 7-8 min, and volume filtered was typically around 300 m 3 . Volume filtered per tow was very similar among years. Tows were collected on reciprocal compass courses set at 90" to the wind direction. Fish larvae were removed from each replicate and enumerated by species, with the exception of osmerids, agonids, cottids, and cyclopterids, which were identi- fied only to family. Osmerids were not identified to species because larvae of eulachon Thaleichthys paciftcus and capelin Mallotus villosus, the two spe- cies common in the Auke Bay area, are very similar. The other three families (Agonidae, Cottidae, and Cyclopteridae) lack comprehensive identification guides to the species level. Mean densities of each taxon were calculated as the number/m 2 of surface. Results Total number of larvae collected annually ranged from 6087 in 1988 to 18,655 in 1986 (Table 1). Most of the interannual variation was due to differences in catches of osmerids. The five most-abundant taxa in all years were osmerids, Pacific sandlance Ammodytes hexapterus, walleye pollock Theragra chalcogramma, flathead sole Hippoglossoides elassodon, and rock sole Pleuronectes bilineatus. We did not include cottids in this summary, as they include at least eight species, none of which was exceptionally abundant; whereas the osmerids were very abundant, and included two species. Total abundance of all larvae, excluding osmerids, peaked in March or early April of 1986-88 and in May 1989 (Fig. 4). Osmerids were excluded from total abun- dance estimates because in 1986 and 1987 their abun- dance obscured patterns associated with seasonal cycles of other species. In all years, osmerid abundance peaked from late May through June (Fig. 5). Such consistency in time of appearance in Auke Bay was typical of most species. Larvae present in late March or early April were well in advance of either the spring phytoplankton bloom or the herbivorous copepod maximum. These early peaks in abundance were due primarily to high numbers of Pacific sandlance and rock sole (Figs. 6, Haldorson et al : Spring abundance patterns of marine fish larvae 39 Table 1 Taxa of larval fishes collected in Auke Bay Alaska in the spring. 1986-89, with tota number collected land rank order, in parentheses) of the five most frequently collected taxa in each year. The num ber of weekly sam Dies, each consisting of five replicates is indicated below each year. 1986 1987 1988 1989 12 16 14 13 Clupeidae Clupea harengus 67 125 128 283 Osmeridae 11975 (1) 9704 (1) 336 (4) 1006 (3) Ammodytidae Ammodytes hexapterus 1926 (21 2829 (2) 2295 (1) 611 (4) Bathylagidae Leuroglossus schmidti 121 401 155 326 Gadidae Theragra chalcogramma 1696 (3) 856 (3) 1453 (2) 4618 ID Gadus macrocephalus 4 2 Stichaeidae Anoplarchus msignis 221 108 103 114 Lumpenella longirostris 108 63 110 210 Lumpenus sagitta 175 178 75 150 Ptilichthyidae Ptilichthys goodei 5 3 7 6 Cryptacanthodidae 17 35 7 18 Cottidae* 541 531 306 860 Agonidae 396 404 216 334 Cyclopteridae 26 34 14 31 Pleuronectidae Hippoglossoides elassodon 474 (4) 428 (5) 303 (5) 2741 (2) Pleuronectes bilineatus 409 (5) 522 14) 406 (3) 449 (5) Pleuronectes asper 146 104 1 5 Pleuronectes vetulus 84 131 24 2 Platiehthys stellatus 171 142 85 427 Psettiehthys melanostictus 71 93 22 122 Unidentified 26 33 41 47 Total species. 18655 16724 6087 12360 *Not ranked; included at least 8 7). Two less abundant species — longsnout prickleback Lumpenella longirostris and slender cockscomb Anoplarchus insignis — also appeared early in 1987 and 1988, but had maximum density in May of 1989. Two of the most abundant species, walleye pollock and flathead sole, consistently appeared in May (Figs. 8, 9) and were well synchronized with maximum den- sity of copepods. Less common larvae that also tended to reach maximum density in May were starry flounder Platiehthys stellatus and poachers (agonids) (Figs. 10, 11). Discussion It has been observed that many fish larvae occur in approximate synchrony with maximum zooplankton densities (Sherman et al. 1981 and 1984, Townsend 1984, Jenkins 1986). The strategy of synchronizing pro- duction of larvae to high abundance of prey has obvi- ous adaptive value, and is the prerequisite of high recruitment in Cushing's (1975) match-mismatch hypothsis. Fishes with this strategy were termed "synchronous" by Sherman et al. (1984). An alternate 40 Fishery Bulletin 91(1). 1993 ioo-| T 80- 1986 60- 40- 20- 80-1 • 60- 1967 A 40 / \ — 20- J v^ M ^-^ ^— ■> E o - O c ~* 3 - 1988 ) I DENSITY 4 \ 8- 6- 1989 r\ 4 - / \ 2- / w s, 60 90 120 150 180 MARCH APRIL MAY JUNE Figure 5 Mean densities of osmerid larvae in Auke Bay, Alaska during spring and early summer, 1986-69. Error bars are 1SE; where no bars are visible, they are obscured by the point symbol. 40 - 1986 30 - 20 - 10 - 20- 1987 10 - Oi O 20 - C 1988 DENSITY / V 3- I 1989 2 h \ ^A 90 120 150 1 MARCH A p R | L MAY JUNE )0 Figure 6 Mean densities of sandlance Ammodytes hexapterus larvae in Auke Bay, Alaska during spring and early summer, 1986-89. Error bars are 1SE; where no bars are visible, they are ob- scured by the point symbol. / k ^A 1986 4 - 3 - 2- _J^ /V 1987 3 2- 1 1988 3- 2- 0- -J \a — O- A 1989 - Figure 7 Mean densities of rock sole Pleuronectes bilineatus larvae in Auke Bay, Alaska during spring and early summer, 1986-89. Error bars are 1SE; where no bars are visible, they are ob- scured by the point symbol. strategy, characterized by prolonged production of lar- vae, has been termed "bet-hedging" (Lambert & Ware 1984) or "ubiquitous" (Sherman et al. 1984). This strat- egy is described as adaptive in situations where prey availability is unpredictable. In Auke Bay, fish species reproducing in the spring appear to follow two strategies: One group, typified by walleye pollock and flathead sole, is clearly synchro- nous, in the sense described above; whereas Pacific sandlance and rock sole are examples of species that could be termed "early" The early group can be de- fined as those species that produce their larvae prior to the peak in the spring phytoplankton bloom (before mid-April in Auke Bay). It is possible that early spe- cies in Auke Bay are following the "bet hedging" strat- egy discussed above, as they could have been produc- ing larvae throughout the winter. In that case our sampling would have coincided with the end of their production period. From mid-March through June, conditions in Auke Bay are rapidly changing as the system passes through two of the production phases — spring phytoplankton bloom and herbivorous copepod maximum — that typify nearshore subarctic marine environments (Smetacek et al. 1984). In the pre-bloom period, the water column Haldorson et al Spring abundance patterns of marine fish larvae 41 30- 1989 /N 20- A 10- 0- — n p m — °"~^l- n ii- J s s. Figure 8 Mean densities of walleye pollock Theragra chalcogramma larvae in Auke Bay, Alaska during spring and early summer, 1986-89. Error bars are 1SE; where no bars are visible, they are obscured by the point symbol. > en Figure 9 Mean densities of flathead sole Hippoglossoid.es elassodon lar- vae in Auke Bay, Alaska during spring and early summer. 1986-89. Error bars are 1SE; where no bars are visible, they are obscured by the point symbol. z Q Figure 10 Mean densities of starry flounder Platichthys stellatus larvae in Auke Bay, Alaska during spring and early summer, 1986- 89. Error bars are 1SE; where no bars are visible, they are obscured by the point symbol. z LU Q Figure 1 1 Mean densities of agonid larvae in Auke Bay, Alaska dur- ing spring and early summer, 1986-89. Error bars are 1SE; where no bars are visible, they are obscured by the point symbol. 42 Fishery Bulletin 91(1), 1993 is well mixed and uniformly cold, with mean tempera- ture below 5 C. With onset of the phytoplankton bloom, the Bay stratifies, with rapid warming of the mixed layer to over 10°C by June (Bruce et al. 1977, Ziemann et al. 1991). Zooplankton are in low density until the end of the phytoplankton bloom, and are comprised of relatively large plankton such as overwintering copepedids and some meroplankton such as barnacle larvae (Wing & Reid 1972, Coyle & Paul 1990, Paul et al. 1991). The initiation of the herbivorous copepod maximum marks the start of a period with relatively high densities of smaller zooplankton, especially cope- pod nauplii in the size range (150-350(0.) utilized by synchronous species such as walleye pollock and flathead sole larvae (Fig. 12; data from Paul et al. 1991). It seems clear that fish larvae hatched prior to the phytoplankton bloom must be adapted to a very different set of conditions than those that occur syn- chronously with the herbivorous copepod maximum. Larvae spawned in winter apparently employ vari- ous foraging strategies while utilizing similar ener- getic principles. Bailey (1982) concluded that Pacific hake larvae use energy slowly and grow slowly while passively hunting large prey. In Long Island Sound NY, larvae of American sandlance Ammodytes americanus hatched in winter are herbivores that sur- MI - 40- 1986 A 30H y\ J 20 - X \ / / \ / 10- a ^°~ a ~ < \ bOi 40 - 1987 30- 20- 10- 50- 40 - 1988 30i 20 - 10- 0- i , . . , bO - 40 - 1989 30 - 20 - 10- MARCH Figure 1 2 Densities of copepod nauplii ( 150-350|j.) averaged over depths of 5, 10, and 15 m in Auke Bay, Alaska during spring. 1986- 89 (data from Paul et al. 1991). vive nonfeeding in cold temperatures until the spring phytoplankton bloom commences (Monteleone et al. 1987). In both of these strategies, larvae survive by relying on their ability to withstand long periods of low food availability, largely as a result of low meta- bolic rates at cold temperatures. Sherman et al. (1984) identified the synchronous strategy as a major adaptive tactic for many north- west Atlantic Ocean fishes. In Auke Bay, several taxa including osmerids, walleye pollock, and flathead sole consistently appeared at about the time copepod abun- dance was maximized. Although prey is relatively plen- tiful during the herbivorous copepod maximum, the numbers of predatory invertebrates are rising (Smetacek et al. 1984, Coyle & Paul 1990); conse- quently, mortality rates of fish eggs and larvae are probably rising rapidly. Higher temperatures during this period may also limit the time available for larvae to encounter suitable conditions by causing high meta- bolic rates and more rapid depletion of energy reserves. These species are probably most sensitive to interannual variation in the production cycle and may demonstrate the type of recruitment fluctuation de- scribed by the match-mismatch hypothesis (Cushing 1975) or the critical-period hypothesis (Hjort 1914). Patterns of abundance observed in Auke Bay could result from hatching of larvae in the Bay or from ad- vective events that carried larvae into the Bay from other areas. Auke Bay has only one deep (>20m) en- trance, just east of Coghlan Island (Fig. 1). Water in the Bay is quite persistent, with an average replace- ment time of the water mass once a month or longer during the March-June period (Nebert 1990). In other studies of growth, length-frequencies were determined for walleye pollock and flathead sole larvae (Haldorson et al. 1989, 1990). Cohorts of both species first ap- peared in Auke Bay as small larvae comprising length- frequency modes that could be followed for several weeks. Over the 4 years of the study there were very few cases in which length-frequency modes occurred that could not be identified in preceding weeks. Conse- quently, we conclude that most of the larvae sampled in this study originated from hatching within or near the Bay, with possible exception of osmerids. The osmerids in our study are most likely eulachon Thaleichthys pacificus and capelin Mallotus villosus. Eulachon is an anadromous species that spawns dem- ersal, adhesive eggs in rivers. After hatching, the lar- vae are carried into nearby marine waters. The most likely source of eulachon larvae in Auke Bay is the Mendenhall River, a glacier-fed stream about 2 km east of Auke Bay. The fresh and turbid waters from the Mendenhall River form a surface lens that projects out into nearby Fritz Cove and often intrudes into the eastern edge of Auke Bay through a narrow passage Haldorson et al.: Spring abundance patterns of marine fish larvae 43 northeast of Spuhn Island (Fig. 1). In 1987 we studied the depth distribution of fish larvae in Auke Bay and found that osmerids were always concentrated above the pycnocline and moved to the surface at night (unpubl. data). Most other species were found at or below the pycnocline and tended to move deeper at night. Therefore, interannual and seasonal variation in osmerid abundance may reflect variation in the amounts of river water reaching Auke Bay. Among the 4 years of the study, 1989 was distin- guished by relatively high densities of larvae during May. Walleye pollock and flathead sole were mark- edly more abundant in 1989 than in the previous 3 years, as were less-common synchronous species such as starry flounder. This increase could have resulted from increased egg production in the bay or from exceptionally high survival of eggs and larvae. We have no data on density of fish eggs in Auke Bay; however, in 1989 invertebrate predators were present in the lowest abundance observed in the 4-year study (Coyle & Paul 1990). It seems possible that reduced predatory mortality contributed to the exceptionally high densities offish larvae that occurred in May of 1989. However, 1986 also had relatively high abun- dances of fish larvae in May, and did not have re- duced numbers of invertebrate predators (Coyle & Paul 1990). Acknowledgments This study was part of the APPRISE program, a col- laborative research effort by the School of Fisheries and Ocean Sciences (University of Alaska, Fairbanks) and the Oceanic Institute (Waimanalo, Hawaii). The program was supported by Contract no. NA-85-ABH- 00022 from the US Department of Commerce, National Oceanic and Atmospheric Administration. Support pro- vided by staff at the Juneau Center for Fisheries and Ocean Sciences was essential to completion of this study. In particular, we appreciate assistance provided in the laboratory by Amanda Arra, Karen Besser, and Lynette McNutt, and in the field by Donald Erickson and Russell Sandstrom. Citations Bailey, K. 1982 The early life history of the Pacific hake Merluccius productus. Fish. Bull, U.S. 80:589-598. Bruce, H.E., D.R. McLain, & B.L. Wing 1977 Annual physical and chemical oceanographic cycles of Auke Bay, southeastern Alaska. NOAA Tech. Rep. NMFS SSRF-712, 11 p. Coyle, K.O., & A.J. Paul 1990 Interannual variations in zooplankton population and biomass during the spring bloom in an Alaskan subarctic embayment. In Ziemann, D.A., & K.W. Fulton-Bennett (eds.), APPRISE— Interannual Vari- ability and Fisheries Recruitment, p. 179-228. The Oceanic Institute, Honolulu. Coyle, K.O., & T.C. Shirley 1990 A review of fisheries and oceanographic re- search in Auke Bay, Alaska and vicinity, 1966- 1985. In Ziemann, D.A., & K.W. Fulton-Bennett (eds.), APPRISE— Interannual variability and fisheries recruitment, p. 1-74. The Oceanic Insti- tute, Honolulu. Coyle, K.O., A.J. Paul, & D.A. Ziemann 1990 Copepod populations during the spring bloom in an Alaskan subarctic embayment. J. Plankton Res. 12:759-797. Cushing, D.H. 1975 Marine ecology and fisheries. Cambridge Univ. Press, 278 p. Haldorson, L., A.J. Paul, D. Sterritt, & J. Watts 1989 Annual and seasonal variation in growth of lar- val walleye pollock and flathead sole in a southeast Alaskan Bay. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 191:220-225. Haldorson, L., M. Pritchett, D. Sterritt, & J. Watts 1990 Interannual variation in the recruitment poten- tial of larval fishes in Auke Bay, Alaska. In Ziemann, D.A., & K.W. Fulton-Bennett (eds.), APPRISE— Interannual variability and fisheries recruitment, p. 319-356. The Oceanic Institute, Honolulu. Hjort, J. 1914 Fluctuations in the great fisheries of northern Europe viewed in the light of biological re- search. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 20:1-228. Houde, E.D. 1987 Fish early life dynamics and recruitment vari- ability. In Hoyt, R.D. (ed.), 10th annual larval fish conference, p. 17-29. Am. Fish. Soc. Symp. 2, Bethesda. Jenkins, G.P. 1986 Composition, seasonality and distribution of ichthyoplankton in Port Phillip Bay, Victoria. Aust. J. Mar. Freshwater Res. 37:507-520. Lambert, T.C, & D.M. Ware 1984 Reproductive strategies of demersal and pelagic spawning fish. Can. J. Fish. Aquat. Sci. 41:1565- 1569. Monteleone, D.M., W.T. Peterson, & G.C. Williams 1987 Interannual fluctuations in the density of sand lance. Ammodytes americanus, larvae in Long Island Sound, 1951-1983. Estuaries 10:246-254. Nebert, D.L. 1990 Marine circulation in Auke Bay, Alaska. In Ziemann, DA., & K.W. Fulton-Bennett (eds.), AP- PRISE — Interannual variability and fisheries recruit- ment, p. 75-98. The Oceanic Institute, Honolulu. 44 Fishery Bulletin 91(1), 1993 Paul, A.J., K.O. Coyle, & L. Haldorson 1991 Interannual variations in copepod prey of larval fish in an Alaskan bay. ICES J. Mar. Sci. 48:157- 165. Pepin, P., & RA. Myers 1991 Significance of egg and larval size to recruitment variability of temperate marine fish. Can. J. Fish. Aquat. Sci. 48:1820-1828. Sherman, K., R. Maurer, R. Byron, & J. Green 1981 Relationship between larval fish communities and zooplankton prey species in an offshore spawning ground. Rapp. R-V. Reun. Cons. Int. Explor. Mer 178:289-294. Sherman, K., W. Smith, W. Morse, M. Berman, J. Green, & L. Ejsymont 1984 Spawning strategies of fishes in relation to circu- lation, phytoplankton production and pulses in zoop- lankton off the northeast United States. Mar. Ecol. Prog. Sen 18:1-19. Smetacek, V., B. von Bodungen, B. Knoppers, R. Peinert, F. Pollehne, P. Stegmann, & B. Zeitzschel 1984 Seasonal stages characterizing the annual cycle of an inshore pelagic system. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 183:126-135. Townsend, D.W. 1984 Comparison of inshore zooplankton and ichthyo- plankton populations in the Gulf of Maine. Mar. Ecol. Prog. Ser. 15:79-90. Williamson, R.S. 1978 Phytoplankton and productivity in Auke Bay, Alaska. Manuscr. rep. MR-F 157, Auke Bay Lab, NMFS Alaska Fish. Sci. Cent., Auke Bay AK, 15 p. Wing, B.L., & G.M. Reid 1972 Surface zooplankton from Auke Bay and vicinity, southeastern Alaska, August 1962 to January 1964. Data rep. 72, Auke Bay Lab., NMFS Alaska Fish. Sci. Cent., Auke Bay AK, 764 p. Ziemann, D.A., L.D. Conquest, K.W. Fulton-Bennett, & P.K. Bienfang 1990 Interannual variability in the physical environ- ment of Auke Bay, Alaska. In Ziemann, D.A., & K.W. Fulton-Bennett (eds.), APPRISE— Interannual vari- ability and fisheries recruitment, p. 99-128. The Oce- anic Institute, Honolulu. Ziemann, D.A., L.D. Conquest, M. Olaizola, & P.K. Bienfang 1991 Interannual variability in the spring phytoplank- ton bloom in Auke Bay, Alaska. Mar. Biol. (Berl.) 109:321-334. AbStraCt-Tautog Tautoga onitis are gaining popularity in Virginia's coastal waters as a recreational and food fish. Adult tautog are season- ally abundant on inshore hard- bottom habitats (l-10m) and inhabit offshore areas ( 10-75 ml year-round. Juveniles, especially newly-settled recruits, inhabit vegetated areas in shallow water (usually 25 cm, >3-4 yr old) migrate to offshore, overwintering areas. Adult tautog re- turn from offshore wintering areas to inshore spawning sites with onset of increasing temperatures in the spring (Chenoweth 1963, Cooper 1966, Stolgitis 1970, Olla et al. 1974, Briggs 1977, Olla et al. 1979). In Vir- ginia, tautog populations on near- shore sites also undergo seasonal fluctuations. However, seasonal movements of tautog in Virginia's coastal waters may not be as well defined as the migration patterns noted for tautog in more northern ar- eas. In our study area, not all tautog migrated to inshore areas during 45 46 Fishery Bulletin 91(1), 1993 springtime spawning periods. Tautog in peak spawn- ing condition were collected on both inshore and off- shore sites throughout the spring-summer period (late April-early August). Additionally, it was not unusual to observe large fish (>25cm) at inshore sites during winter, especially in deeper areas. In more northern areas, other researchers (Olla & Samet 1977, Eklund & Targett 1990) have also noted that some adult tautog in the population remain off- shore throughout the year. Tautog are rapidly gaining popularity as a prized food and sport fish with recreational anglers and spearfishermen in Virginia waters (Bain 1984, Arrington 1985). Recreational angling for this species in Virginia has received increased interest since the recent capture of a world record tautog (24 lb, 10.89 kg) by an angler fishing off the eastern shore of Virginia (IGFA 1990). This increasing popularity has also been reflected in the number of awards issued annually to recreational fishermen by the Virginia Salt Water Fish- ing Tournament for outstanding catches [tautog weigh- ing 4.1kg (9 lbs) or more]. Awards for outstanding catches of tautog increased from mean values of 122/ yr for the period 1976-80 to 282/yr for 1981-86 (C. Bain, Virginia Saltwater Fishing Tournament, VMRC, Virginia Beach VA 23451, pers. commun. ). Most re- cently, however, citations for outstanding catches have decreased to 106/yr (range 91-130/yr) for 1987-91. Commercial catches from 1922 (Hildebrand & Schroeder 1928) to the present (E. Barth, Deputy Chief, Fish. Manage. Div., VMRC, Newport News VA 23607, pers. commun. 3 Jan. 1991) show little annual varia- tion in weight of reported catches in landings for this species. Reported commercial landings of tautog in Vir- ginia from 1973 to 1988, for example, ranged from 234 to 3586 lb/yr (x = 1840 lb/yr). However, these landings are insignificant compared with unreported catches by commercial and recreational rod-and-reel fishermen. According to statistics compiled by the Marine Recre- ational Fishery Statistics Survey conducted by the Na- tional Marine Fisheries Service (NMFS), estimated rec- reational catches of tautog in the mid-Atlantic Bight from 1979 to 1989 ranged from 70,000 (1982) to 815,000 (1984) fish/yr (x=383,200 fish/yr). In 1985, landings of tautog taken by recreational anglers in Virginia alone were estimated to be 743,600 lb, representing 3.6% of the total estimated poundage for fishes taken by recre- ational fishermen in Virginia (VMRC 1985). Aspects of the tautog's biology, including its associa- tion with hardbottom areas (Bigelow & Schroeder 1953, Cooper 1967, Olla et al. 1974, 1977, 1979), which are limited and generally discontinuous in Virginia, and its relatively slow growth (Cooper 1967), render this species susceptible to overexploitation (Briggs 1977). This situation is further exacerbated by recent techno- logical advances in LORAN and recording depth finders used by fishermen that have simplified locating even the smallest, isolated substrates. Populations of slow- growing fish species concentrated in reef areas can be severely depleted by fishing pressure exerted by recre- ational interests (Briggs 1977, Turner et al. 1983, Manooch & Mason 1984, Matheson and Huntsman 1984, Moore & Labisky 1984, Harris & Grossman 1985). Based on informal surveys of charterboat cap- tains, recent declines in citation awards for large fish, and personal field experiences, the catch-per-unit- effort of tautog has already decreased in Virginia, and the relative abundance of tautog (especially larger-sized individuals) has been detrimentally affected in the more popular fishing areas. In coastal waters of Rhode Island and New York, Cooper (1966), Olla et al. (1974), and Briggs (1977) noted a seasonal inshore-offshore spawning migration in tautog with no significant north-south component. None of these studies reported the recovery of fish from areas outside the general area in which they were tagged, indicating that little, if any, mixing of adult tautog takes place between fish inhabiting even rela- tively proximate areas (Rhode Island and New York waters). This study was undertaken to estimate age struc- ture of the population, growth, longevity, and seasonal patterns of reproduction for tautog occurring in lower Chesapeake Bay and nearby coastal waters of Virginia. Growth parameters estimated for tautog collected in Rhode Island 25+ years ago may not be applicable for the population! s) occurring in more southerly waters. Also, in coastal waters of Rhode Island and Virginia where tautog occur, environmental parameters (pri- marily seasonal temperature regimes) and habitat availability are different, and these factors may influ- ence growth rates in different populations or segments of the same population. Occurrences of large speci- mens, such as the current (IGFA 1990) and previous (IGFA 1986) world-record tautogs in coastal waters of Virginia, may also indicate that fish in the lower Mid- Atlantic Bight comprise a separate population, with growth characteristics different from their northern counterparts. Materials and methods Tautog were collected opportunistically from March 1979 to July 1986 by spearfishing, rod-and-reel, com- mercial fish pots, and as bycatch in trawl tows from other research studies. During the course of the study, fish were taken at 19 different locations, at depths of 2-35 m and representing a wide variety of ecological conditions characteristic of tautog habitat within Hostetter and Munroe Age. growth, and reproduction of Tautoga onitis 47 Figure 1 Collection localities (marked by stars) in lower Chesapeake Bay and Virginia coastal waters where Tautoga onitis were collected for age, growth, length, and weight param- eters. Stars indicate general collecting areas, with most indicating more than one collec- tion per area. Arrows indicate locations of the Chesapeake Bay Bridge Tunnel (CBBT) and of previous studies on tautog from Rhode Island (by Cooper, see Citations) and from New York and northern New Jersey (by Briggs & Olla and co-workers, see Citations). Chesapeake Bay and offshore hardbottom areas (CBBT, wrecks, artificial reefs) of coastal Virginia (Fig. 1). Each specimen was measured to the nearest 1 mm for standard length (SL) and total length (TL), and weighed (WT) to the nearest 5g. Length data from both sexes were combined to generate a regression equation for SL on TL (TL = 4.78+1.20 SL). High cor- relation (r 2 =0.98) between these measurements prompted the use of TL in analyses, since this was the more easily and reliably obtained measurement. For each fish, an initial determination of sex was made by examining several external characters that have been previously shown to be dimorphic (Cooper 1967, Olla & Samet 1977). Adult males usually have a blunt forehead with a more massive mandible, com- pared with that of adult females which have a less massive man- dible and more anteriorly-ta- pered profile of the head. Larger males are typically gray and have distinctly visible (especially un- derwater) white markings on the caudal, pelvic, and dorsal fins, and also on the chin region. Fe- males (all sizes) and smaller males tend to have a mottled brown coloration without white markings on the fins and chin region. After initial determination of sex based on external character- istics, gonads were then excised, staged macroscopically for matu- rity state, and weighed (gonad weight = GW) to the nearest O.lg (gonad weights were not available for all fish). Maturity stages were defined as fol- lows: Immature — gonads un- differentiated; Mature — gonads obviously differentiated; Ripe — gonads enlarged, containing sperm or ova; Running ripe — sperm or ova expressed when slight pressure applied to abdo- men; Spent — large, flaccid go- nads, often bloody in appearance, with no obvious signs of sperm or ova upon dissection. A go- nadosomatic index (GSI) for fe- males for all years combined was calculated using the formula GSI=GW x 100/WT. Scales, saccular otoliths, and opercles were collected and com- pared to determine the best method for ageing tautog. Whole unsectioned opercles (Fig. 2) were prepared fol- lowing procedures used by Cooper (1967). The articu- lar apical center, as defined by Le Cren (1947), McConnell (1952), Bardach (1955), and Cooper (1967), is the center of the high ridge projecting from the me- dial surface of the opercle. Opercular radius (OR), de- fined as the distance from the articular apex center to the midpoint of the posterior margin of the opercle, was measured with dial calipers to 0.1mm (Fig. 2). Similarly, measurements (to 0.1mm) were made along this axis to each annulus to determine annual growth increments. An annulus was defined as the sharp tran- sition from a translucent (hyaline) to an opaque zone on the opercle. Only discernible annuli continuous from 48 Fishery Bulletin 91(1), 1993 Figure 2 Measurements on each opercle of Tautoga onitis used in age analyses. Illustration depicts whole left opercle from Tautoga onitis. Orientation of the opercle is: (A) anterior; (P) poste- rior; (D) dorsal; (V) ventral. Successive annuli (3-15; first two annuli obscured by articular apex) are indicated; OR is the opercular radius; AA is the articular apex of the opercle. anterior to posterior margins of the bone were con- sidered to be annuli and counted to determine age. Other horizontal marks such as incomplete bands ( = false checks) were not included in the annuli counts. All annuli continuous within contours of the opercle were counted on both right and left opercles using transmitted light. Initially, annuli on both opercles from each fish were counted with 90% agreement between counts. If differences were noted, the age estimate from the opercle with the most clearly de- fined annuli was used. All age estimates were made by the senior author, then a subset (n = 100) of those opercles were re-read by the second author. All ini- tial estimates by both readers were in agreement, thus age estimates (by Hostetter) were used in sub- sequent analyses. Annuli were counted for both right and left opercles using transmitted light. A total of 24% (167/706) of the opercles were counted six times; four counts were made at lx and two were made at 6x magnifica- tion. Since there was close agreement between all counts regardless of magnification, the remaining 76% of opercles were aged twice under lx magnifi- cation. Marginal increment, the seasonal growth of the opercle, was measured by plotting increment width from the last annulus (A) against date of cap- ture (Fig. 3). Standard least-squares linear regression (Sokal & Rohlf 1981) was used in Lotus-R spreadsheet format (Jeanty 1984) to describe TL:SL, TL:OR, and TL:WT relationships. Determination of time of annulus for- mation was adopted from Nose et al. (1955) and Coo- per (1967). Mean back-calculated TL-at-age (Table 1) was computed for males and females separately and for sexes combined, following Bagenal & Tesch ( 1978). Slope and intercept values from these equations were used in the back-calculated length equation of Ricker (1975). Calculated lengths by sex were independently determined through substitution of logarithmic val- ues for average TL and OR by age-class. Analysis of covariance (ANCOVA) using SPSS-X (Norusis 1985) was used to compare age-at-length between sexes. Back-calculated mean lengths-at-age were used to de- velop von Bertalanffy curves (Gulland 1976). Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH Figure 3 Mean seasonal incremental growth (in mm I of opercles of Tautoga onitis from coastal waters of Virginia. Vertical bars represent the range; middle horizontal bars are the mean; darkened portions represent the standard deviation; and num- bers at top of each line are sample sizes. Hostetter and Munroe Age. growth, and reproduction of Tautoga onitis 49 Table 1 Back-calculation formulae for male and female Tautoga onitis from coastal waters of Virginia (TL=mm total length; OR=mm opercular radius I; L„=total length at year n; R n =opercle ra- dius at year n; R,=opercle radius at capture; b=slope of body-bone regression; where L n = log TL+b(log R„-log R,). (males) log TL = 1.2916+0.860 log OR (rc=398; r-'=0.968) (females) log TL = 1.2889+0.864 log OR (n=281; r 2 =0.967) Results Ageing technique and validation Opercles were found to be the best structure for esti- mating age of tautog. In fish with less than four an- nuli, there was close agreement between annuli on scales and those on opercles. However, abrasion, sur- face-area damage, and ring compaction along the outer margin of scales precluded using scales for ageing older fish. Comparisons of age estimates from otoliths and opercle bones from 27 fish indicated close agreement in readings of annuli from each structure, especially in younger fish. However, otoliths from larger fish had a thickened nuclear core, which made it difficult to discern any microstructure in this region. Sanding and sectioning of otoliths proved difficult because of the small size of the otoliths, and in most cases these pro- cedures also blurred or removed annuli on the outer edge of the otolith. The first annulus is also sometimes difficult to detect on opercles of large tautog because of thickening of the buttress zone at the articular apex. In these instances (<5% of the fish aged), we assigned a distance to the first annulus (7.4 mm) based on mean measurements from opercles with clearly defined first annuli. Opercle radius and TL (Fig. 4) were linearly related (r 2 =0.97; sexes combined). Age estimates based on the first and fourth readings (at lx ) of individual opercles were in close agreement (81%, 135/167). All age dis- agreements were within lyr (n=31), except one (2yr). Marginal increment analysis (Fig. 3) revealed mini- mum growth distal of the last annulus during May, June, and early July for all age-groups. There was no indication of formation of a second mark during the year for any fish examined. Age and growth A total of 712 tautog measuring 51-765 mm TL was collected (Fig. 5). Of these, 701 (398 males, 282 fe- males, and 21 immature fish of unknown sex) were used to estimate age distribution and growth rates. 70- x< X 60- Y = -0.353 r2 = 0.97 + n 0.009X = 706 X "£50- 3 40- Q < Q. uj 30- o UJ £2°- XT** !!r X 10- >* ~r. 1 i 100 200 300 400 500 600 700 800 TOTAL LENGTH (mm) Figure 4 Relationship of opercular radius (OR) on total length (TL) for Tautoga onitis (sexes combined) from coastal waters of Vir- ginia. 1979-86. Specimens not included in age analyses had missing data or damaged opercles, which precluded their use in the analyses. Mature males (rc=364) ranged in size from 198 to 765mmTL, weighed 138-6895 g, and were aged 3-25yr. Mature females (n=247) were comparable in size (232-750 mm TL) and weight (130-7392 g), and were aged 3-21. Few males (9%, n=32) and females (5%, n=14) in our samples were older than age-13. Immature fish ranged from 51 to 265mmTL, weighed 5-410 g, and were aged 0-3 yr. Length-weight rela- tionships (Figs. 6,7), calculated separately for males and females, indicated that mean total lengths and 70- 60 - [L 50- n = 712 -i -■ x 40- *30- I ~| -1 20 - T-| 10- JrhJi ln __ 100 200 300 40 500 600 700 TOTAL LENGTH (mm) Figure 5 Length-frequencies of Tautoga onitis within 25mmTL size- classes collected in coastal waters of Virginia. 1979-86. 50 Fishery Bulletin 91(1), 1993 4.0- d 3.8- y - -4.54 + 2.94x X XxXX r2 = 0.97 n = 396 >*«£ & 3.6- ^-~. 3 3.4- r^ x g 3.2- x x *3.0- x >*^af^ fc < B2.8- £jg< >< X t- *p g ) 2.6- j^jap^ x _J 2.4- X >*^ X 2.2- ] ; 1 2.3 2.5 2.6 2.7 Log TOTAL LENGTH (mm) 2.8 2.9 Figure 6 Length-weight relationship for male Tautoga onitis collected in coastal waters of Virginia, 1979-86. weights (Tables 2,3) increased with age for both sexes (P<0.01). Estimates of mean back-calculated size-at-age (Tables 4,5) suggest that growth for male and female fish is similar. However, we analyzed the data by sex to com- pare with previously reported values. Greatest incre- mental growth in TL for both sexes occurred during the first year and then declined rapidly. Growth in the second year was only 40-49% of that recorded for the first year for both sexes (Tables 4,5). Only small differ- ences in back-calculated lengths-at-age occurred be- tween the sexes to age-13 (Tables 4,5), and these were not statistically significant. Males usually had a slightly 3.8- 9 y = -4.58 + 2.96x x x 3.6- v»*X r 2 = 98 n = 290 >\^» 3 3.4- &JT o 3.2- J&gP l 3.0 J < O 2-8_ i- o*2.6. 2.4. 2.2_ X x X 2.3 2.4 2.5 2.6 2.7 2.8 2.9 5 TOTAL LENGTH (mm) Figure 7 Length-weight relationship for female Tautoga onitis collected in coastal waters of Virginia, 1979-86. larger growth increment at each successive age throughout the life span. Estimates of empirical length-at-age (Table 6) com- pared favorably with both back-calculated estimates (Tables 4,5) and observed growth (Tables 2,3). K-va\ue for male tautog (0.090) was greater (Table 6) than that calculated for females (#=0.085). Males were also larger in size (TL) when compared with females of compa- rable age (Fig. 8), although the differences were not statistically significant. Males and females achieved 50% of L„ between ages 6 and 7, and 75% between ages 14 and 15. ANCOVA analyses indicated no sig- nificant differences between slopes of regression equa- tions of length-at-age for male and female tautog (/ r =2.600, P>0.05) or for homogeneity of means around regression slopes (F=2.979, P>0.05). Derived length-at-age estimates from von Bertalanffy growth equations were later used in regression equa- tions to calculate weight-at-age. Although correlation coefficients were high (r=0.81) in the analysis of WT on TL, variation in weight-at-age within age-groups was considerable, and estimates of growth based on weight were less reliable than estimates based on TL. Sexual dimorphism and reproductive biology We observed two different morphological males in fish we examined. Approximately 15% of the fish we classi- fied initially as females, based on external characteris- tics, were later determined upon dissection to be males. Generally, these non-dimorphic males were fish smaller than 550mmTL and less than age-10. However, sev- eral approached the largest sizes observed for other males. Pigmentation of non-dimorphic males was a dull mottled brown, with remnants of disrupted lateral bands, and was similar to that noted for females. In contrast, dimorphic males were typically grayish with distinctive white markings on ventral and dorsal mar- gins of pectoral and caudal fins and on the chin. The anterior skull and rostral region were also blunter and more massive in dimorphic males than for those noted in females and non-dimorphic males. Both types of morphological males were considered as males in analy- ses of age and growth and sex ratios. Gonadal maturation was evident in both sexes by age-3. Age-2 fish, collected only in late March and early April, were immature with undeveloped gonads. GSI values for females plotted against date of capture (Fig. 9) indicated peak spawning from April through June, with the highest GSI recorded in May. GSI values de- clined rapidly after July. Although not shown, the GSI of tautog collected inside and at the mouth of Chesa- peake Bay peaked somewhat earlier, in mid-May, and started to decrease by mid-June, whereas a small per- centage (usually -20%) of running ripe fish were Hostetter and Munroe: Age. growth, and reproduction of Tautoga onitis Table 2 Sample sizes (n ) means (x), and standard deviations (SD) for total and s tandard length s (mm) and weight (g) by age for male To j toga onitis from coastal waters of Virginia. Age Total length Standard length Weight n X SD n X SD n X SD 1 11 246 21 11 201 17 11 306 98 2 23 276 33 23 225 28 22 491 156 3 52 292 40 52 240 33 51 542 261 4 38 329 42 38 273 41 34 740 306 5 29 357 57 29 295 47 27 982 459 6 30 375 46 30 313 43 30 1145 447 7 31 414 52 31 340 52 31 1577 711 8 40 443 42 40 364 34 40 1795 511 9 39 481 47 39 398 47 39 2397 694 10 30 501 48 30 416 48 30 2693 794 11 18 533 41 18 438 36 18 2911 731 12 20 564 47 20 466 38 20 3663 839 13 6 545 22 6 451 22 6 3410 466 14 7 574 25 7 474 23 7 3861 573 15 1 595 1 491 1 4340 16 5 586 29 5 489 30 5 4202 777 17 2 587 27 2 479 20 2 4860 30 18 8 655 48 8 532 31 7 5549 766 19 3 614 15 3 504 10 3 4452 62 20 1 550 1 460 1 3090 21 1 613 1 508 1 4750 22 2 655 25 2 538 22 2 4995 725 23 — — — — — — 24 — — — — — — 25 1 672 1 577 1 6568 Table 3 Sample sizes (n) means (jc), 3nd standard deviations (SD) for total and sta idard lengths (mm), and weight (g) by age for female Tautoga onitis from Virgi nia. Age Total length Standard length Weight n X SD n X SD n X SD 1 12 224 33 12 185 28 12 256 107 2 23 268 37 23 222 32 23 444 182 3 31 286 40 31 236 33 31 526 240 4 34 334 30 34 274 26 34 822 226 5 27 350 35 27 284 33 27 910 245 6 23 378 47 23 317 44 23 1205 468 7 37 401 38 37 334 33 37 1395 479 8 16 425 43 16 351 38 16 1551 438 9 20 453 58 20 375 46 20 1960 700 10 13 519 42 13 428 37 13 3114 876 11 13 524 54 13 436 46 13 3033 873 12 11 545 42 11 451 35 11 3660 1063 13 8 525 39 8 437 31 8 3227 643 14 3 551 64 3 464 54 3 3677 1188 15 2 518 30 2 429 20 2 2715 395 16 1 485 1 395 1 2530 17 1 575 1 479 1 4220 18 5 628 80 5 517 47 5 4961 1700 19 — — — — — — 20 1 557 1 485 1 3385 21 1 660 1 570 1 5100 52 Fishery Bulletin 91 11). 1993 fc. ^6S 2 c to he H co co ^ t- CO t- to CO to to CO to CM ■x to *tf to CD CM to CM CM CO 00 in to in in to en c- CO to CO o in to CO — CO CM t^ CO CM CO in in to o to CO to — CO CO CO o CD ?] to m CO en m en CM to CO to CM *tf O -* CD 1 X to - in CO in m en m to to X CO CO to " 00 00 m CM m to in to in in o to to CO m en oo en ^-< m en in m to CO to 00 m m CO CO m to in CO CO m CO m in •-H 00 00 -H m CO 00 m in m in m 1 — 1 to CO m m 00 CM m CO in en • — in CO to CD -h m CM CO W o m CM CO m CO in m 05 in CM in CM en 00 o m en CM in 35 r- to tf ^ m in CO -tf in CM in m in o in o in CM m to m en in o in in to *te —i en CO m "tf lO in CO CM lO CO CM ■tf m o 03 00 CM in CM m -* >* CO CO in CO -tf CM CO CM CM m CM C in CO CO in en en — en ■tf CO CO ■tf CM CO — - X CM . — 1 m to cc CM in CO CO to -<* en tf en cm tf- - CO -tf 05 77 in LTD CO CO c- C35 in •tf m CO m m CO CO CM en CO CO CO CO en CO in co t~ CM CO -tf CO ■tf CO CO -tf CO CM CO "tf ~ -tf CO CO "tf CO o — en -** CO l> CO o en CO I— 1 in CO CM CO in cm tf CO tf "tf •tf CO CO ■tf - — ■tf ■tf ■tf -tf CM -tf o o "tf CO 00 CO en CO CO — CO CO CM in CO CO m CO o 00 CO c- in CO in oo CM cm -tf - tf — O ■tf x ■tf CO ■tf in CO 71 o CO co 00 CO CO / in CO CO to CO CO CM in CO to CM CO c~ CM CO en CO en CO CO C- CO en co CO CM X CO c~ CO CO CO m CO o CO CO -r CO CO CO tf- tf CO CM CO CO 00 ■tf CO 00 CM CO CO CO o CO CO o o CO en CO CM CO CM CM CO -tf CO CO CO CO — CO CM CO CO CM -tf CO en CM CO o tf CO CO CO CO CO CO CO CO CO - CM CO CO CO CO 71 CO - CO 00 en CM CM CM CO CD CM o en CM CO en CM -H co CO CO CO / CO CO CO CO OS CM CO CM 05 CM X- CO 00 en CM (35 CM in en CM 00 ■tf en CM ■tf CO CM CO A CN 00 CM in CM to CM en CO CM CO CM CO CM CM l> CM CO ^ en tf CM 00 CO CM CM X -1 CO m CM — in CM ■tf in CM i- in CM IK CM CM ~ CM CO ■tf CM in CM — CM CO CM CM X CM en in CM CM en CM CM en o CM 00 CM CM CO CM CO CO CM -tf O in in CM en CM 77 - CM - CM IN — zz CM CO = CM CM ~- W en CO 00 tf 77 ^1 CM CO CM CO 1 - to m in to o CM CM 00 ■tf CO o in CM CO tf CM tf — — — -tf — CM IG i- CM >- lO — 1 - Cn — in — CO 1- — IC CM CO 00 CO en to CD -r — *tf CM t— in •-< CN 00 00 CN lO CM CO CO in CO 05 CO X ■tf O tf -tf [ - ■tf c in en in X lO X CO CO in CM en in 00 m CO OJ in CO to to X 00 m CM CM to en tr- ee CM to ^H CO ^ CM CM in 00 CO Cn o CO CO o — CO o CO 00 X CM CO c~ . m CM 00 CO . CM ,_, ,_, c c p oj o> a t ■— iNW^iOCObOOOOr- ' iM W tJ" lO CD [^ by S C0050^CNCC^lO> F t HrHCNMWCNNCN>C Hostetter and Munroe: Age, growth, and reproduction of Tautoga onitis 53 •Q in -2 o J 5 ^ ■a c rt S ^ ft _. CD u 0) — bJD H O M ifi !£ lO lO CO CO o m co cm m tj- ex cm lO CD W | CO CO Tt O N 00 Tf CO CM -h CO lO CO CD 00 CO t~- ■— < GO ft O l> CM CM O CM cd co m co co CO CM CM CD o r- I> CO OS CO O O CO CM m lO CD lO CD id CMcotr-ror-Tr'Om H^r-Tj-cocxcoiN m io io co -^ m tn HOOiOtTCDCOCOCD^ OOlCOifDCMiOCOCO'— t lOOCDCO^COCOCMf* iocnt^ocoococ-cM OTfTflClOCOTj-lOlO iooc-uocoomco loasocOfCOf'ft ^ ^ ii: in Tf ic io lOX'toiOTf-f-Triocoicoico •fOCMiccocococoaiTj-aif > icicifi't'j'jTj'mcomTit COCMcOCD-rfCOCMCOOfiCDCDcy: f'OiXOCOftCOmcOt-CMQOf' lO^TtmTfTjt'fl'Tj'lOCOmTj' Ifl -^ Tf 't ■* Tf ^TtO^iOCMCDCD COf«CCOmOCOCM co ^ 't m co w ^ Co^OftCOCOtOOOCMOCOCOOOO ^COCDCOCMCDCOCOOOOr^COr-Tj'CM "tTfTj'Tj'^'tCOCOCOlTf^COTl'Tt CM O UO O "tf* t- C- h O] rt CO O Ol ■* ■"3* Tf tJ" rj* rj- CO ^ CO f < t"- CO CM CO t^ CO t- CDCOCDNlCHCOHrM COCOCOCOTj-COTf"^ c^ajcoc7J-^'aiocMai'^'Oc--ot>ocoir^ CCCOXOOt^^CMCOO^^^CTicjlCOCM COCOCO^t-rJ'COCO^t'COCOCOCOTj'CMCOCO COOCOCOXCOCX^OtXJ^OiOCOOi'^'aiCM ^^iCiOt-t'^'tOHCOONHNiOlOCO COCOCOCOCOCOCOCO^J'COCMCOCO^f'CMCOeO NONtDCOOWHOlXtCCOHQO^QOO^^ COTfCMtrT}'Tj.[s.c£) r » N ^Q^ c o^^ 1 CNCOCOCOCOCOCOCO^^COCOCOlOlNNHCNincOTj-COCO i'OiCOCOWOlHNOCNCNC^OCOCOCOlO^'lOt-O C^^J'OOCO^C~-OCMl£)CMCMCM"^'iO'— tftC--— 'OOmCD HiNMcococo , t , t'!j'ioiommmiowio;D!Dioco O^tOiCMCMCMCMfi^ftftfi CD CD •— »cMcoTtuocotr~-cocyiOi— lojco^intflt-coajo O! 5: O 54 Fishery Bulletin 9! [I). 1993 Table 6 Estimated parameters and standard errors (SE) of the von Bertalanffy growth equation for Tautoga onitis from coastal waters of Virginia (L ,=mean asymptotic total length in mm; K=growth coefficient; t =time (yr) at which total length would theoretically be zero). Asymptotic Parameter Estimate SE Males u 732.24 9.124 K 0.090 0.003 to -1.64 0.132 Females L, 733.61 28.362 K 0.085 0.009 to -1.74 0.324 Sexes combined L, 742.37 9.051 K 0.085 0.003 to -1.816 0.144 9 f 8-' 65 7-' 6-' 5 i/i 3 " ^1 43 30 13 2-H 3 1 - 1 Tru o 4 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MONTH Figure 9 Seasonal gonadal development (gonadosomatic index, GSI) for female Tautoga onitis collected in coastal waters of Vir- ginia, 1982-84. Numbers over bars are sample sizes. present in samples from offshore collecting sites until late July and early August. Fish occurring on wrecks further offshore (and usually deeper) generally had higher GSI values later into the season (early July-early August) than those collected from inshore areas. Chi-square analysis of sex ratios for 701 tautog di- vided into 10 cm length-groups indicated significant de- viations cj- 10 - Q 5 - - -5 Virginia Rhode Isl X 6 7 Month 10 11 12 Figure 1 Average monthly water temperatures ( 1979-86 ) for lower York River, Virginia (VIMS oyster pier, taken lm below surface at 2.0-2.5m depth; G. Anderson, Coll. William & Mary, VIMS, Gloucester Point VA, pers. commun.), and for bottom-water temperatures at Fox Island, Narragansett Bay, Rhode Island iH.P. Jeffries, Univ. Rhode Island, Grad. School Oceanogr., Narragansett RI, pers. commun.). Solid squares and open triangles represent mean monthly values for York River and Narragansett Bay, respectively. 10° C nearly a full month longer in Narragansett Bay (not until mid- April) than in the York River (usually mid-March). And in the fall, temperatures again de- cline below 10° C during mid- to late October in Narragansett Bay, whereas in Virginia temperatures remain above 10° C usually well into mid-November or early December. These temperatures, which are con- ducive for somatic growth in juvenile tautog, are ex- tended seasonally in coastal waters of Virginia com- pared with those of more northern areas. Recently, D.L. Martin and T.E. Targett (Grad. Coll. Mar. Stud., Univ. Delaware, Lewes DE 19958, unpubl. data) found in laboratory growth experiments that young-of-the- year tautog from high-latitude populations (Rhode Is- land) show no genetic compensation for a shorter grow- ing season when compared with tautog from Delaware Bay and Virginia waters. These data further support our hypothesis that observed differences in growth of young tautog from northern (Rhode Island) and south- ern (Virginia) areas of the species range are due pri- marily to environmental factors between the two ar- eas (i.e., duration of optimal temperatures for growth is longer in Virginia coastal waters compared with those of coastal areas in the northern end of the species range ). Estimated values of L„ for tautog from Virginia are also considerably higher than those estimated for tau- tog from Rhode Island (Table 8). A calculated L„ of 733mmTL (data for sexes combined; 733mm for fe- males, 732 mm for males) as derived from the von Bertalanffy equation in our study is close to the ob- served maximum TL of 765 mm. Growth equations es- 56 Fishery Bulletin 91 1 1), 1993 Table 8 von Bertalanffy growth functions derived for Tautoga onitis collected in Virginia (1979-86) and Rhode Island (from Coo- per 1965). Virginia tautog (males, rc=398) K to L, (females, n=281) K to L, Rhode Island tautog (males, rc=1041) L, (females, n=1119) L, 0.090. L. = 732 -1.644 732 [l-e^" 90 " 416441 ] 0.085, L,., = 733 -1.743 733[ 1-e -"""■•"•' 7,J, I _ 66411-e -009108 "* 1 66238 '] = 506[l-e J " 51,,i " uo95220, ] timated for tautog in Rhode Island (Cooper 1965) indi- cated appreciably smaller L„ values (506 mm and 664 mm) for both sexes compared with those estimated for tautog in Virginia. The present study also found that males grow at a somewhat faster rate than females. Cooper ( 1965) found a slightly faster growth rate in females when com- pared with males (Table 8). However, in both studies estimated if-values are comparable. Similarities in K- values between tautog occurring in Rhode Island and Virginia support the contention that growth rate (K) is an intrinsic value for the species, largely independent of geographic location. Although Cooper (1965) found that females initially grew at a faster rate than males, he reported diver- gence in growth between sexes favoring faster growth in males after age-3. We also found similar divergence in growth of males, but unlike Cooper's study, this difference was apparent for males at all ages. In our study area, males live longer than females. The oldest fish examined were a male estimated to be age-25 and a female estimated to be age-21. Cooper ( 1967) indicated a life span of 34 yr for males and 22 yr for females in the population he studied in Rhode Is- land, and suggested that females may reach senes- cence at an earlier age than males. Although longevity estimates of tautog based on actual data range be- tween 25 and 34 yr, most fish aged were considerably younger than this, and it seems that claims of fish more than a half century old (Reiger 1985) are exag- gerated. Average age for tautog in this study was just over 7yr; 82% of the fish were age-10 or younger, only \' r were age-20 or older. Cooper (1967) found a similar age structure in the population of tautog residing in Narragansett Bay just over 25 yr ago, where approxi- mately only 157c and 8% of males and females, respec- tively, were older than age-13. The current world record for tautog (IGFA 1990) was a fish taken by rod-and- reel off Virginia measuring 819mmTL (-10.89 kg). Within constraints of the von Bertalanffy equation and length-weight relationship derived for tautog from Vir- ginia waters, we estimate an age for this fish of -30 yr, which is comparable to the maximum age (34yr) re- ported for the species (Cooper 1967). Weight-at-age estimates for tautog, as a measure of growth, were much more variable than length esti- mates. This variation is attributed to different stages in ontogenetic development, as well as differences in sex, maturity, and age. Geographic location and asso- ciated environmental conditions, such as seasonality (date and time of capture), stomach fullness, disease and parasite loads (Le Cren 1951, Bagenal & Tesch 1978), can also affect weight-at-age estimates. Since these factors contribute significant variation to regres- sion relationships of WT on TL, interpretation of dif- ferences in these data between populations must be viewed with caution. Direct comparisons of our length-weight data with those of Cooper (1967) were not feasible due to sam- pling differences (Cooper used eviscerated weights of fish). However, in a study of tautog from coastal waters off New York, Briggs ( 1969 ) calculated a length- weight relation based on uneviscerated weights of over 3000 fish collected during several seasons (May- November) over a 3yr period. Although Briggs did not present data for individual sexes, comparisons of data combined for both sexes are still possible (Table 9). From these comparisons, it is evident that the length- weight relationship for tautog from Virginia waters is similar to that estimated for tautog from off New York. Table 9 Comparison of estimated length-weight relationships between Tautoga onitis collected in Virginia (uneviscerated weights, sexes combinec ; Log W (g)= -4.632+2.979 Log L; n=687l and New York (from Bri ;gs 1969; uneviscerated weights, sexes combined: Log W(oz = -5.992+2.916 Log L; n=3156). Length (mm) Estimated weight Ig) New York Virginia 150 76.5 70.9 200 119.1 112.2 250 334.4 324.7 300 567.0 558.9 350 890.2 884.7 400 1215.4 1316.8 450 1854.1 1870.3 500 2520.3 2560.0 550 3328.3 3400.5 600 4289.4 4406.7 650 4941.4 5096.1 Hostetter and Munroe: Age, growth, and reproduction of Tautoga onms 57 Reproduction and growth Maximum GSI values calculated in the present study indicate that spawning commences in late April and continues to early June for tautog taken at inshore sites in Virginia. Finding fish in spawning condition in Virginia in late April is somewhat earlier than reported previously for tautog from more northern inshore ar- eas, and undoubtedly reflects the warmer tempera- tures in coastal waters of Virginia in early spring. Based on laboratory (Olla et al. 1980) and field obser- vations (Chenoweth 1963, Olla et al. 1974, Eklund & Targett 1990, this study) of ripe fish, spawning in tau- tog generally commences when water temperatures reach 11° C or above. In Massachusetts, tautog spawn from mid-May to early August (Stolgitis 1970), while peak spawning was reported to occur from late May to early June in tautog collected within shallow wa- ters of Narragansett Bay, Rhode Island (Chenoweth 1963). Near Long Island, tautog eggs have been collected in the plankton from May through early September (Perlmutter 1939, Wheatland 1956, Austin 1973); however, the effective spawning season may be somewhat shorter since few larvae were col- lected when water temperatures exceeded 21.0° C (Aus- tin 1973). Highest GSI values for tautog from Virginia occurred over a longer seasonal period than reported for tautog collected inside Narragansett Bay (Chenoweth 1963). We attribute this to the fact that we collected fish over a broad range of sites spread over a much wider geo- graphic area, including many deepwater offshore sta- tions. On hardbottom areas 22-37 km offshore of Mary- 1 7 . i> I hM 6- 8 30 5- 4- 16 l Jan Feb Mar Apr III) May Jun Jul Aug Sep cl Nov De c MONTH Figure 1 1 Relationship demonstrating co-occurrence of time of annulus formation (mean marginal increment, MMI) and peak spawn- ing season Igonadosomatic index, GSI) for Tautoga onitis from coastal waters of Virginia. land and northern Virginia, Eklund & Targett (1990) noted significantly higher GSI values for female tau- tog (24-50 cm TL) from May through the beginning of August, with spawning taking place during summer (May-July). This time schedule is similar to what we observed in tautog collected at offshore habitats in southern Virginia, and, as was pointed out by Eklund and Targett, this seasonality also corresponds with the occurrence of tautog eggs and yolksac larvae in Mid-Atlantic Bight plankton samples (Colton et al. 1979). Annulus formation in tautog collected in Virginia occurs in May or June commensurate with gonadal maturation (Fig. 11). Formation of an annulus on the opercle concomitant with spawning was noted also by Cooper (1967) for tautog from Rhode Island. Decline in physiological condition during gonadal development, presumably representing disruption in somatic growth, was observed in tautog by Chenoweth (1963). Such disruption in somatic growth could contribute to the slow-growth phase observed on opercles during this time. However, annulus formation occurs in sexually immature fish during this same time, indicating that inherent physiological factors other than those associ- ated with spawning also influence annulus formation in these fish. Annulus formation in tautog collected in Rhode Is- land occurred in late or middle May at the start of the spawning season (Cooper 1967). In Virginia, we ob- served that annuli formed over a longer time-period (May-July). Warmer water temperatures earlier in spring in Virginia may cause smaller fish to form an- nuli slightly earlier than larger fish. This growth pat- tern would agree with observations that younger and smaller fish are more active than larger fish at lower water temperatures (Olla et al. 1974, 1980). Also, since these smaller fish do not usually participate in spawn- ing activities, all growth during spring would be re- flected as increases in somatic rather than gonadal growth. Variation in the estimate of time of annulus forma- tion may have resulted from analyzing data indepen- dent of year of collection. Interannual differences in environmental conditions would be expected when these data are combined. Also, we sampled tautog from vari- ous geographically separated inshore and offshore lo- cations, and small-scale variations in time of annulus formation may be expected in fish collected from these diverse areas. One other source of variation could re- sult from annulus formation during the spawning pe- riod. Since spawning commences earlier and is appar- ently more protracted for fish in southern areas of the species range, the period for annulus formation would also be extended in tautog occurring in Virginia com- pared with those occurring further north. 58 Fishery Bulletin 91(1). 1993 Growth and sexual strategy The present study, and previous ones conducted in more northern waters (Chenoweth 1963, Cooper 1967, Stolgitis 1970, Briggs 1977), found that male tautog mature by age-3 and females by age-4. Similar age-at- maturation for tautog from different portions of the species range for the present population of tautog liv- ing in Virginia's coastal waters, and that reported by Cooper 25 years ago (1967) for tautog from Narragansett Bay, may reflect demographics of popu- lations that have not sustained intensive exploitation. It would be valuable to compare age-growth data for the present-day population of tautog residing in Narragansett Bay with historical data in Cooper (1967) to test this hypothesis. It is unknown what percentage of tautog in a popu- lation mature precociously, under what environmental or social situations, or even if smaller fish are sexually active and reproduce successfully. In May 1985, the second author collected a 180mmTL, age-2 gravid fe- male north of the confluence of the Taunton River and Mount Hope Bay, Massachusetts. In the present study, no sexually mature females smaller than 230mmTL or younger than age-3 were found. We note, though, that our data are limited for age-2 fish, especially for fish of this age-group during their second summer and fall. Olla & Samet ( 1977) also reported collecting sexu- ally mature tautog "which were of a much smaller size and younger age [no sizes or ages reported] than has previously been reported" from coastal waters of New York and northern New Jersey. It is interesting to note that precocious individuals have only recently been reported, and these were tautog occurring in northern waters where sport and localized commercial fisheries have operated historically since the 1800s (Goode 1884) and have undoubtedly intensified since that time. Ear- lier maturation is a common compensatory response in fish populations subjected to intensive exploitation (Goodyear 1980) and may explain the appearance of precocious individuals in tautog populations inhabit- ing northern portions of the species range where ex- ploitation has occurred for a longer time. Sex ratios for tautog divided into 10 cm length-groups were found to deviate significantly from a 1:1 ratio in the larger size-classes, with larger (and older) size- classes of tautog being comprised predominantly of males. Based on a smaller sample size, Eklund and Targett ( 1990 ) also reported a sex ratio (0.86:1) skewed in favor of males. Small sample size in their study, however, precluded breakdown of sex ratios over the size range studied. Among other factors, skewed sex ratios in larger (and older) fish may be attributed to differential growth and longevity of males or slowing of growth (measured as TL) with age in females, or possibly as a result of sex reversals. Faster growth coupled with greater longevity for male tautog found in Virginia's coastal waters may reflect the higher en- ergetic costs of reproduction and subsequent earlier senescence and differential mortality for females, as suggested by Cooper (1967). Larger size may also be selected for in males. Observations of courtship and spawning reveal that a size-related male dominance hierarchy is one reproductive mode (group spawning without a dominance hierarchy is the other) occurring in tautog with dominant males exhibiting strong terri- toriality and performing a protracted courtship with females, culminating in pair-spawning (Olla & Samet 1977). In reproductive strategies involving pair-spawn- ing, territoriality, and dominance social hierarchies, size selection for large males would be advantageous. Such strong size and sexual selection is known in other labrids, including bluehead wrasse Thalassoma bifasciatum (Warner et al. 1975), California sheeps- head Semicossyphus pulcher (Warner 1975), cunner Tautogolabrus adspersus (Johansen 1925, Pottle & Green 1979a,b), and others (Warner & Robertson 1978), where females primarily select larger (older) males as spawning partners. Although diandric male phases are prevalent in both tropical and temperate labrids (Robertson & Choat 1974, Warner & Robertson 1978, Dipper & Pullin 1979), diandric male tautog were not reported in earlier stud- ies (Chenoweth 1963, Cooper 1967). Olla & Samet (1977), following Cooper (1967), noted in their study on spawning behavior that "tautog were easily identi- fiable with respect to their gender by the sexually di- morphic mandible, which is more pronounced in males." However, in that same paper, unpublished data of Olla & Bejda noted the occurrence of sexually-mature young tautog of both sexes, without any sexual dimorphism. Olla & Samet (1977) suggested that mandibular di- morphism in tautog may develop ontogenetically, be- coming apparent only in older, larger fish. In contrast, we found wide overlap (up to ~50cmTL) in body size between the two male forms, rendering it unlikely that age alone controls development of secondary male char- acteristics in tautog. Olla & Samet (1977) also discussed the possibility that younger, mature fish may represent a different sexual stage than that of older fish. They pointed out that nothing was known of behavior or gonadal devel- opment of these young fish, and that it was even re- motely possible that tautog might be hermaphroditic. Since two spawning strategies have been observed in male tautog (Olla et al. 1977, 1981), it is possible that diandric males are those that utilize different re- productive strategies. Since non-dimorphic males have coloration patterns reminiscent of females, they may increase spawning opportunities through sneak or in- Hostetter and Munroe: Age, growth, and reproduction of Tautoga onitis 59 terference spawning during activities of territorial males. Reproductive behavior of individual male tau- tog is flexible and influenced by both size and sexual composition of the population. For example, in some situations with co-dominant males, or when males greatly outnumbered females, group spawning occurred even among dimorphic males that in previous experi- ments exhibited strong territoriality and more typi- cally attempted only exclusive pair-spawning with fe- males (Olla et al. 1981). Diandric males, each with different reproductive behaviors, have been reported for hermaphroditic labrids (Robertson & Choat 1974, Warner 1975, Warner & Robertson 1978, Dipper & Pullin 1979, Pottle & Green 1979a) and scarids (Warner & Downs 1977, Robertson & Warner 1978). However, diandric males not resulting from sexual inversions, but with different reproductive strategies, have also been reported (Pottle & Green 1979a,b) in cunner Tautogolcibrus adspersus, another temperate species of wrasse co-occurring throughout most of the geo- graphic range of the tautog. Plasticity of male reproductive behavior, presence of diandric males in the population, and skewed size and sex-ratios indicate that reproduction in tautog is more complex than recognized previously. Many of these same characteristics are paralleled in protogynous her- maphroditic labrids. In fact, protogynous hermaphro- ditism is one of the more common reproductive strate- gies utilized by labrids (Roede 1972, Warner & Robertson 1978). Earlier researchers (Chenoweth 1963, Cooper 1967) did not consider that tautog might be hermaphroditic; others (Olla & Samet 1977) recog- nized such possibilities, but, as yet, no evidence based on histological examination of gonads exists to prove or disprove the occurrence of hermaphroditism in this species. In view of the complex reproductive biologies of other labrids, further study on reproductive biol- ogy of tautog is warranted and is currently under investigation. Growth rates of tautog, other labrids, and reef fishes Few published age-growth studies on labrids exist, un- doubtedly because most are tropical species that have proven difficult to age reliably and few have commer- cial or recreational value. However, growth rates avail- able for temperate labrids from the eastern and west- ern Atlantic and eastern Pacific indicate slow growth rates and generally extended longevities in these spe- cies (Fig. 12), similar to those reported for tautog. It is possible that slow growth and extended longevities are characteristic not only of relatively large-sized tem- perate labrids, but also may be an inherent feature of growth patterns in large-sized labrids in general. E E 600- •i 500- J 400- o S 300- < u 200 L bergylta T. adspers 10 15 20 AGE (yrs) Figure 1 2 von Bertalanffy growth curves for selected species of temper- ate and subtropical species of wrasses (Family Labridae). Spe- cies represented and information sources for data presented in figure are: TX=Tautoga) onitis , Rhode Island (Cooper 1965), Virgina, this study; SX=Semicossyphus) pulcher, Warner 1975; LX=Labrus) bergylta. Dipper et al. 1977; TX=Tautogolabrus) adspersus, Serchuk & Cole 1974. Coefficients derived from the von Bertalanffy growth equation provide insights into ecological strategies, es- pecially in direct comparisons among diverse taxa (Table 10). Manooch (1979) considered fishes such as the bluefish Pomatornus saltatrix, Atlantic menhaden Brevoortia tyrannus, and king mackerel Scomber- omorus cavalla, which have relatively high /C-values (0.23-0.39) indicative of fast growth rates, as the coastal pelagic guild. Species with slower growth rates (K usually <0.22) and generally longer lived, on the other hand, were grouped together as the snapper- grouper guild. These fishes represent a wide spectrum of distantly related taxa including temperate labrids, reef-dwelling snappers and groupers, and other dem- ersal fishes such as tilefish Lopholatilus chamaelonticeps which inhabits burrows on the conti- nental shelf. Based on categories of growth coefficients adopted by Manooch (1979), we include the tautog in the snapper-grouper guild. This type of comparison, which crosses phylogenetic and demographic lines, sug- gests similarities in selection patterns for growth rates among species inhabiting areas where spatial resources may be limited. Conclusions and management considerations Growing recreational and commercial fisheries for tau- tog, limited amounts of natural habitat available in 60 Fishery Bulletin 91(1), 1993 Table 10 Comparison of growth coefficients (K-values) and longevity for selected species coastal fishes (age in years; L,, in mm). of labrids and other Species Source Age L, K Snowy grouper Epinephelus niveatus Matheson & Huntsman ( 1984) 17 1255 0.07 Tautog Tautoga onitis this study (males) this study (females) Cooper 1 1965 1 ( males ) (females) 25 22 27 22 732 733 664 506 0.09 0.09 0.09 0.15 Ballan wrasse Labrus bergylta Dipper et al. (1977) 29 405 — Calif, sheepshead Semicossyphus pulcher Warner (1975) >20 800 — Lane snapper Lutjanus synagris Manooch & Mason ( 1984) 10 501 0.13 Speckled hind Epinephelus drummondhayi Matheson & Huntsman ( 1984) 15 967 0.13 Mutton snapper Lutjanus analis Mason & Manooch (1985) 14 862 0.15 Tilefish Lopholatitus chamaeleonticeps Turner et al. (1983) 35 960 0.16 Red snapper Lutjanus campeehanus Nelson & Manooch (1982) 16 975 0.16 Scamp Myeteroperca phenax Matheson et al. (1986) 21 985 0.17 Cunner Tautogolabrus adspersus Serchukfc Cole (1974) 6 285 0.20 Black sea bass Centropristis striata Wenneretal. (1986) 10 341 0.23 Bluefish Pomatomus saltatrix Wilk(1977) 9 — 0.23 Blue runner Caranx crysos Goodwin & Johnson (1986) 11 412 0.35 King mackerel Scomberomorus cavalla Normura & Rodriques 1 1967) 14 — 0.35 Atlantic menhaden Brevoortia tyrannus Schaaf & Huntsman ( 1972 ) 0.39 the southern Mid-Atlantic Bight, and the slow growth and reproductive characteristics of this species, sug- gest a need for a fisheries management plan to main- tain the present stocks of tautog in Virginia's coastal waters (and elsewhere). It has been shown that in- tense fisheries directed at species exhibiting slow growth rates and a habitat-restricted ecology affect populations detrimentally (Manooch & Mason 1984, Matheson & Huntsman 1984, Moore & Labisky 1984, Harris & Grossman 1985, Matheson et al. 1986). Strong habitat preferences (hardbottom with structural relief), slow growth rates, extended longevities (to 25+ yr), and relatively long time to reach sexual maturity (3+ yr), indicate that strategies applied to reef spe- cies — the snapper-grouper cohort of Manooch (1979) — may be applicable in managing tautog populations as well. We suggest as a first step in managing stocks of tautog in Virginia the imposition of size limits on fish taken by recreational as well as commercial fishermen, Hostetter and Munroe: Age, growth, and reproduction of Tautoga onitis since the recreational fishery is the primary harvester of tautog. A minimum size limit of approximately 300mmTL (12 in.) is recommended for fish taken by recreational or commercial fisheries to insure that all females have at least one opportunity to spawn before being harvested (Briggs 1977). Imposing a 12 in. size limit for tautog should also insure the maintenance of a quality recreational fishery. Currently, a 12 in. mini- mum size limit is required for tautog taken by recre- ational and commercial fishermen in Rhode Island, Massachusetts, and Connecticut waters. To maximize any management plan for this species, it is also critical that the reproductive biology of tau- tog be well understood. Directed fishing pressure, dis- ruptive to size or sex ratios by the selective removal of dominant pair-spawning males (usually larger indi- viduals), could affect reproductive success in localized populations. Musick & Mercer (1977) concluded that heavy fishing pressure on black sea bass Centropristes striata may impact reproduction through changes in sex ratios in the population. Tautog populations in Virginia and elsewhere can also be enhanced by continued development of artifi- cial reefs (Feigenbaum & Blair 1986). Reef develop- ment is especially important in Virginia since suit- able, naturally-occurring substrate appears to be limited both in size and occurrence. Placement of arti- ficial structures over wide geographic areas also dis- perses fishing pressure, since competition for fishing space on presently-available isolated wrecks can at times be intense. Acknowledgments Portions of this study comprised an M.S. thesis (by the first author) presented to the Biology Department, Old Dominion University (ODU). Data collected by the sec- ond author comprised a graduate research project at the Virginia Institute of Marine Science (VIMS). We thank M. Armstrong, H. Brooks, M. Bucy, R. Crabtree, J. Colvocoresses, J. Desfosse, D. Estes, L. Gillingham, M. Harrel, M. Hodges, A. Hodges, K. Hodges, J. Lascara, J. Musick, B. Parolari, I. B. Ratnose, G. Sedberry, J. Smith, S. Smith, T. Sminkey, G. Susewind, W. Susewind, J. Sypeck, D. Thoney, G. van Hausen, and D. Wright for assisting with field collections, data gathering, and providing specimens. C. Hostetter provided much encouragement during early phases of this study R. Birdsong served as the- sis advisor and provided much encouragement and sup- port while the senior author attended Old Dominion University. J. Merriner (formerly VIMS) and H. Aus- tin (VIMS) arranged funding for the second author to intercept fishermen and work up catches. D. Munroe funded frequent purchases of large quantities of tau- tog from commercial fishermen. Students and staff at VIMS purchased eviscerated carcasses after sampling, thereby regenerating funds critical for additional pur- chases of samples. E. Barth, VMRC, provided fishery data on tautog in Virginia. Members of the Peninsula Saltwater Fishing Club provided specimens and sportfishery information. J. Stephens collected many large tautog in his commercial fishpot catches that enhanced this study. D. Hata and J. Loesch (VIMS), and D. Schmidt and G. Sedberry (Marine Resources Research Institute, Charleston, SC) assisted with com- puter analyses. G. Anderson (VIMS), H.P Jeffries (Uni- versity of Rhode Island), and D.G. Mountain (NMFS, Woods Hole) provided water temperature data. D. Mar- tin and T Targett (University of Delaware) provided information on laboratory growth experiments of juve- nile tautog. J. Howe, J. Nestlerode, M. Nizinski, T. Orrell, B. Parolari, K. Rhyu, and J. Vieira assisted with figure preparation. D. Vaughan (NMFS, Beau- fort) provided a critical reference. Earlier drafts of this manuscript benefited from comments by H. Austin, R. Birdsong, P. Briggs, B. Collette, D. Dauer, J. Merriner, J. Musick, R. Rose, and J. Smith. Citations Arrington, J. 1985 The wrecking crews of Wachapreague. Salt Wa- ter Sportsman 46( 10):37-39. Austin, H.M. 1973 Distribution and abundance of ichthyoplankton in the New York Bight during the fall in 1971. NY. Fish Game J. 23:58-72. Bagenal, T.B., & F.W. Tesch 1978 Age and growth. In Bagenal, T.B. led.). Meth- ods for assessment of fish production in fresh waters, p. 101-136. IBP (Int. Biol. Prog.) Handb. 3. Bain, CM. Ill (editor) 1984 The Virginia-North Carolina Fishing Report 8(8):1. Bardach, J.E. 1955 The opercular bone of the yellow perch, Perca flavescens, as a tool for age and growth studies. Co- peia 1955:107-109. Bearden, CM. 1961 List of marine fishes recorded from South Caro- lina. Bears Bluff Lab., Wadmalaw Island, SC, 12 p. Bigelow, H.B., & W.C Schroeder 1953 Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53:1-577. Bleakney, J.S. 1963 Notes on the distribution and reproduction of the fish Tautoga onitis in Nova Scotia. Can. Field Nat. 77:64-65. 62 Fishery Bulletin 91(1). 1993 Briggs, P.T. 1969 A length-weight relationship for tautog from the inshore waters of eastern Long Island. N.Y. Fish Game J. 16:258-259. 1977 Status of tautog populations at artificial reefs in New York waters and effect of fishing. N.Y. Fish Game J. 24(21:154-167. Briggs, P.T., & J.S. O'Connor 1971 Comparison of shore-zone fishes over naturally vegetated and sand-filled bottoms in Great South Bay. N.Y. Fish Game J. 18:15-41. Chenoweth, S.B. 1963 Spawning and fecundity of tautog, Tautoga onitis (L.). Unpubl. M.S. thesis, Univ. Rhode Island, N. Kingston, 60 p. Colton, J.B., W.G. Smith, A.W. Kendall, D.L. Berrien, & M.P. Fahay 1979 Principal spawning areas and times of marine fishes. Cape Sable to Cape Hatteras. Fish. Bull., U.S. 76:911-915. Cooper, R.A. 1965 Life history of the tautog, Tautoga onitis (Linnaeus), from Rhode Island. Unpubl. Ph.D. diss., Univ. Rhode Island, N. Kingston, 153 p. 1966 Migration and population estimation of the tautog, Tautoga onitis (Linnaeus), from Rhode Island. Trans. Am. Fish. Soc. 95:239-247. 1967 Age and growth of the tautog, Tautoga onitis (Linnaeus), from Rhode Island. Trans. Am. Fish. Soc. 96:134-142. Dipper, F.A., & R.S.V. Pullin 1979 Gonochorism and sex-inversion in British Labridae (Pisces). J. Zool. (Lond.) 187:97-112. Dipper, F.A., C.R. Bridges, & A. Menz 1977 Age, growth and feeding in the ballan wrasse, Labrus bergylta Ascanius 1767. J. Fish. Biol. 11:105- 120. Eklund, A.M., & T.E. Targett 1990 Reproductive seasonality of fishes inhabiting hard bottom areas in the Middle Atlantic Bight. Copeia 1990:1180-1184. Feigenbaum, D., and C. Blair 1986 Artificial reef study— a final report. VMRC-83- 1185-616, Va. Mar. Resour. Comm., Newport News, 93 p. Goode, G.B. 1884 Natural history of useful aquatic animals. Pt. 3. The food fishes of the United States. In Fishery in- dustries of the United States. Sec. 1:169-549, 610- 612, 629-681. Goodwin, J.M. IV, & A.G. Johnson 1986 Age, growth, and mortality of blue runner, Caranx crysos, from the northern Gulf of Mexico. Northeast Gulf Sci. 8:107-114. Goodyear, C.P. 1980 Compensation in fish populations. In Hocutt, C.H., & J.R. Stauffer Jr. (eds.), Biological monitoring of fish, p. 253-280. Lexington Books, D.C. Heath & Co., Lexington MA. Gulland, J.A. 1976 Manual of methods for fish stock assess- ment. Part 1. Fish population analysis. FAO Fish. Ser. 3:36-41. Harris, M.J., & G.D. Grossman 1985 Growth, mortality, and age composition of a lightly exploited tilefish substock off Georgia. Trans. Am. Fish. Soc. 114:837-846. Hildebrand, S.P., & W.C. Schroeder 1928 Fishes of Chesapeake Bay. Fish. Bull, U.S. 43:1-366. Hostetter, E.B. 1988 Age and growth of the tautog, Tautoga onitis (Pi- sces, Labridae), from lower Cheasapeake Bay and coastal waters of Virginia. Unpubl. M.S. thesis, Old Dominion Univ., Norfolk VA, 65 p. IGFA (International Game Fish Association) 1986 World record game fishes. Int. Game Fish Assoc, Fort Lauderdale FL, 320 p. 1990 World record game fishes. Int. Game Fish Assoc, Fort Lauderdale FL, 336 p. Jeanty, P. 1984 Performing linear regressions with 1-2-3. PC Magazine 3(20):243-248. Johansen, F. 1925 Natural history of the cunner (Tautogolabrus adspersus Walbaum). Contrib. Can. Biol, (n.s.) 2(17)423-468. LeCren, E.D. 1947 The determination of the age and growth of the perch, (Perca fluviatilus), from the opercular bone. J.Anim. Ecol. 16:188-204. 1951 The length-weight relationship and seasonal cycle in gonad weight and condition in the perch (Perca fluviatilus). J. Anim. Ecol. 20:201-219. Leim, A.H., & W.B. Scott 1966 Fishes of the Atlantic Coast of Canada. Bull. Fish. Res. Board Can. 155, 485 p. Manooch, C.S. HI 1979 Recreational and commercial fisheries for king mackerel, Scomberomorus cavalla, in the south At- lantic bight and Gulf of Mexico, U.S.A. In Nakamura, E.L., & H.R. Bullis Jr. (eds.), Colloquium on the Span- ish and king mackerel resources of the Gulf of Mexico. Proc. Gulf States Mar. Fish Comm. 4:33-41. Manooch, C.S. Ill, & D.L. Mason 1984 Age, growth, and mortality of lane snapper from southern Florida. Northeast Gulf Sci. 7:109-115. Mason, D.L., and C.S. Manooch III 1985 Age and growth of mutton snapper along the east coast of Florida. Fish. Res. ( Amst. ) 3:93-104. Matheson, R.H. HI, & G.R. Huntsman 1984 Growth, mortality, and yield-per-recruit models for speckled hind and snowy grouper from the United States South Atlantic Bight. Trans. Am. Fish. Soc. 113:607-616. Matheson, R.H. HI, G.R. Huntsman, & C.S. Manooch III 1986 Age, growth, mortality, food and reproduction of Hostetter and Munroe: Age. growth, and reproduction of Tautoga onitis 63 scamp, Mycteroperca phenax, collected off North Caro- lina and South Carolina. Bull. Mar. Sci. 38:300-312. McConnell, W.J. 1952 The opercular bone as an indicator of age and growth in the carp, Cyprinus carpio Linnaeus. Trans. Am. Fish. Soc. 81:138-149. Moore, CM., & R.F. Labisky 1984 Population parameters of a relatively unex- ploited stock of snowy grouper in the lower Florida Keys. Trans. Am. Fish. Soc. 113:322-329. Musick, J.A. 1972 Fishes of Chesapeake Bay and the adjacent coastal plain. In Wass, M.L. (ed.), A checklist of the biota of the lower Chesapeake Bay. Va. Inst. Mar. Sci. Spec. Sci. Rep. 65:175-212. Musick, J.A., & L.P. Mercer 1977 Seasonal distribution of black sea bass, Centropristis striata, in the Mid-Atlantic Bight with comments on the ecology and fisheries of the species. Trans. Am. Fish. Soc. 106:12-25. Nelson, R.S., & C.S. Manooch III 1982 Growth and mortality of red snapper in the west- central Atlantic Ocean and northern Gulf of Mex- ico. Trans. Am. Fish. Soc. 111:465-475. Normura, H., & M.S.S. Rodrigues 1967 Biological notes on king mackerel, Scomberomorus cavalla (Cuvier), from northeastern Brazil. Arq. Estac. Biol. Mar. Univ. Fed. Ceara 7(l):79-85. Norusis, M.J. 1985 SPSS-X Advanced statistics guide. McGraw-Hill, NY, 505 p. Nose, Y., T. Sayako, M. Koya, & H. Yoshio 1955 A method to determine the time of ring forma- tion in hard tissue of fishes, especially for the age determination of Pacific tunas. Rec. Oceanogr. Works Jpn. 2:9-18. Olla, B.L., & C. Samet 1977 Courtship and spawning behavior of the tautog, Tautoga onitis (Pisces: Labridae), under laboratory conditions. Fish. Bull, U.S. 75:585-599. Olla, B.L., A.J. Bejda, & A.D. Martin 1974 Daily activity, movements, feeding, and seasonal occurrence in the tautog, Tautoga onitis. Fish. Bull., U.S. 72:27-35. 1979 Seasonal dispersal and habitat selection of cun- ner, Tautogolabrus adspersus, and young tautog, Tautoga onitis, in Fire Island Inlet, Long Island, New York. Fish. Bull., U.S. 77:255-261. Olla, B.L., A.J. Bejda, & A.L. Studholme 1977 Social behavior as related to environmental fac- tors in the tautog, Tautoga onitis. In The behavior of marine organisms: Social behavior and communi- cation; navigation; and development of be- havior. Proc. Annu. Northeast Reg. Meet. Animal Behav. Soc. 1977. Plen. Pap. Mar. Sci. Res. Lab. Tech. Rep. 20, Mem. Univ. St. John's, Newfoundland, p.47- 99. Olla, B.L., A.L. Studholme, A.J. Bejda, & C. Samet 1980 Role of temperature in triggering migratory be- havior of the adult tautog Tautoga onitis under labo- ratory conditions. Mar. Biol. (Berl.) 59:23-30. Olla, B.L., C. Samet, & A.L. Studholme 1981 Correlates between number of mates, shelter availability, and reproductive behavior in the tautog Tautoga onitis. Mar. Biol. I Berl. ) 62:239-248. Perlmutter, A. 1939 An ecological survey of young fish and eggs iden- tified from tow-net collections. In A biological sur- vey of the salt waters of Long Island, Part II, p. 11-71. NY State Conserv. Dep., Albany. Pottle, R.A., & J.M. Green 1979a Field observations on the reproductive behav- ior of the cunner, Tautogolabrus adspersus (Walbaum), in Newfoundland. Can. J. Zool. 57:247-256. 1979b Territorial behaviour of the north temperate labrid, Tautogolabrus adspersus. Can. J. Zool. 57:2337-2347. Reiger, G. 1985 Fishfinder. Salt Water Sportsman 46(10):112- 113. Richards, C.E., & M. Castagna 1970 Marine fishes of Virginia's Eastern Shore (inlet, marsh, and seaside waters). Chesapeake Sci. 11:235- 248. Ricker, W.E. 1975 Handbook of computations for biological statis- tics of fish populations. Bull. Fish. Res. Board Can. 191. 382 p. Robertson, D.R., & J.H. Choat 1974 Protogynous hermaphroditism and social systems in labrid fish. In Proc, Second Int. Symp. Coral Reefs 1:217-225. Robertson, D.R., and R.R. Warner 1978 Sexual patterns in the labroid fishes of the west- ern Caribbean. II. Parrotfishes (Scaridae). Smithson. Contrib. Zool. 255:1-26. Roede, M.J. 1972 Color as related to size, sex, and behavior in seven Caribbean labrid fish species (Genera Thalassoma, Halichoeres, and Hemipterinous). Stud. Fauna Curacao Other Caribb. Is. 42:1-266. Schaaf, W.E., & G.R. Huntsman 1972 Effects of fishing on the Atlantic menhaden stock: 1955-1969. Trans. Am. Fish. Soc. 101:290-297. Scott, W.B., & M.G. Scott 1988 Atlantic Fishes of Canada. Can. Bull. Fish. Aquat. Sci. 219, 731 p. Sedberry, G.R., & H.R. Beatty 1989 A visual census of fishes on a jetty at Murrells Inlet, South Carolina. J. Elisha Mitchell Sci. Soc. 105:59-74. Serchuk, F.M., & C.F. Cole 1974 Age and growth of the cunner, Tautogolabrus adspersus (Walbaum) (Pisces: Labridae) in the Weweantic River estuary, Massachusetts. Ches- apeake Sci. 15:205-213. Sokal, R.R., & F.J. Rohlf 1981 Biometry, 2nd ed. WH. Freeman, San Francisco, 859 p. 64 Fisher/ Bulletin 91(1), 1993 Stolgitis, J.A. 1970 Some aspects of the biology of the tautog, Tautoga onitis (Linnaeus), from the Weweantic River Estuary, Massachusetts, 1966. Unpubl. M.S. thesis, Univ. Mass., Amherst, 48 p. Turner, S.C., C.B. Grimes, & K.W.Abie 1983 Growth, mortality, and age/size structure of the fisheries for tilefish Lopholatilus chamaelonticeps in the middle Atlantic-Southern New England regions. Fish. Bull, U.S. 81:751-763. VMRC (Virginia Marine Resources Commission) 1985 Marine recreational fishing in Virginia. 1985. VMRC Tech. Rep. 87-01, Va. Mar. Resour. Comm., Newport News, 91 p. Warner, R.R. 1975 The reproductive biology of the protogynous her- maphrodite Pimelometopon pulchrum (Pisces: Labridae). Fish. Bull., U.S. 73:262-283. Warner, R.R., & I.F. Downs 1977 Comparative life histories: growth vs. reproduc- tion in normal males and sex-changing hermaphro- dites in the striped parrotfish, Scarus croicensis. In Proa, 3rd Int. Symp. Coral Reefs 1 (Biol. 1:275-282. Warner, R.R., & D.R. Robertson 1978 Sexual patterns in the labroid fishes of the west- ern Caribbean, I: The wrasses. Smithson. Contrib. Zool. 254:1-27. Warner, R.R., D.R. Robertson, & E.G. Leigh Jr. 1975 Sex change and sexual selection. Science (Wash. DC) 190:633-638. Wenner, C.A., W.A. Roumillat, & C.W Waltz 1986 Contributions to the life history of black sea bass, Centropristis striata, off the southeastern United States. Fish. Bull., U.S. 89:723-741. Wheatland, S. 1956 Pelagic fish eggs and larvae. In Oceanography of Long Island Sound, 1952-1954. Part VII. Bull. Bingham Oceanogr. Collect. Yale Univ. 15:234-314. Wilk, S.J. 1977 Biological and fisheries data on bluefish, Poma- tomus saltatrix (Linnaeus). Tech. Sen Rep. 11, Sandy Hook Lab., NMFS Northeast Fish. Sci. Cent., High- lands NJ 07732, 56 p. Abstract. — Studies were per- formed to determine effects of envi- ronment and physiology on the for- mation of daily increments in winter flounder otoliths. Otoliths from embryonic to 1-yr-old laboratory- raised winter flounder Pleuronectes amehcanus and young-of-year wild- caught specimens were examined, and growth patterns were deter- mined from photographs taken on light and scanning electron micro- scopes. Behavioral observations were made from hatching through meta- morphosis. Daily growth increments of oto- liths from larval winter flounder were enumerated, and a growth curve was derived describing the first 2 months of life. Growth was best described by a Gompertz-type curve. The relationship between sagitta size and fish length was ex- ponential for larvae, but linear dur- ing the remainder of the first year. Sagittae were compared with fish length for both wild and laboratory- reared juveniles and exhibited the same relationship for each. The change in relationship between sagitta size and fish length coincided with changes in dimensional growth of the fish. During metamorphosis, swim- ming and feeding modes changed from tail-propelled, upright swim- ming and frequent sudden feeding lunges in larvae, to bottom-resting and creeping accompanied by infre- quent feeding gulps in juveniles. This change reflected the transfer from pelagic to benthic habitat and anatomical transformation to asym- metrical form. In general, juveniles maintained lower activity levels than did larvae. Behavioral and anatomi- cal changes are summarized. Early growth, behavior, and otolith development of the winter flounder Pleuronectes amehcanus Ambrose Jearld Jr. Woods Hole Laboratory, Northeast Fisheries Science Center National Marine Fisheries Service. NOAA 1 66 Water Street. Woods Hole. Massachusetts 02543-1 097 Sherry L. Sass Division of Marine Fisheries, 1 8 Route 6A Sandwich, Massachusetts 02563 Melinda F. Davis Biology Department. Fort Valley State College Fort Valley, Georgia 3 1 030 Manuscript accepted 4 November 1992. Fishery Bulletin, U.S. 91:65-75 1 1993). Many aspects offish development are reflected in otolith structure. Short- and long-term changes in growth rate may be caused by either environmen- tal fluctuations or life history changes (e.g., metamorphosis, spawning), and these events may also be incorporated into the otolith record of sagittae. Hyaline bands have been used for decades to estimate age. Daily growth increments have also been discovered in fish otoliths (Pannella 1971, 1974) and are proving a powerful tool to study larval population dynamics. One daily growth increment in- cludes both a calcium-rich aragonite layer in a protein-poor matrix (the "incremental zone") and a poorly cal- cified protein-rich matrix layer (the "discontinuous zone") (Watabe et al. 1982). Increments are entrained in response to a 24 h light/dark cycle (Taubert & Coble 1977, Tanaka et al. 1981, Radtke & Dean 1982) as well as influenced by other cues (Campana & Neilson 1982, Campana 1984a,b). As a result of the daily cycle in cal- cium deposition, otoliths often reflect fish age, irrespective of growth, al- though this has not always been found to be the case (Geffen 1982, Campana 1983, Jones 1984). Differ- ences in width and other features of daily growth increments have been correlated to life-history transitions, changes in environmental conditions such as temperature and ration size, and physiological factors (for review see Campana & Neilson 1985, Jones 1986). Flounders have a particularly com- plex first year because, not only their habitat, but also much of their body form and behavior changes drasti- cally at metamorphosis ( 35-56 d af- ter hatching). Hatching as symmetri- cal larvae that feed planktonically, they begin to frequent the bottom as their dorsoventral dimension in- creases, the notocord tip bends, and the adult-shaped caudal fin develops (Klein-MacPhee 1978). Finally, the fish spends less time swimming in the water column and becomes benthic. This occurs as one eye mi- grates across the dorsum to the op- posite side of the head and the juve- nile flounder orients at a 90" angle to its previous alignment. Otoliths do not change their position in the head during this transformation (Piatt 1973, Policansky 1982), so the sagittae end up lying one over the other. Evidence of the fish's orien- tational change may be reflected in the otolith depositional pattern. 65 66 Fishery Bulletin 91(1), 1993 Furthermore, physiological changes as well as the change in food source might also register on the otolith. This study is the first in a continuing sequence of investigations on the effects of environment and physi- ology on the formation of subannual increments in win- ter flounder otoliths. Relationships between behavioral and anatomical changes were defined and correlated with baseline data on daily growth increments. Growth rates of both wild and laboratory-reared larvae were determined. Increment counts and morphological changes from known-age laboratory-reared fish were compared with those of wild fish to establish anatomi- cal markers in the otolith during the first year. Materials and methods Acquisition of eggs and larval rearing Adults caught in Narragansett Bay served as gamete sources. Eggs fertilized in the laboratory were acquired from the National Marine Fisheries Service (NMFS) and Environmental Protection Agency (EPA) laborato- ries at Narragansett, Rhode Island during February and March 1981. Larvae hatched 15 March 1981 (termed the Mar 15 group) were reared in static trays using methods of Klein-MacPhee et al. (1980). The light cycle was maintained at 11:13 (light/dark). Light in- tensity in the growth trays varied between 256 and 1777 lux. A separate 114 L tank was maintained for behavioral observations. Light intensity in this tank was 847-4780 lux. Salinity range was 32-33 %,, tem- perature was kept at 5-10° C (±1°C) without diel varia- tion until July, when temperature was allowed to roughly follow seasonal patterns (Fig. 1). Larvae were fed once daily unicellular green algae Tetraselmis souscii and rotifers Brachionus sp., beginning at 3-5 d before yolksac absorption began. Rotifer concentrations of at least 1/mL were maintained, although concentra- tions ranged to over 20/mL. When larvae were 40 d old, newly-hatched brine shrimp were added to main- tain prey concentrations of 1/mL. After several months, fish were fed frozen brine shrimp with the addition of chopped mussels at irregular intervals. Heavy mortalities (>90%) reduced larval populations such that only 30 fish survived through metamorpho- sis ( 40-60 d posthatch). Of the metamorphosed fish, 11 lived 1 yr and then were killed for otolith examination. Due to physiological effects induced by natural mor- tality, only otoliths from sacrificed fish were examined. Sampling Initially, 10 fish per day selected at random were re- moved and preserved in 95% ethanol. After hatching, i 'i '< '1 'i '1 '1 '1 '1 '1 ', '1 ', ', ', ', >981 1982 Figure 1 Water temperatures showing mean and range of bimonthly temperatures. Arrows represent hatch-group dates (day of 50% hatch) of winter flounder Pleuronectes americanus. fish were sampled at 1-6 d intervals until metamor- phosis. In an effort to conserve the 30 fish surviving metamorphosis, additional samples of laboratory-reared larvae in 95% ethanol were provided for ageing by the NMFS Narragansett Laboratory (hereafter referred to as the LR hatch group). These samples were of larvae 0-55 d old and were reared using the same methods as the Mar 15 hatch group. Standard length measurements were obtained from preserved fish. To check for shrinkage in body length, larvae were measured before and after 1 week's pres- ervation and percent shrinkage determined. Collections Wild young-of-the-year (YOY) winter flounder were col- lected throughout summer 1981 using beach seines from estuaries along Cape Cod. Otoliths were dissected from these specimens. Otolith preparation and analyses Body lengths of preserved larvae were measured on glass slides. In larvae (hatch to -30 d), sagittae could not be distinguished from the asterisci or lapilli; there- fore all otolith pairs were removed and aged. In fish Jearld et al.: Early growth, behavior, and otolith development of Pleuronectes amencanus 67 . M < Wjt Figure 2 Light microscope (A) and SEM (B) photographs of winter flounder Pleuronectes americanus otoliths showing similarity of detail visible in both. Scale bars represent 10 p.. 30 d and older, sagittae were measured and used for increment counts. Otoliths were measured to the nearest micron (under a com- pound microscope) along the longest axis through the central core and along the axis perpen- dicular to that dimension using an optical micrometer at 200- lOOOx (depending on otolith size). Most increment counts were done on photographs at 1000 X. All increments visible in at least two places on an otolith were counted. Varying the focus changed the resolution of incre- ments; therefore the maximum number of increments seen in a series of pictures taken at slightly-varying focal planes was counted. Two or three separate counts by two age-readers were averaged. If the two readers dis- agreed by more than two incre- ments or the photographs were considered unclear, that otolith set was not used in daily growth- increment calculations. Incre- ments formed prior to yolksac absorption were either absent or difficult to resolve and were not included in the total count. Based on the work of Radtke & Scherer (1982), a correction factor of 10 was added to the number of ob- served increments in order to es- tablish each larva's estimated age in days from hatch. For scanning electron micros- copy (SEM) viewing, some of the larger ( 600-840 u along the larg- est dimension) sagittae were pre- pared according to the methods in Radtke & Dean (1982). Light microscope pictures of an otolith were compared with SEM photo- graphs of the same specimen (Fig. 2) and counts were found to be comparable. Behavioral observations Larval behavior was observed from hatching through metamor- 68 Fishery Bulletin 91(1). 1993 as 35 10 15 americaiuis. phosis to correlate behavior with physiological (as registered in otolith development) and mor- phological changes. Observations of the larvae were made in hold- ing tanks throughout the period they were reared. Individual lar- vae were also observed in small containers under low magnifica- tion to verify anatomical changes as well as details of small move- ments. On the 48th day after hatch- ing, individuals were moved into an observation tank to facilitate observation. An undergravel filter bed was placed in the re- frigerated observation tank to minimize disturbance resulting from maintenance procedures. Temperature, diets, and light- cycle conditions were the same as those for separate tray-reared larvae. Light intensity was higher in the observation tank than in trays because overhead lights were supplemented with tank lights. Fish were observed for lOmin periods twice daily, at 10 a.m. and 4 p.m. One fish chosen at random was followed as long as it could be seen; if it moved out of sight, another individual was selected for the remainder of the obser- vation period. Behaviors recorded included swimming (duration, vertical and horizontal direction, body ori- entation in relation to the bottom, and fin usage); feed- ing, both before and after adding food (frequency, loca- tion, sequence of body motions, success); resting (duration, location, body position); and interactions be- tween individuals. Observations were terminated sev- eral weeks after fish had metamorphosed and behav- ior patterns had stabilized (i.e., assumed a typical adult sedentary behavior pattern). Results and discussion Larval growth rates From analysis of 113 preserved larval winter flounder ranging from 2.5 to 9.0mmSL, growth was best de- scribed by a Gompertz-type curve (Fig. 3). Previous uses of the Gompertz growth curve and methodology for fitting the curve are described in Pennington ( 1979) and Bolz & Lough (1988). The variance was stabilized by using the natural log form of the growth equation, and parameters were derived by nonlinear estimation techniques resulting in the relationship: u 9 - » - — *- 8 7- 6 5 - " . • ^* ^^-—r7~T~~ ! 4 3 a > 1 .■■ b -0.0741Age -2.5329e SL 8.8994e 2 n 113 r 2 0.8106 1 1 i iiit 20 25 30 35 40 Estimated Age (d) 45 50 55 60 Figure 3 Gompertz growth curve and equation fitted to plot of standard length and estimated age in days Ino. of otolith increments + 10) for 113 larval winter flounder Pleuronectes ln(L) = -0.3469+2. 5329(l-e- r-=0.8106, (1) where L = standard length in mm, and R = estimated age (increments +10) in days. The predicted length of 2.66 mm at yolksac absorp- tion compares favorably with that found by Radtke & Scherer (1982) for wild larvae (2.5 mm). The asymptotic length of 8.9 mm probably delineates mean length at metamorphosis and falls within the range (7-13 mm) given by Fahay (1983). The average growth rate (from Eq. 1) for the period under study was 0.31 mm/d, which is slightly less than that observed in the 1982 study using preserved lengths by Radtke & Scherer (0.38 mm/d). Shrinkage Little shrinkage was observed for larvae 4-35 d old as has been reported by other researchers (Radtke & Waiwood 1980, Theilacker 1980). Fresh lengths were 2.8-5. 0mm; preserved lengths, 2. 5-5. 0mm (n=28) with average shrinkage of 4.2<7c (SE=0.6). For 91-112d old flounder, fresh lengths were 5.9-13.8 mm; preserved, 5.8-13.6 mm (n=19) with average shrinkage of 8.6% (SE=0.9). Radtke & Scherer (1982) found no shrinkage in small larvae (<4.7mm) and only minimal shrinkage (4%) in older fish. Though we observed slightly larger shrinkage than Radtke & Scherer (1982), the growth rates during the 55 d agree roughly with larval flounder growth rates found in their study. Jearld et al.: Early growth, behavior, and otolith development of Pleuroneaes amencanus 69 Figure 4 (A) Otoliths taken from winter flounder Pleuronectes amerkanus embryo 13d after spawning, showing primordial granules in center. (B) Rings in otolith taken from em- bryo 16 d after spawning. Magnification 1000 x. Otolith development Pre-hatch formation All three pairs of otoliths were present in embryos as early as 13 d after spawning. Premordial granules of material were evident at this time, clumped together in the otolith core (Fig. 4A). The periphery of the otolith forms a rather irregular sphere. Up to four growth rings could be seen on some embryonic otoliths (Fig. 4B). Similar formations have been found on embryonic otoliths of several other spe- cies (Taubert & Coble 1977, Brothers & McFarland 1981, Radtke & Dean 1982, Geffen 1983, Brothers 1984), but their periodicity or significance has not yet been determined. Shape change at metamor- phosis At the time of eye mi- gration (40-50 d posthatch), the sagittae of winter flounder un- derwent a profound change in shape. Clumps of what seemed to be amorphous calcareous ma- terial accumulated at the otolith periphery. These accessory growth centers developed irregu- larly, sometimes appearing two or three on an otolith, often forming at 90° intervals around the circumference of the previ- ously round otolith (Fig. 5A). Similar observations have been made in other Pleuronectids (Brothers 1984, Campana 1984c). It is significant that accessory growth centers were found only on otoliths of flounders during and after metamorphosis. Lar- vae with symmetrically-placed eyes did not show these irregu- lar formations on their sagittae, even as late as 73 and 76 d posthatch (Fig. 5B). A photo- graph of a non-metamorphosing 73d-old larval otolith without accessory growth centers is com- pared with that of the typical otolith from a metamorphosing individual in Figure 5. The ap- pearance of asymmetrical forma- tions on sagittae of metamor- phosing flounders coincides with the change from vertical to hori- zontal orientation (i.e., dorsal- side uppermost to right-side up- permost) which accompanies the shift to an asymmetrical form and a benthic habitat. Thereaf- ter, accretion again seems to proceed by increments which coincide with age in days, but which continue an asymmetrical deposition until the adult shape is stabi- lized. If the formation of accessory growth centers is found to occur at the time of metamorphosis in other flounder species (Brothers 1984, Campana 1984c), this may prove useful in marking the point of habitat change within the otolith record. It is possible that accurate otolith counts could then begin with the juvenile stage rather than the earlier, less easily prepared, and counted larval otoliths. Fish length/otolith length relationship during the first year The relationship between sagitta size (larg- 70 Fishery Bulletin 91(1). 1993 <> B Figure 5 (A) Accessory growth centers on sagitta of metamorphosing winter flounder Pleuronectes amerieanus. (B) Symmetrical otolith from 73 d-old winter flounder that had not yet under- gone metamorphosis, showing continuing lack of accessory growth centers. est dimension) and fish length was nonlinear for pre- metamorphic larvae raised under laboratory conditions (Fig. 6). The best fit equation was exponential, Y = 7.8e 03 * (r 2 =0.87), where x is standard length and Y is otolith length. Because larger larvae had an average shrinkage of 8.69? as compared with 4.2% for smaller larvae, the parameter estimates in the above equation may not be bias-free. By the end of the first year, however, the relationship was linear (Fig. 7). Covariate analysis in- dicated that the regression lines for laboratory-reared and wild YOY winter flounders were not significantly different, so data pairs were pooled. The resulting re- gression line for YOY flounders (about 3 mo or older) using the same variables as above was Y = 0.10+0.29x (r 2 =0.95). Both linear and allometric relationships between oto- lith size and larval fish length have been reported in the literature (Taubert & Coble 1977, Brothers & McFarland 1981, Methot 1981, Radtke & Dean 1981). Although this relationship has been reported for starry flounder (Campana 1984c), it has not been previously reported for larval winter flounder. It is not surpris- ing, however, that otolith growth exceeds growth in body length. Addition to body depth is enhanced as the body form alters towards the adult shape. This feature may compensate for the decline in growth in length at this time (Pearcy 1962, Laurence 1975). Investigating the relationships between length, depth, otolith dimen- sion, otolith mass, and fish mass may in future studies elucidate the relationship between larval flounder so- matic growth and otolith growth. Wild vs. laboratory-reared fish Otoliths from wild juvenile samples showed the same diameter/fish length relationship as otoliths from our laboratory-reared fish. Otoliths from wild fish exhibited a more regular and somewhat sharper depositional pattern of increments. Therefore, these were used more frequently for SEM analysis. The superior clarity and regularity of otolith incremental patterns from wild fish over laboratory- reared fish have been discussed in the literature (Blaxter 1975, Uchiyama & Struhsaker 1981, Radtke & Dean 1982, Radtke & Scherer 1982) and is discussed in detail by Campana & Neilson (1985). Early larvae: Hatch to 40 d Under a temperature regime of 5-7°C, larvae hatched 14-18 d after being spawned. Hatching was accompanied by intermittent writhing and vibrating motions of the embryos. Hatch- lings sank to the tray bottom when not in motion. Swimming began immediately after hatching. Lar- vae swam with rapid lateral tail-whips and could con- trol direction. All swam away from disturbances caused by a pipette tip. Swimming became stronger and more sustained over the first 10 d posthatch. On the third day after hatching, a series of brief (l-8s) upward swims, followed by a 20 s to lmin passive period re- sulting in head-down sinkings, were first noted. This intermittent swimming behavior may be adaptive in Jearld et al.: Early growth, behavior, and otolith development of Pleuronectes amencanus • • i 200 / • / • / a. • / g 150 _ /• 5 / < i/> 2 o 1 CO / • z 1 Ui s 1 100 I H / * S3 / • o cc 3 50 / • • • 3456789 10 STD LENGTH ( MM ) Figure 6 Exponential relationship between sagitta size and standard fish length of winter flounder Pleuronectes americanus. Y=7.8e 03x , r^O.87. Curve represents all points to 94 d posthatch for Feb. 27 hatch data only. 2.50 6 7 STD LENGTH (CM) Figure 7 Regression of standard fish length on sagittae size for pooled young-of-the-year winter flounder Pleuronectes americanus (3 mo or older). Pooled data include both laboratory-reared and wild collected data (n=28). Y=0.10+0.29X, r 2 =0.95. protecting young larvae from extensive transport by surface currents (Sullivan 1915, Pearcy 1962). Sullivan reported that swimming appeared to be periodically inhibited by a factor other than fatigue. By 7d, larvae swam for 3-10 s followed by a 2-10 s passive period. After about 25-30 d, swimming was constant during the day except when interrupted by feeding behaviors. For the first 10 d posthatch, swimming larvae were concen- trated at the surface; after about 20 d, larvae were dispersed throughout the water column. There were noticeable aggre- gations of larvae at tank corners and along sides, but no schooling behavior was observed. Swimming larvae avoided bright microscope lights. Lunging — defined as a sudden, rapid thrusting motion of the body which propels the larva less than a centimeter for- ward but which is faster and more abrupt than swimming activity — was observed as early as 5-7 d posthatch (depend- ing on hatch group), before the mouth was completely formed. Once the mouth parts had formed (at the time of yolksac absorption), lunges included rapid and wide jaw gape and snap. Incidence of lunging increased from less than once a minute initially to once every 10 s or more frequently by 40 d posthatch. Systematic observations of lunges began at day 48 (Fig. 8). Prey items were not always visible but were seen often enough that these movements were assumed to be feed- ing lunges. No lunges were observed after day 72 coincident with metamorphosis. A sigmoid coiling of the body, called an "S motion" here, often preceded the lunge. This motion could be slow or fast, and when rapid often included a single side lunge as the body was pulled backwards and the head whipped from one side of the "S" to the other. The rapid "S" was first observed 5-9 d posthatch, while the slow one was not noted until 30- 50 d posthatch. Larval feeding by such an "S strike" motion has been described for other species in the literature (Rosenthal & Hempel 1970, Hunter 1972). The slow "S motion" we observed in older larvae is speculated to be related to the greater accuracy of striking prey facili- tated by experience. Successful feeding, defined by observation of at least one food particle in the gut of sampled larvae, began at 9-14 d posthatch, at or just after yolksac absorption (8-12d posthatch). However, growth has been reported to slow or stop for several days after absorption of the yolksac (Cetta & Capuzzo 1982). Passive, nonswimming yolksac larvae sank in a head- down position in the water column until they hit bot- tom or abruptly resumed swimming towards the sur- face. When on the tray bottom, they lay on either side, or on top, of their yolksac. As swimming duration in- creased, time on the tray bottoms decreased until, af- ter 20 d (when the yolksac was no longer present), few were seen on the bottom. Passive, nonswimming be- havior in the water column was, however, observed past 20 d. After yolksac absorption (-12 d posthatch) 12 Fishery Bulletin 91(1). 1993 26 - 22 - 18 14 10 _4 e A A \ 2 *v A 52 64 68 72 POSTHATCH AGE, DAYS - Figure 8 Incidence of winter flounder Pleuronectes amerieanus feeding behaviors, average num- ber of lunges in water column lA) versus average number of gulps on bottom ( ). Note change in feeding behavior from frequent pre-metamorphic lunging to less-frequent post-metamorphic gulping. No lunging was observed after day 72. the larvae began to maintain their bodies in a horizon- tal position as they sank, instead of sinking in the vertical, passive, nonswimming position. Behavior at metamorphosis: 40-60 d posthatch Eight- een fish were placed in the observation tank on day 48 so that their behavior could be more closely monitored. The following account is based on those observations. By 40-50 d posthatch, larval body form began to change. The body widened dorsoventrally, pigment de- veloped (especially over the head, jaw, gut, and fins), and the end of the notocord began to bend as the adult caudal fin formed. These physical changes, described more fully by other researchers (Sullivan 1915, Breder 1922), took place concurrently with the behavioral changes described below. The most obvious behavioral changes were seen in swimming and resting patterns. Larvae up to 50-60 d posthatch swam upright using the tail-whipping mo- tion. The body was positioned with dorsal fin upper- most, and the eye had not yet migrated. Beginning at 48 d, occasional interruption of swimming was noted as fish drifted in the water column, usually maintain- ing an upright posture but not moving fins or tail. At 55 d, the first observation of canted swimming was recorded. Fish rose off the bottom in response to dis- turbances and swam at about a 60" angle to the side, then sank to the bottom again. Only larvae whose eye was in the process of migration (asymmetrical placement) were seen to swim at an angle in this way, and only infrequently were these individuals observed swim- ming in the water column. Fish with obviously widened bodies and adult-shaped tails be- gan the transition to bottom habitat just prior to eye migra- tion ( -40-50 d posthatch). Eight fish were observed lying on the bottom by 42 d posthatch, either on their ventral or left sides. Ap- parently, some of these individu- als returned to larval swimming patterns, since only three were not in the water column on day 48. From day 42 until the last record of fish seen in the water column (day 72), swimming fish were seen to sink to the bottom, either head-down or horizontally oriented, and usually rested on their left side for varying peri- ods before swimming up from the sand, again with an upright lar- val swimming posture (Fig. 9). Other observations of left-side resting have been reported as early as 10- 12 d posthatch (Sullivan 1915). Eye migration was difficult to observe precisely. It has been reported (Sullivan 1915) to occur over an interval of several days, and we observed that it ap- peared to occur shortly after establishment of the de- veloping larva on the tank bottom. Once eye migration was completed, fish were never seen to swim upright. However, they occasionally entered the water column lying horizontally on the left side, slowly rippling their dorsal and ventral fins rhythmically. Newly metamorphosed fish were relatively inactive, lying on the bottom for 10 min or more at a time, occa- sionally moving their eyes. Activity and metabolic lev- els have been reported elsewhere to decrease dramati- cally at this point in flounder development (Blaxter & Staines 1971, Laurence 1977). The first slow bottom swimming with rippling dorsal and ventral fins or "creeping" activity was noted on day 55 during meta- morphosis. Bottom-resting fish also "darted" at intervals, swim- ming in a rapid burst propelled by tail beats. A dart appeared somewhat like a long lunge in its sudden- ness and in its generation by caudal body motions rather than fin motions. Darting was also observed in unmetamorphosed fish in the water column as an in- frequent reaction to a disturbance. The earliest obser- Jearld et al.: Early growth, behavior, and otolith development of Pleuronectes amencanus 73 vation of darting in bottom-resting fish occurred on day 54. Darting persisted as a sporadic activity in metamorphosed ju- veniles. Metamorphosing fish fed in the water column as did unmetamorphosed larvae. Feeding behavior was observed through- out the daylight period. Other research- ers have noted the strictly diurnal feed- ing behavior of young winter flounder (Laurence 1977). Newly metamorphosed fish resting on the bottom were observed "gulping" on day 55. Gulping was a relatively inactive feeding behavior, with the jaw gape and snap found in the lunge but not accom- panied by other body movements. Post- metamorphic juveniles increased the in- cidence of gulps with age, sometimes combining a short creep and a gulp but often showing no other sign of active feed- ing. After metamorphosis is complete, in- creased feeding efficiency, coupled with decreased metabolic requirements, re- sults in comparatively low energy expenditures associ ated with feeding behavior (Blaxter & Staines 1971 Laurence 1977). Conclusions In this study, larval winter flounder otoliths were found to reflect internal and external changes indirectly. Daily growth increments did not begin immediately after hatching, although some individuals exhibited otolith rings at hatch. Daily growth increments were not vis- ible beginning at yolksac absorption. Rather, incre- ments were visible beginning at a point midway be- tween yolksac absorption and the beginning of metamorphosis. These increments may have reflected internal changes presaging metamorphosis, although such changes are not yet externally evident. The pe- riod during which these increments appeared was also the period during which swimming behavior during the day became constant, except when it was inter- rupted by feeding behavior. Shortly after this time- period, the slow "S motion" was first exhibited as a feeding behavior, possibly correlated with experience and better feeding efficiency. Metamorphosis resulted in obvious behavioral changes as well as anatomical ones. Swimming began with an upright position in which the tail was whipped back and forth to provide propulsion. Larvae then went through a period of canted swimming before settling into swimming on their sides using a rippling of their '/.UNMETAMORPHOSED LARVAE RESTING ON BOTTOM % OEVELOPING {EYE ASYMMETRICAL) OR METAMORPHOSEO FLOUNDER RESTING ON BOTTOM 70 80 90 POSTHATCH AGE, DAYS - 100 Figure 9 Swimming vs. resting behavior of pre-metamorphosed versus developing (eye asymmetrical I or post-metamorphosed winter flounder Pleuronectes americanus. 3-18 fish were observed in each 10 min period. Average number observed/period was 7. fins for most propulsion. Horizontal swimming with rippling fins was associated with the change to a more benthic existence. The horizontal swimming style oc- curred less frequently than the earlier upright one, which was consistent with a decrease in overall activ- ity levels. Individuals spent increasing amounts of time resting on the bottom as metamorphosis progressed. At this time, slow bottom "creeping" was first exhib- ited along with the relatively inactive "gulping" feed- ing behavior. The external changes in body shape at metamor- phosis corresponded with internal change in sagittae shape. Otoliths from metamorphosed juveniles were found to exhibit accessory growth centers. Sagittae from older fish that had not yet undergone metamorphosis did not exhibit these centers and consequently were still spherical. The connections between fish growth, behavioral ecol- ogy, and physiology are still not well understood. There is clearly a connection between behavior and physiol- ogy, although the causality of change in morphology, growth, and behavior is still unclear. This study at- tempted to correlate some of the behavioral and ana- tomical changes. Acknowledgments We gratefully acknowledge the technical assistance pro- vided by the staff of the Fishery Biology Investigation of the Northeast Fisheries Science Center, National 74 Fishery Bulletin 91 |l). 1993 Marine Fisheries Service, Woods Hole, Massachusetts. We especially appreciate the aid provided by Louise Dery. We would like to acknowledge the student train- ees that made the project possible. Our sincere appre- ciation to Dr. Werner Graf of Rockefeller University, New York City, who provided his expertise and sugges- tions. Our thanks also to Dr. Grace Klein-MacPhee, Graduate School of Oceanography, University of Rhode Island, Narrangansett Bay Campus, Kingston, and Alphonse Smigielski, National Marine Fisheries Ser- vice, Narrangansett, Rhode Island, for their assistance and suggestions on the proper maintenance of speci- mens. Finally, we wish to gratefully acknowledge George R. Bolz, Dr. Michael Pennington, NMFS, Woods Hole, and Dr. Bruce Collette, NMFS, National Sys- tematics Laboratory, National Museum of Natural His- tory, Washington, DC, for their reviews and valuable technical assistance. Citations Blaxter, J.H.S. 1975 The eyes of larval fish. In Ali, M.H. (ed.), Vision in fishes; New approaches in research, p. 427- 443. Plenum, NY. Blaxter, J.H.S., & M.E. Staines 1971 Food searching potential in marine fish larvae. In Crisp, D.T. (ed.), Fourth European ma- rine biology symposium, p. 467-485. Cambridge Univ. Press, Cambridge. Bolz, G.R. & R.G. Lough 1988 Growth through the first six months of Atlantic cod, Gadus morhua, and haddock, Melanogrammus aeglefinus, based on daily otolith increments. Fish. Bull., U.S. 86:223-235. Breder, CM. Jr. 1922 Some embryonic and larval stages of the winter flounder. Bull. Bur. Fish. 1922:311-315. Brothers, E.B. 1984 Otolith studies. In Moser, et.al. (eds.), Ontog- eny and systematics of fishes, p. 50-57. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol. Allen Press, Lawrence KS. Brothers, E.B., & W.N. McFarland 1981 Correlations between otolith microstructure, growth, and life history transitions in newly recruited French grunts (Haemulon flavolineatum ) (Dersmarest Haemulidae). Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:369-374. Campana, S.E. 1983 Feeding periodicity and the production of daily growth increments in otoliths of steelhead trout (Salmo gairdneri) and starry flounder (Platichthys stellatus). Can. J. Zool. 61:1591-1597. 1984a Interactive effects of age and environmental modifiers on the production of daily growth incre- ments in otoliths of plainfin midshipman, Porichthys notatus. Fish. Bull., U.S. 82:165-177. 1984b Lunar cycles of otolith growth in the juvenile starry flounder Platichthys stellatus. Mar. Biol. (Berl. 180:239-246. 1984c Microstructural growth patterns in the otoliths of larval and juvenile starry flounder, Platichthys stellatus. Can. J. Zool. 61:1507-1512. Campana, S.E., & J.D. Neilson 1982 Daily growth increments in otoliths of starry flounder (Platichthys stellatus) and the influence of some environmental variables in their pro- duction. Can. J. Fish. Aquat. Sci. 39:937-942. 1985 Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. Cetta, CM., & J.M. Capuzzo 1982 Physiological and biochemical aspects of embry- onic and larval development of the winter flounder Pseudopleuronectes americanus. Mar. Biol. (Berl.) 71:327-337. Fahay, M.P. 1983 Guide to the early stages of marine fishes occur- ring in the western North Atlantic Ocean, Cape Hatteras to the southern Scotian Shelf. J. North- west Atl. Fish. Sci. 4:1-423. Geffen, A.J. 1982 Otolith ring deposition in relation to growth rate in herring (Clupea harengus) and turbot (Scoph thai mum maximus) larvae. Mar. Biol. (Berl.) 71:317-3;26. 1983 The deposition of otolith rings in Atlantic salmon, Salmo salar L., embryos. J. Fish Biol 23:467-474. Hunter, J.R. 1972 Swimming and feeding behavior of larval anchovy, Engraulis mordax. Fish. Bull., U.S. 70:821-838. Jones, CM. 1984 Using earstones to determine age in very young fish. Maritime 28(41:11-12. 1986 Determining age of larval fish with the otolith increment technique. Fish. Bull., U.S. 84:91-103. Klein-MacPhee, G. 1978 Synopsis of biological data for the winter flounder, Pseudopleuronectes americanus. Walbaum. NOAA Tech Rep. NMFS Circ. 414, FAO Fish. Synop. 117, 43 p. Klein-MacPhee, G., W.H. Howell, & A.D. Beck 1980 Nutritional value of five geographical strains of Artemia to winter flounder Pseudopleuronectes americanus larvae. In Personne, G.V., P. Sorgeloos. O. Roels, & E. Jaspers (eds.), The brine shrimp Artemia, vol. Ill, p. 305-312. Universa Press, Wetteren, Belgium. Laurence, G.C 1975 Laboratory growth and metabolism of the winter flounder, Pseudopleuronectes americanus, from hatch- ing through metamorphosis at three tem- peratures. Mar. Biol. (Berl.) 32:223-229. 1977 A bioenergetic model for the analysis of feeding and survival potential of winter flounder, Pseudo- pleuronectes americanus, larvae during the period Jearld et al.: Early growth, behavior, and otolith development of Pleuronectes americanus 75 from hatching to metamorphosis. Fish. Bull., U.S. 75:529-546. Methot, R.D. Jr. 1981 Spatial convariation of daily growth rates of lar- val northern anchovy, Engraulis mordax, and north- ern lampfish, Stenobrachius leucopsarus. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:424-431. Panella, G. 1971 Fish otoliths: daily growth layers and periodical patterns. Science (Wash. DC) 173:1124-1127. 1974 Otolith growth patterns: an aid in age determi- nation in temperate and tropical fishes. In Bagenal, T.B. (ed.), Proa, Int. symp. ageing of fish, p. 28- 39. Unwin Bros., Surrey, England. Pearcy, W.G. 1962 Ecology of an estuarine population of winter flounder, Pseudopleuronectes americanus (Wal- baum). Bull. Bingham. Oceanogr. Collect. Yale Univ. 18:39-78. Pennington, M.R. 1979 Fitting a growth curve to field data. In Ord, J.K., G.P. Patil, & C. Taille (eds.l, Statistical distribu- tions in ecological work, p. 419-428. Int. Coop. Publ. House, Fairfield MD. Piatt, C. 1973 Central control of postural orientation in flatfish I. Postural change dependence on central neural changes. J. Exp. Biol. 59:491-521. Policansky, D. 1982 Influence of age, size and temperature on meta- morphosis in the starry flounder, Platichthys stellatus. Can. J. Fish. Aquat. Sci. 39:514-517. Radtke, R.L. & J.M. Dean 1981 Morphological features of the otoliths of the sail- fish, Istiophorus platypterus, useful in age deter- mination. Fish. Bull., U.S. 79:360-367. 1982 Increment formation in the otolith of embryos, larvae, and juveniles of the mummichog, Fundulus heteroclitus. Fish. Bull., U.S. 80:201-215. Radtke, R.L. & M. Scherer 1982 Daily growth of winter flounder (Pseudo- pleuronectes americanus) larvae in the Plymouth Har- bor estuary. In Bryan, C.F, J.U. Connors, & F.M. Truesdale, (eds.), Louisiana Coop. Fish Res. Unit, p. 105. Louisiana State Univ., Baton Rouge. Radtke, R.L. & K.G. Waiwood 1980 Otolith formation and body shrinkage due to fixation in larval cod iGadus morhual). Can. Tech. Rep. Fish. Aquat. Sci. 929, 10 p. Rosenthal, H. & G. Hempel 1970 Experimental studies in feeding and food require- ments of herring (Clupea harenqus L.) larvae. In Steele, J.H. (ed.), Marine food chains, p. 334- 364. Univ. Calif. Press, Berkeley. Sullivan, W.E. 1915 A description of the young stages of the winter flounder {Pseudopleuronectes americanus Wal- baum). Trans. Am. Fish. Soc. 44:125-136. Tanaka, K., Y. Mugiya, & J. Yamada 1981 Effects of photoperiod and feeding on daily growth patterns of juvenile Tilapia nilotica. Fish. Bull., U.S. 79:459-466. Taubert, B.D., & D.W. Coble 1977 Daily rings in otoliths of three species of Lepomis and Tilapia mossambica. J. Fish. Res. Board Can. 34:332-340. Theilacker, G.H. 1980 Changes in body measurements of larval north- ern anchovy, Engraulis mordax, and other fishes due to handling and preservation. Fish. Bull., U.S. 78:685-692. Uchiyama, J.H., & P. Struhsaker 1981 Age and growth of skipjack tuna, Katsuwonus pelamis, and yellowfin tuna, Thunnus albacares, as indicated by daily growth increments of sagittae. Fish Bull., U.S. 79:151-162. Watabe, N.K. Tanaka, J. Yamada, & J.M. Dean 1982 Scanning electron microscope observations of the organic matrix in the otolith of the teleost fish, Fundulus heteroclitus (Linnaeus) and Tilapia nilotica (Linnaeus). J. Exp. Mar. Biol. Ecol. 58:127-134. AbStraCt.-The larval develop- ment of three roughy species com- plexes, Paratrachichthys sp., Aulo- trachichthys sp., and Optivus sp., is described and illustrated using lar- vae collected from Tasmanian and New South Wales waters. Larvae were identified using meristic and morphological characters and are characterized by differences in head and dermal spination, size at cau- dal flexion, and size and pigmenta- tion of the pelvic fins. Head spination is well developed in Aulotra- chichthys, and weak in Optivus and Paratrachichthys. Dermal spination is well developed in postflexion Aulo- trachichthys and flexion Optivus, but absent in Paratrachichthys. Devel- opment of a luminous organ and an- terior migration of the anus occur much earlier in Aulotrachichthys than Paratrachichthys and are no- tably absent in Optivus. The use of these larval characters in tra- chichthyid systematics, and the pos- sible reasons for the absence in our samples of larvae attributable to or- ange roughy Hoplostethus atlanticus, are discussed. Larval development of three roughy species complexes (Pisces: Trachichthyidae) from southern Australian waters, with comments on the occurrence of orange roughy Hoplostethus atlanticus Alan R. Jordan CSIRO Division of Fisheries, GPO Box I 538, Hobart, Tasmania 700 1 . Australia Present address Sea Fisheries Research Laboratories, Department of Primary Industry and Fisheries, Crayfish Point, Taroona, Tasmania 7053, Australia Barry D. Bruce South Australian Department of Fisheries, GPO Box 1 625, Adelaide 500 1 . South Australia Present address: CSIRO Division of Fisheries, GPO Box Australia 538, Hobart, Tasmania 7001. Manuscript accepted 2 November 1992. Fishery Bulletin, U.S. 91:76-86 (1993). The family Trachichthyidae (order Beryciformes) consists of some 31 species, of which at least 15 occur in southern temperate waters of Aus- tralia (May & Maxwell 1986). Seven genera are known from Australian waters: Hoplostethus, Paratra- chichthys, Aulotrachichthys, Optivus, Gephyroberyx, Trachichthys, and Sorosichthys. The various species in- habit depths from nearsurface to greater than 1200 m, with most oc- curring in depths greater than 200 m. There is considerable confusion re- garding the taxonomy of the group, and a number of the species occur- ring in Australian waters are un- described. The genus Hoplostethus comprises at least three Australian species — H. intermedins, H. latus, and orange roughy H. atlanticus (May & Maxwell 1986)— which sup- ports a recently developed fishery. Paratrachichthys is represented in Australian waters by the sandpaper fish, Paratrachichthys sp., an un- described species that is closely re- lated to, and has only recently been distinguished from, the New Zealand endemic P. trailli (M. Gomon, Mus. Victoria, Melbourne, pers. commun., Nov 1990). Specimens have been re- corded from New South Wales, Victoria, South Australia, Tasmania, and southern Western Australian waters (May & Maxwell 1986). Aulotrachichthys and Optivus each contain two closely-related unde- scribed Australian species (M. Gomon, pers. commun. ). Aulotrachichthys sp.l occurs in shallow waters of South Aus- tralia, whereas Aulotrachichthys sp.2 occurs in deeper water off the east coast (May & Maxwell 1986). Sim- ilarly, Optivus has both eastern and western Australian representa- tives: Optivus sp.l along the east coast as far north as southern Queensland, and Optivus sp.2 off southern Western Australia (May & Maxwell 1986). The remaining three genera are monotypic, represented by Gephyroberyx darwini, Trachichthys australis, and Sorosichthys anannassa. Little is known about the ecology or early life history of trachichthyids, and there is nothing in the litera- ture on the early life history of Aus- 76 Jordan and Bruce: Larval development of three roughy species 77 tralian species. Parr (1933) and Johnson (1970) de- scribed 19mm and 21.5mm juvenile Korsogaster, re- spectively, a genus subsequently synonymized with Hoplostethus by Woods & Sonoda (1973). Crossland (1981) illustrated a trachichthyid larva, possibly Optivus elongatus, from northeastern New Zealand. Robertson ( 1975) described an egg tentatively ascribed to Paratrachichthys trailli. Kotlyar (1984) described juveniles of four species of Hoplostethus (including a 36mm H. atlanticus), the smallest of his specimens being a 15 mm H. melanopterus. Okiyama (1988) figured and briefly described single specimens of an unidentified Hoplostethus (10.7mm), Gephyroberyx japonicus (11.0 mm ), and Paratrachichthys prosthemius (27.5 mm). Comparatively little is known of trachichthyid lar- val characters; however, common characters include precocious pelvic fin development, heavy pigment, a stocky body form approaching the shape of adults (in larger larvae), a myomere count of 26-30, and the pres- ence of minute spines over the body surface ( Keene & Tighe 1984. and references therein). Recent commercial interest in the orange roughy Hoplostethus atlanticus has emphasized the need for information on the early life history of this species. As yet, despite considerable effort in ichthyoplankton sam- pling, no H. atlanticus larvae have been reported from Australian waters. The current study describes the larval development of three trachichthyid species com- plexes — Paratrachichthys sp., Aulotrachichthys sp., and Optivus sp. — in specimens obtained from plankton samples collected primarily in Tasmanian and New South Wales coastal waters from 1984 to 1986. These descriptions are presented in order to further define larval characters that may be of use in trachichthyid systematics and future identification of other trachichthyid larvae, including//, atlanticus. Materials and methods Specimens were largely obtained from ichthyoplankton samples collected in 1984-86 by the CSIRO Division of Fisheries, Hobart, Tasmania, as part of a study aimed at documenting the distribution and abundance of lar- val fishes in Tasmanian coastal and neritic waters. De- tails of sampling locations and protocol are provided by Thresher et al. (1989). Larvae were obtained from ob- lique tows to a depth of 200 m (bottom depth permit- ting) at a series of stations covering shelf and slope waters, using aim diameter ring net (500u mesh). Additional material was obtained from samples collected with identical gear in New South Wales shelf and slope waters by A. Miskiewicz (Water Board, Environ. Proj. Group, P*0 Box A53, Sydney South, 2000). Larval samples were fixed in either 10% formalin buffered with sodium tetraborate, or 959c ethanol. Mor- phometric analysis and illustrations were based on for- malin-fixed specimens of Paratrachichthys and Optivus. However, only ethanol-fixed material was available for Aulotrachichthys. No allowance was made for shrink- age or distortion in preservative. Larvae were examined using a Wild M5 dissecting microscope, and all drawings were made with the aid of a camera lucida. Larvae were identified using exist- ing literature (Keene & Tighe 1984, Okiyama 1988) by comparison with juvenile and adult features of identi- fied species, and by the establishment of developmen- tal series. Comparisons with similar larvae (e.g., zeids) were made with material from the CSIRO samples. All unspecified body lengths refer to notochord length in preflexion and flexion larvae, and to standard length in postflexion larvae and juveniles. We define snout to anal-fin length as the horizontal distance from the tip of the snout to the anterior origin of the anal fin or anal-fin anlagen. Body depth at anus is the vertical distance between body margins through the center of the anal opening. Body depth at pectoral is equivalent to 'body depth' of Leis & Rennis (1983). Other definitions, such as body shape, follow Leis and Trnski ( 1989 1. Nomenclature of head spination follows that of Moser & Ahlstrom (1978). Larval measurements were made using an ocular micrometer. Juveniles were measured with vernier calipers. Results During 18 months of sampling, 119 Paratrachichthys, 147 Optivus, and 25 Aulotrachichthys larvae were col- lected. The distribution of larvae is detailed in Figure 1. No larvae that could be attributed to Hoplostethus were collected. A representative series of each species was deposited in the ISR Munro Fish Collection (CSIRO, Hobart, Tasmania). Reference numbers: Optivus, CSIRO L179-184; Paratrachichthys, CSIRO L185-190; Aulotrachichthys, CSIRO L191-196. Identification In larger specimens of two of the series, the anus is located between the pelvic fins. Only three trachich- thyid genera have this character: Paratrachichthys, Aulotrachichthys, and Sorosichthys (May & Maxwell 1986). Sorosichthys is separated easily on the basis of a pelvic count of 1,5, compared with the 1,6 of the other two genera (Table 1). The only character reported in the literature to distinguish adults of Aulo- trachichthys from Paratrachichthys is the presence in 78 Fishery Bulletin 91(1), 1993 Figure 1 Distribution of trachichthyid larvae sampled in southeast Aus- tralian waters. (•) Paratrachichthys sp., (*) Optivus sp., ( ) Aulotrachichthys sp. the former of striated silvery tissue on the bases of the pectoral fin, on the isthmus beneath the gill cover, and in a narrow strip along the ventral edge of the body (May & Maxwell 1986). However, examination of juve- nile and adult specimens also reveals a difference in anal fin-ray counts: Paratrachichthys with a count of Table 1 Meristic characters of trachichthyid genera present in southe rn Australian waters. D A PI P2 Vertebrae Paratrachichthys V.13 111,10 12-14 1,6 27-29 Aulotrachichthys V.13 111,8 12-14 1.6 27-29 Optivus IV, 11 111,9 10-12 1,6 27-29 Hoplostethus V-YI 11,12-18 111,9-11 12-20 1,6 25-30 Gephyroberyx VIII, 13-14 111,11 14 1,6 26-27 Sorosichthys IX-X.8-9 11,8 13 1,5 Unknown Trachichthys IV, 10-14 111,9-11 11-14 1,6 27 111,10 and Aulotrachichthys with 111,8. Both series had a pelvic count of 1,6. An anal count of 111,8 and stri- ated pectoral tissue occurred in the largest specimen of only one series; on that basis, we assign the series with the largest specimen to Aulotrachichthys and the other to Paratrachichthys. We were unable to deter- mine whether or not the Aulotrachichthys and Optivus series represented more than one species. Optivus lar- vae were distinguished on the basis of a dorsal count of IV,11, an anal count of 111,9, and the position of the anus, which remains static, immediately anterior to the anal fin. Larval development Paratrachichthys sp. (Fig. 2) Morphology Head length is about equal to body depth at pectoral until flexion, after which body depth in- creases to approximately 50% of body length (Table 2). The mouth is large, reaching to approximately the cen- ter of the eye in our smallest specimen (3.2mm) and beyond the eye in larvae greater than 4.5 mm (Fig. 2A-C). The body depth at anus increases mark- edly during flexion, associated with the anterior mi- gration of the anus during this period. The gas blad- der is inflated and prominent in all specimens. There are 27-29 myomeres. Initially the gut is straight and tube like. It quickly thickens, coils, and becomes triangular by approxi- mately 5.0 mm. The anus begins to migrate anteriorly by 6.5mm and is in the adult location (between the pelvics) by 7.8mm (Fig. 2E). The light organ surround- ing the anus first appears in 5.4 mm larvae as an unpigmented, thickened ring. By 6.1mm the light or- gan is lightly pigmented; by 6.9 mm the organ is heavily pigmented and rugose. Notochord flexion commences at about 5.9 mm and is complete by 7.6 mm. Fin development Development of the pelvics is pre- cocious. Slight swellings on either side of the gut are present in our smallest specimen (3.2 mm). Distinct buds are present by 3.9 mm. The pelvics develop rapidly, hav- ing a full complement of 7 elements by 5.6 mm, and reaching up to 347c body length by 7.6 mm. Anlagen of both dorsal and anal fins are present by 4.3 mm. The anlagen first appear as hyaline zones within the fin folds, connected to the body by a series of filamentous extensions in- serted at each myoseptum (Fig. 2B). Bases of the anal and dorsal fins are present by 4.7 mm, and posterior incipi- Jordan and Bruce: Larval development of three roughy species 79 Table 2 Bodv proportions of larvae and juveniles of Paratrachichthys sp. (expressed as mean proportions of body length with standard deviations in parentheses). Specimens between dashed lines were undergoing notochord flexion. Characters lacking standard deviation are baset on one individual only. n = number of in dividuals. Size range Snout to Preanal Body depth Body depth Head Eye Pelvic immi n anal fin length i at pectoral ) iat anus) length diameter fin length 3.32 1 0.62 0.19 0.10 0.22 0.13 _ 3.90 1 - 0.68 0.22 0.09 0.28 0.12 0.06 4.01-4.50 5 _ 0.59(0.03) 0.26(0.02) 0.12)0.01) 0.29(0.01) 0.12(0.01) 0.11(0.04) 4.51-5.00 7 0.63(0.02) 0.62(0.02) 0.28(0.03) 0.14)0.011 0.30(0.01) 0.14(0.01) 0.18(0.02) 5.01-5.50 8 0.63(0.02) 0.62(0.02) 0.34(0.02) 0.19(0.04) 0.31(0.02) 0.14(0.01) 0.25(0.03) 5.51-6.00 3 0.61(0.02) 0.57(0.01) 0.35(0.02) 0.26(0.04) 0.33(0.02) 0.15I0.0U 0.27(0.02) 6.01-6.50 2 0.60(0.05) 0.54(0.07) 0.38(0.03) 0.29(0.05) 0.33(0.01) 0.15(0.01) 0.29(0.01) 6.95 1 0.70 0.55 0.43 0.39 0.35 0.14 0.30 7.60 1 0.74 0.43 0.53 0.43 0.41 0.17 0.34 10.00 1 0.71 0.48 0.42 0.42 0.34 0.17 0.26 35.80 1 0.58 0.38 0.38 0.38 0.34 0.14 0.18 39.70 1 0.60 0.37 0.39 0.39 0.36 0.12 0.18 ent rays first appear above these bases by 5.4 mm. Incipient rays appear in the pectoral fin shortly there- after (5.5 mm). Ossification of dorsal, anal, pelvic, and pectoral fins occurs during flexion, with full comple- ments in all fins present by 7.8 mm. Spination Paratrachichthys larvae have only weakly developed head spination. A low supraocular ridge is present at 3.3 mm, developing 1-2 spines by 3.9 mm (Fig. 2A). The number of supraocular spines increases to 2-3 by 4.0 mm, reaching a maximum of 5-6 just prior to flexion. During flexion, the supraocular spines disappear. The single opercular spine is present by 6.9 mm and is retained in the adult. Similarly, single preopercular and posttemporal spines are present by 8.7mm and are retained. Cranial ridges are present by 5.4 mm: however, even by 10.0 mm these have not yet become denticulate as they are in juveniles and adults. Pigmentation Paratrachichthys larvae are moderately to heavily pigmented (with the exception of the last 2-8 myomeres, including the notochord tip) through- out the entire larval period. Pigment tends to be con- centrated over the dorsal and ventral surfaces of the body and the dorsal surface of the gut. Otherwise, there are few useful distinguishing features based on pat- terns of pigmentation. The pelvic fins are heavily pigmented by 4.2 mm and remain so in our largest postflexion larva ( 10.0 mm). Some variation was introduced by the obvi- ously faded pigment of certain specimens, the result being a series of heavily-pigmented and a series of moderately- to lightly-pigmented individuals. Because the major pigment concentrations, morphology, and meristic information were otherwise identical for the two series, it is unlikely that variations in the inten- sity of pigmentation indicate the presence of more than one species. Scalation Juvenile and adult Paratrachichthys have small, adherent, ctenoid scales covering the body and a series of strong ventral scutes between the anus and the anal fin (Woods & Sonoda 1973). Our largest larva (10.0 mm ) has no sign of scalation and lacks the minute dermal spines of other trachichthyid genera, although a weak fleshy ridge develops along the ventral midline between the anus and the anal fin by 7.6 mm, prob- ably a precursor to the characteristic ventral scutes of juveniles and adults. The 39.7 mm juvenile examined was, in effect, a minature adult having completed scalation, including the ventral scutes. Aulotrachichthys sp. (Fig. 3) Morphology Head length is about equal to body depth at the pectoral fin until flexion, after which body depth increases to 509c body length (Table 3). The mouth is moderately large, reaching to the posterior margin of the eye in our smallest specimen (2.8 mm), falling to just short of the margin in our largest specimen (7.9 mm). The body depth at the anus increases mark- edly prior to flexion as the anus migrates anteriorly. The gas bladder is inflated and prominent in all speci- 80 Fishery Bulletin 91(1), 1993 mens. There are 27-29 myo- meres. The gut is a convoluted tube in our smallest specimen (2.8mm). It quickly thickens, coils, and becomes triangular by 4.4 mm. The anus begins to mi- grate by 3.9 mm and is in the adult location (between the pel- vic fins) by 4.9mm (Fig. 3C,D). A light organ that surrounds the anus first appears in 3.6 mm lar- vae and is well developed, rugose in appearance, and heavily pig- mented by 4.4 mm. Ventral stri- ated tissue, characteristic of adults, is present in the 7.9 mm specimen. Insufficient specimens were available to fully document flexion; however, flexion was just about to commence in a 4.9 mm specimen, was well underway in a 5.7 mm specimen, and had been completed in a 7.9 mm specimen. Fin development Development of the pelvic fins is precocious. Even our smallest specimen has a full pelvic complement of 7 ele- ments (although ossification is not completed until 4.4 mm). The pelvic fins are large, up to 35% body length at 4.9 mm, and reach beyond the anal-fin origin in all specimens. The limited number of specimens precluded docu- menting initial dorsal- and anal- fin anlagen development; how- ever, the separation from the body of the posteriormost anal base in our 4.1 mm specimen sug- gests a finfold development simi- lar to Paratrachichthys. Both dor- sal and anal bases are present by 3.94.1mm, and incipient rays appear above these bases by 4.4 mm. Incipient rays appear in the pectoral fin shortly thereaf- ter (4.9 mm). Ossification of the dorsal, anal, and pectoral fins commences in flexion-stage lar- vae, with a full complement in all fins present by 7.9 mm. Figure 2 Development stages of Paratrachichthys sp.: (A) 3.9mm, (B) 4.3mm, (C) 4.7mm, (D) 5.5 mm, (E) 7.8 mm. Arrows indicate location of anus. Jordan and Bruce: Larval development of three roughy species Table 3 Body proportions of larvae of Aulotrachwhthys sp. (expressed as mean proportions of body length with standard deviations in parentheses 1. Specimens between dashed lines were undergoing notochord flexion. Characters lack- ing standard deviation are based on one individual only. rc=number of individuals. Size range (mm) n Snout to anal fin Preanal length Body depth (at pectoral) Body depth (at anus) Head length Eye diameter Pelvic fin length 2.51-3.00 3.01-3.50 3.51-1.00 4.01^.50 4.91 5 3 3 2 1 0.62(0.01) 0.59 0.59(0.05) 0.63(0.01) 0.60(0.03) 0.54(0.01) 0.38 0.25(0.03) 0.25(0.01) 0.34(0.03) 0.34(0.02) 0.39 0.14(0.02) 0.14(0.01) 0.18(0.04) 0.22(0.02) 0.37 0.24(0.02) 0.27(0.02) 0.31(0.02) 0.31(0.01) 0.34 0.11(0.01) 0.13(0.01) 0.13(0.01) 0.13(0.01) 0.13 0.23(0.01) 0.27(0.03) 0.33(0.03) 0.31(0.00) 0.35 5.66 1 0.64 0.30 0.35 0.51 0.30 0.16 0.28 7.85 1 0.72 0.44 0.50 0.50 0.43 0.17 0.31 Spination Aulotrachichthys larvae have well-developed head spination. Our smallest specimen has a low supraocular ridge with a single spine. The number of supraocular spines increases to 3-4 by 4.4 mm (Fig. 3C); they become quite robust and reach a maxi- mum number of 8-9 during flexion. Generally, the pos- terior group of these spines are the largest and are recurved. Preopercular spines are present by 4.4 mm, with an anterior preopercular series added by 4.9 mm. By 5.7 mm, preopercular spination is quite robust, with secondary ridging and branching of the largest spines (particularly at the angle) (Fig 3D). Nuchal, supra- cleithral, and posttemporal spines, as well as nasal and cranial ridges, are developed prior to 5.7 mm. During flexion a hard bony plate forms in the region of the posttemporal, extending posteriorly to the level of the opercular margin. This plate extends beyond the opercular margin by 7.9 mm, and is retained in the adult. Available specimens are insufficient to deter- mine if this plate results from the fusion of the supracleithral and posttemporal series. During flexion, spines also develop on the dentary and infraorbital. Several cranial and opercular ridges appear at this stage. A single spine is present immediately posterior to the anus by 4.9 mm. Dermal spines are present on the pelvic bases by 5.7 mm, and by 7.9 mm a cluster of spines is also present immediately anterior to the anus. The 7.9 mm specimen has well developed dermal spination in longitudinal rows over the entire body surface and on the dorsal- and anal-fin bases, although there is no sign of scalation. Additionally, this speci- men has a row of strong spines extending posteriorly along the ventral midline from the anus towards the anal fin, probably precursors to the ventral scutes of adults. Fine villiform teeth are present in both jaws. Pigmentation Aulotrachichthys larvae are heavily pig- mented (with the exception of the posteriormost 5-6 myomeres, including the notochord tip) throughout the larval period. The pelvic fins are heavily pigmented in the smallest specimen and remain so in the 7.9 mm specimen. Aulotrachichthys larvae are more heavily pigmented than Paratrachichthys, although, as with Paratrachichthys, there are few useful distinguishing characters based on pigment pattern. Optivus sp. (Fig. 4) Morphology Body depth increases to a maximum of 49% body length during flexion (Table 4). Body depth at anus increases only slightly compared with Paratrachichthys and Aulotrachichthys, because the anus position in Optivus remains static. Head length increases from 36% body length in preflexion larvae to 44% in juveniles. Eye diameter remains relatively con- stant. The mouth is moderate to large, reaching to the center of the eye in our smallest specimen (2.5mm) and beyond the eye in larvae greater than 4.0 mm. The gut, which is initially straight, quickly thickens, coils, and becomes broadly triangular by 3.5 mm. Optivus larvae do not develop a light organ. Notochord flexion commences at about 4.0 mm and is complete by I 7.1mm. There are 27-29 myomeres. Fin development Pelvic fins first appear in larvae of 3.0 mm as slight swellings on either side of the gut. 82 Fishery Bulletin 91(1), 1993 Figure 3 Development stages of Aulotrachichthys sp.: (A) 2.9mm, (B) 3.4 mm, (C) 4.4 mm, (D) 4.7 mm, (E) 5.7 mm (note: pectoral fin missingl, (F) 7.9 mm. Arrows indicate location of anus. and these develop rapidly. Distinct buds are present by 3.2 mm, and the developing fin reaches up to 25% body length by 6.2mm (Fig. 4A-C). Ossification com- mences by 5.1mm, and a full complement of seven elements is present by 8.0 mm. Anlagen of both dorsal and anal fins are present by 2.7 mm and appear as hyaline zones located within the median finfolds, as in Paratrachichthys. Bases are first visible in both fins by 3.5 mm, and incipient rays are present by 4.0 mm. Incipient rays appear in the pectoral fin by 4.5 mm. Ossification of the dorsal, anal, and pectoral fins com- mences in early-flexion-stage larvae, with a full comple- ment in all fins present by 7.1 mm. Spination Head spination is only weakly developed in preflexion Optivus larvae. A low supraocular ridge is present by 2.7 mm, with 4-5 spines developing by 3.4mm (Fig. 4A). By 4.5 mm, these supraocular spines have disappeared. Cranial ridges are present by 4.7mm, and a series of spines develops on the preopercular margins by 5.1mm. The preopercular, opercular, and posttemporal spines characteristic of adults are present by 23.0 mm ( Fig. 4E ). Scalation Small dermal spines appear on the body by 4.7 mm and develop in longitudinal rows over the en- tire body and dorsal- and anal-fin bases by 5.1 mm. By 8.0 mm the base of each single dermal spine has trans- Jordan and Bruce: Larval development of three roughy species 83 Table A Body proportions of larvae and juveniles of Optivus sp. (expressed as mean proportions of body length with standard deviations in parentheses). Specimens between dashed lines were undergoing notochord flexion. Charac- ters lacking standard deviation are based on one individual only. n= number of individuals. Size range i mm I n Snout to anal fin Preanal length Body depth (at pectoral) Body depth (at anus) Head length Eye diameter Pelvic fin length 2.51-3.00 3 0.66(0.03) 0.68(0.04) 0.34(0.011 0.18(0.01) 0.36(0.011 0.13(0.01) _ 3.01-3.50 8 0.66(0.03) 0.66(0.03) 0.38(0.031 0.20(0.02) 0.35(0.02) 0.14(0.01) 0.05(0.01) 3.51^.00 9 0.69(0.08) 0.65(0.05) 0.40(0.051 0.20(0.02) 0.36(0.03) 0.14(0.02) 0.08(0.02) 4.01-4.50 11 0.67(0.04) 0.65(0.04) 0.44(0.05) 0.24(0.04) 0.40(0.031 0.15(0.02) 0.12(0.02) 4.51-5.00 5 0.69(0.04) 0.67(0.04) 0.45(0.04) 0.26(0.05) 0.40(0.041 0.15(0.01) 0.14(0.01) 5.01-5.50 3 0.74(0.01) 0.70(0.04) 0.45(0.041 0.30(0.03) 0.45(0.041 0.14(0.04) 0.20(0.02) 6.01-6.50 2 0.72(0.03) 0.69(0.06) 0.49(0.011 0.29(0.02) 0.46(0.011 0.16(0.02) 0.25(0.02) 7.15 1 0.71 0.70 0.48 0.30 0.43 0.15 0.20 8.00 1 0.72 0.69 0.47 0.31 0.45 0.15 0.20 10.01-10.50 2 0.70(0.07) 0.69(0.081 0.44(0.02) 0.30(0.01) 0.44(0.01) 0.15(0.01) 0.21(0.03) 15.00 1 0.62 0.60 0.39 0.28 0.36 0.13 0.23 17.40 1 0.63 0.61 0.37 0.28 0.36 0.13 0.19 17.80 1 0.62 0.60 0.38 0.27 0.36 0.13 0.20 18.40 1 0.65 0.64 0.38 0.27 0.35 0.14 0.21 19.60 1 0.63 0.62 0.37 0.27 0.34 0.13 0.18 23.00 1 0.65 0.63 0.37 0.27 0.35 0.12 0.19 formed into a small ctenoid scale (Fig. 5A). Not all scales develop spines at the same stage; by 10.2 mm, scales have 1-3 spines (Fig. 5B). Three spines appear to be present on all scales by 23.0 mm. A row of larger spines appear on the ventral surface between the anus and the pelvic fins by 7.2 mm and form the character- istic ventral scutes of juveniles and adults by 15.0 mm. Pigmentation Pigmentation in preflexion Optivus lar- vae is moderate and concentrated on the dorsal surface of the gut, as well as the dorsal and ventral surfaces of the trunk. Pigment is absent from the posteriormost 5-6 myomeres (including the notochord tip) as in Paratrachichthys and Aulotrachichthys. During flexion, the entire body and head become moderately pigmented and the dorsal surface of the gut becomes heavily pig- mented. The entire body and head is evenly pigmented in the largest specimen (23.0mm, Fig 4E). The pelvic fins are moderately pigmented by 4.7 mm; the pigment contracts towards each base during flexion and disap- pears by 23.0 mm. Discussion Considerable confusion exists in the systematics of beryciform fishes at the species level. Current classifi- cations are based almost entirely on adult characters. Keene & Tighe (1984) noted the usefulness of includ- ing early-life-history characters in these studies, but the lack of such data at that time for ten of the beryciform families precluded an adequate appraisal. The three genera featured here together share charac- ters common to other described trachichthyid larvae, including moderate to heavy pigment, a mod- erate to large mouth, a stocky body form, precociously- developing and heavily-pigmented pelvic fins, cranial ridges, opercular spination, and a myomere count of 26-30. Pelvic-fin pigmentation is most pronounced in Aulotrachichthys and Paratrachichthys and is least de- veloped in Optivus. Dermal spination is well devel- oped in Aulotrachichthys and Optivus, although ab- sent in Paratrachichthys. Cranial ridges and opercular spines are present in all of our series; however, Aulotrachichthys develops by far the most pronounced head spination of the three and perhaps of any re- ported trachichthyid larva. Small trachichthyid larvae can be confused with zeids and some gadoid larvae that also have precocious, heavily-pigmented pelvic fins. However, zeid larvae are more evenly pigmented, have pigment extending into the finfolds in small larvae, generally have a higher myomere count (29-42, Tighe & Keene 1984), and have a more tightly coiled gut with consequently a longer postanal length. The sequence in which fin-ray ele- 84 Fishery Bulletin 91(1). 1993 Figure 4 Development stages of Optivus sp.: (A) 3.4mm, (B) 4.7mm. (C) 5.1 mm. (D) 7.2mm, (E) 23.0 mm unite: scale spination not figured). Arrows indicate location of anus. ments form also distinguishes zeid larvae. In zeid larvae exam- ined during this study, the anteriormost bases and rays were the first dorsal-fin elements to form. In trachichthyids, how- ever, the middle or posterior ele- ments are the first to form. When present, supraocular spination was also a useful feature to dis- tinguish trachichthyids from zeids (zeids examined did not de- velop supraocular spines until af- ter flexion). Although this may be useful locally, some zeid spe- cies (e.g., Zeus faber) have supra- ocular spines at sizes similar to trachichthyid larvae (Sanzo 1931). Larger zeid larvae are eas- ily distinguished from trachich- thyid larvae by their longer dorsal- and anal-fin bases, often with elongate anterior rays, a larger mouth, and a rhom- boid, laterally-compressed body shape. Small gadoid larvae with precocious, heavily-pigmented pelvics (e.g., Gaidropsarus) dif- fer from trachichthyids in hav- ing a higher myomere count (>40), pelvics set higher on the body, and a more slender post- anal body form. Larger gadoid larvae are easily distinguished by morphology, fin meristics, and pigment (see Dunn & Matarese 1984, for details). Hoplostethus species, and in particular orange roughy H. atlanticus, are by far the most abundant trachichthyids in Tas- manian waters. Despite exten- sive sampling throughout the year covering pelagic environ- ments from nearshore to mesope- lagic and epipelagic zones (see Thresher et al. 1989, for details), no larvae of the genus Hoplo- stethus have been identified. New Zealand researchers also have been unable to locate H. atlan- ticus larvae (Pankhurst & Con- roy 1987), even though spawn- Jordan and Bruce: Larval development of three roughy species 85 Figure 5 Development of dermal spination on scale of Opticus sp.: (A) 8.0mm, (B) 10.2mm. ing aggregations have been located (Beardsell 1984). The 26mm Hoplostethus atlanticus specimen, caught in a demersal trawl at 400-950 m off St. Patricks Head, east- ern Tasmania (CSIRO H1141), repre- sents the smallest H. atlanticus reported to date (Fig. 6). Kotlyar ( 1984 1 previ- ously recorded a 36 mm H. atlanticus taken by bottom trawl in the Atlantic Ocean at a depth of 965-990 m. The 26 mm juvenile has characters that are common to the other trachichthyids identified, such as a deep body, large mouth, cranial ridges and opercular spination, heavily-pigmented gut and pelvic fins, and distinct anal- and dorsal-ray bases. It is highly likely that such characters are retained in the larvae of H. atlanticus, and Hoplostethus larvae in general. Several scenarios may explain why Hoplostethus atlanticus lar- vae have not yet been located. 1 Bimonthly sampling frequency is too coarse to capture larvae during their pelagic stage. Although this cannot be discounted due to the lack of information on larval duration, the likelihood of com- pletely missing all larvae seems low. 2 Larvae occur further off the shelf or slope than sampled (>18km from the shelf break). 3 Larvae may occur in greater depths than those sampled in 'standard' ichthyoplankton surveys. This may be the most reason- able scenario, and has previously been suggested by Kotlyar (1984) for Hoplostethus less than 15-19 mm (based on the capture of three H. melanopterus juveniles 15.0-18.2 mm in Isaacs Kid trawls at 1000-1500 m in the Sulu Sea). Larvae may also occur close to the bottom on the continental slope, as supported by the capture of our 26 mm specimen and that by Kotlyar ( 1984). Hoplostethus atlanticus adults occur in depths of 500-1200 m (May & Maxwell 1986 ). Spawning has been confirmed from both the east coast of Tasmania (Lyle et al. 1989) and New South Wales (Williams 1989). Hoplostethus atlanticus eggs (2. 12-2.45 mm in diameter with a conspicuous orange oil droplet) have been collected in plank- ton tows above 400m in Tasmanian (Lyle et al. 1989) and New Zealand waters (Beardsell 1984), and are presumably bouyant. The lack of larvae in surface waters, however, suggests that egg density, the presence of a thermocline, or a combination of both factors may confine eggs and larvae to deep water. Although deep-water plankton sampling is logistically more difficult than shallow- water sampling, such sampling should be carried out if we are to fully understand the early life histories of certain deep-water species. Figure 6 Hoplostethus atlanticus juvenile, 26 mm. Acknowledgments We are grateful to R. Thresher, J. Gunn, J. Leis, J. Paxton, P. Last, and two anonymous re- viewers who made many useful suggestions for improving the manuscript. We thank A. Miskiewicz and A. Steffe for supplying additional specimens. This work was supported in part by FIRDC research grant 87/129. 86 Fishery Bulletin 91(1), 1993 Citations Beardsell, M. 1984 Thick orange roughy spawning schools re- corded. Catch, Sept. 1984, p. 24. Crossland, J. 1981 Fish eggs and larvae of the Hauraki Gulf, New Zealand. N.Z. Minist. Agric. Fish Res. Bull. 23, 61 P- Dunn, J.R., & A.C. Matarese 1984 Gadidae: Development and relationships. In Moser, H.G., et al. (eds.). Ontogeny and systematics of fishes, p. 283-299. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol. Allen Press, Lawrence, KS. Johnson, R.K. 1970 A second record of Korsogaster nanus Parr (Beryciformes: Korsogasteridae). Copeia 1970:758- 760. Keene, M.J., & K.A. Tighe 1984 Beryciformes: Development and relation- ships. In Moser, H.G., et al. (eds.), Ontogeny and systematics of fishes, p. 383-392. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol. Allen Press, Lawrence, KS. Kotlyar, A.N. 1984 A description of the fry of four species of Hoplostethus (Trachichthyidae, Beryciformes). Byull. Mosk. Ova. Ispyt. Prir. Otd. Biol. 89(3):33-39 [in Russ.]. Leis, J.M., & D.S. Rennis 1983 The larvae of Indo-Pacific coral reef fishes. New South Wales Univ. Press, Sydney, 269 p. Leis, J.M., & T. Trnski 1989 The larvae of Indo-Pacific shorefishes. New South Wales Univ. Press, Sydney, 371 p. Lyle, J., J. Kitchener, & S. Riley 1989 Orange roughy bonanza off Tasmania. Aust. Fish. 48(12 ):20-24. May, J.L., & J.G.H. Maxwell 1986 Field guide to trawl fish from temperate waters of Australia. CSIRO, Hobart, Tasmania, p. 216-222. Moser, H.G., & E.H. Ahlstrom 1978 Larvae and pelagic juveniles of blackgill rock- fish, Sebastes melanostomus, taken in midwater trawls off Southern California and Baja California. J. Fish. Res. Board Can. 35:981-996. Okiyama, M. 1988 An atlas of early stage fishes in Japan. Tokai Univ. Press, Tokyo, 1150 p. [in Jpn.]. Pankhurst, N.W., & A.M. Conroy 1987 Size-fecundity relationships in the orange roughy, Hoplostethus atlanticus. N.Z. J. Mar. Freshwater Res. 21:295-300. Parr, A.E. 1933 Deep-sea Berycomorphi and Percomorphi from the waters around the Bahama and the Bermuda Islands. Bull. Bingham Oceanogr. Collect. Yale Univ. 3(6). Robertson, D.A. 1975 A key to the planktonic eggs of some New Zealand marine teleosts. Fish. Res. Div. Occas. Publ. (N.Z.) 9. Sanzo, L. 1931 Uova e larve di Zeus faber L. Arch. Zool. Ital. 15:475. Thresher, R.E., B.D. Bruce, D.M. Furlani, & J.S. Gunn 1989 Distribution, growth and advection of larvae of the southern temperate gadoid, Macruronus no- vaezelandiae (Teleostei: Merlucciidae), in Australian coastal waters. Fish. Bull, U.S. 87:29^8. Tighe, K.A., & M.J. Keene 1984 Zeiformes: Development and relationships. In Moser, H.G., et al. (eds.), Ontogeny and systematics of fishes, p. 393-398. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol. Allen Press, Lawrence, KS. Williams, M. 1989 Orange roughy research in Australia: a case study for research co-ordination. Search 20:130-134. Woods, L., & P. Sonoda 1973 Order Berycomorphi (Beryciformes). In Fishes of the western North Atlantic. Mem. Sears. Found. Mar. Res. 1 (6):263-396. Abstract.— A manned submers- ible was used in the eastern Gulf of Alaska to observe spatial distribu- tions of Pacific ocean perch Sebastes alutus and other Sebastes spp.. and count rockfish for comparison with bottom-trawl catch rates. Twenty submersible dives were completed in 1988 and 1989 at depths of 188- 290 m. Approximately 80<7f of the 5317 rockfish observed from the sub- mersible were Pacific ocean perch. Most adult Pacific ocean perch were in groups of 2-200 over flat, pebble substrate. Fish within a group were 1— 4m apart, usually oriented into the current, and 0-7 m above bot- tom. Most juvenile Pacific ocean perch, and juveniles and adults of other Sebastes spp., were associated with rugged habitat (cobble, boul- ders, pinnacles, and coral). Densi- ties of Pacific ocean perch estimated from bottom-trawl catches were ap- proximately twice those observed from the submersible, indicating that the bridles and otter doors herded fish into the trawl. Bottom- trawl surveys may overestimate Pacific ocean perch abundance be- cause of this possible herding effect and the preference of adult Pacific ocean perch for smooth (trawlable) substrate. Distribution and abundance of rockfish determined from a submersible and by bottom trawling Kenneth J. Krieger Auke Bay Laboratory, Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 1 1305 Glacier Highway, Juneau, Alaska 99801-8626 Manuscript accepted 29 September 1992. Fishery Bulletin, U.S. 91:87-96 ( 1993). Pacific ocean perch Sebastes alutus is a commercially important rockfish found along the North American coast from southern California to the Bering Sea, and along the Asiatic coast from Cape Navarin to the Kuril Islands (Balsiger et al. 1985). Prima- rily an offshore species that inhabits the outer continental shelf and up- per slope regions, it is caught with bottom trawls at depths of 165-290 m (Hart 1973). In the Gulf of Alaska, Pacific ocean perch were heavily exploited in the 1960s by the Soviet and Japanese trawl fleets. Foreign catches of rock- fish (consisting mainly of Pacific ocean- perch) peaked in the Gulf of Alaska in 1965 at 350,000 metric tons (t). By the late 1970s, rockfish catches had declined to less than 10,000 1, and catch-per-unit-effort (CPUE) had decreased by 80 c i (Balsiger et al. 1985). Domestic trawling replaced foreign trawling in the Gulf of Alaska in the 1980s. Although rockfish stocks remain depressed from overfishing, domestic rockfish catches (mainly Pa- cific ocean perch) have increased in the Gulf of Alaska from 1000 1 in 1985 to 20,000 1 in 1990 (Heifetz & Clausen 1990). Reliable stock assessments are needed for managing depressed stocks of Pacific ocean perch. Cur- rent stock assessments, based prima- rily on catch rates from bottom-trawl surveys, are questionable because (1) the catch efficiency of bottom trawls on Pacific ocean perch is unknown, (2) Pacific ocean perch are not sampled in areas where the habitat is too rugged for bottom trawling, and (3) information is limited on Pacific ocean perch spatial distribution, mi- gration, and off-bottom movement. Leaman & Nagtegaal (1986) noted their dissatisfaction with rockfish bot- tom-trawl surveys because of wide confidence intervals associated with the biomass estimates, and because alternate assessment techniques in- dicate that the estimates themselves may be grossly in error. Balsiger et al. ( 1985) reported that bottom-trawl surveys probably underestimate populations because Pacific ocean perch occupy the water column above the opening of the trawl. An understanding of both the be- havior and habitat of Pacific ocean perch and the catch efficiency of bot- tom trawls is needed to improve bio- mass estimates. In this study, a two- man submersible was used for in situ observations of rockfish and a bot- tom trawl to sample rockfish at sub- mersible dive sites. The main objec- tives were to describe the spatial distribution of rockfish, visually de- termine their abundance in trawlable and untrawlable habitat, and deter- mine the efficiency of bottom trawls for capturing rockfish. Pacific ocean perch was the target species of this study. 87 Fishery Bulletin 91 f I). 1993 58N 56N Materials and methods Study area This study was conducted in Au- gust 1988 and 1989 on the outer continental shelf in the eastern Gulf of Alaska (Fig. 1). Study sites extended over a 300 km range to include a wide range of habitats and population densities of Pacific ocean perch. Dive sites were selected from coordinates of previous trawl and sonar stud- ies: (1) Sites north of lat. 56° N from studies by the Auke Bay Laboratory, Alaska Fisheries Sci- ence Center, National Marine Fisheries Service, in July 1987 and 1988 using bottom trawls, midwater trawls, and sonar to locate high and low densities of rockfish at high-relief (un- trawlable) and low-relief (trawl- able) sites; and (2) sites south of lat. 56° N from studies by Leaman & Nagtegaal ( 1986) who used sonar to locate high densi- ties of rockfish at untrawlable sites. Submersible The submersible Delta was char- tered for all dives. This battery- powered two-man submersible is 4.7 m long, dives to 365 m, and travels 2-6 km/h for 2-4 h. It is equipped with ten 150 W exter- nal halogen lights, internal and external video cameras, a 35 mm external camera, magnetic compass, directional gyro compass, underwater telephone, and transponder that allowed tracking of the submersible from the surface vessel Wm. A. McGaw. Submersible transects were charted from the Wm. A. McGaw. The surface vessel tracked the sub- mersible and recorded LORAN fixes at the beginning and end of a transect, and every 5-15 min during a transect. In 1988, a transect consisted of a single com- pass heading followed for 60 min. The transect pattern was changed in 1989 to facilitate trawl comparisons, and consisted of four parallel compass headings fol- lowed for 15 min each. The four parallel transects were each separated by 5 min of travel. CAPE SPENCER DIXON ENTRANCE 140W Locations of submers 1989 (•). 136W 132W Figure 1 ble dive sites in the eastern Gulf of Alaska, 1988 ▲ and A pilot and one of the two observers were aboard the submersible on each dive. The pilot maintained the submersible within 0.5 m of bottom at 3-4 km/h while the observer made observations through a star- board porthole; external cameras were mounted on the starboard side, and the side portholes provided the widest range of view. Observations included rockfish species identification, number, size, grouping behavior, orientation, position relative to the sea bottom, habi- tat affiliations, and reactions to the submersible. Addi- tional observations included identification and enu- meration of other fish species, and estimates of current direction and velocity based on the bending of sea pens and drift of silt. The pilot sat above the observer in a Kneger Distribution and abundance of Sebastes spp 89 tower with a panoramic view and assisted in counting fish above the observer's view and in monitoring fish behavior completely around the submersible. All observations were audio- and video-recorded for sub- sequent analysis and verification. All dives were dur- ing daylight between 0600 and 1900 h. Rockfish densities were derived from the number of fish counted and the seafloor area searched. The sea- floor area searched was the distance traveled (1.7- 2.2 km/dive) times estimates of the lateral distance from the submersible at which rockfish were visible. Lat- eral distance estimates varied between 5 and 6 m be- cause of changes in water clarity; illumination was provided entirely by the submersible lights and re- mained constant. Estimated distances were compared with true distances using three methods: (DA length of pipe marked at meter intervals was laid on the seafloor on two dives, (2) a hand-held sonar gun pro- vided distance readouts to rock formations on six dives, and (3) the submersible's sonar provided distance read- outs to the seafloor during descent and ascent. Esti- mates of distance were consistently within lm of true distances. About 10 m was monitored above the sub- mersible: 5-6 m by the observer and an additional 4-5 m by the pilot. Dive sites were classified as trawlable, marginally trawlable, or untrawlable based on bottom type and extent of relief. Trawlable sites contained pebble sub- strate interspersed with cobble <0.5 m in diameter on flat bottom; marginally trawlable sites contained pebble substrate interspersed with cobble and boulders of 0.5-5.0 m in diameter on low-relief bottom; untrawlable sites contained mainly bedrock substrate with a vari- ety of rugged habitats including boulders, coral, ledges, rocky outcroppings, and pinnacles on high-relief bottom. Bottom trawling Bottom trawling was used in 1989 to identify fish spe- cies at submersible dive sites and to derive population densities from catch rates. Trawling was conducted from the NOAA ship RV Townsend Cromwell, using a 400-mesh Eastern otter trawl equipped with 1.5x2.1 m doors, each weighing 386 kg. Trawling was during day- light, usually within 4h after completion of a submers- ible dive. The time between dives and trawls depended on ship's operations, including how long it took to pro- cess the catch. The sampling strategy was to trawl at 6.0km/h, intersecting the four parallel submersible transects at each dive site. Trawl catches were processed for total number and weight by species. The fork lengths of rockfish were measured to the nearest centimeter. Fish density esti- mates were derived from the number of fish captured and the seafloor area swept by the net. The seafloor area swept was the distance trawled (0.93-1.35 km/ haul) times the horizontal opening of the net (14m, based on measurements between the wing tips, 12.2- 14.3 m for the 400-mesh Eastern otter trawl; NMFS 1990). Measured vertical openings were 1.4—1.8 m. Data analysis The off-bottom distance monitored from the submers- ible was about 10 m, whereas the trawl sampled to about 2 m off bottom. Correlation between the percent composition offish species observed from the submers- ible and captured with bottom trawls was determined using correlation analysis (Sokal & Rohlf 1981). Cor- relations were determined for rockfish, flatfish, short- spine thornyhead Sebastolobus alascanus, and wall- eye pollock Theragra chalcogramma. Correlation analysis also was used to examine the correlation between densities of rockfish observed from the submersible and densities derived from trawl catch rates. Ratio estimates (Cochran 1977) of observed and trawl densities were then used to determine the catch efficiency of bottom trawls for rockfish. For these analy- ses, Pacific ocean perch, sharpchin rockfish S. zacentrus, redstripe rockfish S. proriger, and harlequin rockfish S. variegatus >25cm long were categorized as "large"; whereas those <25cm were categorized as "small." Most "large" rockfish observed from the sub- mersible were identified as adult Pacific ocean perch (based on symphyseal knob and body shape), whereas "small" rockfish could not be consistently identified. Rockfish were visually categorized as either large or small from the submersible, whereas trawl-caught fish were measured. Other rockfish species observed from the submersible included redbanded S. babcocki, rosethorn S. helvomaculatus, dusky S. ciliatus, silvergray S. brevispinis, yelloweye S. ruberrimus, and greenstriped S. elongatus. These solitary, demersal rockfish were identified from the submersible by their distinct color patterns, and categorized as "other" rock- fish. Results and discussion Submersible dives Six submersible dives were completed in 1988 and four- teen in 1989 at 188-290 m depths. Thirteen of the dive sites were classified as trawlable, three as marginally trawlable, and four as untrawlable. Of the 9278 fish observed from the submersible, 5317 were rockfish (Table 1). Rockfish were the most abundant fish on 11 90 Fishery Bulletin 91(1). 1993 dives and second in abundance on 9 dives. Of the rock- fish, 76% were large, 21% small, and 3% other. Other commonly observed fish were shortspine thornyhead, flatfish, and walleye pollock (Table 1). Distribution of large rockfish Large rockfish were solitary or in groups of 2-200 fish. Of the 4020 large rockfish observed, 998 (25%) were solitary and 3022 (75%) were divided about equally among groups of 2-10, 11-50, and 51-200 (Table 2). Although 53% of the rockfish were in groups of more than 10 fish, the number of these groups were few: 56 groups of 11-50 fish and 9 groups of 51-200 fish, com- pared with 303 groups of 2-10 fish (Table 2). The in- frequent occurrence of large groups probably contrib- utes to the high variability of trawl catch rates encountered during rockfish surveys. Regardless of group size, large rockfish were sepa- rated by 1-4 m, were similar in size within each group, Table 1 Numbers of rockfish and other fish observed at 20 submersible dive sites in the eastern Gulf of Alaska, 1988 and 1989. Distance between Surveyed Number offish Site sites Depth area Site type* (km) (ml (10%r) Rockfish Thornyheads Flatfish Pollock Other Total 1 MT 2.6 204-208 13.3 80 96 59 3 34 272 2 MT 3.5 204-209 13.8 304 221 7 16 548 3 MT 44.1 188-211 15.2 151 319 49 1 106 626 4 T 0.6 196-198 14.5 104 129 24 48 17 322 5 T 9.3 190-193 13.7 124 79 45 47 10 305 6 T 55.2 202-207 10.0 42 135 19 3 199 7 T 2.8 203-207 15.3 198 178 42 9 5 432 8 U 1.7 226-290 9.0 44 86 2 5 137 9 T 220-227 13.2 47 89 12 4 152 0.7 10 T 3.2 210-213 13.6 83 121 40 15 259 11 T 1.1 204-210 13.9 376 88 50 35 6 555 12 T 8.2 207-210 12.1 234 122 58 53 2 469 13 T 3.2 192-192 12.0 1015 37 60 60 10 1182 14 T 8.0 195-200 12.3 760 58 53 42 4 917 15 T 24.6 213-221 10.5 98 92 20 6 216 16 T 2.8 197-208 14.8 863 45 40 15 4 967 17 T 4.8 192-207 12.9 136 23 96 37 4 296 18 U 128.0 192-211 11.8 298 188 10 10 506 19 U 9.1 201-259 12.0 248 260 12 9 529 20 U v-trawlable 195-259 13.2 Totals site; T = trawlable site; U 112 250 13 14 389 5317 2616 = untrawlable site. 711 353 281 9278 *MT = = marginal! Krieger: Distribution and abundance of Sebastes spp Table 2 Numbers of solitary and grouped rockfish and distribution by group size of large and small rockfish observed at 20 submersible dives sites in the eastern Gulf of Alaska, 1988 and 1989. Solitary Grouped No. 2-10 11- ■50 51- -200 Total No. No. No. No. No. No. No. No. Site fish groups fish groups fish groups fish groups fish Large rockfish 1 38 4 9 4 9 2 71 13 35 2 32 15 67 3 77 15 33 15 33 4 33 7 20 7 20 5 90 9 20 9 20 6 20 7 71 33 116 33 116 8 11 4 9 4 9 9 18 8 24 8 24 10 57 7 21 7 21 11 75 28 80 1 36 1 160 30 276 12 34 14 57 5 98 19 155 13 172 48 90 17 283 3 300 68 673 14 114 40 131 10 197 2 233 52 561 15 17 10 50 2 27 12 77 16 38 29 93 17 373 3 325 49 791 17 24 12 41 1 41 13 82 18 14 8 21 1 20 9 41 19 10 8 31 8 31 20 14 6 16 6 16 Totals 998 303 897 56 1107 9 1018 368 3022 Small rockfish 1 16 3 14 3 14 2 51 3 9 3 69 6 78 3 13 6 17 6 17 4 16 7 30 7 30 5 5 2 5 2 5 6 18 2 4 2 4 7 1 1 3 1 3 8 17 2 7 2 7 9 4 10 1 3 1 3 11 16 3 6 3 6 12 2 1, 4 1 35 2 39 13 49 14 58 3 56 17 114 14 26 8 19 2 33 10 52 15 1 2 1 2 16 12 1 2 1 18 2 20 17 12 3 18 3 18 18 1 10 40 5 144 15 184 19 50 14 47 1 49 1 60 16 156 20 51 8 29 8 29 Totals 360 90 317 16 404 1 60 107 781 and were usually motionless and facing the current (Fig. 2). Their vertical distribution ranged 0-7 m above bottom. Rockfish in groups of 1-5 rockfish were usu- ally 0-1 m above bottom, whereas rockfish in larger groups were 0-7 m above bottom. No large rockfish were seen above 7 m, either by the observer while as- cending, descending, or traveling off the bottom, or by the pilot who searched to 10 m above bottom. This 92 Fishery Bulletin 91(1). 1993 Figure 2 Photograph from the submersible of adult Pacific ocean perch Sebastes alutus spaced about 1 m apart and facing into the current. Current direction is indicated by the bend in the sea pens. observation is supported by results of previous studies that used echosounding equipment to locate off- bottom rockfish, and midwater and bottom trawls to capture fish. In Queen Charlotte Sound, British Co- Table 3 Densities ( no./lOOOm 2 ) of large uid s mall rockfish at trawlable (T), marginally-trawlable (MT) and untrawlable (U) sites ob- served from a submersib. e. Site Large rockfish Small rockfish T MT U T MT U 1 3.5 2.3 2 10.0 9.3 3 7.2 2.0 4 3.6 3.2 5 8.0 0.7 6 2.0 2.2 7 12.2 0.3 8 2.2 2.7 9 3.2 0.3 10 5.7 0.2 11 25.3 1.6 12 15.6 3.4 13 70.4 13.6 14 54.9 6.3 15 9.0 0.2 16 56.0 2.2 17 8.2 2.3 18 4.7 15.7 19 3.4 17.2 20 2.3 6.1 Mean densities 21.1 6.9 3.2 2.8 4.5 10.4 lumbia in 1976, Pacific ocean perch were the domi- nant species in bottom-trawl catches, but were not a significant component of midwater-trawl catches tar- geting off-bottom fish echosignals (Gunderson & Nelson 1977). In the Gulf of Alaska in 1987 and 1988, Pacific ocean perch were not captured in midwater-trawl hauls that targeted fish echosignals 10-30 m off bottom, but Pacific ocean perch were abundant in bottom-trawl hauls (NMFS Auke Bay Lab., Juneau, unpubl. cruise reps. JC 87-04 and JC 88-03). The highest densities of large rockfish observed from the submersible were at trawlable sites. Densities av- eraged 21.1 rockfish/lOOOm 2 of seafloor area at the 13 trawlable sites compared with 6.9 rockfish/1000 m 2 at the 3 marginally-trawlable sites and 3.2 rockfish/ 1000 m 2 at the 4 untrawlable sites (Table 3). About 90% (3565) of the large rockfish were associated with pebble substrate on flat or low-relief bottom. The re- maining 455 large rockfish were among a variety of rugged habitats: 226 (6%) over cobble at trawlable sites, 138 (47%) over cobble and boulders at margin- ally-trawlable sites, and 91 (62%) among ledges, coral, etc., at untrawlable sites (Table 4). The preference of trawlable substrate by Pacific ocean perch is supported by two other studies: Westrheim ( 1970) mentions that best trawl catches of Pacific ocean perch were on "good bottom"; and Matthews et al. (1989) used sunken gill nets and caught 231 Pacific ocean perch (38% of the rockfish catch) on trawlable bottom, but only 25 (2% of the rockfish catch) on untrawlable bottom. Only trawlable sites contained high densities of large rockfish, but densities varied considerably among trawlable sites. For example, sites 6 and 7 were 55.2 km apart and had a sixfold difference in large-rockfish densities; adjacent sites 16 and 17, only 2.8 km apart, had a sevenfold difference in large-rockfish densities (Table 3). These variations in abundance may have been related to the distribution of cobble habitat: All groups of more than 30 large rockfish were within 20 m of cobble habitat, although cobble averaged only 10% of the habitat at trawlable sites. The cobble was in patches <30m 2 and the cobble size was <0.5m in diameter. Distribution of small rockfish Of the small rockfish, 68% (781) were in groups of 2- 60 individuals and 32% (360) were solitary (Table 2). Distribution by group size of 2-10, 11-50, and 51-200 rockfish was 41%, 52%, and 7%, respectively. Individu- als within a group were usually separated by 25cm) and 219 juveniles (<25cm). After trawl-caught rockfish were standardized to the submersible rockfish categories, large rockfish, small rockfish, and other rockfish comprised 90%, 9%, and 1% of the trawl-caught rockfish, respectively. Other commonly occurring fish in the trawl hauls were shortspine thornyhead (17%), flatfish (13%), and walleye pollock (3%) (Table 5). The composition of fish captured in trawls and observed from the submersible were highly correlated; correlation coefficients (r) for all rockfish combined, shortspine thornyhead, flatfish, and walleye pollock were 0.93, 0.86, 0.79, and 0.72, respectively (Fig. 4). These high correlations indicate that bottom trawls sampled the same fish species observed from the submersible. Catch efficiency on rockfish Catch densities of large rockfish were highly correlated with observed densities from the submersible (r=0.88). Catch densities were Figure 3 Submersible transect courses ( — ) and trawl paths (-) at nine submersible sites where rockfish counts made from a submersible were compared with CPUE of bottom-trawl hauls, August 1989. Num- bers refer to sites in Figure 1. 94 Fishery Bulletin 91(1). 1993 Table 5 Numbers of rockfish and other fish captured wi th bottom trawls at nine submersible dive sites in the eastern Gulf of Alaska. Area Site trawled Thornv- Site type* ( 10 3 m 2 ) Rockfish heads Flatfish Pollock Other Total 1 MT 13.7 75 32 94 4 6 211 3 MT 15.3 138 239 43 52 472 7 T 16.1 342 112 53 6 16 529 10 T 18.9 358 448 58 6 20 890 11 T 16.4 1148 321 139 66 15 1689 12 T 16.4 863 355 210 24 22 1474 13 T 13.0 2918 188 320 90 50 3566 14 T 14.0 893 90 183 20 20 1206 17 T 15.5 427 87 391 142 5 1052 Totals 7162 able site; T 1872 1491 = trawlable site. 358 206 11,089 *MT = = marginally-trawi merits in response to trawling have been reported for other semi-demer- sal fish species (Wardle 1983 and 1986, Ona & Godo 1990). The low catch rates at the two marginally- trawlable sites may reflect decreased trawl efficiency in rugged habitat. Catch densities of small rockfish were also highly correlated with ob- served densities from the submers- ible (r=0.89). Catch densities were similar to observed densities at all nine sites (Fig. 5); the ratio estimate of catch-to-observed densities was 1.3:1 (SE=0.3). Two possible reasons that ratio estimates were lower for small rockfish than for large rock- fish are that ( 1 ) small rockfish es- cape through net meshes at a greater higher than observed densities at the seven trawlable sites (Fig. 5) and lower at the two marginally- trawlable sites; the ratio estimate of catch to observed densities was 2.2:1 (SE=0.4). This high ratio estimate indicates that bottom trawls are very efficient for cap- turing large rockfish, resulting in density estimates approximately twice those observed from the submersible. Herding of large rockfish to- ward the net opening may ex- plain the high catch rates of large rockfish. Submersible counts in- cluded rockfish to 7 m off bottom, whereas the trawl sampled to only 2 m off bottom. Rockfish ap- parently moved downward in re- sponse to the trawl gear. Down- ward movements alone would have resulted in a catch-to- observed ratio of about 1:1. The ratio of 2.2:1 indicates large rock- fish also moved inward toward the net opening to avoid the bridles and otter doors. The bridles and doors extend approxi- mately 7 m on each side of the 14m horizontal net opening, in- creasing by twofold the seafloor area swept by the trawl gear. Downward and inward move- THORNYHEAD 13 SAMPLEO 7 10 11 12 13 14 17 STTE OBSERVED • SAMPLED FLATFISH r. .79 POLLOCK r- .72 SITE Figure 4 Percent composition of fish groups observed from a submersible and sampled with bot- tom trawls, and correlation coefficients (r) at nine sites in the eastern Gulf of Alaska. Krieger Distribution and abundance of Sebastes spp. 95 STTE STTE OBSERVED t SAMPLED I OBSERVED + SAMPLED Figure 5 Densities of rockfish observed from a submersible and estimated from bottom-trawl catch rates, and correlation coefficients (r) at nine sites in the eastern Gulf of Alaska, August 1989. rate than large rockfish, and (2) most small rockfish use rugged habitat, which bottom trawls do not sample as effectively as smooth habitat. The 2.2:1 ratio for large rockfish and 1.3:1 for small rockfish should be considered preliminary, because these are based on only nine comparisons and the area sampled by the trawl may be underestimated. If rock- fish were captured during trawl retrieval, the area sampled was underestimated and the ratios would be less. Studies are planned to determine the sampling capabilities of the trawl during retrieval, and to in- crease the number of trawl-to-submersible compari- sons. Reliability of fish counts Fish densities were determined from the submersible by counting fish within an estimated distance. Esti- mates to within 1 m were possible because of uniform illumination and minimal change in water clarity be- tween sites. Large rockfish were ideal for counting within the illuminated area because they were brightly colored, solitary or loosely grouped, not obstructed by rugged habitat, and usually motionless. The only move- ments were by individuals moving out of the direct path of the submersible, and a few larger groups mov- ing toward the submersible. These fish swam slowly and maintained their spacing and orientation. The pi- lot observed similar rockfish behavior completely around the submersible. The species and size of rock- fish captured in the trawls con- firmed that most large rockfish observed from the submersible were adult Pacific ocean perch; 81% of the rockfish catch were Pacific ocean perch, of which 92% were >30 cm long. Besides large rockfish, short- spine thornyhead were ideal for counting because they were mo- tionless on the bottom and not obstructed by rugged habitat. Counts were biased for other fish species observed from the sub- mersible. Small rockfish and "other" rockfish were under- estimated because some were blocked from view by rugged habitat. Flatfish were underesti- mated because they blended into the bottom and were difficult to see more than about 3 m from the submersible. Walleye pollock and sablefish Anoplopoma fimbria re- acted both positively and negatively to the submers- ible, and the accuracy of their counts could not be determined. Application to bottom trawl assessments Estimates of rockfish abundance derived from bottom trawl assessments are based on the assumptions that rockfish densities at untrawlable sites are similar to their densities at trawlable sites, and that the seafloor area sampled is determined from the horizontal open- ing of the net. Results from this study indicate these assumptions are incorrect for Pacific ocean perch. High densities of adult Pacific ocean perch were observed only at trawlable sites; hence, extrapolation of catch rates from trawlable substrate to untrawlable substrate would overestimate their abundance. Also, seafloor area sampled may include area swept by the bridles and otter doors, resulting in additional overestimates of abundance. Acknowledgments I thank the crews of the submersible Delta and the support vessel Wm. A. McGaw for completing safe and successful dives under adverse weather conditions. I thank the crew and scientists aboard the NOAA RV Townsend Cromwell for their detailed sampling that allowed comparisons of bottom-trawl catches to sub- 96 Fishery Bulletin 91(1). 1993 mersible observations. I also thank Victoria O'Connell of the Alaska Department of Fish and Game and John Karinen for participating in the submersible diving, and Dr. Richard Carlson and Nancy Maloney for their advice and assistance in writing this paper. The Delta was chartered by NOAA's National Undersea Research Program (NURP). Citations Balsiger, J.W., D.H. Ito, D.K. Kimura, D.A. Somerton, & J.M. Terry 1985 Biological and economic assessment of Pacific ocean perch (Sebastes alutus) in waters off Alas- ka. NOAA Tech. Memo. NMFS F/NWC-72, NMFS Alaska Fish. Sci. Cent., Seattle, 210 p. Carlson, H.R., & R.R. Straty 1981 Habitat and nursery grounds of Pacific rockfish, Sebastes spp., in rocky coastal areas of southeastern Alaska. Mar. Fish. Rev. 43(7):13-19. Cochran, W.G. 1977 Sampling techniques, 3rd ed. John Wiley, NY, 428 p. Gunderson, D., & M.C. Nelson 1977 Preliminary report on an experimental rockfish survey conducted off Monterey, California and in Queen Charlotte Sound, British Columbia during Au- gust-September, 1976. Interagency Rockfish Survey Coord. Comm. Avail. Inst. Mar. Sci., Univ. Wash., Seattle, 82 p. Hart, J.L. 1973 Pacific fishes of Canada. Fish. Res. Board Can. Bull. 180, 740 p. Heifetz, J., & D.M. Clausen 1990 Slope rockfish. In Gulf of Alaska Groundfish Plan Team (eds.), Stock assessment and fishery evalu- ation report for the 1991 Gulf of Alaska groundfish fishery, p. 140-161. N. Pac. Fish. Manage. Counc, P.O. Box 103136, Anchorage AK 99510. Leaman, B.M., & D.A. Nagtegaal 1986 Biomass survey of rockfish stocks in the Dixon Entrance-southeast Alaska region, July 5-22, 1983 (R/V G.B. Reed and MV Free Enterprise No. 1). Can. Tech. Rep. Fish. Aquat. Sci. 1510, 56 p. Matthews, K.R., J.R. Candy, L.J. Richards, & CM. Hand 1989 Experimental gill net fishing on trawlable and untrawlable areas off northwestern Vancouver Island, from the MV Caledonian August 15-28, 1989. Can. Manuscr. Rep. Fish. Aquat. Sci. 2046, 78 p. NMFS (National Marine Fisheries Service) 1990 ADP Code Book. Resource Assess. & Conserv. Eng. (RACE) Div., NMFS Alaska Fish. Sci. Cent, Se- attle, 78 p. Ona, E., & O.R. Godo 1990 Fish reaction to trawling noise: the significance for trawl sampling. In Developments in fisheries acoustics. Rapp. P.-V Reun. Cons. Int. Explor. Mer 189. Pearcy, W.G., D.L. Stein, M.A. Hixon, E.K. Pikitch, W.H. Barss, & R.M. Starr 1989 Submersible observations of deep-reef fishes of Heceta Bank, Oregon. Fish. Bull., U.S. 87:955-965. Sokal, R.R., & F.J. Rohlf 1981 Biometry, 2nd ed. W.H. Freeman, NY, 859 p. Straty, R.R. 1987 Habitat and behavior of juvenile Pacific rockfish {Sebastes spp. and Sebastolobus alaseanus) off south- eastern Alaska. NOAA Symp. Ser. Undersea Res. 2(2):109-123. Wardle, C.S. 1983 Fish reactions to towed fishing gears. /;; Macdonald, A.G., & I.G. Priede (eds.), Experimental biology at sea, p. 167-195. Academic Press, NY. 1986 Fish behaviour and fishing gear. In Pitcher, T.J. (ed.), The behaviour of teleost fishes. Chap. 18, p. 463-495. Croom Helm, London. Westrheim, S.J. 1970 Survey of rockfishes, especially Pacific ocean perch, in the Northeast Pacific Ocean, 1963-66. J. Fish. Res. Board Can. 27:1781-1809. AbStraCt.-Along the U.S. east coast, the bluefish Pomatomus saltatrix spawns in offshore conti- nental shelf waters during at least two distinct periods: spring and summer. Juveniles migrate to in- shore nurseries where they complete the first growing season. Previous studies have shown that diet during the oceanic larval stage consists of copepods, while older juveniles cap- tured inshore feed largely on teleost prey. To determine timing of the on- togenetic shift in diet to piscivory, we examined the feeding habits of 189 early-juvenile bluefish (18- 74mmTL). Samples were collected from continental shelf waters of the Middle Atlantic Bight (MAB) during spring and summer of 1988 and 1989. Spring- and summer-spawned P. saltatrix differed in body size, prey size, and in the proportions of prey types consumed. Copepods were the most common prey type in fish <60mm. Teleost prey appeared ini- tially in the diet of 30 mm in- dividuals and became the major di- etary item in spring-spawned fish >40mmTL. Gut fullness and inci- dence of piscivory peaked in late af- ternoon and were positively corre- lated with daylight hours. There was no evidence of an abrupt increase in mouth width associated with this on- togenetic shift in diet. Because juve- nile bluefish migrate inshore soon after becoming piscivores, their im- pact as predators on the abundance of other young fishes is probably fo- cused on inshore/estuarine, rather than offshore species. Ontogenetic shift in the diet of young-of-year bluefish Pomatomus saltatrix during the oceanic phase of the early life history* Rick E. Marks Marine Sciences Research Center, State University of New York Stony Brook. New York 1 1 794-5000 Present address: National Fisheries Institute, Inc , 1 525 Wilson Boulevard. Suite 500, Arlington. Virginia 22209 David O. Conover Marine Sciences Research Center. State University of New York Stony Brook, New York 1 I 794-5000 The bluefish, Pomatomus saltatrix (Linnaeus), occurs along the Atlantic coast of North America from Florida to the Gulf of Maine. Throughout its range the species is found from shal- low coastal waters to the outer, con- tinental shelf at various times of the year (Bigelow & Schroeder 1953, Lund & Maltezos 1970, Kendall & Walford 1979, Nyman & Conover 1988). Pomatomus saltatrix is an im- portant component of the recreational fishery along the east coast of North America (NMFS 1991). Two major spawning concentra- tions exist in the western Atlantic. Spring spawning occurs in the South Atlantic Bight (SAB) during March- May, with a peak in April. Summer spawning occurs in the Middle Atlantic Bight (MAB) during the months of June-September, with a peak in July (Kendall & Walford 1979, Nyman & Conover 1988, McBride & Conover 1991). The spring-spawned larvae move northeastward in waters associated with the Gulf Stream. Juveniles cross shelf waters at an age of 40-70 d, and enter bays and estuaries of the Manuscript accepted 4 November 1992. Fishery Bulletin, U.S. 91:97-106 ( 1993). ■"Contribution 866 of the Marine Sciences Re- search Center, State University of New York at Stony Brook. mid-Atlantic coast in late spring (Kendall & Walford 1979, Nyman & Conover 1988). Summer-spawned lar- vae may either spend most of the summer at sea or inhabit the inshore nursery areas of the MAB for a brief period before the onset of autumn, when both cohorts move southward to wintering grounds in the SAB (Kendall & Walford 1979, Nyman & Conover 1988). Once inshore, P. saltatrix feed al- most exclusively on piscine prey (Lassiter 1962, Richards 1976, Naughton & Saloman 1978, McDer- mott 1983, Smale & Kok 1983, Smale 1984, Olla et al. 1985, Friedland et al. 1988). Very little is known, how- ever, about the diet of young P. saltatrix at sea. In the only published account, the diet of newly-hatched P. saltatrix was found to consist mainly of copepods (Kendall & Nalpin 1981). Hence, at some point during ontogeny there must be a shift in diet from zooplankton to fish. If the diet shift occurs early in development, then juvenile P. saltatrix may be im- portant predators of larval fishes on the shelf. On the other hand, the shift from a zooplankton to a fish- dominated diet may occur coinciden- tally with the habitat shift to estua- rine waters. 97 98 Fishery Bulletin 91 ( 1993 In this paper we examine the feeding ecology of young juvenile P. saltatrix prior to their arrival in the nursery areas of the MAB. We document the predator size at which teleost prey initially appear in the diet and whether or not this shift to piscivory occurs coin- cidentally with the habitat shift inshore. Since fish are considered gape-limited predators (Hartman 1958, Ross 1978, Hunter 1980, Roberts et al. 1981), we examine mouth width for abrupt changes that may accompany an ontogenetic shift in diet. Methods Field collections Spring- and summer-spawned juvenile P. saltatrix were sampled from transects across the MAE in 1988 and 1989. The area sampled extended along the eastern U.S. coast from Cape May, New Jersey northeast to Montauk, New York and to 125 miles offshore. Cruises were conducted during two time-periods: 4 April 1988-8 August 1988, and 4 April 1989-8 August 1989. A total of 275 stations were sampled during 16 cruises. Juvenile P. saltatrix were collected at 46 of these sta- tions (Figs. 1,2). Offshore stations were located lOnmi (18.52 km) apart, and several coastal stations were lo- cated within lmi (1.85 km) of the shoreline (for com- prehensive cruise track maps, see Hare & Cowen 1991, Marks 1991). Cruise duration was 3-7 d with tows con- ducted at regular intervals, 24 h/d. Tows consisted of a 10 min net deployment filtering approximately 3500 m :! of seawater, at a ship speed of 3-4 kn. Samples were collected using a modified Methot Frame Trawl (Methot 1986). The opening of the trawl was 5 m 2 and the mesh was 2 mm. Total net length was 13 m. Because juvenile P. saltatrix are often found near the surface (Kendall & Nalpin 1981, Collins & Stender 1987), trawls were conducted with the top 30 cm of the net above the water surface. All speci- mens were preserved in 70% ethanol. ' • iOOn, - 41" N 75" W 7J"W 71" W Figure 1 Station locations off the coast of New Jersey where spring-spawned bluefish Pomatomus saltatrix were caught during 1988 and 1989. Filled circles represent station locations (/i=9l where piscivorous bluefish i»=25) were collected i.v TL=51.6mm, SD=10.60). Open circles represent station locations (;i = 13l where non- piscivorous fish (n=33) were collected t.r TL=37.1 mm, SD=14.04). Marks and Conover. Ontogenetic shift in diet ofyoung-of-year Pomatomus saltatrix 99 1 1 1 1 1 . ,.,j£/^^**~ a ~ J BLKkhbnd Sea Girl / -41°N -m ° ° f -'" '"'■tOOa,--' ■■>i?l ° New Jersey '^feW '" tS§\ • M) ° W - , £y ° 9 /, J J3 , Vgfc. Cape May .••"' ^ A / -39°N i i i i 75° W 73 » w 71° W Figure 2 Station locations where summer-spawned bluefish Pomatomus saltatrix were caught during 1988 and 1989. Filled circles represent station locations l«=9) where piscivorous bluefish (rc=18) were collected (x TL=41.0mm, SD=7.71). Open circles represent station locations (/i = 16) where non-piscivorous fish (rc=80) were collected (.rTL=33.7mm, SD=7.42). Laboratory analysis Preserved P. saltatrix were wet weighed (±0.001 g) and measured (total length, TL, to ±0.1 mm). Fish in the 18-74 mm size-range were examined for diet compari- sons. The minimum size represents the stage at which metamorphosis into the juvenile, phase is completed (Norcross 1974). The maximum fish size represents the largest specimen collected. Stomachs were removed from the pharynx and an- terior to the intestine, cut longitudinally, and the con- tents transferred to a petri dish. The gut cavity was then washed to remove any adhering particles. Prey items were identified to species or lowest taxon pos- sible, enumerated, and measured for total length using a dissecting scope equipped with an ocular micrometer. Weighted mean prey length was calculated for each stomach to provide an accurate representation of prey length consumed. Individual bluefish often con- sumed several small prey (e.g., copepod) and one large prey (e.g., teleost). Computation of the weighted mean is equivalent to adding up all the original mea- surements and dividing the sum by the total number of measurements. Hence, the most common prey type will influence the weighted mean length in pro- portion to its numerical occurrence (see Sokal & Rohlf 1981). Protocol for determining various prey length was defined as follows: (1) largest diameter for hydrated oocytes, (2) metasome + cephalosome for copepods, (3) anterior edge of head to anterior edge of caudal rami for fish lice {Caligus spp.), (4) anterior edge of eye to anterior edge of uropod for amphipods, (5) base of rostrum to anterior edge of telson for other Crusta- cea, and (6) total length for polychaetes, ostracods, pteropods, squid (Loligo spp. ), and teleost prey. Fish scales were present in <5% of the guts and were not regarded as a prey type, but were used to indicate the occurrence of piscine prey. The presence of any teleost part (e.g., otolith, spine, fin ray) was recorded as piscine prey. 100 Fishery Bulletin 91(1). 1993 Prey dry weights were obtained either from the lit- erature or measured directly after oven-drying for 24 h at 60° C. Where individual prey weights were very low or existed as exoskeleton material, dry weights were obtained by drying a known number of fresh prey items to obtain an average weight per prey item (see Grossman 1980, Ryer & Orth 1987). Dry weights of 49 bluefish (17-74 mm) with empty guts were determined by oven-drying at 60° C until recording constant weight. The time required for a constant measure was 24-48 h, depending on individual fish size. The regression equation (Log Y=3.128 x Log X-6.059) was used to predict the dry weights of all bluefish in this study (Marks 1991). Gut fullness was measured using a ratio of prey dry weight to indi- vidual P. saltatrix dry weight. Mouth width measurements were taken to detect mor- phological changes during ontogeny. Mouth width was measured as the width at the posterior tip of the maxillaries using digital calipers (±0.01 mm) (see Hartman 1958, Ross 1978, Hunter 1980, Hunter & Kimbrell 1980). Additionally, body depth of prey fish was measured at the widest location. Diet analysis Diet was analyzed using the methods outlined by Hyslop (1980) (see also Lassiter 1962, Naughton & Saloman 1978, Friedland et al. 1988): Total num- ber of stomachs in which a food item occurred divided by the total number of stomachs (%F); total number of individu- als of a taxon divided by the total num- ber of food items (%N); and total dry weight of a taxon divided by the total dry weight of all food items (%W). Bluefish were grouped by spring- and summer-spawned cohorts based on size and date of capture (Kendall & Walford 1979, Nyman & Conover 1988, McBride & Conover 1991). Trophic ontogeny was examined by arbitrarily splitting blue- fish into size groupings and using per- cent dry weight of prey categories to as- sess dietary importance. Results General diet description A total of 189 P. saltatrix were exam- ined for gut content (Table 1). Approxi- mately 84% had food present in the stom- achs. Spring-spawned bluefish consumed an average of 31 prey items per individual compared with 85 prey items for summer-spawned fish. Spring-spawned fish were found to be significantly larger (Table 1; ^-test for means with unequal vari- ances, ^=5.24, df=156, P<0.001,) and had a greater mouth width (0.01 100 km from shore (Fig. 1). Piscivorous individuals from this cohort were larger in size (x TL=51.6 mm) and generally captured closer to shore than non-piscivorous in- dividuals (x TL=37.1mm). The difference between the mean TL for piscivorous fish versus non-piscivorous fish was significant (P<0.001, df=50, *„=6.22). In comparison, piscivorous bluefish from the summer-spawned cohort were more evenly dispersed across the shelf (Fig. 2). Mean TL of piscivorous individuals (41.0 mm) was significantly greater than that of non-piscivorous fish (x TL=33.7, P<0.001, df=92, t s =5.04). Pomatomus saltatrix tended to feed on larger teleost prey with increasing size. A linear regression of prey fish length on preda- tor length was highly significant (r 2 =0.46, n=25, P<0.001; Fig. 5). With increasing body size, P. saltatrix also tended to feed on multiple teleost prey. More than one fish prey was found only in P. saltatrix >49 mm. The maximum number of intact teleost prey found in an individual predator (49 mm) was three. A t-test for the difference in means between the size of fish that con- sumed multiple fish prey (57.1 mm) and the size of those that fed on single prey (44.8 mm) was significant (P<0.05, df=30, ^=3.265). Diel cycle Gut fullness pooled across all fish examined in the study versus time of day is depicted in Fig. 6. Gut fullness was evaluated as the ratio of total prey dry weight to individual P. saltatrix dry weight (X100). Feeding peaked during 1600-2000h (the time of sunset during the study period ranged from 1945 to 2020 h). A nonparametric Spearman Rank Correlation test (Sokal & Rohlf 1981) resulted in a significant, positive relationship (r s =0.5818, P<0.001, /? = 189) between time of day and gut fullness. The high proportion of food in the gut at approximately 2400 h may be attributed to food that was consumed during the evening crep- uscular period, since prey from this time-period were generally in an advanced stage of digestion. Experimental laboratory data 25 E 20 £j 15 i, Y=J.5107+(.1902)X r 2 -.462 n-25 (P<0.001) 20 30 40 50 60 70 BLUEFISH TOTAL LENGTH (mm) Figure 5 Regression of prey fish total length (mm) on bluefish Pomatomus saltatrix total length (mm I. Marks and Conover: Ontogenetic shift in diet of young-of-year Pomatomus saltatrix 103 46 22 1) 47 3 4 9 U 1 ] 31 1 , ' ] ib 400 600 1200 1600 2000 2400 TIME 124 hours) Figure 6 Gut fullness expressed as a ratio of total prey dry weight igi divided by individual dry weight (g) ( x 100) for bluefish Pomatomus saltatrix, as a func- tion of time of day. Squares represent the mean; vertical lines represent ±1SE. Sample sizes are pro- vided above each vertical line. showed that time to 90% digestion at 21°C is 5-7 h post-feeding in 57-199 mm bluefish (Marks 1991). A two-way test for independence (adjusted G-statistic, G od =5.98, P<0.05; Sokal & Rohlf 1981) showed that the occurrence of fish in the stomachs depended on time of day. Piscivory was restricted to daylight hours, with some incidence recorded prior to 2400 h but none during 2400-0400 h. A slight increase (12%) was observed at 0800 h; the maximum evidence of piscivory (42%) occurred at 1600 h. Mouth morphology Teleost prey had the greatest maximum thick- ness of all prey types. The relationship be- tween thickness offish prey at the widest point and P. saltatrix mouth width was positive, but the regression was marginally non-significant (P=3.91, ra=23, 0.0560 mm (Fig. 4). The appearance of crab larvae as the third most-abun- dant prey item is not surprising. The majority of crab larvae were found in spring-spawned bluefish captured close to shore. This may be a result of higher concentrations of crab larvae at frontal zones in the nearshore environment. Friedland et al. (1988) reported a high abundance of crustaceans in the diet of inshore juveniles. 104 Fishery Bulletin 91(1). 1993 There was a marked increase in the maximum size of prey consumed as spring-spawned bluefish size in- creased beyond 30mm (Fig. 3). This is in agreement with the generalization that prey size increases with increasing predator size in fishes (Brooks & Dodson 1965, Tyler 1972, Ross 1977 and 1978, Hunter & Kim- brell 1980, Roberts et al. 1981, Smale 1984, Ryer & Orth 1987, Wetterer 1989, Persson 1990). There was, however, only a moderate increase in the maximum prey sizes consumed by summer-spawned bluefish. The reasons for this discrepancy are not clear, but may be a function of prey availability and differences in size at inshore migration. Summer-spawned bluefish ap- pear inshore at smaller sizes than do spring-spawned bluefish (McBride & Conover 1991). The shift to larger prey items in summer-spawned fish may occur after inshore migration. The onset of piscivory in the 30-70 mm size-range is similar to that reported for other teleosts: 40-60 mm for characids and 60-80 mm for pimelodids (Winemiller 1989), 23mm for Micropterus salmoides (Keast 1985), and 60-100 mm for Gaclus morhua (Bowman & Michaels 1984). Piscivory in larval stages has been reported in some fishes, but apparently does not occur in bluefish. Hunter & Kimbrell (1980) found Scomber japonicus to be pis- civorous at 8 mmSL. This could perhaps be attributed to the large mouth and strong, well-developed jaws that are characteristic of larval scombrids (Fahay 1983). Hunter (1980) also reported Sphyraena argentea to be piscivorous at 4.4mm. Houde (1972) found evidence for piscivory in S. borealis at 9 mm. Despite having well-developed dentition at 4.3mm (Fahay 1983), lar- val and postlarval P. saltatrix may lack either the jaw structure, visual acuity, or swimming speed necessary to feed on other fishes. Pomatomus saltatrix are known to be visual preda- tors (Olla & Marchioni 1968, Olla et al. 1970, Van der Elst 1976). It follows that reduced light levels after the evening crepuscular period should reduce feeding efficiency. Olla & Marchioni (1968) documented that P. saltatrix detect and attack prey visually, so it is not surprising that feeding appears to be correlated with daylight periods. Kjelson et al. (1975) also reported finding the lowest proportion of food in the gut of post- larval fishes during the evening hours. Fish are considered "gape-limited" predators and are ultimately restricted by mouth size (Hartman 1958, Ross 1978, Hunter 1980, Roberts et al. 1981). Ontoge- netic shifts in diet may be related to morphological changes in mouth size during development that allow for consumption of larger prey (Ross 1978, Grossman 1980, Roberts et al. 1981). Mouth width in P. saltatrix, however, appears to increase isometrically with body size. The inclusion offish in the diet at a size of 30 mm may be attributed to changes in feeding behavior with growth, or simply a result of the mouth reaching a size that permits fish ingestion. The size at which teleost prey constitute a substan- tial portion of the diet is about 40-70 mm. This is also the size-range in which P. saltatrix juveniles recruit to the inshore waters of the MAB (Nyman & Conover 1988, McBride & Conover 1991). Hence, the dietary shift is largely coincident with a habitat shift. This is further supported by the observation that virtually all piscivorous spring-spawned P. saltatrix were captured close to shore. The limited occurrences of piscivory in the summer-spawned cohort were more evenly dis- persed across the shelf. However, summer-spawned bluefish migrate inshore at a smaller size than do spring-spawned fish (McBride & Conover 1991) and may do so largely before the onset of piscivory. Our results suggest that the overall impact of predation by young bluefish on the abundance of other fishes is prob- ably focused more on inshore rather than offshore species. Acknowledgments We thank Robert Cowen, Jonathan Hare, and numer- ous members of their laboratory for sharing samples, ship time, and information. Special thanks to Francis Juanes for his insight. We also acknowledge the help- ful comments offered by Linda Jones, Ronald Hardy, and two anonymous reviewers. An earlier version of this manuscript was submitted by R.E. Marks to the Graduate School of the State University of New York at Stony Brook in partial fulfillment of the require- ments for a Master of Science degree in Marine Envi- ronmental Sciences. This work was supported under grant NA86AA-D-SG045 to the New York Sea Grant Institute. Citations Bigelow, H.B., & W.C. Schroeder 1953 Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Bowman, R.E., & W.L. Michaels 1984 Food of seventeen species of northwest Atlantic fish. NOAA Tech. Memo. NMFS-F/NEC-28, NMFS Northeast Fish. Sci. Cent., Woods Hole MA, 183 p. Brooks, J.L., & S.I. Dodson 1965 Predation, body size and the composition of plankton. Science (Wash. DC) 150:28-35. Collins, M.R., & B.W. Stender 1987 Larval king mackerel (Scomberomorus cavalla), Spanish mackerel (S. maculatus), and bluefish (Poma- Marks and Conover Ontogenetic shift in diet of young-of-year Pomatomus saltatnx 105 tomus saltatrix) off the southeast coast of the United States, 1973-1980. Bull. Mar. Sci. 41:822- 834. Cowen, R.K., J.A. Hare, & M.P. Fahay In press Beyond hydrography: Can physical processes explain larval fish assemblages within the Middle Atlantic Bight. In Moser, H.G., & RE. Smith (eds.), Fish larva assemblages and oceanographic boundaries. Contrib. Sci. (Los Angel. ). Fahay, M.P. 1983 Guide to the early stages of marine fishes occur- ring in the western north Atlantic ocean, Cape Hatteras to the southern continental shelf. J. North- west Atl. Fish. Sci. 4, 423 p. Friedland, K.D., G.C. Garman, A.J. Bejda, A.L. Studholme, & B. Olla 1988 Interannual variation in diet and condition in juvenile bluefish during estuarine residency. Trans. Am. Fish. Soc. 117:474-479. Grossman, G.D. 1980 Ecological aspects of ontogenetic shifts in prey size utilization in the bay goby (Pisces: Gobiidae). Oecologia 47:233-238. Hare, J A, & R.K. Cowen 1991 Expatriation of Xyrichtys novacula (Pisces: Labridae) larvae: Evidence of rapid cross-slope ex- change. J. Mar. Res. 49:801-823. Hartman, G.F. 1958 Mouth size and food size in young rainbow trout (Salmogairdneri). Copeia 1958:233-234. Houde, E.D. 1972 Development and early life history of the north- ern sennet, Sphyraena borealis DeKay (Pisces: Sphyraenidae) reared in the laboratory. Fish. Bull., U.S. 70:185-196. Hunter, J.R. 1980 The feeding behavior and ecology of marine fish larvae. In Bardach, J.E., J.J. Magnuson, R.C. May, & J.M. Rheinhard (eds.i. Fish behavior and its use in the capture and culture of fishes, p. 287- 330. ICLARM (Int. Cent. Living Aquat. Resour. Man- age. ) Conf. Proc. 5. Hunter, J.R., & C A. Kimbrell 1980 Early life history of pacific mackerel. Scomber japonicus. Fish. Bull, U.S. 78:89-101. Hyslop, E.J. 1980 Stomach contents analysis-A review of methods and their application. J. Fish. Biol. 17:411-429. Keast, A. 1985 The piscivore feeding guild of fishes in small freshwater ecosystems. Environ. Biol. Fish. 12:19- 129. Kendall, AW. Jr., & NA. Nalpin 1981 Diel-depth distribution of summer ichthyo- plankton in the Middle Atlantic Bight. Fish. Bull.. U.S. 79:705-726. Kendall, A.W. Jr., & L.D. Walford 1979 Sources and distribution of bluefish larvae and juveniles off the east coast of the United States. Fish. Bull., U.S. 77:213-227. Kjelson, MA, D.S. Peters, G.W. Thayer, & G.N. Johnson 1975 The general feeding ecology of postlarval fishes in the Newport River Estuary. Fish. Bull., U.S. 73:137-144. Lassiter, R.R. 1962 Life history aspects of the bluefish, Pomatomus saltatrix (Linneaus), from the coast of North Carolina. M.S. thesis, N.C. State Univ., Raleigh, p. 60-103. Last, J.M. 1980 The food of twenty species of fish larvae in the west-central North Sea. North Sea Fish. Res. Tech. Rep. 60 [ISSN-0-308-5589], 44 p. Lund, WA Jr., & G.C. Maltezos 1970 Movements and migrations of the bluefish, Pomatomus saltatrix, tagged in waters off New York and southern New England. Trans. Am. Fish. Soc. 99:719-725. Marks, R.E. 1991 Ontogenetic shift in the diet of spring- and sum- mer-spawned bluefish iPomatomus saltatrix) during the oceanic phase of the early life history. M.S. the- sis, State Univ. New York, Stony Brook, 73 p. McBride, R., & D.O. Conover 1991 Comparative growth and abundance of spring versus summer-spawned young-of-year bluefish, Pomatomus saltatrix, recruiting to the New York Bight. Mar. Ecol. Prog. Ser. 78:205-216. 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 McLachlan, A., & T Erasmus (eds.), Sandy beaches as ecosystems, p. 529-537. W. Junk, The Hague, Netherlands. Methot, R.D. Jr. 1986 Frame trawl for sampling pelagic juvenile fish. Calif. Coop. Oceanic Fish. Invest. Rep. 27:267-278. Morse, W.W. 1989 Catchability, growth, and mortality of larval fishes. Fish. Bull., U.S. 87:417-449. Naughton, S.P., & C.H. Saloman 1978 Food of bluefish {Pomatomus saltatrix) from the U.S. south Atlantic and Gulf of Mexico. NMFS Panama City Lab., Southeast Fish. Sci. Cent., 37 p. NMFS (National Marine Fisheries Service) 1991 Fisheries of the United States 1990. Current fish. stat. 8900, 111 p. Norcross, J.J. 1974 Development of young bluefish (Pomatomus saltatrix) and distribution of eggs and young in Virgin- ian coastal waters. Trans. Am. Fish. Soc. 3:477-497. Nyman, R.M., & D.O. Conover 1988 The relation between spawning season and the recruitment of young-of-the-year bluefish iPomatomus saltatrix) to New York. Fish. Bull., U.S. 86:237-250. Olla, B.L., & W. Marchioni 1968 Rhythmic movements of cones in the retina of bluefish, Pomatomus saltatrix, held in constant darkness. Biol. Bull. (Woods Hole) 135:530-536. 106 Fishery Bulletin 91 1 1993 Olla, B.L., H.M. Katz, & A.L. Studholme 1970 Prey capture and feeding motivation in the blue- fish, Pomatomus saltatrix. Copeia 1970:360-362. Olla, B.L., A.L. Studholme, & A.J. Bejda 1985 Behavior of juvenile bluefish, Pomatomus saltatrix, in vertical thermal gradients: Influence of season, temperature acclimation and food. Mar. Ecol. Prog. Ser. 23:165-177. Persson, L. 1990 Predicting ontogenetic niche shifts in the field: What can be gained by foraging theory? NATO ASI (Adv. Sci. Inst.) Ser. G20:303-320. Richards, S.W. 1976 Age and growth of bluefish (Pomatomus saltatrix), from east-central Long Island Sound from July through November 1975. Trans. Am. Fish. Soc. 4:523-525. Roberts, D., E. DeMartini, C. Engel, & K. Plummer 1981 A preliminary evaluation of prey selection by ju- venile-small adult California halibut (Paralichthys californicus ) in nearshore coastal waters off southern California. In Caillet, G.M., & CA. Simenstad (eds.), Gutshop 1981: Fish food habit studies, p. 173- 178. Proc, 3rd Pacific workshop. Wash. Sea Grant Prog., Univ. Wash., Seattle. Ross, S.T. 1977 Patterns of resource partitioning in searobins (Pisces: Triglidae). Copeia 1977:561-571. 1978 Trophic ontogeny of the leopard searobin, Prionotus scitulus, (Pisces: Triglidae). Fish. Bull., U.S. 76:225-234. Ryer, C.H., & R.J. Orth 1987 Feeding ecology of the northern pipefish, Syngnathus fuscus, in a seagrass community of the lower Chesapeake Bay. Estuaries 10:330-336. Smale, M.J. 1984 Inshore small-mesh trawling survey of the Cape south coast. Part 3. The occurrence and feeding of Argyrosomus hololepidotus, Pomatomus saltatrix and Merluccius capensis. J. Zool. S. Afr. 19:170-179. Smale, M.J., & H.J. Kok 1983 The occurrence and feeding of {Pomatomus saltatrix) (elf) and (Lichia amia) (leervis) juveniles in the Cape south coast estuaries. J. Zool. S. Afr. 18:337-342. Sokal, R.R., & F.J. Rohlf 1981 Biometry. W.H. Freeman, San Francisco, 859 p. Tyler, A.V. 1972 Food resource division among northern marine demersal fishes. J. Fish. Res. Board Can. 29:997- 1003. Van der Elst, R. 1976 Gamefish of the east coast of southern Africa. I. The biology of the elf, Pomatomus saltatrix (Linneaus), in the coastal waters of Natal. S. Afr. Assoc. Mar. Biol. Res. 44, 59 p. Wetterer, J.K. 1989 Mechanisms of prey choice by planktivorous fish: Perceptual constraints and rules of thumb. Anim. Behav. 37:955-967. Wilkinson, L. 1987 SYSTAT The system for statistics. SYSTAT Inc., Evanston IL. Winemiller, K.O. 1989 Ontogenetic diet shifts and resource partitioning among piscivorous fishes in the Venezuelan Llanos. Environ. Biol. Fishes. 26:177-199. Abstract. -The spatial distribu- tions of marine biota are frequently patchy. Samples taken from these populations are characterized by val- ues which are mostly small, relative to the population mean, and a few that are very large. It is therefore difficult to estimate stock size using conventional methods. We performed Monte Carlo simulations based on trawl data for Dungeness crab Can- cer magister and compared the be- havior of three estimators of central tendency: sample mean, geometric mean, and a lognormal estimate. Al- though the sample mean is unbi- ased, results indicate that single es- timates of the population mean (and thus population estimates obtained using area-swept) may be overly sen- sitive to extreme values; confidence intervals are large and capture the true value at a level well below that prescribed. Estimates of the geomet- ric mean exhibit more stable behav- ior about its parameter, with mixed results for the lognormal estimate. We propose a conservative approach based on comparison of trends found in each of the three estimators. Moreover, we suggest that abun- dance of aggregated stocks should be indexed with an estimator that has more desirable statistical prop- erties, such as the geometric mean. This may reduce error associated with conventional fisheries stock- assessment practices and thus pro- vide for more effective management of overdispersed stocks. Trawl survey estimation using a comparative approach based on lognormal theory* Robert A. McConnaughey School of Fisheries. WH-10 University of Washington, Seattle, Washington 98 1 95 Present address. Alaska Fisheries Science Center, National Marine Fisheries Service. NOAA. 7600 Sand Point Way NE, Seattle, Washington 981 1 5-0070 Loveday L. Conquest Center for Quantitative Science, University of Washington Seattle. Washington 98195 Effective scientific management of fishery resources is dependent upon reliable measures of stock abundance. To this end, research trawl surveys are routinely used in concert with fishery catch statistics to provide es- timates of population parameters. The analytical procedures used often rely on the assumption that statisti- cal methods based on normal prob- ability theory are appropriate and, as such, that the individuals compris- ing the population are not aggregated in space (Elliott 1977). However, ma- rine biota are commonly overdis- persed, and frequently it is the loga- rithms of abundance (or biomass) which conform to the normal or Gaussian distribution (reviewed by McConnaughey 1991). Rather than an artifact of sampling, in many cases this spatial attribute is the product of behavioral responses and/or physi- cal processes affecting dispersal (e.g., Epifanio 1987, Dew 1990). Samples taken from these populations are characterized by mostly small values relative to the population mean, and a few very large ones. Under these circumstances, single estimates of the population mean from the arithmetic mean (sample average) may be too low because very large values are of- Manuscript accepted 28 October 1992. Fishery Bulletin, U.S. 91:107-118 ( 1993) ■"Contribution 853 of the School of Fisheries, University of Washington. ten underrepresented at the levels of sampling effort common to research trawl surveys. When large catches are present in a sample, variance es- timates may become excessively high (e.g., Otto 1986). This may introduce a high degree of uncertainty into the resource management process which, if ignored, can have potentially seri- ous repercussions (Ludwig & Walters 1981). We investigated two alternative measures of central tendency and compared their statistical behavior with that of the arithmetic mean. These alternatives are the geometric mean and a model-based estimate of the arithmetic mean based on log- normal theory. An evaluation of trends based on a comparison of the three estimators is proposed. This approach may identify error associ- ated with the conventional index of abundance, thereby reducing the like- lihood of false conclusions concern- ing trends in stock abundance. Data and methods Monthly trawl surveys of Dungeness crab Cancer magister abundance along the southern Washington coast provided representative values of density (rc/ha) for analysis with Monte Carlo techniques. Density data 107 Fishery Bulletin 91(1). 1993 such as these are commonly expanded according to area-swept procedures to produce estimates of popula- tion size. Both nearshore and estuarine populations were sampled, and we examined data collected during four consecutive years. Survey design and methodol- ogy are discussed in Armstrong & Gunderson (1985) and Gunderson et al. ( 1990). The spatial dispersion of a population determines the relationship between its mean abundance and vari- ance, and this information may be used to select an advantageous data transformation (Elliott 1977). Strong linear relationships between the means and standard deviations of our density data (r 2 =0.98 and 0.70, with P«0.001 and P«0.001 for the coastal and estuarine areas, respectively) and graphical analysis of log-transformed data suggested a logarithmic trans- formation would be appropriate. In order to test this assumption, the Kolmogorov test for normality was applied to the density data for each cruise, both before and after transformation (Table 1). These preliminary analyses suggested that the density data were lognor- mally distributed and, as such, that individuals in the crab population were aggregated in space. However, lognormal theory cannot be applied directly to any sample that contains a zero value, since the logarithm of zero is undefined. Since our data exhibit only the occasional zero catch, we used the common In (X+l) transformation to normalize the data. An alternative approach, when a significant fraction of the data con- sists of zero catches, would be to use the A-distribution (Pennington 1983 and 1986, Smith 1988), which is es- sentially a lognormal distribution with a proportion (A) of zeros. The lognormal distribution and parameter estimation The lognormal distribution may be represented as a Gaussian distribution of logarithmic data or, equiva- lent^, as a right-skewed distribution of untransformed data (Aitchison & Brown 1969). A brief review of the density function and relevant parameters for the log- normal distribution appear in the Appendix. There, and throughout the text, we use the following notation to distinguish between untransformed and transformed scales and between population parameters and their estimates: X represents untransformed density val- ues, while X (the ordinary sample mean) and s~ x (the Table 1 Goodness-of-fit probabilities from Kolmogorov tests for normality with means (F) and standard deviations (s Y ) for log-transformed Cancer magister ibundanc e data . Cruise refers to sequential trawl surveys n=num ber of samp es) in coastal and estuarine areas along the ^ outhern Washing- ton coast over a consecutive 4-yeai period. Cruis Coast Estuary 3 ;i Raw 1 Log 2 Y s v n Raw Log y s, 1 35 .006 .964 4.20 2.35 20 .211 .890 6.96 1.19 2 41 .000 .740 4.64 2.50 20 .129 .816 6.82 1.10 3 42 .000 .687 4.95 2.86 20 .909 .651 6.88 .66 4 38 .000 .738 6.65 2.13 20 .133 .207 6.59 .79 5 42 .000 .932 5.30 2.26 20 .167 .434 6.34 .56 6 44 .014 .287 3.34 1.97 16 .373 .411 4.90 1.72 7 40 .109 .230 3.08 1.72 20 .004 .418 6.04 1.21 8 41 .000 .800 3.59 1.60 20 .088 1.000 5.65 1.21 9 44 .000 .164 2.91 2.08 20 .013 .573 5.68 1.36 10 43 .016 .528 3.90 1.78 20 .140 .878 5.45 .96 11 44 .007 .102 3.16 2.10 20 .172 1.000 4.64 1.16 12 44 .002 .829 3.62 1.93 20 .129 .705 5.81 1.01 13 44 .000 .759 4.55 2.15 20 .435 .979 6.61 1.04 14 44 .000 .507 3.90 2.49 20 .311 .960 5.93 1.06 15 44 .000 .306 5.05 2.20 20 .336 .981 6.42 1.10 16 44 .000 .484 3.62 3.84 20 .027 .931 6.43 1.91 17 44 .000 .471 3.44 4.94 20 .434 .922 6.93 .82 18 44 .000 .370 3.69 4.73 19 .336 .868 6.82 1.18 19 44 .000 .949 4.75 6.89 20 .388 .980 6.43 .92 20 43 .000 .528 rmed data. 3.90 6.48 20 .010 .964 6.23 1.29 'P-val ue for untransfo -P-val ue for log-transformed data. McConnaughey and Conquest: Comparative trawl-survey estimation based on lognormal theory 109 sample variance) are estimators of u and o 2 , the popu- lation mean and variance of the untransformed data. Letting Y = In (X), Yand s 2 Y are estimators of u LN and o 2 ln the population mean and variance of the log- transformed data. Note that in Eq. (A3) and (A4) both u and o 2 for the lognormal distribution are functions of two parameters, u LN and a 2 LN , making the former parameters difficult to estimate. In particular, any es- timate of u involves both location and dispersion pa- rameters. Therefore, variation in estimating u will come from two sources: variation in estimating u LN and variation in estimating a 2 LN . The arithmetic mean (AM) The ordinary sample mean is an unbiased estimator of u regardless of the under- lying frequency distribution. When the underlying dis- tribution is normal, the sample mean is also the mini- mum variance unbiased estimator (MVU, the one with the smallest variance of all unbiased estimators) of u. However, the sample mean does not have this MVU property when the underlying distribution is lognor- mal (Gilbert 1987). Moreover, the AM is sensitive to the presence of one or more large data values, particu- larly for small sample sizes. For lognormal data, these extreme values are not outliers; they simply reflect the right-skewed nature of the distribution. Finney (1941) demonstrated the inefficiency of the sample mean when the variance of the natural logarithms is greater than 0.69, and Koch & Link (1970) suggested using the sample mean only when the coefficient of variation is believed to be less than 120%. For highly- skewed distributions such as the lognormal, sample sizes in excess of 200 may be necessary to invoke the Central Limit Theorem, which justifies use of the sample mean for inferences about means of popula- tions that are not normally distributed (Sissenwine 1978, Jahn 1987). The Finney-Sichel estimator (FM) Among alternative estimators that have been investigated is an MVU es- timator of u (Finney 1941, SicheJ 1952), which also has been described as equivalent to a maximum- likelihood estimator for lognormal data (Aitchison & Brown 1969). The Finney-Sichel method adjusts the geometric mean upwards and is commonly used in gold and trace-mineral assay work, where ore concentra- tions are typically lognormally distributed ( Sichel 1952 ). If Yand s 2 Y represent the ordinary sample mean and variance of the log-transformed values, the Finney- Sichel estimate for \i is 1 (n-l)t 1 + + (n-l)H 2 (n-l)V FM = exp(Y)y n (r) (1) where n is the sample size and \j/ n (t) is the infinite series 2\n-(n+l) 3ln 3 (n+l)(n+3) (n-Vt 4 4!n 4 (n+l)(rc+3)(n+5) (2) + .. with t = —^— . The function \|/ n (t) is defined such that E[\|/ n (s 2 )] = exp f-^p o 2 ) and ^^ [\j/ n (s 2 >] = exp(o 2 ); it is used extensively with the lognormal distribution (Smith 1988). In their book, Aitchison & Brown (1969) included tables of y„ for computing the Finney-Sichel estimate. More extensive tables are provided in Link et al. (1971), who claim that linear interpolation be- tween tabled values gives close approximation for esti- mates of u. They also include a FORTRAN program for calculating the \\i n function, which we used in com- puting FM, the Finney-Sichel estimate of the popula- tion mean. (A version of this program may be obtained from the authors. ) Confidence limits for the lognormal mean are not symmetric because of the skewed nature of the under- lying distribution. Hence, it becomes necessary to com- pute separate upper and lower confidence limits. Land (1971, 1975) obtained upper one-sided 100(1-00% and lower one-sided 100a% confidence limits for the log- normal mean, where a is the frequency of type I error: C/L, exp Y + s 2 Y s Y H h LL a = exp Y + — + S v Vn-1 (3) (4) The quantities H^, and H„ [functions of a, (n-1) and s Y ] are obtained from tables in Land ( 1975) for sample sizes of n>3. The geometric mean (GM) The geometric mean, e^, will be a biased estimate of |a (Appendix) but may be more precise with respect to its population parameter than will be the case for estimators of the population mean. (Actually, £(e ? ) = exp(u LN + - — - so the GM is 2/JOln biased even for e^ L ' N ', but this bias decreases rapidly as n increases.) When exponentiated, the population mean of the transformed data, u LX . is the geometric mean catch and, equivalently, the median catch for lognor- mal data. It remains unaffected by skewness, a func- tion of [exp(o 2 LN -D]. It is less affected by large values of X, owing to the nature of a logarithmic transforma- tion; hence, its sampling distribution is less skewed Fishery Bulletin 91 fl), 1993 than that for the AM. Aitchison & Brown (1969) note that "since the arithmetic mean involves both the lo- cation and dispersion parameters, it is not a pure mea- sure of the [response variable] under the lognormal hypothesis: for this the geometric mean or median is to be preferred." Monte Carlo simulations based on crab trawl data Monte Carlo simulations consist of calculations made on data sets whose elements are randomly selected from specified probability distributions. This approach permits an evaluation of various point-estimation pro- cedures on the basis of expected outcomes. It also al- lows a closer examination of individual cases than is possible with a purely analytical approach and per- mits an evaluation of the effects of sample size. For this investigation, single values of mean density and standard deviation were calculated for each cruise in the two trawl locations along the Washington coast. The means of these statistics were used to define two representative lognormal distributions, which are iden- tified as lognormal (4,2) for the coastal area and log- normal (6,1) for the estuarine area. These distribu- tions have means of 4.0 and 6.0, and standard deviations of 2.0 and 1.0, respectively, for the log- transformed variable; they will be referred to as LOGN (4,2) and LOGN (6,1) (Fig. 1). From these two prob- ability distributions, we created 1000 sets of simu- 0.03 LOGN (4,2) — LOGN (6,1) • GEOMETRIC PROBABILITY o O ARITHMETIC 0.01 300 600 900 1200 15CM ) X Figure 1 Probability density functions for the representative lognor- mal distributions. Included are reference marks to indicate values of the respective arithmetic and geometric means. lated density data for each of 13 sample sizes (2,4,6,8,10,15,20,25,30,35,40,45,50) using a pseudoran- dom number generator (Minitab, Inc., University Park PA). Sample sizes were selected to encompass the range of values associated with ongoing trawl surveys. Table 2 presents descriptive statistics for each of these data sets. (These data sets are archived on magnetic tape, and access can be arranged through the authors. ) We investigated three methods of estimating cen- tral tendency. The AM method consisted of computing arithmetic means and traditional confidence intervals (e.g., at 90% confidence) based on the Student's /-distribution. The FM method used the Finney-Sichel estimator for the mean of a lognormal distribution as presented in Eq. (1) and (2). For confidence limits, the method by Land (1971, 1975) as presented in Eq. (3) and (4) was used. The GM method used e 1 as an esti- mate of e^ LN , the geometric mean (or median) in the untransformed scale. A 90% confidence interval was derived as follows: exp Y-L 1y_ exp y+ '-' t (5) This method estimates a different parameter (the me- dian rather than the population mean) than the first two methods. However, because the median is asymp- totically a function of only a single parameter, u LN , the GM method tends to give more stable estimates of its parameter, and it is worthwhile to compare its perfor- mance as another index of central tendency to the first two methods. Comparison measures to evaluate performance of the estimators We used the following measures of comparison to evalu- ate the performance of the estimators: root mean squared error (RMSE), deviation of the estimate from the true parameter (BIAS), average length of the 90% confidence interval (AVL), standard deviation of the 90% confidence interval length (SDCI), and percent containment of the parameter by the confidence inter- val estimate (PERCON). These were estimated as follows: RMSE= V (estima ted pa ra meter ^ from ; lh data set — true value)' 2 . 1000 Root mean squared error is a measure of the average variation in the estimated mean relative to the true McConnaughey and Conquest: Comparative trawl-survey estimation based on lognormal theory 1 I 1 Table 2a Descriptive statistics for 1000 simulated trawl data sets representati se of coastal popu- lations of Cancer magister in Washington [Lognormal (4,2)]. Trimmed means calculated using the central 90** of the individual data sets. Min/Max refer to the minimum and maximum values in the data set. Standard n Arithmetic mean Trimmed mean deviation Min Max Raw' Log b Raw Log Raw Log 2 431 4.104 192 4.106 1,949 2.026 <1 48,860 4 382 4.003 163 4.004 1,714 1.999 <1 64,248 6 398 3.987 162 3.988 1.946 2.007 <1 62,131 8 370 3.985 157 3.984 1,818 1.982 <1 71,789 10 386 4.001 161 4.002 1,964 1.996 <1 110,761 15 378 4.002 163 4.007 1,723 2.002 <1 82,492 20 391 4.004 162 4.003 1,920 1.997 <1 122,967 25 394 3.999 159 4.002 2,313 1.991 <1 160,546 30 407 4.002 161 4.004 2,603 2.000 <1 236,341 35 422 4.016 162 4.015 3,301 1.992 <1 380,743 40 419 3.999 161 3.999 7,609 1.997 <1 1,479,353 45 397 3.989 162 3.989 2,167 2.010 <1 226,970 50 411 4.003 163 4.003 2,525 2.008 <1 323,734 True value of\i is 403.43 = exp n/ha). K) i see Appendix). * untransformed density b log-transformed density (rc/hal. Table 2b Descriptive statistics for 1000 simulated trawl data sets representative of estuarine popul ations of Dungeness crab Cancer magister in Washington [Lognormal (6,1 )|. Trimmed means calculated using the central 90% of the individual data sets Min/Max refer to minimum and maximum values in the data set. Standard n Arithmetic mean Trimmed mean deviation Min Max Raw' Log" Raw Log Raw Log 2 633 5.968 528 5.971 739 .994 10 5,703 4 676 6.007 552 6.007 860 1.012 10 10,525 6 682 6.026 557 6.024 924 .993 11 24,259 8 690 6.010 553 6.006 955 1.016 12 20,499 10 659 5.993 541 5.995 835 1.004 6 13,333 15 650 5.986 537 5.986 816 .993 12 18,877 20 655 5.992 538 5.993 825 .998 8 13,756 25 667 6.000 544 5.999 868 1.003 4 19,785 30 664 5.995 543 5.995 857 1.004 5 21,060 35 664 5.999 541 5.998 901 .995 7 30,309 40 657 5.996 540 5.995 844 .993 9 20,555 45 666 6.004 547 6.004 869 .997 6 30,655 50 664 6.001 544 6.002 856 .998 7 18,725 True value of]i is 665.14 = exp n/ha). [ 2 J {see Appendix). untr ansformed density b log-transformed density (n/ha). I 12 Fishery Bulletin 91(1). 1993 mean density and, as such, is a measure of accuracy. For any unbiased estimator (e.g., the AM or FM, where the expected value of the estimator is the parameter itself), the RMSE is the same as the variance of the estimator, in terms of expected value. For a biased estimate (recall that GM has positive bias in estimat- ing e MLN , which decreases as n increases), RMSE incor- porates both bias and variance. BIAS= y (estimated parameter ^ from i' h data set — true value) 1000 = (average value of estimated parameter — true parameter). Results Monte Carlo simulations Root mean squared error The RMSE was consis- tently lower for the GM than for the other measures of central tendency (Fig. 2). The FM provided point esti- mates of p that were consistently more accurate (ex- cept at very small sample sizes) than the AM method, particularly as skewness of the density data (Fig. 1) increased. The RMSE of GM estimates and of p ob- tained with the FM declined steadily as sample size increased, whereas that for the AM, although gener- ally declining, was somewhat less regular and much more erratic (see Fig. 2a, n=40). Closer inspection of the LOGN (4,2) data set revealed a single extreme The bias is the average amount by which the estimate tends to "miss" its respective parameter. 1000 1000 £ (UL-LL\ £ length, AVL = 1000 1000 where (UL-LL), = length of a single 90% confidence interval for the i ,h data set. The average length is a measure of precision. SDCI £ (length, -AVL) 2 1000 - 1 The standard deviation is a measure of the spread of the confidence-interval lengths around the average length. An estimator with the most reproducible esti- mate of the precision of the estimated mean would have confidence intervals of relatively low variability in length. The frequency with which a confidence interval in- cludes the true value of the parameter defines the con- tainment rate, PERCON. If the assumptions of sam- pling and the appropriateness of statistical model are met, 90% confidence intervals should contain the den- sity parameter being estimated approximately 90% of the time. The three estimators of central tendency (AM, FM, GM) and their confidence intervals were also calcu- lated for actual density data obtained during the monthly trawl surveys. Two large systems, termed the Coast and the Estuary, were considered. 20 30 SAMPLE SIZE Figure 2 Comparison of root mean-square error (RMSE) values associ- ated with the three estimators of central tendency according to sample size, (a) LOGN (4,2) data representative of coastal population of Cancer magister used for Monte Carlo simula- tions. See text for outlier explanation, (b) LOGN (6,1) data representative of estuarine population of C. magister used for Monte Carlo simulations. McConnaughey and Conquest: Comparative trawl-survey estimation based on lognormal theory 1 13 value out of 40,000 data points that caused a consider- able increase in the RMSE associated with the AM estimate. It is noteworthy that the magnitude of this simulated density value is in keeping with extreme values observed in the field. Accuracy of GM estima- tion improved dramatically as skewness increased, in contrast to the FM and AM responses wherein accu- racy decreased as skewness increased. Deviation of the estimator from the parameter (bias) Overall, the most extreme deviations were as- sociated with the smallest sample sizes; this disparity decreased as sample sizes increased (Fig. 3). GM esti- mates deviated less, stabilized at smaller sample sizes, and, despite the positive bias, converged much more predictably to e MLN than did AM and FM in estimating u. In general, estimates of u oscillated about the para- metric value and converged as sample size increased. The absolute values of the deviations from u are smaller for the FM than for the AM method in 17 of the 26 cases examined, without an obvious trend related to the skewness of the data. It is noteworthy that esti- mates of u obtained with the AM and FM methods are equivalent when n=2. Average length of the interval estimate The aver- age length of the 1000 909c confidence intervals (CIs) calculated for each sample size was consistently shorter for the GM (which only estimates one parameter) than for the intervals of the AM or FM (Fig. 4). Intervals calculated using the FM method were consistently larger than those for the AM method. Overall, the de- gree of difference between the three estimators decreased as sample size increased and as skewness decreased. Average lengths were inordinately large at the smallest sample sizes and decreased rapidly there- after. The average CI length for the GM decreased as skewness increased, in contrast with the behavior of CI lengths for u. 150 20 30 SAMPLE SIZE Figure 3 Comparison of degree of deviation from the parameter (bias; scaled to 0) for the three estimators of central tendency ac- cording to sample size, (a) LOGN (4,2) data representative of coastal population of Cancer magister used for Monte Carlo simulations, (b) LOGN (6,1) data representative of estuarine population of C. magister used for Monte Carlo simulations. 5,000 •5- 4,000 .c, U 3,000 S? o o> LU O 2,000 £ LU < 1,000 8,000 € 6,000 o g -J.ooo LU (3 < DC £ 2,000 < I ARITHJVIETIC FINNEY GEOMETRIC (a) • 10 20 30 40 50 1 I ARITHJVIETIC FINNEY GEOMETRIC ' ^^T--^ 5 — 1 — 1 (b) • T * I ^ 1" 1 — 1 10 20 30 SAMPLE SIZE 40 50 Figure 4 Comparison of average length of 90<7( confidence interval for the three estimators of central tendency according to sample size, (a) LOGN (4,2) data representative of coastal popula- tion of Cancer magister used for Monte Carlo simulations, (b) LOGN (6,1) data representative of estuarine population of C. magister used for Monte Carlo simulations. 14 Fishery Bulletin 9 1 ( I ). 1993 Standard deviation of the confidence interval length The standard deviation of the 1000 90% CIs calculated for each sample size was consistently lower for the GM than for the AM or FM, both in an absolute sense and relative to the average CI length (Fig. 5). At smaller sample sizes, the standard deviations for the AM were less than those for the FM. However, this pattern was reversed at larger sample sizes such that the FM had the more precise interval estimates (note the crossovers at rc=35 and n=25 in Figs. 5a and 5b, respectively). Overall, the precision of the interval es- timates declined as sample size decreased and as skew- ness increased; the effect was most pronounced for the FM method. The GM response was unique in that pre- cision increased as skewness increased. Of particular note is the dramatic loss of precision of the AM inter- val estimate apparent in Fig. 5a («=40) which, upon investigation, was attributed to a single extreme value. Parameter containment within the interval estimate The GM parameter e^ occurred within its interval estimates, as did u within the intervals ob- tained by using FM, at the prescribed 90% confidence level (Fig. 6). The rate of GM containment oscillated within 1-2% of this level under all circumstances. Con- fidence intervals for the FM contained u at the rates of 89.1-92.3% (Fig. 6a) and 89.1-93.6% (Fig. 6b); the high- est percentages were associated with the smallest sample size, perhaps due to their relatively greater lengths (Fig. 5). In contrast, AM interval estimates contained u at rates of 47.7-65.8% (Fig. 6a) and 76.9-85.3% (Fig. 6b), well below the prescribed level of confidence. Dungeness crab trawl survey data We also computed the three estimators for actual C. magister density data to assess the gain in informa- 10,000 "£ 8,000 o o 6,000 Z o § 4,000 £ Q a 2,000 i- V) 3,000 ARITHMETIC FINNEY GEOMETRIC v.k (a) * * • — ♦■-...> «,. — • — -• * • < 10 20 30 40 50 ARITHJVIETIC FINNEY GEOMETRIC 20 30 SAMPLE SIZE Figure 5 Comparison of standard deviation of the length of 90% confi- dence interval for the three estimators of central tendency according to sample size. (a) LOGN (4,2) data representa- tive of coastal population of Cancer magister used for Monte Carlo simulations, (b) LOGN (6,1) data representative of estuarine population of C. magister used for Monte Carlo simulations. 100 20 30 SAMPLE SIZE Figure 6 Comparison of percent occurrence of the parameter in the 909c confidence interval for the three estimators of central tendency according to sample size of Cancer magister. (a) LOGN (4,2) data representative of coastal population of C. magister used for Monte Carlo simulations, (b) LOGN (6,1) data representative of estuarine population of C magister used for Monte Carlo simulations. McConnaughey and Conquest Comparative trawl-survey estimation based on lognormal theory 1 15 tion over using any single estimator alone. For the Estuary data, trends in abundance routinely paral- leled one another, differing only by their relative mag- nitude (Fig. 7a). Characteristically, estimates of u from the survey data obtained with the FM method exceeded those of the AM, which, in turn, exceeded estimates of the GM parameter (e MLN ). In some cases, trends in the Coast estimates were diametrically opposed (Fig. 7b). As expected, the GM estimate was consistently lower than both AM and FM, reflecting the difference in the population parameter being estimated. Noteworthy was the reversal in the relative magnitudes of the AM and FM estimates during the interval between Cruise 2 and Cruise 4. Discussion Conventional analysis of catch data and alternatives Population estimates are routinely generated using untransformed catch data and arithmetic mean calcu- 1,400 1,200 "a & 1 ,000 800 v> UJ D 3 600 O 400 200 (a) * ARITHMETIC FINflEY GEOMETRIC, , • • • 12 13 14 15 14,000 12,000 10,000 t 8,000 (0 z w 6,000 CD < 4,000 O 2,000 (b) A ARITHMETIC FINNEY GEOMETRIC fr ^-^- ■■» » 12 3 4 5 CRUISE Figure 7 Comparison of three measures of central tendency calculated for Cancer magister using monthly cruise data for (a) the estuarine area (year 3) and (b) the coastal area (year 1). (The order of the coast/estuary figures is deliberately reversed here to illustrate certain results; see text.) lations (e.g., BIOMASS procedure of the U.S. NMFS, Gunderson et al. 1978; STRAP procedure of Can. Dep. Fish. & Oceans, Smith & Somerton 1981). Several methods for reducing the variance associated with these estimates of abundance have been used, often despite recognizable limitations. These fall broadly into two categories: (1) model-based approaches, which model the underlying distribution of the data, and (2) design- based approaches, which rely upon probability sam- pling and large sample results. Smith (1990) compared the two approaches for estimating resource abundance with trawl surveys and concluded with an example of a model-based predictive estimate using additional in- formation (salinity, temperature, depth). Other ex- amples of model-based estimation in fisheries applica- tions include use of a weighted negative binomial distribution (Zweifel & Smith 1981), the delta distri- bution (Pennington 1983), and the geostatistical tech- nique of kriging (Conan 1985). Stratification of the sampling frame is a common example of a design-based approach. Although this is theoretically appealing, Gavaris & Smith ( 1987) have demonstrated that strati- fied random sampling may be inferior to a simple ran- dom design, because of suboptimal allocation of sta- tions to strata. They suggest that a decrease in the number of strata used in the eastern Scotian Shelf groundfish survey would provide for more flexible allo- cation of total sampling effort in the future. Unfortu- nately, many of the problems attendant with specify- ing stratum boundaries will persist; these include interannual variability in distribution and abundance of stocks related to environmental factors and the typi- cal multi-species scope of most research trawl surveys. Because of these difficulties, catch data are commonly stratified after sampling is completed (Picquelle & Stauffer 1985, Otto 1986). However, post-stratification (i.e., a priori examination of catch data for the purpose of assigning strata) is not a valid approach and is not recommended (Cochran 1977). Other methods for estimation of abundance are ex- pedient, yet may be based on the specious assumption that extreme values are "outliers" and are therefore not integral to the data set. Included is the practice of eliminating extreme values or the use of trimmed (or Winsorized) means (Halliday & Koeller 1981, Bates 1987, Harding et al. 1987, Smith 1981). Ignoring in- stances where human error is involved, these ad hoc procedures may introduce substantial negative bias to estimates of the true population mean (compare u and the trimmed means in Table 2), thereby contributing to misleading conclusions about trends in the data. Design-based and model-based approaches A strict probability sampling approach (i.e., design- based and without any underlying models) requires 1 1( Fishery Bulletin 91(1). 1993 that resulting estimates be normally distributed according to the Central Limit Theorem. However, Hansen et al. (1983) state, "When surveys use rela- tively small samples, the samples may be too small for the application of the theory [for large samples] to be essentially assumption-free." Our approach is model- based. As long as one is restricted to samples that may not be considered "acceptably large" (and further ham- pered by considerable skewness caused by extreme val- ues), use of a model-based approach is not unwarranted (Little 1983). With regard to robustness, Myers & Pepin (1990) argue that exclusive use of a lognormally-based esti- mate can be sensitive to model assumptions, leading to possible bias and reduction in efficiency. Because contamination of a lognormal distribution with data from similarly-shaped distributions (e.g., Weibull or gamma) is difficult to detect for sample sizes less than 40, they suggest using lognormally-based estimators of abundance only when there is evidence that the underlying population is lognormal. Obviously, use of transformations and model-based estimators is a "double-edged sword," and these procedures should not be applied indiscriminately. When appropriate (e.g., Table 1 ), however, significant improvement in the rela- tive efficiency of the sample average and, in particu- lar, the estimated variance, can be realized (e.g., Finney 1941, Koch & Link 1970, Myers & Pepin 1990). A comparative approach If nothing is known about the spatial distribution of an organism, the sampling plan must be designed to determine distribution patterns as well as population size. Knowledge of the distribution pattern aids in se- lection of the proper estimation procedure. Based on the arguments presented above, combined with the rather ubiquitous nature of overdispersion in the ma- rine environment, we prefer an approach based on three estimators, namely the arithmetic mean, the geomet- ric mean, and the Finney-Sichel estimator of u. By taking a comparative approach, one may be reason- ably certain of apparent trends in the data if the trend is consistent for the three estimators. For the estua- rine crab population illustrated in Figure 7a, the par- allel behavior of the estimates corresponded to chang- ing values of Y coupled with nominal changes in s 2 Y (Table 1). In this case, there is no evidence to suggest that conventional analysis of catch data (i.e., using the AM method) was less than adequate. However, trends in the estimates may, on occasion, be opposed to one another, as was demonstrated for the coastal crab popu- lation (Fig. 7b). The AMs suggest a precipitous drop in abundance occurred during the interval between Cruises 3 (with two extreme values) and 4, whereas both the FM and GM procedures indicated a moderate increasing trend during the same period. From the behavior of the three estimators, we conclude that be- tween Cruises 2 and 4, u lN (and u) may have increased slightly, but cr 2 LN (and thus the skewness of the distri- bution) probably increased and then decreased, affect- ing the FM and AM estimates (the latter more strongly) but not the GM estimate. This is verified by checking the Y and s 2 Y values in Table 1. Changes in skewness relate directly to the size of the larger catches and, hence, the degree of spatial aggregation in the popula- tion. Plotting the three estimators and relating the trends back to changes in Y and changing s 2 Y has yielded some insight into the behavior of the estima- tors. It has also allowed us to extract more informa- tion about the crab population than if we had used only one estimator, the AM. In cases such as this, where there is significant disagreement among the estima- tors, the data set should be carefully evaluated as to its underlying probability distribution and the most appropriate index selected. If the lognormal distribu- tion is reasonable, the GM may well be the preferred estimator (Aitchison & Brown 1969); use of the GM may be advantageous because it is relatively insensi- tive to extreme values ( particularly so for highly-skewed data) in terms of accuracy and precision. Since catch coefficients are not routinely considered with trawl- survey data of this type, the resulting stock-size estimates are, strictly speaking, indices of abundance (Caddy 1986). Under these circumstances, it may be advantageous to use an alternative estimator of central tendency, such as the GM, to generate the index. Acknowledgments We thank Drs. David Armstrong and Donald Gun- derson of the University of Washington School of Fisheries for access to unpublished data and for their comments and suggestions. Their research program was supported by an institutional grant from Wash- ington Sea Grant (#NA86AA-D-SG044 Project R/F-68) and the U.S. Army Corps of Engineers (#DACW67-85- C-0033). In addition, we thank the following for help- ful comments and suggestions on previous drafts of this paper: Dr. Robert Crittenden (University of Washington, School of Fisheries), Charles Douglas Knechtel (FishStat Statistical Helps Service, Seattle), Stephen J. Smith (Marine Fish Division, Bedford Institute of Oceanography), and two anonymous reviewers. McConnaughey and Conquest: Comparative trawl-survey estimation based on lognormal theory Citations Aitchison, J., & J.A.C. Brown 1969 The lognormal distribution, with special refer- ence to its uses in economics. Cambridge Univ. Press, Cambridge, 176 p. Armstrong, D.A., & D.R. Gunderson 1985 The role of estuaries in Dungeness crab early life history: A case study in Grays Harbor, Wash- ington. In Melteff. B.R. led), Proceedings of the sym- posium on Dungeness crab biology and management, p.145-170. Alaska Sea Grant Prog. Rep. 85-3, Univ. Alaska. Anchorage. Bates, R.D. 1987 Estimation of egg production, spawner biomass and egg mortality for walleye pollock, Theragra chalcogramma, in Shelikof Strait from ichthyo- plankton surveys during 1981. NWAFC Proc. Rep. 87-20, NMFS Alaska Fish. Sci. Cent., Seattle, 192 p. Caddy, J.F. 1986 Stock assessment in data limited situations: The experience in tropical fisheries and its possible rel- evance to evaluation of invertebrate resources. In Jamieson, G.S., & N. Bourne leds.). North Pacific workshop on stock assessment and management of invertebrates, p. 379-392. Can. Spec. Publ. Fish. Aquat. Sci. 92. Cochran, W.G. 1977 Sampling techniques. John Wiley, NY, 428 p. Conan, G.Y. 1985 Assessment of shellfish stocks by geostatistical techniques. ICES ( Int. Counc. Explor. Sea) CM/K 30, 24 p. Dew, C.B. 1990 Behavioral ecology of podding red king crab, Parahthodes camtschatica. Can. J. Fish. Aquat. Sci. 47:1944-1958. Elliott, J.M. 1977 Statistical analysis of samples of benthic inverte- brates. Freshwater Biol. Assoc. Sci. Publ. 25, 157 p. Epifanio, C.E. 1987 The role of tidal fronts in maintaining patches of brachyuran zoeae in estuarine waters. J. Crustacean Biol. 7:513-517. Finney, D.J. 1941 On the distribution of a variate whose logarithm is normally distributed. J. R. Stat. Soc. Suppl. 7:155- 161. Gavaris, S., & S.J. Smith 1987 Effect of allocation and statification strategies on precision of survey abundance estimates for At- lantic cod (Gadus morhua) on the eastern Scotian Shelf. J. NorthwestAtl. Fish. Sci. 7:137-144. Gilbert, R.O. 1987 Statistical methods for environmental pollution monitoring. Van Nostrand Reinhold, NY, 320 p. Gunderson, D., D. Worlund, & B. Gibbs 1978 Users guide to BMS09: Estimation of bio- mass and population size by the area-swept method. NWAFC internal doc, NMFS Northwest Fish. Sci. Cent, Seattle 98112, 23 p. Gunderson, D.G., D.A. Armstrong, Y. Shi, & R.A. McConnaughey 1990 Patterns of estuarine use by juvenile English sole (Parophrys vetulus) and Dungeness crab (Cancer magister). Estuaries 13:59-71. Halliday, R.G., & P.A. Koeller 1981 A history of Canadian groundfish trawling sur- veys and data usage in ICNAF divisions 4TVWX. In Doubleday, W.G., & D. Rivard (eds.), Bottom trawl surveys, p. 27-41. Can. Spec. Publ. Fish. Aquat. Sci. 58. Hansen, M.H., W.G. Madow, & B.J. Tepping 1983 An evaluation of model-dependent and probabil- ity-sampling inferences in sample surveys. J. Am. Stat. Assoc. 78:776-793. Harding, G.C., J.D. Pringle, W.P. Vass, S. Pearre Jr., & S.J. Smith 1987 Vertical distribution and daily movements of lar- val lobsters Homarus americanus over Browns Bank, Nova Scotia. Mar. Ecol. Prog. Ser. 41:29-41. Jahn, A.E. 1987 Precision of estimates of abundance of coastal fish larvae. //; Hoyt, R.D. (ed.), 10th annual larval fish conference, p. 30-38. Am. Fish. Soc. Symp. 2. Koch, G.S. Jr., & R.F. Link 1970 Statistical analysis of geological data. John Wiley, NY, 375 p. Land, C.E. 1971 Confidence intervals for linear functions of the normal mean and variance. Ann. Math. Stat. 42:1187-1205. 1975 Tables of confidence limits for linear functions of the normal mean and variance. Sel. Tables Math. Stat. 3:385-419. Link, R.F., G.S. Koch Jr., & J.H. Schuenemeyer 1971 Statistical analysis of gold assay and other trace- element data. U.S. Bur. Mines Rep. Invest. 7495, 127 p. Little, R.A. 1983 Comment on: An evaluation of model-dependent and probability sampling inferences in sample sur- veys. J. Am. Stat. Assoc. 78:797-799. Ludwig, D., & C.J. Walters 1981 Measurement errors and uncertainty in param- eter estimates for stock and recruitment. Can. J. Fish. Aquat. Sci. 38:711-720. McConnaughey, R.A. 1991 Factors affecting Dungeness crab (Cancer magister) year class strength along the Washington coast. Ph.D. diss., School Fish., Univ. Wash., Seattle, 162 p. Myers, R.A., & P. Pepin 1990 The robustness of log-normal estimators of abundance. Biometrics 46:1185-1192. Otto, R.S. 1986 Management and assessment of eastern Bering Sea king crab stocks. In Jamieson, G.S., & N. Bourne 1 18 Fishery Bulletin 91(1). 1993 (eds.), North Pacific workshop on stock assessment and management of invertebrates, p. 83-106. Can. Spec. Publ. Fish. Aquat. Sci. 92. Pennington, M. 1983 Efficient estimators of abundance for fish and plankton surveys. Biometrics 39:281-286. 1986 Some statistical techniques for estimating abun- dance indices from trawl surveys. Fish. Bull., U.S. 84:519-525. Picquelle, S., & G. Stauffer 1985 Parameter estimation for an egg production method of northern anchovy biomass estimation. In Lasker, R. (ed.), An egg production method for esti- mating spawning biomass of pelagic fish: Application to the northern anchovy, Engraulis mordax, p. 7— 15. NOAA Tech. Rep. NMFS 36. Sichel, H.S. 1952 New methods in the statistical evaluation of mine sampling data. Trans. Inst. Mining Metal. (Lond.) 61:261-288. Sissenwine, M.P. 1978 Using the USA research vessel spring bottom trawl survey as an index of Atlantic mackeral abundance. Int. Comm. Northwest Atl. Fish. Sel. Pap. 3:49-55. Smith, S.J. 1981 A comparison of estimators of location for skewed populations, with applications to groundfish trawl surveys. In Doubleday, W.G., & D. Rivard (eds.), Bot- tom trawl surveys, p. 154-163. Can. Spec. Publ. Fish. Aquat. Sci. 58. 1988 Evaluating the efficiency of the A-distribution mean estimator. Biometrics 44:485^193. 1990 Use of statistical models for the estimation of abundance from groundfish trawl survey data. Can. J. Fish. Aquat. Sci. 47:894-903. Smith, S.J., & G.D. Somerton 1981 STRAP: A user-oriented computer analysis sys- tem for groundfish research trawl survey data. Can. Tech. Rep. Fish. Aquat. Sci. 1030, 66 p. Zweifel, J.R., & P.E. Smith 1981 Estimation of abundance and mortality of larval anchovies (1951-75): Application of a new method. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:248-259. Appendix A random variable X is considered to be lognormally distributed when the natural logarithm of X, Y=ln(X), has a normal distribution. Specifically, if Y is normally distributed with mean u^ and standard deviation ct ln , then X=e Y is lognormally distributed with density function (Aitchison & Brown 1969): 1 The kth moment about zero, E(X k ) is expressed as exp f-(lnZ-u LN ) 2 2o 2 , (AD £(X k ) = E (e kY ) = exp ^u^ + 2 J 25 roti- fers/mL). For each experimental temperature, 10 lar- vae were sampled 10 d past first-feeding. Length (SL), dry weight, and mouth width (an estimate of gape size) were obtained from preserved larvae. Daily in- stantaneous rates of increase in length and weight were calculated according to Ricker ( 1975:207). Length-weight relationships were used to estimate sizes separating early-life-history periods (Balon 1984). Larvae were measured live, and dry weights were ob- tained from larvae that were preserved by freezing. For each species, an iterative process was used to de- termine the break point (SL) between two regressions that described the length/weight relationship. Itera- tions were done with break point values of 8-14 mmSL at 1 mm intervals. The two regressions that resulted in the minimal mean square error were chosen. An ANCOVA was used to compare the biphasic regressions. Statistical testing was done by using the General Linear Model procedure (SAS 1985). Unless otherwise noted, ANOVA was used to test differences. Each treat- ment consisted of a group of larvae in a common tank. The mean measurement (e.g., mean length or mean weight of the group) was used in the statistical analy- sis. Each experiental replicate was treated as an inde- pendent observation. Differences between main effects were considered significant at a=0.05. For interaction between factors, differences were considered signifi- cant at a=0.10 to reduce the chance of a type-II error. Results Atlantic menhaden eggs were larger in diameter and nearly two-fold heavier than gulf menhaden eggs (Table 1, 2). The volume of yolk was greater for Atlan- tic menhaden, but oil globule volumes were similar (Table 1,2). There was no discernible change in egg size during the laboratory spawning season (Fig. 1), and the level of variability throughout the season was smaller for gulf (SD=0.02-0.04) than Atlantic (SD=0.04-08) men- haden. There was a trend for field-collected Atlantic Table 1 Summary of analysis of variance results ex pressed as P-values (Pr>F) for gulf and Atlantic menhaden. S = species, T = temperatures, SxT = species x temperature interactions, F = food, SxF = species x food interaction, TxF = tempera- ture x food interaction, and SxFxT = species x food x temperature interaction. Factor Class S T S i 1 > 1 a a LU 1 40 37-38 38-39 39-40 40-41 4 1-42 42-43 LATITUDE (degrees) Figure 2 Relationship between egg size (.f±SD) and lati- tude of capture for Brevoortia tyrannus. Num- bers in parentheses indicate number of eggs measured. at lower incubation temperatures. On the other hand, incubation temperatures in the laboratory did not af- fect the weight of larvae at hatching (Table 1, 3). Yolk reserves were larger for Atlantic menhaden at hatching, but oil reserves were greater for gulf men- haden (Table 1, 3). Temperature did not affect the amount of oil or yolk reserves for recently-hatched At- lantic or gulf menhadens. After hatching, both species utilized yolk and oil reserves at an exponential rate (Table 4). Atlantic menhaden utilized yolk reserves at a higher rate than gulf menhaden, but rates of oil depletion did not differ (Table 1, 4). Temperature af- fected rates of yolk and oil utilization. At first-feeding, the oil globule and yolk were virtually depleted (Table 3). Yolk reserves at this time were similar for the two species and were independent of rearing tem- perature. The age at which anatomical features asso- ciated with first-feeding appeared (eye pigment, fore- and midgut, and functional mouth) did not differ, but the interval between the appearance of these features and first-feeding was shorter for Atlantic menhaden (Table 5). At first-feeding, Atlantic menhaden were significantly longer and younger than, but of similar weight to, gulf menhaden (Table 1, 3). Although length-at-hatching in- creased with decreasing temperature, temperature did not influ- ence length of larvae at first-feeding. Age-at-first-feeding decreased with increasing temperatures for both species (Table 1, 3). First-feeding Atlantic menhaden, which are longer than first- feeding gulf menhaden, may be slightly more resistant to starva- tion (Fig. 3). Atlantic menhaden lived 1-2 d longer than gulf men- haden. The survivorship curves of the two species were similarly shaped at the same temperature. At the highest temperature (24°C) survivorship declined rapidly after the third day, while at the lowest temperature ( 16° C) it was nearly linear with time. The amount of biomass gained during the early stages did not differ between species (Table 1, Fig. 4). Population biomass of both menhadens was relatively low at 16° C, as compared with 20° C and 24° C when food was not limiting. Food concentrations of <1.0 rotifer/mL similarly limited the growth and survival of both species. Low temperature (i.e., 16° C) affected biomass gained by larvae of both species during this 7d feeding and growth study. Interactions between species and temperature were observed for all growth experiments (Table 1, Fig. 4). Atlantic menhaden growth was lower at 16° C than at 20° and 24° C. Gulf menhaden growth did not appear to differ between temperatures. Atlantic menhaden exhibited higher growth rates at 20° and 24° C, and, in like manner, were larger than gulf menhaden 10 d past first- feeding; however, mouth gape at this time did not differ between species (Table 1, 3). Atlantic menhaden may attain its size-threshold for metamor- phosis (i.e., Balon's 1984 concept) earlier than gulf menhaden (Figs. 5, 6). Length-weight relationships expressed during early- Powell: Early-life-history traits of Brevoortia tyrannus and B. patronus 123 Table 3 Summary of data U±SD) for larval Atlantic and gulf me nhaden. Values for specific temperatures are given only when temperature had a statis ically-significant la=0.05) influence on the trait. An asterisk (*) preceding the trait indicates a statistically-significant (a=0.05) difference between species. Values in parentheses indicate the number of replicates. Means ire calculated from replicate means. Temperature Atlantic Gulf Trait (°C) menhaden menhaden *Size at hatching (mmSL) 16 3.4 ±0.2 (4) 3.1 ±0.2(4) 20 3.2 ±0.2 (4i 2.6 ±0.1 (4) 24 3.2 ±0.1 (3i 2.8 ±0.2 (4) *Dry weight at hatching (ug) — 49.8 ±5.0 (11) 39.3 ±5.9 (12) *Yolk volume at hatching (mm 3 ) — 0.1512 ±0.0220 (15) 0.1205 ±0.0259 (11) *Oil volume at hatching (mm 1 ) — 0.0023 ±0.0005 (15) 0.0029 ±0.0004 (11) Yolk volume at first-feeding (mm 1 ) — 0.0002 ±0.0000 (15) 0.0002 ±0.0000(11) Oil volume at first-feeding (mm 1 ) — 0.00(15) 0.00(11) *Age at first-feeding (d) 16 5.2 ±0.4 (8) 5.7 ±0.1 (6) 20 3.0 ±0.3 (8) 3.9 ±0.1 (7) 24 2.4 ±0.3 (8) 2.9 ±0.1 (7) *Size at first-feeding (mmSL) — 4.8 ±0.2 (9) 4.3 ±0.3 (10) Dry weight at first-feeding (ug) — 37.8 ±2.6 (9) 35.1 ±1.6(10) Growth rate (10di in length (In mm/d)** 16 0.027 ±0.010 (3) 0.038 ±0.013 (3) 20 0.047 ±0.000 (3) 0.037 ±0.005 (3) 24 0.049 ±0.006 (3) 0.042 ±0.004 (3) *Growth rate ( lOd) in weight (In mg/d)** 16 0.051 ±0.029(3) 0.061 ±0.029(31 20 0.103 ±0.016(3) 0.062+0.013(3) 24 0.102 ±0.013 (3) 0.062 ±0.005 (3i "Length 10 d past first-feeding (mmSL)** 16 6.6 ±0.7 (3) 6.5 ±0.8(3) 20 7.9 ±0.1 (3) 6.2 ±0.3(3) 24 7.7 ±0.0 (3) 6.4 ±0.3(3) *Dry Weight lOd past first-feeding (ug)** 16 62.0 ±19.2 (3) 60.3 ±15.8 (3) 20 115.3 ± 19.6 (3) 63.3 ±8.1 (3) 24 100.7 ±13.4 (3) 71.0 ±3.5 (3) Mouth gape 10 d past first-feeding (mm) 16 0.3 ±0.02 (3) 0.3 ±0.01 (3) 20 0.3 ±0.01 (3) 0.3 ±0.01 (3) 24 3 observed. 0.3 ±0.02 (3) 0.3 ±0.03 (3) **A species X temperature interaction wa Table 4 Linear regression equations (Y=B„ +B,X,) between log e yolk and oil volume in mm 3 (Y) and age in days (X,) for gulf and Atlantic menhaden. N = number of replicates. Species °C N Slope (B,) Intercept (B ) r 2 Yolk Gulf menhaden 16 4 -1.220 -2.089 0.96 20 4 -1.734 -1.733 0.97 24 4 -2.228 -2.118 0.98 Atlantic menhaden 16 5 -1.270 -1.765 0.97 20 5 -2.065 -1.712 0.96 24 5 -2.595 -1.723 0.97 Oil Gulf menhaden 16 4 -1.432 -5.082 0.81 20 4 -1.819 -4.985 0.86 24 4 -2.834 -4.814 0.81 Atlantic menhaden 16 5 -1.222 -5.392 0.92 20 5 -2.041 -5.232 0.83 24 5 -2.017 -5.714 0.84 124 Fishery Bulletin 91(1), 1993 Table 5 The mean time (d) before first-feeding when 580% of gulf and Atlantic menhaden arvae attained eye pigment 10mmSL and gulf menhaden >12mmSL, there were no differ- ences between slopes or intercepts. For Atlantic menhaden <10 mmSL and gulf menhaden <12 mmSL, there were no differences in slopes between the two species, although their intercepts differed. Beginning at the size- threshold in both species (>10mmSL for Atlantic menhaden, >12mmSL for gulf menhaden), there is a larger change in weight in relation to a given change in length, reflecting morphogenic changes. 100 80 60 40 20 100 80 60 40 20 o / ' 1 "\ 16° C 1 2 3 4 5 6 7 8 20° C x 01 2345678 lOOf--,. 80- 60- 40 20 _i_ 1 2 3 4 5 6 7 8 DAYS PAST FIRST FEEDING Figure 3 Time to starvation of unfed Brevoortia tyrannus !•) and B. patronus ( ) in relation to tem- perature. Approximately 25 larvae were used at each experimental temperature. N denotes number of replicates. Discussion Egg size and its influence on early-life-history characteristics have received considerable attention. My results, in general, are in accord with other studies of diverse fishes and other organisms. Larger eggs are positively correlated with size of larvae at hatching, yolk reserves, resistance to star- vation, and size-at-first-feeding, e.g., Blaxter & Hempel 1963 (clupeoid stocks), Crump 1984 (amphibia), Wallace & Aasjord 1984 (salmonids), Knutsen & ATLANTIC MENHADEN r\ GULF MENHADEN 2500 * \ A £ 2000 a Ik / ' ^v \ ? 1500 a a. + * 10 1000 / / > < s$ / 9 500 m m m J ... - 4 ^ *rrS ' ^t- - -2.0 -1.0 0.0 1.0 2.0 -2 -1.0 0.0 1.0 2.0 LOO NUMBER OF ROTIFERS / ML Figure 4 Amount of biomass gained (mean dry weight gained/larvae X number of survivors) over a 7d period for early-larval Brevoortia tyrannus and B. patronus at 16 (■). 20 • and 24°C ( ). Values are means of two replicates for food levels of 5, 25. and 50 rotifers/mL and means of three replicates for all other food levels. Powell: Early-life-history traits of Brevoortia tyrannus and B patronus 125 Log, Dry Wt.= -2.398+ 4.775(log, SL) 4.0 r 2 =0.98 / 3 3.0 j X G 111 5 f I 2.0 f\ o 6" o _j / ! 1.0 Log )0 Dry Wt.= -0.714 + 3.085(log, SL) r2 = 0.95 0.2 0.4 0.6 0.8 1.0 12 1.4 L0G, STANDARD LENGTH (mm) Figure 5 Length-weight relationship of larval Brevoortia tyrannus showing a change in body morphology. Breakpoint occurs at lOmmSL. Log, Dry Wt.= -2.787 + 5.1 43 M 1 9 1 v 15 \\4\ 7 V FL ATLANTIC ^ 28 6 / ^20 «21 ^ hs \ 27 4 \ 26 Jf GULF OF MEXICO 3 yv 25 1 '2 i 24 Figure 1 NMFS Statistical Areas in the Gulf of Mexico and Atlantic. Vessels were recruited with the assistance of NMFS port agents, NOAA Sea Grant Marine Advisory agents, regional shrimp associations, and industry contacts. All participating vessels received appropriate federal authorization to use TEDs in only half the trawls when a NMFS observer was on board. Twenty-six quad-rigged vessels (two trawls towed/side) and one twin-rigged vessel (one trawl towed/side) were used in the study. Areas Beginning in March 1988, observers were placed on shrimp vessels in each of the four major Gulf of Mexico offshore fishing areas (Louisiana, Texas, south Florida, and Alabama-Mississippi) and in the Atlantic off Florida, Georgia, and North Carolina. Higher levels of observer effort were allocated for areas which histori- cally had higher shrimp production. Of 600 planned observer days, 240 were scheduled for Louisiana, 200 for Texas, 50 each for east and west Florida, and 60 for Mississippi-Alabama. One-hundred observer days were also scheduled for Georgia and North Carolina waters. Observer days were targeted for peak regional shrimping seasons in each area, although this sched- ule was not always implemented due to constraints of voluntary participation by the shrimp industry. The U.S. coasts of the Gulf of Mexico and Atlantic Ocean are divided into Statistical Areas (Fig. 1) by NMFS for analytical purposes. Areal groupings for analyses in this study were Statistical Areas 1-8 (West Florida), 9-12 (Florida Pan- handle, Alabama, and Missis- sippi), 13-17 (Louisiana), 18-21 (Texas), 28 (Cape Canaveral), 30- 31 (East Florida and Georgia), and 34-35 (North Carolina). The study depended on shrimpers volunteering to allow NMFS personnel to collect data onboard their vessels. Due to lim- ited response by shrimpers, data came from virtually any vessel whose owner or captain would al- low NMFS aboard. Since one of the principal objectives of this study was to evaluate the effect of the use of TEDs on commer- cial shrimping, the shrimpers de- cided where and when to fish and which certified TED to use. Our only stipulations were that the shrimper had to use federally ap- proved TEDs, allow gear special- ists to properly adjust the TEDs, and keep catches from all nets of Renaud et al.: Shrimp loss by TEDs in U.S. coastal waters 131 a tow separate to facilitate data collection on deck. The conditions under which the data were collected were assumed to be representative of commercial fishing conditions. Gear tuning and control tows The fishing efficiency of all nets used in this study was standardized by NMFS or Sea Grant gear specialists during the initial trip of a participating vessel. Prior to installation of TEDs, control tows were made using standard nets. Lazy line, tickler chain, and float ad- justments were made to each net until approximately equal amounts of shrimp were caught by every net. A Georgia TED FRONT Super Shooter TED FRONT SIDE Vessel captains were instructed by gear specialists on the proper installation of TEDs. Once TEDs were installed, the gear specialist modified the rigging for the proper operation of the TED. This procedure usu- ally required 2-3 d. The captain then was responsible for later gear tuning. Differences in the tuning ability of captains may contribute to variations in the catch data. All Super Shooter TEDs were constructed with accelerator funnels (Fig. 2), i.e., mesh in the shape of a funnel sewn into the net directly in front of the TED. Funnels accelerate water flow through the TED and into the cod end of the net. Georgia TEDs were tested with and without funnels. Data collection C. TED with accelerator funnel installed in shrimp trawl TED GRID ACCELERATOR FUNNEL Figure 2 Schematics of Georgia and Super Shooter TEDs and accelerator funnel. Every phase of the operation was ex- plained to vessel captains by NMFS personnel to insure that all data could be collected. Aside from sampling the catch and working up the data, ob- servers did not interfere with normal fishing activity. The primary require- ment of the study was that catches from each net be kept separate from all others so the shrimp from each trawl could be weighed and recorded. If necessary, the back deck of the ves- sel was partitioned with wooden beams to prevent catches from mixing. Cap- tains of the vessels were requested to examine the data collected by the NMFS observer and to sign the data sheets to verify their accuracy. Shrimp catch on observer vessel A random sample weighing 50-70 lb was shoveled from the contents of each trawl into standard-sized plastic shrimp baskets. Thus, a quad-rigged vessel produced four samples per tow and a twin-rigged vessel two samples per tow. Shrimp were separated from each sample and total weight (to the nearest lb) of brown, pink, and white shrimp (Penaeus sp. ) combined was re- corded for every net of each tow. No analysis by species was possible or pro- posed by this study. If the shrimper discarded small shrimp, observers were instructed to include only the size-range of shrimp retained by the shrimpers for their weights. Catch was recorded as heads-on or heads-off. Heads-off weight = (0.63 heads-on weight). 132 Fishery Bulletin 91 [1). 1993 For each tow, shrimp CPUE (heads-off lb/h/100 ft of headrope towed) from all TED-equipped nets were av- eraged and compared against the average shrimp CPUE of all standard nets, to provide one TED- standard data pair per tow. Unless otherwise stated, shrimp CPUE will refer to heads-off lb/h/100 ft of headrope. The average CPUEs of two TED-equipped and two standard nets were paired for each tow for 26 quad-rigged vessels and 1 twin-rigged vessel. However, if one net was excluded from the analysis due to unac- ceptable operation (refer to Gear Performance), then the CPUE value from the remaining net was paired with the average of CPUEs from the other two nets. If both nets of a given gear type malfunctioned, all data from that tow were deleted from the analysis. Stan- dard and experimental nets were compared on twin- rigged vessels and these data pooled with those from quad-rigged vessels. Biological models Deterministic population models were produced for brown shrimp Penaeus aztecus, white shrimp P. setiferus, and pink shrimp P. duorarum by linking a Ricker-type yield-per-recruit model to recruit- ment estimates that were independent of parent stock (Ricker 1975, Nichols 1984, Nance & Nichols 1988). Recruitment level was set at the geometric mean for the complete data set (1960-88). Estimates for 1986- 89 fishing mortality rates (F) were derived from vir- tual population analysis, and the average was used as the baseline for current conditions. Yield estimates were made for all three species for a range of "F-multiplier" values of 0-2 by 0.02 increments. Tables of these yield estimates were used to determine effects of TED- equipped nets on the shrimp yield in the Gulf of Mexico. This was possible because yield estimates (Y t ) are a direct result of fishing mortality rates (Royce 1972). The yield model was Commercial shrimp catch Effort data for a given tem- poral and spatial area were calculated by taking the average trip CPUEs (heads-off lbs/24 h day/4 nets), ob- tained by interviewing vessel captains, and extrapo- lating to total effort by using the total-pounds value from dealers' records. Fishing-effort data on the shrimp fleet have been collected in this manner since 1960. These data were compared with CPUEs (heads-off lbs/ 24 h day/4 nets) from our observer trips. The assump- tion that shrimp CPUEs were equal, both for vessels from this study and from the commercial fleet fishing during the same seasons and in the same Statistical Areas, was tested using a paired £-test with a prob- ability level of 0.05. Y, = F t N t W, dt , where N, is the number of animals (R) in a cohort subject to fishing (F) and natural (M) mortality at a given time (t), using the formula N, = Re-' F+M,lt -V' F, = fishing mortality at a given time, W t = average weight of an individual at time t, estimated from growth equations. Fishing mortality rate (F) is the product of two sepa- rate variables, a catchability coefficient (q) and directed nominal fishing effort (f): Gear performance Each net was characterized by an operation code based on its performance in the water. Codes were used to describe successful tows or prob- lems encountered, such as tangling of trawl doors, gear fouling, twisted cables, bag choking, etc. Two codes were occasionally required to describe trawl performance. Data collected from the problematic tows not related to TEDs, e.g., cod end coming untied, gear not fishing properly, torn nets, and broken cables, were not in- cluded in the analyses. Chi-square (P<0.05) analysis was used to determine if the problematic tows were independent of net type (e.g., TED-equipped nets or standard nets) by area (Gulf of Mexico or Atlantic). F = qf. TED-equipped nets influence fishing mortality (F) by affecting shrimp catchability (q), and not fishing effort (f). Any percentage change in shrimp catchability caused by TED-equipped nets was assumed to be di- rectly reflected in an equal percentage change in fishing mortality. This is based on an assumption of direct proportionality between change in CPUE and change in q. Thus, any change in CPUE as a result of TED use is translated into a proportional change in q. Results Statistical analyses Paired t-tests Paired <-tests were performed to test the hypothesis of equal CPUE of shrimp by standard and TED-equipped trawls. Data were paired by tow. Confi- dence intervals (95% ) on CPUE were also calculated. Descriptive data summary Paired data In the Gulf of Mexico, 589 data pairs were collected using Georgia TEDs equipped with ac- celerator funnels, 59 pairs from Georgia TEDs without funnels, and 50 pairs from Super Shooter TEDs with Renaud et al.: Shrimp loss by TEDs in US coastal waters 133 Frequency of Georgia TED paired tows for stc s with funnel (GF) Table 1 ndard nets and nets equipped with Super Shooter TEDs with funnel (SF), and Georgia TEDs without funnels (G) by season and area. Areas* Winter Spring Summer Fall SF GF G SF GF G SF GF G SF GF G WFL 2 17 _ 15 79 10 _ - - - (1-8) MAFP 28 11 3 20 39 - (9-121 LA 60 22 55 25 21 104 (13-17) TX 3 5 1 - 88 - 67 23 (18-211 CCFL 60 _ _ _ _ _ _ _ _ (28) EFLG 30 _ _ 21 163 35 - (30-31) NC _ _ _ 186 _ _ _ (34-35) Totals 2 138 Florida Canave 65 ). 9- ral), 48 138 10 186 154 184 245 23 12 (Florida Panhandle. Alabama, Mississippi), 13-17 (Louisiana), 18-21 30-31 (East Florida and Georgia), and 34-35 (North Carolina). * Areas 1-8 (West (Texas), 28 (Cape funnels. There were 86 and 223 data pairs in the At- lantic for Georgia TEDs with and without accelerator funnels, respectively, and 186 pairs for Super Shooter TEDs with funnels. Frequencies of data collection by geographic area and season (winter: December-Feb- ruary, spring: March-May, summer: June-August, fall: September-November) are presented in Table 1. Performance of TED-equipped and standard nets Data were collected from 5937 nets during the 2.5 yr study. Frequency of net problems was tabulated by TED type. The most frequent problems included clogging of the net, twisting of trawl doors and cables, and torn web- bing. In the Gulf of Mexico, no problems occurred dur- ing 86%, 87%, 75%, and 87% of the tows for nets equipped with Georgia TEDs with and without fun- nels, Super Shooter TEDs and standard nets, respec- tively (Table 2). In the Atlantic, the values were 96%, 90%, 89%, and 95% for the respective gear types (Table 2). A variety of problems, including but not lim- ited to those with trawl doors, cables, bogging-down of nets, etc., were shown to be net-type independent (e.g., TED-equipped nets or standard nets) in the Gulf of Mexico and net-type dependent in the Atlantic (chi- square, P<0.05). Testing of paired tows Reduction of shrimp CPUE associated with use of TEDs Mixtures of brown and white shrimp were cap- tured in all areas of the Gulf and Atlantic, except for Table 2 Comparison of net types and gear-related problems n the Gulf of Mexico and Atlantic for Georgia TEDs with (GF) and without (G) funnels. Super Shooter TED with funnel (SF), and standard shrimj > nets (STD). Sample includes all nets used from all vessels during the study. Values represent the percent of nets in each category; totals may not equal 100% due to rounding. STD G GF SF (n=2356) (n=199) 0.05) from CPUEs on commercial vessels without observers. Mean differences ranged from a 6.2 lb/h gain by standard nets on TED observer vessels to a 4.9 lb/h gain by standard nets on other commercial vessels. In three of seven season/area combinations, shrimp CPUE from TED-observer vessels was higher than CPUEs of other commercial vessels. Since there were no significant differences in net size during our study, we assumed that this was the case for the rest of the commercial fleet. TED-observer vessels were apparently represen- tative of other commercial vessels in the fleet fishing in similar areas during the same season. Similar analy- ses for the Atlantic fishery could not be made since catch information was not available on a trip-by-trip basis. Biological yield models Ricker-type yield models (Ricker 1975) developed for each of the three major shrimp species show the same basic curve shape (Fig. 3; Nance & Nichols 1988). The curves are asymptotic where yield estimates are plot- ted for current fishing mortality rates (F-multiplier = 1.0). Thus, with current fishing patterns and current fishing mortality rates, little increase or decrease in yield is predicted with the minor reductions in F that MILLIONS OF POUNDS j O O O O BROWN SHRIMP / WWTE SHRIMP U PINK SHRIMP 0.0 0.5 1.0 1.5 2.0 F-MULTIPLIER Figure 3 Yield models for brown Penaeus aztecus, white P. setiferus, and pink P. duorarum shrimp. would be expected due to small losses of shrimp by TEDs. Yield estimates were calculated in the model by vary- ing the F-multiplier in increments of 0.02. Mean shrimp loss with TED-equipped vs. standard nets varied from 1 to 14% by TED type. A decrease of 5% in F would result in an undetectable change in annual yield in the brown or white shrimp fisheries and a 1% reduc- tion in the annual yield of the pink shrimp fishery in the U.S. Gulf of Mexico. Discussion Our data were collected by NMFS observers during cooperative cruises with shrimp industry participants. Since this was a voluntary program, TED type, area, and season of sampling were controlled by industry participants. Data came from virtually any vessel whose owner or captain would allow NMFS observers aboard. Not all federally approved TED types were tested. If a shrimper could not maintain TED efficiency during a trip, the trip was aborted by the shrimper or the TED was not used again. This resulted in nominal imbalances in the data by area, season, and TED type, including some data sets too small for analysis. Mean shrimp catch rates in TED-equipped nets were lower than those in standard nets, varying from a loss of 1.4% with Super Shooter TEDs to a loss of 13.6% for Georgia TEDs without funnels. Nets equipped with Georgia TEDs without a funnel were used mainly dur- 136 Fishery Bulletin 9 1 [ I ). 1993 ing the first 6 months of this study. Higher losses of shrimp from these nets may be due to (1) initial inex- perience by shrimpers using TEDs, (2) high losses of shrimp in rough-bottom areas, and (3) absence of a funnel in the TED. The lack of an accelerator funnel to assist shrimp movement past the escape opening of the TED could also account for some shrimp loss. The Georgia TED with an accelerator funnel exhibited a 3.6% reduction in shrimp CPUE compared with 13.6% by the Georgia TED without a funnel. Nets equipped with the Super Shooter TED exhibited the lowest re- duction (1.4%) in shrimp CPUE when compared with the standard nets. This may have been due to (1) shrimpers having more experience with TEDs when this model was introduced during the second year of the study, and (2) more effective shrimp retention by the TED. The Super Shooter design also reduces clog- ging of TED bars by seagrasses and algae and may reduce shrimp loss. Although this TED exhibited the lowest reduction in shrimp CPUEs, it accounted for more problems during trawling than the other TEDs. These problems evidently did not affect shrimp catchability, since there was no significant difference between its catch rate and that of the paired standard net. Areal differences in shrimp abundance may be con- founded with CPUEs due to different types of TEDS and standard nets (flat nets, semiballoon nets, mon- goose nets, etc.). Some TEDs work better on hard- bottom than on soft-bottom or with different types and abundances of bycatch. Georgia TEDs with funnels were the most common TED tested in Texas, Louisi- ana, and Florida. Super Shooter TEDs with funnels were used in North Carolina. The effectiveness of the TED type does influence the catch rates of shrimp. Phares ( 1978), in describing the selectivity of shrimp nets, indicated that loss rates varied by area and sea- son and affected an extensive size-range of lost shrimp. We have assumed ( 1 ) that shrimp escaping through either a TED-equipped net or a standard net will not die because of that episode, and (2) that escaping shrimp will grow and experience the same subsequent natural and fishing mortality as the rest of the stock. Thus, survival rates of shrimp escaping through the cod end of a standard net should be the same as those of shrimp escaping through the cod end of a TED net. Shrimp escaping through TED openings probably are not injured and are subject to subsequent recapture. Although decreases in CPUE may impact a given fisherman on any particular tow, these lost shrimp will still be available to fishermen for capture by suc- ceeding tows. Mathematical models indicated that a TED-induced decrease of 5% in F would result in an undetectable change in yield in the brown or white shrimp fisheries and a 1% reduction in the annual yield of the pink shrimp fishery in the U.S. Gulf of Mexico. Because of the asymptotic nature of the yield curves, only slight decreases in yield would be observed in some shrimp fisheries even if loss rates from TEDs were in the 10-20% range. With a 10% loss rate, we calculated a reduction from the pink shrimp fishery of 2% and no decreases in yield from either the white or brown shrimp fisheries. A 20% loss rate would result in a 4% reduction of the annual yield of pink shrimp and a 1-2% reduction for brown and white shrimp fisheries. Acknowledgments We would like to acknowledge several organizations and their personnel for assistance in securing vessels to participate in this study: Gary Graham and Hollis Forrester, Texas A&M University Sea Grant Marine Extension Service; David Harrington and Paul Chris- tian, University of Georgia Sea Grant Marine Exten- sion Service; Bill Hogarth, North Carolina Fish and Wildlife; numerous NMFS Port Agents; Texas and Loui- siana Shrimp Associations, and the Gulf and South Atlantic Fisheries Development Foundation. Wil Seidel, John Watson, Windy Taylor, Dale Stevens, and James Barber with NMFS Pascagoula assisted in vessel re- cruitment, TED construction, gear tuning and back- ground information on the development of the TED, its installation and proper use. Jo Williams and Frank Patella, NMFS Galveston, prepared figures for the manuscript and assisted in various statistical analy- ses. Much credit also goes to the NMFS Galveston observers who painstakingly collected the data for this project. Finally, we would like to thank the shrimpers who participated in the study. Without their coopera- tion, the study could not have been conducted. Citations Federal Register 1987 52(1241:24244-24262. Klima, E.F., G.A. Matthews, & F.J. Patella 1986 Synopsis of the Tortugas pink shrimp fishery, 1960-1983, and the impact of the Tortugas Sanc- tuary. N. Am J. Fish. Manage. 6:301-310. Magnuson, J.J., K.A. Bjorndal, W.D. DuPaul, G.L. Gra- ham, D.W. Owens, C.H. Peterson, P.C.H. Pritchard, J.I. Richardson, G.E. Saul, & C.W. West 1990 Decline of the sea turtles: Causes and pre- vention. Natl. Res. Counc. Natl. Acad. Sci. Press, Wash. DC, 190 p. Nance, J.M., & S. Nichols 1988 Stock assessment for brown, white and pink shrimp in the U.S. Gulf of Mexico, 1960-1986. NOAA Renaud et al.: Shrimp loss by TEDs in US coastal waters 137 Tech. Memo. NMFS-SEFC-203, Galveston Lab., NMFS Southeast Fish. Sci. Cent., 64 p. Nichols, S. 1984 Updated assessments of brown, white and pink shrimp in the U.S. Gulf of Mexico. Pascagoula Lab., NMFS Southeast Fish. Sci. Cent, Pascagoula, 53 p. Phares, P.L. 1978 An analysis of some shrimp trawl mesh size se- lection data. Miami Lab., NMFS Southeast Fish. Sci. Cent., Miami, 37 p. Flicker, W.E. 1975 Computation and interpretation of biological sta- tistics offish populations. Bull. Fish. Res. Board Can. 191:1-382. Royce, W.F. 1972 Introduction to the fishery sciences. Academic Press, NY, 351 p. AbStraCt.-Food habits data from 415 sandbar sharks collected in the area between Cape Hatteras and Georges Bank (Great South Chan- nel) were examined. Mean fork length (FL) and body weight (BW) were 55.0 cm and 1.72 kg for pups. 123.0cm and 23.0kg for juveniles, and 166.0 cm and 52.3 kg for adults. Of all juvenile and adult stomachs, 49% contained prey, primarily fish (teleosts and skates). Of stomachs from pups, 80% held food remains consisting almost exclusively of soft blue crabs. The mean percentage of stomach content volume to BW is 1.16 for pups, and 0.42 for juveniles and adults. Daily ration estimates as percentage of mean BW are 1.43 for pups, and 0.86 for juveniles and adults. Annual food consumption is estimated to be 5.1 times the mean BW for pups, and 3.1 times for juve- niles and adults. Food habits of the sandbar shark Carcharhinus plumbeus off the U.S. northeast coast with estimates of daily ration* Charles E. Stillwell Nancy E. Kohler Narragansett Laboratory. Northeast Fisheries Science Center National Marine Fisheries Service, NOAA Narragansett. Rhode Island 02882 The sandbar shark Carcharhinus plumbeus is a medium-sized species found in temperate and subtropical waters of the world's oceans and the Mediterranean Sea. It occurs from nearshore out to a depth of at least 250m (Springer 1960, Garrick 1982). Evidence of its occurrence over deep water is provided by Springer (1960) who reports the capture of three specimens taken in midwater over depths of 1000-1800 m. Distribution of the sandbar shark along the U.S. east coast extends from Massachu- setts to the Florida Keys in the sum- mer and from the offings of the Caro- linas to Cape Canaveral during the winter months (Bigelow & Schroeder 1948, Springer 1960). From May through September, newborn pups and small juveniles (<100cm fork length, FL) are common to abundant in shallow bays and estuarine sys- tems along the coast from Long Is- land, New York to Cape Canaveral, Florida. With the approach of autumn, young sharks migrate offshore and south to winter at depths approaching 137m (Springer 1960, Medved & Marshall 1983). Casey ( 1976) and Casey et al. ( 1985) showed that when the juvenile sand- bar sharks attain a size of about HOcmFL, they no longer frequent Manuscript accepted 19 August 1992. Fishery Bulletin, U.S. 91:138-150(1993) 'MARMAP Contribution FED/NEFC 86-05. the shallow nursery areas but remain off the coast, demonstrating more ex- tensive seasonal migrations with in- creasing size. The most detailed publications to date on food and feeding in the sand- bar shark come from Springer ( 1960), Medved & Marshall (1983), Medved (1985), and Medved et al. (1985, 1988). These papers are important contributions to our knowledge of the diet, feeding behavior, and daily ra- tion of young sandbar sharks and what impact they have on prey re- sources in the estuaries and near- shore areas. The first study to esti- mate digestion rate in the sandbar shark was conducted by Wass (1973) in a seawater enclosure at the Ke- walo Basin facility in Hawaii. The purpose of this paper is to present data on the food and feeding habits of sandbar sharks occurring from Georges Bank (Great South Channel) to Cape Hatteras, to define dietary differences and energy needs of pups, juveniles, and adults, and to estimate their daily ration. Methods Stomachs were sampled from 1972 through 1984 during ( 1 ) shark fishing tournaments held at several coastal ports from Rhode Island to southern New Jersey, (2) on cruises using 138 Stillwell and Kohler. Food habits of Carcharhmus plumbeus off US. northeast coast 139 42°N 40 38 - 36 - 34 A. US*^ CHINCOTEAGUE BAY' 7 8°W Figure 1 Fishing area off the U.S. northeast coast where 415 sandbar shark Carcharhmus plumbeus pups, juveniles, and adults were caught and examined for food habits studies, 1972-84. The 100 m depth contour separates the nearshore and offshore sampling areas. offshore (>100m) (Fig. 1). The Chincoteague sample was further separated into two distinct age classes: newborn pups (esti- mated <3d-old) and small juve- niles (>3 d-3+ yr). Newborn pups were distinguished by pale, un- pigmented edges on the fins, unhealed or partially-healed um- bilical openings, and the presence of large cream-colored livers that floated slightly above the surface when placed in seawater. "Older" pups and small juveniles had liv- ers that were reduced in size, varied in color from tan to gray- green, and sank slowly or floated just beneath the water surface. In addition, their umbilical open- ings were completely healed, vis- ible only as white streaks 5- 6 mm long. Juveniles and adults of both sexes were separated, based on a minimum reproduc- tive size of 150cmFL (Casey et al. 1985). longline gear aboard research and commercial fishing vessels from Cape Hatteras to Georges Bank, and (3) during a 6d period of fishing at the end of June 1983 with rod-and-reel in Chincoteague Bay, Virginia (Fig. 1). Collections and examinations of all stomachs were made during March to September, with the ma- jority being taken in June and July. Stomachs were excised and the volume of the contents (liquid and solids) measured as soon after capture as possible. Solid remains were drained, sorted, and identified to the lowest taxon possible, then enumerated and measured volumetrically by water displacement in a graduated beaker. A conversion of lmL=lg was used to convert volume to weight for comparisons with shark body weights. Major forage categories were expressed as per- centages by number of particular prey items, as total volume of the prey items, and as frequency of their occurrence (number of stomachs). Maximum capacity was estimated by filling the stomachs with water un- der low pressure, then measuring the volume of water in a graduated container. The maximum capacities of stomachs from Chincoteague Bay sharks were not determined because a pressurized water system was not available. Analysis of the data for differences in prey, food volumes by area, and daily ration was accomplished by separating the samples into three groups: Chincoteague Bay, nearshore (<100m), and Results and discussion Stomachs from 415 sandbar sharks were examined, including 321 from nearshore (268) and offshore (53) waters between Cape Hatteras, North Carolina and Georges Bank, and 94 from Chincoteague Bay, Virginia. Analysis of nearshore and offshore samples In the nearshore area, juvenile males and females and adult females were represented by almost equal num- bers, i.e., 81, 84, and 89, respectively, whereas only 12 adult males were sampled. Offshore, juvenile males were most abundant (37), with adult males represented by four individuals. Females were limited to four adults and eight juveniles. Mean fork length (FL) and body weight (BW) of sandbar sharks for the whole sample were 138cm (range 69.0-212.0) and 34.0kg (3.0-145.0) (Table 1). Offshore, only juvenile males were numer- ous enough in the sample to derive reliable mean values. Prey analysis Prey consumed by sandbar sharks in the study area consists primarily of benthic and demersal species, both vertebrate and invertebrate (Table 2). Of the 40 different prey types observed in 140 Fishery Bulletin 91 1 1993 Table 1 Average fork lengths and body weights for 321 juvenile and adult sandbar sharks Carcharhinus plumbeus examined from nearshore KlOOm and offshore (>100ml waters of the U.S. northeast coast between Cape Hatteras and Georges Bank, 1972-84. Overall mean Adults Juveniles Overall mean by sex Male Female All Male Female All Male Female "N 321 134 185 110 16 94 211 118 91 Total sample "FL 138.0 125.0 147.5 166.0 156.3 167.6 123.0 120.8 126.0 N 288 107 180 108 15 93 180 92 87 C BW 34.0 23.8 40.3 52.0 40.0 54.3 23.0 21.0 25.0 N 268 93 173 102 12 90 166 81 83 Nearshore FL 142.3 130.5 149.0 165.0 157.0 166.0 128.3 126.5 130.4 N 252 83 168 101 12 89 151 71 79 BW 35.6 26.5 40.2 50.0 40.0 52.0 25.7 24.2 27.2 N 53 41 12 8 4 4 45 37 8 Offshore FL 114.5 112.6 120.5 178.3 153.5 203.2 103.0 108.0 79.0 N 36 20 12 7 3 4 29 21 8 BW sharks. 22.8 14.0 40.3 79.8 39.0 110.5 9.0 10.6 5.2 *jV = number of h FL = mean fork length in cm. BW = mean body weight in kg. the stomachs, only six occurred in both the near- and offshore areas (Tables 3, 4), including squids, skates, skate egg cases, goosefish Lophius americanus, blue- fish Pomatomus saltatrix, and Bothidae (flatfish). Sum- marizing the prey into major food groups (Fig. 2) shows that 43.0% (by frequency of occurrence) of the food was composed of teleosts, followed by elasmobranchs (16%), cephalopods (3.0%), and miscellaneous organ- isms and trash (pebbles, seagrass, paper scraps; 5.0%). The size of prey ingested appears to be an important factor in its selection, since the majority of prey items observed in the stomachs were small enough to be swallowed whole. Those that were consumed as bite- sized portions included larger skates, goosefish, blue- fish, and smooth dogfish Mustelus canis and spiny dog- fish Squalus acanthias. These food items were eaten by the larger juveniles and adults only. Earlier reports by Bigelow & Schroeder (1948), Springer (1960), Bass et al. ( 1973), and Lawler ( 1976) also indicate that small fish and invertebrates are most common in the diet. Springer (1960) adds that fresh fish is preferred over stale or decomposed fish and mammal flesh. Teleosts The food group 'All Teleosts' (Fig. 2) was composed of species ranging from sedentary (goosefish) to actively-swimming forms (bluefish, mackerel Scomber scombrus). Flatfish (flounders) from the fami- lies Bothidae and Pleuronectidae occurred with the highest (10.0%) frequency overall (Fig. 2). Predation on these two families was most evident in sharks col- lected nearshore (Table 3). Goosefish comprised the second most-important fish in the diet by frequency of occurrence (6.0%) and was consumed by juvenile and adult sharks (Fig. 2). Goosefish remains varied from small (4cmTL) to medium-sized (45cmTL) individuals that were eaten in chunks. Remains of this prey item occurred most often in sandbar shark stomachs col- lected off the Long Island (NY) and New Jersey coasts. Bluefish occurred in nine stomachs (S.O'X), seven of which were from females captured nearshore. Gadids consisted principally of hakes digested beyond species recognition, except for one silver hake Merluccius bilinearis that was relatively fresh. Scombrids occurred in seven stomachs (Fig. 2) and consisted almost exclu- sively of identifiable remains of common mackerel. The occurrence of fast-swimming scombrid species in stom- achs agrees with the reported occurrence of bonito Sarda sarda [and weakfish Cynoscion regalis] by Bigelow & Schroeder (1948). "Other Teleosts" (Fig. 2), comprising 20.0% of the food by frequency of occur- rence, is a group composed of at least 12 species from Table 1, each occurring infrequently in the diet but representative of local availability. Of this food cat- egory, 14% (by frequency of occurrence) also included fish remnants that could not be identified to family or species. Stillwell and Kohler: Food habits of Carcharhinus plumbeus off U.S. northeast coast 141 Table 2 Stomach contents from 321 sandbar sharks Carcharhinus plumbeus captured in nearshore (<100m) and offshore <>100m) waters between Cape Hatteras and Georges Bank. Food Items Vol. (mL) N Stomachs N Arthropoda Cancridae Cancer sp. Umdent. crab Isopoda Cephalopoda Gonatidae Ommastrephidae Illex illecebrosus Unident. squids Echinoderma Scutellidae (sand dollars) Elasmobranchs Squalus acanthias Mustelus canis Raja ennacea Raja sp. Dasyatidae Skate eggs Teleosts Congridae Ophichthus cruentifer Clupeidae Chauliodontidae Lophius amencanus Synodontidae Gadidae Merlucaus bilmearis Carangidae Cottidae Labridae Ophidiidae Pomatomus saltatn.x Scombridae Scomber scombrus Pepnlus triacanthus Triglidae Zoarcidae Macrozoarces amencanus Bothidae Limanda ferruginea Pleuronectidae Teleost unident. Miscellaneous Clam Shells Marine mammal flesh Animal remains Trash Totals 93 0.20 3 1.03 3 0.93 1 0.00 1 0.34 1 0.31 8 0.02 10 3.46 1 0.31 120 0.26 1 0.34 1 0.31 6 0.01 1 0.34 1 0.31 70 0.15 2 0.69 2 0.62 141 0.30 9 3.11 7 2.18 13 246 2365 6785 5190 175 192 925 55 80 75 7445 75 1975 100 50 965 25 30 2792 50 1410 100 14 410 350 1330 4627 2780 4163 55 50 390 1 46,127 0.02 0.53 5.12 14.71 11.25 0.37 0.41 2.00 0.11 0.17 0.16 16.14 0.16 4.28 0.21 0.10 2.09 0.05 0.06 6.05 0.97 3.05 0.21 0.03 0.88 0.75 2.88 10.03 6.02 9.02 0.12 0.10 0.80 0.00 1.38 6 2.07 1 0.34 17 5.88 26 9.00 1 0.34 11 3.80 1 0.34 13 4.49 1 0.34 1 0.34 23 7.95 3 1.03 9 3.11 1 0.34 1 0.34 6 2.07 1 0.34 1 0.34 9 3.11 1 0.34 6 2.07 2 0.69 1 0.34 1 0.34 1 0.34 5 1.73 21 7.26 20 6.92 60 20.76 2 0.69 1 0.34 4 1.38 1 0.34 0.93 4 1.24 1 0.31 13 4.05 25 7.78 1 0.31 7 2.18 1 0.31 2 0.62 1 0.31 1 0.31 20 6.23 3 0.93 6 1.86 1 0.31 1 0.31 3 0.93 1 0.31 1 0.31 9 2.88 1 0.31 6 1.86 1 0.31 1 0.31 1 0.31 1 0.31 5 1.55 8 2.49 18 5.60 47 14.60 2 0.62 1 0.31 4 1.24 1 0.31 289 142 Fishery Bulletin 91(1). 1993 Table 3 Stomach contents from 268 sandbar sharks Carcharhinus plumbeus captured in nearshore (<100m) waters between Cape Hatteras and Georges Bank. Food Items Stomachs Vol. ImL) % N % N % Arthropoda Cancridae Cancer sp. 93 0.22 3 1.37 3 1.10 Unident. crab 1 0.00 1 0.45 1 0.37 Cephalopoda Gonatidae 120 0.28 1 0.45 1 0.37 *Unident. squid 40 0.09 5 2.28 3 1.11 Echinoderma Scutellidae (sand dollars) 13 0.03 4 1.82 3 1.11 Elasmobranchs Squalus acanthias 180 0.42 2 0.91 2 0.75 Mustelus canis 2365 5.64 1 0.45 1 0.37 Raja erinacea 6785 16.20 17 7.76 13 4.80 *Raja sp. 5140 12.27 25 11.41 24 8.95 *Skate eggs 186 0.44 9 4.10 6 2.20 Teleosts Congridae 925 2.20 1 0.45 1 0.37 Clupeidae 80 0.19 1 0.45 1 0.37 Chauliodontidae 75 0.17 1 0.45 1 0.37 *Lophius americanus 5870 14.02 20 9.13 17 6.30 Synodontidae 75 0.17 3 1.36 3 1.10 Gadidae 1450 3.46 4 1.82 4 1.50 Merluccius btlinearis 100 0.23 1 0.45 1 0.37 Carangidae 50 0.11 1 0.45 1 0.37 Cottidae 965 2.30 6 2.73 3 1.10 Labridae 25 0.05 1 0.45 1 0.37 Ophidiidae 30 0.07 1 0.45 1 0.37 *Pomatomus saltatrix 2765 6.60 8 3.65 8 3.00 Scomber scombrus 1410 3.36 6 2.73 6 2.20 Peprilus triacanthis 100 0.23 2 0.91 1 0.37 Zoarcidae 410 0.97 1 0.45 1 0.37 Macrozoarces americanus 350 0.83 1 0.45 1 0.37 *Bothidae 730 1.73 4 1.82 4 1.50 Limanda ferruginea 4627 11.05 21 9.58 8 3.00 Pleuronectidae 2780 6.64 20 9.13 18 6.71 Teleost unident. 3676 8.78 41 18.72 37 13.80 Miscellaneous Clam shells 55 0.13 2 0.91 2 0.75 Marine mammal flesh 50 0.11 1 0.45 1 0,37 Animal remains 340 0.81 3 1.36 3 1.10 Trash Totals 1 0.00 ated offshore. 1 219 0.45 1 0.37 41,862 items duplic Note: Asterisk indicates prey Stillwell and Kohler: Food habits of Carcharhinus plumbeus off U.S. northeast coast 143 Table 4 Stomach contents from 53 sandbar sharks Carcharhmus plumbeus captured offshore (>100m) be- tween Cape Hatteras, North Carolina, and Georges Bank. Food Items Stomachs Vol. (mL % N % N % Arthropoda Isopoda 8 0.18 10 14.28 1 1.90 Cephalopoda Ommastrephidae 6 0.14 1 1.42 1 1.90 Illex illecebrosus 70 1.64 2 2.85 2 3.77 *Unident. squid 101 2.36 4 5.71 4 7.54 Elasmobranchs Squalus acanthias 66 1.54 4 5.71 2 3.77 *Raja sp. 50 1.17 1 1.42 1 1.90 Dasyatidae 175 4.10 1 1.42 1 1.90 *Skate eggs 6 0.14 2 2.85 1 1.90 Teleosts Ophichthus cruentifer 55 1.28 13 18.57 2 3.77 Gadidae 525 12.50 5 7.14 2 3.77 *Lophius americanus 1575 36.92 3 4.28 3 5.66 *Pomatomus saltatrix 27 0.63 1 1.42 1 1.90 Scombridae 450 10.55 1 1.42 1 1.90 Triglidae 14 0.32 1 1.42 1 1.90 *Bothidae 600 14.06 1 1.42 1 1.90 Teleost unident. 487 11.41 19 27.14 10 18.86 Miscellaneous Animal remains Totals 50 1.17 cated nearshore. 1 70 1.42 1 1.90 4265 items dupl Note: Asterisk indicates prey Elasmobranchs Elasmobranchs ranked second to te- leosts as a major food group, accounting for 16.0% of the food by frequency of occurrence (Fig. 2). Skates of the family Rajidae were the principal representatives in this food group, with Raja erinacea occurring most frequently. Unspecified skate remains described as Raja spp. in Table 2 most likely included R. erinacea and possibly R. eglanteria, both which commonly occur in the sampling area. Eleven skate egg cases were also found in seven stomachs. These were generally torn and old looking. However, a few contained yolk mate- rial suggesting they were ingested as a food source rather than by accident. Based on frequency of occur- rence, spiny and smooth dogfish sharks are relatively unimportant in the sandbar shark's diet when com- pared with the importance of skates in the diet (Table 2). Bigelow & Schroeder (1948), Bass et al. ( 1973), and Lawler ( 1976) also report the occurrence of shark remains in sandbar shark stomachs. Springer ( 1960), however, after examining several hundred sand- bar sharks from the Florida coast, reported finding very few stomachs with shark remains. The high fre- quency of occurrence of skates in the sandbar shark's diet can be attributed to their general abundance over the continental shelf ( Waring 1986). Cephalopods From this study, cephalopods (squids and octopus) appear to be generally unimportant in the sandbar shark's diet by evidence of their low num- ber, volume, and frequency of occurrence (Fig. 2). Ear- lier studies by Springer (1960), Clark & von Schmidt (1965), Lawler (1976), and Branstetter (1981) also showed low occurrences of squid in the sandbar shark's diet, but all were based on specimens obtained from inshore areas of low squid abundance. Our findings, however, suggest that predation on this food source is 144 Fishery Bulletin 91(1), 1993 FLATFISH GOOSEFISH BLUEFISH GADIDS SCOMBRIDS OTHER TELEOSTS ALL TELEOSTS ELASMOBRANCHS CEPHALOPODS MISCELLANEOUS — ■— * □ NUMBER VOLUME □ OCCURRENCE T i i ^^^^r !=? !b 10 20 30 40 PERCENT 50 60 70 Figure 2 Major food categories consumed by 321 juvenile and adult sandbar sharks Carcharhinus plumbeus from the U.S. northeast coast, 1972-84. probably linked to prey density, since 7 of the 11 stom- achs containing squid were from sharks captured off- shore where squid are most abundant. It is also pos- sible that the four sharks captured nearshore with squid in their stomachs had moved inshore after eat- ing the squid. Areal comparisons Flatfish and cephalopods were the only food groups to show significant differences (P<0.05, X 2 test) in importance between the areas. Flatfish oc- curred most often in nearshore stomach samples, prob- ably as a result of their high summer abundance in shoaler waters along the coast (Bigelow & Schroeder 1953). Cephalopods occurred more often in stomachs offshore because of their high abundance off the U.S. northeast coast (Lange & Sissenwine 1980, Lange 1982), and hence their availability probably accounts for their appearance in the stomachs examined in this study. Nearshore there was significantly (P<0.05, x 2 test) more predation on elasmobranchs and goosefish by fe- male sharks than by males. Offshore, juvenile males consumed significantly (P<0.05, x 2 test) more 'Other Teleosts' than females. 'Other Teleosts' was the only food group in this area for which there was a differ- ence between sexes. Overall, there was no difference in predation rates on the major food groups with respect to shark size (juveniles or adults). Food volumes Overall, 49% of examined stomachs contained food. Wass ( 1973) found that 45% of stomachs from sandbar sharks captured by hook-and-line off Hawaii contained food remains. In other studies, averages have been lower, but up to 29%* have been observed for the sandbar shark (Springer 1960, Bass et al. 1973, Lawler 1976). Our findings show that stom- ach content volumes ranged from trace amounts to a maximum of 3102 mL, with a mean of 144 mL. The mean for adults was 175.4 mL and 125.2 mL for juve- niles. Stomachs from adult and juvenile females contained more food (184.0 and 165.0 mL) on the average than their male counter- parts (125.0 and 97.0 mL); how- ever, differences were not signifi- cant at the oc=0.05 level (Mest). The ratio of stomach content volume to mean body weight (.vBW) varied between different groups of the population, from a low of 0.30% for adult males to a high of 0.83% for juvenile males offshore. The mean for adults and juveniles was 0.33% and 0.55%, respectively, with an overall sample mean of 0.42%. Means for adults of both sexes were similar for the whole sample and nearshore, ranging from 0.30 to 0.36%. Juvenile males and females from nearshore differed the most, with percentages of 0.42 and 0.66, respectively. The highest stomach content values were from an adult and a juvenile female. Stomach con- tents in these sharks amounted to 5.35 and 5.34% of their body weight, respectively. The adult's stomach contained a whole smooth dogfish and gadid remains. The stomach from the juvenile contained 10 yellowtail flounder Limanda ferruginea and a small goosefish. The flounders (x size=13.4cm) may have been con- sumed as natural prey or obtained as culls from a trawl catch. However, other sandbar sharks caught in the area on the same day contained only 1 or 2 flounder, suggesting that this juvenile female was more success- ful in obtaining natural prey. Overall mean stomach volume in terms of liquid capacity for the sharks in this study was 2.62 L which was 7.7% of the *BW (34.0kg). For adults, the mean was 5.15 L, amounting to 10.0% of their .tBW (52.3 kg). Stillwell and Kohler: Food habits of Carcharhinus plumbeus off U.S. northeast coast 145 A measure of stomach fullness was determined by calculating the ratios of food volume to maximum liq- uid capacity. The mean food volume ( 144 mL) was 5.5% of the mean maximum capacity. Dividing the sample by size-class showed that the percent stomach fullness was 5.2 and 13.5 for adults and juveniles, respectively. One stomach from a juvenile female was filled to 50.0% capacity, while two others approximated 40.0%. Just over half (53.0%) of the stomachs contained less than 10.0%. All adults had less than 10.0%, except for two that ranged from 15.0 to 19.0%. Chincoteague Bay sample The Chincoteague sample was composed of pups and young juvenile sandbar sharks captured at six fishing stations located in the lower bay estuaries. The overall mean fork length and body weight for these sharks was 55 cm and 1.72 kg, respectively. The mean for 65 newborn pups (39 males, 26 females) was 50.7cm and 1.38 kg, while the mean for 29 (16 males, 13 females) "older" pups and small juveniles was 63.7 cm and 2.48 kg. There was no difference in mean fork length or body weight between the sexes within each size- class. Food analysis Food items consisted of crustaceans and fish. By frequency of occurrence, these contrib- uted 82.0 and 13.8%, respectively. Crustaceans were represented primarily by soft blue crabs (75.5%), with the remainder (6.3% ) consisting of lady crabs and man- tis shrimp. Fish prey consisted of small flounder, an- chovy, Atlantic silver sides, mullet, and one smooth dogfish (48cmTL) eaten in three pieces. A more com- plete prey list for young sandbar sharks captured in Chincoteague Bay during the summer of 1983 is given in Medved et al. (1985). Previous studies of young sand- bar sharks along the Virginia coast also showed that their diets consisted of small fish and crustaceans but was dominated by soft blue crabs (Hoese 1962, Medved & Marshall 1981; V.J. Lascara, Jonathan Corp., Nor- folk VA, pers. commun. 1987). Food volumes Stomachs from 75 (79.8%) sharks con- tained food varying from trace amounts to a maximum of 125 mL. Nineteen stomachs (20.2%) were empty. Stomachs from 236 sharks caught by gillnets in Chincoteague Bay during the same time-period (Medved et al. 1985) showed that 85.6% (202) held food remains, while 14.4% (34) were empty. The mean food volume for sharks considered to be newborn pups was 16.6 mL (1.2% of .fBW); for "older" pups and small juveniles, it was 27.0 mL ( 1.1% of .fBW). The whole sample mean was 20.0 mL or 1.2% of the .fBW. Estimates of daily ration and annual food consumption Daily ration Reviews are available of studies and tech- niques for determining stomach evacuation rates (Windell 1978, Fange & Grove 1979) and daily ration (Davis & Warren 1971, Conover 1978, Mann 1978) for several species of teleosts. Comparable types of stud- ies for sharks are lacking in the literature, primarily because the technology for maintaining sharks in a healthy "normal" condition in the laboratory has not been perfected (Gruber & Keyes 1981). A few excep- Table 5 A comparison of feeding-related variables for sandbar shark pups Carcharhinus plumbeus, an caught bv different gear types in Chincoteague Bay and in the nearshore (<100m) and offshore U.S. northeast coast, 1972-84. J juveniles and adults > 100 m ) waters of the N Capture method x BW (kg) .v Stomach contents xMeal size Est. daily ration Source 76' 00' Figure 1 Chesapeake Bay geography. The four fishery locations are indicated by x's peared that year. In 1990, contacts were primarily made in the first half of the fishing season; we used that data (Fig. 3) to describe when Spanish mackerel ap- peared and became abundant that year. Catch-size in- formation for 1989 and 1990 was related to daily sur- face-water temperature records at Kiptopeke and at the Chesapeake Bay Bridge Tunnel, provided by the National Ocean Survey (Fig. 1). The accuracy of our estimates of catch size varied from data con- tact to contact, because the esti- mates were not always simple catch data. Most estimates were quite accurate, particularly when catches were zero or very small (e.g., records being "none caught," "few caught," etc.), when, as usual, the fishermen were will- ing to give a specific estimate ("n boxes caught"), or when the catch was stacked in boxes on pallets for shipment and could be counted by us. In some cases, our esti- mates were verbal (e.g., records being "larger than last week," etc.). To compare estimates, the size of the catch from each data contact in 1989 and 1990 was scored in the following catego- ries: (0) no Spanish mackerel caught, (1) <1 22.6 kg box offish caught, (2) 1-5 boxes, (3) >5-10 boxes, (4) >10-20 boxes, and (5) >20 boxes. Adjacent categories may show some overlap, because the original records are inexact. However, these categories permit- ted a distinct separation of zero or small catches (categories 0,1) from large catches (categories 3.4,5), but a less-distinct separa- tion of intermediate-sized catches. We feel the error of these esti- mates is small and does not af- fect the broad spatial and tempo- ral patterns described. To evaluate temporal distribu- tions, differences in monthly catches in 1989 were tested for each location using a Kruskal- Wallis one-way nonparametric analysis of variance (Table 1) af- ter ranking the scores (SAS 1988). This was supported by Tukey's multiple com- parisons tests (Table 2), applied to the ranked scores to evaluate specific monthly differences. Similar pro- cedures were followed to evaluate spatial distributions. We interpret significance tests on spatial differences with caution, because the number of nets varied among locations and information does not exist to standard- ize nets and nominal effort. We feel this has little NOTE Chittenden et al.: Spatial and temporal occurrence of Scomberomorus maculatus 153 CO LU X o m I o h- < o 10-20 5-10 1-5 <1 — r A EASTERN SHORE ° i ° 9° >20-i 10-20 5-10 1-5 <1 — - 0-~ 1 10 20 30 9 19 29 9 19 29 MAR APR MAY 1 °-2°-|C REEDVILLE 5-10 1-5- B LOWER YORK RIVER • • • LAST o cpo p rtp o o • • .I 1 P°- ,° P -P — , , , I 18 28 8 18 28 7 17 27 6 16 26 6 16 26 5 15 25 JUNE JULY AUG SEP OCT NOV -rfi

20-i D LYNNHAVEN • • 50-10C •• • m • 10-20- • 5-10- FIRST I • • • LAST 1 1-5- l • • • .. i < 1 n^ o nm I m f> i ~~1 1 i i i 1 r • i i i — i" i ? i q t -f* 2 - QO lj> 1 10 20 30 9 19 29 9 19 29 8 18 28 8 18 28 7 17 27 6 16 26 6 16 26 5 15 25 MAR APR MAY JUNE JULY AUG SEP OCT NOV C 90 30 20 10 yu- E WATER TEMPERATURE LAST FIRST . yrvtir'^^T*^^ *~*z^*~~^\ -^5". 1 70- I /2ri ^"^ ■TV* y^t^S"' ^"'""""^S 50- -i 1 1 1 1 1 1 1 r - KIPTOPEKE ••BRIDGE-TUNNEL i i i i i i 1 I 1 1 1 1 1 i 1 1 10 20 30 9 19 29 9 19 29 8 18 28 8 18 28 7 17 27 6 16 26 6 16 26 5 15 25 MAR APR MAY JUN JUL AUG SEP OCT NOV DATES Figure 2 Estimates of daily Spanish mackerel Scomberomorus maculatus catches (no. of 22.6 kg boxes landed) in 1989 in Chesapeake Bay: (A) Eastern Shore, (B) lower York River, (C) Reedville, and (D) Lynnhaven, with daily water temperatures (E) at Kiptopeke and off the Chesapeake Bay Bridge-Tunnel. All data contacts are indicated. Zero catches are indicated by open circles. One box = 22.6 kg fish. "First" and "Last" indicate dates when first and last fish were captured; this is not specified when a long time-lapse in sampling occurred before the "First" or after the "Last" record. effect on temporal trends, however, because the same number of nets was generally used in a fishery through- out the season. To make significance tests for differences between areas, we converted the raw catch records to catch- per-unit-effort (C/f) by using the nominal number of nets (Table 1) to estimate effort. The resulting C/f records were then scored into the categories described above. This procedure does not change the original scores for records of "no catch" or "<1 box"; it does tend to lower scores for larger catches, thereby making it more difficult to declare significance. 154 Fishery Bulletin 91(1). 1993 Table 1 Summary, by location, of Kruskal-Wallis one-way nonparamet- ric significance tests for monthly differences in Spanish mack- erel Scomberomorus maculatus catches in 1989. n = number of records at one location, df+1 = number of months sampled. Location Nets n r df Prob. Reedville 2 31 21.83 6 0.0013 York River 7 20 16.37 7 0.0219 Lynnhaven 5 43 31.86 7 0.0001 Eastern Shore 7 23 16.94 5 0.0046 F -90 C . Kiptopeke Temperature WATER 8 >20- ° BBT Temperature RRST 75 115 Catches Made a No Catches a * — "Zf&Jl^ - 80 -70 ^ H 03 /W*«E^b m W 10-20- O ^__J -60 2 TJ m -"-^ -50 -10 TJ — 5 10t > ATC -40 C o o- 0* a a arjoj 10 20 30 9 19 29 9 19 29 8 e MAR APR MAY JUNE Figure 3 Estimates of daily Spanish mackerel Scomberomorus maculatus catches (no. of 22.6 kg boxes landed), early 1990 at Lynnhaven in Chesapeake Bay, with daily surface-water temperatures at Kiptopeke and off the Chesapeake Bay Bridge-Tunnel (BBT). "First" indicates date when first fish were captured. No records collected mid-May to early June. Table 2 Summary, by location, of Tukey's multiple comparisons tests to evaluate specific monthly differences in Spanish mackerel Scomberomorus maculatus catches in 1989. Mean ranks (of scores for catch sizes; see Methods) without the same letters are significantly different at a=0.05. Mean Mean Month n rank Significance Month n rank Significance Reedville York River Jun 5 25.3 a Jul 2 19.5 a Aug 6 23.1 a Jun 1 18.0 a Jul 2 20.0 a b Aug 2 16.0 a Sep 6 15.7 a b c Sep 4 11.0 a b May 5 9.6 be Mar 2 7.0 b Apr 2 7.0 c Apr 5 7.0 b Oct 5 7.0 c May 1 7.0 b Oct 3 7.0 b Lynnhaven Eastern Shore Jun 6 36.8 a Jun 4 19.3 a Jul 4 36.3 a Jul 4 18.5 a b Aug 5 29.1 a b Aug 2 14.0 a b c Sep 4 21.1 b c May 8 8.7 be May 9 16.7 b c Apr 1 5.5 c Oct 7 12.4 c Oct 4 5.5 c Apr 4 10.0 c Nov 3 10.0 c Spatial distribution Results Spanish mackerel become widely distributed in sum- mer throughout Virginia waters of the Chesapeake Bay. In 1989, we observed large catches, at least on occa- sion, at Lynnhaven, off the lower York River, on the Eastern Shore, and at Reedville (Fig. 2). Catches were consistently large in June, July, and early August off Lynnhaven and apparently off the lower York River, although records were not as complete there. Com- paratively low catches were consis- tently made at Reedville and on the Eastern Shore. We formed the dis- tinct impression from our data and observations that Spanish mackerel were much more abundant in the summer along the lower Western Shore of Chesapeake Bay in 1989 than either along the Eastern Shore or upbay at Reedville. Our interpretation of spatial pat- terns in Spanish mackerel abun- dance is supported by significance tests that evaluate the null hypoth- esis, within months, of no difference in C/f between areas. Kruskal- Wallis non-parametric tests for 1989 showed significant differences in C/f between areas in June and July (Fig. 4), when peak abundance occurred, but there were no sig- nificant differences in the other- months when abundance was lower (Table 3). Tukey's multiple compari- sons tests (Table 4) showed signifi- cantly higher C/f in July at Lynnhaven and the lower York River than at Reedville or the East- ern Shore. In June, these tests showed significantly higher C/f at Lynnhaven than on the Eastern Shore. Reedville C/f in June was in- termediate and not significantly dif- ferent from either Lynnhaven or the Eastern Shore; data from the Lower York River were not included in the Tukey's test presented because only one data contact was made there in June. Discussion Spanish mackerel primarily occur in the lower Chesapeake Bay, i.e., NOTE Chittenden et al.: Spatial and temporal occurrence of Scomberomorus maculatus 155 10-20- X X o m 5-10 -| x« X XX * X x • ~— , o 1-5- XX •O X A X • X <1- o- e> e • A • o o go • • t J x X A i 10 i 20 30 10 I 20 30 9 19 29 JUN JUL AUG Figure 4 Estimates of daily Spanish mack- erel Scomberomorus maculatus catch-per-unit-effort (C/f, no. of 22.6kg boxes landed), June- August 1989 by location. Com- paratively few or no fish were landed March-May and Septem- ber-November, (x) Lynnhaven, (▲> lower York River, (•) Reedville, (O) Eastern Shore. Table 3 Summary, by month of Kruskal-Wallis one-way nc nparamet- ric significance tests for differences between areas in Spanish mackerel Scomberomorus maculatus catches in 1989. ;; = num- ber of records in one month, df+1 = lumber of areas. Month /? X 2 df Prob. Apr 12 none caught Mav 23 1.51 3 0.6799 Jun 16 8.34 3 0.0395 Jul 12 9.68 3 0.0215 Aug 15 1.86 3 0.6011 Sep 14 0.53 2 0.7673 Oct 19 1.71 3 0.6338 Table 4 Summary, by month, of Tukey's multiple comparisons tests to evaluate specific differences between areas in Spanish mack- erel Scomberomorus maculatus catches in 1989. Mean ranks (of scores for catch-per-unit-effort, see Methods) without the same letters are significantly different at a=0.05. There were no significant differences in months not tabulated. Area Mean rank Significance June Lynnhaven Reedville Eastern Shore July Lynnhaven York River Reedville Eastern Shore 11.50 6.40 4.75 9.63 9.25 3.50 3.50 a a b b the waters of Virginia. We found regular occurrences at Reedville near the Potomac River mouth, occasion- ally in high numbers as noted by Uhler & Lugger (1876), although Hildebrand & Schroeder (1928) re- ported few occurrences north of the Rappahannock River. Many fish may enter Maryland waters in years of abundance (Butz & Mansueti 1962), such as in 1880 when landings were 8.2 1 (Earll 1883). However, catches there have always been small compared with those in Virginia, where landings made up 97-99% of the re- ported bay- wide catch in 1880 (Earll 1883), in 1920 (Hildebrand & Schroeder 1928), in 1887-1967 (Lyles 1969), and in 1968-76 (Trent & Anthony 1979), and in 1978-90 (from annual printouts, "(Year) landings for the U.S.," provided by the NMFS Office of Data Infor- mation Management to VIMS library). Spanish mackerel may be abundant throughout much of the Chesapeake Bay in Virginia. Large pound and gillnet fisheries existed for it in the 1880s off Gloucester and Mathews counties on the Western Shore, off the Eastern Shore from Cape Charles to Crisfield MD, and off Tangier Island VA (Earll 1883 and 1887, McDonald 1887). Fish also enter the more- saline, lower parts of tributaries like the Potomac and York Rivers (Baird cited in Goode 1888, Hildebrand & Schroeder 1928). Though they may be useful for management and environmental impact assessment, little data exist to describe in fine detail the spatial distribution of Span- ish mackerel in Chesapeake Bay. Such data would be difficult and probably expensive to obtain without man- datory catch-reporting by all the commercial fisheries, because this is a pelagic, fast-swimming, and widely- distributed species that is not well suited to most fishery-independent collecting programs. However, the large-scale distributional patterns of this species ap- parently have been stable for over 100 years. Our data, biological notes, and anecdotal information from the years 1870-80 (Uhler & Lugger 1876, Earll 1883 and 1887, McDonald 1887) and 1920-60 (Hildebrand & Schroeder 1928, Butz & Mansueti 1962), and long- term landings data from Maryland and Virginia indi- cate that this species primarily occupies waters which, according to Lippson & Lippson (1984), are of poly- haline salinity ( 18-30 ppt) and the saltier portions of mesohaline waters (5-18 ppt). 156 Fishery Bulletin 91(1). 1993 Temporal distribution Results Spanish mackerel occur in Chesapeake Bay from late April to early October. They first appeared in the catches on 15 May 1989 (Fig. 2) when three fish were taken off Lynnhaven, on 26 April 1990 (Fig. 3) when two fish were taken there, and, anecdotally according to those fishermen, on about 10 May 1991. Though fishing continued well afterwards, the last catches were on 3 October 1988 and 2 October 1989, dates when we recorded only a few individuals at the lower York River and Lynnhaven fisheries, respectively. Peak abundance of Spanish mackerel in Chesapeake Bay occurs from early or mid-June through mid- August, based on pound-net records. After the first appearance in 1989, catches at Lynnhaven rapidly rose to high levels in early to mid-June and remained high through mid- to late August (Fig. 2). Combined catches at Reedville, the lower York River, and the Eastern Shore showed the same pattern. Comparatively few fish were captured in any area after late August or early September in that year. After fish appeared in late April 1990, catches at Lynnhaven remained low through at least early May (Fig. 3) when our observa- tions temporarily ceased. Catches were large at Lynnhaven by early to mid-June when observations were made again. Our interpretation of temporal patterns in Spanish mackerel abundance is supported by significance tests that evaluate the null hypothesis of no difference in catch between months within locations. Significant dif- ferences in catch were found between months at each location in 1989 (Table 1). Catches were significantly higher at each location in summer months (June, July, August) than in early spring (March, April) or late fall (October, November) (Table 2). As typically occurs with multiple comparisons tests, intermediate-size catches in May and September were or were not significantly different from adjacent periods of higher or lower catch; the trend of increasing abundance to midsummer and decreasing abundance into fall is the most important feature here. Spanish mackerel abundance during the season fol- lows a unimodal pattern. Catches in 1989 at Lynnhaven especially, along the Eastern Shore, and off the lower York River show a roughly bell-shaped distribution (Fig. 2). Catches may be bimodal at Reedville near the upbay margin of the range. Spanish mackerel occur in Chesapeake Bay when water temperatures near the Bay mouth exceed about 17° C. The first fish were taken at Lynnhaven when temperatures had risen to 17° C in 1989 (Fig. 2), and, in 1990, 19° C after a period of rapid temperature in- crease in late April (Fig. 3). Large catches began in late May in 1989, soon after temperatures rose to 20° C (Fig. 2), and catches remained large through midsum- mer at 21-27°C. The last fish were taken at Lynnhaven on 2 October 1989 when temperatures decreased and remained below 21°C. Discussion Our data on the temporal occurrence of Spanish mack- erel in Chesapeake Bay agree with Earll (1883) and Hildebrand & Schroeder (1928) in that (1) the overall period of occurrence in this species is generally mid- May through early October, (2) peak abundance is early June through mid-August or mid-September, (3) the last records of catches are all in early October, and (4) the initial records of appearance are generally in mid-May (10, 15 May in our records; 12, 20 May in previous records), though fish may appear consider- ably earlier (26 April, our records). Our late- April record may reflect the early, rapid temperature increase that occurred in 1990. The period when Spanish mackerel occur in Chesapeake Bay is shorter than their late- April to early-November distribution off North Caro- lina (Earll 1883, Smith 1907, Roelofs 1951) but is some- what longer than their distribution off New York and New Jersey, variously reported as late May or late July to late September-early October (Earll 1883, Bean 1903, Nichols & Breder 1926). The bell-shaped distri- bution of catches that we, and apparently Hildebrand & Schroeder (1928), observed for Chesapeake Bay dif- fers from a bimodal distribution (i.e., peak abundance in spring and fall) reported for North Carolina (Smith 1907, Hildebrand & Cable 1938, Roelofs 1951). Pre- sumably, this reflects a north-south migration by part of the population(s) through North Carolina waters in spring and fall, in contrast to a summer residence in the Chesapeake. Munro (1943) reported that the genus Scom- beromorus is subtropical and tropical in distribution, the optimum range of all species being within the 20° C ocean isotherm in summer. Our findings agree, in that Spanish mackerel initially appear in Chesapeake Bay at temperatures of about 17-19° C and become abun- dant at about 20° C. Other reports also support that value (Earll 1883, Manooch 1984, Goode 1888). Perret et al. (1971) captured one fish at 10°C, but that ap- pears unusual. Beaumariage (1970) related the 20° C ocean isotherm to Spanish mackerel distribution and suggested Long Island would be near their northern limit in August. Indeed, they are uncommon off Mas- sachusetts (Nichols & Breder 1926, Bigelow & Schroeder 1953). The time-period when temperatures are above 20° C decreases with increasing latitude, and NOTE Chittenden et al.: Spatial and temporal occurrence of Scomberomorus maculatus 157 that probably explains why, as noted above, this spe- cies occurs for respectively shorter periods in the sum- mer off New Jersey-New York, in Chesapeake Bay, and off North Carolina. Timing of the appearance and disappearance of Span- ish mackerel in Chesapeake Bay is probably regulated, in part, by temperature differences between the Bay and ocean. Bay waters warm up faster than the ocean in spring and cool faster in fall, due to their different volumes. Cooler ocean temperatures probably limit the time when fish arrive in Chesapeake Bay in spring, and cooler Bay temperatures probably limit the length of time they remain there in fall. Ocean isotherms off the Bay mouth in May and September-October show slightly warmer water along the southern (e.g., West- ern) shore of the Bay (Anonymous 1989a,b,c). In spring, this might encourage fish to initially enter the Bay along the Western Shore as our records suggest. In fall, it might encourage them to leave that area last. Acknowledgments We are indebted to Mr. S. Lyles, National Ocean Sur- vey, for providing records of water temperatures at the Chesapeake Bay mouth, and to the fishermen in many pound net fisheries who helped us in gathering infor- mation. H. Austin, J. Musick, and D. Sved reviewed the manuscript. Financial support was provided by the College of William & Mary, Virginia Institute of Ma- rine Science, by Old Dominion University, Applied Ma- rine Research Laboratory, and by a Wallop/Breaux Pro- gram Grant from the U.S. Fish and Wildlife Service through the Virginia Marine Resources Commission for Sport Fish Restoration Project F-88-R3. L.R. Barbieri was partially supported by a scholarship from CNPq, Ministry of Science and Technology, Brazil (pro- cess no. 20358 1/86-OC). Citations Anonymous 1989a East Coast SST-monthly mean. NOAA, Natl. Weather Serv., Natl. Environ. Satellite Data Inf. Serv. (NESDIS), Natl. Ocean Serv., Oceanogr. Monthly Summ. 9(6):21. 1989b East Coast SST-monthly mean. NOAA, Natl. Weather Serv., Natl. Environ. Satellite Data Inf. Serv. (NESDIS), Natl. Ocean Serv., Oceanogr. Monthly Summ. 9(6):21. 1989c East Coast SST-monthly mean. NOAA, Natl. Weather Serv., Natl. Environ. Satellite Data Inf. Serv. (NESDIS), Natl. Ocean Serv., Oceanogr. Monthly Summ.9(10):21. Bean, T.H. 1903 Catalogue of the fishes of New York. N.Y. State Mus. Bull. 60, 784 p. Beaumariage, D.S. 1970 Current status of biological investigations of Florida's mackerel fisheries. Proc. Gulf Caribb. Fish. Inst., Annu. Sess. 22:79-86. Berrien, P., & D. Finan 1977 Biological and fisheries data on Spanish mack- erel, Scomberomorus maculatus (Mitchell). Sandy Hook Lab. Tech. Ser. Rep. 9, NMFS Northeast Fish. Sci. Cent., Highlands NJ, 52 p. Bigelow, H.B., & W.C. Schroeder 1953 Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Butz, G., & R.J. Mansueti 1962 First record of the king mackerel, Scomberomorus cavalla, in northern Chesapeake Bay, Maryland. Chesapeake Sci. 3:130-135. Chittenden, M.E. Jr. 1991 Operational procedures and sampling in the Chesapeake Bay pound net fishery. Fisheries (Bethesda) 16(51:22-27. Chittenden, M.E. Jr., L.R. Barbieri, & CM. Jones In review Fluctuations in abundance of Spanish mackerel in Chesapeake Bay and the middle Atlantic region. N. Am. J. Fish. Manage. Collette, B.B., & J.L. Russo 1984 Morphology, systematics, and biology of the Span- ish mackerels {Scomberomorus, Scombridae). Fish Bull., U.S. 82:545-692. Collette, B.B., J.L. Russo, & L.A. Zavala-Camin 1978 Scomberomorus brasiliensis, a new species of Spanish mackerel from the Western Atlantic. Fish. Bull, U.S. 76:273-280. Earll, R.E. 1883 The Spanish mackerel, Cybium maculatum (Mitch.) Ag.; its natural history and artificial propa- gation, with an account of the origin and develop- ment of the fishery. U.S. Comm. Fish. Rep. Comm. 1880. App. E. Pt. 8:395-424. 1887 Maryland and its fisheries. In Goode, GB. (ed.), The fisheries and fishery industries of the United States, p. 421-448. U.S. Comm. Fish. Sec. II. Goode, G.B. 1888 American fishes. W.A. Houghton, NY, 496 p. Hildebrand, S.F., & L.E Cable 1938 Further notes on the development and life his- tory of some teleosts at Beaufort, N.C. Bull. U.S. Bur. Fish. 48(241:505-642. Hildebrand, S.F., & W.C. Schroeder 1928 Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43, 388 p. Lippson, A.J., & R.L. Lippson 1984 Life in the Chesapeake Bay. Johns Hopkins Univ. Press, Baltimore. 229 p. Lukens, R.R. (editor) 1989 Spanish mackerel fishery management plan (Gulf of Mexico). Gulf. States Mar. Fish. Comm. 19. 158 Fishery Bulletin 91 1 1993 Lyles, C.H. 1969 The Spanish mackerel and king mackerel fisheries. U.S. Fish Wildl. Serv., Bur. Comm. Fish. Hist. Fish. Stat. C.F.S. 4936, 21 p. Manooch, C.S. Ill 1984 Fishes of the southeastern United States. N.C. State Mus. Nat. Hist., Raleigh, 362 p. McDonald, M. 1887 Virginia and its fisheries. In Goode, G.B. (ed.), The fisheries and fishery industries of the United States, p. 449-473. U.S. Comm. Fish. Sect. II. Munro, I.S.R. 1943 Revision of Australian species of Scombero- morus. Mem. Queensl. Mus. 12, Pt. 2:65-95. Musick, J .A. 1972 Fishes of Chesapeake Bay and the adjacent coastal plain. In Wass, M.L. (compiler), A check list of the biota of lower Chesapeake Bay, p. 175-212. Va. Inst. Mar. Sci. Spec. Sci. Rep. 65. Nichols, J.T., & CM. Breder Jr. 1926 The marine fishes of New York and southern New England. Zoologica (NY) 9:1-192. Perret, W.S., W.R. Latapie, J.F. Pollard, W.R. Mock, G.B. Adkins, W.J. Gaidry, & C.J. White 1971 Fishes and invertebrates collected in trawl and seine samples in Louisiana estuaries. In Perret, W.S., et al. (eds.) Cooperative Gulf of Mexico estuarine in- ventory and study, Louisiana. Phase 1, Area descrip- tion; Phase 4, Biology. La. Wildl. Fish. Comm., New Orleans. Reid, G.K. Jr. 1955 The pound net fishery in Virginia. Part I - His- tory, gear description, and catch. Commer. Fish. Rev. 17(5):1-15. Roelofs, E.W. 1951 The edible finfishes of North Carolina. //; Taylor, H.F. (ed.), Survey of marine fisheries of North Carolina, p. 109-139. Univ. N.C. Press, Chapel Hill. Ryder, J.A. 1882 Development of the Spanish mackerel (Cybium maculatus). Bull. U.S. Fish. Comm. 1:135-163. SAS 1988 SAS/STAT users guide, release 6.03 ed. SAS Inst., Inc., Cary NC, 1028 p. Smith, H.M. 1907 The fishes of North Carolina. N.C. Geol. Econ. Surv., Raleigh, Vol. 2, 453 p. Trent, L., & E A. Anthony 1979 Commercial and recreational fisheries for Span- ish mackerel, Scomberomorus maculatus. In Proc, Mackerel Colloq., p. 17-32. Gulf States Mar. Fish. Comm. 4. Uhler, P.R., & O. Lugger 1876 List of fishes of Maryland. Rep. Comm. Fish. Md. 1876, 176 p. Uncoupling of otolith and somatic growth in Pagrus auratus (Sparidae) Malcolm R Francis Maryann W. Williams Andrea C. Pryce Susan Pollard Stephen G. Scott Fisheries Research Centre. MAF Fisheries PO. Box 297. Wellington. New Zealand Slow-growing fish tend to have heavier, larger otoliths than fast- growing fish of the same length, be- cause otoliths continue to grow even when somatic growth has slowed or stopped (e.g., Templeman & Squires 1956, Mosegaard et al. 1988, Reznick et al. 1989, Secor & Dean 1989, Secor et al. 1989, Campana 1990, Pawson 1990). This uncou- pling has important implications for the back-calculation of fish lengths from check marks in the otoliths. If the relationship between an otolith dimension and fish length varies with growth rate, the back- calculated lengths may be biased (Campana 1990). This bias may be largely overcome by specifying a "biological intercept" (such as oto- lith and somatic size-at-hatching) and incorporating time-varying growth (as measured by daily in- crement widths) into the back- calculation equation (Campana 1990). Pagrus auratus (Bloch & Schneider 1801) is a commercially- important sparid fish that ranges through most of the temperate to subtropical Western Pacific Ocean (Paulin 1990). It has been reported previously under a variety of syn- onyms, especially P. major (Japan), Chrysophrys auratus (Australia and New Zealand), and C. unicolor (Aus- tralia) (Paulin 1990). The common name for P. auratus in New Zealand and Australia is "snapper," though it is not a true snapper (Lutjanidae). Uncoupling of otolith and somatic growth has been demonstrated in reared larval and presettlement juvenile P. auratus from Japan (Secor et al. 1989). In this study, we report uncoupling of otolith and somatic growth in wild, post- settlement, juvenile New Zealand snapper. We also discuss the impli- cations this has for back-calculation of juvenile snapper lengths using otolith daily increments. Methods Snapper were caught using a small otter trawl net equipped with a 20 mm stretched-mesh codend. Samples were collected near Kawau Island, Hauraki Gulf, New Zealand (36°25'S, 174°46'E), January 1987 to March 1989. Fish were chilled on capture, and frozen within 24 h. After thawing, snapper were mea- sured to the nearest mm fork length (FL). Trial measurements before and after freezing and thawing showed that shrinkage was mini- mal (mean shrinkage=2.03%, SD=1.09%, n=42), thus no length corrections were made. In New Zealand, snapper have a prolonged summer spawning season from October to February (Scott & Pankhurst 1992), and we follow Paul (1976) in taking the theoreti- cal birthday as 1 January. Each year-class was numbered after its first full year; e.g., snapper spawned during the 1986-87 austral summer were assigned to the 1987 year- class. Age-0+ fish were identified from length-frequency modes; they grow to about 80-140 mmFL at the end of their first year (Paul 1976; M.R Francis, unpubl. data). Sagittae were removed, and one of each pair was weighed and measured for maximum length ( anterioposterior axis) and height (dorso- ventral axis). For snapper <200 mmFL, transverse sections 1 were prepared from a subsample of sagittae, and sulcal width was mea- sured as the distance between the sulcal side of the metamorphic mark (Francis et al. In press) and the sagitta margin at the ventral edge of the sulcus. This measurement was used in preference to total sagitta width because the antisulcal face of sagittae varied considerably in shape, making it a poor refer- ence surface, and because most of the increase in sagitta width oc- curred on the sulcal surface. The collective term "size variables" is used here when referring to sagitta weight, length, height, and sulcal width. A series of analyses of covariance (ANCOVA) were used to investigate the effects of year-class ( 1987 and 1988) and seven sampling periods (Table 1) on the relationship be- tween the four size-variables and FL in 0+ snapper. Data from snapper samples col- lected in Periods 2 and 3 (Table 1) were used to determine whether sagitta size at any given FL depends 'Terminology used to describe otolith planes and ageing follows Wilson et al. (19871. Manuscript accepted 28 October 1992. Fishery Bulletin, U.S. 91:159-164 (1993) 159 160 Fishery Bulletin 91(1), 1993 Table 1 Sampling periods for age-0+ snapper Pagrus auratus, 1987 and 1988 year-classes. Period 1987 1988 27 January 2 March 28 April 30 June 24-27 August 19 October 14 December 4 February 15 March 6 April 30 May-7 June 25 July^t August 31 October 29 Nov-20 Dec on somatic growth rate. Regressions were fitted to plots of size-variables vs. FL, and the residuals were plotted against somatic growth rate. The latter was estimated by the equation Somatic growth rate = (FL-8)/(post-metamorphic age), where the constant 8 represents approximate mean length of snapper at metamorphosis (Fukuhara 1985 and 1991, Foscarini 1988, Battaglene & Talbot 1992). Post-metamorphic age-at-capture was estimated from transverse sections by counting daily increments be- tween the metamorphic mark and the section margin (see Francis et al. [In press] for validation of daily increments). Post-metamorphic age was used rather than post-hatch age because only about 10% of our sections contained cores; use of post-hatch age would have severely limited sample sizes. Similar analyses were not performed on data from Period 1 because of small sample size, nor on data from Periods 4-7 be- cause daily increments deposited during winter are not resolvable with a light microscope (Francis et al. In press). Results Figure 1 shows plots of sagitta size-variables vs. FL for all sampling periods and age-classes combined. Sagitta weight increased exponentially with FL (Fig. 1A). Plots of sagitta length, height, and sulcal width vs. FL were convex, with slopes decreasing over the range 35-300 mmFL (Fig. IB). Data for 0+ snapper of the 1987 and 1988 year- classes collected in Periods 1-7 were extracted for fur- ther analysis by ANCOVA. Because only linear rela- tionships can be analyzed by ANCOVA, sagitta weight and FL were log,,, transformed before the relationship between them was analyzed. Relationships between the other three size-variables and FL are clearly 180 -| A • 150- -v: O) • • > E 120 - . «: f £ *i~' 5 90 - i /■ CO 1 60- Q) CO CO 30 - J 1 '' mm ^^^ 50 100 150 200 250 300 12 - B Length . ,i - 3.5 E E, 10- -3.0 f E ■? 8- Sit:' -2.5 ~ ngth or he 13 - 2.0 S a -1.5 I «! 4- JL*t*" CO cs JEr r 10 £ Sagiti 3 M .fir u >*"&' Sulcal width CO ^0.5 "> 50 100 150 200 250 300 Fork length (mm) Figure 1 Plots of (A) sagitta weight vs. fork length, and (B) sagitta length, height, and sulcal width vs. fork length for Pagrus auratus. Data for all sampling periods and age-classes combined. nonlinear (Fig. IB). However, ANCOVAs fit linear re- gressions to individual samples (i.e., sagittae of snap- per from one year-class caught in one period), which span only short segments of the lower end of the FL range shown in Fig. IB. All samples were tested for nonlinearity by regressing size-variables against FL, and plotting the residuals against FL. There were no trends in the residuals, so the untransformed data were used in the ANCOVAs. The first set of four ANCOVAs (one for each size- variable) tested the effects of year-class and sampling period on sagitta size, using FL (or log„)FL) as the covariate. There were no significant interaction terms involving year-class, and the year-class factor itself was not significant (p>0.05) in any ANCOVA. There- fore, data for the two year-classes were pooled for sub- sequent analyses. NOTE Francis et al.: Uncoupling otolith and somatic growth in Pagrus auratus 161 A second set of four ANCOVAs tested the effect of the seven sampling periods on sagitta size. In each case, the slopes of the regression lines differed significantly among sampling periods (Table 2). Slope coefficients declined mark- edly between Periods 4 and 5 (Table 3); consequently, a third set of four ANCOVAs was limited to data for Periods 1-4. Whereas slopes did not differ significantly for sagitta length, height, or sulcal width, the intercepts did (Table 2). The three size-variables increased relative to FL between time-periods, i.e., snapper sampled later in the year had larger sagittae than those sampled earlier (Fig. 2Bl. The only sample-pairs that did not differ were Periods 1 and 2 for sagitta height and sulcal width measurements Table 2 Summary of results of ANCOVAs of sagitta size-variable data for snapper Pagrus auratus, 1987 and 1988 yeai •-classes combined. Separate analyses were performed with the variables length, height. sulcal width , and log,,, (weight). The covariate was fork length for the first three analyse 3 and log,,, (fork length for the last. NS = not significant; *p<0.05 , **p<0.01. Sagitta Test for Test for Periods for which size- slope intercept intercepts did not variable Periods differences differences differ (p>0.05)' Length 1-7 ** 1-4 NS ** Nil Height 1-7 ** 1-4 NS ** 1&2 Width 1-7 * 1-4 NS ** 1&2 Logi weight) 1-7 1-4 mer test. 'Conditional Tukey-Kra rable 3 Regression s opes for the relationsh ps between sagitta size- variables and fork length for snapper Pagrus auratus during seven samph ng pen ods. Sagitta length, height, and sulcal width were regres *ed against fork length, and sagitta log, (weight) against log,,, (fork length). Data for Periods 1-4 are shown in Fig. 2. N = sample size. Period Length H eight Width Log( weight) N Slope N Slope TV Slope N Slope 1 17 0.045 22 0.032 14 0.0035 18 2.69 2 83 0.042 92 0.031 44 0.0039 88 2.30 3 38 0.040 58 0.033 53 0.0036 40 2.14 4 84 0.041 86 0.033 42 0.0036 83 2.32 5 61 0.037 65 1)028 22 0.0033 62 2.08 6 71 0.037 72 0.027 20 0.0027 72 2.05 7 45 0.038 47 0.029 16 0.0024 44 2.12 (Table 2). In the ANCOVA of sagitta weight vs. FL, slopes differed significantly among the four periods; thus the intercepts could not be tested (Table 2). However, sagitta weight followed the same trend as the other size-variables, being greater in snapper sampled later in the year than in those sampled earlier (Fig. 2A). Periods 2 and 3 data were used independently to investigate the effect of growth rate on size- variables within sampling periods. The data rep- resent juveniles with estimated post-metamorphic ages of 53.5-136.0 d, and lengths of 43-96 mm FL. Estimated growth rates, averaged over the whole juvenile life, ranged from 0.54 to 0.93 mm/d. Re- siduals from regressions of Period-2 size-variables vs. FL were negatively correlated with somatic growth rate (r=-0.87, -0.70, -0.74, and -0.54 for sagitta weight, length, height, and sulcal width, respectively; p<0.01 in all cases). Therefore, sagittae were heavier and larger (in all dimen- sions) in slow-growing than in fast-growing snap- per. Sagitta weight residuals are plotted against growth rate in Fig. 3. Residuals from regressions of Period-3 size- variables vs. FL were also negatively correlated with somatic growth rate (r=0.45, -0.39, -0.13, and -0.26 for sagitta weight, length, height, and sulcal width, respectively). However, only the sagitta weight correlation was significant (p<0.05). To determine whether differences in somatic growth rate might explain the observed differences in sagitta size variables between sample periods (Fig. 2), an analysis of variance was performed on growth-rate estimates for Period-2 and -3 snap- per. Variances for the two periods were homoge- neous (F 4660 =1.15, p>0.05. Period-2 snapper had significantly higher growth rates (.f 0.81 mm/d, range 0.68-0.94 mm/d) than Period-3 snapper (x 0.66 mm/d, range 0.55-0.82 mm/d, F u06 = 128.0, p<0.001). Discussion Residuals analysis showed that in Period 2 and over the length range 43-90 mmFL, slow-growing snapper had larger sagittae (relative to FL) than fast-growing snapper. In Period 3, a similar but weaker pattern was found. This study, there- fore, demonstrates uncoupling of sagitta and so- matic growth in wild age-0+ snapper, and extends a previous report of such uncoupling in reared Japanese snapper up to 30mmSL (Secor et al. 1989). 162 Fishery Bulletin 91(1). 1993 en E CD '5 3 at as CO 20- A 15- ° Period 1 • Period 2 - Period 3 " Period 4 10- 5 - - I i -aft i i i i i i 30 50 70 90 110 E E £ CD '5 C o I 2 en to CO - B Length r o Period 1 • Period 2 • Period 3 ° Period 4 0% o ariwr^' He| 9 h ' k {;Pr' width i i i i i 1.4 1.2 1.0 0.8 0.6 0.4 0.2 ca C a> ca CO 30 50 90 70 Fork length (mm) 110 Figure 2 Plots of (A) sagitta weight vs. fork length, and (B) sagitta length, height, and sulcal width vs. fork length for age-0+ Pagrus auratus, 1987 and 1988 year-classes, sampled during Periods 1-4 (see Table 1). 0.30- | r = -0.87 1 0.15- •--. Residual (mg) p b o * r • • ••• • -0.15 - • • 0.6 i i i i 0.7 0.8 0.9 1.0 Growth rate (mm.day' 1 ) Figure 3 Residuals from a log 10 (sagitta weight) vs. log,,, (fork length) regression of the Period-2 data from Figure 2A plotted 'against Pagrus auratus post-metamorphic growth rate. Increases in sagitta size-variables between Periods 2 and 3 are also consistent with a growth rate effect: snapper with lower growth rates (Period 3) had larger sagittae than those with higher growth rates (Period 2). Snapper growth rate generally declines be- tween summer and winter (Paul 1976), so the pattern of increasing sagitta size over Periods 1-4 (summer- winter) is also consistent with a growth-rate effect. Between Periods 4 and 5, the slopes of single-sample plots of size-variables vs. FL decreased. A reduction in slope during the winter months suggests that somatic growth slows faster in small than in large snapper, leading to relatively larger sagittae in the former. The reduction in slope is responsible for the curvilinear trends observed when data from all periods are pooled (Fig. IB). A similar effect of season on sagitta-somatic relationships has been reported in other species (Reay 1972, Thomas 1983). When sagitta and somatic growth rates are un- coupled, back-calculated lengths may be biased (Campana 1990). To reduce this bias, Campana (1990, eq. 4) connected the growth trajectory end-points (i.e., sagitta and somatic sizes-at-capture) with a "biological intercept" (he suggested sagitta and somatic size- at-hatching). Snapper larvae are about 2mmSL (equivalent to ~2.5mmFL) at hatching, and have cir- cular sagittae that are 0.010-0.012 mm in diameter (M.P Francis, unpubl. data). These values would form an appropriate biological intercept for daily in- crement back-calculations using measurements in ei- ther the anterio-posterior (length) or dorso-ventral (height) axes. Francis (1990) reviewed back-calculation methods, but was not aware of Campana's (1990) study. Francis identified two back-calculation hypotheses: scale (=sagitta) proportional, and body (=somatic) propor- tional. He pointed out that the commonly used Fraser-Lee equation follows neither hypothesis, and recommended that it be replaced with an equation that does. Campana's equa- tion 4 is a modification of the Fraser-Lee equation, and also does not follow scale- or body-proportional hypotheses. This is easily shown by considering the point at which growth trajectories converge. For scale-proportional methods, this point is on the body-size axis; for body-proportional meth- ods, the point is on the scale-size axis; for Campana's method, the point is at the biological intercept which will usually have some small, positive value on both axes (Campana 1990, Francis 1990). Campana's method, therefore, repre- sents a third back-calculation hypothesis, which is based on the idea that the proportional relationship between scale and body size is initiated at some growth stage, such as hatching. (The Fraser-Lee equation was also based on this idea, but, in practice, most authors using that equation cal- culated the intercept from a regression line rather than from biological data [Francis 19901). NOTE Francis et al.: Uncoupling otolith and somatic growth in Pagrus auratus 163 The key factor that must be considered when decid- ing which back-calculation method to use is the accu- racy with which it estimates back-calculated lengths. Comparison of mean back-calculated lengths with mean observed lengths can detect only gross errors (Francis 1990), and is not a good test for accuracy. Campana (1990) used simulations to show that his method re- moved much of the bias associated with a sagitta- somatic growth-rate effect. The existence of a strong growth-rate effect in juvenile snapper suggests that Campana's method should be used to overcome the expected bias. Campana's (1990) equation 4 corrects for growth- rate variability among fish, while assuming linear sagitta-somatic trajectories for individual fish. The need for the latter assumption can be overcome by incorpo- rating time-varying growth into the model (Campana 1990, eq. 7). However, there are two obstacles to use of the time-varying model for snapper: First, the model takes no account of sagitta and somatic size-at- capture, which limits its use to back-calculation of mean lengths; second, the model requires width mea- surements from all daily increments between the bio- logical intercept and capture, plus a proportional rela- tionship between increment width and somatic growth. For snapper, the relationship between increment width and somatic growth is unknown. Furthermore, recent work on other species has shown that changes in in- crement width may lag or be unrelated to changes in somatic growth (Molony & Choat 1990, Wright 1991). For these reasons, we recommend that back-calcula- tion of snapper lengths from daily increments be done using Campana's equation 4. Acknowledgments We thank the University of Auckland for providing research facilities and technical help at the Leigh Ma- rine Laboratory. In particular, we thank M. Kampman, B.S. Doak, and W Jackson for assistance in the field. R.I.C.C. Francis advised on data analysis. Helpful com- ments on the manuscript were given by J.D. Neilson, D.H. Secor, J.M. Kalish, R.I.C.C. Francis, M.J. Kingsford, and an anonymous reviewer. Citations Battaglene, S.C., & R.B. Talbot 1992 Induced spawning and larval rearing of snapper, Pagrus auratus (Pisces: Sparidae), from Australian waters. N.Z. J. Mar. Freshwater Res. 26:179-183. Campana, S.E. 1990 How reliable are growth back-calculations based on otoliths? Can. J. Fish. Aquat. Sci. 47:2219-2227. Foscarini, R. 1988 A review: Intensive farming procedure for red sea bream (Pagrus major) in Japan. Aquaculture 72:191-246. Francis, M.P., M.W. Williams, A.C. Pryce, S. Pollard, & S.G. Scott In press Daily increments in otoliths of juvenile snap- per, Pagrus auratus (Sparidae). Aust. J. Mar. Fresh- water Res. 43(5). Francis, R.I.C.C. 1990 Back-calculation of fish length: A critical re- view. J. Fish Biol. 36:883-902. Fukuhara, O. 1985 Functional morphology and behaviour of early life stages of red sea bream. Bull. Jpn. Soc. Sci. Fish. 51:731-743. 1991 Size and age at transformation in red sea bream, Pagrus major, reared in the laboratory. Aquaculture 95:117-124. Molony, B.W., & J.H. Choat 1990 Otolith increment widths and somatic growth rate: The presence of a time-lag. J. Fish Biol. 37:541-551. Mosegaard, H., H. Svedang, & K. Taberman 1988 Uncoupling of somatic and otolith growth rates in Arctic char (Salvelinus alpinus) as an effect of dif- ferences in temperature response. Can. J. Fish. Aquat. Sci. 45:1514-1524. Paul, L.J. 1976 A study on age, growth, and population struc- ture of the snapper, Chrysophrys auratus (Forster), in the Hauraki Gulf, New Zealand. N.Z. Fish. Res. Bull. 13, 62 p. Paulin, CD. 1990 Pagrus auratus, a new combination for the spe- cies known as "snapper" in Australasian waters (Pisces: Sparidae). N.Z. J. Mar. Freshwater Res. 24:259-265. Pawson, M.G. 1990 Using otolith weight to age fish. J. Fish Biol. 36:521-531. Reay, P.J. 1972 The seasonal pattern of otolith growth and its application to back-calculation studies in Ammodytes tobianus L. J. Cons. Cons. Int. Explor. Mer 34:485- 504. Reznick, D., E. Lindbeck, & H. Bryga 1989 Slower growth results in larger otoliths: An experimental test with guppies (Poecilia reticu- lata). Can. J. Fish. Aquat. Sci. 46:108-112. Scott, S.G., & N.W. Pankhurst 1992 Interannual variation in the reproductive cycle of the New Zealand snapper Pagrus auratus (Bloch & Schneider ) ( Sparidae ). J. Fish Biol. 4 1 :685-696. Secor, D.H., & J. M. Dean 1989 Somatic growth effects on the otolith-fish size re- lationship in young pond-reared striped bass, Morone saxatilis. Can. J. Fish. Aquat. Sci. 46:113-121. Secor, D.H., J.M. Dean, & R.B. Baldevarona 1989 Comparison of otolith growth and somatic growth 164 Fishery Bulletin 91(1), 1993 in larval and juvenile fishes based on otolith length/ Wilson, C.A., R.J. Beamish, E.B. Brothers, K.D. Car- fish length relationships. Rapp. P.-V. Reun. Cons. lander, J.M. Casselman, J.M. Dean, A. Jearld, E.D. Int. Explor. Mer 191:431-438. Prince, & A. Wild. Templeman, W., & H.J. Squires 1987 Glossary. In Summerfelt, R.C., & G.E. Hall 1956 Relationship of otolith lengths and weights in (eds.), Age and growth of fish, p. 527-529. Iowa State the haddock Melanogrammus aeglefinus (L. ) to the Univ. Press, Ames, rate of growth of the fish. J. Fish. Res. Board Can. Wright, P.J. 13:467-487. 1991 The influence of metabolic rate on otolith incre- Thomas, R.M. ment width in Atlantic salmon parr, Salmo salar 1983 Seasonal variation in the relationship between L. J. Fish Biol. 38:929-933. otolith radius and fish length in the pilchard off South- West Africa. S. Air. J. Mar. Sci. 1:133-138. A new method of oocyte separation and preservation for fish reproduction studies* Susan K. Lowerre-Barbieri Luiz R. Barbieri The College of William & Mary, School of Marine Science Virginia Institute of Marine Science Gloucester Point, Virginia 23062 Studies on the reproduction of multiple-spawning fishes often in- volve estimates of batch fecundity and oocyte size (Hunter & Goldberg 1980, DeMartini & Fountain 1981, Hunter et al. 1985, Brown-Peterson et al. 1988). Because data collection and laboratory analysis are rarely concurrent, oocytes which are pre- served and hardened are generally used for these analyses. It is criti- cal, therefore, to have a method of oocyte preservation which does not damage or destroy oocytes and has a determinate effect on oocyte size. The preferred oocyte preservative (Bagenal & Braum 1978, Snyder 1983, Cailliet et al. 1986) has been a modified Gilson's solution: 100 mL 60% ethanol or methanol, 880 mL water, 15 mL 80% nitric acid, 18 mL glacial acetic acid, and 20g mercuric chloride (Snyder 1983). The benefit of using Gilson's is its ability to harden oocytes while chemically separating them from ovarian tissue. However, a number of problems are associated with this procedure, including degenera- tion of hydrated oocytes ( Hunter et al. 1985, Schaefer 1987, Brown- Peterson et al. 1988); substantial and continuous oocyte shrinkage, reported to range from 15% to 24% (DeMartini & Fountain 1981, Schaefer 1987, Witthames & Greer Walker 1987); a relatively long fixa- tion period of several days to a few weeks (Cailliet et al. 1986); and the extreme toxicity of mercuric chlo- ride (West 1990). Formalin solution (4-10%) has also been used to preserve whole fish ovaries (Bagenal & Braum 1978, Hunter 1985, Cailliet et al. 1986). It is recommended by Hunter et al. (1985) as the only preserva- tive appropriate for use with the hydrated oocyte method. This is be- cause the hydrated oocyte method estimates batch fecundity by calcu- lating the number of hydrated, unovulated oocytes in gravid ova- ries, and Gilson's destroys hydrated oocytes (Hunter et al. 1985). Formalin preservation has the advantages over Gilson's of ( 1 ) pre- serving hydrated as well as other oocytes over long periods of time, (2) having a short fixation period and (3) low shrinkage rates, varying from to 7% (Hiemstra 1962, Fleming & Ng 1987, Hislop & Bell 1987), and (4) relative ease of handling (Hunter et al. 1985, Cailliet et al. 1986, Schaefer 1987, West 1990). Its greatest disadvan- tage is that oocytes and ovarian tis- sue may become fixed into a hard mass, making it extremely difficult and tedious to separate oocytes without damage (Schaefer & Or- ange 1956, Bagenal & Braum 1978, Cailliet et al. 1986). In this paper we propose a new, two-step method to obtain hard- ened, separated oocytes for fish re- production studies. Oocytes are physically separated before being preserved in formalin, thus main- taining the advantages of formalin fixation and preservation while also providing well-separated oocyte samples. The objectives of this pa- per are to (1) describe this new method and evaluate its effective- ness, (2) determine the shrinkage rates of weakfish Cynoscion regalis oocytes separated and preserved by this method, after 3-4 and 6-7 mo preservation, and (3) assess the ap- propriateness of this method for use with the hydrated oocyte method of estimating batch fecundity ( Hunter et al. 1985). Methods Twenty-eight weakfish Cynoscion regalis, with ovaries in the hydrated but unovulated developmental stage, were collected in the sum- mer of 1991. Fresh (unpreserved) oocytes were removed from the right ovary of each fish and spread onto a microscope slide. Twenty hydrated oocyte diameters were then mea- sured, after a minimum sample size of 15 oocytes was determined using the iterative method described in Sokal & Rohlf (1981) (S=0.05, a=0.05; P=0.90, 5=0.06 mm). An ocular micrometer in a dissecting microscope was used to measure oocyte diameters to the nearest 0.038 mm (1 micrometer unit at a total magnification of 24 x). Mea- surements were taken along the median axis of the oocyte, parallel to the horizontal micrometer gra- dations (Macer 1974, DeMartini & Fountain 1981). Ten of these fish were also used to estimate batch fecundities gravi- *Contribution 1780 of the College of Wil- liam & Mary, School of Marine Science, Virginia Institute of Marine Science Manuscript accepted 28 January 1993. Fishery Bulletin, U.S. 91:165-170 (1993). 165 166 Fishery Bulletin 91(1), 1993 metrically (Bagenal & Braum 1978), using the hydrated oocyte method (Hunter et al. 1985). A 0.2 g subsample of fresh oocytes was taken from the middle of the right ovary, and all hydrated oocytes in each subsample were counted under a dissecting microscope at a magnifica- tion of 24 x. Oocytes were separated from one another and the ovarian membrane through a washing process. Each ovary was slit longitudinally, turned inside out, and held under vigorously flowing tapwater. This flushed the oocytes out of the ovarian membrane and into a 0.01 mm mesh sieve, which was held beneath the ovary. Oocytes collected in the sieve were again rinsed with fully-flowing tapwater to help separate them from one another. The whole procedure took 5-10 min per ovary After draining the water, oocytes were transferred to containers where they were preserved in 2% neu- trally-buffered formalin. This formalin concentration was chosen because it was the lowest possible con- centration that would ensure proper oocyte preserva- tion while minimizing changes in oocyte size and appearance. The equipment necessary for the washing process is very basic. We used two standard faucets (2 cm diam- eter), with flow rates of 133 and 286 mL/s, respectively. Both faucets had sufficient hydraulic pressure to dis- lodge oocytes of all stages from ovarian tissue. How- ever, the faucet with the higher flow rate, and thus greater water pressure, worked best. Any sieve with mesh small enough to retain less-developed oocytes, and deep enough to keep them from being flushed over the edge during washing, can be used as a collecting sieve. We used a sieve made from a piece of nylon plankton net (0.01 mm mesh) inserted between two sections of 10 cm diameter PVC pipe, with a depth (from lip to the mesh layer) also of 10 cm (Fig. 1). Preserved oocytes were measured 3-4 mo after col- lection and, again, 6-7 mo after collection. Samples were stirred before oocytes were removed to reduce bias due to settling differences caused by oocyte size or density. Oocytes were then dipped out of the forma- lin with a spoon and placed in a gridded petri dish. The first 20 undamaged hydrated oocytes were mea- sured along the median axis as described for fresh oocytes. Oocyte damage, due to the washing process, was evaluated by assessing the percentage of dam- aged oocytes in subsamples of 50 hydrated oocytes from each of 10 preserved samples. We considered as dam- aged those oocytes which were partially collapsed and thus not appropriate for diameter measurements. Batch fecundities were also estimated gravimetri- cally from preserved samples (after 3-4 and 6-7 mo preservation) using oocyte samples from the same 10 fish originally used to estimate batch fecundities from fresh samples. Oocytes were stirred, decanted into a sieve, drained of formalin, and washed with tapwater. Oocytes were removed from the sieve, spread on the bottom of a petri dish, and blotted dry with tissue paper. A 0.2 g subsample was then transferred to a gridded petri dish. A small amount of tapwater was added to keep the oocytes moist and to help distribute them evenly over the bottom of the dish. One-way analysis of variance (ANOVA) was used to evaluate differences between fresh and preserved oocyte diameters and batch fecundity estimates. Indi- vidual females were used as blocks to remove the ef- fect of variation among females. To compare batch fecundities based on fresh samples with those based on preserved samples, it was important to evaluate the within-ovary positional effect. This was necessary because fresh oocyte samples were taken from the middle of the ovary, whereas preserved samples came from mixed areas (due to the washing process). Hy- drated oocytes were counted in 0.2 g oocyte samples taken from the anterior, middle, and posterior areas of 28 fresh ovaries. Mean oocyte shrinkage was calculated for each of the 28 ovaries after 3-4 and 6-7 mo preservation. Mean oocyte shrinkage was then plotted against mean fresh hydrated oocyte diameter to evaluate whether oocyte shrinkage was consistent over the size-range of hy- drated oocytes. All data were analyzed using statistical methods available through the Statistical Analysis System (SAS 1988). Model assumptions were evaluated by exami- nation of residuals (Draper & Smith 1981). Batch fe- cundity data was log ln -transformed to meet the as- sumption of homogeneity of variances. 10 cm PVC pipe 0.01 mm mesh Figure 1 Schematic representation of the sieve used to collect weakfish Cynoscion regahs oocytes dislodged during the washing process. NOTE Lowerre-Barbien and Barbien Oocyte separation and preservation for reproduction studies 167 Results We successfully used this separation technique on ova- ries in all stages of development and observed little or no damage to the oocytes (Figs. 2, 3). The percentage of damaged oocytes ranged from to 6%, with an aver- age of 29c. None of these, however, were structurally damaged, i.e., no empty chorions were found. Because such low percentages of slightly-degenerated hydrated 1 Figure 2 Appearance of weakfish Cynoscion regalis oocytes in various developmental stages: (top) fresh and (bottom) after hydraulic separation and fixation in 29c formalin. Bars=l mm. oocytes can also be found in fresh oocyte samples — i.e., some hydrated oocytes are never ovulated and will be resorbed, and some ovulated oocytes are never spawned (e.g., Clark 1934, DeMartini & Fountain 1981)— we considered oocyte damage due to the washing process to be negligible. Oocytes in all stages (primary growth to hydrated) were obtained in sufficiently large numbers and cor- rect proportions to develop oocyte size-frequency dis- tributions (Fig. 2). For most ova- ries it was possible to flush virtually all oocytes out of the ovarian membrane. However, we found it was easier to dislodge oocytes in well-developed ovaries than in early-developing or resorbing-phase ovaries. Formalin (2%) successfully fixed and then preserved weak- fish oocytes for over 6 months with minimal effect on their ap- pearance and size. There was no need for a separate, higher- concentration fixative. Atretic and hydrated oocytes were more opaque after preservation, but much less so than when kept at higher formalin concentrations. After 6 months, hydrated oocytes were still easily recognized by their larger size and greater translucence than were less- developed oocytes (Fig. 3). Most oocytes, in all stages, retained their spherical shape. Hydrated oocytes had a highly significant decrease in diameter after preservation, with a range of 0—11% shrinkage after 3-4 mo in preservative (F=223.25, N= 560, P<0.01). The average oocyte shrinkage, however, was only 5% and after more than 6 months, oocyte shrinkage had not signifi- cantly increased (F=1.91, N=560, P=0.17). Mean shrinkage of hy- drated oocytes preserved for 3-4 and 6-7 mo showed no relation- ship with their original mean fresh diameters (Fig. 4), indicat- ing that an oocyte's stage in the hydration process did not affect its rate of shrinkage. Batch fecundities estimated from fresh oocyte samples were • 168 Fishery Bulletin 9! |l). 1993 Figure 3 Appearance of weakfish Cynoscion regalis hydrated oocytes after 6-7 mo preservation in 2% formalin. Bar=l mm. Discussion not significantly different (F=0.0027, N=10, P=0.14) from those estimated from samples preserved for both 3-4 and 6-7 mo (Table 1). Batch fecundities esti- mated from fresh and preserved samples could be com- pared because no positional effects were found between counts from different areas (ANOVA, F=0.91, N=28, P=0.41). 15-i 3 months A A 6 months Percent shrinkage 01 o 1 . 1 A A A • A a • a '.. •■■ - : a a a aa • • a 0.7 0.8 0.9 1.0 Fresh hydrated oocyte diameter (mm) Figure 4 Mean shrinkage of hydrated oocytes of weakfish Cynoscion regalis after 3-4 and 6-7 mo preservation in 2 C >> formalin. There is need for a reliable method of separating and pre- serving fish oocytes. Generally, separated and preserved oocytes are used to estimate oocyte size and develop oocyte size-frequency distributions, as well as in the hydrated oocyte method of de- termining batch fecundity. Re- searchers often want to evaluate changes in egg size over the spawning season, or between spawning seasons (e.g., DeMar- tini 1990), and oocyte size- frequencies are used to assess whether fish have determinate or indeterminate fecundity (Hunter & Macewicz 1985). The hydrated oocyte method appears to be the easiest and most accurate way to determine batch size in serial spawners (Hunter et al. 1985). All of these types of analyses are integral to reproductive studies of multiple-spawning fishes, and yet their accuracy de- pends on using either fresh oocyte samples or oocyte samples separated and preserved in a reliable fashion. Although still widely used, researchers are begin- ning to recognize a number of problems with the use of Gilson's solution as a preservative. It degenerates hydrated oocytes (Hunter et al. 1985, Schaefer 1987, Brown-Peterson et al. 1988), making it impossible to use the hydrated oocyte method to estimate batch fe- cundity (Hunter et al. 1985). It causes a high rate of Table 1 Batch fecundities of weakfish Cynoscion regalis estimated from fresh oocyte samples and from oocyte samples preserved for 3-4 and 6-7 mo in 2% formalin. Fresh After 3 mo After 6 mo Fish# count preservation preservation 1 159.400 140.400 154,700 2 101,700 104,800 95,700 3 190,000 212,700 208,300 4 158,900 176,300 156,800 5 171,100 174,000 181,400 6 155,100 149,000 143,300 7 208,400 167,800 223,100 8 161,000 154,200 191.800 9 144,700 127,200 148,300 10 209,000 172,200 216,100 NOTE Lowerre-Barbieri and Barbieri: Oocyte separation and preservation for reproduction studies 169 oocyte shrinkage (DeMartini Fountain 1981, Schaefer 1987, Witthames & Greer Walker 1987), which could mask gaps found naturally in oocyte size-frequency distributions. Gilson's solution also causes continuous shrinkage over time (Witthames & Greer Walker 1987), which could make any comparisons of egg diameter during the spawning season, or between consecutive years, meaningless unless all samples were preserved for the same amount of time. Formalin at low concentrations (3-5%) meets the requirements of both an oocyte fixative and preserva- tive (Markle 1984). It prevents microbial activity, with minimal effect on shape, cell contents, and osmolality. As a preservative, it maintains this state, is relatively mild, stable, and long-lasting (Snyder 1983, Markle 1984). Although formalin is commonly used to pre- serve ichthyoplankton samples (Snyder 1983), it has not been commonly used for adult fish-reproduction studies. This is due to the tendency for formalin to fix the whole ovary into a hard mass, from which it is difficult to separate individual oocytes ( Schaefer & Or- ange 1956, Bagenal & Braum 1978). By physically separating the oocytes before preser- vation in formalin, our method overcomes the problem of oocyte separation while maintaining the advantages of using formalin as a preservative. This method is inexpensive, quick, and much less toxic than Gilson's, providing researchers with undamaged oocytes of all stages, with little effect on appearance or size. This new method has been successfully used on weakfish and two other sciaenids (Atlantic croaker Micro- pogonias undulatus, and black drum Pogonias cromis; unpubl. data), and, given the similarity of teleost ova- ries, should be applicable to a wide range of species. Additionally, because oocytes are preserved in a low concentration of formalin, similar to the preservation of most plankton samples, hydrated oocytes processed in this fashion would be comparable to those collected and preserved during plankton studies. This would make it possible to better link adult fish-reproduction studies with those from egg surveys. Acknowledgments We would like to thank Sonny Williams for his ex- traordinary help in obtaining hydrated females. Rogerio Teixeira provided helpful insight in the preliminary stages of developing the method. We would also like to thank Mark E. Chittenden Jr., James Colvocoresses, John Graves, John Hunter, Beverly Macewicz, and two anonymous reviewers for reviewing and commenting on the manuscript. Financial support was provided by the College of William and Mary, Virginia Institute of Marine Science and by a Wallop/Breaux Program Grant from the U.S. Fish and Wildlife Service through the Virginia Marine Resources Commission for Sport Fish Restoration, Project No. F-88-R3. L.R. Barbieri was partially supported by a scholarship from CNPq, Min- istry of Science and Technology, Brazil (process no. 203581/86-OC). Citations Bagenal, T.B., & E. Braum 1978 Eggs and early life history. In Bagenal, T. (ed.), Methods for assessment of fish production in fresh- water, p. 165-201. IBP (Int. Biol. Programme) Handb. 3. Brown-Peterson, N„ P. Thomas, & C.R. Arnold 1988 Reproductive biology of the spotted seatrout, Cynoscion nebulosus, in south Texas. Fish. Bull., U.S. 86:373-388. Cailliet, G.M., M.S. Love, & A.W. Ebeling 1986 Fishes. Wadsworth Publ. Co., Belmont CA, 194 p. Clark, F.N. 1934 Maturity of the California sardine (Sardina caerulea), determined by ova diameter measure- ments. Calif. Fish Bull. 42:1-49. DeMartini, E.E. 1990 Annual variations in fecundity, egg size and con- dition of the plainfin midshipman iPorichthys nota- tus). Copeia 1990:850-855. DeMartini, E.E., & R.K. Fountain 1981 Ovarian cycling frequency and batch fecundity in the queenfish, Seriphus politus: Attributes represen- tative of serial spawning fishes. Fish. Bull., U.S. 79:547-560. Draper, N.R., & H. Smith 1981 Applied regression analysis, 2d ed. John Wiley, NY, 709 p. Fleming, I. A., & S. Ng 1987 Evaluation of techniques for fixing, preserving, and measuring salmon eggs. Can. J. Fish. Aquat. Sci. 44:1957-1962. Hiemstra, W.H. 1962 A correlation table as an aid for identifying pe- lagic fish eggs in plankton samples. J. Cons. Perm. Int. Explor. Mer 27:100-108. Hislop, J.R.G., & M.A. Bell 1987 Observations on the size, dry weight and energy content of the eggs of some demersal fish species from British marine waters. J. Fish Biol. 31:1-20. Hunter, J.R. 1985 Preservation of northern anchovy in formalde- hyde solution. In Lasker, R. (ed.), An egg production method for estimating spawning biomass of pelagic fish: Application to the northern anchovy, Engraulis mordax, p. 63-64. NOAA Tech. Rep. NMFS 36. Hunter, J.R., & S.R. Goldberg 1980 Spawning incidence and batch fecundity in north- ern anchovy, Engraulis mordax. Fish. Bull., U.S. 77:641-652. 70 Fishery Bulletin 91(1), 1993 Hunter, J.R., & B.J. Macewicz 1985 Measurement of spawning frequency in multiple spawning fishes. In Lasker, R. (ed.), An egg produc- tion method for estimating spawning biomass of pe- lagic fish: Application to the northern anchovy, Engraulis mordax, p. 79-94. NOAA Tech. Rep. NMFS 36. Hunter, J.R., N.C.H. Lo, & R.J.H. Leong 1985 Batch fecundity in multiple spawning fishes. In Lasker, R. (ed.), An egg production method for esti- mating spawning biomass of pelagic fish: Application to the northern anchovy, Engraulis mordax, p. 67- 77. NOAA Tech. Rep. NMFS 36. Macer, C.T. 1974 The reproductive biology of horsemackerel, Tra- churus (L.), in the North Sea and English Channel. J. Fish Biol. 6:415-438. Markle, D.F. 1984 Phosphate buffered formalin for long term pres- ervation of formalin fixed ichthyoplankton. Copeia 1984:525-528. SAS 1988 SAS/STAT user's guide, release 6.03 ed. SAS Inst, Inc., CaryNC, 1028 p. Schaefer, K.M. 1987 Reproductive biology of black skipjack, Euthynnus lineatus, an eastern Pacific tuna. Int. Am. Trop. Tuna Comm. Bull. 19:169-260. Schaefer, M.B., & C.J. Orange 1956 Studies of the sexual development and spawning of yellowfin tuna (Neothunnus macropterus) and skip- jack (Katsuwonus pelamis) in three areas of the east- ern Pacific Ocean, by examination of gonads. Int. Am. Trop. Tuna Comm. Bull. 6:211-231. Snyder, D.E. 1983 Fish eggs and larvae. In Nielsen, LA., & D.L. Johnson (eds.), Fisheries techniques, p. 165-197. Am. Fish. Soc, Bethesda. Sokal,R.R., & F.J. Rohlf 1981 Biometry, 2d ed. W.H. Freeman, San Francisco, 859 p. West, G. 1990 Methods of assessing ovarian development in fishes: A review. Aust. J. Mar. Freshwater Res. 41:199-222. Witthames, P.R., & M. Greer Walker 1987 An automated method for counting and sizing fish eggs. J. Fish Biol. 30:225-235. Vertical and horizontal movements of adult Chinook salmon Oncorhynchus tshawytscha in the Columbia River estuary Alan F. Olson School of Fisheries WH- 1 0, University of Washington Seattle, Washington 98195 Present address: EA Engineering, Science and Technology Inc., 8520 1 54th Ave. NE, Redmond, Washington 98052 Thomas P. Quinn School of Fisheries WH-1 0, University of Washington Seattle. Washington 98195 Maturing salmon leave oceanic feed- ing grounds and migrate towards their natal rivers, converging on coastal and estuarine waters. Al- though the passage through an es- tuary represents a physical and physiological milestone during the homing migration of salmon and is often a period of heavy commercial and sport harvest, relatively little is known about how oceanographic processes might affect the distribu- tion of salmon. Estuaries are tran- sition zones between coastal and riverine waters, and are areas of rapidly changing temperature, sa- linity, and current regimes which may present migrating fish with osmo- and thermoregulatory chal- lenges. Furthermore, estuaries may also represent a transition zone for the orientation mechanisms salmon use to find their natal stream (McKeown 1984). Several investigators have ob- served the horizontal movements of Atlantic salmon Salmo salar (Stasko 1975), sockeye salmon On- corhynchus nerka (Groot et al. 1975), and chinook salmon 0. tshawytscha (Fujioka 1970) in es- tuaries, and observed both passive and active movements with and into tidal currents. More recent track- ing studies of maturing Atlantic salmon, sockeye salmon, chum salmon O. keta, and steelhead trout O. mykiss in coastal waters have demonstrated that their vertical movements may be related to the local vertical stratification of the water column (Westerberg 1982, Soeda et al. 1987, Quinn et al. 1989, Ruggerone et al. 1990). No studies are presently available which de- scribe both the vertical and hori- zontal movements of salmon within an estuary. The following study was designed to describe the short-term move- ments of adult chinook salmon in the Columbia River estuary outfit- ted with pressure-sensitive ultra- sonic tags to (1) relate these move- ments to tidal currents and the temperature and salinity structure of the water column, and (2) exam- ine how these movements might be explained by their physiology and the need for orientating clues. Materials and methods Study site description The Columbia River has a large es- tuary with tidal influence extend- ing approximately 161 km upriver from the mouth, although salt in- trusion extends no more than 48 km upriver along the bottom (Si- menstad et al. 1984). Average monthly river flows from 1969 to 1982 were 7460 mVs with a range of 4070 m 3 /s in September to 10,530 m 3 /s in June (Simenstad et al. 1984). This estuary has mixed semidiurnal tides; that is, each tidal day has two high and two low tides of unequal size (Jay 1984). The mean tidal range (mean high water to mean low water) measured over 138 tides in 1958 was 2.31m at North Jetty (Fig. 1; Jay 1984). Ultrasonic telemetry Chinook salmon were captured dur- ing the morning of each tracking day with short (~5min) drifts us- ing 90-180 m of 21cm stretched- mesh commercial gillnet (-12 m in depth) which fished the entire wa- ter column. When a fish was de- tected, the net was immediately re- trieved, and the fish removed and placed in a 100 L cooler filled with surface water. If more than one chinook was captured, one was se- lected for tracking based on scale retention, lack of scars, and gen- eral activity level. Total length was measured to the nearest cm, and a numbered disc tag was attached be- low the dorsal fin. A pressure- sensitive (74 mm long X 16 mm in diameter) ultrasonic transmitter (Vemco Ltd.), weighing 13 g in wa- ter and calibrated within ±1 m to a conductivity/temperature/depth probe (CTD; InterOcean model 513) prior to the track, was inserted into the stomach of the unanesthetized fish. The fish was placed in the boat's partially-filled watertight fish locker (2. 5x 1.5x0.5 m) for recovery (-30-45 min). The holding tank al- lowed the fish to reach the surface, gulp air, and inflate its swim- bladder. All fish were captured in Manuscript accepted 15 September 1992. Fishery Bulletin, U.S. 91:171-178 (1993). 171 172 Fishery Bulletin 91(1). 1993 Figure 1 Study area and track maps of hori- zontal movements by chinook salmon Oncorhyrwhus tshawytscha tracked in the Columbia River es- tuary. Sampling during flooding ( ) and ebbing (•) tides. Each circle represents 30min of track- ing time. 'H' indicates extended holding period occurred. relatively shallow water (about 5 m) on the south side of Sand Island (except Fish 1 which was captured on the north side of Desdemona Sands), and all fish were released at Buoy 21 (Fig. 1 ). A single fish was released each day and followed primarily during daylight hours from the gillnet ves- sel Midnight Gambler. Transmitted signals were re- ceived by a directional hydrophone and tunable re- ceiver/decoder (Vemco Ltd.). During tracking, the boat typically stayed 50-400 m away from the fish, and the following data were collected: (1) boat position every 5 min from a loran C receiver; (2) water depth beneath the boat every 5 min from a fathometer; (3) fish depth every lmin from the decoder; (4) approximately every 30 min the fish was more closely approached (usually to within 50 m, based on triangulation and signal strength), and secchi disk and CTD casts were made while the boat drifted. CTD casts took about 5 min to perform and measured the conductivity and tempera- ture at intervals of 1 or 2 m, usually to within 4 m of the bottom. In deeper waters, casts were generally limited to 12 m to avoid losing the fish. Except for fish swimming close to the bottom, this range always encompassed the depth at which the fish was swimming and any large changes in temperature or salinity. Data analysis Boat positions were used to reconstruct each fish's path on a horizontal track map and to determine ground speed. A 15 min sampling interval was chosen to cal- culate ground speeds because shorter intervals may overestimate fish speed due to extraneous boat move- ments, and longer intervals may underestimate fish speed because calculations based on a straight line between positions may mask shorter-scale movements. Water and fish depths were used to reconstruct each fish's path on a vertical track map. Conductivity was converted to salinity (Perkin & Walker 1972) for con- struction of temperature and salinity profiles. To determine whether salmon showed preferences for ranges of temperature or salinity, the salinity and temperature of the water experienced by each fish were determined indirectly by substituting the appropriate values from the temperature and salinity profile for the depth at which the fish was swimming during each observation. Salinities and temperatures between the measured depth-intervals were determined by linear interpolation. The range of temperatures and salini- ties available to each fish was determined from tem- perature and salinity profiles separated into 1-unit (°C or %o) intervals. The fraction of the water column that each unit of temperature or salinity occupied within the sampled depth was calculated and multiplied by the time-interval of the representative temperature and salinity profile. Each temperature and salinity profile was assumed to represent water conditions over a time- interval midway between consecutive profiles. Fish that swam near the bottom sometimes exceeded the depth of the CTD casts, and these observations were omitted from analysis of salinity or temperature preference. Frequencies of temperature and salinity were summed over all profiles for each track to obtain the salinity and temperature distribution available to each fish. These distributions were tested statistically by good- ness-of-fit analysis to determine if the distributions of available and experienced conditions were similar. Dif- ferences were assumed to indicate fish were display- ing non-random vertical movements, presumably to se- lect for a favorable combination of environmental factors. NOTE Olson and Quinn: Vertical and horizontal movements of adult Oncorhynchus tshawytscha 173 Results Eight chinook salmon were tracked in the Columbia River estuary from 27 August to 5 September 1987, resulting in 56:39 h of tracking time over more than 127 km (Table 1). Mean river flow over Bonneville Dam during the study period was 2910 m'Vs (range 2370- 3430 m : Vs: Fish Passage Center, Corvallis OR). Secchi disc measurements taken intermittently during all tracks had a pooled average depth of 2.47m (range 1.43-4. 12m for individual tracks). In general, signal reception in the estuary was good and no fish were lost during the tracking period. Tracking of a fish was terminated owing to danger of vessel stranding on mudflats (Fish 1), high waves at the river entrance sandbar (Fish 2,4,8), darkness (Fish 3), or fish move- ment into the ocean (Fish 6,7). Only Fish 5 was fol- lowed during periods of darkness (l:09h). Five of the eight fish (Fish 2,5,6,7,8) had dark or dusky skin color, indicative of lower-river stocks known as tules. Bright- skinned fish (Fish 1,3,4) may have derived from either tules or upriver brights. All upriver brights enter the river with a more "oceanic" appearance and return to spawning grounds and hatcheries primarily near the Hanford Reach (Howell et al. 1984); however, some tules also enter the river in bright ocean-type condition. Horizontal movements Fish usually moved in the direction of the prevailing tidal current, and reversals in direction and a milling/ holding behavior were often associated with changing tides (Fig. 1). The average ground speed (weighted by the number of sampling intervals) for tracked fish was 2.33 km/h (range 1.28-3. 17km/h for individual fish (Table 1). Ground speeds are the resultant of two vectors: velocities (speed and direction) of the tidal current and of the tracked fish. When analyzed by tidal stage, mean ground speeds for individual fish ranged from 0.74 to 4.08 km/h (2.60 overall) during ebbing tides, and 0.91 to 3.12 (2.04 overall) during flood tides (Table 2). Two chinook salmon were recovered after the track- ing period. Fish 2 was recaptured 14 d after release during test fishing operations 93 km from the river mouth, and Fish 7 was recaptured 9 d after release by a sportsman about 80 km from the river mouth. These fish had net travel rates of 6.0 and 7.8 km/d, respec- tively, after release. Vertical movements Mean fish depth was 5.5 m, and mean water depth beneath the boat was 13.4 m (Table 3). Vertical pro- files of temperature and salinity indicated extremely dynamic hydrographic regimes. Within a single track, some profiles indicated nearly uniform temperatures and salinities over all depths, while others revealed strong haloclines and thermoclines. Vertical track maps (Fig. 2), and fish-depth frequency distributions rela- tive to mean temperature and salinity profiles (Fig. 3) for Fish 4 and 5, show two observed patterns of verti- cal movement: Some salmon swam in brackish sur- face waters with large vertical gradients of salinity and temperature and made occasional excursions into uniform bottom waters (Fish 2,6,7,8), whereas others demonstrated periods of swimming in the water col- umn and near the bottom (Fish 1,3,4,5). Some vertical track maps show fish that appear to be deeper than Table 1 Summary statistics for tracks of adult chinook sa mon Oncorl vnchus shawvtseha in the Columbia River estuary. Gross dista nces travel ed and average speeds were based on 15 min sampling periods. Fish Gross Mean total Time distance ground Release Release length tracked traveled speed Fish date time (cm) (h:min) (km) (km/h) Reason for ending track 1 Aug. 27 11:12 91 7:18 11.73 1.89 Possible vessel stranding 2 Aug. 28 12:53 84 6:29 18.52 2.96 High waves at river entrance 3 Aug. 29 12:19 86 7:44 9.75 1.28 Darkness 4 Sept. 1 10:55 76 7:20 16.16 2.23 High waves at river entrance 5 Sept. 2 10:12 96 10:52 24.41 2.26 Darkness 6 Sept. 3 10:44 83 4:26 12.29 2.89 Movement into ocean 7 Sept. 4 09:57 76 4:33 14.27 3.17 Movement into ocean 8 Sept. 5 09:40 81 7:57 20.56 2.65 High waves at river entrance Mean 84 7:05 15.96 2.33 Total 56:39 127.69 174 Fishery Bulletin 91(1). 1993 Table 2 Mean ground speeds and sample sizes during ebb and flood tides based on 15min sampl ng inte rvals fo r chinook salmon Oncorh ynchus tshawytscha tracked in the Columbia River estuary. Ebb Flood Ground Sample Ground Sample Fish speed size speed size 1 0.74 15 3.12 14 2 4.08 9 2.34 16 3 1.86 12 0.19 19 4 2.98 17 1.17 12 5 1.71 5 2.54 21 5 2.09 17 6 3.01 15 2.01 2 7 4.09 11 1.73 7 8 3.10 15 2.24 16 Pooled 2.60 116 2.04 107 structure within the estuary, the frequency distributions of available salinities and temperatures were different for all fish tracks (log- likelihood test, Zar 1984; p<0.001). Hence, it was impossible to com- pare the distributions of temperature and salinity experienced by individual fish. No analysis was made on the depth data transformed to salinity and temperature for portions of fish tracks below depths sampled by the CTD, because the available frequency distributions of salinity and temperature could not be calculated for these depths and the distance from the fish to the bottom could not be accurately determined. Due to these problems, an average of 83.2% (range 50.9- 100%) of the depth observations for individual tracks were converted to experienced salinity and temperature. Fish 3 was not analyzed for temperature and salinity preference because it spent nearly all its time below depths sampled with the CTD. However, the frequency distributions of temperatures and salinities occupied by fish showed modes between 14° and 16°C for five of seven fish, and 17 and 19%c for four of seven fish (e.g., Fish 4 and 5, Fig. 4). The log-likelihood test indicated that all fish occupied different distributions of tem- perature and salinity than they would have experienced by random vertical movements in their environments (p<0.001). the bottom. This resulted from record- ing water depth under the boat, which generally followed a short distance be- hind the fish rather than directly above it. Although this discrepancy makes it impossible to accurately determine the distance of the fish from the bottom, Fish 1 and Fish 3 spent the majority of their time close to or on the bottom, and Fish 4 spent approximately 35% of its time near the bottom. Fish encountered a wide range of sa- linities ( 7. 8-33. 69c c) and temperatures (8.9-22.9°C; Table 3). Due to the dynam- ics of tidal currents and vertical water Discussion In general, the tracked fish moved with tidal currents, milled during periods of low current velocity, and reversed their direction of move- ment with the change of tides. The results suggest that tidal cur- rents are a major component to horizontal fish movements in the Columbia River estuary. Chinook salmon had higher mean ground speeds during ebbing tides than during flooding tides, presumably because tidal and riverine flows are additive during ebbing tides and antagonistic during flooding tides. These findings tend to agree with other estuarine tracking studies (Groot et al. 1975, Fujioka 1970) of Pacific salmon. Fujioka (1970) found that the position of chinook salmon tracked in the Duwamish River estuary was dependent on the tidal stage, with fish generally Table 3 Mean, maximum, and sample size of fish-depth observations; mean, minimum, and maximum water depth beneath the tracking boat; and fish depth observations transformed to salinity and temperature experienced by tracked chinook salmon Oncorhynchus shawytscha within the Columbia River estuary CTD = conductivity/temperature/depth probe. Fish depth I m ) Water depth ( m I Salinity (%< ) Temperature °C) Max. CTD Fish Mean(SD) Max. N Mean(SD) Min. Max. N Mean(SD) Min. M.i\ MeanlSDl Min. Max. depth (ml 1 4.6(1.9) 10.2 406 5.8(2.7) 1.2 12.5 89 12.9(3.6) 7.8 19.8 17.0(2.0) 13.0 22.9 12 2 2.0(1.0) 8.1 374 12.1 (2.9) 6.7 18.3 78 16.0(3.4) 8.0 27.4 16.8(1.9) 11.8 20.0 12 3 17.1(5.7) 24.9 440 16.5(3.6) 7.9 23.8 92 10 4 7.9(5.6) 22.3 405 14.6(4.9) 3.7 29.3 89 18.4(5.5) 9.1 32.7 15.0(2.1) 8.9 18.1 12 5 2.1 (1.6) 24.2 618 14.2 (7.31 4.3 30.5 129 20.4 (3.8) 13.0 32.6 14.4(1.3) 9.4 16.8 14 6 2.3(1.7) 9.5 249 16.2(5.6) 7.9 29.3 51 25.5(3.4) 16.8 32.3 13.6(1.5) 10.5 16.0 12 7 3.3(3.6) 16.1 235 15.4(6.4) 7.0 30.2 55 25.3(5.2) 17.4 33.6 13.0(1.8) 8.9 15.5 10 8 2.8(1.2) 10.8 437 13.7(2.9) 6.7 19.8 92 18.7(2.3) 10.6 31.3 15.6(1.1) 10.0 19.1 12 Pooled 5.5(3.3) 24.9 2675 13.4(4.9) 1.2 30.5 675 19.4(3.8) 7.8 33.6 15.1 (1.6) 8.9 22.9 NOTE Olson and Quinn Vertical and horizontal movements of adult Oncorhynchus tshawytscha 175 *Apf"~v\ - FISH DEFTH ■+■ WATER DEFTH Fish 6 9/02/ B7 1 ^f^ I E r I L » K 1 6V pnrij Figure 2 Cross-section of intestinal loop through gonad of unparasitized Argopecten gibbus. Note the ciliated columnar epithelium (IE). Few macrophagous hemocytes are present in the hemolymph sinus (S) surrounding the intestine, and the lumen (ID con- tains only a slight amount of particulate matter. OV = ovar- ian acini, C = cilia, CT = connective tissue. Mallory's trichrome stain. NM h . vSti' its • '^^^; Sir. Figure 3 Cross-section of intestinal loop through gonad of a parasit- ized Argopecten gibbus. Note that the epithelium (IE) shows a loss of integrity, and the cells are more squamous than columnar. Ciliation has disappeared. The sinus surrounding the intestine contains numerous macrophagous hemocytes (M). The lumen (ID is filled with non-cellular debris (NM). T = testicular acini. CT = connective tissue. Mallory's trichrome stain. A. gibbus. The gonadal indices of non-parasitized A. gibbus during late gametogenesis and spawning were significantly higher (ANOVA, p=0.0001) than those of parasitized individuals in the same gametogenic state (Table 1). Discussion Past reports of larval Echeneibothrium hosts have been limited to species ofVenerupis staminea, inhabiting the west coast of North America (Sparks & Chew 1966). Therefore, this report presents the first evidence for the occurrence of this genus in a bivalve from the east coast of North America, and the first indication of a scallop host. It is presently unclear how Echeneibothrium has Table 1 Tukey's Studentized Range Test of mean gonadal indices (GI) among uninfected (January-April 19901 and infected (No- vember 1990 and February 19911 Argopecten gibbus in the same reproductive state. Means not significantly different are underlined. Pairwise comparisons of GI Feb '90 Mar Jan Apr Feb '91 Nov Mean SD 21.15 ±6.10 17.13 ±2.96 15.93 ±3.05 14.74 ±4.28 9.93 6.11 ±1.97 ±2.08 NOTE Smghas et al.: Occurrence of Echeneibothnum in the calico scallop 181 migrated to eastern coastal waters. It is possible that parasitized intermediate and/or final host species ei- ther were introduced or migrated into the area. Invasion of molluscan gonadal tissue by parasitic flatworms has been described as a secondary invasion, with the hepatopancreas (digestive gland) serving as the primary site. The resulting damage to the gonad, including atrophy and eventually destruction of the germinal epithelium, is believed to be a combination of mechanical pressure and nutrient deprivation (Cheng 1967). However, within the gonads of parasitized Argopecten gibbus, these changes were not apparent. Developing and mature eggs in the acini were identi- cal in morphology to the eggs seen in non-parasitized tissue, and did not differ significantly in diameter (Singhas 1992). The only obvious difference was the higher numbers of hemocytes surrounding the germi- nal epithelium in infected scallops. The most exten- sive tissue damage in parasitized A. gibbus occurred in the epithelium of the intestinal loop. This damage consisted of alterations in intestinal epithelial cell shape and structural integrity, similar to those de- scribed by Cheng (1967) for hepatopancreatic tissue. Failure of the local commercial crop of A. gibbus coincided with the appearance of Echeneibothrium in 1991 (P. Phalen, N.C. Div. Mar. Fish., Morehead City NC 28557-0769, pers. commun.). It is uncertain if the Echeneibothrium infestation contributed directly to this failure, because this species is known to undergo dra- matic fluctuations in number (Moyer & Blake 1986). However, the commercial failure of A. gibbus in Florida during January-February 1991 was attributed to a Protoctistan parasite, Marteilia sp. (Blake & Moyer, 1992). Parasitism may therefore be an important fac- tor in the population biology of A. gibbus, and merits further investigation. Acknowledgments We would like to gratefully acknowledge Dr. Thomas Cheng for his identification of the organism described in this Note, and for his editorial assistance with the manuscript. Thanks also to Dr. Charles Singhas for technical and editorial assistance, and to Dave Taylor of the N.C. Division of Marine Fisheries for his assis- tance in collections. Citations Cake, E.W. 1977 Larval cestode parasites of edible mollusks of the NE Gulf of Mexico. In Shumway, S.E. (ed.), Scal- lops: Biology, ecology, and aquaculture, p. 482- 483. Elsevier Sci. Publ., Amsterdam. Blake, N.J., & Mj\. Moyer 1992 Mass mortality of calico scallops, Argopecten gibbus, resulting from an Ascetosporan infection [abstract]. In 25th Annu. Meet., Soc. Invertebr. Pathol., 16-21 Aug. 1992, Heidelberg, Germany. Cheng, T.C. 1967 Marine molluscs as hosts for symbiosis. Adv. Mar. Biol. 5, 424 p. Cummins, R. Jr., A.L. Rhodes, J. Easley, B. Anderson, J.C. Cato, F. Prochaska, P. Fricke, F. Munden, & B. Palmer 1981 Profile of the calico scallop fishery in the South Atlantic and Gulf of Mexico. Sect. 8.0-12.0, South Atl. Fish. Manage. Counc, Charleston SC. Humason, G.L. 1962 Animal tissue techniques. W.H. Freeman, San Francisco, 560 p. Moyer, MA., & N.J. Blake 1986 Fluctuations in calico scallop production (Argo- pecten gibbus). In Proc, Eleventh annu. trop. and subtrop. fish. conf. of the Americas. Texas Agric. Ext. Serv., Texas A&M Univ., Dep. Anim. Sci. Singhas, L.S. 1992 Reproductive periodicity and gonadal develop- ment of the calico scallop, Argopecten gibbus in North Carolina. Master's thesis, East Carolina Univ., Greenville, 104 p. Sparks, A.K., & K.K. Chew 1966 Gross infestation of the littleneck clam, Venerupis staminea, with a larval cestode (Echeneibothrium sp.). J. Invertebr. Pathol. 8:413^16. Superintendent of Documents Subscriptions Order Form I I YES, enter my subscription as follows: Order Processing Code: *5178 Charge your order. It's Easy! To fax your orders (202) 512-2233 subscriptions to FISHERY BULLETIN (FB) for $24.00 per year ($30.00 foreign). The total cost of my order is $ . Price includes regular domestic postage and handling and is subject to change. Please Choose Method of Payment: I I Check Payable to the Superintendent of Documents □ GPO Deposit Account I [ I i I I 1 1 — | | (Company or Personal Name) (Please type or print) (Additional address/attention line) (Street address) (City, State, ZIP Code) (Daytime phone including area code) I I VISA or MasterCard Account 1 1 1 MM ,„ _,. . . , , Thank you for (( rerlit carrl eypiratinn rla(p) your order! (Purchase Order No.) YES NO May we make your name/address available to other mailers? [~ (Authorizing Signature) 11 Mail To: New Orders, Superintendent of Documents P.O. Box 371954, Pittsburgh, PA 15250-7954 The National Marine Fisheries Service (NMFS) does not approve, recommend or en- dorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS. or to this publication furnished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS ap- proves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirect- ly the advertised product to be us°d or purchased because of this NMFS publication. Fishery Bulletin Guidelines for Contributors Form of manuscript The document should be in the following sequence: Title Page, Abstract (not re- quired for Note), Text, Acknowledgments, Citations, Text footnotes, Appendices, Tables, Figure legends, and Figures. Title page should include authors' full names and mailing addresses and the senior author's telephone and FAX numbers. Abstract Not to exceed one double- spaced typed page. Should include a sentence or two explaining to the gen- eral reader why the research was under- taken and why the results should be viewed as important. Abstract should convey the main point of the paper and outline the results or conclusions. No footnotes or references. Text A brief introduction should por- tray the broad significance of the paper. The entire text should be intelligible to readers from different disciplines. All technical terms should be defined, as well as all abbreviations, acronyms, and symbols in text, equations, or formulae. Abbreviate units of measure only when used with numerals or in tables and figures to conserve space. Measure- ments should be expressed in metric units, with other equivalent units given in parentheses. Follow the U.S. Govern- ment Printing Office Style Manual. 1984 ed., and the CBE Style Manual, 5th ed. Fishery and invertebrate no- menclature should follow the American" Fisheries Society Special Publication 12 (for fishes), 16 (for mollusks), and 1 7 (for decapod crustaceans). Text footnotes should be numbered in Arabic numerals and typed on a sep- arate sheet from the text. Footnotes are not used for reference material or per- sonal communications, but rather to explain or define terms in the text and for contribution numbers on the title page. Informal sources Personal communi- cations, unpublished data, and untitled manuscripts in preparation are noted parenthetically in the text (full name, affiliation, brief address including zip code, and month and year when appro- priate). Acknowledgments Gather all ac- knowledgments into a brief statement at the end of the text. Give credit only for exceptional contributions and not to those whose contributions are part of their normal duties. Citations All titled sources should be listed in the Citations section, including unpublished and processed material. In text, cite as Smith and Jones (1977) or (Smith and Jones 1977); if more than one citation, list chronologically (Smith 1936, Jones 1975, Doe In press). All sources cited in the text should be listed alphabetically by the senior authors' surnames under the heading CITA- TIONS. Abbreviations of periodicals and serials should conform to Serial Sources for the BIOSIS Data Base™. Indicate whether sources are in a langu- age other than English. For informal literature, include address of author or publisher. Authors are responsible for the accuracy of all citations. Tables should supplement, not dupli- cate, the text. Each table should be numbered and cited consecutively, with headings short but amply descriptive so that the reader need not refer to the text. For values less than one, zeros should precede all decimal points. In- dicate units of measure in column head- ings; do not deviate from the unit of measure within a column. Table foot- notes should be noted consecutively in Roman letters across the page from left to right and then down. Since all tables are typeset, they need not be submitted camera-ready. Figures Photographs and line draw- ings should be of professional quality- clear and concise— and reducible to 42 picas for full-page width or to 20 picas for a single-column width, and to a max- imum 55 picas high. All graphic ele- ments in illustrations must be propor- tioned to insure legibility when reduced to fit the page format. Line weight and lettering should be sharp and even. Let- tering should be upper and lower case, and vertical lettering should be avoided whenever possible (except for vertical, y, axis). Zeros should precede all deci- mal points for values less than one. Re- productions of line artwork are accepted in the form of high-quality photographic prints from negatives or photomechan- ical transfer (PMT). Halftones should be sharply focused with good contrast. Micron rules should be inserted on elec- tron micrographs, even when magnifi- cation is included in the figure legend. There should be clear distinction be- tween identifying letters (press-on or overlay) and background of photograph. Label each figure in pencil on the back. Send only xerox copies of figures to the Scientific Editor; originals or photo- graphic prints will be requested later when the manuscript is accepted for publication. Copyright Government publications are in the public domain, i.e., they are not protected by copyright. Submission of manuscript Disks Authors are encouraged to re- tain manuscripts on word-processing storage media (diskettes, floppy disks) and submit a double-spaced hardcopy run from the storage media. Submit disks as MS-DOS "print" or "non- document" files (often called "ASCII files"). If a disk cannot be converted to an ASCII file, the author should indicate on the disk the source computer and software language along with the file name. Either 5V4-inch or 3 l /2-inch disks from IBM-compatible or Apple/Macin- tosh systems (non-graphics only) can be submitted, double-sided/double-density or high-density, limiting each file to 300 kilobytes. All 8-inch word-processing disks (e.g., Wang or NBI) must be con- verted onto 5V4- or 3V2-inch MS-DOS print disks. Send original hardcopy and two dupli- cated copies to: Dr. Ronald W. Hardy, Scientific Editor Northwest Fisheries Science Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, WA 98112-2097 Copies of published articles and notes The senior author and his/her organiza- tion each receive 50 separates free-of- charge. Additional copies may be pur- chased in lots of 100. U.S. Department of Commerce Volume 91 Number 2 April 1993 Fishery Bulletin AUG 3 1 U U.S. Department of Commerce Ronald H. Brown Secretary National Oceanic and Atmospheric Administration D. James Baker Under Secretary for Oceans and Atmosphere National Marine Fisheries Service Scientific Editor Dr. Ronald W. Hardy Northwest Fisheries Science Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, Washington 98 1 1 2-2097 Editorial Committee Dr. Andrew E. Dizon National Marine Fisheries Service Dr. Linda L. Jones National Marine Fisheries Service Dr. Richard D. Methot National Marine Fisheries Service Dr. Theodore W. Pietsch University of Washington Dr. Joseph E. Powers National Marine Fisheries Service Dr. Tim D. Smith National Marine Fisheries Service The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115-0070. Second class postage is paid in Seattle, Wash., and additional offices. POSTMASTER send address changes for subscriptions to Fishery Bulletin, Super- intendent of Documents, Attn: Chief, Mail List Branch, Mail Stop SSOM, Washington, DC 20402-9373. Although the contents have not been copyrighted and may be reprinted entire- ly, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this period- ical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $24.00 domestic and $30.00 foreign. Cost per single issue: $12.00 domestic and $15.00 foreign. See back page for order form. Managing Editor Nancy Peacock National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE. BIN C 1 5700 Seattle, Washington 981 15-0070 The Fishery Bulletin carries original research reports and technical notes on investiga- tions in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Begin- ning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 91 Number 2 April 1993 Fishery Bulletin ie Biological Laboratuiy LIBRARY AUG 3 1 1993 lA/oods Hole, Mass. Contents 183 Anganuzzi, Alejandro A. A comparison of tests for detecting trends in abundance indices of dolphins 195 Choat, J.H., Peter J. Doherty, BA. Kerrigan, and J.M. Leis A comparison of towed nets, purse seine, and light-aggregation devices for sampling larvae and pelagic juveniles of coral reef fishes 210 Comyns, Bruce H., and George C. Grant Identification and distribution of Urophyas and Phyas (Pisces, Gadidae) larvae and pelagic juveniles in the U.S. Middle Atlantic Bight 224 Davis, Tim L.O., and Grant J. West Maturation, reproductive seasonality, fecundity, and spawning frequency in Lutjanus vittus (Quoy and Gaimard) from the North West Shelf of Australia 237 Ebert, Thomas A., Stephen C. Schroeter, and John D. Dixon Inferring demographic processes from size-frequency distributions: Effect of pulsed recruitment on simple models 244 Fitzhugh, Gary R., Bruce A. Thompson, and Theron G. Snider III Ovarian development, fecundity, and spawning frequency of black drum Pogonias cromis in Louisiana 254 Forward, Richard B., Leslie M. McKelvey, William F. Hettler, and Donald E. Hoss Swimbladder inflation of the Atlantic menhaden Brevoortia tyrannus 260 Jackson, George D. Seasonal variation of reproductive investment of the tropical loliginid squid Loligo ch/nensis and the small tropical sepioid Idiosepius pygmaeus Fishery Bulletin 9 1 (2). 1993 271 Kimura, Daniel K., Allen M. Shimada, and Sandra A. Lowe Estimating von Bertalanffy growth parameters of sablefish Anoplopoma fimbria and Pacific cod Gadus macrocephalus using tag-recapture data 28 1 Lough, R. Gregory, and David C. Potter Vertical distribution patterns and diel migrations of larval and juvenile haddock Melanogrammus aeglefinus and Atlantic cod Gadus morhua on Georges Bank 304 O'Connell, Victoria M., and David W. Carlile Habitat-specific density of adult yelloweye rockfish Sebastes rubernmns in the eastern Gulf of Alaska 3 1 Prager, Michael H., and Alec D. MacCall Detection of contaminant and climate effects on spawning success of three pelagic fish stocks off southern California. Northern anchovy Engraulis mordax. Pacific sardine Sardinops sagax. and chub mackerel Scomber japonicus 328 Shepherd, Gary R., and Josef S. Idoine Length-based analyses of yield and spawning biomass per recruit for black sea bass Centropristis striata, a protogynous hermaphrodite 338 Smith, Wallace G., and Wallace W. Morse Larval distribution patterns: Early signals for the collapse/recovery of Atlantic herring Clupea harengus in the Georges Banks area 348 Squire, James L. Jr. Relative abundance of pelagic resources utilized by the California purse-seine fishery: Results of an airborne monitoring program, 1 962-90 362 Tringali, Michael D., and Raymond R. Wilson Jr. Differences in haplotype frequencies of mtDNA of the Spanish sardine Sardmella aurita between specimens from the eastern Gulf of Mexico and southern Brazil Notes 371 Bartoo, Norman, and David Holts Estimated drift gillnet selectivity for albacore Thunnus alalunga 379 Haynes, Evan B. Stage-I zoeae of laboratory-hatched Lopholithodes mandtu (Decapoda, Anomura, Lithodidae) 382 Lai, Han-Lin Optimal sampling design for using the age-length key to estimate age composition of a fish population 389 McBride, Richard S., Jeffrey L. Ross, and David O. Conover Recruitment of bluefish Pomatomus saltatrix to estuaries of the U.S. South Atlantic Bight Abstract. —An analysis of alter- native methods for detecting trends in a series of abundance indices is carried out through simulation. The alternative procedures explored have been applied to analysis of relative abundance indices of dolphins in the eastern Pacific Ocean. They include a linear test over a moving time- period, and a nonparametric proce- dure based on smoothing of the time- series of abundance indices. Results indicate that the nonparametric pro- cedure outperforms the linear tests in most of the situations tested. A comparison of tests for detecting trends in abundance indices of dolphins Alejandro A. Anganuzzi Inter-American Tropical Tuna Commission 8604 La Jolla Shores Drive, La Jolla. California 92037- 508 Manuscript accepted 26 January 1993. Fishery Bulletin, U.S. 91:183-194 ( 1993). An important part of the analysis of any set of abundance indices is the application of an objective procedure or test to determine whether changes in the estimates are due to random fluctuations in conditions of the sam- pling procedure or to actual changes in the population size. Such a proce- dure must exhibit certain properties in order to be effective. Among these properties, perhaps most important is the power of the test for a given significance level. In deriving conclusions about changes in the size of a population, we can fall into two types of error. First, we can erroneously conclude that population size has changed when, in fact, differences in estimates are due to random errors. This is usually known as a Type-I error. A Type-II error occurs when we con- clude that the estimates reflect ran- dom fluctuations when, in fact, there has been a change in population size. The probability of falling into a Type-I error is usually referred to as the significance level of the test. The power of a test is defined as 1 minus the probability of a Type-II error. An ideal test will minimize the trade- offs between both types of error. An- other desirable property of a test is robustness to underlying assump- tions about the populations. For ex- ample, tests commonly carried out to detect changes in population size are based on specific assumptions about the error structure of the estimates (e.g., normality) and the model that would best describe the population size as a function of time (e.g., lin- ear, exponential; see, for example, Gerrodette 1987). In the case of dolphin stocks in- volved in the tuna fishery in the east- ern Pacific Ocean (EPO), it has been recommended that their management should include both estimates of ab- solute abundance, derived from re- search-vessel data (RVD), and analy- sis of trends in relative abundance, derived from tuna-vessel observer data (TVOD) (IWC 1992:218). In the case of EPO dolphin stocks, the use of TVOD seems the natural choice for analyzing trends, given the vast amount of low-cost information avail- able from the observer programs. However, for this analysis to be ef- fective, it is necessary to obtain abun- dance estimates with a constant bias over the years, or, at least, a bias that shows no trend over the years. Procedures developed by the Inter- American Tropical Tuna Commission (IATTC) to analyze the TVOD, de- scribed in Buckland & Anganuzzi (1988) and Anganuzzi & Buckland (1989), were specifically aimed at re- ducing the magnitude of year-to-year fluctuations in the estimates due to changing biases. These procedures were complemented with more spe- cific analyses when there were rea- sons to suspect that biases might be changing, for example, due to wide- spread use of high-resolution radar for the detection of birds (Anganuzzi et al. 1991). However, in spite of the robustness of the methods, randomly fluctuating biases (an extra source of 183 184 Fishery Bulletin 9 1(2), 1993 variability) may still affect estimates from year to year. This problem may not be exclusive to the TVOD esti- mates; interannual variability also seems to affect estimates of relative abundance derived from research- vessel data (Wade & Gerrodette 1992). These biases and imprecise estimates will affect the performance of statistical tests designed to detect trends and, ulti- mately, our ability to draw conclusions about the sta- tus of populations. For the analysis of trends in the EPO dolphin stocks, Buckland & Anganuzzi (1988) applied a linear test for trends over a moving period of 5 yr, although they ex- pressed concern about the low power of such a test. Edwards & Perkins (1992) extended the moving time- frame to 10 yr to increase the power. However, such a test still shows some undesirable properties. Given the inadequacy of the tests based on linear regressions, Buckland et al. (1992) proposed a different procedure, based on a nonparametric regression, which addresses some of the problems exhibited by the linear test. In this paper, the characteristics of these tests are discussed and compared by analyzing their performance in a number of simulated scenarios. Current tests for trends Linear tests Buckland & Anganuzzi ( 1988) tested for linear trends over successive 5 yr periods by carrying out a weighted linear regression of abundance index vs. time. Each individual estimate was weighted by the inverse of its variance, calculated by applying a bootstrap procedure. The null hypothesis for the test is that no change has occurred in the population, i.e., that the slope of the regression is equal to zero. As the authors noted, the test has low power since it estimates precision from the deviations of only five estimates from a straight line. Power can be increased by extending the moving time-period to incorporate more years in the test. Un- fortunately, this also increases the probability of vio- lating the assumption of a constant rate of change implicit in the linear model being fitted to the esti- mates (Edwards & Perkins 1992). The linear test also fails to consider the precision of estimates adequately. Variances of the estimates are not taken into account except as weights in the regres- sion. As a consequence, only the ratios of the variances between estimates are relevant, and not their absolute values. For example, if for any given series of esti- mates we double the variance of each individual esti- mate, the results of the test will remain unchanged. Weighting by the inverse of the variance can also present other problems. Suppose, for example, that the variance of the estimate is not independent of the estimate itself, but that the variance is correlated with the estimate, i.e., the coefficient of variation (CV=ratio of standard error to point estimate) is con- stant. In this case, a very low estimate (and especially in the case of an outlier) with a correspondingly small variance will become an influential observation. A linear test for trends will then indicate that there was a decline in the population if that estimate is at the end of the moving period, or a significant in- crease if it is at the beginning. An example from the EPO dolphin abundance estimation is the case of the 1983 index of relative abundance for the northern stock of offshore spotted dolphin Stenella attenuate!, which was an anomalous index as a result of a very strong El Nino event (Fig. 1). In such cases, where the error distribution of the estimates seems to be better approximated by a lognormal distribution, it would be more appropriate to apply weights (wt) defined as wt = ln(l+CV 2 ) 1 . (1) For the comparisons in this paper, two versions of the linear test are applied: the original 5yr linear test with inverse variance weighting applied by Buckland & Anganuzzi (1988), and a 10 yr linear test with weights as described by Eq. 1. Smoothed trends The approach taken by Buckland et al. (1992) differs considerably from the method just described. First, they replaced the assumption of a parametric model for the underlying change in population with a nonparamet- ric model. Among the many possible choices for such a model, they selected the smoothing algorithm known as '4253H, twice' (Velleman & Hoaglin 1981) on the basis of a comparison described by K. L. Cattanach and S. T Buckland (SASS Environ. Model. Unit, MLURI, Craigiebuckler, Aberdeen, Scotland, unpubl. manuscr. ). The adoption of a nonparametric model in- creases the robustness of the test to model mis- specification, a problem that affects the linear test. Furthermore, the procedure, which involves the use of a compound running median, incorporates the infor- mation from nearby years into calculation of the smoothed estimate for a particular year, therefore re- ducing the influence of possible outliers and increas- ing the precision of each smoothed estimate. The smoothed test also provides a different way of looking at the trend. Instead of the trend being described by a single parameter (the slope of a linear regression), the sequence of smoothed estimates constitutes the best estimate of the underlying trend. Anganuzzi: Detecting trends in dolphin abundance indices 185 6000 5000 " 4000 ! 3000 • 2000 1000 1992 Figure 1 Smoothed trends in abundance of the northern offshore stock of spotted dolphin Stenella attenuata in the eastern tropical Pacific Ocean. Broken lines indicate -85% confidence limits. Horizontal lines correspond to 85 r t confidence limits for the 1990 estimate. If both the 1990 confidence limits lie above the upper limit for an earlier year, abundance has increased significantly between that year and 1990 lp<0.05); if both limits lie below the lower limit for an earlier year, abundance has decreased significantly. is often not met, but it can be shown that results are robust-to- moderate departures from it. The third condition is not met for es- timates that are close in time, due to the correlation introduced by smoothing, and the relative importance of this effect is dis- cussed further below. Simulation comparison between tests To illustrate the difference in per- formance between methods, a simulation study was carried out. Series of estimates were simu- lated by assuming different sce- narios of underlying trends in the population over a period of 25 yr. The following scenarios were chosen. To obtain confidence intervals of the smoothed esti- mates, Buckland et al. (1992) combined smoothed esti- mates and bootstrap replication using the following procedure. First, they obtained 79 bootstrap estimates of the abundance index for each year. Next, they built bootstrap replicates of the series of estimates by tak- ing, for example, the first bootstrap estimate for each year to obtain the first replicated series. They smoothed each replicated series, thereby obtaining 79 smoothed estimates for each year. Finally, they sorted the smoothed estimates within each year and obtained 85% confidence intervals based on the percentile method (Buckland 1984). The median of the smoothed boot- strap replicas is considered to be the best smoothed estimate. The confidence intervals thus calculated allow a di- rect comparison between estimates. If the confidence intervals for two estimates do not overlap, then they are significantly different at a level of -5%. An ex- ample based on estimates of relative abundance for the northern stock of offshore spotted dolphin is shown in Fig. 1. Since the last and first smoothed estimates are too variable, Buckland et al. (1992) recommend excluding them from the comparisons. The significance level is approximate, since it depends on normality of the estimates, homogeneity of the variances of esti- mates, and their independence. The second condition Stable population Population exhibits no trend over the simu- lated period. This scenario pro- vides us with an estimate of the probability of a Type-I error de- tecting a trend when, in fact, there is none, or obtain- ing a "false positive." Under this scenario, a percent- age of detected trends close to 5% would be expected for a test based on a significance level of 5%. Rapid decline Population remains at a constant level for a period of time, and then declines sharply over a period of 3yr to 50% of its previous level. After the decline, the population recovers at a rate of 5% per yr. Steady decline Population declines exponentially at an annual rate of 10%. Steady cycle Population follows a sinusoidal change over the simulated period, completing one cycle over the 25 yr. Amplitude of the cycle is 30% of the original population size. These scenarios are intended as a means of high- lighting properties of the different tests and not as an exhaustive list of possible situations. Each scenario was replicated 100 times for different combi- nations of sources of variation. Variability in the esti- mates was assumed as coming from two different sources: (1) Interannual variability, resulting in a point estimate for each year t of I -Np' ■Hit — i> t e ) 186 Fishery Bulletin 91 12), 1993 where z is a random variable distributed as N (0, a 2 ), and (2) precision of the estimate, represented in the distribution of the simulated bootstrap replicas for year t, I bt , as I bt = I t e v ,forb=l, . ..,79, where v is a random variable distributed as N (0, e 2 ). The rationale for a setup with independent control of two sources of variation and a lognormal error struc- ture lies in the properties of the estimates and in the fact that the linear and smoothed tests deal differ- ently with both components of variation. The choice of a lognormal error structure can be justified by consid- ering that the abundance estimates are naturally con- strained to be positive. The choice of error structure for the simulation can be further justified by an analy- sis of the available TVOD estimates (Fig. 2), which shows that the main target stocks tend to have con- stant CVs, particularly in recent years when observer coverage of the tuna fleet increased and more in- formation was available for abundance estimation. There is considerable variability in some stocks, due to changing levels of targeting from the purse-seine fleet that result in unequal sample sizes from year to year. The argument for including two sources of variation in the estimates is based on the possibility that actual relative abundance indices are affected by random bi- ases from year to year. Under standard assumptions, o 2 and e 2 should be equal. However, estimates of dol- phin abundance may exhibit an additional variability, represented in this setup as o 2 >e 2 . This can be attrib- uted to randomly-fluctuating biases, such as those in- troduced by changing environmental conditions or differences in the way the purse-seine fishery oper- ates. It is important to separate these two components since, for example, in the case of the linear test, the results are affected only by the variability represented byo 2 . Therefore, it was assumed in the simulation that estimates are lognormally distributed around the un- derlying trend with a constant coefficient of variation. Figure 3 illustrates one simulated series for each of the different scenarios. The simulations were repeated for different combinations of CV a and CV f , the coeffi- cients of variation for both sources of variation. To compare the test for trends, two diagnostics were used. / / SCOM / t : '■ ' ■' ; * \ v ■' '■ / * : \ / » A ; i ~/K" 1 / 1 ^ NCOM .-■ i CCOM Figure 2 Coefficients of variation in relative abundance estimates of dolphin stocks in the eastern Pacific Ocean as a function of time. NOFF=northern stock of the offshore spotted dolphin Stenella attenuata; SOFF=southern stock of the offshore spot- ted dolphin; EAST=eastern stock of the spinner dolphin Stenella hngirostns; WHBL=whitebelly stock of the spinner dolphin; NCOM=northern stock of the common dolphin Del- phinus delphis; CCOM=central stock of the common dolphin; SCOM=southern stock of the common dolphin. Number of detected trends For the linear trends, this is the number of 10 yr periods with slopes signifi- cantly different from zero at the 5% significance level. For the smoothed test, it is the number of significant differences between the next-to-last estimate and the estimate 10 yr earlier. In this way, the comparison is based on the same number of tests for each method. Since there are 25 yr simulated in each replica, a total of 1500 tests were carried out in the simulation of each scenario. Anganuzzi: Detecting trends in dolphin abundance indices 187 ST4BLF. POPULATION ff- ? X SI * f^trr Ifl Hf -H RAPID DECLINE i i \ s i A\ jr.*'.. < '•" ~ "J £ fit 1 1 irlvt u i . . u Li^t 1 2 3 4 5 6 7 8 9 101112131415)6171819 20 2122 232425 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Yeor STEADY CTCLE 1 2 3 4 5 6 7 8 9 1011 1213141516171819 20 2122 23 24 25 1 2 3 4 5 6 7 8 9 1011 1213141516171819 20 2122 23 2425 Figure 3 Example of a simulated series of relative abundance estimates for each of four scenarios. Solid line represents true underlying trend in the population. Dashed line connects simulated point estimates. Distribution of simulated bootstrap estimates is represented by the distribution of dots. Ratio between estimates As a way of assessing how well each method describes the underlying trend in the population, an estimated rate of change was ob- tained. For the smoothed test, this is the ratio of two smoothed estimates separated by 10 yr. For the linear test, it is the ratio of the corresponding estimates cal- culated from the linear regression. These estimated rates of change were then compared with the true rates of change and the discrepancies summarized as aver- age absolute error. Correlation in the smoothed estimates One of the problems of the smoothed test is that the smoothing procedure induces a correlation between es- timates. This lack of independence affects the results of the comparison between estimates close in time, and it is therefore important to assess the magnitude of this correlation and how it is reduced as the separa- tion in time between estimates increases. To investi- gate this, the following Monte Carlo procedure was carried out on the series of relative abundance esti- mates for dolphin stocks in the EPO reported by Anganuzzi et al. (1992). 1 For each year, 79 estimates were sampled with replacement from the distribution of bootstrap esti- mates of relative abundance. The 79 estimates were available from the standard bootstrap procedure used to estimate confidence bounds in the relative abun- dance estimation (Anganuzzi & Buckland 1989). 2 Each of the 79 trajectories obtained in the previ- ous step were smoothed, and 85^ confidence limits for the resulting smoothed estimates were obtained based on the percentile method. This step is essentially the application of the smoothed test. 3 Steps 1 and 2 were repeated 100 times, therefore obtaining 100 estimates of the lower and upper confi- dence bounds for the smoothed estimates for each year. Fishery Bulletin 9 1(2). 1993 4 A correlation matrix between years for both up- per and lower limits of the confidence bounds was esti- mated on the basis of results of the previous step. 5 Estimates of correlation as a function of the dis- tance in time between estimates were obtained. This was done by averaging over all correlation coefficients between estimates separated by a given number of years, that is, by taking the average of elements in the subdiagonals of the correlation matrix obtained in step 4. beginning of the period both the linear and smoothed tests have similar proportions of detected trends, close to the nominal significance level. As the underlying trend in the period included in the tests departs from linearity, the smoothed test tends to outperform both linear tests. For CV n =CV ( =0.2, the smoothed test indi- cates a maximum of almost 80% significant trends in Results Number of detected trends The results of this analysis are shown in Figs. 4-7, as the num- ber of detected trends each year in 100 simulations for the differ- ent scenarios. The underlying trends are shown on an arbitrary scale to relate changes in perfor- mance of the tests to changes in population trajectory. Stable population This sce- nario can be used to assess the actual level of significance of the tests. An ideal procedure for de- tecting trends would indicate sig- nificant trends in -5% of the tests under this scenario, given that the significance level is set at 5%. Results from the simula- tions are shown in Fig. 4. Both linear tests show an actual sig- nificance level close to the ex- pected value. These results were relatively robust to the different values of CVs tested. The smoothed test also performs well in all cases, except when interannual variation exceeds the precision of the estimate. For ex- ample, for CV„=0.2 and CV E =0.3, the percentage of detected trends was -10''. Rapid decline In this scenario, different trade-offs of the tests are illustrated by their perfor- mance along the simulated pe- riod (Fig. 5). For CV =CV £ , at the Smoothed test 10 yf linear led 6 yf linear lest cvO.3 cv € :0.2 cv, : 0.2 -0.3 I 15 Year 20 Smoothed tesl 10 yr linear lest S yr linear test Figure 4 Percentage of trends detected in 100 simulations of the 'stable population' scenario for two levels of precision: (top) CV„=0.2, and (bottom) CV„=0.3. The two lines for each smoothed test represent different levels of interannual variation, CV,. Broken line ( ) represents underlying trend in the population on an arbitrary scale. Anganuzzi Detecting trends in dolphin abundance indices 189 comparison with <60% for the linear test. In the recov- ery phase of the trajectory (starting when the tests cover the period between years 12 and 22), the linear test improves its performance relative to the smoothed test. In absolute terms, the performance of both tests seems to be poor between years 20 and 23, but this is the result of the small difference between the first and =V°2 8 - O _ w C © TJ O © (O TS CV E . This re- sult seems to be a consequence of higher Type-I error probabili- ties, as suggested by the higher number of detections in years 11 and 12. Steady decline The power of both types of tests improves under this scenario relative to the previous one, due to the smoother nature of the underly- ing trend (Fig. 6). For CV =CV E the smoothed test outperforms the linear tests for all levels of variability. The percentage of significant trends detected by the smoothed test ranges from over 95% for CVs of 0.20 to -80% for a CV of 0.30. The power of the 10 yr linear test seems to be more affected by increasing vari- ability in the estimates, falling to -50% detections for CV=0.40. For CV a >CV E , the power of the smoothed test seems to increase although, as before, this is prob- ably the result of greater Type-I error rates. The 5 yr linear tests show low power for all levels of variability. Steady cycle The results of this set of simulations are very simi- lar to those from the 'rapid de- cline' scenario (Fig. 7). Overall performance for both tests im- 190 Fishery Bulletin 91|2). 1993 proves relative to that scenario, due to the smoother underlying trend, reaching a maximum of 90% for the smoothed test for CV=0.2. This performance falls rapidly, as the amount of variabil- ity increases, to a maximum of slightly over 20% when CV=0.40. For CV„>CV E , the smoothed test again shows an apparent increase in power related to high Type-I error probabilities. Once more, the 5yr linear test shows very low power throughout the series. Ratio between estimates The purpose of this analysis is to assess sensitivity of the estimated rates of change, obtained by smoothing or linear procedures, to departures from linearity of the underlying trend. It also mea- sures the ability of the procedures to reconstruct the true underly- ing population trajectory. The re- sults of this analysis are shown for each of the simulated trajec- tories in Fig. 8. Results indicate that the estimated rates of change obtained through the smoothing procedures are better, in terms of the average absolute error, than estimates obtained by any of the linear methods, even when the simulated trend is linear ('stable' scenario, Fig. 8, page 192). Esti- mates from the 5 yr linear regres- sions are consistently worse than estimates from the 10 yr regres- sions, as expected due to the greater number of points on which the latter is based. The only ex- ception is for the scenario with cyclic fluctuations, where the 10 yr linear estimates are poorer than the 5 yr estimates. This is a consequence of the period of the sinusoidal cycle in the underlying population that can be better approximated by a linear model over a short period of time. For longer periods in the cycle, 10 yr linear estimate should improve, as suggested by its performance in the scenario with an expo- nential decline. Results for the smoothed procedure are consistent over the sets of scenarios, indicating its C-V0.2 8 - cu -0° cv c :03 Smoothed lest N — -"■" """ cv < ' ° 2 tOytlineef test O _ CD \^/ X \Av^X^ =v°- 3 c £ TJ o _ « to I 0) ■D s O N. 0) \ 1 ' " S. V^ 0. \.^ o _ ~~~"~~"~'^. V ** ,-'' \ S **cv € :0.2 - ™5*^j. __-, '' "-. cv c :0.3 o - l I l l 1 J 5 10 15 20 25 Year CV a :0.3 8 - ______^ smoom.oi.st cv : 3-^ ,- — -_ ^ _„...„ — __ 10 yr linear test /*V / ""'"'-- w -*'"\ cv € -0.4' o _ CO C X" VX/N 'V 3 T? O _ I © /V^ ^^—^ _~ % O CD - J '" V\ 04 I ° - § \ 3 Q. O _ '~-.^ CM ___TT^T»-s~^_ --> -', , ~~.cv,:0.3 o - I I l l l l 5 10 15 20 25 Year Figure 6 Percentage of trends detected in 100 simulations of the 'steady decline' scenario for two levels of precision: (top) CV„=0.2, and (bottom) CV„=0.3. The two solid lines for each smoothed test represent different levels of interannual variation, CV r Broken line ( 1 represents underlying trend in the population on an arbitrary scale. robustness to departures from linearity in the popula- tion trajectory. Correlation between smoothed estimates Results from this analysis are shown in Figs. 9 and 10 (pages 193, 194), which indicate that the correlation Anganuzzi. Detecting trends in dolphin abundance indices CV o - = 0.Z I - Smoothed tost ^^—v o _ GO 5yr finaar lest / \ ■D s ■o o ID (O s /— x if X centage ol 40 |v^ \/ Ol Q. If / \ \\\ . c. ( :0.Z O CM o - c» £ :0.2 1 1 I 1 I — ' 3 5 10 15 20 25 Year 0^0.3 o o - Smoothed lest o _ Percentage of detected trends 20 40 60 1 I i //I-''''' 0\ i c» t -0.3 o - cVO.4 L.J!' " = :".-''-'^_V---^.^_ S^NS-^ cv e :0.4 cv 6 : 0.3 ^ ^ ( ' l l l | 5 10 15 20 25 Year Figure 7 Percentage of trends detected in 100 simulations of the 'steady cycle' scenario for two levels of precision: (top) CV o =0.2, and (bottom) CV a =0.3. The two solid lines for each smoothed test represent different levels of interannual variation, CV t . Broken line ( ) represents underlying trend in the population on an arbitrary scale. A characteristic of the smoothed test is that, while the smoothing procedure induces a correlation between estimates, the correlation across years be- tween fixed percentiles of the dis- tribution of smoothed estimates is lower. This is a result of the (implicit) sorting of estimates that removes some of the depen- dency. To illustrate this, the average correlation between smoothed estimates obtained be- fore sorting is also shown in Figs. 9 and 10, suggesting that the reduction in correlation due to sorting is -30% for estimates close in time. Discussion declines rapidly as distance between the estimates in- creases, approaching very low values as the separa- tion between estimates exceeds 4 yr. This suggests that the validity of the test will not be seriously compro- mised by the correlation induced from the smoothing procedure, when the comparison is carried out on esti- mates that are separated by at least 4yr. In summary, the two types of tests represent different ways of looking at the data, and a com- parison between them based on the same criteria only partially reflects these differences. The smoothed test provides more in- formation since it allows a graphic comparison between in- dividual estimates. It is based in a more robust assumption about the underlying trend of the popu- lation, since it assumes only that the change has been smooth over a short time-period. It also in- corporates the lack of precision of the estimates in a much more effective way than the linear test. In some cases, however, the linear test might be more suit- able. For example, if the amount of interannual variation is low relative to the precision of the estimates, or if changes are closely approximated by a linear function, the linear test should perform better. Also, if large changes in population size occur over a very short time-period, smoothing the series will tend to underestimate the rate of change. Despite limitations of the comparison, the results from simulations indicated that the smoothed test out- performs the linear test in most situations. The only exception is the tendency to detect spurious trends 192 Fishery Bulletin 91(2). 1993 Stable population Rapid decline Average absolute error 3 2 4 6 8 1.0 5- year linear ^^^ 10-year linear ^ — •""""^ „_ — - Smoothed Average absolute error 00 02 04 06 08 10 " 5-year linear ^ — 1 0-year linear _ - - Smoothed 1 111! 2 3 4 5 Inter-annual vanatton {cv(b)} Steady decline 02 03 04 05 Inter-annual vanatlon (cv(b)) Steady cycle Average absolute error )0 2 4 6 8 10 ___ -w ^— 5-year linear Average absolute error 00 02 04 06 08 10 _ ** 10-year linear 5-year linear - - Smoothed Average a simulated r r- i 2 3 4 5 Inter-annual variation {cv(b)) Fig bsolute error as a function of different degrees of inter ure 8 annual varia 2 3 4 5 Inter-annual variation {cv(b)) tion, represented by CV„ for each of the four scenarios when the amount of interannual variability exceeds precision of the estimates. In other words, in the in- evitable trade-off between Type-I and Type-II errors, the smoothed test has lower Type-II error rates at the expense of an increase in the Type-I error rate. From the management point of view, this is a safer compro- mise than the one posed by the linear test, since the probability of failing to detect a significant trend is smaller with the smoothed test. The increase in Type-I error rates can be related to the amount of smoothing done by the particular algorithm chosen. An algorithm that would smooth the estimates more would be less prone to this problem, but it would have less power to detect trends in the estimates in certain situations. Such an algorithm would also induce more correlation between smoothed estimates, and the separation in time between them would have to be greater in order not to compromise validity of the comparisons. An al- ternative would be a smoothing algorithm that can adaptively change the amount of smoothing done on the estimates, either by cross-validation techniques or by controlling the amount of smoothing through incor- porating auxiliary information, such as birth and death rates, in the procedure; in other words, by building a model of the population dynamics. Acknowledgments I would like to thank Bill Bayliff, Steve Buckland, Martin Hall, and Tim Gerrodette for their valuable comments. Anganuzzi Detecting trends in dolphin abundance indices 193 Northern offshore spotted dolphin Southern offshore spotted dolphin - Unsorled smoothed estimate Lower confidence bound Upper confidence bound ■- — - ' Unaorted smoolbed estimate Lower confidence bawl Upper confidence bound Distance between estimates (in years) Northern and southern offshore combined Distance between estimates (in years) Eastern spinner dolphin estimate bound - — — - Upper confidence \ **'. \ bound \ v.. N. ^^^__^^ - — Unsorted smoothed estimate --- Lower confidence botxid - Upper confidence bound Distance between estimates (in years! Distance between estimates (in years) Figure 9 Correlation between smoothed estimates as a function of the separation in years between estimates for four stocks of dolphin in the eastern Pacific Ocean: Offshore spotted Stenella attenuata and spinner dolphin Stenella longirostris. Citations Anganuzzi, A.A., & S.T. Buckland 1989 Reducing bias in estimated trends from dolphin abundance indices derived from tuna vessel data. Rep. Int. Whaling Comm. 39:323-334. Anganuzzi, A.A., S.T. Buckland, & K.L. Cattanach 1991 Relative abundance of dolphins associated with tuna in the eastern tropical Pacific, estimated from tuna vessel sightings data for 1988 and 1989. Rep. Int. Whaling Comm. 41. Anganuzzi, A.A., K.L. Cattanach, & S.T. Buckland 1992 Relative abundance of dolphins associated with tuna in the eastern tropical Pacific in 1990 and trends since 1975, estimated from tuna vessel sightings data. Rep. Int. Whaling Comm. 42:541-546. Buckland, S.T. 1984 Monte Carlo confidence intervals. Biometrics 40:811-817. Buckland, S.T, & A.A. Anganuzzi 1988 Trends in abundance of dolphins associated with tuna in the eastern tropical Pacific. Rep. Int. Whal- ing Comm. 38:411-437. Buckland, S.T, K.L. Cattanach, & A.A. Anganuzzi 1992 Estimating trends in abundance of dolphins asso- ciated with tuna in the eastern tropical Pacific Ocean, using sightings data collected on commercial tuna vessels. Fish. Bull, U.S. 90:1-12. Edwards, E.F., & P.C. Perkins 1992 Power to detect linear trends in dolphin abun- dance: Estimates from tuna-vessel observer data, 1975-89. Fish. Bull., U.S. 90:625-631. Gerrodette, T. 1987 A power analysis for detecting trends. Ecology 68:1364-1372. IWC (International Whaling Commission) 1992 Report of the sub-committee on small ceta- ceans. Rep. Int. Whaling Comm. 42:178-228. 194 Fishery Bulletin 91(2). 1993 - Unsorted smoothed estimate - Lower confidence bound Upper confidence bound Distance between estimates (In years) Central common dolphin Distance between estimates On years) Southern common dolphin Unsorted smoothed estimate Lower confidence bound Upper confidence bound Unsorted smoothed estimate Lower confidence bound — Upper confidence bound Distance between estimates {in years) Distance between estimates (In yearsl Figure 1 Correlation between smoothed estimates as a function of the separation in years between estimates of four stocks of dolphin in the eastern Pacific Ocean: Spinner dolphin Stenella longirostris and common dolphin Delphinus delphis. Velleman, P.F., & D.C. Hoaglin 1981 Applications, basics and computing of exploratory data analysis. Duxbury Press, London. Wade, P.R., & T. Gerrodette 1992 Estimates of dolphin abundance in the tropi- cal Pacific: Preliminary analysis of five years of data. Rep. Int. Whaling Coram. 42:533-539. Abstract. — We compared sam- pling performance of four nets and two aggregation devices for larval and pelagic juvenile coral-reef fishes. The six sampling devices were de- ployed simultaneously over three nights near a coral reef at Lizard Island, northern Great Barrier Reef, Australia. The resulting 83 samples captured 57,701 larval and pelagic juvenile fishes of 70 families (exclud- ing clupeoids which were not con- sidered in this analysis). The bongo net took the most families, and the light-trap the fewest. In all meth- ods, a few families dominated the catch. Dominance was least in the Tucker trawl catches and greatest in light-trap catches, where poma- centrids constituted 939c of the catch. Composition of catches was similar for the four nets. Catches from the light-trap were markedly- different from those taken by net; catches taken by light-seine showed similarities to those taken by both net and light-trap. For four abun- dant families (Apogonide, Gobiidae, Lutjanidae, Pomacentridaei. the bongo net gave the overall highest density estimates, although those from purse-seine were frequently equivalent to bongo-net estimates. The Tucker trawl provided the low- est density estimates in most cases. Catches of bongo, neuston, and seine nets were similar in size structure and were dominated by small lar- vae; overall, however, bongo nets col- lected the greatest size-range of fishes. The Tucker trawl did not col- lect small larvae well nor did it col- lect significantly greater densities of large larvae and pelagic juveniles than the bongo net. Fishes collected by aggregation devices were gener- ally larger than those taken by net, and light-traps caught very few fish <5mm. Light-traps collected greater numbers of large pomacentrids (>6mm) than other methods. In an extended sampling period of five nights, both aggregation devices showed obvious peaks in the den- sity of large pelagic pomacentrids and mullids; these patterns were not detected by the nets. A comparison of towed nets, purse seine, and light-aggregation devices for sampling larvae and pelagic juveniles of coral reef fishes* John H. Choat Department of Marine Biology, James Cook University of North Queensland Townsville, Queensland 4811. Australia Peter J. Doherty School of Australian Environmental Studies, Griffith University Nathan, Queensland 4111, Australia Present address Australian Institute of Marine Science. PMB No 3, Townsville M.C.. Queensland 4810. Australia Brigid A. Kerrigan Department of Marine Biology, James Cook University of North Queensland, Townsville. Queensland 4811, Australia Jeffrey M. Leis The Australian Museum, PO. BoxA285 Sydney South, NSW 2000, Australia Manuscript accepted 24 February 1993. Fishery Bulletin, U.S. 91:195-209 ( 1993). Almost all species of marine teleost fishes have a pelagic phase in the early part of their life history (Moser et al. 1984). Size, morphology, and behavior of larval and pelagic juve- nile phases vary greatly (Moser 1981), and this makes accurate sam- pling of these fishes problematical (Murphy & Clutter 1972, Frank 1988, Suthers & Frank 1989, Brander & Thompson 1989). The problem is ex- aggerated in tropical waters due to high taxonomic and developmental diversity and the presence of many demersal species with extended pe- lagic phases (Leis & Rennis 1983, Leis & Trnski 1989, Leis 1991b). Studies of the pelagic phase can pro- vide important information on popu- lation biology of reef fishes. Despite its brevity, the high mortality and dispersion characteristic of this phase can have important demographic con- sequences for many species (Victor 1986). There is now a widespread in- terest in the process of recruitment in coral reef fishes (Doherty & Wil- liams 1988, Warner & Hughes 1989), and sampling techniques which cover the full size-range of the pelagic phase are needed. A number of different methods are available to sample this complex as- semblage of early-life-history stages, including towed nets, purse-seines, and various types of aggregation de- vices which attract fish into collec- tion sites or traps. These methods differ in their method of deployment and capture, and each has its own set of advantages and disadvantages. All have biases in number, identity, and sizes of pelagic fishes collected (Clutter & Anraku 1968, Clarke 1983 and 1991). For the pelagic phase of reef fishes, there have been few at- tempts to evaluate the relative bias of different sampling methods. Re- cent studies have provided informa- tion on the comparative performance Contribution of the Lizard Island Research Station. Authorship alphabetical. 195 196 Fishery Bulletin 91|2). 1993 of nets and light-traps (Gregory & Powles 1988), nets and plankton pumps (Brander & Thompson 1989), and towed nets and purse-seines (Kingsford & Choat 1985), but have dealt with the less-diverse fauna of temper- ate waters. The purpose of this study was to compare several types of towed and seine nets and an automated light- trap (Doherty 1987) in terms of taxa, numbers, and sizes of larvae and pelagic juveniles of coral reef fishes captured. These methods represent the range of sam- pling devices currently used to collect larval and pe- lagic juvenile fishes. For the towed nets, we used di- mensions and mesh size normally employed to sample larval and pelagic juvenile fishes. We used designs of purse-seine and light-trap which had been subject to thorough field testing (Kingsford & Choat 1985 and 1986, Kingsford et al. 1991, Doherty 1987). For each sampling device we obtained the following informa- tion: ( 1 ) Taxonomic composition of samples at the level of family; (2) patterns of density and size structure in selected taxa; and (3) temporal patterns in the density of selected taxa over short time-periods. The program also provided information on the logistic constraints associated with each sampling method. Our findings will be useful to those designing sam- pling programs for larval and pelagic juvenile stages of demersal fishes in tropical and other areas, and should have some generality because the taxa sampled included a wide variety of body shapes and swimming capabilities. Among the taxa studied are families of great importance in coral reef ecosystems as adults (Apogonidae, Atherinidae, Callionymidae, Gobiidae, Labridae, Pomacentridae), and several are also impor- tant in commercial, sport, or subsistence fisheries throughout the tropics (Carangidae, Lethrinidae, Lutjanidae, Mullidae, Nemipteridae, Platycephalidae, Scaridae). All are abundant in ichthyoplankton sam- ples in tropical coastal areas, especially in the Indo- Pacific. Materials and methods Sampling and identification procedures We sampled at 150-600 m off the fringing reefs at Watsons Bay on the NW side of Lizard Island in the lagoon of the northern Great Barrier Reef, Australia (145°26'E, 14°40'S). Water depth was 20-30 m over a sandy bottom (Fig. 1). This site was chosen for its proximity to the logistic support offered by the Lizard Island Research Station, a base for much work on the pelagic phase of coral reef fishes (Leis 1991b). Also, it offered relatively sheltered conditions from the 15- 25 kn southeasterly winds present during the sampling LIZARD ISLAND Prevailing Trade Wind >^ 500m. Palfrey Island sS-A Nv />-'' \ / y-^*\. _^^"\ ' Mrs. Watsons _ \« ^A_^^ -„.-- J 5<* Ba y \ '"'"' / Kf ^ '•'<' Cr\ Purse C \^ i. 1. • -./ \ Seine ^^\\. ^ Light Traps | ?7 \ Tow Path^\^>~^ >v , x / ,- -. f w 5O0m, < Figure 1 Lizard Island, Great Barrier Reef, Australia, showing loca- tion of study area and position of sampling sites for light- traps, towed nets, and purse-seines at Watsons Bay. Coral reefs are shown as broken lines. Lizard Island ( 145 : 26'E. 14°40'S) is located 30 km off the eastern coast of mainland Australia. period. This was particularly important for the conti- nuity of sampling over a number of nights. We sampled on the nights of 2, 3, 5, 6, and 7 Decem- ber 1986, starting at a minimum of 1.25 h after sun- set. Sampling never continued past 0200 h. New moon was on 2 December 1986. Nocturnal sampling reduces potential bias due to vertical distribution because ichthyoplankton show little vertical stratification at night in the study area (Leis 1986, 1991a). In addi- tion, the nets should operate at peak efficiency at night due to lessened visual avoidance. Finally, the aggrega- tion devices are effective only at night because they depend on self-generated light to attract fishes. We concentrated our analyses on data from 3, 5, and 6 December because we were able to take and process all planned samples from all gears only on these nights. For some gears, it was possible to examine temporal trends over the full sampling period. Six different sampling devices were deployed each night. Three nets were towed from the 14 m catama- Choat et al Comparison of ichthyoplankton sampling methods 197 ran RV Sunbird at lm/s along a fixed 1km path. The towed nets were fitted with flowmeters and were washed with pumped seawater. Details of each collec- tion device are as follows. 1 A neuston net of mouth dimensions 1.0x0.3 m with 0.5 mm mesh was rigged to sample water between the bows of the catamaran. Typically, the net sampled to a depth 0.1m and filtered 187-312 nvVtow. Four tows were taken per night. 2 A bongo net (McGowan & Brown 1966) of 0.85 m mouth diameter per side, and with 0.5 mm mesh, was towed from an "A'-frame at the stern. The RV Sun- bird draws lm, and the net was towed so its top was lm below surface and on the vessel's centerline in wa- ter which had not been disturbed by the passage of its twin hulls. The volume of water filtered for each side of the net was 498-673 m 3 tow. Samples from only the port-side net were analyzed. Four tows were taken per night. 3 A Tucker trawl (Tucker 1951) with nominal mouth dimensions of 2x2 m and of 3mm mesh was towed in the same position as the bongo net. At a towing speed of 1 m/s, a diver estimated that the bottom bar of the net trailed the top bar by -0.5 m, so the effective mouth area was -3.8 m 2 . Between 3240 and 4570 m 3 of water were filtered per tow. Four tows were taken per night. Both the bongo net and the Tucker trawl used the same depressor. Time constraints and the logistics of rigging and deploying each net precluded randomising the order of bongo and Tucker trawl tows, so they were taken in blocks of four, with the order alternating from one night to the next. Neuston net samples were taken during the Tucker trawl tows. 4 A plankton mesh purse-seine of 14x2 m (Kingsford & Choat 1985) of 0.28 mm mesh was used to take samples of -32 m 3 each. This estimate was based on the ideal cylinder of water enclosed by the net at the beginning of pursing and made no allowance for herd- ing of fishes during deployment or loss during pursing. There was no estimate of variation in the volume en- closed by the net sets. The net was deployed from a 4 m dinghy adjacent to the northern end of the tow path (Fig. 1). Wind conditions precluded effective de- ployment of this net at greater distances offshore. Two to four samples were taken per night. 5 Two automated light-traps (Doherty 1987) were de- ployed from an anchored boat adjacent to the center of the tow path and -700 m from the purse-seine site. Traps were positioned at ~10m apart. Entries into the trap were at 0.5-1 m below surface. The second trap began to sample 30 min after the first, and both traps sampled for hourly intervals, resulting in continuous sampling in overlapping, 1 h segments. The trap de- ployment was staggered to allow for clearing and pro- cessing of each trap after the 1 h fishing period. Eight to nine 1 h light-trap samples were taken per night. 6 A battery-powered fluorescent light source identi- cal to that in the trap (Doherty 1987) was deployed from a second boat anchored at the purse-seine site. After 1 h in the water, the light was set adrift and the water around it immediately sampled by the same purse-seine used in (4) above. Our estimates of what was attracted to the light included only those indi- viduals that were within -2 m (i.e., radius of the seine at pursing) of the light at the time of seining. Four to five light-seine samples were taken per night. Purse- seine (no light, (4) above) and light-seine samples were interspersed during the night. Our goal was to sample simultaneously using six methods in the same location over several nights, so as to avoid confounding comparisons of methods with temporal or spatial variation. The purse-seine, light- seine, and light-trap samples were taken throughout the nightly sampling period. At the same time, the RV Sunbird sampled with the towed nets. Logistic prob- lems required two compromises in this program. Bongo tows and Tucker trawl tows (and simultaneous neus- ton tows) were done in sequential blocks of four each night as discussed in (3) above. The purse-seine and light-trap samples were taken 700 m apart because it was not possible to duplicate these devices and thus randomize their positions. The RV Sunbird tow track covered the area between these two. Fishes from the towed nets, purse-seines, and light- seines were immediately fixed in 10% formalin seawa- ter. Samples from the light-traps were maintained alive until returned to the Research Station where they were subsequently fixed in 100% ethanol or 10% formalin seawater. All fish were transferred to 70% ethanol for at least a month prior to measurement. For light-traps and light-seines, density is expressed as number per sample. Catches from the towed net and purse-seine collections were standardized to the number of fishes/1000 m 3 on the basis of flowmeter records or purse-seine geometry. All fishes were removed from samples and identified to family following Leis & Rennis (1983) and Leis & Trnski ( 1989). Standard lengths were measured to the nearest 0. 1 mm using a Bioquant software package that allows for measurement of enlarged camera lucida im- ages offish and accommodates curvature of specimens. The accuracy of electronic measurement was monitored by measuring subsamples manually with calipers and eye-piece micrometers. In a few samples with very large numbers of certain taxa such as gobiids, the catch was subsampled and a minimum of 10% of the sample mea- sured. For some analyses, fishes were divided into small (<6mm) and large (>6mm) size-groups. This was done because, on the basis of results reported here, the light- 198 Fishery Bulletin 91(2), 1993 trap captures few larvae <6mm, and we wished to compare density estimates among gears for the sizes of fishes captured by the light-trap. Damaged fish (-3% of total) were excluded from the length analysis. The terminology of early-life-history stages of fishes is complex and ultimately arbitrary, whether based on morphological or ecological criteria (Kendall et al. 1984, Kingsford 1988, Leis 1991b). We were primarily inter- ested in taxa of which the adults are benthic on coral reefs, but did not want to exclude semipelagic reef- associated taxa by use of an ecological term like 'presettlement', nor did we wish to exclude partially- or fully-transformed but still pelagic individuals of benthic taxa by the use of a morphological term like 'larva'. Therefore, we use the terms 'larvae' and 'pe- lagic juveniles' for the fishes collected during this study, or refer to them collectively as 'pelagic fishes'. Larval, transforming, juvenile, and adult clupeoid fishes of several types (including Spratelloides spp., Dussumeria sp., Stolephorus sp., and probably Her- klotsichthys sp.) were captured in large numbers, mainly by light attraction. These clupeoid fishes rep- resented a distinct assemblage of fishes with a differ- ent age and size structure and adult habitat than the reef species of primary interest to us. These clupeoids are not considered here, but will be dealt with in a separate publication. Reduction of data sets and analytical procedures Sampling produced a data set comprising 70 families of fishes (exclusive of the Clupeidae and Engraulidae) collected from the sampling nights of 3, 5, and 6 De- cember by six methods. For ease of analysis and un- ambiguous interpretation, it was necessary to reduce the number of families treated. We initially removed from consideration any family which did not consti- tute at least 1% of the catch of at least one method. The removal of taxa of this level of rarity would be unlikely to influence the outcome of the analyses (Green 1979). This excluded 51 families, leaving 19 (referred to as 'abundant families') for analysis beyond simple listing of numbers of families sampled (e.g., Table 1). Relative-abundance information obtained by all six sampling methods for the 19 abundant families was subjected to Principal Component Analysis (PCA) us- ing the variance-covariance matrix. As a check, the same analysis was run incorporating the next 10 most- abundant families; this generated identical patterns. Reducing the data set from 29 to 19 families did not change the resulting pattern. The PCA analysis identified patterns in the complex data set of 19 families sampled by six methods. Many of these 19 families were relatively rare and contrib- uted little to the variation in the data set. A detailed examination of the factors contributing to these pat- terns required factorial analyses such as multivariate analysis-of-variance (MANOVA). These procedures are best carried out with a reduced number of variables, which allows a clearer interpretation of trends in the data. This called for a further reduction in the number of families analyzed. To achieve this reduction, the data set of 19 families collected by nets was subjected to a PCA, which iden- tified the taxa that contributed most substantially to the variation in the data set. This PCA identified apogonids, atherinids, gobiids, lethrinids, mullids, and pomacentrids as major contributors (95.2%) to the variation in the data set. These six taxa were used in a MANOVA. This design provided sufficient degrees of freedom for testing and interpreting the significance of method and night of sampling. The analysis was carried out on samples from nets only. For graphic display of trends in sampling by nets, the eight most-important taxa from the PCA were de- picted. These were apogonids, atherinids, gobiids, lethrinids, lutjanids, mullids, pomacentrids, and labrids. Labrids were included in this group at the expense of schindleriids, as they were an abundant reef-associated taxon of considerable interest to reef fish biologists. This substitution did not affect the cu- mulative variance accounted for by the eight families. Unlike nets, aggregation devices did not allow for adjustment offish densities to a common volume. More- over, aggregation devices collected a different set of fishes. An additional PCA run on light-trap and light- seine data identified atherinids, gobiids, labrids, lethrinids, mullids, and pomacentrids as taxa, which explained over 90% of the variability in the data set. The families selected showed a strong relationship to the overall abundance ranking, although two relatively rare taxa (lethrinids and mullids) were included. Aggregation devices sample an unknown volume of water. Because catches by aggregation devices could not be standardized to number offish per unit volume, we made separate comparisons of nets and aggrega- tion devices. The variables used were mean number/ 1000 m a for nets, and mean number/sample for aggre- gation devices. A factorial analysis was designed to test for differences in sampling method (fixed) and time (random). For factorial analyses, residual analysis was performed (Snedecor & Cochran 1980) to check assump- tions of normality and homogeneity of variance. Taylor's Power Law (Taylor 1961) was used to determine the appropriate transformation. Canonical Discriminant Analysis and Tukey's Stu- dentized Range Test (HSD) were used to display the differences detected. For MANOVA, the multivariate test statistic (Pillai's Trace) was used because it is Choat et al.: Comparison of ichthyoplankton sampling methods 199 less likely to involve Type-I error and is more robust to heterogeneity of variance than comparable tests (Green 1979). All analyses were performed using SAS Version 6 (SAS 1987). A more subjective procedure was used to select taxa for size-frequency measures. For meaningful compari- sons, it was necessary to select taxa that were well represented in the collecting devices and that covered a reasonable size-range (>8mm) within each method. Apogonids, gobiids, lutjanids, and pomacentrids met these criteria and also accounted for over 95% of the variation in the main data set from net sampling. Catches for nets and aggregation devices were analyzed separately. For net catches, density was expressed as mean number/ 1000 m ! within 2 mm size-classes among the different methods and compared by one-way ANOVAs. With aggregation devices, the variable was the number of fish per sample and comparisons were made by <-tests. Results The 83 samples contained a total of 57,701 fishes of 70 families, excluding clupeoids (Table 1). Table 2 lists families which con- stituted at least 1% of the individuals taken by any sampling method and records their size-ranges by method. We refer to these as 'abundant families'. Taxonomic composition and size structure of the samples There were marked differences in taxonomic com- position of the samples among methods. The bongo net collected the largest number of families overall (Table 1), including all of the abundant families and a wide size-range within most families (Table 2). The light-trap collected the fewest families overall and only Table 1 Number of samples, total individuals, and numbers of families of fishes (clupeoids excluded) taken by six sampling methods on the nights of 3, 5, and 6 December 1986 off Lizard Island, Great Barrier Reef. Volume of water sampled by aggregation devices is unknown. Volume of Sampling Number of Number of water sampled Number of method samples fish Im'l families Light-trap 26 7624 unknown 20 Seined light 14 2707 unknown 37 Purse-seine 7 812 224 25 Neuston net 12 2418 2861 31 Bongo net 12 43417 6833 63 Tucker trawl 12 723 47100 29 Total 83 57701 — 70 Table 2 Numbers and size ranges of the 19 families of fishes which made up >l% of the catch of at least one method on 3. 5, and 6 December 1986 off Lizard Is land, Great Barrier Reef. Clupeoids are excluded. Size-range in mmSL, and total number of individuals within the taxon [n). Family Sampling method Light-trap Light- seine Bong o net Purse-seine Neuston net Tucker trawl SL n SL n SL n SL /! SL n SL n Apogonidae 5.4-9.3 4 1.6-9.8 211 1.6-15.5 10295 1.6-6.8 86 1.7-6.2 491 2.3-5.1 99 Atherinidae 6.7-19.1 20 7.6-61.7 135 6.0-25.2 14 6.8-24.7 2 16.0-56.3 110 15.2-39.3 36 Bothidae 3.2-5.3 3 1.4-7.7 76 3.0-10.0 10 Callionymidae 1.3-3.5 35 1.1-4.9 1003 1.3-2.9 11 1.6-3.9 94 1.9-4.5 6 Carangidae 1.9-57.4 19 1.8-7.6 1555 1.9-4.0 7 1.8-4.5 63 2.2-14.2 13 Ephippididae 1.7-8.7 81 5.8-7.5 14 Gobiidae 3.7-10.5 235 1.2-17.7 643 1.1-10.1 8386 1.4-8.6 487 1.4-20.3 1207 1.9-9.0 258 Labridae 5.1-8.8 48 1.5-13.1 47 1.6-6.0 876 1.7-5.9 21 2.0-5.3 27 2.2-4.1 9 Lethrinidae 8.4-16.6 45 1.9-18.0 24 1.8-4.7 380 2.6-3.3 3 1.9-4.4 17 2.6-11.3 9 Lutjanidae 2.1-5.2 76 1.8-6.6 2740 2.1-7.4 33 1.8-4.9 105 2.5-8.4 48 Microdesmidae 1.5-4.8 10 2.0^.3 100 2.2-3.2 9 3.3-5.4 6 2.9-6.3 7 Monacanthidae 46.6 1 1.5-23.3 13 1.2-4.6 608 1.9-3.3 3 1.8-3.7 11 2.0-6.3 22 Mullidae 11.2-21.9 51 21.5-39.7 54 2.4-4.9 8 5.1-23.6 2 22.4-30.2 10 Nemipteridae 6.4-9.3 28 1.8-12.3 42 1.5-5.6 1548 1.8-5.2 15 1.6-5.0 75 4.2-4.8 4 Pinguepididae 2.0 1 1.4-6.5 30 1.3-5.6 2838 1.4-4.6 20 1.7-5.5 109 2.3-4.8 9 Platycephalidae 2.1-3.1 6 1.6-8.3 469 2.8-5.5 6 2.4-4.2 3 Pomacentridae 5.3-14.9 7124 1.8-25.1 1248 1.0-14.6 496 1.9-9.4 22 1.8-11.7 30 6.4-14.6 68 Scaridae 1.6-4.4 30 1.7-4.6 136 2.2^.0 34 2.5-7.7 10 Schindleriidae 2.0-16.2 219 3.1-8.3 8 4.1-10.7 25 4.4-17.7 79 200 Fishery Bulletin 91(2). 1993 the larger individuals of most families. Analysis of the catch by method (Tables 1,2) suggests that the appar- ent selectivity of the light-trap reflects size-specific rather than taxonomic biases. The absence of certain taxa from the light-trap during the sampling period may mean that few large individuals were in the sam- pling area. Table 3 shows that, with the exception of bothids, schindleriids and carangids, taxa not caught by the light-trap were represented by relatively small individuals in the catch by other methods. Whether large carangids were present in more than trivial num- bers is unclear. A single 57.4 mm carangid was taken by the light-seine, but the next-largest carangid taken by other methods was 14.2mm. The question of selec- tivity by light-traps must be resolved by more compre- hensive sampling. The light-seine and Tucker trawls captured most of the abundant families in all sizes. The neuston net and purse-seine captured the same abundant taxa, with size-ranges similar to one another. The exceptions were mullids, microdesmids, gobiids, and atherinids, for which the neuston net captured larger individuals. For the mullids and microdesmids, size distributions pro- duced by the two methods overlapped slightly. Catches by all methods were dominated by a few abundant families of fishes. The first five most- Table 3 Comparison of maximum size of the 19 abundant taxa (Table 2). Maximum size captured by light-trap is compared with maximum size captured by five other methods tested on 3. 5, and 6 December 1986 off Lizard Island. Great Barrier Reef. Taxa listed in increasing order of maximum size captured by 'other methods' (maximum size captured by the next-best 'other method"). Taxon Maximum size (mml captured by Light-trap Other methods Callionymidae not caught 4.9(4.5) Microdesmidae not caught 6.3(5.4) Pinguepididae 2.0 6.5(5.6) Scaridae not caught 7.7(4.6) Platycephalidae not caught 8.3(5.5) Lutjanidae not caught 8.4(7.4) Ephippididae not caught 8.7(7.5) Bothidae not caught 10.0(7.7) Nemipteridae 9.3 12.3(5.6) Labridae 8.8 13.1(6.0) Apogonidae 9.3 15.5(9.8) Schindieriidae not caught 17.7(16.2) Lethrinidae 16.6 18.0(11.3) Gobiidae 10.5 20.3(17.7) Monacanthidae 46.6 23.3(6.3) Pomacentridae 14.9 25.1(14.6) Mullidae 21.9 39.7(30.2) Carangidae not caught 57.4(14.2) Atherinidae 19.1 61.7(56.3) abundant families listed in Table 2 accounted for 80% or more of the catch by all methods. The Tucker trawl was the most equitable in terms of abundance distri- butions, and the light-trap the least. However, the rank order of abundant families was not the same for all methods (Fig. 2). The dominant families for all towed nets and the purse-seine were gobiids and apogonids. For light-trap and light-seine the dominant families were pomacentrids, followed by gobiids. Small apo- gonids, although consistently abundant in net samples, were not captured by light-aggregation devices. In light- trap catches, a single family — the Pomacentridae — accounted for 93% of individuals collected. For most collecting methods, there was a high de- gree of consistency among samples. Results of PCA (Fig. 3) showed that samples taken by light-trap were 0) o c D X) c < o Q. c D Bongo Net n=43417 4 ■*! Purse Seine n = 812 ■6k .4 .2 ttCb I 2 3 4 5 6 7 8 9 10 I 2 15 4 10 8 3 6 7 16 Neuston Net n=2418 Light Seine n=2707 Etk I 2 II 3 4 7 6 5 10 10 12114178673 Tucker Trawl n = 723 rm^ 2 12 10 4 II 9 13 5 14 Light Trap n=7624 10 I 17 8 18 6 19 II 20 21 Family Figure 2 Mean proportional abundance i±l SE, vertical axis, shown only upward) and ranked taxonomic categories of fishes (clupeoids excluded) collected by six sampling methods off Lizard Island, Great Barrier Reef on 3, 5. and 6 December 1986. Other sample data are given in Table 1. Key to taxa: 1 Gobiidae, 2 Apogonidae, 3 Pinguepididae, 4 Lutjanidae. 5 Carangidae, 6 Nemipteridae. 7 Callionymidae. 8 Labridae, 9 Monocanthidae, 10 Pomacentridae, 11 Atherinidae, 12 Schindieriidae, 13 Ephippididae, 14 Bothidae, 15 Scaridae, 16 Microdesmidae, 17 Mullidae, 18 Lethrinidae, 19 Synodontidae, 20 Scnmbridae, 21 Blenniidae. Choat et al.: Comparison of ichthyoplankton sampling methods 20! PC 2 + \ \ 0.3 \o\ o; • bongo nels o neuslon nels ▼ tucker trawls O purse seines D light seines ^ light trops ■O. •-..-": .0.1 ■T ;:-■ 1 •0.6 •> T '-. D : «• - : •".- ▼ : • •■'.. t: •• •/• ; ▼ 0.2 Q D 0.4 , ; / d \~ .-- ...0 J?/ ■4 /' PC1 "• T ""'-.. ▼"■- •0.3 Figure 3 Results of Principal Components Analysis on proportional abundances of 19 families of fishes collected by six sampling methods on 3, 5, and 6 December 1986 off Lizard Island, Great Barrier Reef. Principal Components 1 and 2 are plot- ted. Differences between number of replicate samples and number of symbols for each method are due to overlap of some symbols. distinct from net samples, and that samples taken by light-seine were intermediate between net and light-trap samples. Tucker trawl samples were almost completely distinct from bongo, neuston, and seine net samples. Bongo net samples formed a more discrete group than did the neuston and seine net samples. The data sets for size analysis were heterogenous. Therefore, we attempted only to test for differences in density among methods within selected size-classes using single-factor ANOVA(df 3,39; p<0.05). The power of these tests to detect differences among methods was low. For apogonids, gobiids, lutjanids, and poma- centrids, there were sufficient numbers for statistical comparisons across the first three size-classes (i.e., <6mm, Fig. 4). For all four families, density estimates provided by the bongo net were as high as, and in many cases higher than, those provided by the other nets. The Tucker trawl provided the lowest density estimates. For the larger sizes (>6mm), low or zero catches in some size-classes precluded statistical tests in most cases. We compared the Tucker trawl, which is de- signed to capture such large stages with the bongo net. The few tests that were possible show that in no instance did the Tucker trawl provide higher density estimates than the Bongo net (Fig. 4). Two taxa, pomacentrids and gobiids, were sufficiently abundant to allow for comparisons of density by 2 mm size-classes between the aggregation devices. For pomacentrids we tested the 7-15 mm size-classes. Light-traps caught significantly higher numbers of pomacentrids in the 7, 9, and 11mm size-classes than the light-seines (Fig. 4B). The two aggregation devices provided similar estimates of numbers for the 13 and 15mm size-classes (Fig. 4B). The difference in overall density for pomacentrids sampled by light-traps and light-seines is due to the greater number of poma- centrids in the 7, 9, and 11mm size-classes in the light-trap catches. Pomacentrid larvae >14mm were collected by the light-seine on one night only. Although we did not statistically test the gobiid data, the light-seine appeared to collect greater numbers of smaller (<4mm), and the light-trap greater numbers of larger (>8 mm), individuals (Fig. 4B). The light-seine collected few gobiids >6 mm and the light-trap almost no gobiids <6 mm. Sizes of apogonid and lutjanid fishes sampled by the light-seine were similar to those of the purse-seine (Fig. 4C). No lutjanids and only four apogonids were collected by the light-traps. Results of pooled samples from three nights for eight taxa (Materials and methods) by the different nets (Fig. 5) reflect both entry of fish into nets and subsequent extrusion. Most of the fishes taken by all nets were small (Table 2, Fig. 4). Bongo nets consistently provided the highest estimates of density of small fishes, espe- cially gobiids, apogonids, lutjanids, labrids, and lethrinids. This reflects both the low-avoidance and high- retention properties of this fine-mesh net. The purse- seine filtered only small volumes of water, but provided high estimates of density, especially for gobiids, apogonids, and lutjanids (Fig. 4). Extrusion is probably minimal, due to the passive mode of filtering and the very fine mesh of this seine. Neuston nets provided low estimates of density for all families except two that concentrate in the surface layer — atherinids and mullids (Leis 1991a). Density estimates from the Tucker trawl were low for all families, most probably due to the loss of smaller larvae through its large mesh. Both atherinids and mullids, which attained large size (Table 2), were also poorly represented in Tucker trawl catches, possi- bly because the Tucker trawl did not sample the neustonic habitat of these taxa. For aggregation devices, we compared densities of the important families identified by PCA, excepting apogonids and lutjanids which were rare or absent from light-traps. Light-traps collect mainly large indi- viduals, so the samples were subdivided by size 202 Fishery Bulletin 91(2). 1993 APOGONIDAE GOBIIDAE 17 ps 17 PS LUTJANIDAE POMACENTRIDAE S 'andi 3rd 17 ' ps **y 17 PS GOBIIDAE POMACENTRIDAE c. ( ^»6 mmSL (Large ). *0.05>p>0.01;NSp>0.05. Family Size Light-seine Light-trap P Atherinidae S 0.29± 0.22 L 9.36± 1.98 0.65+ 0.25 * Gobiidae S 45.50±10.13 0.12± 0.08 * L 0.43± 0.23 8.92+ 3.98 ns Labridae S 1.57+ 0.62 0.04+ 0.04 ns L 0.21± 0.11 1.54± 0.64 ns Lethrinidae S 0.43± 0.23 L 1.29± 0.34 1.38± 0.77 ns Mullidae S L 3.86± 1.61 1.65± 0.68 ns Pomacentridae S 1.36+ 0.52 0.27± 0.16 ns L 87.79 ±13.10 273.38±32.63 Table 5 Multivariate analysis of variance of density data for apogonids, atherinids, gobiids, lethrinids, mullids, and pomacentrids (see Materials and methods) from off Lizard Island, Great Barrier Reef. Factors include sampling methods (purse-seine, bongo net, neuston net, Tucker trawl) and nights (3, 5, and 6 December 1986). Data are ln(x+l) transformed. Test statistic used is Pillai's trace. Significance levels: **0.01>p>0.001; ***p<0.001. Source F Numerator df Denominator df Method Night Method < Night 11.53 4.05 1.65 18 12 36 Z.A *** 186 ** dominant families; neuston, by higher numbers of atherinids, a neustonic group. The significant interac- tion is attributable largely to the purse-seine result. 204 Fishery Bulletin 91 12). 1993 Table 6 Standardized canonical coefficients from the Canonical Dis- criminant Analysis of density of fishes over each method by night combination, from samples taken off Lizard Island. Great Barrier Reef on 3, 5, and 6 December 1986. Data were lni x+ 1 1 transformed. Family CAN 1 CAN 2 Apogonidae 5.031* 0.675 Atherinidae -1.129 1.961* Gobiidae 1.463 0.585 Lethrinidae -1.005 -1.279 Mullidae 0.177 0.595 Pomacentridae 0.184 -0.736 Canonical variate Proportion Cumulative 1 0.793 0.793 2 0.134 0.927 * Consistently high values in total, between and within canonical structure. These variables contribute significantly to the discriminatory power of the canonical variate. Data from all five nights provided more information on patterns of temporal change for some taxa (Fig. 7). We focused on the comparative ability of the different methods to detect changes over time in numbers of the larger (>6mm) individuals of some families because we wished to know the best methods for identifying temporal pulses of large larvae and pelagic juveniles of reef fishes. Large pomacentrids and mullids serve as appropriate examples. Although absolute numbers of fishes taken by nets and aggregation devices could not be directly compared, temporal changes in pat- terns of density could be evaluated among these meth- ods. Comparisons were made using all methods, al- though bongo net data were available for the nights of 3, 5, and 6 December only. Data from the two aggregation devices indicated that large pomacentrids increased in density from the 2nd to a peak on the 5th, and decreased over the 6th and 7th (Fig. 7). This pattern was not present in the data from nets, each of which provided a different temporal pattern of density. \J*y purse seines \j|jl}/ bongo nets neuston net3 •V Tucker trowls + ATHERINIDAE 6 + APOGONIDAE 10 Figure 6 Results of Canonical Discriminant Analysis of density data tnumbers/lOOOm') for apogonids, atherinids, gobiids, lethrinids. mullids, and pomacentrids taken by four net types on the nights of 3, 5, and 6 December 1986 off Lizard Island. Great Barrier Reef. Factors analyzed were net type and night of sampling. Canonical variates 1 and 2 are displayed. Numbers superimposed on circles refer to the day of sample. Choat et al. Comparison of ichthyoplankton sampling methods 205 POMACENTRIDAE MULLIDAE 2 3 Sompling Dote 4 5 6 7 - December 1986 ▲ light traps • light seines tucker trawls A bongo nets o neuston nets • purse seines 2 3 4 5 6 7 Sampling Date — December 1986 Figure 7 Changes in mean density i±SE) of large !>6mml pelagic pomacentrids and mullids sampled by six methods over six nights, 2-7 December 1986 off Lizard Island, Great Barrier Reef. Density estimates for the aggregation devices are not ad- justed for volume sampled. Some methods did not collect large pomacentrids or mullids. The aggregation devices indicated that large mullids were rare or absent until the 5th, and increased greatly in density on the 7th (Fig. 7). This trend was not present in data from the nets. Only the neuston net caught large mullids, but in low and variable numbers. Discussion The taxonomic composition obtained when sampling for larval and pelagic fishes is highly method- dependent. The bongo net captured the largest num- ber of families, many of which were rare in the samples. Among abundant taxa, the four nets provided similar estimates of taxo- nomic composition. The light-trap, however, was more selective, and its catch differed in composition from that of the nets. Taxonomic composition of the light-seine samples was interme- diate between the trap and nets, an expected result given its mode of op- eration. Our results suggest that capture by the light-trap is dependent on fish size: larger pelagic stages are more likely to be attracted to the light and to swim into the trap than are small stages. However, trap performance may also be time-dependent. For ex- ample, apogonids, carangids, lutjanids, and scarids, which were rare or ab- sent in light-trap catches during this study, have been captured during ex- tended light-trap sampling around Liz- ard Island (M. Milicich, Griffith Univ., Nathan, Queensland, pers. commun.). The absence from light-traps at par- ticular times may simply indicate that large or well-developed individuals of some families were not present at that time. However, our study provides evi- dence that pelagic stages of some families may not be photopositive or enter traps, thus indicating some se- lectivity by the aggregation devices. Schindleriids were present in the net samples to adult size, yet were not captured with either of the light- aggregation methods. The net samples may have included the largest pelagic individuals of callionymids, and per- haps platycephalids and bothids, because they leave the pelagic environment (i.e., settle) at a relatively small size (see Table 3). These families were not present in the light-trap catches. The size-distribution and density estimates of pelagic fishes captured also differ among nets. The bongo net, neuston net, and purse-seine captured predominantly smaller fishes. For abundant families, density estimates by the bongo net and purse-seine were generally simi- lar, neuston net estimates were somewhat lower, and the Tucker trawl provided still lower estimates. The bongo net provided the highest abundance estimates for most sizes of most families. The Tucker trawl 206 Fishery Bulletin 91(2). 1993 undersampled smaller individuals, but was no better than the bongo net at capturing larger larvae and pe- lagic juveniles. This is consistent with the results of Kendall et al. (1987) and Clarke (1991), who compared bongo nets and larger trawls. The light-seine captured a wide size-range of fishes because it combined the sampling characteristics of both a purse-seine and an aggregation device. Mesh size is an important determinant of catch com- position because extrusion varies with mesh size. For a given mesh size, extrusion is a function of body shape and pressure across the net mesh (Clarke 1983 and 1991, Gartner et al. 1989). Body shape is species-spe- cific, which emphasizes the importance of taxon-spe- cific factors in methodological studies. Our results cover a comprehensive range of body shapes, from slender (gobiids) to deep bodied (apogonids and pomacentrids) to moderately deep with elongate fin spines (lutjanids), and should have general application. Purse-seines ap- pear to herd planktonic organisms, while towed nets actively filter, often under considerable pressure; thus extrusion will vary between these two gear types re- gardless of mesh size. As our primary interest was in comparing a series of sampling devices in their normal working configuration, we did not attempt to test the effects of different mesh sizes within gear types. Although vertical stratification is minimal at night in the study area (Leis 1986, 1991a), vertical distribu- tion of the fishes could have affected apparent perfor- mance of the samplers because each method sampled somewhat differently in the vertical plane. Towed nets were deployed at fixed depths. Experience elsewhere has suggested that light-traps draw their catch from a relatively narrow depth stratum, the upper 5m (P.J. Doherty, unpubl.). However, only in the neuston net can we confidently attribute greater catches (especially of atherinids) to vertical stratification. For this study, we assumed that vertical distribution of the fishes did not affect our evaluation of the other methods. Horizontal or temporal variations in density may also have confounded comparisons. A position effect was possible because the aggregation devices were op- erated at fixed positions about 700 m apart (Fig.l). A temporal effect is possible because the bongo net and Tucker trawl tows were run in blocks and not random- ized during each night's sampling, although the order of blocks was alternated among nights. Absolute sampling efficiency of the nets was not mea- sured. Our estimates of sampling performance were relative, because we did not obtain unbiased estimates of the true densities of small pelagic fishes. We did not attempt to use the methods of Somerton & Kobayashi (1989) to correct our net catches because we felt some of the assumptions required, especially those relating to patch size and consistency through time, were not appropriate in the case of our study. The smaller bongo net seemed to have equal or greater sampling effi- ciency than the larger Tucker trawl at night for large pomacentrids. A comprehensive comparison of the six sampling methods would require two things. First, we would need to standardize all results as number of organ- isms per unit volume of water sampled. Second, we would require an estimate of the sampling precision of each device. For towed nets, both could be obtained because flowmeters provided estimates of the volume filtered for each tow. In the case of the purse-seine, it was not possible to obtain reliable estimates of the volume of water filtered during each deployment of the net. Minor variations in the deployment procedure can modify the dimensions of the volume enclosed by the net. At present, we have no reliable way of esti- mating this; therefore, for the purse-seine we have a general estimate of water filtered based on idealized dimensions of the deployed net. Volumes sampled by aggregation devices cannot be estimated at this time, but preliminary calculations (below) suggest they may be large. The bongo net as operated in this study will sample -4000 nvVh, the Tucker trawl -14,000 m 3/ h, and we estimate the light- aggregation techniques could sample tens of thousands of m3/h. Therefore, light-aggregation techniques may be the best way to capture sufficient numbers of rarer, larger stages for useful analyses. Aggregation meth- ods may offer considerable advantages in studies of settlement-stage reef fishes, but one must accommo- date the characteristic taxonomic selectivity and un- known sample volume. Two alternatives may explain the apparent dispar- ity in numbers of larger pomacentrids estimated by the bongo net (average 6.9/1000 m 3 ; Tucker trawl catches averaged 1.49/1000 m 3 ) and the light-trap ( average 273/h ): ( 1 ) The bongo net undersamples these larger pelagic stages relative to the light-trap, or (2) the light-trap samples larger volumes of water. As- suming the two methods sample large pomacentrids with equal efficiency, the light-traps sample volumes on the order of 40,000m"h. This requires the trap to capture, with efficiency equal to that of the net, pho- topositive stages within a 7-50 m radius (to 5m depth) of the trap, depending on the current speed (average in the area is 15cm/s; Leis 1986) and geometry of the light field. It is not possible to choose between alterna- tives without a better measure of the effective volume swept by traps. Work in progress will help resolve this question. Short-term temporal variation in the density of par- ticular families was more obvious in the results of some methods than others. For the smaller size-classes, neu- ston, bongo, and Tucker nets gave consistent results Choat et al.: Comparison of ichthyoplankton sampling methods 207 over short time-periods (Fig. 6). Catches from the purse- seine were more variable within a sampling period and showed greater variability among nights of sam- pling than did the towed nets. This reflects the local- ized sampling area and small sample volume of the purse-seine. For larger mullids and pomacentrids, simi- lar trends in density over five nights were identified by the aggregation devices. These trends were not ap- parent in the data from the towed nets. Thus, the aggregation devices seem particularly suited to stud- ies of short-term temporal variation in the larger (>6mm) size-classes. The rapid and independent changes in density of the larger individuals of these two families suggest that larger pelagic stages are not present in the water at all times at a location. The alternative, that there are short-term taxon-specific changes in catchability due to changes in behavior of the fishes, seems less likely, but cannot be dismissed without further study. A number of other studies have compared sampling methods for planktonic and pelagic assemblages. Purse- seines were found to be superior to towed nets for sampling larval anchovies (Murphy & Clutter 1972). Larger, faster, more-transparent nets may minimize net avoidance (Clutter & Anraku 1968). However, Smith & Richardson (1977) suggest that increased net size and towing speed may intensify the disturbance in front of the net and increase net avoidance. All towed nets in these cited studies employed towing bridles, which are a source of water disturbance and, thus, net avoidance by fishes. Towing bridles were not used in the present study, which may be why our conclusions differ from those of Clutter & Anraku (1968) and Murphy & Clutter ( 1972 ). We agree, however, with Clarke (1991) who made detailed comparisons of the effectiveness of two types of bongo nets and a midwater trawl in capturing reef- fish larvae. He suggested that the bongo nets (0.7m diameter with 0.183 mm mesh, and 1.25 m diameter with 2.5 mm mesh) sampled larvae as well or better than a 3 m Issacs-Kidd trawl (6 mm mesh). Clarke con- cluded that when densities of larvae were high, 0.7 m and 1.25 m bongo nets were the most effective meth- ods for sampling small and large larvae, respectively. Although larger nets are assumed to capture more and larger fishes due to lessened avoidance (Clarke 1983 and 1991, Methot 1988), this was not true in our study nor is it always true in other pelagic groups (Barnes & Tranter 1965, Sands 1978, Pillar 1984). One other significant study compared catches from a light-trap with those from a towed net. Gregory & Powles (1988) investigated a relatively simple plank- tonic assemblage of freshwater fishes. Based on a com- parison of taxonomic composition and size of fishes, they concluded that both sampling methods should be used to avoid selectivity biases. An interesting conclu- sion that differs from our results was that the light- trap provided a better representation of size-classes, including smaller individuals, than did the towed net. This emphasizes the taxon-specific and, perhaps, habi- tat-specific nature of gear-performance measures. We agree with Omori & Hamner (1982) that the sampling device and program selected must be ques- tion-driven (Kingsford 1988). In order to assist in the choice of appropriate methods, we summarize the per- formance and sampling properties of the six methods employed in this study (Table 7). Surveys of larval fishes are best accomplished with a bongo net. This will cover a significant portion of the size-range in many important taxa, including larger individuals, at least at night. No extra benefits were apparent from using the larger Tucker trawl. A major advantage of bongo nets is the relative ease with which they may be deployed and retrieved. As expected, neuston nets fo- cused on neustonic fishes. Surprisingly, the purse-seine provided results com- parable to the bongo net despite the small volumes sampled. Among-sample variances were predictably Table 7 Sampling characteristics of six method s used to collect pk nktonic and pelagic fishes at the Lizard Island study site. Great Barrier Reef. Performance Bongo Neuston Tucker Purse Light- Light- criterion net net trawl seine trap seine Size selectivity Wide size- Samples larger Samples larger Primarily Primarily Wide size-range. range; modal individuals of sizes; no more small large values at lower some taxa; effective than individuals. individuals. size. modal values at lower size. bongo net at night. Taxonomic Least-selective. Neustonic Slender taxa Captures only Selective; Combines light selectivity taxa only. and small shallow living dependent on selectivity with individuals taxa; taxon characteristics extruded. undersamples rare taxa. behavior. of purse-seine. 208 Fishery Bulletin 91(2), 1993 higher than those of towed nets. Sampling of local- scale surface features requires the degree of spatial precision and replication provided by small purse-seines (Kingsford & Choat 1985 and 1986, Kingsford et al. 1991), but purse-seines cannot sample deeper than the upper few meters of the water column, and are diffi- cult to operate in any but the best conditions. Local- ized replicated sampling may also be obtained by free- fall plankton nets (Kobayashi 1989) which, however, obscure vertical patterns and also have a small vol- ume sampled. Investigation of the patch size of pelagic organisms requires the ability to sample simultaneously over sev- eral spatial scales. Large-scale deployment of arrays of automated light-traps will increase replication and allow investigation of phenomena at several spatial scales without risk of temporal confounding, provided the traps can be retrieved over the same time-period. Also, both light-traps and purse-seining with aggrega- tion devices may detect temporal pulses in the density of larger larvae and pelagic juveniles with greater reli- ability and precision than towed nets. In addition to the sampling properties of the differ- ent devices, there are a number of more pragmatic considerations. Sorting and identification of samples may be a major bottleneck. This will be influenced by the size of the sample, the amount of organic material included, and condition of the fishes themselves. In this context, large samples taken by finer-mesh nets may be particularly difficult to process. Smaller or more selective samples are more readily processed, and those from purse-seines and light-traps yield living fishes suitable for rearing and experimentation. Further, the smaller the larva the more difficult it is to identify; thus, methods like the light-trap, which samples larger fishes, simplify identification. It is clear that studies of the biology of small pelagic fishes require the use of both nets and aggregation devices either separately or in combination, depending on the type of question posed. No single method can provide a comprehensive picture of the larval and pe- lagic juvenile fish fauna, and few programs could cover the expense and logistic effort of the simultaneous de- ployment of a variety of methods. The picture one ob- tains of the larval and pelagic juvenile fish fauna is highly method-dependent. Which picture or combina- tion of pictures is suitable for answering a given ques- tion varies with the question, the taxon, and the size- range of the fishes. Acknowledgments This project was supported by funds from the URG Griffith University to PJD, the Australian Marine Sci- ences and Technology funding panel to JHC and JML, and internal funds from the Australian Museum (JML) and James Cook University (JHC). Field facilities were provided by the Lizard Island Research Station and the Australian Museum. We wish to thank M. and M. Jumelet (RV Sunbird) and M. Milicich, M. Meekan, M. McCormick, N. Preston, and M. Doherty for assis- tance with the sampling program; D. Furlani, R. Birdsey, S. Reader, and S. Thompson for assistance in the lab; and M. Kingsford and M. Milicich for dis- cussion of the program and critical reading of the manu- script. The manuscript benefited from the comments of anonymous reviewers. Citations Barnes, H., & D. J. Tranter 1965 A statistical examination of the catches, numbers, and biomass taken by three commonly used plankton nets. Aust. J. Mar. Freshwater Res. 16:293-306. Brander, K., & A. B. Thompson 1989 Diel differences in avoidance of three vertical pro- file sampling gears by herring larvae. J. Plankton Res. 11 (4):775-784. Clarke, T. A. 1983 Comparison of abundance estimates of small fishes by three towed nets and preliminary results of the use of small purse seines as sampling devices. Biol. Oceanogr. 2:311-340. 1991 Larvae of nearshore fishes in oceanic waters near Oahu, Hawaii. NOAA Tech. Rep. NMFS 101, 19 p. Clutter, R. I., & M. Anraku 1968 Avoidance of samplers. In Zooplankton sampling, p. 57-76. UNESCO Monogr. Oceanogr. Method. 2. Doherty, P. J. 1987 Light-traps: selective but useful devices for quan- tifying the distributions and abundances of larval fishes. Bull. Mar. Sci. 41:423-431. Doherty, P. J., & D. McB. Williams 1988 The replenishment of coral reef fish popula- tions. Oceanogr. Mar. Biol. Annu. Rev. 26:487-551. Frank, K. T. 1988 Independent distribution of fish larvae and their prey: natural paradox or sampling artifact. Can. J. Fish. Aquat. Sci. 45:48-59. Gartner, J. V. Jr., W. J. Conley, & T. L. Hopkins 1989 Escapement by fishes from midwater trawls: a case study using lantern fishes. Fish. Bull., U.S. 87:213-222. Green, R. H. 1979 Sampling design and statistical methods for environmental biologists. Wiley Interscience, NY, 257 p. Gregory, R. S., & P. M. Powles 1988 Relative selectivities of Miller high-speed samplers and light traps for collecting ichthyoplankton. Can. J. Fish. Aquat. Sci. 45:993-998. Choat et al Comparison of ichthyoplankton sampling methods 209 Kendall, A. W. Jr., E. H. Ahlstrom, & H. G. Moser 1984 Early life history stages of fishes and their characters. In Moser, H.G., et al. (eds.), Ontogeny and systematics of fish, p. 11-22. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol. Allen Press, Lawrence KS. Kendall, A. W. Jr., M. E. Clark, M. M. Yoklavich, & G. W. Boehlert 1987 Distribution, feeding and growth of larval walleye pollock, Theraga ehalcogramma, from Shelikof Strait, Gulf of Alaska. Fish. Bull., U.S. 85:499-521. Kingsford, M. J. 1988 The early life history of fish in coastal waters of northern New Zealand: A review. N.Z. J. Mar. Fresh- water Res. 22:463-479. Kingsford, M. J., & J. H. Choat 1985 The fauna associated with drift algae captured with a plankton-mesh purse seine net. Limnol. Oceangr. 30:618-630. 1986 The influence of surface slicks on the distribution and movements of small fish. Mar. Biol. (Berl.) 91:161-171. Kingsford, M. J., E. Wolanski, & J. H. Choat 1991 Influence of tidally induced fronts and Langmuir circulations on distribution and movements of presettlement fishes around a coral reef. Mar. Biol. (Berl.) 106:167-180. Kobayashi, D. A. 1989 Fine-scale distribution of larval fishes: Patterns and processes adjacent to coral reefs in Kaneohe Bay, Hawaii. Mar. Biol. (Berl.) 100:285-293. Leis, J. M. 1986 Vertical and horizontal distribution of fish larvae near coral reefs at Lizard Island, Great Barrier Reef. Mar. Biol. (Berl.) 90:505-516. 1991a Vertical distribution offish larvae in the Great Barrier Reef Lagoon, Australia. Mar. Biol. (Berl.) 109:157-166. 1991bThe pelagic stage of reef fishes: the larval biology of coral reef fishes. In Sale, PF (ed.). The ecology of fishes on coral reefs, p. 183-230. Academic Press, San Diego. Leis, J. M., & D. S. Rennis 1983 The larvae of Indo-Pacific coral reef fishes. N.S.W. Univ. Press, Sydney, and Univ. Hawaii Press, Hono- lulu, 269 p. Leis, J. M., & T. Trnski 1989 The larvae of Indo-Pacific shore fishes. N.S.W. Univ. Press, Sydney, and Univ. Hawaii Press, Hono- lulu, 371 p. McGowan, J. A., & D. M. Brown 1966 A new opening-closing paired zooplankton net. Ref. 66-23, Univ. Calif. Scripps Inst. Oceangr., La Jolla, 56 p. Methot, R. D. 1988 Frame trawl for sampling pelagic juvenile fish. Calif. Coop. Oceanic Fish. Invest. Rep. 27:267-278. Moser, H. G. 1981 Morphological and functional aspects of marine fish larvae. In Lasker, R. (ed.), Marine fish larvae: morphology, ecology and relation to fisheries, p. 89- 131. Univ. Wash. Press, Seattle. Moser, H. G., W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr., & S. L. Richardson (editors) 1984 Ontogeny and systematics of fishes. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol., Allen Press, Lawrence KS, 760 p. Murphy, G. I., & R. I. Clutter 1972 Sampling anchovy larvae with a plankton purse seine. Fish. Bull., U.S. 70:789-798. Omori, M., & W. M. Hamner 1982 Patchy distribution of zooplankton: behavior, popu- lation assessment and sampling problems. Mar. Biol. (Berl.) 72:193-200. Pillar, S. C. 1984 A comparison of the performance of four zooplank- ton samplers. S. Afr. J. Mar. Sci. 2:1-18. Sands, N. J. 1978 Ecological studies on the deep-water pelagic com- munity of Korsfjorden, western Norway. Comparison of the catch of the more numerous species by two different nets. Sarsia 63:237-246. SAS 1987 Guide for personal computers, Version 6 ed. SAS Inst., Cary NC. Smith, P. E., & S. L. Richardson 1977 Standard techniques for pelagic fish eggs and larva surveys. FAO Fish. Tech. Pap. 175, 100 p. Snedecor, G. W., & W. G. Cochran 1980 Statistical methods, 7th ed. Iowa State Univ. Press, Ames, 593 p. Somerton, D. A., & D. R. Kobayashi 1989 A method for correcting catches of fish larvae for the size selection of plankton nets. Fish. Bull., U.S. 87:447-455. Suthers, I. M., & K. T. Frank 1989 Inter-annual distribution of larval and pelagic ju- venile cod (Gadus morhua) in southwestern Nova Scotia determined with two different gear types. Can. J. Fish. Aquat. Sci. 46:591-602. Taylor, L. R. 1961 Aggregation, variance and the mean. Nature (Lond.) 189:732-733. Tucker, G. H. 1951 Relation of fishes and other organisms to the scat- tering of underwater sound. J. Mar. Res. 10:215- 138. Victor, B. C. 1986 Larval settlement and juvenile mortality in a recruitment-limited coral reef fish population. Ecol. Monogr. 56:145-160. Warner, R. R., & T. P. Hughes 1989 The population dynamics of reef fishes. Proc. Sixth Int. Coral Reef Symp., Australia, vol. 1:149-155. Abstract.— Analysis of surface and subsurface plankton collections in the Middle Atlantic Bight (MAB) yielded larvae and juveniles of Phycis chesteri and five species of Urophycis. Identification was based on numbers of epibranchial gill rak- ers, abdominal vertebrae, and fin rays (dorsal, caudal, pelvic), patterns of pterygiophore interdigitation, and morphometric characters including body depth at the vent and a ratio between height of the pelvic-fin base and length of the mandible. Urophycis tenuis accounted for 99% of the Urophycis larvae and pelagic juveniles collected during spring off Virginia and New Jersey and was most abundant offshore. Urophycis tenuis larvae were smallest at off- shore stations and increased in size as collections proceeded shoreward. Urophycis chuss was found in sum- mer and fall collections off the coasts of New Jersey and Virginia, with abundances highest at midshelf sta- tions. Urophycis chuss was the only species of hake found during August and early September, and it domi- nated summer ichthyoplankton col- lections. Urophycis regia was found primarily in midshelf areas off Vir- ginia during fall, but was also col- lected offshore from both Virginia and New Jersey during winter. Phycis chesteri, also found in fall and winter collections, was restricted to offshore stations. Southern species, found exclusively in offshore winter collections, included U. floridana and U. cirrata. Identification and distribution of Urophycis and Phycis [Pisces, Gadidae) larvae and pelagic juveniles \n the U.S. Middle Atlantic Bight* Bruce H. Comyns Virginia Institute of Marine Science, The College of William & Mary Gloucester Point. Virginia 23062 Present address. Gulf Coast Research Laboratory. PO. Box 7000 Ocean Springs. Mississippi 39564 George C. Grant Virginia Institute of Marine Science, The College of William & Mary Gloucester Point, Virginia 23062 Manuscript accepted 4 December 1992. Fishery Bulletin, U.S. 91:210-223 (1993). Species of the gadid genera Urophycis (Gill) and Phycis (Artedi), collectively referred to as 'hakes', are abundant on the continental shelf and slope of the northwest Atlantic Ocean. Six species of Urophycis and one species of Phycis are found in this area (Svetovidov 1948, Wenner 1983): U tenuis (Mitchill), U. chuss (Walbaum), U. regia (Walbaum), U. floridana (Bean and Dresel), U. earlli (Bean), U. cirrata (Goode and Bean), and P. chesteri (Goode and Bean). Larval hake are present at all times of the year in the Middle Atlantic Bight (MAB) and dominate summer plank- ton collections (Comyns 1987), but persistent taxonomic problems ( Dunn & Matarese 1984) have hindered the accumulation of ecological data on these important components of offshore ichthyoplankton communities (Kendall & Naplin 1981, Hermes 1985). Newly hatched U. chuss and U. regia of known parentage were described by Hildebrand & Cable (1938), Miller & Marak ( 1959), Barans & Barans (1972), and Serebryakov (1978). Although these sources de- scribe pigmentation differences be- tween the two species, this informa- tion alone is insufficient to positively identify field-caught larvae. Older larvae and juveniles of U. chuss, U. regia, U. floridana, and a single juvenile specimen of U. earlli were described by Hildebrand & Cable (1938). Larvae and juveniles of U. regia were collected off Beau- fort NC and were identified by the presence of relatively few second dorsal-fin rays and lack of dark ven- tral-fin pigment. A second larval morph collected off Beaufort was tentatively identified as U. floridana because adult U. floridana was the only other species of Urophycis com- monly found in the collection area. These specimens differed from U. re- gia in having darkly-pigmented ven- tral fins and more second dorsal-fin rays. A single juvenile specimen (37 mm) collected off Beaufort was identified as U. earlli because this specimen was darker than U. regia and U. floridana, and possessed smaller scales. Specimens of a fourth morph, collected off Cape Henry VA were identified as U. chuss because they possessed dark ventral-fin pigment, were relatively slender- bodied, and it was assumed that Contribution 1799 of the Virginia Institute of Marine Science and School of Marine Sci- ence. The College of William & Mary 210 lomyns and Grant. Urophyas and Phyas larvae and pelagic juveniles U. floridana would not be found as far north as Cape Henry. Methven (1985) presented a size-dependent key to the identification of young U. chuss, U. tenuis, and P. chesteri from the Northwest Atlantic. Identifications were based on body depth, numbers of epibranchial gill rakers (Musick 1973, Wenner 1983), and numbers of caudal-fin rays. Material for Methven's study came primarily from Canadian waters, and he did not en- counter U. cirrata, U. earlli, U. floridana, or U. regia. As a result, Methven's key is of limited use in more southerly locations where these species occur. The objective of this paper is to describe additional morphometric and meristic characters that aid in the identification of Urophycis and Phycis larvae, and to describe the spatial and temporal distribution of these larvae collected 1975-77 off Virginia and New Jersey in the Middle Atlantic Bight. Materials and methods Sampling locations and shipboard procedures Sampling extended from October 1975 until August 1977 and was conducted quarterly at 12 stations off Virginia and New Jersey (Fig. 1, Table 1). Neuston samples were collected with a floating sampler devel- oped at Woods Hole Oceanographic Institution (Bartlett & Haedrich 1968, Craddock 1969). The net, constructed with 505 (im mesh Nitex, was lm wide and fished to a depth of 12 cm in calm seas. Tows were of 20min duration at a ship speed of ~2kn. The net was deployed from a boom and the towing course followed a widely-circular track to prevent sampling in the ship's wake. A single neuston tow was made at 3h intervals over a 24 h period at each station, re- sulting in eight samples per station during each cruise. Two oblique tows between nearsurface and bot- tom were made at all stations with 60 cm opening- closing bongo systems (McGowan & Brown 1966), the first with paired 202 um Nitex nets and the second with paired 505 tim nets. To prevent surface contami- nation, all nets were closed during passage through the surface layer (upper meter). Both bongo and neu- ston nets were equipped with flowmeters (General Oceanics, Inc.). The flowmeter attached to the under- side of the neuston frame provided an estimate of horizontal distance relative to sea surface fished by the net. Estimates of volume filtered by the neuston sampler were determined by multiplying distance fished by net area fished (lmxl2cm). In calm seas the neuston sampler consistently fished to a depth of 12 cm, but in rough seas the net opening would occa- >x /\l / ^1 v *"■; _ _ __ - ^s~ f \ - ^\ 1 " "~s ^- -^ \ ^> a ^-j 1" N \ . \ ^ * \JOOO \\ / ' 1 \" \\ i A2 ' ' i • ' ' V B5 ' ill * \ 1 / y - -»o 1 ,,' i ^ 1 » ""■ ^r J ^ \ 1 ^^C I NJ t C1 \ / \ 1 ^^ {x?U • N \ D . 1 >Vf \r N3 , --"\ • , E3 \ n * 'A ' i /, j y '!" >\ / i'i t \ Hi > \ "i r j \ y%> ! l \ ,^ j '■'■ ^ \ s ' \ ^ DE ^N 1 ' \ lr"^ J ^' 'vl 1 ^ MD j 1 ) V I \ s \ ) ^k > < L <3 '/ ; \ VO m™ A" 3 M A, I { LI \ !■>. 1 \' \ 1 \ ^s^^ • t s L . 4 L6' IS Figure 1 Icht.hyoplankton sampling locations off New Jersey and Virginia. Stns. F2, Jl, A2, L4. and L6 are considered offshore stations. sionally be almost completely filled or empty (the flowmeter was always submerged). This variability decreased precision of neuston volume estimates but was not expected to bias volume estimates. Compari- sons between neuston and bongo collections were per- formed only to emphasize the relative importance of the surface layer to larval hakes. Patterns of the spa- tial and temporal distribution of larvae were based only on comparisons among neuston collections be- cause most specimens were collected with this gear type. Laboratory procedures Fish larvae were sorted from whole collections. All specimens of Urophycis and Phycis were cleared and stained (Dingerkus & Uhler 1977, Potthoff 1984, Tay- 212 Fishery Bulletin 91(2). 1993 Table 1 Sampling schedu e for twelvs stations occupied off N ew Jersey (NJ) and Virgi nia (VA) , 1975- -77. 1975 1976 1977 Aug Feb Stn. Oct Feb Jun Sep Nov Mar May Aug CI (NJ) X X X X X X X X Dl X X X X X X X X N3 X X X X X X X X E3 X X X X X X X X F2 X X X X X X X X Jl X X X X X X X X B5 (NJ) X X X X A2 X X X X LI (VA) X X X X L2 X X X X L4 X X X X L6 X X X X lor & Van Dyke 1985), except those occurring in collec- tions taken during August-September 1976 («>16,000) and August 1977 («>4000). During these periods of high abundance, subsamples of over 2000 larvae from Au- gust-September 1976 and >900 larvae from August 1977 were randomly selected and similarly processed. Herein specimens <18mmSL are arbitrarily termed larvae, whereas fish >18mm are termed juveniles (Markle et al. 1982). Fish <~12mm were measured with an ocular micrometer, while lengths of larger specimens were mea- sured with a dial caliper ruler. The largest pelagic juve- niles found were -40 mmSL. The following morphometric criteria were used in the analysis: ( 1 ) height of pelvic fin/vertical distance from base of pelvic fin to ventral mar- gin of body; (2) mandible length/distance from anterior tip of the dentary to posteroventral tip of the angular; and (3) body depth at anus/ver- tical distance from anterior end of anal-fin base to dorsal surface immediately above. Morpho- metric measurements were made with an ocu- lar micrometer. The first interneural space was defined as the space anterior to the first neural spine. Hake larvae and juveniles possessing the adult meristic complement were initially iden- tified using published and unpublished meristic data (Table 2). Meristic observations included epibranchial gill rakers (left side examined), ab- dominal vertebrae, and fin rays (dorsal, caudal, pelvic). Observations were taken from both cleared and stained material and from radio- graphs of juvenile and adult museum specimens (App. Table 1). Identification of smaller larvae was facilitated by using morphometric criteria, patterns of interdigitation between pterygiophores supporting the median fins and the neural or haemal spines, and by defining the size at which larvae attained various stages of fin-ray, ver- tebral, and gill-raker development. Unfortunately, faded pigmentation caused by specimen storage in formalin and subsequent clearing and staining prevented use and further description of larval pigmentation in the present study. Table 2 Ranges of meristic characters in Phycis chesteri and six species of Urophycis. Numbers in parentheses indicate meristic ranges that were extended by the present study. Data sources are (1) Svetovidov 1948, (2) Hildebrand & Cable 1938, (3) Bigelow & Schroeder 1953, (4) Leim & Scott 1966, (5) Miller & Jorgenson 1973, (6) Musick 1973, (7) Hoese & Moore 1977, (8) Markle 1982, (9) Fahay 1983, (10) Wenner 1983, (11) Methven 1985, (12) J.A. Musick, pers. commun., VIMS, Gloucester Point VA 23062. U. tenuis U. chuss U. regia U. floridana U. earlli U. cirrata P. chesteri Caudal-fin rays 1st dorsal-fin rays 33-38(40) 9-10(12i 28-34 9-11(12) 28-32(34) 8-10 28-32(34) 10-13 27-30(31) 8-11 28-33 9-11(12) 28-35(361 8-11(12) 2nd dorsal-fin ravs 50-59(62) (52153-64 43-51(52) 54-63 57-63(68) 54-68 50-63 Anal-fin rays 41-52(53) 45-57 41-50(52) 45-54(55) 49-56(60) 46-58 43-54 Vertebrae ( total ) 47-50(51) 45-50(51) (44)45-47(48) 44-50(51) 45-47(48) 47-53 45-52 Caudal vertebrae Abdominal vertebrae (32)34-35 13-17 33-36 14-17 (30)31-33(34) 13-15 30-34(35) 14-17 31-33(34) 14-15 32-37 15-17 31-37 13-16 Pelvic-fin rays b Epibranchial gill Rakers (1st arch) 3 2 3 3 3 3 3 2 3 2 3 3 3 4-5 Data source 1,3,6,8.11 1,2,4,5,6,8,11 1,2.3,4,5,8 1.7,12 1.5,12 1.12 1.10,11,12 In our material (n=205) U. tenuis never possessed <15 abdominal vertebrae. b The third pelvic-fin ray in adult Urophycis and Phycis is rudimentary. ' Urophycis tenuis occasionally has three epibranchial gill rakers. Comyns and Grant Urophycis and Phycis larvae and pelagic |uveniles 213 Results Identification of Urophycis and Phycis larvae and juveniles Meristics Epibranchial gill rakers (Table 3) A complete size- series of all species was not available, but size (mm) at which U. regia, U. chuss, and U. tenuis larvae attain the adult complement of epibranchial gill rakers and other meristic elements is shown in App. Table 2. The following sections are abbreviated to avoid repeating in the text what the tables and figures succinctly show. Phycis chesteri does not attain the adult complement of epibranchial gill rakers until 16-18 mm (Methven 1985), but by 13 mm the third gill raker has developed and serves to separate larvae of this species from U. tenuis, U. earlli, and U. floridana. Occasionally U. chuss and U. regia possess two or four epibranchial gill rak- ers on one side, but most of these specimens have the normal complement of three gill rakers on the other side. Urophycis tenuis occasionally possesses a 3rd epibranchial gill raker, but only very rarely is this third gill raker found on both sides of a specimen. Caudal-fin rays (Table 3) All but two specimens of U. tenuis (rc=195) had more caudal-fin rays than all other species of Urophycis. Numbers of caudal-fin rays of U. tenuis overlapped those of P. chesteri, but more than half of our U. tenuis were distinct in having >36 rays, and over 40% of P. chesteri (n-56) differed from U. tenuis in having <34 rays. As few as 28 caudal-fin rays have been reported in P. chesteri (Wenner 1983) and U. cirrata (J.A. Musick, VA Inst. Mar. Sci., pers. commun.), but this may be because some of the small procurrent rays are not easily seen in radiographs of larger fish. No U. earlli specimens (n=31) possessed >31 caudal- fin rays, while all other hake commonly have >31 rays. Dorsal-fin rays (Table 3) Despite overlapping ex- tremes, numbers of first dorsal-fin rays helped distin- guish U. floridana from other species of hake. In our material, U. regia and U. earlli never possessed >10 and 11 first dorsal-fin rays, respectively, while over 80% of U. floridana (ra=45) had >11 rays. One-third of U. floridana specimens examined possessed 13 first dorsal-fin rays, delimiting these from all other hake taxa. The relatively low number of second dorsal-fin rays in U. regia separated this species from P. chesteri and other Urophycis species with little overlap. Urophycis chuss and U. regia with incomplete development of second dorsal-fin rays were delimited by numbers of pterygiophores supporting these rays at sizes as small as 6mm (Fig. 2). Although extremes in numbers of second dorsal-fin rays overlapped in all other taxa, many of the available specimens of U. earlli and U. cirrata were distinct in possessing >63 rays. Abdominal vertebrae (Table 3) Numbers of ab- dominal vertebrae cannot be used alone to identify individual specimens because of overlapping extremes, but this meristic character is useful when identifying collections comprised entirely off/, tenuis or U. chuss. Urophycis tenuis larvae, identified by numbers of epibranchial gill rakers and caudal-fin rays, were found in the MAB only in spring and accounted for 99% of the Urophycis collected at this time ( U. regia juveniles accounted for the other 1%). Urophycis larvae <10mm (n=154) that were present in spring collections had not yet developed the adult complement of caudal-fin rays, but these larvae (>4 mm) had developed the adult complement of abdominal vertebrae and were identi- fied as U. tenuis because their frequency-distribution of numbers of abdominal vertebrae was identical to that found in larger U. tenuis; 88% of the larvae had 16 abdominal vertebrae, and no specimens were found with <15. It is unlikely that any of these small larvae were U. floridana or U. cirrata, two other species with similar numbers of abdominal vertebrae, because these two southern species were found only in offshore winter collections and most specimens were pelagic juveniles. Urophycis chuss >4 mm (n=448) possessed 14-16 ab- dominal vertebrae, but the majority of specimens (n=391) had 15. In all other species of Urophycis the count of 15 occurred in <20% of the specimens; and although extremes of U. chuss and P. chesteri were similar, P. chesteri commonly had 14 or 16 abdominal vertebrae (17%). Consequently, in the MAB during late summer when U. chuss larvae are extremely abun- dant (and, in this study, were the only species of hake found at this time), complete meristic counts to check for species other than U. chuss need be performed only on those specimens that do not have 15 abdominal vertebrae. If species other than U. chuss are found in late-summer collections, numbers of abdominal verte- brae are no longer taxonomically useful and complete meristic counts are necessary to identify larvae. Urophycis regia (n=698) had 13-15 abdominal ver- tebrae, but only eight specimens had 15, and seven of these specimens had an anomalous 15th abdominal vertebra. This anomalous vertebra possessed one short transverse process characteristic of abdominal verte- brae and one long transverse process typical of caudal vertebrae. Because 99.9% of U. regia examined had <15 normal abdominal vertebrae, it was assumed that specimens with >15 abdominal vertebrae were not U. regia. This meristic character aided in the separation of small (<6mm) U. chuss and U. regia in fall collec- 214 Fishery Bulletin 91(2), 1993 Table 3 Frequency distribution of meristic values "ecordec in Phycis :hesteri and six species of Urophycis. Asterisks denote that data from juvenile and adult specimens are included. All larva e had attained the adult meristic comp] ement. Epibranchial gill rakers on left first gill arch. Slash separates nu nbers on left and right sides. *P. chesteri 2 3/2 3 4/3 4 5 24 8 U. chuss 8 596 8 2 U. regia 4 631 4 2 *U. cirrata 19 U.. tenuis 160 6 1 *U. floridana 44 *U. earlli 32 Number of caudal-fin rays U. tenuis 29 30 31 32 33 34 35 36 37 38 39 40 2 28 56 65 34 9 1 *P chesteri 1 8 15 19 10 3 U. regia 1 19 34 16 1 U. chuss 1 1 22 13 10 3 *U. cirrata 3 8 2 *U. floridana 5 13 21 14 2 *U. earlli 12 13 6 Number of first dorsal-fin rays *U. floridana 8 9 10 11 12 13 8 22 15 *U. cirrata 2 8 3 U. chuss 2 33 51 9 U. tenuis 10 24 26 3 *P. chesteri 1 19 38 13 2 *U. earlli 6 21 5 U. regia 14 61 7 Number of second dorsal-fin rays *U. cirrata 44 45 46 47 48 49 50 51 52 53 54 . r >:> 56 57 58 59 60 61 62 63 64 65 66 67 68 1 1 112 2 5 1 *U. earlli 1 2 8 8 3 7 11 1 U. floridana 6 6 7 9 6 5 3 3 *P. chesteri 3 8 5 14 13 5 5 3 1 1 U. chuss 1 3 4 6 16 16 16 20 9 6 6 3 U. tenuis 1 3 9 7 8 3 12 5 4 1 1 2 U. regia 2 1 15 22 30 36 29 16 2 Number of abdominal vertebrae U. tenuis 13 14 15 16 17 20 180 5 *U. cirrata 2 10 1 *U. floridana 1 7 38 3 U. chuss 50 391 7 *P. chesteri 8 57 4 *U. earlli 27 4 U. regia 66 324 8 Note: Although 8 specimens of U. regia had 15 abdominal vertebrae . in only one of these speci nens (0.1%) was the 15"' vertebra normally developed. Number of anal-fin pterygiophores anterior to first haemal jpine *U. earlli 3 4 5 6 7 8 9 4 14 8 1 *U. floridana 11 23 7 1 V. regia 8 156 202 14 U. chuss 18 173 152 7 U. tenuis 3 36 25 1 *U. cirrata 1 8 2 *P. Clh 3 28 34 5 Interneural space into which projects the pterygiophore supporting the first ray of tr e second dorsal fin U. tenuis 7 8 9 10 8 52 6 *U. floridana 1 10 25 6 *P. chesteri ■', 31 20 1 U. chuss 4 203 220 4 U. regia 141 41 *U. earlli 15 12 Note: If pterygiophore was aligned with the tip of i neural spine, it was arbitrarily recorded as pointing into the space posterior to the spine in question. Comyns and Grant Urophycis and Phycis larvae and pelagic juveniles 215 60 >50 c !40 30 » CO — 2 o cm en a o d d cnm i ' ipi □ mm cm ODD O CD □ rm OO a n en en a a cd □ a a a a a a 3 a q Urec N=I0I U ChuSS • N = I88 9 10 II 12 Standard Length (mm) 13 14 16 Figure 2 Development of second dorsal-fin pterygiophore number in Urophycis chuss and U. regia. tions when numbers of second dorsal-fin pterygiophores were not yet taxonomically useful. Numbers of abdominal vertebrae may help sepa- rate U. earlli from U. floridana and U. cirrata, the other two southern species of Urophycis. Over 80% of U. floridana and U. cirrata possessed 16 or 17 ab- dominal vertebrae, but U. earlli has never been re- corded with this many. Anal-fin pterygiophores (Table 3) The number of anal-fin pterygiophores positioned anterior to the first haemal spine helps distinguish P. chesteri, U. cirrata, and U. tenuis from U. earlli, U. floridana, U. regia, and U. chuss. Only one specimen of U. tenuis 7 anal- fin pterygiophores positioned an- terior to the first haemal spine, but 45% of U. chuss (n=350) and over half of U. earlli (rc=27), U. floridana (rc=42), and U. regia («=380) had at least 7 of these pterygiophores. More than 60% of U. tenuis, U. cirrata, and P. chesteri had <6 anterior anal-fin pterygiophores, whereas <2% of U. regia and no U. earlli or U. floridana had this few. Second dorsal-fin pterygiophores (Table 3) The interneural space into which points the first pterygiophore of the second dorsal fin helped separate U. chuss from U. regia, and U. floridana from U. earlli. In over half of U. chuss examined (n=431) the first pterygiophore of the sec- ond dorsal fin pointed into the 9th or 10th interneural space, whereas in all U. regia examined (?? = 182) this pterygiophore pointed into the 7th or 8th interneural space. In >75% of U. regia this pterygiophore pointed into the 7th interneural space, whereas <1% of U. chuss showed this pattern. In >70% of U. floridana exam- ined (n=42) the first pterygiophore of the second dorsal fin pointed into the 9th or 10th interneural space, whereas in all juvenile and adult U. earlli examined («=27) this pterygiophore pointed into the 7th or 8th interneural space. In over half of U. earlli examined, the first pterygiophore of the second dorsal fin pointed into the 7th interneural space, but in only 2% of U. floridana did this pterygiophore project this far for- ward. Morphometries Body depth at anus (Fig. 3) Body depth at the anus separated some species of hake larvae at sizes >12-13mm. Extremes of body depth as percent of standard length for cleared and stained P. chesteri, U. tenuis, and U. chuss were 21.0-23.4, 19.0-21.1, 25 P chesteri O n = l7 U tenuis D n = 22 C 24 (J. floridana & n =3 0) a w U regia n =27 53 «- D _ U chuss • n =42 oC3» O D_ ° to j- 22- 1-3 < 20- 3 a " ° o a a , »■ • • • • a D • • • „. • a „ ° n n ^' 9 • • • °- 2 18 • . . ' a w • • • > o l7 - • T3 & '* 6 7 8 9 10 M 12 13 14 15 16 17 18 19 20 21 22 23 24 Standard Length (mm) Figure 3 Body depth at anus as a percent of standard length plotted against standard length for larvae and juveniles of Phycis chesteri and four species of Urophycis. 216 Fishery Bulletin 91(2), 1993 Table 4 Ranges of pelvic-fin-base height as percent of mandible length for Phycis chesteri and five Urophycis. Ranges of values are given for different size-intervals of larvae. ND = no data. species of Size-interval (mm) 5-9 10-14 15-19 20-24 25-29 30-34 35^5 U.regia (re=31) 21-30 19-33 19-25 16-28 12-17 ND ND U. floridana (n=19) ND ND 29-36 23-37 23-28 24-29 ND U. chuss (n=38) 20-39 23-33 24-36 19-22 15-16 16 16 U. cirrata (n=4) ND ND ND 39 31 ND 19-31 P. chesteri (n=29) 44-74 52-61 54-61 46-61 52-64 50-59 26-57 U. tenuis (n=39) 28-42 24-42 33-40 29-37 26-30 32 ND and 17.6-19.7, respectively. Body depth oft/, floridana, however, was found to overlap extremes of U. tenuis and U. regia, while U. regia exhibited the greatest variation in this character, overlapping the extremes of P. chesteri and all other species of Urophycis studied. Mandible length and height of pelvic fin fTable 4, Fig. 4) Height of the pelvic fin plotted against mandible length separated larval P. chesteri from other hake at sizes between ~6 and 35 mm. At sizes > 3 5 mm P. chesteri was similar to Urophycis with respect to this character because P. chesteri became more slen- der-bodied and the pelvic-fin origin moved closer to the ventral margin of the body. Ranges of pelvic-fin height as percent of mandible length in cleared and stained larvae ranging in length from 6 to 35 mm varied from 44 to 74% in P. chesteri in-29), but the highest value of this ratio in five species of Urophycis (n = 131) was only 42%. Distribution and abundance of hake larvae Urophycis chuss Urophycis chuss was found only in summer and fall plankton collections from the MAB, and was the only species of larval hake found in Au- gust and early September. Densities of U. chuss in summer collections off the coast of New Jersey were up to two orders of magnitude greater than densities found off Virginia (Fig. 5). In October 1975 and No- vember 1976, U. chuss were still present off both Vir- ginia and New Jersey, but were far less abundant than during summer. Densities of larval U. chuss also varied with dis- tance from shore (Fig. 5), particularly during summer when lowest densities occurred inshore and highest densities were found in midshelf regions in water depths of 40-120 m. Variations in larval density with both latitude and water depth were not well defined in 1 1 collections. Phvcis * chesteri n=3 3 Uroohvcis • * tenuis chuss n=40 n=4 5 * * * reaia n=32 * floridana n=20 • * » cirrata n=4 * • * ; • - * • • mm i • . "• * • ..'11*1 - -1 • MJ* - ' "" E II *> > »- o ' o CD 0) ». (/) o ■2 S, en P "° 2- u_ to = 1 to o Q 2 3 4 5 Length of Mandible (mm) Figure 4 Height of the pelvic fin plotted against mandible length for larvae and juveniles of Phycis chesteri and five species of Urophycis. An increase in mean size of U. chuss was evident in fall col- lections (Fig. 6). As larval size increased in fall collections, the number of larvae collected with bongo nets decreased greatly. More than 1300 specimens were collected in October 1975 and November 1976, but only 25 of the larvae were collected with bongo gear. Onshore-offshore variation in size of U. chuss was most evident in fall collections off both Virginia and New Jersey; size tended to increase with de- creasing water depth. Urophycis regis Urophycis re- gia was collected in the MAB from October until May, with highest densities of larvae occur- ring in fall collections off the Vir- Comyns and Grant: Urophyas and Phycis larvae and pelagic juveniles 217 NJ 38°42-39°2l' AUGUST 197 7 37°05'-37°3l' 4 □ n = l794 E3n = l906 □ n = 27l5 nn=2550 * * * LT- •'''•'•* 2 ''■'•*•' an 9 -: : x : : * * 1- n ■:•:•:•: •:•:-:■; ;•:•:':• AUGUST-SEPTEMBER 1976 + I' O O I- □ n = 27504 Hn=2770 OCTOBER 1975 Dn=i73 3i EDfl = 249 13 1 n NOVEMBER 1976 □ n=!46 9H = n Dfl on 762 -4 H"vKt.toj CI 01 N3 E3 F2 J Station LI L2 L4 L6 Figure 5 Mean abundance of Urophycis chuss in neuston and bongo collections at stations off Virginia and New Jersey, October 1975-August 1977. n = number of larvae collected. Neuston catches are denoted by clear histograms; bongo catches by stippled histograms. NS = no samples taken. Star denotes bongo catches exceed neuston catches. ginia coast at midshelf station L2 (Fig. 7). Densities of U. regia were much lower in collections taken in Feb- ruary and March, and most specimens were pelagic juveniles found at offshore stations off both Virginia and New Jersey. By May, U. regia was scarce; only seven neustonic juveniles were found at offshore sta- tions. Urophycis tenuis Apart from an occasional U. regia juvenile found at offshore stations, U. tenuis was the only species of hake present in spring plankton collec- tions off Virginia and New Jersey. Abundance of U. tenuis in May 1977 was up to one order of magnitude greater than abundances in June 1976 (Fig. 8). Larvae were collected at all but inshore stations off both Vir- ginia and New Jersey, but were most abundant at off- shore stations. Larvae were smallest at offshore sta- tions and increased in size as collections proceeded inshore (Fig. 9, page 220). Urophycis floridana and U. cirrata Urophycis floridana (ra=41, 13-32 mmSL) and U. cirrata (n=5, 20-42 mmSL) were found exclusively in offshore winter collections (Fig. 10, page 221). With the exception of a single juve- nile U. floridana (23.0 mmSL) captured in a bongo tow, all specimens were found in neuston samples. Phycis chesteri Phycis chesteri larvae first appeared in fall neuston collections from the Middle Atlantic Bight; 16 larvae 6-13 mm in length were collected in November 1976 at offshore stations off Virginia (Fig. 11, page 221). Phycis chesteri larvae and pelagic juveniles remained in surface waters during winter and were found in water deeper than -100 m off both Virginia and New Jersey (n=41). All specimens were collected with the neuston net. Discussion Because of similarities between larvae of the seven hake species found in the MAB, a dichotomous key is not a practical tool with which to identify hake larvae in this area. However, the specific identification of lar- val and pelagic juvenile hake is feasible using a suite of diagnostic characters (App. Table 2). Identifications in this study were based on comparison of larval meristics with adult meristics. Further examination of larvae revealed diagnostic characters comprised not only of meristic information, but also morphometric and pterygiophore interdigitation data. Spawning sea- son and capture location were not used as 'characters' to identify larvae in this study. Methven (1985) had limited success using pigment characters to separate U. chuss and U. tenuis >7-8 mm. Problems will persist with the identification of small hake larvae until ontogenetic pigment patterns of all species have been described. These ontogenetic pig- ment patterns, when used in concert with meristic char- acters, will hopefully enable relatively routine identifi- cations of these taxa. The only species of Urophycis not found in the present study was U. earlli. Adult U. earlli are rare and larvae remain undescribed, but they are expected to co-occur with U. floridana (Hildebrand & Cable 1938). Both species are similar in having two epi- branchial gill rakers, but numbers of first dorsal-fin rays, abdominal vertebrae, and caudal-fin rays delimit most specimens of these two species. Larval and juvenile Urophycis or Phycis were present in the MAB throughout the year, and patterns of spa- tial and temporal distribution of larvae were consis- tent during both years of this study. Urophycis ch uss larvae were found in summer and fall collections, with greatest abundances occurring during summer in the 218 Fishery Bulletin 9 1(2). 1993 August 1977 T - t. t, August-September 1976 NS tot 3^1 55 NS October 1975 00 NS November 1976 N3 E3 Station F2 Figure 6 Range and mean size of Urophycis chuss in neuston and bongo collections at stations off Virginia and New Jersey. Solid and dashed lines indicate neuston and bongo collections, respectively. Horizontal lines show mean values. Two neuston ranges are shown if more than one size-mode is present. NS = no samples taken; numbers = no. larvae/lOOOm 1 . central and northern MAB where water depth was 40-60 m. Urophycis chuss was the only species of lar- val hake found in summer collections, and accounted for 80% of all hakes collected during this 2-year study. Most U. regia in the present study were collected in November, but some larvae or neustonic juveniles were collected from October to May. Urophycis regia in the MAB is reported to spawn from late Sep- tember through November, and possibly to February, with peak activity in October (Barans & Barans 1972). Urophycis regia was most abundant during fall in the southern MAB in the rela- tively shallow 1 4 1-43 m) midshelf area. Size range of Urophycis re- gia collected in this area was 2- 34 mm, and although some of the larger specimens may have drifted from deeper water, small larvae were most likely spawned on the shallower central shelf. Evidence of U. regia spawning in shallow water was also found in October 1975 off New Jersey where larvae as small as 4mm were found in water as shallow as 12 m. However, not all speci- mens off New Jersey originated in shallow water; a second group of larvae 6-23 mm in length was found offshore. The offshore distribution of U. regia became quite distinct in winter collections, with abun- dances being greatest at offshore stations in February 1976, Feb- ruary-March 1977, and May 1977. These U. regia were prob- ably spawned in offshore waters of the MAB or in offshore waters of the South Atlantic Bight and transported northward. Larval U. regia have been found in abundance in winter collections from offshore waters off North Carolina in the South Atlantic Bight (Fahay 1975, Powles & Stender 1976). Late-summer spawning by U. tenuis occurs in shallow wa- ter of the southern Gulf of St. Lawrence and the Scotian Shelf (Markle et al. 1982). Fahay & Able (1989) suggest the existence of a second stock of U. tenuis that spawns in deep water during early spring on the slope of Georges Bank, and probably also along the slopes of the Scotian Shelf, southern New England, and the MAB. The present study found direct evidence of spring spawn- L2 L4 Comyns and Grant Urophyas and Phyas larvae and pelagic juveniles 219 39°2l-26' NJ 38°42-39°2l OCTOBER 1975 Dn=IIO Qn = VA S7*0S-37"S|' NOVEMBER 1976 O O o □ n = i42 ] Qn = 2i K □ n=eo3i E3 n = 15 FEBRUARY 1976 □ n □ n = 29 = FEBRUARY-MARCH 1977 a n = 233 Q n = 2 K , ? Dn=563 0n = o ,On=0 E3 n = o CI Dl N3 E3 F2 Station , On = 2 Qn=o =1=1 LI L2 L4 L6 Figure 7 Mean abundance of Urophycis regia in neuston and bongo collections at stations off Virginia and New Jersey, October 1975-May 1977. n = actual number of larvae collected. Neus- ton catches denoted by clear histograms, bongo catches de- noted by stippled histograms. NS = no samples taken. ing by U. tenuis in deep water of the MAB; in May 1977 U. tenuis larvae as small as 3-4 mm were found over the continental break and slope off both New Jer- sey and Virginia. In June 1976 U. tenuis found at offshore stations were 16-38 mm in length. Based on estimated larval and pelagic juvenile growth rates of 10-22 mm/mo (Markle et al. 1982) and demersal juve- nile growth rates of =30 mm/mo (Fahay & Able 1989), these fish were probably spawned in late April and May. Fahay & Able (1989), studying young U. tenuis in the Georges Bank area, found a shoreward migration with growth. Recruitment to nearshore areas was also indicated in the present study by the increasing size of U. tenuis as collections proceeded shoreward. Neustonic juveniles (35-53 mm) were captured in water as shal- low as 32 m off the coast of New Jersey. Urophycis floridana and U. cirrata, two southern species of hake, were found off New Jersey and Vir- ginia only in offshore winter collections. The large size NJ 39°2l'-2s' D n = 57 □ n = □ n = l30 0n = 5 MAY 1977 m □ n = 294 ran=i D n = 33 oa n=o Dl N3 E3 F2 Station 1 1 1 1 LI L2 L4 L6 Figure 8 Mean abundance of Urophycis tenuis in neuston and bongo collections at stations off Virginia and New Jersey, June 1976 and May 1977. n = actual number of larvae collected. Neus- ton catches denoted by clear histograms, bongo catches de- noted by stippled histograms. NS = no samples taken. and offshore distribution observed for both species sug- gest that these pelagic juveniles may have been trans- ported northward into the study area. Larvae of U. earlli, another species found south of the MAB, are rare and remain undescribed, but this species may also occur occasionally in offshore waters of the MAB during winter. Hildebrand & Cable (1938) expected U. earlli to be a winter spawner after collecting three juveniles (37, 75, 103 mm) in March and April off North Carolina, and Fahay (1975) collected a few neustonic U. earlli in winter in the South Atlantic Bight. Phycis chesteri larvae and pelagic juveniles appeared at offshore stations in fall and winter off Virginia and New Jersey. This larval distribution concurs with Wenner ( 1983) who found adult P. chesteri generally at depths >183m on the continental slope from 36°N to 47°N in the western North Atlantic, and noted that spawning off Virginia took place between late Septem- ber and April, with peak spawning occurring in De- cember and January. Methven & McKelvie (1986) col- lected 51 P. chesteri larvae and pelagic juveniles along the edge of the continental shelf in the MAB, Grand Bank, and Labrador Shelf, and based on estimated growth rates suggested that most spawning occurs in October. This study has shown the spatial and temporal dis- tribution of hake larvae in the MAB to be more com- plex than previously thought. Additional taxonomic characters, particularly ontogenetic pigment patterns, are still needed in order to routinely identify small hake larvae, and more research is needed to explain the observed patterns of larval distribution. Of par- ticular interest is an understanding of the processes that result in the northward transport of larvae and 220 Fishery Bulletin 91(2). 1993 40- 6 30- 20 10 3 1 1 1.0 N S5 June 1976 NS NS B5 A2 N3 E3 Station F2 Figure 9 Range and mean size of Urophycis tenuis in neuston and bongo collections at stations off Virginia and New Jersey. Solid and dashed lines indicate neuston and bongo collections, respectively. Horizontal bars show mean values. Two neuston ranges are shown if more than one size-class is present. NS = no samples taken; numbers = no. larvae/1000m :i . pelagic juveniles of southern species into the MAB. Assuming that these individuals are transported north- ward by the Gulf Stream, it remains to be shown how they leave the influence of this current and move shoreward. Acknowledgments Collections serving as the basis of this research were supported by the U.S. Dept. of the Interior, Bureau of Land Management Contracts 08550-CT-5-42 and AA550-CT6-62. We thank J. Musick, J. Olney, C. Baldwin, J. Lyczkowski-Shultz, and especially M. Fahay for reviewing the manuscript. Adult meris- tic data was provided, in part, by J. Musick and D. Cohen. This work initially comprised a portion of a thesis submitted as partial re- quirement for the MA degree at the College of William and Mary. Citations Barans, C. A., & A. C. Barans 1972 Eggs and early larval stages of the spotted hake, Urophycis regius. Copeia 1972(11:188-190. Bartlett, M. R., & R. L. Haedrich 1968 Neuston nets and South Atlantic larval blue marlin. Copeia 1968(31:469-474. Bigelow, H. B., & W. C. Schroeder 1953 Fishes of the Gulf of Maine. U.S. Fish. Wildl. Serv. Fish. Bull. 74 (vol. 53), 577 p. Comyns, B. C. 1987 Identification and distri- bution of Urophycis (Gill) and Phycis (Artedi) larvae and pe- lagic juveniles in the Middle Atlantic Bight. M.A. thesis, Va. Inst. Mar. Sci., Coll. Wil- liam & Mary, Williamsburg, 130 p. Craddock, J. E. 1969 Neuston fishing. Oceanus 15:10-12. Dingerkus, G., & L. Uhler 1977 Enzyyme clearing of alcian blue stained small verte- brates for demonstration of cartilage. Stain Technol. 52:229-232. Dunn, J. R., & A. C. Matarese 1984Gadidae: Development and relationships. In Moser, H.G., et al. (eds.), Ontog- eny and systematics of fishes, p. 283-289. Spec. publ. 1, Am. Soc. Ichthyol. Herpetol. Allen Press, Lawrence KS. Fahay, M. P. 1975 An annotated list of larval and juvenile fishes cap- tured with surface-towed meter net in the South At- lantic Bight during four R/V Dolphin cruises between May 1967 and February 1968. NOAA Tech. Rep. NMFS SSRF-685, 39 p. 1983 Guide to the early stages of marine fishes occur- ring in the western North Atlantic Ocean, Cape Hatteras to southern Scotian Shelf. J. Northwest Atl. Fish. Sci. 4, 423 p. Fahay, M. P., & K. W. Able 1989 White hake, Urophycis tenuis, in the Gulf of Maine: spawning seasonality, habitat use, and growth in young of the year and relationships to the Scotian Shelf population. Can. J. Zool. 67:1715-1724. L2 L4 L6 Comyns and Grant Urophycis and Phycis larvae and pelagic j uvemles 221 39°2l' 26' NJ 38°42' 39°2l' U. floridona FEBRUARY 1976 Dn=i H n = o VA 37°Os' 37°3l' O o o z FEBRUARY-MARCH 1977 D n=i E] n = o Dn=28 E3 n=i | ■■*"■ E3 n = 4=n □ n=o E3 n=o U. cirrota FEBRUARY-MARCH 1977 □ n = 2 EJ n = o o n=3 m n = o I I I CI Dl N3 E3 F2 J Station LI L2 L4 L6 Figure 10 Mean abundance of Urophycis flondana and U. cirrata in neuston and bongo collections at stations off Virginia and New Jersey, February 1976 and February-March 1977. n = actual number of larvae collected. Neuston catches denoted by clear histograms, bongo catches denoted by stippled histo- grams. NS = no samples taken. Hermes, R. 1985 Distribution of neustonic larvae of hakes Urophycis spp. and fourbeard rockling Enchelyopus cimbrius in the Georges Bank Area. Trans. Am. Fish. Soc. 114:604-608. Hildebrand, S. F., & L. E. Cable 1938 Further notes on the development and life history of some teleosts at Beaufort, N.C. Bull. U.S. Bur. Fish. 48:505-642. Hoese, H. D., & R. H. Moore 1977 Fishes of the Gulf of Mexico, Texas, Louisiana, and adjacent waters. Texas A&M Univ. Press, Col- lege Station, 327 p. Kendall, A. W. Jr., & N. A. Naplin 1981Diel-depth distribution of summer ichthyoplankton in the Middle Atlantic Bight. Fish. Bull., U.S. 79:705-726. Leim, A. H., & W. B. Scott 1966 Fishes of the Atlantic Coast of Canada. Fish. Res. Board Can., Bull. 155, 485 p. Leviton, A. E., R. H. Gibbs Jr., E. Heal, & C. E. Dawson 1985 Standards in herpetology and ichthyology: Part 1. Standard symbolic codes for institutional resource col- lections in herpetology and ichthyology. Copeia 1985 (3):802-832. Markle, D. F. 1982 Identification of larval and juvenile Canadian At- lantic gadoids with comments on the systematics of gadid subfamilies. Can. J. Zool. 60:3420-3438. 39°2l-28' NJ 3B°42-3Sfl2l' NOVEMBER 1976 V* 37°0S'-37V E ° o o o E FEBRUARY 1976 n=23 FEBRUARY- MARCH 1977 _l a CI Dl N3 E3 F2 Station LI L2 L4 L6 Figure 1 1 Mean abundance of Phycis chesteri in neuston collections at stations off Virginia and New Jersey, during February 1976, November 1976, and February-March 1977. n = actual num- ber of larvae collected. Neuston catches denoted by clear his- tograms, bongo catches denoted by stippled histograms. NS = no samples taken. Markle, D. F., D. A. Methven, & L. J. Coates-Markle 1982 Aspects of spatial and temporal cooccurrence in the life history states of the sibling hakes, Urophycis chuss (Walbaum 1792) and Urophycis tenuis (Mitchill 1815) (Pisces:Gadidae). Can. J. Zool. 60:2057-2078. McGowan, J. A., & D. M. Brown 1966 A new opening-closing paired zooplankton net. Ref. 66-23, Univ. Calif., Scripps Inst. Oceanogr., La Jolla, 56 p. Methven, D. A. 1985 Identification and development of larval and juve- nile Urophycis chuss, U. tenuis, and Phycis chesteri (Pisces, Gadidae) from the Northwest Atlantic. J. Northwest Atl. Fish. Sci. 6:9-20. Methven, D. A., & D. S. McKelvie 1986 Distribution of Phycis chesteri (Pisces: Gadidae) on the Grand Bank and Labrador Shelf. Copeia 1986 (41:886-891. Miller, D., & R. R. Marak 1959 The early larval stages of the red hake, Urophycis chuss. Copeia 1959 (3 ):248-250. Miller, G. L., & S. C. Jorgenson 1973 Meristic characters of some marine fishes of the western Atlantic Ocean. Fish. Bull., U.S. 71:301-312. Musick, J. A. 1973 A meristic and morphometric comparison of the hakes, Urophycis chuss and Urophycis tenuis (Pisces, Gadidae). Fish. Bull, U.S. 71:479-488. Potthoff, T. 1984 Clearing and staining techniques. In Moser, H.G, et al. (eds.), Ontogeny and systematics of fishes, p. 35-37. Spec. publ. 1, Am. Soc. Ichthyol. Herpetol. Allen Press, Lawrence KS. 222 Fishery Bulletin 91(2). 1993 Powles, H., & B. W. Stender 1976 Observations on composition, seasonality and dis- tribution of ichthyoplankton from MARMAP cruises in the South Atlantic Bight in 1973. Tech. Rep. 11, MARMAP Contrib. 118, NMFS MARMAP Prog. Of- fice, Contract 6-35147, and S.C. Wildl. Mar. Resour. Dep., Columbia, 47 p. Serebryakov, V. P. 1978 Development of the spotted hake, Urophycis regius, from the Northwestern Atlantic. J. Ichthyol. [Engl, transl. Vopr. Ikhtiol] 18:793-799. Svetovidov, A. N. 1948Gadiformes. Fauna of the USSR. Vol IX (4), 304 p. [Engl, transl. from Russ. by Israel Prog. Sci. Transl. for Natl. Sci. Found., Wash. DC, 1962.] Taylor, W. R., & G. C. Van Dyke 1985 Revised procedures for staining and clearing small fishes and other vertebrates for bone and cartilage study. Cybium 9:107-119. Wenner, C. A. 1983 Biology of the longfin hake, Phycis chesteri, in the western North Atlantic. Biol. Oceanogr. 3:41-75. Appendix Table 1 Sources of material, collection data, and lengths of hakes used in radiographic analyses of meristics and pterygiophore interdigitation. Standard acronyms for resource collections follow Leviton et al. (1985). Species Collection # Location No. specimens SL(mm) U. earlli USNM 025295 N. Carolina 1 124 USNM 155746 32°34'N,79 o 05'W 1 55 USNM 155747 Wilmington, NC 2 50-60 USNM 226521 32°29'N,79°42'W 3 88-129 USNM 226522 32°29N,79 c '4rW 3 91-122 USNM 226523 32°29'N,79°4rW 1 113 USNM 226524 33°14'N,78°24'W 1 130 USNM 226525 34°14'N,78°24'W 1 82 USNM 226526 32 28'N,79 42'W 4 91-132 USNM 226530 32°29'N,79°40 , W 4 96-157 USNM 226531 32°29'N,79"4rW 5 138-166 USNM 226543 28°48'N,80 o 38'W 1 74 VIMS 06557 Gulf of Mexico 1 195 U. floridana USNM 073010 Key West, FL 1 63 USNM 116729 Beaufort, NC 16 35-49 USNM 131586 26°18 , N,83°09 , W 1 59 USNM 155738 Texas 1 77 USNM 155782 Cape Canaveral, FL 1 86 USNM 155783 St. Augustine, FL 1 67 USNM 156146 Pelican-Stn. 120-5 1 94 USNM 214118 Brickhill Creek, GA • 5 64-75 VIMS 03756 Silver Bay 1 165 VIMS 04142 Brunswick Sound, GA 5 78-113 VIMS 04152 Silver Bay 4 52-85 VIMS 04192 Pensacola, FL 2 81-109 VIMS 04193 Cumberland Id., GA 2 129-184 VIMS 04194 N.Cumberland R„ GA 2 133-157 VIMS 04195 Santa Rosa Sound, FL 1 66 VIMS 04196 Oregon S646 1 185 U. cirrata GCRL 433 29 o 09'N,88°33'W 1 281 GCRL 436 29°22'N,87 o 30'W 1 290 GCRL 525 29°11 , N,88°07 , W 3 318-343 GCRL 2783 Louisiana 4 109-130 GCRL 17534 28°27 , N,90°38 , W 1 145 USNM 115686 22°23 , N,91°45 , W 1 141 USNM 116929 Tortugas, FL 1 140 USNM 155642 29 c 04'N,88 o 44'W 1 114 USNM 218169 29°18'N,88 5rW 1 108 USNM 218192 28 C 58 , N,84°44'W 1 109 USNM uncat. 24°32'N,83 C '36'W 1 197 USNM uncat. 28 59'N,88 48'W 1 198 USNM uncat. 28 : 35'N,91 12'W 2 186-220 P. chesteri USNM 025903 Newport, RI 17 73-98 USNM 026081 Martha's Vineyard. MA 5 68-83 USNM 026097 No data 1 79 USNM 028732 No data 9 58-76 USNM 083821 GA, SC 12 54-65 USNM uncat. Atlantic Arctus Exped. 6 105-147 USNM 092695 No data 1 63 VIMS 05238 36°43 , N,74°39 , W 4 67-150 lomyns and Grant Urophyas and Phycis larvae and pelagic juveniles 223 O 0) c o 0) - t- CO ix 43 cO s o s ^5 O) £ o (N CO c c i ti o. a- cd o s£ ■5, m c bo co co '5-° s cc o CO J ■S E* •Q — ffl i -a c o ! §sf Be! en a, < CD c CO en en >, o £ to ^ — - en -73 CO 3 "- cO C CJ CC '3. 'So B e ■> 5S :§ 00 — 00 _ SSi E rt 33 _~ Al w in VI E CO E2 E _ si3 E E cm •* i I CM CO E °> E7 CM CO E e?s S E^fe Cjl — oi^ E » Al is rf VI ESS ■* Al w E _ £ « E C oo ^ I CO CM O '"' 02 I E S# ^ °° A E^._ c vi — E ^ u en V ~ eSh E (S E ■> — c° I? I CO CO =f> M CM E E co .5 co 2 i .J, CN E 2 E7 CM <> * 2a CC. SB CM — V5 co o T? I Al — CO ^-. A ^-^ E 24 E co E c^ 1 E i CM P CM o E CO W-C ffl tn c o CO *-S fc. c -.2 c3- O c^ r- > « •= eg -JJ a, fu a = 2 ; a p & = - c at ' ^ 'Ej ^ in -^- m I- Of C tn CO >i tx £ o ^ en ^h Is o-co § £%* _V — co M 2 a. t- a co •o;S •a 8 > r m O §•» CJ N CU " in be 'S « 2 ■s-s § 1^1 1; > — S 2 « or a; P3 a> > CD Eh 3 -^ C Cu rs eu co si? CO P u fe o 3 0) o S Ju co C3 ^, ^ CO _• i- O = 13 CO -» co CS O 3]' -a S! c ^) o lis c" « s O .2 «« cn m ° t- CD cfi 1) r >■ I s % z e l u CU CU '2 .5 o T3 R c_i T3 0/ _Q _c *"• ■§ S 3 en ■s I S6o «S © J £ 1 5a ^ »•! (? c S3 -s s g 5 m 2 ^ ? CC s m » - •- cc „° •S S.^ j "C ™ O CD o — c a °- M D 5 3 O u . r c C 5 ^ j5 C =0 o> Q-'o *" .2 O 220mm were mature (Fig. 5). The Gompertz curve was fit to the proportion of mature females by 10 mm length groupings: Y(t)=yj> -e-gll - <„' (ll where Y(t) = proportion mature at length t, and y„, g, and t„ are parameters. This provided the best fit with the least number of parameters. The length at which 50% of females were mature (L 050 ) was calculated to be 154mm. L 050 is reached at age-1 (Davis & West 1992), and all females at age-3 and older were mature (Fig. 5). Spawning season Only data on fish >200 mm were used to describe sea- sonal changes. The main spawning season appears to be September-April (Fig. 6). The presence offish with hydrating oocytes throughout the year indicates the spawning season is protracted, although only four fish with hydrated oocytes were collected in June and Au- gust 1983. Some mature fish had only unyolked oo- cytes in their ovaries during April-August (Fig. 6). The mean relative gonadal index (RGI) peaked in September-October and then declined gradually, reach- ing its lowest value in June (Fig. 6). There was a marked and significant increase in mean RGI during August-October (2x2 factorial ANOVA, month effect, F=309, df 1,339, p<0.001), with the same increase in both years (interaction effect not significant: F=2.4, p>0.1), and there was a weak but significant differ- ence between years (F=4.3, p<0.05). There were sig- nificant differences in RGI between the four periods sampled during the main spawning season, Septem- ber 1982-April 1983 (ANOVA, F=50.5, df 3,706, p<0.001), predominantly due to a linear component (F=130.1, df 1,706, p<0.001) caused by a decline in RGI over this period (slope -0.236, SE 0.021). On the other hand, there was no significant difference in pro- portion of ripe fish throughout this period (likelihood ratio x 2 = 6.3, df 3, p=0.10), nor any indication of a decline in the proportion. This suggests that spawning activity remained at about the same level during the main spawning period but that gonads of individual fish were depleted by successive spawnings. Lunar periodicity in spawning activity Ovaries of mature fish (rc=374) sampled in November- December 1982 were examined histologically for evi- dence of spawning activity. Data were grouped into 1 d periods according to moon age. Two measures of spawn- ing were used: proportion of fish with postovulatory 228 Fishery Bulletin 9 1(2), 1993 >. o c cr CD 60 - 40 20 30 20 - 10 - -I 20 10 - Unyolked hydrated (translucent) M late migratory nucleus (becoming translucent) M yolk granule (opaque) E23 yolk vesicle (translucent) □ perinucleolus (transparent) ■.VAjl Yolked r t m Yolked Ripe Ripe 10 5 - 10 5 - 100 200 300 400 500 600 700 800 900 1000 Oocyte diameter (urn) Figure 3 Oocyte size-frequency distribution and oocyte stage, by 20 um intervals, in Lutjanus vittus ovaries representing the developmental sequence of maturation (see Table 1 for details). An ovary classified as Ripe indicates that it contained late-migratory nucleus- stage oocytes or ripe oocytes. Only oocytes >200um diameter were measured in Ripe stage-5 ovaries, whereas in other ovaries all oocytes >100(im were measured. follicles, and proportion of fish with late-migratory nucleus or hydrated-oocyte stages. The former showed a clear cyclical pattern with two peaks of spawn- ing activity during the month of sampling (Fig. 7). Various sinu- soidal curves were fit to the pro- portion of mature fish with postovulatory follicles sampled each day. The following model, which has a period of 29 d and allows one or two peaks of possi- bly unequal heights per period, was chosen: y = A+B sin (x) + C cos (x) + D sin (2x) + E cos (2x), (2) where t= t hs X 2jt, and t = moon age (d). A regression weighted by the number (n) in each sample was fit to arcsine (angular) root- transformed data. Proportions of 1 were replaced by n-'A divided by n . The fitted model accounted for 85.99r of variance in the data. The smaller peak occurred 3 d af- ter the new moon, and the larger peak 6d after the full moon. No simple pattern of spawn- ing was apparent from the pro- portion of fish with late migra- tory nucleus or hydrated oocyte stages captured each day, pre- sumably because such a pattern was confounded by the effects of time of day (see next section). As postovulatory follicles prob- ably persist for a day, and maybe longer in other species (Hunter & Goldberg 1980, Hunter et al. 1986), their detection does not depend on the time of day of sampling. Diel periodicity in spawning activity The November-December 1982 samples were also examined for evidence of diel periodicity in spawning. Only mature fish Davis and West Reproductive biology of Lutjanus vittus from North West Shelf of Australia 229 1200 - 1000- ~ 800- E =t q 600- "a) o jfSfll » O 400- • »• • • 200- 0- •••/ . • • o ripe • yolked unyolked 00 200 300 400 Length (mm) Diameter of Figure 4 the largest oocyte IMOD) of Lutjanus I'lttus by fish lengt h. Ovaries classified as Yolked and Ripe indi- cate that fis i are mature. Data were collected at height of the spawnin % season (October-February). caught during days of major spawning activity (lu- nar days 2-10 and 18-29) were considered. Two measures that proved useful in detecting changes in spawning activity on temporal scales of less than a day were the proportion of mature fish with ripe- stage ovaries based on whole-oocyte staging, and maximum oocyte diameter (MOD). A clear diel cycle of spawning was evident (Fig. 8). Proportions of ripe-stage fish in samples taken at different times throughout the day were significantly different (likelihood ratio x 2 =69, df 7, p<0.001). Proportions of ripe fish were highest be- tween 08:00 and 14:00 h, and no ripe fish were present by 16:00 h. The mixture of ripe and unripe ovaries between 11:00 and 15:00 h could indicate that spawning for that day had already begun and that some of the fish were spent or that only a portion of fish spawned each day. The temporal dis- tribution of MOD showed a similar pattern. Fish about to spawn that day were clearly separable from other fish by MOD at ll:00h. The MOD's of all but one fish sampled after 15:00 h were the same as nonspawning fish, suggesting that most spawning occurred between 11:00 and 15:00 h on these days, which more or less coincided with rising daytime tides. Postovulatory follicle data from the same subset offish also showed a diel pattern (Fig. 9, page 232). 1.0 0.8 CD i_ | 0.6 o 1 0.4 O 0.2 0.0 J r-r .IfTTTTTTTlJ i i i i ' ' ' i i 1 1 1 ' i i ' 1 1 i ' 1 1 100 150 200 250 300 350 Length (mm) 1.0 0.8 S D « 0.6 E c o t 0.4 o Q- O °- 0.2 0.0 12 3 4 5 6 Age (years) Figure 5 Proportion of total female Lutjanus vittus that are mature by 10 mm length-classes and by age-class during the height of the spawning season. Shown are 95% binomial confidence limits. Age data from Davis & West ( 19921. - r T T r The proportion of fish with early- or late-stage post- ovulatory follicles differed significantly with time of sampling (early-stage likelihood ratio .r 2 =130, df 7, p<0.001; late-stage likelihood ratio .r 2 =131, df 7, p<0.001). Fish with early-stage postovulatory follicles were first detected at 12:00 h. By 17:00 h the propor- tion of fish with early-stage postovulatory follicles had reached its highest level (91%). Thereafter the propor- tion declined, until by 04:00 h none were present in fish sampled. Late-stage postovulatory follicles followed a similar temporal pattern, with a peak that lagged the early stage by 12-14h. The very few late-stage postovulatory follicles in samples at 17:00 h may have resulted from spawning early that day or late spawn- 230 Fishery Bulletin 91(2), 1993 unyolked yolked □ ripe 100 S> >. c To « ra c o CD > . Oven-dried weights of hydrated oocytes were deter- mined on 9 fish. Egg weight did not vary with fish length (F=0.28, df 1,8, p=0.61) or with maximum oo- cyte diameter (F=0.88, df 1,8, p=0.37). The mean dry weight of individual hydrated oocytes was 0.012 mg (SE 0.0005). tr o Q. 1.0 0.8 0.6 0.4 - 0.2 0.0 J -L 9 1 2 2 33 30 12 13 62 18 22 1200 -i F a. 1000 - CD t- m o 800 - a> ^ o o O 600 - F 3 t 400 - 200 J « ft - ! t~:l ' } . i I |« 4:00 8:00 12:00 16:00 20:00 24:00 Time of day Figure 8 Proportion of mature female Lutjanus vittus with ripe-stage ovaries, and maximum oocyte diameter (MOD) in individuals caught at different times of the day. Samples taken during periods of major spawning activity (lunar days 2-10 and 18— 29) during November-December 1982. Oocytes have been as- signed to ovarian maturity stages: Ripe ( ) and Yolked (•). Shown are 95% binomial confidence limits for proportions and number in each sample. Discussion The smallest ripe female observed in this study was 142 mm long, which is close to the length-at-first- maturity of female L. vittus from New Caledonia (Loubens 1980a). In a review of size-at-first-maturity, Grimes (1987) found consistent differences between in- sular and continental species of lutjanids and between shallow and deepwater groups in the ratio of length- at-first-maturity to maximum length. Female L. vittus clearly fall into the continental shallow group that ma- ture early at -44% of their maximum length. The spawning season on the NW Shelf is pro- tracted: While L. vittus spawned throughout the year, the major spawning period appeared to be September- April. Loubens (1980a) determined from GSI and visual staging that L. vittus spawn during October- February in New Caledonia (Loubens 1980a). Lutjanus 232 Fishery Bulletin 91(2). 1993 1.0 CO a> o 1 0.8 CD to i °- 6 H I I 0.4 ■c o a. I 0.2 0.0-1 30 33 (b) 3 -4-M rr 12 18 62 '■Hi 4:00 8:00 12:00 16:00 20:00 24:00 Time of day Figure 9 Proportion of mature female Lutjanus vittus with (a) early and (b) late-stage postovulatory follicles plotted against time of sampling. Samples taken during periods of spawn- ing (lunar days 2-10 and 18-29) during November-De- cember 1982. Shown are 95% binomial confidence limits for proportions and number in each sample. 1.0- 18 T on ripe i i 6 I' 1 g. 0.4- o a. 13 27 0.2- 57 1 ^ 0.0- f 1200- 1 J [ 1 O O o | 1000- E CO o 800- 8 ° R ° o 600- E 3 1 4 00 " CO 2 o I : : 1 ! 200- 4:00 8:00 12:00 16:00 20:00 24:00 Time of day Figure 1 Proportion of mature female Lutjanus vittus with ripe ova- ries, and maximum oocyte diameter (MOD) of individuals caught at different times of the day. Samples taken during 6- 10 October 1988. Oocytes have been assigned to ovarian ma- turity stages: Ripe ( ) and Yolked (•). Shown are 95'i binomial confidence limits for proportions and number in each sample. vittus tends to comply with Grimes' (1987) generaliza- tion that continental species, regardless of latitude, have a restricted spawning season. The pattern of restricted spawning in L. vittus may be linked with the production cycle on the NW Shelf, as has been suggested for other continental species of lutjanids (Grimes 1987). The nutrient source is slope- water washing up onto the NW Shelf in summer when the Leeuwin Current is no longer flowing (Holloway et al. 1985, Tranter & Leech 1987). Enrichment is great- est between December and April. In winter the south- east trade winds blow, there is little stratification of the water column, and the plankton are dispersed (Tranter & Leech 1987). These winds abate by late August or early September, enabling the water col- umn to become highly stratified and the plankton more concentrated (Tranter & Leech 1987). This would re- sult in improved feeding conditions for larvae and co- incides with the start of major spawning activity. During the major spawning period, individual L. vittus spawn a number of times. Serial spawning has been inferred in a number of lutjanids: L. purpureus (De Moraes 1970), L. kasmira (Rangarajan 1971), L. griseus (Campos & Bashirullah 1975), L. synagris (Erhardt 1977), Prist ipomoides multidens and P. typus (Min et al. 1977), Rhomboplites aurorubens (Grimes Huntsman 1980), P. filamentosus (Ralston 1981), Etelis carbunculus (Everson 1984), E. coruscans and Aprion virescens (Everson et al. 1989). While serial spawning appears to be commonplace in lutjanids, the number of batches of eggs spawned each season has not been determined conclusively for any species (Grimes 1987). Our data suggest that L. vittus spawn about 22 times/ mo during late November and early December. If this spawning intensity were maintained throughout the whole spawning period (October-April), then most in- dividuals would spawn about 150 times/yr. Even if spawning intensity were half this rate for the remain- der of the season, thenL. vittus would spawn about 90 times/yr. Greatest spawning activity in L. vittus was shortly after the full and new moons. A lunar rhythm in spawn- Davis and West Reproductive biology of Lutjanus vittus from North West Shelf of Australia 233 80 when fed surplus food once each day in the laboratory (Menzel 1960). Assuming a some- what generous daily ration of 4% for L. vittus and a mean prey energy content of 4.2 kJ /g wet weight (Crisp 1971) would result in an approximate annual repro- ductive efficiency of 4.8% for 90 spawnings and 8% for 150 spawnings. This is similar to an annual reproduc- tive effort in Engraulis rnordax of 8-11% (Hunter & Leong 1981), a much smaller, shorter-lived species. The reproductive efficiency of the goby Pomatoschistus microps is considered to be among the highest ever calculated (Rogers 1988), although its efficiency was calculated using energy consumed over the 16 wk pe- riod of spawning and not the whole year. The reproduc- tive efficiency of P. microps, calculated using annual energy intake, would be -8.7-13.5%. The reproductive efficiency of L. vittus is comparatively high considering that this effort is sustained over many years. Acknowledgments We thank P. Binni and D. Le for laboratory assistance, and all the people who assisted in the fieldwork on the North West Shelf Program. K. Haskard provided sta- tistical advice and modeled lunar periodicity in spawn- ing activity. S. Blaber, J.S. Gunn, R.E. Johannes, and two anonymous referees reviewed the manuscript and suggested many improvements. Citations Alheit, J., V. H. Alarcon, & B. J. Macewicz 1984 Spawning frequency and sex ratio in the Peruvian anchovy, Engraulis ringens. Calif. Coop. Oceanic Fish. Invest. Rep. 25:43-52. Campos, A. G., & A. K. M. Bashirullah 1975 Biologia del pargo Lutjanus griseus (Linn.) de la Isla de Cubagua, Venezuela. II. Maduracion sexual y fecundidad. Bol. Inst. Oceanogr. Univ. Oriente Cumana 14:109-116. Clarke, T. A. 1987 Fecundity and spawning frequency of the Hawai- ian anchovy or nehu, Encrasicholina purpurea. Fish. Bull., U.S. 85:127-138. Colin, P.L., & I.E. Clavijo 1988 Spawning activity of fishes producing pelagic eggs on a shelf edge coral reef, southwestern Puerto Rico. Bull. Mar. Sci. 43:249-279. Crisp, D. J. 1971 Energy flow measurements. In Holmes, N.A., & A.D. Mclntyre (eds.), Methods for the study of ma- rine benthos, p. 197-323. IBP (Int. Biol. Prog.) Handb. 16. Davis, T. L. O., & G. J. West 1992 Growth and mortality of Lutjanus vittus (Quoy and Gaimard) from the North West Shelf of Australia. Fish. Bull., U.S. 90:395-404. De Moraes, N. U. A. 1970Sobre a desova e a fecundidade do pargo, Lutjanus purpureus Poey no nordeste Brasiliero. Bol. Estud. Pesca 10(11:7-20. Erhardt, H. 1977 Beitrage zur biologie von Lutjanus synagris (Linnaeus 1758) an der Kolumbianischen Atlantikkuste. Int. Rev. Gesamten Hydrobiol. 62:161-171. Erikson, D. L., Hightower, J. E., and G. D. Grossman 1985 The relative gonadal index: an alternative index for quantification of reproductive condition. Comp. Biochem. Physiol. 81A ( 1 1:117-120. Everson, A. R. 1984 Spawning and gonadal maturation of the ehu, Etelis carbunculus, in the northwestern Hawaiian Islands. In Grigg, R.W., & K.Y. Tanoue (eds.), Proa, Second symposium on resource investigations in the Northwestern Hawaiian Islands, vol. 2, p. 128- 148. UNIHI-SEAGRANT-MR-84-01, Univ. Hawaii, Honolulu. Everson, A. R., H. A. Williams, & B. M. Ito 1989 Maturation and reproduction in two Hawaiian eteline snappers, uku, Aprion virescens, and onaga, Etelis coruscans. Fish. Bull., U.S. 87:877-888. Forberg, K.G. 1983 Maturity classification and growth of capelin, Mai- Intus villosus ( M. ), oocytes. J. Fish. Biol. 22:485-496. Gladstone, W, & M. Westoby 1988 Growth and reproduction in Canthigaster valentini (Pisces, Tetraodontidae): a comparison of a toxic reef fish with other reef fishes. Environ. Biol. Fishes 21:207-221. Grimes, C. B. 1987 Reproductive biology of the Lutjanidae: a re- view. In Polovina, J.J., & S. Ralston (eds.). Tropical snappers and groupers: biology and fisheries man- agement, p. 239-294. Westview Press, Boulder. Grimes, C. B., & G. R. Huntsman 1980 Reproductive biology of the vermilion snapper, Rhomboplites aurorubens (Cuvier), from North and South Carolina waters. Fish. Bull, U.S. 78:137-146. Hilge, V. 1977 On the determination of the stage of gonad ripe- ness in female bony fishes. Meeresforschung 25:149- 155. Hislop, J. R. G., & M. A. Bell 1987 Observations on the size, dry weight and energy content of the eggs of some demersal fish species from British marine waters. J. Fish. Biol. 31:1-20. Holloway, P. E. 1983 Tides on the Australian North-West Shelf. Aust. J. Mar. Freshwater Res. 34:213-230. Davis and West: Reproductive biology of Lutjanus vittus from North West Shelf of Australia 235 Holloway, P. E., S. E. Humphries, M. Atkinson, & J. Imberger 1985 Mechanisms for nitrogen supply to the Australian North-West Shelf. Aust. J. Mar. Freshwater Res. 36:753-764. Hunter, J. R., & S. R. Goldberg 1980 Spawning incidence and batch fecundity in north- ern anchovy, Engraulis mordax. Fish. Bull., U.S. 77:641-652. Hunter, J. R., & R. J. H. Leong 1981 The spawning energetics of female northern an- chovy, Engraulis mordax. Fish. Bull., U.S. 79:215- 230. Hunter, J. R., & B. J. Macewicz 1980 Sexual maturity, batch fecundity, spawning fre- quency, and temporal pattern of spawning for the northern anchovy, Engraulis mordax, during 1979 spawning season. Calif. Coop. Oceanic Fish. Invest. Rep. 21:139-149. Hunter, J. R., N. C. H. Lo, & R. J. H. Leong 1985 Batch fecundity in multiple spawning fishes. In Lasker, R. (ed.), An egg production method for esti- mating spawning biomass of pelagic fish: application to the northern anchovy, Engraulis mordax, p. 67- 77. NOAA Tech. Rep. NMFS 36. Hunter, J. R., B. J. Macewicz, & J. R. Sibert 1986 The spawning frequency of skipjack tuna, Kat- suwonus pelamis, from the South Pacific. Fish. Bull., U.S. 84:895-903. Jernakoff, P., & K. J. Sainsbury 1990 CSIRO's northern demersal finfish stock assess- ments: 1980 to 1990. Inf. pap. IP/6/90. Bur. Rural Resour., Dep. Primary Ind. & Energy Australia, 169 p. Johannes, R. E. 1978 Reproductive strategies of coastal marine fishes in the tropics. Environ. Biol. Fishes 3:65-84. Kimura, D. K. 1977 Statistical assessment of the age-length key. J. Fish. Res. Board Can. 34:317-324. Leis, J. M. 1987 Review of the early life history of tropical groupers (Serranidae) and snappers (Lutjanidaei. In Polovina, J.J., & S. Ralston (eds.), Tropical snappers and groupers: biology and fisheries management, p. 189- 237. Westview Press, Boulder. Loubens, G. 1980a Biologie de quelques especes de poissons du lagon neo-caledonien. II. Sexualite et reproduction. Cah. Indo-Pac. II (l):41-72. 1980b Biologie de quelques especes de poissons du lagon neo-caledonien. III. Croissance. Cah. Indo-Pac. II (2):101-154. Menzel, D. W. 1960 Utilization of food by a Bermuda reef fish Epine- phalus guttatus. J. Cons. Perm. Int. Explor. Mer 25:216-222. Min, T. S. T., T. Senta, & S. Supongpan 1977 Fisheries biology of Pristipomoides spp. (Family Lutjanidaei in the South China Sea and its adjacent waters. Singapore J. Primary Ind. 5(21:96-115. Mizenko, D. 1984 The biology of the Western Samoan reef-slope snap- per populations of: Lutjanus kasmira, Lutjanus rufolineatus, and Pristipomoides multidens. M.S. the- sis, School Oceanogr., Univ. Rhode Island, Kingston, 66 p. Mori, K. 1984 Early life history of Lutjanus vitta (Lutjanidaei in Yuya Bay, the Sea of Japan. Jpn. J. Ichthyol. 30(4):374-392. Ralston, S. V. D. 1981 A study of the Hawaiian deepsea handline fishery with special reference to the population dynamics of opakapaka, Pristipomoides filamentosus (Pisces: Lutjanidaei. Ph.D. diss., Univ. Wash., Seattle. 204 p. Randall, J. E., & V. E. Brock 1960 Observations on the ecology of epinephaline and lutjanid fishes of the Society Islands, with emphasis on food habits. Trans. Am. Fish. Soc. 89:9-16. Rangaragan, K. 1971 Maturity and spawning of the snapper, Lutjanus kasmira (Forskal) from the Andaman Sea. Indian J. Fish. 18:114-125. Reshetnikov, Y. S., & R. M. Claro 1976 Cycles of biological processes in tropical fishes with reference to Lutjanus synagris. J. Ichthyol. 16:711— 723. Robertson, D. R., C. W. Peterson, & J. D. Brawn 1990 Lunar reproductive cycles of benthic-brooding reef fishes: reflections of larval biology or adult biol- ogy? Ecol. Monogr. 60(3):311-329. Rogers, S. I. 1988 Reproductive effort and efficiency in the female common goby, Pomatoschistus microps (Kroyer) (Tel- eostei: Gobioidei). J. Fish. Biol. 33:109-119. Schnute, J. 1981 A versatile growth model with statistically stable parameters. Can. J. Fish. Aquat. Sci. 38:1128-1140. Starck, W. A. II 1970 Biology of the gray snapper, Lutjanus griseus. In Starck, WA. II, & R.E. Schroeder (eds.), Investiga- tions on the gray snapper, Lutjanus griseus, p. 1- 150. Stud. Trop. Oceanogr. (Miami) 10. Suzuki, K., & S. Hioki 1979 Spawning behavior, eggs and larvae of the lutjanid fish, Lutjanus kasmira, in an aquarium. Jpn. J. Ichthyol. 26:161-166. Takita, T., T. Iwamoto, S. Kai, & I. Sogabe 1983 Maturation and spawning of the dragonet, Cal- lionymus enneactis, in an aquarium. Jpn. J. Ichthyol. 30:221-226. Thresher, R. E. 1984 Reproduction in reef fishes. T.F.H. Publ., Nep- tune City NJ, 399 p. Tranter, D. J., & G. S. Leech 1987 Factors influencing the standing crop of phyto- plankton on the Australian Northwest Shelf seaward of the 40m isobath. Continental Shelf Res. 7(2):115- 133. 236 Fishery Bulletin 91(2), 1993 Wallace, R. A., & K. Selman 1981 Cellular and dynamic aspects of oocyte growth in teleosts. Am. Zool. 21:325-343. West, G. 1990 Methods of assessing ovarian development in fishes: a review. Aust. J. Mar. Freshwater Res. 41:199-222. Wicklund, R. 1969 Observations on spawning of the lane snap- per. Underwater Nat. 6:40. Yamamoto, K. 1956 Studies on the formation offish eggs. Annual cycle in the development of ovarian eggs in the flounder, Liopsetta obscura. J. Fac. Sci. Hokkaido Univ. Ser. 6, 12:362-373. Young, P. C, & K. J. Sainsbury 1985 CSIRO's North West Shelf program indicates changes in fish populations. Aust. Fish. 44(31:16- 20. Young, P. C, J. M. Leis, & H. F. Hausfeld 1986 Seasonal and spatial distribution of fish larvae in waters over the North West Continental Shelf of West- ern Australia. Mar. Ecol. Prog. Ser. 31:209-222. Abstract. -The shape of a size- frequency distribution is the result of age- or size-specific rates of growth and survival, their variability, and seasonal and interannual variation in recruitment. Simulation of size distributions can be used to gain in- sight into the underlying processes that give rise to observed size struc- ture of organisms in the field, but the utility of this approach depends critically on underlying assumptions. Incorrect judgment of the signifi- cance of assumptions can lead to erroneous conclusions concerning the causes of bi- or polymodal distributions. Using the Brody-Bertalanffy growth model and a constant sur- vival rate, bi- and polymodal distri- butions can be generated when re- cruitment is pulsed. Even with as many as 10 recruitment episodes per year, size distributions show several modes. A sampling of the literature indicates that most fish and marine invertebrates have pulsed rather than continuous recruitment; thus, when very little is known about a species, pulsed rather than continu- ous recruitment would be the better assumption when interpreting the shapes of size distributions. Our simulations differ from those conducted by Barry & Tegner (1990) who assumed continuous and con- stant recruitment and focused on changing growth and survival param- eters to explain bimodal size struc- ture. These authors also suggested that their analysis was appropriate for interpreting the dynamics of red sea urchins Strongylocentrotus franciscanus. We have been docu- menting settlement of both red and purple (S. purpuratus) sea urchins. At La Jolla, California, neither species showed continuous settle- ment; rather, both species had pulses of settlement in spring 1990 and 1991. Although age-specific variation in growth or mortality parameters can result in bimodal size distributions, it is more likely that such distribu- tions are caused by seasonal pulses of recruitment. Inferring demographic processes from size-frequency distributions: Effect of pulsed recruitment on simple models Thomas A. Ebert Stephen C. Schroeter John D. Dixon Department of Biology, San Diego State University San Diego. California 92182-0057 Manuscript accepted 22 January 1993. Fishery Bulletin, U.S. 91:237-243 ( 1993). For many organisms, size data are easy to gather and size-frequency dis- tributions are common in the litera- ture. In many cases, they provide the only clues to the underlying dynam- ics of growth, survival, and recruit- ment. Thus, it is understandable that an extensive literature exists con- cerning their analysis. One general research approach has focused on the separation of size distributions into components (e.g., Harding 1949, Cassie 1950, Bhattacharya 1967, Young & Skillman 1975, Macdonald & Pitcher 1979). A second approach has attempted to use size data ei- ther to estimate mortality when growth parameters are known (e.g., Beverton & Holt 1956, Smith 1972, Van Sickle 1977ab, Ebert 1981 and 1987, Sainsbury 1982) or to estimate both growth and mortality param- eters (e.g., Green 1970, Ebert 1973 and 1987, Saila & Lough 1981, Fournier & Breen 1983, Pauly 1987). A third approach has modeled size distributions to gain insight into the underlying processes that give rise to observed distributions (e.g., Craig & Oertel 1966, DeAngelis & Coutant 1982, Barry & Tegner 1990, Hartnoll & Bryant 1990). Simulations of size distributions are metaphors of the dynamic processes that give rise to actual size distributions. The utility of simulation depends critically on the underlying assumptions. If the significance of any of the assumptions is wrongly judged, one may be led to erroneous conclusions concerning un- derlying dynamics. As an approach to explaining bi- modal size distributions, Barry & Tegner (1990) presented a determin- istic model for the development of size distributions that has seven as- sumptions: (1) Brody-Bertalanffy growth, (2) constant rate of mortal- ity, (3) constant and continuous re- cruitment, (4) strict determinism for growth, so ct=0 for all sizes at an age, (5) strict determinism for sur- vival, so rr=0 for numbers at an age, (6) population growth rate per indi- vidual, r, equal to 0, and (7) a stable size distribution equivalent to a stable age distribution. Bimodal size distributions are not possible with these seven assumptions, yet bimo- dality is commonly observed. Accord- ingly, one or more of the assumptions must be violated. Barry & Tegner focused on the assumptions con- cerning growth and survival and concluded that "...bimodality can develop only from an increase in survivorship with age or an increase in the growth coefficient with age, or both." In particular, they argued that size distributions of red sea urchins Strongylocentrotus franciscanus re- quired age- or size-specific changes of the growth-rate constant, K, in the Brody-Bertalanffy equation, the mor- tality coefficient, Z, in an exponen- tial survival curve, or both. 237 238 Fishery Bulletin 91(2), 1993 There are three issues that we would like to ex- plore: (1) Possible causes of bimodal size distribu- tions and, in particular, the consequences of pulsed recruitment, (2) applicability of the Barry & Tegner model to sea urchins in California, and (3) general applicability of the Barry & Tegner model. Size-distribution simulation We simulated several size distributions to show how the Barry & Tegner model works. Growth was mod- eled using the Brody-Bertalanffy equation for individual growth: S t = Sjl-be Kt ) where S t = size at time t after birth or settlement S„ = asymptotic size K = growth rate coefficient , S„-S B (1) (2) S R = size at t=0 when organisms begin to grow according to Eq. 1. Sometimes Eq. 1 is written S t = Sjl-e- Bt -V) where t = time at which size would be b = e Kt °. (3) (4) Cohort survival was modeled so that the mortality rate was constant: N t = N e" zt (5) where N t = number remaining in a cohort at time t N tl = initial number in a cohort Z = mortality rate coefficient. ~ 1.0 CO 2> 0.8- 5 °6 E CD £ 0.4 2 o B 0.0 2 Size at t Equations 1 and 5 were used to generate a number- density distribution that was integrated over segments of arbitrary size to produce a size-frequency distribu- tion. The first step was to calculate sizes at particular ages (Eq. 1) and then to estimate numbers in a cohort that survived to each age (Eq. 5). The number surviv- ing to a specific size was generated using a constant time-interval and Z 0.1 c 0) ZS 0.0 i- ( 0.4- 1 4 8 12 ( 1 4 8 12 0.3- c D | 0.2- 1 0x/year 1 1 OOx/year /■ 0.1 - lxmI 1 1 n n - 4 8 12 4 8 12 Size Figure 2 Integration of N, vs. S, over intervals of 1.0 size unit. Dotted line emphasizes general shape of the envelope of each distribution. (A) 1 recruitment epi- sode/yr; envelope of the distribution has a negative slope. (B) 2 recruitment episodes/yr; envelope of the distribution has high points at smallest and largest sizes. (C) 10 recruitment episodes/yr; general envelope has a positive slope but has modes at small sizes. (D) 100 episodes/yr; envelope has a positive slope. P(x)=^-r(a,t + a,t 2 + a,t 3 ) + e(x) (7) er* 1+Px (8) (9) 110) with a l = 0.4361836, a, = -0.1201676, a 3 = 0.9372980, p = 0.33267, and efxklO 5 . The area under the normal curve, A, from s to s+As is A = P(xL,,-P(xL (11) Areas under the normal curve for each co- hort were reduced by multiplying each area, A, for a cohort by e~ Zt according to Eq. 5. The size-frequency distribution was produced by establishing a 1-unit size- interval and summing parts of all cohorts in each interval. With one recruitment episode/yr (Fig. 3A), the distribution is polymodal. Because the cluster of individuals that are >1 yr is bi- Simulated size distributions take on an appearance much closer to distributions seen in the field when individual sizes are dis- tributed around mean size-at-age (Fig. 3). A coefficient of variation of 0.1 was used for simulation, so rr=0.1|i. Mean sizes were calculated using Eq. 1, and areas under the normal curve were estimated out to 4ct in units of o710. Areas for each size segment were determined by successive subtraction of terms ob- tained from a polynomial ap- proximation of the area under the normal curve and based on a program for the normal distri- bution given by Poole & Borchers (1979), who used an algorithm from Hastings (1955) (Function 26.2.16 in Abramowitz & Stegun 1972). P(x) is the area under the normal curve from the mean, u, to a size, s, given a standard devia- tion of a: 0.4 o> 0.3- ) 4 c 8 12 1 ) 4 D 8 12 0.2- 1 Ox/year : 1 OOx/year 0.1 - 12 Size 12 Figure 3 Results of a simulation with Z=0.5, K=1.0, S~=10.0, and s=0.1xmean. Simulations differ with respect to number of recruitment episodes/yr (range 1-100/yr). (A) polymodal with general envelope with negative slope; (B) polymodal with general negative slope; (C) polymodality still evident but envelope has a positive slope; (D) unimodal with general positive slope. 240 Fishery Bulletin 91 12), 1993 modal, it is clear that this distribution would always be bi- or polymodal. With two recruitment events/yr (Fig. 3B), the distribution again is polymodal and would always be so. With ten recruitment events/yr (Fig. 3C), the modes begin to disappear but the distribution still is weakly polymodal with modes at 0-1, 4-5, and 8-9 size units. With 100 recruitment events/yr (Fig. 3D), the distribution is unimodal. The general shapes are similar to the distributions in Fig. 2, with slopes that initially are negative (Fig. 3A) switching to positive (Fig. 3C, 3D). A combination of pulsed recruitment, coupled with a decaying exponential growth pattern and low mortality, can lead to size distributions with a wide range of shapes. A panoply of size-distribution shapes can be produced with identical values for Z, K, and S. using different frequencies of recruitment. riod of about 3 wk starting in late March 1990, and over a longer period in spring 1991 (Fig. 4). Timing of settlement was the same for both S. purpuratus and S. franciscanus. A few S. purpuratus settled in June 1990, but in terms of influencing the structure of a size-frequency distribution, settlement in 1990 can be considered a single event of short duration. Settlement in 1991 began in late February and continued into early June. The important point is that sea urchin settlement at Scripps Pier was seasonal. Settlement at other sites in California, as well as inside and out- side kelp beds, all showed seasonal settlement (Ebert et al, in prep.). Discussion Observed settlement of S. franciscanus and 5. purpuratus Barry & Tegner (1990) used their model specifically to address bimodal size structure in red sea urchins S. franciscanus. We disagree with their assumption of continuous recruitment and base this on observed settlement data. Starting in late February 1990, we deployed settle- ment collectors at a number of sites along the Califor- nia coast. Wood-handled scrub brushes (model #0115 National Brush Co., Aurora ID were used to evaluate temporal and spatial variability in settlement. Brushes were attached as pairs to a line with two pairs per line. The bottom pair of brushes was suspended lm from the bottom, and the second pair was attached ~20cm farther up the rope. Brushes were tended on a weekly basis at sites within the California Bight and in northern California. One of the sites in southern California was off the end of the pier at Scripps Insti- tution of Oceanography, La Jolla. Following weekly collection, brushes were placed in a sonic cleaner with seawater for ~3 min to remove animals. Newly-metamor- phosed sea urchins have a diam- eter of -500 m; thus, following soni- cation, the water and sediment in the sonicator were strained through 436 m Nitex. Material retained by the screen was then examined us- ing a dissecting microscope, and newly-settled sea urchins were iden- tified and counted. Settlement at Scripps Pier in San Diego County was confined to a pe- Bi- or polymodal size distributions and pulsed recruit- ment are common in the literature. We examined 69 papers that included larval distribution, settlement, or recruitment information for fish, molluscs, anne- lids, bryozoans, crustaceans, and echinoderms. Out of 216 species, only 8 could be considered to have con- tinuous recruitment, and of these only five spider crab species (Hines 1982) appeared to have constant re- cruitment; that is, the same number/mo at all seasons. About 98% of the species failed to meet the assump- tion of constant and continuous recruitment made by Barry & Tegner (1990). As shown by our simulations, pulsed recruitment produced bimodal distributions that were not transi- tory in the sense that distributions showed bimodality at all times between recruitment events. However, were a population characterized by pulsed recruitment, sam- pling could be done in such a manner that the relative magnitude of Z and K could be deduced from a simple 20 10 4-i ° St rongy locen trot us purpuratus 3 00 o St rongy locen trot us froncisconus MAMJJASONDJFMAMJ 1990 1991 MAMJJASONDJFMAMJ 1990 1991 Figure 4 Settlement of purple and red sea urchins, Strongylocentrotus purpuratus and S. franciscanus, on eight scrub brushes: four suspended lm from bottom and four at 1.2m off bottom at Scripps Pier, Scripps Inst. Oceanogr., La Jolla CA (32°52'N). Solid circles indicate animals not identified to species. Ebert et al Size distributions with pulsed recruitment 241 model such as Eq. 6. The changing shapes of the size- frequency distributions for a species with pulsed re- cruitment could be summed and so be made to ap- proximate the shape that would be obtained with continuous recruitment. To obtain a reasonable approxi- mation, it would be necessary to ( 1) take many evenly- spaced samples between recruitment events, and (2) weight the samples with the survival rate, e~ Zt , from the time of recruitment, t. Weighting could be accomplished if accurate estimates of density were known, which, of course, would be the same as know- ing survival. An obvious variant would be the case in which the same area was sampled each time and all individuals were measured. Such a procedure would result in the largest N for the sample immediately following recruitment and the smallest TV for the sample just prior to the next recruitment episode. All samples would be pooled before the size-frequency distribution would be constructed. Approximation of a species with pulsed recruitment to a continuous form could be the same as distribu- tions shown in Figs. 2 and 3. For example, if a species had a single pulse of recruitment and was sampled 10 evenly-spaced times during a year, and each sample was weighted according to the survival rate, then the summed frequency distribution would be C in Figs. 2 and 3. However, if size data were gathered in such a manner that weighting was not automatic, survival rate would have to be obtained by some other tech- nique before size distributions could be summed to approximate continuous recruitment. Techniques for obtaining survival rate, the weighting factor, from size data include those presented by Ebert (1973, 1987), Saila & Lough (1981), Fournier & Breen (1983), and Pauly ( 1987). It must be noted that analysis of a series of size distributions to obtain the weighting factor would provide information on growth as well as sur- vival, and so there would be scant motivation for con- structing a summed distribution. It is not possible to infer the causes underlying an observed size distribution from a single sample or even from several samples that are widely spaced in time. For example, bimodal size distributions can arise from intra-cohort (e.g., Shelton et al. 1979, Timmons et al. 1980) or inter-cohort (Johnson 1976) competition, and the simple models examined here and in Barry & Tegner (1990) demonstrate that similar size distribu- tions can result from very different mechanisms. In a time-series of size distributions, when the smallest mode shifts through time, the simplest explanation for bimodality is pulsed recruitment (e.g., McPherson 1965, Hickman 1979, Dafni & Tobol 1986/87, Davoult et al. 1990). If sampling is adequate and the smallest mode of a bimodal distribution does not shift during the year (e.g., Gladfelter 1978), the most probable explanation is continuous recruitment coupled with high mortality rates for the smallest animals and improved survival with increased size, which is a case that fits the expla- nation for bimodality provided by Barry & Tegner (1990). When size distributions are bi- or polymodal and are presented without a time-series (e.g., Tegner & Dayton 1981, Stein & Pearcy 1982, Wilson 1983), reasonable hypotheses can be formulated, but testing requires additional data. There are numerous examples of pulsed recruitment for sea urchins in California. Size-frequency distribu- tions gathered for purple sea urchins at Papalote Bay, Baja California, Mexico (31°42") (Pearse et al. 1970) may indicate multiple settlement events each year dur- ing 1962-69 because samples from January, April, and June-November all had a mode < 1.0 cm (Pearse et al. 1970, Ebert 1983). However, if growth was very slow at Papalote Bay, as also indicated by the size-frequency distributions, a single settlement episode would ex- plain the data because individuals with a mode at 0.5 cm were observed only in summer and fall samples. Published size data for sea urchins at Whites Point (33°43'N) and Point Vicente (33°44'N) during 1966 and 1967 (Pearse et al. 1970) show recruitment pulses for both species of Strongylocentrotus and for Lytechinus. Recruitment was better in 1966 than in 1967, and small individuals were collected in September 1966 as well as in July and August 1967. Recruitment was not continuous at either Whites Point or at Point Vicente. Finally, our results showing pulsed settlement for red and purple sea urchins corroborate the observa- tion of a single spike of settlement at Naples Reef (34°25'N) off Santa Barbara in May 1986 (Rowley 1989) and the report by Harrold et al. (1991) of two recruit- ment events during a year in central California. The pulsed nature of recruitment means that analysis of size-frequency distributions of Strongylocentrotus spp. should not be based on a model that explicitly requires continuous and constant recruitment (Eq. 6). We have demonstrated that by using fixed growth and survival parameters, it is possible to generate a wealth of size-distribution shapes merely by changing the number of recruitment episodes/yr. We have inten- tionally focused on this aspect of size-distribution shape because we believe that it forms the stumbling block to the application of the Barry & Tegner model. In effect, their model does not provide a convenient way of gaining insight into demographics because, in order to use it to divine the relative magnitude of param- eters, it would be necessary to demonstrate the pat- tern of recruitment for the population being studied. Since the preponderance of field evidence indicates that recruitment generally is pulsed, one cannot "...draw inferences concerning the demographic dynamics of a population. ..simply by observing the shape of its size- 242 Fishery Bulletin 91(2), 1993 frequency distribution" (Barry & Tegner 1990). Fur- thermore, classifying populations as "growth domi- nated" or "mortality dominated," as these authors have done, introduces terms that obscure rather than illu- minate the analysis of size distributions, much in the manner of r- and K-selection comparisons. Size data should be part of every demographic study because they contain a record of the recent past his- tory of a population. Such data ultimately can be used to estimate parameters, such as Z in Eq. 5, that fre- quently are difficult to obtain, or to test assumptions concerning annual variability in recruitment or mor- tality. There currently is no substitute for population studies that include not only size data but also inde- pendent estimates of growth and, where possible, survivorship. Acknowledgments Support for this work was provided by the National Science Foundation, landing tax funds from commer- cial sea urchin fishermen administered through the California Department of Fish and Game, and San Diego State University in the form of a sabbatical leave for the senior author. Citations Abramowitz, M., & I. A. Stegun (editors) 1972 Handbook of mathematical functions with formu- las, graphs, and mathematical tables., 9th ed. U.S. Gov. Printing Off., Wash. DC. Barry, J. P., & M. J. Tegner 1990 Inferring demographic processes from size-fre- quency distributions: Simple models indicate specific patterns of growth and mortality. Fish. Bull, U.S. 88:13-19. Beverton, R. J., & S. J. Holt 1956 A review of methods for estimating mortality rates in exploited fish populations, with special reference to sources of bias in catch sampling. Rapp. P.-V. Cons. Int. Explor. Mer 140:67-83. Bhattacharya, C. G. 1967 A simple method of resolution of a distribution into Gaussian components. Biometrics 23:115-135. Cassie, R. M. 1950 The analysis of polymodal frequency distributions by the probability paper method. N.Z. Sci. Rev. 8:90- 91. Craig, G. Y., & G. Oertel 1966 Deterministic models of living and fossil popula- tions of animals. Q. J. Geol. Soc. Lond. 122:315- 355. Dafni, J., & R. Tobol 1986/87 Population structure patterns of a common Red Sea echinoid {Tripneustes gratilla elatensis). Isr. J. Zool. 34:191-204. Davoult, D., F. Gounin, & A. Richard 1990 Dynamique et reproduction de la population d'Ophiothrix fragilis (Abildgaard) du detroit du Pas- de-Calais (Manche orientale). J. Exp. Mar. Biol. Ecol. 138:201-216. DeAngelis, D. L., & C. C. Coutant 1982 Genesis of bimodal size distributions in species cohorts. Trans. Am. Fish. Soc. 111:384-388. Ebert, T. A. 1973 Estimating growth and mortality rates from size data. Oecologia 11:281-298. 1981 Estimating mortality from growth parameters and a size distribution when recruitment is peri- odic. Limnol. Oceanogr. 26:764 -769. 1983 Recruitment in echinoderms. In Jangoux, M., & J.M. Lawrence (eds.), Echinoderm studies 1, p. 169— 203. AA. Balkema, Rotterdam. 1987 Estimating growth and mortality parameters by nonlinear regression using average size in catches. In Pauly, D., & G.R. Morgan (eds.), Length-based meth- ods in fisheries research, p. 35-44. ICLARM (Int. Cent. Living Aquat. Resour. Manage.) Conf. Proc. 13, Manila, Philippines, and Kuwait Inst. Sci. Res., Safat, Kuwait. Fournier, D. A., & P. A. Breen 1983 Estimation of abalone mortality rates with growth analysis. Trans. Am. Fish. Soc. 112:403-411. Gladfelter, W. B 1978 General ecology of the Cassiduloid urchin Cassidulus caribbearum. Mar. Biol. 47:149-160. Green, R. H. 1970 Graphical estimation of rates of mortality and growth. J. Fish. Res. Board Can. 27:204-208. Harding, J. P. 1949 The use of probability paper for the graphical analy- sis of polymodal frequency distributions. J. Mar. Biol. Assoc. U.K. 28:141-153. Harrold, C, S. Lisin, K. H. Light, & S. Tudor 1991 Isolating settlement from recruitment of sea urchins. J. Exp. Mar. Biol. Ecol. 147:81-94. Hartnoll, R. G., & A. D. Bryant 1990 Size-frequency distributions in decapod Crustacea — The quick, the dead, and the cast-offs. J. Crusta- cean Biol. 10:14-19. Hastings, C. Jr. 1955 Approximations for digital computers. Princeton Univ. Press, Princeton. Hickman, R. W. 1979 Allometry and growth of the green-lipped mussel Perna canaliculus in New Zealand. Mar. Biol. 51:311-327. Hines, A. H. 1982 Coexistence in a kelp forest: size, population dy- namics, and resource partitioning in a guild of spider crabs (Brachyura, Majidae). Ecol. Monogr. 52:179- 198. Johnson, L. 1976 Ecology of Arctic populations of lake trout. Ebert et al.: Size distributions with pulsed recruitment 243 Salvelinus namaycush, lake whitefish, Coregonus clupeaformis, Arctic char, S. alpinus, and associated species in unexploited lakes of the Canadian North- west Territories. J. Fish. Res. Board Can. 33:2459- 2488. Macdonald, P. D., & T. J. Pitcher 1979 Age-groups from size-frequency data: A versatile and efficient method of analyzing distribution mixtures. J. Fish. Res. Board Can. 36:987-1001. McPherson, B. F. 1965 Contributions to the biology of the sea urchin Tripneustes ventricosus. Bull. Mar. Sci. 15:228-244. Pauly, D. 1987 A review of the ELEFAN system for analysis of length-frequency data in fish and aquatic in- vertebrates. In Pauly, D., & G.R. Morgan (eds.). Length-based methods in fisheries research, p 7- 34. ICLARM (Int. Cent. Living Aquat. Resour. Man- age.) Conf. Proc. 13, Manila, Philippines, and Kuwait Inst. Sci. Res., Safat, Kuwait. Pearse, J. S., M. E. Clark, D. L. Leighton, C. T. Mitchell, & W. J. North 1970 Final report. Marine waste disposal and sea ur- chin ecology. Appendix. In Kelp habitat improvement project, annu. rep. (1 July 1969-30 June 1970) p. 1- 93. Calif. Inst. Technol., Pasadena. Poole, L., & M. Borchers 1979 Some common BASIC programs, 3rd ed. OSBORNE/McGraw-Hill, Berkeley CA. Rowley, R. J. 1989 Settlement and recruitment of sea urchins (Strongylocentrotus spp.) in a sea-urchin barren ground and a kelp bed: Are populations regulated by settlement or post-settlement processes? Mar. Biol. (Berl.i 100:485-494. Saila, S. B., & R. G. Lough 1981 Mortality and growth estimation from size data — an application to some Atlantic herring larvae. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:7-14. Sainsbury, K. J. 1982 Population dynamics and fishery management of the Paua, Haliotis iris. II. Dynamics and man- agement as examined using a size class popula- tion model. N.Z. J. Mar. Freshwater Res. 16:163- 173 Shelton, W. L., W. D. Davies, T. A. King, & T. J. Timmons 1979 Variation in the growth of the initial year class of largemouth bass in West Point Reservoir, Alabama and Georgia. Trans. Am. Fish. Soc. 108:142-149. Smith, S. V. 1972 Production of calcium carbonate on the mainland shelf of southern California. Limnol. Oceanogr. 17:28-41. Stein, D. L., & W. G. Pearcy 1982 Aspects of reproduction, early life history, and bi- ology of macrourid fishes off Oregon, U.S.A. Deep- Sea Res. 29:1313-1329. Tegner, M. J., & P. K. Dayton 1981 Population structure, recruitment and mortality of two sea urchins (Strongylocentrotus franciscanus and S. purpuratus) in a kelp forest. Mar. Ecol. Prog. Ser. 5:225-268. Timmons, T. J., W. L. Shelton, & W. D. Davies 1980 Differential growth of largemouth bass in West Point Reservoir, Alabama-Georgia. Trans. Am. Fish. Soc. 109:176-186. Van Sickle, J. 1977a Mortality rates from size distributions. The ap- plication of a conservation law. Oecologia 27:311- 318. 1977b Mortality estimates from size distributions: A critique of Smith's model. Limnol. Oceanogr. 22:774- 775. Wilson, G. D. 1983 Variation in the deep-sea isopod Eurycope iphthima (Asellota, Eurycopidae): depth related clines in ros- tral morphology and in population structure. J. Crus- tacean Biol. 3:127-140. Young, M. Y. Y, & R. A. Skillman 1975 A computer program for analysis of polymodal fre- quency distributions (ENORMSEP). FORTRAN IV Fish. Bull, U.S. 73:681. Abstract.— Patterns in oocyte de- velopment, batch fecundity, and spawning frequency were assessed for black drum Pogonias cromis from Louisiana. We identified histological oocyte stages present throughout a protracted breeding season in 1986- 87. We observed vitellogenesis be- ginning in November, and first postovulatory follicles were detected in February. Atresia of yolked oo- cytes was complete in May. We de- tected recruitment of vitellogenic eggs after the onset of spawning, suggesting indeterminant total fe- cundity. Mean batch fecundity for a 6.1kg female (mean size sampled with hydrated oocytes) was calcu- lated to be 1.6 million hydrated oo- cytes. A field estimate of spawning frequency was 0.311, indicating that a female spawns on average once ev- ery 3d during the breeding season. Sex ratios were skewed with respect to sampling gear during the breed- ing season, suggesting segregation of actively-spawning fish on the spawning grounds. Ovarian development fecundity, and spawning frequency of black drum Pogonias cromis in Louisiana Gary R. Fitzhugh Department of Zoology. Box 76 1 7. North Carolina State University Raleigh. North Carolina 27695-76 1 7 Bruce A. Thompson Coastal Fisheries Institute, Center for Coastal, Energy and Environmental Resources Louisiana State University, Baton Rouge, Louisiana 70803-7503 Theron G. Snider III Department of Veterinary Pathology, School of Veterinary Medicine Louisiana State University, Baton Rouge. Louisiana 70803-8420 Manuscript accepted 28 January 1993. Fishery Bulletin. U.S. 91:244-253 ( 1993). The black drum Pogonias cromis ranges from Argentina to the Bay of Fundy (Sutter et al. 1986) and is the largest sciaenid, up to 66 kg (Hildebrand & Schroeder 1928). A maximum age of 43 yr was recorded for black drum in the northern Gulf of Mexico (Beckman et al. 1990). Fishing pressure on black drum was historically very low, but has in- creased with commercial landings in the Gulf of Mexico rising from 1.9 million kg in 1982 to 4.8 million kg in 1987 (NMFS Natl. Fish. Stat. Of- fice, New Orleans LA 70130). Many temperate fishes are serial spawners with variable production of clutch sizes (Hunter & Goldberg 1980, DeMartini & Fountain 1981, Conover 1985). In particular, large long-lived species may exhibit high variability in reproductive output (Ware 1982). An important management objective with long-lived species is to identify changes in population egg production associated with the harvest loss of older age-classes. For black drum, the potential for exploitation of older age- classes is increased with development of the commercial fishery (NMFS 1986). Despite the commercial value of black drum, relatively little is known of many life-history aspects of this species (Sutter et al. 1986). In the northern Gulf of Mexico, the spawn- ing season has been reported from late-fall to spring, based upon egg and larval distributions ( Jannke 1971, Holt et al. 1985, Ditty 1986) and oc- currence of gravid females (Cody et al. 1985). Peters & McMichael (1990) reported peak spawning in March from Tampa Bay, Florida, based on distribution of larvae and juveniles. Murphy & Taylor (1989) computed size-at-maturation to be 590 mm and 650 mmFL for males and fe- males, respectively, although there have been accounts of small fe- males (<350mmFL) with developing ovaries (Simmons & Breuer 1962, Pearson 1929). Spawning locations have been reported to occur within estuarine bays and in open coastal waters (Pearson 1929, Simmons & Breuer 1962, Jannke 1971, Peters & McMichael 1990). There has been only one estimate of fecundity determined from a single female (Pearson 1929), and no previous estimate of spawn- ing frequency has been made from the adult stock. Our objective is to provide baseline reproductive information from 1986- 87 to be used in assessing popula- 244 Fitzhugh et al.: Reproductive biology of Pogonias cromis in Louisiana 245 tion changes and potential egg production to maintain future fishing harvests. We characterize ovarian de- velopment, seasonal spawning duration, and frequency as determined by ovarian histology and batch fecun- dity in Louisiana coastal waters. Materials and methods We sampled black drum monthly from commercial land- ings during March, June, July, October, November, and December 1986 and July 1987 to obtain reproductive information. We increased sampling effort during the period of reported peak seasonal reproductive activity and sampled 25 commercial landings during February, March, April, May, and June 1987. We also sampled recreational hook-and-line landings during March and July 1986 and April 1987. Landings sampled from in- shore waters (bays and sounds) were primarily taken by gillnet, haul-seine, and hook-and-line. Landings sampled from offshore waters were taken by trawl and purse-seine. In order to contrast size-at-maturity with other stud- ies, we made gross visual classifications of gonads dur- ing sampling. Macroscopic characteristics for classify- ing gonads as mature correspond to Bagenal (1968) and Nielson & Johnson (1983). Female characteristics included the presence of eggs visible to the naked eye and light-yellow to reddish appearance from increased vascularization of the ovary. Characteristics for ma- ture males included white appearance and relative enlargement of testes within the body cavity. Measure- ments included fork length (FL), sex, gutted (viscera removed) body weight (BW), and gonad weight (GW; wet weight blotted dry to nearest 0.1 g). We documented reproductive development by expressing gonad weight as a function of body size using the gonosomatic index (GSI) (Htun-Han 1978, Nielson & Johnson 1983). We held gonads in ice up to 24 h after sampling and then fixed gonads in 10% formalin. One tissue sample was randomly selected from the preserved ovary and placed in an OmniSette tissue cassette. For histologi- cal observation, tissue samples were dehydrated, em- bedded in paraffin, sectioned, stained with Gill's he- matoxylin, and counter-stained with eosin*. We classified oocytes from the prepared histology slides following Wallace & Selman (1981), DeVlaming ( 1983), and Selman & Wallace (1986). These stages include primary growth (PG), cortical alveoli (CA), vitellogen- esis (V), and hydration (H). * Preparation of histological slides, including washing, embedding, sectioning and staining, were completed by the Louisiana State Uni- versity School for Veterinary Medicine, Department of Pathology. To determine relative frequency of oocyte stages, we located a random starting point on a histological sec- tion and counted and staged all identifiable oocytes within a field before moving to a new microscope field using manual stage drive. Field movement was in- ward along the ovigerous lamellae, from the outer tu- nica albuginea toward the center of the ovary, with realignment along a vertical axis. To be counted, >50% of an oocyte must have been within a field of view. We counted and staged a minimum of 200 oocytes from each female and expressed tallies of the four oocyte stages as a percentage of the total count ( Htun-Han 1978, Holdway & Beamish 1985). The Bioquant IV image analysis system software, IBM PC, and Hous- ton Instrument digitizing pad (Hipad model DT-11) were used in conjunction with an Olympus microscope (with video attachment) to facilitate counts and measurements. In addition to relative frequency of developmental stages, we classified each histological section for the presence of postovulatory follicles (POF) and atretic oocytes to aid in determination of spawning frequency. Our atresia classification was modified from that for northern anchovy (Hunter & Macewicz 1985). If no atresia of yolked oocytes was observed, we denoted the ovary as atretic state 0. Tissue sections exhibiting yolked oocytes undergoing atresia at <50%, >50%, and 100% were classed as atretic states 1, 2, and 3, respectively. To estimate oocyte development rate, we related our histological observations of actively-spawning black drum to published accounts of spawning time of black drum (Mok & Gilmore 1983, Holt et al. 1985) and rates of hydration and postovulatory follicle degenera- tion in sciaenids (DeMartini & Fountain 1981, Brown- Peterson et al. 1988). With this time-calibrated histo- logical information, we estimated the hours from spawning for females displaying yolk coalescence, hy- dration, postovulatory follicles, and atresia, and clas- sified day-0, day-1, and nonspawning females. Our es- timate of seasonal spawning frequency was determined by taking the average of the fractions of day-0 and day-1 spawning females relative to total females ob- served histologically with vitellogenic oocytes. Batch fecundity, the number of hydrated oocytes which com- prise the leading "batch" of eggs immediately prior to spawning, was determined from formalin-fixed tissue samples taken from each visibly-hydrated ovary. We took replicate ovarian tissue samples (l-2g) from an- terior, mid-, and posterior regions of left and right lobes. In order to obtain 100-300 hydrated oocytes, tissue subsamples of 90-100 mg (weighed to the near- est 0.05 mg) were placed on a slide, glycerin added, and hydrated oocytes counted (Hunter et al. 1985). After observation of histological sections, any ovaries 246 Fishery Bulletin 9 1(2), 1993 with postovulatory follicles, indicating onset of spawning and possible shedding of eggs, were eliminated from further analysis of fecundity. To determine the precision of batch fecundity meth- odology (Hunter et al. 1985), we compared oo- cyte counts per unit weight within ovaries using a two-way analysis-of-variance model, SAS GLM procedure (SAS 1985). Results Sampling, sex ratio, and maturity We recorded information on length, sex, and gear for 236 male, 108 female, and 36 immature black drum. In addition, we collected capture data, ovar- ian samples, and measurements, including somatic and gonad weights used for GSI analysis, from 198 males and 296 females. Ovarian histology sec- tions were made from 234 mature females ran- domly sampled through June 1987, and were used to determine relative frequencies of oocyte stages (Fig. 1). Of these females, 25 were visibly hy- drated, possessed no postovulatory follicles, and 100 -, o 100 H 3.8 3.8 8 ° O 100 ra 35.3 45.6 28.9 3.7 18.5 17.3 1n . ^3 « ^ ^ 1 M „ ^ ^ 93.3 ?S=9 -,¥£ 91.5 i PG 64.9 liiiliil Oct Nov Dec Feb Mar Apr May Jun 1986 1987 Figure 1 Percent oocyte stage by month for 1986-87 based on point counts of -200 oocytes/female black drum Pogonias cromw. Stages include primary growth (PG), cortical alveolar (CA), vitellogenic (V), and hydrated (H). Number of females examined = 22 Oct, 23 Nov, 23 Dec, 69 Feb, 32 Mar, 24 Apr, 22 May, 19 June. were used to estimate batch fecundity. Our sample mean length was 761mmFL, with adults measuring 650- 900 mmFL comprising 89% of individuals used in the study. We found apparent differences in sex ratios from inshore landings (gillnet and haul-seine) and landings from an offshore trawl fishery during the reproductive season (Table 1). Trawl catches were dominated by males, while gillnet and haul-seine samples were dominated by females (Table 1). For months just before and after the reproductive season (October, June, July), females also dominated gillnet and haul-seine landings, but ratios were less divergent. The trawl fishery was not active at these times, but samples of offshore fish were taken from a purse-seine landing (June 1986) and numbers of females and males were nearly equal (0.92:1.0). We applied a x' 2 contingency analysis to test sex ratios by gear type. The x 2 statistic was significant during the reproductive season (December-May), leading to rejec- tion of the null hypothesis that gear type and sex ratio are independent (Table 1). We assumed that gears were not selective for sex but reflected actual sex ratios in the locali- ties fished. Therefore, the skewed sex ratio suggests a seg- regation of sexes during the reproductive period. Female: male ratios were more divergent for commercial gears dur- ing the months of November-May than October, June, and July (Table 1). Males and females were first mature at 600-640 mmFL as defined by the size at which individuals exhibit develop- ing and mature gonads from gross visual inspection (Nielsen & Johnson 1983). All black drum >640mm were mature. All fish <590 mm were immature, but sample size during spawning season was small (n = 18 females and 11 males at 460-590 mm). Table 1 Chi-square contingency analysis of black drum Pogonias cromis sex ratios for commercial gears represented by a minimum of 20 indi- viduals. Observed Female:Male ratio Gear Male Female Total Nov, Dec, Feb, Mar, Apr, May Gillnet 42 105 147 1:0.4 Haul-seine 39 65 104 1:0.6 Trawl 267 126 393 0.47:1 Total 348 296 644 Oct, June, July (X 2 =1.18, df 2, 0.950% of the total yolked oocytes from individuals sampled 12 May 1987, probably signaling a decline in spawning (atretic state 2). By this date, all yolked oocytes were undergoing atresia in 12 out of 22 females examined histologically, and all 22 females exhibited some atretic yolked oocytes. By 12 June, atresia of yolked oocytes was complete for all females examined (rc = 19), and only gonadotropin-independent PG oocytes remained (atretic state 3). Table 2 Number of female black drum Pogonias cromis in reproduc- tive condition based on h stological staging for determination of spawning frequency. Day-0 designated females were ac- tively spawn ng or close to onset of spawning. Day-1 desig- nated females have been sampled at least 6h after spawning (see Fig. 3). Day-0 Day-1 spawning spawning Total mature Date females females females* 11/11/86 2 23 12/16/86 5 23 02/03/87 9 15 02/16/87 19 22 02/20/87 18 5 20 02/24/87 2 02/27/87 6 6 12 03/06/87 4 4 03/23/87 13 12 16 03/27/87 4 10 10 04/06/87 1 8 04/24/87 3 2 16 05/12/87 1 22 Total 80 40 193 Proportion of total 0.415 0.207 Average proportion of total 0.311 ?d histologically with vitellogenic oocytes *Total observ The gonosomatic index (GSI), which is gonad weight expressed as a fraction of body weight, is a common measure of gonad development used to document sea- sonal changes (Nielson & Johnson 1983). We observed a marked peak in GSI for both males and females in March (Fig. 2). Females exhibited the most dramatic change in gonad weight as the season progressed, with mean gonad weight increasing to >8% of eviscerated body weight. Both sexes followed a similar pattern with respect to time of onset, peak, and decline of GSI (Fig. 2). The gonosomatic index corresponded well with histological observations (Figs. 1, 2). Females exhib- ited an increase in GSI above "resting" levels in No- vember, when CA and V oocytes were present. Monthly peaks in GSI during February and March resulted from the presence of hydrated oocytes and an increase in the proportion of vitellogenic eggs. Increased atresia of yolked oocytes observed histologically in April and May, associated with decreased spawning, produced a decrease in GSI (Fig. 2). In June, when all yolked oocytes were atretic, female GSI had dropped to 0.87, 248 Fishery Bulletin 91(2), 1993 13 - 36 12 - 11 - 10 - 71 M 9 - 27 »- 8 - 55 G c o 2 7 - 6 - 5 - 4 - 3 - 2 - 1 - - Female / \ 23 23 / t 4 5 22 ,r^ /n\ - u ir 68 24 \38 1 V" 27 20 i 7 i2^rt Ma,e u WjHi i i i i i i i Jun Jul Oct Nov Dec Feb Mar Apr May Jun Jul 1986 1987 Figure 2 Mean value for gonosomatic index (GSI) for adult black drum Pogonias cromis (±1 SD). Sample sizes are indicated by numbers on graph. and female GSI reached the observed minimum value of0.84byJuly(Fig. 2). Spawning frequency In order to estimate proximity to time of spawning and designate females as day-0 and day-1 spawners for spawning frequency calculation (Table 2), we con- structed a time-scale of coalescence-stage oocytes, hy- drated oocytes, and postovulatory follicles based on literature reports and our observations (Fig. 3). From previous work on related sciaenids, it is known that lipid coalescence, germinal vesicle migration, and yolk coalescence occur beginning in morning samples with hydration becoming evident as the day progresses (DeMartini & Fountain 1981, Fitzhugh et al. 1988, Brown-Peterson et al. 1988). Although exact capture times for some black drum were not known, females exhibiting germi- nal vesicle migration and yolk coalescence were commonly taken in haul-seine and gillnet sets which were typically landed during morning hours. A follicle, comprised of an inner layer of epi- thelial granulosa cells and an outer layer of the- cal cells, surrounds each hydrated oocyte. Fol- lowing ovulation, POFs were present as the evacuated follicle remaining in the ovary. Recent POFs were denoted by linear arrangement of the granulosa cell layer and apical location of very prominent nuclei. These cellular arrangements imparted the appearance of a well-defined lu- men and convoluted shape to the POF and were observed from ovaries sampled between 2400 and 0300 h from preliminary samples taken by hook-and-line in March 1986. We con- cluded that spawning occurred earlier that same night, and used these March 1986 samples as an example of recent POFs. Three females sampled from trawl landings in 1987 with recent POFs also had fully-hydrated eggs in the lumen of their ovaries, indicating active spawning and coinciding with reports of onset of spawning after dusk (Mok & Gilmore 1983, Holt et al. 1985). From interviews of commer- cial fishermen, capture of black drum in trawls often occurred after dusk and throughout the night, with fish being placed into ice as they were captured. We routinely sampled black drum the morning following their capture, and therefore it is likely that the recent POFs we observed are from fish captured up to 8 h after spawning (Fig. 3). Over the 1986-87 spawning season, we sampled limited numbers of females bearing POFs (50 females taken from 8 different samples) in- dicating that duration of POFs may be brief. Hydra- tion-stage oocytes, occurring together with visibly- degenerated POFs, were evident in only 19 females. Additionally, only 1 female contained recent POFs as well as degenerating POFs. Older degenerating POFs were similar in appearance to 24h-old follicles illus- trated in Hunter et al. (1986) from skipjack tuna Katsuwonus pelamis spawning at 23-24°C. Therefore, POF duration may be limited to 24-48 h following ovu- lation (Fig. 3). Coalescence I Hydration Ovulation and Spawning □ Day -12 -10 -8-6-4-2 2 4 6 Hours from onset of spawning Day 1 8 10 12 14 16 18 20 22 24 26 28 30 Day 2 32 34 36 I — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — | — i — i — i 0000 0400 0800 1200 1600 2000 Time of Day Figure 3 Generalized time-scale of final oocyte maturation and spawning of black drum Pogonias cromis, determined from observations and lit- erature reports (see text). Fitzhugh et al.: Reproductive biology of Pogonias cromis in Louisiana 249 We estimated spawning frequency by examining ovarian tissues of 193 mature females (vitellogenic oo- cytes present) sampled over the spawning season. We define spawning season as the period during which spawning condition was histologically evident, from first evidence of coalescence of vitellogenic oocytes on 11 November 1986, until atresia of vitellogenic oocytes was occurring in all females encountered on 12 May 1987 (Table 2). The numbers of day-0 and day-1 spawn- ing females was 80 and 40 (frequencies = 0.415 and 0.207) for this period, respectively. This resulted in an overall seasonal frequency of 0.311, or each female spawning on average once every 3d (Table 2). From November 1986 through May 1987, 50 of 193 females (26%) contained POFs. This method of estimating spawning frequency would correspond to a spawning frequency of once every 4d, or a total of 46 times, during the spawning season. Batch fecundity We determined batch fecundity for 25 visibly-hydrated females which possessed no recent POF in histological sections that would indicate egg shedding. The range in batch fecundity was 7.4xl0 5 hydrated oocytes for a 4.3kg female, to 3.8X10 6 hydrated oocytes for a 4.8kg female, indicating wide variation in fecundity based on body size (Fig. 4). The number of hydrated oocytes/ g of ovary weight ranged from 1046 to 4902. Based on eviscerated body weight, the number of eggs/g ranged from 131 to 793. The mean value for batch 5000 - q O c o a> 4000 - 3000 - 2000 - °-^" o o o o (J m c 2 1000 - - -1000 - OO I 1 ' -1 1 — — 1 1- 1 1 4000 6000 8000 Gutted Body Weight (g) 10000 Figure 4 Batch fecundity and eviscerated body weight with 95 r /r confidence i terval for hydrated female black drum Pogonias cromis, sampled February and March 1987. fecundity was 1.6xl0 6 for a mean eviscerated weight of 6. lkg. Location of tissue samples Black drum possess relatively large gonads for teleosts, and we sampled ovaries weighing 0.4-1.6 kg laden with hydrated oocytes. No significant differences were de- tected between positions within a lobe or between right and left ovarian lobes for number of hydrated oocytes/g of ovarian tissue (Table 3). However, counts of hydrated oocytes/g were lower for the anterior posi- tion (jc 1780) than for either mid- or posterior ovarian regions (x 2013 and 1928, respectively) (Table 3). Discussion We noted differences in sex ratios between samples collected by various fishing gears during the period of reproductive development and spawning, November- May. Landings by offshore trawlers (where spawning females were found in February and March) were clearly dominated by males. Females dominated in in- shore samples from all gears (primarily gillnet and haul-seine landings), but this female dominance was not as prevalent outside the breeding season during October, June, and July. These divergent ratios sug- gest sexes are spatially segregated for periods during the reproductive season. This difference in sex ratios has been noted in other fisheries where a higher pro- portion of males attend those females in ac- tive spawning condition on the spawning grounds (DeMartini & Fountain 1981, Hunter & Goldberg 1980). In a synthesis of previous studies, Simmons & Breuer (1962) reported age-at-maturity for black drum to be 2 yr at 320 mm based on scale increments, length frequencies, and gross as- sessment of ripe females (granular roe ob- served) in Texas waters. Pearson (1929) found drum as small as 270 mm with developing ova- ries. Murphy & Taylor (1989) provide evidence for larger sizes-at-maturity, with males ma- turing at 590 mm (4-5yr old) and females ma- turing at 650 mm (5-6 yr old). Our findings of age-at-maturity agree with Murphy & Taylor's (1989) estimates. Mature females and males occurred at a size-range of 600-640 mmFL, cor- responding to ages 3-8 yr (Beckman et al. 1990). A mature female with hydrated eggs was observed as small as 625 mmFL; this indicates that size- or age-at-maturity for northern Gulf of Mexico stocks is greater than previously estimated by Simmons & Breuer 250 Fishery Bulletin 91(2). 1993 Table 3 Effect of location of tissue samples of black drum Pogonias eromis for hy- drated oocyte counts per unit of ovary weight (gl. Locations are anterior (i), mid (ii), and posterior ( iii) of jvarian lobes. Mean, SE of oocytes/g, and number of tissue samples Right lobe Left lobe Both lobes Locations .r SE n * SE n x SE n i 1714 91 25 1846 170 25 1780 96 50 n 2025 245 25 2000 199 25 2013 156 50 iii 1944 149 25 1911 180 25 1928 116 50 Total 1895 100 75 1919 105 75 Two-way ANOVA Source of variance df SS MS F PR>F Lobe 1 22490 22490 0.03 0.87 Segment 2 1385533 692767 0.87 0.42 Interaction 2 213348 106674 0.13 0.87 Error 144 14701071 796535 Total 149 116322441 (1962) and Pearson (1929) for black drum from Texas waters. Multiple oocyte stages were present throughout the 1986-87 spawning season, including primary growth (PG), cortical alveolar (CA), and vitellogenic (V) stages. Primary-growth oocytes were present year-round and are a gonadotropin-independent stage (Wallace & Selman 1981, DeVlaming 1983). Oocyte recruitment from this PG population and onset of seasonal oogen- esis was signaled by the appearance of CA oocytes in October. Vitellogenic oocytes were noted in November, and both CA and V oocytes persisted until May, indi- cating the potential for a protracted spawning season. Descriptions of the breeding season for black drum are varied. Gonad development and possible spawning have been reported during the summer (Pearson 1929, Cornelius 1984 cited in Cody et al. 1985). Our June, July, and October samples did not indicate that spawn- ing was evident. By November, reproductive develop- ment for two females (yolk coalescence and final oo- cyte maturation in histological sections) indicated the potential for spawning to occur. Capture of black drum larvae from offshore Louisiana waters has been re- ported as early as December (Ditty 1986). Pearson (1929) reported the primary spawning season was Feb- ruary to May. Cody et al. (1985) described seasonality of gonad development of black drum in Texas and re- ported gravid females during November-April, with spawning or spent stages predominant dur- ing February-April. From Florida, Murphy & Taylor (1989) and Peters & McMichael (1990) also report the reproductive season ranges from November to April with Feb- ruary-March spawning peaks. All the changes noted with the onset, peak, and decline of the spawning season were reflected both in histological samples of ovaries and in gonad weight changes relative to body weight (GSI) for females. Increase in proportion of vitellogenic oo- cytes was associated with GSI increase in November and December. Appearance of hydrated oocytes and postovulatory follicles in February and March coincided with high- est values for GSI. Subsequent appearance of atretic vitellogenic oocytes in April, and increase in atresia in May and June, were associated with declines in gonad weights reflected by decreasing GSI. The GSI sea- sonal pattern for males coincided with the pattern for females, and suggests a synchrony in development of reproductive states for both sexes. However, mean GSI values must be interpreted with caution. For the same stage of oocyte development, a larger individual may exhibit proportionally larger ovaries and a greater GSI value than a smaller individual (DeVlaming et al. 1982). We apply the GSI here only as a relative measure of changes in reproductive con- dition over the spawning season, and not as a specific measure of reproductive readiness or histological stage. Black drum spawn in inside (estuarine, bay) as well as in outside (coastal waters seaward of inlets) waters. Pearson (1929) indicated spawning took place in Gulf waters off Texas, although Simmons & Breuer (1962) presented evidence for spawning in estuaries. Osburn & Matlock (1984) present evidence from tagging stud- ies for a "quasi-permanent" movement of black drum >4yr from bays to the Gulf, where they may act as spawning stock. Jannke (1971) cited evidence of estua- rine spawning in the Florida Everglades, but also in- dicated that spawning occurred outside the estua- rine portion of the park. Peters & McMichael (1990) report that spawning in the Tampa Bay region was likely to have occurred both inside and outside the bay. We also noted spawning activity over a gradient from offshore to inshore. Examination of black drum in hydrated condition from trawl landings during Feb- ruary and March indicated spawning was occurring in coastal waters off Louisiana. During late March, how- ever, females in hydrated condition were taken by haul- seine from inshore estuarine waters east of the Mis- Fitzhugh et al.: Reproductive biology of Pogonias cromism Louisiana 25! sissippi River. Drumming behavior associated with spawning was noted in Caminada Pass, Louisiana in April, further documenting inshore reproductive activ- ity (Donald Baltz, Coastal Fish. Inst., LA State Univ., Baton Rouge, pers. commun. ). The dynamic of changing spawning locations may have a seasonal component related to water tempera- ture. Mok & Gilmore (1983) analyzed sound produc- tion from black drum in Florida waters and noted "loud drumming," which they associated with spawning, oc- curring at 18-20°C. They also noted cessation of this drumming during a temperature drop to 13-15°C. Pe- ters & McMichael ( 1990) provided more direct evidence for onset of spawning at 16-20°C. By correlating sea- sonal water temperature with larval birthdates, peak births were calculated to have occurred in March when temperatures reached 21-24°C. Although we did not have precise locations and temperatures for commer- cial catches, our samples of hydrated-oocyte and POF- bearing females indicate that spawning may have pre- dominated in outside waters in February (e.g., trawl landings) and moved to inside waters as seasonal tem- peratures increased (haul-seine and gillnet landings in March and April ). Other factors not examined may influence spawning, including moon phase and tidal period (Peters & McMichael 1990). Postovulatory follicles probably last longer than 24 h at sea temperatures encountered in coastal Louisiana waters in February and March (19-22°C). Hunter & Macewicz (1985) found POFs for 3-4 d from northern and peruvian anchovies (Engraulis mordax and E. ringens) spawning at 13-19°C. Based on a higher spawning temperature of 23-24°C, Hunter et al. ( 1986) found 24 h-old follicles in skipjack tuna that appeared similar to those in northern anchovy held 48 h. Our day-1 POFs appeared similar to 24 h POFs for skip- jack tuna shown in Hunter et al. (1986). We estimated black drum follicle duration to be at least 32 h old, due to the presence of recent POFs and older-degenerating POFs together in the same histological sections (i.e., 0-8 h plus 24 h, respectively) which is consistent with spawning on successive nights. If follicles are identifi- able well past 24 h, our estimate of average duration between spawning would increase. Our estimates of spawning frequency, once every 3 or 4d, are similar to other sciaenid species. Tucker & Faulkner (1987) report a daily spawning fraction of 0.35 (once every 3d) for captive spotted seatrout. Brown-Peterson et al. (1988) calculated an average daily percentage of wild seatrout in ripe condition to be 27.5% (indicating spawning once every 3.6 d) over a 6 mo reproductive season. Red drum have displayed a spawning fraction of 0.68 (once every 1.5 d) in captiv- ity over a 76 d period (Arnold et al. 1977). The pattern of appearance of vitellogenic oocytes sup- ports our contention that oocyte recruitment contin- ued during the reproductive season. Therefore, the yolked-oocyte population was not deterministic or rep- resentative of annual fecundity (Hunter et al. 1985). Of females examined histologically in February, 73% showed evidence of recent spawning (day-0 females), yet the proportion of vitellogenic oocytes from females did not reach a maximum until March. In contrast, a species with a determinant oocyte development pat- tern could exhibit multiple spawns but the proportion of vitellogenic oocytes would decrease following onset of spawning (Hunter et al. 1985). With continuous recruitment of batches of oocytes, the traditional method to determine fecundity by enumerating vitellogenic oocytes prior to onset of reproduction (e.g., Bagenal 1968) would underestimate potential fecun- dity. This method necessitates counting of hydrated eggs just prior to ovulation, i.e., determining batch fecundity and number of spawns in the season (Hunter etal. 1985). Previously, little fecundity information has been re- ported for black drum (Sutter et al. 1986). Pearson (1929) estimated fecundity at 6 million eggs for one 110 cm gravid female (107cmFL, 16.4 kg eviscerated weight)* based on extrapolation of wet weight for 60 eggs. Our computation of batch fecundity was 1.6 mil- lion eggs for a 6.1kg female (mean eviscerated weight of 25 females). Figure 4 and comparison with Pearson's result suggest that batch fecundity is a function of body size, but we found this relationship to be quite variable (Fig. 4). While these data do not appear to fit a linear relationship as closely as for smaller sciaenids (DeMartini & Fountain 1981, Brown-Peterson et al. 1988), we only sampled hydrated females measuring 660-876 mmFL. Because females >1000mm are occa- sionally landed (e.g., Beckman et al. 1990), a broader range of sizes could illuminate the functional relation- ship between size and fecundity. Batch size may vary also with the reproductive period and may be higher earlier in the spawning season. Our sample size was too small to detect changes in batch fecundity over the breeding season. However, the pattern of vitellogen- esis and GSI provides evidence that spawning peaked in March. The proportion of females in spawning con- dition was also highest in March. Conover ( 1985) dem- onstrated a quadratic batch-fecundity relationship for Atlantic silversides during the breeding season. He postulated that this pattern may occur when optimal * Calculated from TL-FL regression reported in Murphy & Taylor 1 1989). The relationship between eviscerated body weight (BW) and fork length (FL) is given by the equation: Ln BW = 2.8835 Ln FL - 10.382 17mmTL. The following study will focus on larvae 11-16 mm in length, be- cause the percent inflation can vary with lighting condition. Equal numbers of larvae were not sampled in each size-category ( Fig. 2), as most larvae were 11-12 mmTL and there were few >15mmTL. This size distribution 100- (II) (8) (3) (12) 80- y/m 60- (186) 40- A 2681 (105) 20- (32) -y/-f (79)/ f / "? ,-'f*4) ,-'(65) '('37) (4) V' (9) ',(1) -i 1 1 *— (1) (1) —I 1 f- 1 10 12 14 16 18 Total Length (mm) 20 Figure 1 Percentage of larvae inflating their swimbladder versus size for all larvae in the swimbladder inflation versus light inten- sity experiment (Fig. 3). The dashed line shows the percent- ages before exposure to different light conditions, and the solid line is the experimental percentages. 256 Fishery Bulletin 91(2). 1993 30 /' '\ 20 /■' v\ 10- v\ y/"-i r 1 ' ' — ^^^rs^^ — - F=f , 9 II 13 15 17 19 21 Total Length (mm) Figure 2 Percentage of the total number of larvae of the different total length sampled for the swimbladder inflation vs. light intensities experiment (Fig 3). Dashed line shows larvae sampled before exposed to different light levels (control; total n = 193), whereas the solid line is larvae exposed to different light conditions (experimental; rc=834). was consistent for all experiments. Since the proba- bility of swimbladder inflation was not equal for each size-category (Fig. 1), results will be biased to responses of the most-abundant size if all size- classes are grouped together. Thus, detailed analy- ses should only consider individual size-classes. The most-abundant size (llmmTL; Fig. 2) will be used for this purpose. The percentage of fish inflating their swim- bladders increased as they were exposed to lower light intensities (Fig. 3). The highest light intensity to induce a significant increase in the proportion of fish with inflated swim- bladders (threshold intensity) varied slightly with fish size. For llmmTL larvae (Fig. 3A) the threshold intensity was -6x10' ! photons cnr 2 s _1 , whereas for 12-16 mmTL larvae, it was 1 log unit higher (Fig. 3B). For both size- groups, the proportion filling their swim- bladders at ~10 13 photons cm -2 s _1 and lower light intensities was not significantly different from the proportion in darkness. The variation in swimbladder volume with light intensity was considered in detail for llmmTL larvae (Fig. 4). Mean volume increased as light intensity decreased, but the difference was not significant between the initial mean volume and that in darkness due to large variances (Fig. 4). Similar results were also obtained for 12 and 13 mmTL larvae. In contrast, swimbladder volume increased proportion- ately with larvae size (Fig. 5). When the relationship be- tween mean volume (V) in darkness and total length (L) was expressed as the allometric equation (V=aL b ), the slope of the regression (b) equaled 5.31 (r 2 =0.99,p<0.0001). Means were calculated for larvae of each size in all conditions, because volume did not change with lighting condition (Fig. 4). Since volume changed with larval length, volumes could not be averaged for larvae of different lengths. Timing of swimbladder inflation The timing of swimbladder inflation was measured upon transfer from rearing-light intensity to darkness. By pro- ducing the maximum rate of intensity change, we assumed the maximum rate of inflation should be evoked. Results were combined for larvae 11-16 mmTL because the pro- portion inflating in darkness for 11 mm larvae was not statistically different from the proportion of 12-16 mmTL larvae (Fig. 3). A significant increase in the proportion with inflated swimbladders was evident after 5 min in dark- ness (Fig. 6). The maximum percent inflation was reached within 20 min. The proportion of fish with inflated swim- bladders then remained relatively constant for about the next 1.5 h. Endogenous rhythm in swimbladder inflation The percent inflating prior to placement in darkness re- mained low throughout the 24h sampling interval, which 80 L. 60 0) 1 40 -Q E 20- s (/) o> £ 80- o c 60- A II mm DARK N. INITIAL * B I2-I6mm DARK * .. — — — "^V. ,0 °" 40 ^^\» 20- # INITIAL • I0 10 10" 10* I0 13 io 14 I0 15 Light Intensity (photon cm" 2 s"') Figure 3 Percentage of larvae llmmTL (A) and 12-16 mmTL (B) inflating their swimbladder when exposed to different light intensities and darkness (dark). "Initial" is the percentage sampled shortly after removal from the rearing tank. Average sample sizes for each condition in A and B are 44 and 58, respectively. Asterisk indicates the highest light inten- sity to evoke a response that was significantly (p<0.05l greater than the initial response. Forward et al.: Swimbladder inflation of Brevoortia tyrannus 257 DARK 1 3" t >v T 1 |i I it) (20) o ro | 0.9- 1 (33) \ INITIAL (321 (16) \ o> | 7^ 1 Vo V S 0.5- o E * 0.3- (3) (8) 0.1- (3) io 10 io" io 12 io' 3 io 14 io 15 Light Intensity (photon cnrf 2 s~') Figure 4 Swimbladder volume of llmmTL larvae shortly after removal from the rear- ing tank (initial) and when exposed to different light levels and darkness (dark). Means and standard errors are shown. Number below each plot is the sample size. (Fig. 7A). This response level continued into the time of the next light-phase. Mean swimbladder volume of llmmTL larvae varied over time, but the maxi- mum and minimum means over the first solar day were not significantly different due to the large variances (Fig. 7B). Discussion indicates there was no endogenous rhythm in in- flation in constant high light conditions (Fig. 7A). Placement in darkness induced inflation in -40% of the larvae during the normal light-phase. This percentage increased dramatically to 70^ at the normal time for the beginning of the dark-phase CO O 80- E E "■ 60- Ol E ho- u "O O n E 2.0- S to 1 ." (22) I (30) f (52) i (113) (139) 5) 11 12 13 14 15 1 6 Total Length (mm) Figure 5 Swimbladder volume vs. total length of larvae fron i all lighting conditions. Means and standard errors are Dlot- ted. Number under each plot is the sample size. Atlantic menhaden inflate their swim- bladders in response to a decrease in light intensity. The smallest size observed with an inflated swimbladder was lOmmTL, which is smaller than the mini- mum size of 13mmTL found by Hoss & Blaxter (1982). This difference may re- sult from the large sample size used in the present experiment, since the per- centage of lOmmTL larvae with an inflated swimbladder was -10%. The per- centage of larvae inflating their swim- bladder in response to a decrease in light intensity increased with size and reached 100% at 17mmTL and greater. Swimbladder volume increased with size, which is not surprising, since larger larvae have larger swimbladders. However, within any fish size, the mean volume did not vary significantly with light- ing condition. This result disagrees with the qualitative con- clusion of Hoss et al. (1989) and is likely due to the wide variation in volume. Hoss et al. (1989) also found high vari- ances, but failed to compare mean values statistically. The percentage of larvae inflating their swimbladders in- creased as the light intensity decreased, which clearly indi- cates that the decrease in light intensity cued the response. However, a step function was observed, in that once light was below a particular absolute level, maximum inflation occurred. Future experiments are needed to determine whether inflation is cued by exposure to light intensity be- low an absolute level or to the rate of change in intensity. This information will allow predictions of the time of infla- tion in the field. Hoss et al. (1989) failed to find an endogenous rhythm in swimbladder inflation for larvae held under conditions simi- lar to the present experiment. In contrast, our study showed a clear rhythm during the first day for larvae held under constant light. The percent inflation was low in fish intro- duced to the dark during the time of the light-phase, and nearly doubled at the time the dark-phase began. This high percent response did not return to a low level at the time of the next light-phase. Hoss et al. (1989) did not begin mea- suring swimbladder inflation until after -24 h in constant light, which may be why they failed to detect an endog- 258 Fishery Bulletin 91(2). 1993 _ 80- a> /'ffiSTX. "O 3 60- /an\/ ^^"(66) £ (24) (26) CO g~ 40- /(22) c „ 20- s - (54) 30 50 70 90 Time (min)in Darkness 110 Figure 6 Percentage of larvae 11-16 mmTL filling their swim- bladders after different times in darkness. Number under each point is the sample size, and asterisk is the first time that a proportion was significantly greater (p<0.05l than the proportion of larvae with inflated swimbladders initially in light, which is plotted at time zero. enous rhythm. After this time in constant light, variation in inflation after introduction to dark was not evident in the present study. There are two possible explanations for this result. First, the rhythm could fail to continue be- cause larvae were kept in constant light, a condition that frequently suppresses an endogenous rhythm (Hastings et al. 1991). Second, the rhythm could consist of one cycle in which inflation is suppressed during the light-phase and larvae become "ready" to inflate at the be- ginning of the dark-phase. Readi- ness then continues until a dark cue is received, which resets the endog- enous clock. Clearly, Atlantic menhaden lar- vae are adapted for swimbladder inflation at sunset. Their rhythm in- dicates they are most responsive to a light-intensity decrease at this time and most inflation occurs within 20 min. Such a dramatic re- sponse suggests swimbladder infla- tion has an important functional advantage. Menhaden larvae are negatively buoyant even with a fully inflated swimbladder. Nevertheless, infla- tion reduces their sinking rate (Hoss et al. 1989). Past investigators have suggested that swimbladder inflation acts as an energy- saving mechanism, allowing larvae to expend less energy for maintaining their position in the water column at night when they are not feeding ( Hunter & Sanchez 1976). During the day, a fully inflated swimbladder may reduce the speed of movement and, thereby, the effectiveness of prey capture and predator avoidance. In addition, Uotani ( 1973) proposed that inflation allows larvae to decrease their movement at night, which serves to reduce detection by predators that hunt by vibrations, such as chaetognaths. Field studies show some indication that menhaden larvae undergo reverse diel vertical migra- tion (DVM) in which they descend in the water column near sunset and ascend near sunrise (Hoss et al. 1989). Chaetognaths exhibit the opposite pattern of nocturnal DVM (Pearre 1973, Sweatt & Forward 1985). Reverse DVM is proposed as a mechanism for avoiding zooplankton preda- tors that undergo nocturnal DVM (Ohman et al. 1981, Neill 1990). A slower descent rate at sunset by menhaden larvae due to inflated swimbladders may reduce detection by chaetognaths that are ascending toward the surface. Since the percentage of menhaden larvae with inflated swimbladders increases with size, the importance of re- duced sinking rate for predator avoidance may increase with size. The threat of predation to menhaden larvae is probably reduced during their ascent at sunrise because the descending chaetognaths have been feeding all night. 0800 1000 1200 1400 1600 1800 2000 2200 1000 Time (hrs) Figure 7 Percentage (A) of larvae ll-16mmTL that filled their swimbladders before ( ) and after ( ) exposure to darkness for 2h over the solar day. The swimbladder volume of 11 mmTL larvae (B) after exposure to darkness is also plotted against time in the solar day. Means and standard errors are plotted. Number near each plot is the sample size. Arrow indicates the times of the beginning of the dark phase of the rearing LD cycle. Forward et al.: Swimbladder inflation of Brevoortia tyrannus 259 Acknowledgments This research was part of the South Atlantic Bight Recruitment Experiment (SABRE) sponsored by the NOAA Coastal Ocean Program. We thank Dr. R. Tankersley for his technical assistance. Citations Blaxter, J. H. S., & J. R. Hunter 1982 The biology of clupeoid fishes. Adv. Mar. Biol. 20:1-223. Hastings, J. W., B. Rusak, & Z. Boulos 1991 Circadian rhythms: The physiology of biological timing. //; Prosser, C.L. (ed. ), Comparative animal physiology, 4th ed.. p. 435-546. Wiley-Liss, NY. Hettler, W.F. 1983 Transporting adult and larval gulf menhaden and techniques for spawning in the laboratory. Prog. Fish.-Cult. 45:45-48. Hoss, D. E., & J. H. S. Blaxter 1982 Development and function of the swim bladder- inner-lateral line system in the Atlantic menhaden, Brevoortia tyrannus (Latrobe). J. Fish. Biol. 20:131- 142. Hoss, D. E., & G. Phonlor 1984 Field and laboratory observations on dirunal swim bladder inflation-deflation in larvae of gulf menhaden, (Brevoortia patronus). Fish. Bull., U.S. 82:513-517. Hoss, D. E., D. M. Checkley Jr., & L. R. Settle 1989 Diurnal buoyancy changes in larval Atlantic menhaden (Brevoortia tyrannus). Rapp. P-V. Reun. Cons. Int. Explor. Mer 191:105-111. Hunter, J. R., & C. Sanchez 1976 Diel changes in swim bladder inflation of the lar- vae of the northern anchovy, Engraulis mordax. Fish. Bull., U.S. 74:847-855. McFarland, W. N., & F. W. Munz 1975 The evolution of photopic visual pigments in fish. Vision Res. 15:1071-1080. Munz, F. W. 1958 The photosensitive retinal pigments of fish from relatively turbid coastal waters. J. Gen. Physiol. 42:445-459. Neill, W. E. 1990 Induced vertical migration in copepods as a de- fense against invertebrate predation. Nature (Lond. ) 345:524-526. Ohman, M. D., B. W. Frost, & E. B. Cohen 1981 Reverse diel vertical migration: an escape from invertebrate predators. Science (Wash. DC) 220:1404- 1406. Pearre, S. Jr. 1973 Vertical migration and feeding in Sagitta elegans Verrill. Ecology 54:300-314. Sweatt, A. J., & R. B. Forward Jr. 1985 Diel vertical migration and photoresponses of the chaetognath Sagitta hispida Conant. Biol. Bull. (Woods Hole) 168:18-31. Tracy, H. C. 1920 The membranous labyrinth and its relation to the precoelomic diverticulum of the swimbladder in clupeoids. J. Comp. Neurol. 31:219-257. Uotani, L. 1973 Diurnal changes of gas bladder and behavior of postlarval anchovy and other related species. Bull. Jpn. Soc. Sci. Fish. 39:867-876. Walpole, R. E. 1974 Introduction to statistics. Macmillan, NY. Abstract. —Seasonal variability in gonad growth was investigated for the tropical cephalopods Loligo chinensis and Idiosepius pygmaeus. Statolith ageing techniques enabled a comparison of gonad growth in re- lation to both individual size and age. Age analyses revealed that male and female individuals of L. chinensis matured earlier during the warmer summer period than in the cooler winter period. These prelimi- nary results suggested that matu- rity was governed more by individual size rather than age. Analysis of sea- sonal change in the gonadosomatic index (GSI) revealed that L. chinensis gonad tissue accounted for the greatest percentage of body weight in the month of October. The trend in the nidamental gland/ mantle length index closely paral- leled the trend in GSI values for fe- male individuals of L. chinensis, while growth of the nidamental gland and hectocotylus closely par- alleled growth of the gonad. The sea- sonal variation in reproductive in- vestment for Idiosepius pygmaeus followed a different pattern com- pared with L. chinensis. Slower growing cool-season (spring) indi- viduals lived longer and had com- paratively larger gonads than their warm-season (autumn) counterparts, despite no difference in body size be- tween the two seasons. Idiosepius pygmaeus thus appeared to be em- ploying a 'trade-off in its reproduc- tive strategy by partitioning a greater amount of energy into go- nad tissue over a longer lifespan dur- ing the cooler period of the year. Seasonal variation in reproductive investment in the tropical loliginid squid Loligo chinensis and the small tropical sepioid Idiosepius pygmaeus George D. Jackson Department of Marine Biology, James Cook University of North Queensland Townsville 48 11 . Queensland, Australia Present address: Department of Zoology, University of Western Australia Nedlands, Perth, Western Australia 6009 Manuscript accepted 10 December 1992. Fishery Bulletin, U.S. 91:260-270 1 1993 1. In reviewing cephalopod reproduc- tion, Mangold ( 1987) has emphasised that there are at least as many open questions as there are established facts, and that there are thus many gaps in our knowledge as well as con- tradictory statements. Recent re- search (e.g., Hanlon et al. 1989, Rodhouse & Hatfield 1990, Jackson & Choat 1992) is revealing that cephalopod lifespans are considerably shorter than many estimates made over the last several decades. This past confusion has apparently led to a poor understanding of the re- productive tactics of cephalopods. Clearly, any ideas regarding the lifespan of an organism will influence ideas regarding reproductive events in the individual. Statolith ageing techniques have the potential for resolving some of the discrepancies in our understand- ing of the reproductive tactics of squids and sepioids. By analyzing in- dividual age and maturity status, age-at-maturity and time-specific schedules of gonad growth can be con- structed. The focus of this study was to consider seasonal variation in age- at-maturity and reproductive invest- ment in the tropical loliginid squid Loligo chinensis and the small sepioid Idiosepius pygmaeus. Statolith age- ing techniques have been applied to both /. pygmaeus (Jackson 1989) and L. chinensis (Jackson 1990a, Jackson & Choat 1992). Analysis of seasonal samples for both of these species has revealed that there was considerable seasonal variation in growth (Jack- son & Choat 1992). This study was therefore undertaken to see if there was any seasonal variation in gonad growth or seasonally-induced varia- tion in age-at-maturity in these two species. Loligo chinensis is common in North Queensland waters and can be captured by bottom trawls through- out all months of the year. The species is sexually dimorphic, with males growing longer than females. In North Queensland waters, males are commonly encountered up to -180 mm dorsal mantle length (DML), while females are common up to ~120mmDML. Hatchling size for L. chinensis is -1.4 mm (unpubl. data). In the summer, L. chinensis reaches adult size in ~100d, while winter growth is slower with adult size reached by 140-170 d (Jackson & Choat 1992). Idiosepius pygmaeus is a common neritic cephalopod which is found in surface waters in mangrove, estua- rine, and breakwater habitats (Jack- son 1989). This species, however, is only common in nearshore surface waters between March and Novem- ber, with few specimens observed over the summer months (December- February) (Jackson 1992). Idiosepius 260 Jackson Cephalopod reproductive investment 261 pygmaeus is sexually dimorphic, with females reach- ing much larger sizes than males. In North Queensland waters, males are commonly encountered up to -10 mmDML, while females are commonly encountered between 13 and 18 mmDML. Planktonic hatchlings are ~1 mmDML (pers. observ.). This species has a short lifespan, with maturity reached in <80d (Jackson 1989), and exhibits slower growth during the cooler seasons of the year (Jackson & Choat 1992). Materials and methods Loligo chinensis and /. pygmaeus were captured from tropical waters off Townsville, North Queensland. Preparation and enumeration of statolith growth in- crements for both species were similar to techniques used for Sepioteuthis lessoniana (Jackson 1990b), al- though statoliths of /. pygmaeus were not ground or polished. Individuals of L. chinensis were captured in paired trawl nets (each net had an 11m gape and 3.8cm mesh) which were towed for ~20min. The ma- jority of individuals of L. chinensis were captured in Cleveland Bay (19°11'S,146 56'E) in water depth <20 m. Trawling was undertaken generally for one day each month from February 1988 to November 1989, and up to 15 trawls were taken on each sampling date. Indi- viduals of /. pygmaeus were captured by dip-netting along a breakwater east of the Townsville harbor (19 o 15'S,146 o 50'E; see Jackson 1992). Individuals of L. chinensis used in the age analysis were captured on 12 January 1989 (summer, n=37) and 13 July 1989 (winter, n=21). Individuals of/, pygmaeus were ana- lyzed from autumn and spring. Individuals (n=41) for the autumn sample were captured during four sampling periods over two years: 22 and 23 March 1988; 21 and 22 March 1989. Individuals (n=38) for the spring sample were from six sampling trips over two months: 10, 23, 24 August 1988; 7, 20, 21 Sep- tember 1988. The greatest differences in growth rates and population age structure were observed between these two seasonal periods (Jackson & Choat 1992). Analysis of reproductive structures Specimens of L. chinensis were initially fixed in buff- ered 10% seawater-formalin to preserve the large tis- sue mass and later transferred to 70% alcohol to pre- vent damage to the statoliths. Specimens of/, pygmaeus were preserved immediately in 70% alcohol due to their small body size. Gonads were removed, blotted with paper toweling (L. chinensis) or filter paper (/. pyg- maeus), and weighed. Dorsal mantle length (DML) was measured on both species, and nidamental gland length (NGL) and hectocotylus length was measured on female and male individuals of L. chinensis, respec- tively. Measurements were taken with an eyepiece mi- crometer (/. pygmaeus) or with either callipers or a graduated ruler (L. chinensis). Maturity was deter- mined by the presence of mature oocytes in the ovary along with large nidamental glands in females, and the presence of spermatophores in males. To discern the pattern of growth for gonads, the nidamental gland, and the hectocotylus, measurements of gonad weight, nidamental gland length, and hectocotylus length were plotted against both mantle length and age. Due to the large amount of scatter in many of the plots (especially with /. pygmaeus) and the complex curvilinear relationship between many of the relationships, regression analyses were not carried out. Gonadosomatic and NDL/DML indices The seasonal trend in gonad growth was also exam- ined for L. chinensis. Maturity status was determined for 231 individuals from trawl samples between Feb- ruary 1988 and November 1989 (this analysis included data from individuals which were aged from the Janu- ary 1989 and July 1989 samples). For most samples, all individuals within the adult size-range (>100mm) were used in the gonad analysis, except for several summer samples in which a very large number of indi- viduals >300 mm were captured. Parameters measured for each squid were dorsal mantle length, body weight, gonad weight, and nida- mental gland length for females. The gonadosomatic index (GSI) was calculated for each specimen as gonad weight (g) total body weight (g) x 100. For females, the nidamental gland length/dorsal mantle length (NGL/DML) index was also calculated as nidamental gland length (mm) mantle length (mm) x 100. Analyses of nidamental gland length and hectocotylus length were not carried out for /. pygnmaeus. Results Loligo chinensis Gonad growth with age There were differences in the relationship between gonad weight and age for the two samples of L. chinensis taken within different sea- sonal periods (Fig. 1A,B). The relationship between testis weight and age was similar during both sea- 262 Fishery Bulletin 9 1(2), 1993 08 MALES A s FEMALES B + + SUMMER + SUMMER O WINTER D 5 □ WINTER WEIGHT (g) o o A. a * + „ + 3 I o B £ D S 4 3 + + +■ + 1- U) UJ 1- 02 D ° + □ □ + □ 2 1 + , , ,,+ + B D + ) 20 40 60 80 100 120 140 160 180 c 20 40 60 80 100 120 140 160 180 AGE (days) AGE (days) 0.8 MALES C 6 FEMALES D + H SUMMER 1 SUMMER □ WINTER S 4 3 2 1 □ WINTER TESTIS WEIGHT (g) o o o ++ + 3 t- i a UJ I 1 § a a + D + + 4+ + + + , , % D ,+, + + ) 20 40 60 80 100 120 140 160 180 ) 20 40 60 80 100 120 140 160 180 MANTLE LENGTH (mm) MANTLE LENGTH (mm) Figure Gonad weight/age relationships for (A) males and 1 B) females, and gonad weight/mantle length relationships for (C) males and (D) females of Loligo chinensis collected in summer (January) and winter (July). sons, with the main difference being the shift in the winter curve to the right (Fig. 1A). Because of season- ally-induced differences in somatic growth (Jackson & Choat 1992), males matured later in winter. The rela- tionship between age and testis weight had greater variability in winter, which suggests that age-at- maturity was less well-defined in the winter, with a slower rate of maturity in some individuals. In both winter and summer, males with a testis weight >0.1g had spermatophores present. However, there was variability in age-at-maturity in both sea- sons, with testis weight in apparent immature males of 0.307-0.335 g. The youngest mature male during summer and winter was 83d and Hid, respectively. The oldest immature males for summer and winter were 90 d and 130 d, respectively. There were also seasonal differences in maturity pat- terns of females based on ovary weight (Fig. IB). Fe- males matured in summer at a young age, with the ovary reaching a large size in 65-85 d, although two older specimens were immature. This was especially apparent in the oldest individual (lOOd) which also had very undeveloped nidamental glands. All females in the winter sample had small gonads and nidamental glands, and there were no mature or soon-to-be-mature females. This differed from the pat- Jackson Cephalopod reproductive investment 263 tern observed in males, which had reached maturity in winter, and suggested that female maturity was out of phase with males during this period of the year. All females aged in summer with an ovary weight >1.161g were mature, and the ovary filled much of the mantle cavity. The youngest mature female was 83 d, while the oldest immature female was the oldest fe- male aged, 100 d. Gonad-soma relationships Comparing gonad weight to individual age (Fig. 1A,B) revealed a different pat- tern than in the gonad weight/mantle length analysis (Fig. 1C,D). Despite the fact that there was a notice- able difference in the testis weight/age scatter plot (Fig. 1A), due to the older winter males, the testis weight/mantle length relationship was similar for both seasons (Fig. 1C). Thus gonad increase was propor- tional to the squid length rather than its age. A different pattern emerged when comparing ovary weight/mantle length relationships for female L. chinensis (Fig. ID). Although the lack of maturity was still obvious in winter females, the gonad weight/mantle length relationship for both seasons resulted in a single curvilinear relationship, suggesting that maturity oc- curred at 100-120 mmDML, regardless of age. Fur- thermore, although this ageing study indicated that a large proportion of the winter females were older than their summer counterparts, many of the winter imma- ture females were smaller than squids captured in the summer. As with the males, squid size rather than age may be a better indicator of maturity. CO A MALES N-119 1.2 - 1 - CO 08 O oa + f + T t T + + + i it 4 2 - H — i — I — . — I — i — I — i — I — i --I r— I - 1 1 r- H — i — 1 — i — 1 — r- Feb Apr Jun Aug Oct Dec Feb Apr Jun Aug Oct I ee I 89 MONTH B N-107 i -+- > -+- -i ■+- i ■+- Feb Apr Jun Aug Oct Dec Feb Apr Jun Aug Oct 68 69 MONTH Figure 2 Mean monthly gonadosomatic index for Loligo chinensis over the study period for (A) males and (B) females. Bars = SE. Reproductive indices Although it was possible to de- termine the ages of only a small number of individuals in two seasons (rc=64), changes in the gonad weight/ soma weight relationship throughout the year were examined. For both males and females, a seasonal trend could be detected in the GSI (Fig. 2). In males, GSI values were low, with <1.2% of the total body weight consisting of gonad. In contrast, female GSI values were more variable and generally higher than in males, with the gonad comprising as much as 8% of total body weight. Mature males were found in all months sampled. However, a regular seasonal oscillation in relative go- nad weight was apparent (Fig. 2A). There was an in- crease in relative gonad weight from April to October in both years. Over the two years, the testis accounted for its greatest percentage (>1%) of body weight in October, while its lowest values were recorded in April. A similar pattern of fluctuation also existed with the female GSI values (Fig. 2B). In October, females consistently had GSI values that were considerably higher than for other months. Mature females were present in all months except July 1989. As discussed previously, females aged from the 1989 summer sample showed a considerable range in gonad size and level of maturity, despite similarities in both size and age. This range in gonad size was also reflected in the female GSI values, in that the mean values had large stan- dard errors (Fig. 2B) in nearly every month. This was due to the fact that for many months, a proportion of the individuals was immature. The seasonal trend in the NGL/ML index (Fig. 3) was similar to the trend in the female GSI (Fig. 2B). This index also indicated a greater investment in re- production during October, with lowest values in July. Furthermore, standard errors were generally less for this index than for the GSI. The monthly mean mantle lengths for males and females were plotted for the 2yr period (Fig. 4A,B). Although there was some variation in mantle lengths for the different samples, these could not be related to the seasonal peaks or troughs in the GSI values. For example, the largest females were captured from Feb- ruary to July 1988 (Fig. 4B). However, GSI values dropped considerably over this period. Furthermore, mean mantle length was not highest in October for 264 Fishery Bulletin 9 1 [2). 1993 FEMALES 30- N-107 I '.HI ' 1 h + X 26 UJ a Z 20- _l 2 ' 6 ' *** o z 10- \ 5- Feb Apr i Jun Aug Oct Dec Feb Apr Jun Aug Oct 86 1 89 MONTH Figure 3 Mean monthly index of nidamental gland/mantle length for female Loligo c) linensis collected over the study period. Bars = SE. each year, and although 1989 July values were slightly lower than the other months, the July 1988 values were not. The observed changes in relative gonad weight can thus be considered unbiased by individual size. Nidamental gland and hectocotylus lengths Nida- mental gland length was perhaps the most useful mea- surement in female squids for obtaining an index of maturity. The relationship of nidamental gland length to mantle length and age of female specimens of L. chinensis ( Fig. 5 ) closely resembled the ovary weight/ age and mantle length relationships for this species (Fig. 1B,D). For example, two separate relationships were apparent when nidamental gland length was plotted against age, whereas both seasons' data points produced one curvilinear relationship for nidamental gland length vs. mantle length. As with the ovary data, these data suggest that nidamental gland length was more closely related to mantle length than to age. Similarly for males, hectocotylus length vs. mantle length and age (Fig. 6) exhibited the same pattern ob- served in the testis weight/age and mantle length rela- tionships (Fig. 1A,C). For example, in winter there was a shift along the age axis, producing a separate correla- tion for the summer hectocotylus length/age data. How- ever, there was some indication that at large sizes, faster- growing (summer) males had a shorter hectocotylus than slower-growing (winter) males. This relationship was similar to the testis weight/mantle length relationships for /. pygmaeus ( Fig. 70, D) seen below. Idiosepius pygmaeus Gonad growth with age In contrast to L. chinensis, within-season variability in maturity and age-specific A MALES I |_ 140 " o Z 120 Ul u 1 oo- _l I- 80 z 2 60 - N-119 + 1 + + + t +T - | 40 UJ 2 20 J Feb Apr Jun Aug Oct Dec Feb Apr 1 88 1 MONTH Jun Aug Oct 89 B FEMALES I 140 o Z 120 UJ - 1 100 UJ fd 80 z < 60 2 N-107 + + + + Z 40 < UJ 2 20 Feb Apr Jun Aug Oct Dec Feb Apr I 88 I Jun Aug Oct 89 MONTH Figure 4 Mean monthly mantle length for (A) males and (B) females of Loligo chinensis used in gonad weight analysis and nidamental gland length analysis (Figs. 2,3). trends for /. pygmaeus could be determined with a greater degree of accuracy. This was possible because of the greater number of replicate sub-amples taken during each seasonal period (see Methods). The seasonal pattern of gonad growth was different for /. pygmaeus than for L. chinensis. The seasonal influence on gonad growth and maturation may have been somewhat less for individuals of/, pygmaeus (ana- lyzed for spring and autumn) compared with individu- als of L. chinensis (analyzed for summer and winter). However, individuals captured in autumn would have grown over the warmer period at the end of summer, while individuals from the spring sample would have grown and matured through the colder period, at the end of winter. Due to the considerable scatter in age/length rela- tionships for this species (Jackson & Choat 1992), there was also considerable scatter in the gonad weight/age relationship (Fig.7A,B). While gonad weight/age rela- tionships for L. chinensis resulted in different scatter plots separated on the age continuum (x axis), the pattern was modified differently for /. pygmaeus for Jackson Cephalopod reproductive investment 265 o z UJ _J Q z < _l 2 ° z UJ < a " + SUMMER □ WINTER + + + + + - + + 4- a + + + g a E E Z I- O z UJ _l D Z < Z UJ Q + SUMMER □ WINTER 40 60 80 100 120 140 160 180 AGE (days) + + ao >f° + + a am 20 40 60 80 100 120 140 160 180 MANTLE LENGTH (mm) Figure 5 Nidamental gland length/age relationships and nidamental gland length/mantle length relationships for female individuals ofLoligo chmensis collected during summer (January) and winter (July). the two seasons (i.e., both seasons' data points fell within the same scatter plot). While individuals of I. pygmaeus were older in the spring samples, their gonads reached a proportionally greater weight than did the autumn individuals. Males The relationship for testis weight vs. age (Fig. 7A) was the same for both seasons, with data points clustering on a single testis weight/age con- tinuum, with the exception of one 41 d individual which fell considerably outside the cluster of data points. The major difference in the seasonal component of the data was a clustering of data points for each season at op- posite ends of the testis weight/age continuum, with spring individuals reaching a greater age and possess- ing proportionally heavier testes than their autumn counterparts. The youngest mature males in spring and autumn were 22 d and 37 d, respectively, while the oldest im- 26 25 + SUMMER + SUMMER i E 20 D WINTER ^ + ° E E 20 D WINTER I I 1- -t- I- + a + O + z + Z + UJ 16 UJ 16 •J -1 o . 9 (0 3 & + * o n § daV + + + -1 > + & °° > 10 1- 1° □ a i-io D ° o + n ° n + o + D o D + o D O a i- + 1- + o o UJ 6 UJ 6 I I v 20 40 60 80 100 120 140 180 180 20 40 60 80 100 120 140 160 180 AGE (days) MANTLE LENGTH (mm) Figure 6 Hectocotylus length/age relationships and hectocotylus length/mantle length relationships for male individuals oiLoligo ehinensis collected during summer (January) and winter (July). 266 Fishery Bulletin 91(2). 1993 MALES A FEMALES B 7 60 + AUTUMN D -1- AUTUMN D SPRING D SPRING 6 D 60 a "5 6 o> E E 40 ~-~ 1- G D K X 4 I o + "b+ D ^ 30 HI 5 3 a a <0 K- W , UJ 2 o an £ + O + a t- + + B + + 10 + 1 + ++ + D , +j±i-u., rfff rJQ , o' — 10 20 30 40 60 60 70 20 40 60 80 AGE (days) AGE (days) MALES C 7 f 60 FEMALES D 4- AUTUMN Q +- AUTUMN 6 O SPRING D 50 □ SPRING a O) O) E 6 E ~~ □ a """ 40 - 1- H I Rd X O 4 O 111 s (0 3 1- (0 UJ 1- 2 D D + ^30 ° + + + £ a# < □ + + ip ++ o + * + + 10 + 1 + * + a ruift •© . 0' — 2 4 6 8 10 2 4 6 8 10 12 14 16 MANTLE LENGTH (mm) MANTLE LENGTH (mm) Figure 7 Gonad weight/age relationships for (A) males and (B) females, and gonad weight/mantle length relationships for (C) males and (D) females of Idiosepius pygmaeus collected in autumn (March) and spring (August/September). mature males for both seasons were 32 d and 38 d, respectively. Females A similar relationship to the males also existed in the ovary weight/age relationship, i.e., both seasons' data points tended to cluster along one ovary weight/age continuum (Fig. 7B). However, the scatter was greater than for males, due to the fact that in both seasons there were individuals with very small ovaries. For example, in both seasons, individuals of 40-60 d showed a considerable range in ovary weight. However, as with male testis weight, the spring fe- males also had the heaviest ovaries. Although a num- ber of females had large well-developed ovaries, none of the specimens examined had any mature ova present. Gonad-soma relationships One possibility for the greater gonad weight in spring vs. autumn could have been due to larger individuals being captured in the spring. However, the plot of gonad weight against mantle length for both sexes (Fig. 7C,D) revealed no size difference in males between seasons and only one spring female slightly larger than the other females. Although there is some overlap in the data for the smaller individuals for both seasons, in the larger sizes, Jackson Cephalopod reproductive investment 267 slower-growing spring individuals did, in fact, eventu- ally produce larger gonads than their autumn counter- parts. Discussion The biotic and abiotic influences on maturity are com- plex. Factors such as light (day length), temperature, and food availability can all affect rate and age-at- maturity. Gonad development is directly under hor- monal control, and appears to be influenced by the optic gland (Mangold 1987, Boyle 1990). However, the process of maturity is not completely understood and may also be controlled to a certain degree by indi- vidual genetic factors, apart from outside influences (Mangold 1983). This study gives preliminary insight into the maturation process in tropical squids and sepioids based on age information. Statolith ageing techniques should prove useful in providing a time- scale on the squid maturation process and problems encountered with variation in size-at-maturity. Loligo chinensis Variability in size-at-maturity appears to be a com- mon phenomenon with cephalopods. Hixon (1980) found that some immature Lolligunciila brevis females were as large as other fully-mature individuals. Con- siderable variation in maturity has also been shown to exist in L. opalescens, with females maturing by 81 mmDML while other females remain immature un- til 140mmDML (Hixon 1983). Similar discrepancies in size-at-maturity have also been documented for Se- pia officinalis (Boletzky 1983) and Dosidicus gigas (Nesis 1983). Mangold (1983) has also demonstrated that cultured octopuses reared from the same egg mass reached maturity independent of sibling size or age. The delay in maturity for some L. chinensis females for a given length (e.g., Fig. IB, summer) is similar to observations made for Lolligunciila brevis by Hixon (1980). However, in contrast to the other studies cited above, L. chinensis females do not show a wide differ- ence in size-at-maturity. The fact that maturity was more closely related to length than to age suggests that there might be some physical or physiological mechanisms controlling maturation apart from age. Some female cephalopods have been shown to develop eggs only when a minimum threshold in body size is achieved (Mangold 1987). The maximum size for indi- viduals of L. chinensis obtained in this study is smaller than the maximum size recorded for the species, which is up to 300 mmDML (Roper et al. 1984). This sug- gests that the squids in Cleveland Bay may be a pre- cocious population which is maturing near the mini- mum threshold in body size for the species. This could therefore account for the restricted size-range at maturity. Although male and female L. chinensis were mature in the same age-range in summer, a different situation existed in winter, suggesting that mat