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 ..-•■'" RZ = 0 60 d N = 53 P < 0.001 A 1 0 .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 R2 = 0 71 N = 35 0 P < 0001 1 0 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 (Ft=total number of eggs) by proportion: DeMartim et al Fecundity comparisons of Panuhrus margmatus F. = F. ■ ( w. where Fs = number of eggs in subsample, Ws = 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: F150=0.06, P=0.81) or was deleted from the analysis (F149<0.01, P>0.99). Slopes also were indistinguishable between the two periods at Necker Island (F163=1.17, P=0.28). Intercepts did not differ between periods at Maro Reef (F150=0.22, P=0.64), but the period (intercept) ef- fect at Necker Island was significant (F164= 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 (F159=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% (a2=0.05). CL significantly influenced fecundity at both Maro Reef (Fj 50=70.6, P<0.001; Fig. 1) and at Necker Island (F164=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: F151=0.55, P=0.46). Intercepts also were indistinguishable: egg size was uninfluenced by site (F152=2.75, P=0.10). Carapace length had no Fishery Bulletin 91 1 1993 effect when sites were evaluated separately (F152=0.11, P=0.75). However, female size significantly but weakly (P2=0.08) affected egg size when data for the two sites were pooled (CL effect: FI53=2.104, P=0.04; median egg diameter = 30.2 EPU+0.039 CLmm; 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=aCLb, 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 m3) 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 m3 over the last 2 weeks of sampling. The sum of the weekly mean larval densities over all collections ( 1540 larvae/ 100 m3) 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 0 to 40 lar- vae/100m3 (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 m3) 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=3.7/100m3) 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 m3) 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,0,=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 0 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 0 > • »— 1 1 > O -100 <3 *T3 -200 -»-^ 200 O CD i-H l—i 100 o o -X— — a— o. o o fe°_oooo 0 0:~ — -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~.~°- -O0_©©- -■-?-- &--&- fo0,^'5°0o"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 0 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^-^--—---°"' 0 1 2 3 4 S 6 7 8 25 IS 26"C -> 21"C s • *W> °o >°-£>-£L0 o ° o IS t°o ° Ooo --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 (Q10) 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 Qlft 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 0 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 - 0 On ,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__ 0 c o 1 1 1 1 1 o 1 1 - 0 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 0 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- -•-e3*^ ^^ r* 1987 10 - J knv 8- / "^"•"-rj 6- 1 1 -§■> _ ■ji' v- 4 - »=*= ^r^^^* 1988 o ,0" o """ 8- 0. E 6- Ui 1- 4- ■ — B-* 1989 10 - A. J\ k/ 8- 1 JSsf V 6- J M^^~ I-"*-* 4- •-*— &&Z-+ v^ 60 9° 120 150 18 MARCH APRIL MAY JUNE 0 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./m3 ) 1988 .o\/« "■o---o* o- A -o 1989 600 - r\ •4000 400 - 200- , o-o~.« X? r*-* -3000 -2000 -1000 -0 0 6 ' 90 120 150 It MARCH APHIL MAY jUNE 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 m3. 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/m2 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 0 4 2 0 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 ApR|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 (r2=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; Rn=opercle ra- dius at year n; R,=opercle radius at capture; b=slope of body-bone regression; where Ln = 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; r2=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 (r2=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 0 100 200 300 400 500 600 700 800 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 -■ x40- *30- I ~| -1 20 - T-| 10- JrhJi ln__ 100 200 300 40 0 500 600 700 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 r2 = 0 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 0 1 491 0 1 4340 0 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 0 1 460 0 1 3090 0 21 1 613 0 1 508 0 1 4750 0 22 2 655 25 2 538 22 2 4995 725 23 0 — — 0 — — 0 — — 24 0 — — 0 — — 0 — — 25 1 672 0 1 577 0 1 6568 0 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 0 1 395 0 1 2530 0 17 1 575 0 1 479 0 1 4220 0 18 5 628 80 5 517 47 5 4961 1700 19 0 — — 0 — — 0 — — 20 1 557 0 1 485 0 1 3385 0 21 1 660 0 1 570 0 1 5100 0 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>Ft 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^co^^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; t0=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 9f 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 - 0 - -5 Virginia Rhode Isl X 6 7 Month 10 11 12 Figure 1 0 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-eJ"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 0 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 r2 ■ 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 = 7.8e03* (r2=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 (r2=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=7.8e03x, 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, r2=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 AA \ 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 80ci (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 0 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 0 199 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 0 5 137 9 T 220-227 13.2 47 89 12 0 4 152 0.7 10 T 3.2 210-213 13.6 83 121 40 0 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 0 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 0 10 506 19 U 9.1 201-259 12.0 248 260 12 0 9 529 20 U v-trawlable 195-259 13.2 Totals site; T = trawlable site; U 112 250 13 0 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 0 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 0 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./lOOOm2) 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/lOOOm2 of seafloor area at the 13 trawlable sites compared with 6.9 rockfish/1000 m2 at the 3 marginally-trawlable sites and 3.2 rockfish/ 1000 m2 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 <30m2 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* ( 103m2) Rockfish heads Flatfish Pollock Other Total 1 MT 13.7 75 32 94 4 6 211 3 MT 15.3 138 239 43 0 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 m2 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, ts=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 (r2=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 (rs=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,0 Y=J.5107+(.1902)X r2-.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, God=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 (r2=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 (sY) 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 Raw1 Log2 Y sv 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 o2, the popu- lation mean and variance of the untransformed data. Letting Y = In (X), Yand s2Y are estimators of uLN and o2ln the population mean and variance of the log- transformed data. Note that in Eq. (A3) and (A4) both u and o2 for the lognormal distribution are functions of two parameters, uLN and a2LN, 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 uLN and variation in estimating a2LN. 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 s2Y 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)H2 (n-l)V FM = exp(Y)yn(r) (1) where n is the sample size and \j/n(t) is the infinite series 2\n-(n+l) 3ln3(n+l)(n+3) (n-Vt4 4!n4(n+l)(rc+3)(n+5) (2) + .. with t = —^— . The function \|/n(t) is defined such that E[\|/n(s2)] = exp f-^p o2 ) and ^^ [\j/n(s2>] = exp(o2); 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 \\in 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 + s2Y sY Hh LLa = exp Y + — + Sv Vn-1 (3) (4) The quantities H^, and H„ [functions of a, (n-1) and sY] 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(uLN + - — - 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, uLX. is the geometric mean catch and, equivalently, the median catch for lognor- mal data. It remains unaffected by skewness, a func- tion of [exp(o2LN-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 0 0 300 600 900 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 e1 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, uLN, 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' Logb 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 eMLN, 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 eMLN 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 (eMLN). 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 s2Y (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, ulN (and u) may have increased slightly, but cr2LN (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 s2Y 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 s2Y 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 ctln, then X=eY is lognormally distributed with density function (Aitchison & Brown 1969): 1 The kth moment about zero, E(Xk) is expressed as exp f-(lnZ-uLN)2 2o2, (AD £(Xk) = E (ekY) = 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 (mm3) — 0.1512 ±0.0220 (15) 0.1205 ±0.0259 (11) *Oil volume at hatching (mm1) — 0.0023 ±0.0005 (15) 0.0029 ±0.0004 (11) Yolk volume at first-feeding (mm1) — 0.0002 ±0.0000 (15) 0.0002 ±0.0000(11) Oil volume at first-feeding (mm1) — 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 loge yolk and oil volume in mm3 (Y) and age in days (X,) for gulf and Atlantic menhaden. N = number of replicates. Species °C N Slope (B,) Intercept (B0) r2 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 0 100 80 60 40 20 0 o / ' 1 "\ 16° C 0 1 2 3 4 5 6 7 8 20° C x 01 2345678 lOOf--,. 80- 60- 40 20 0 _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 ■ ■ ^ 0 *rrS ■ ■ ' ^t- - -2.0 -1.0 0.0 1.0 2.0 -2 0 -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,0 Dry Wt.= -2.398+ 4.775(log,0 SL) 4.0 r2=0.98 / 3 3.0 j X G 111 5 f I 2.0 f\ o 6" o _j / ! 1.0 Log)0Dry Wt.= -0.714 + 3.085(log,0 SL) 0 r2 = 0.95 0 0.2 0.4 0.6 0.8 1.0 12 1.4 L0G,0 STANDARD LENGTH (mm) Figure 5 Length-weight relationship of larval Brevoortia tyrannus showing a change in body morphology. Breakpoint occurs at lOmmSL. Log,0 Dry Wt.= -2.787 + 5.1 43 M 0 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 (Yt) 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, = Ft Nt 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, Wt = 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 0 245 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 CBW 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 hFL = 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, X2 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, x2 test) more predation on elasmobranchs and goosefish by fe- male sharks than by males. Offshore, juvenile males consumed significantly (P<0.05, x2 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 <1n^ o nm Im f> i ~~1 1 i i i 1 r • i i i — i" i? i qt -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 /? X2 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 sections1 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 - J1'' mm^^^ 0 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 "> 0 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,0 (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 (F4660=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, Fu06= 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 0 10- 5 - 0 - 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 0 r o Period 1 • Period 2 • Period 3 ° Period 4 0% o ariwr^' He|9h' 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 log10 (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 0 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 logln-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 0 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 2C>> 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 m3/s in September to 10,530 m3/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 3l/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 0 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+CV2)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 85rt 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>te ) 186 Fishery Bulletin 91 12), 1993 where z is a random variable distributed as N (0, a2), and (2) precision of the estimate, represented in the distribution of the simulated bootstrap replicas for year t, Ibt, as Ibt = Itev,forb=l, . ..,79, where v is a random variable distributed as N (0, e2). 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, o2 and e2 should be equal. However, estimates of dol- phin abundance may exhibit an additional variability, represented in this setup as o2>e2. 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 byo2. 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 CVa and CVf, 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 irlvtu 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 CVn=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 CVE=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 CV0=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 CVE. 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 CV0=CVE 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 CVa>CVE, 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„>CVE, 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° cvc: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. __-, '' "-. cvc:0.3 o - l I l l 1 J 5 10 15 20 25 Year CVa:0.3 8 - ______^ smoom.oi.st cv:03-^ ,- — -_ ^ _„...„ — __ 10 yr linear test /*V / ""'"'--w-*'"\ cv€-0.4' o _ CO C X" VX/N 'V 0 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 0 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, CVr 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 CVo- = 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-^ cve:0.4 cv6:0.3 ^ ^ ( ' 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) CVo=0.2, and (bottom) CVa=0.3. The two solid lines for each smoothed test represent different levels of interannual variation, CVt. 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 0 0 2 0 4 0 6 0 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! 0 2 0 3 0 4 0 5 Inter-annual vanatton {cv(b)} Steady decline 02 03 04 05 Inter-annual vanatlon (cv(b)) Steady cycle Average absolute error )0 0 2 0 4 0 6 0 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 0 2 0 3 0 4 0 5 Inter-annual variation {cv(b)) Fig bsolute error as a function of different degrees of inter ure 8 annual varia 0 2 0 3 0 4 0 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 0 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_^^ -„.--J5<* Bay \ '"'"' / 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 m3 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 m2. Between 3240 and 4570 m3 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 m3 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 m3 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 ma 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 0 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: •• •/• ; ▼ 0.2Q 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 0 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 0 L 1.29± 0.34 1.38± 0.77 ns Mullidae S 0 0 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 m3/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 m3; Tucker trawl catches averaged 1.49/1000 m3) 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 rj \ 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.4L6' 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 raysb 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 0 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 = 0 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/lOOOm1. 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 = 0 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 = 0 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 = 0 □ 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 = 0 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 0 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,79o05'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,79c'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 32028'N,79042'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,80o38'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 29o09'N,88°33'W 1 281 GCRL 436 29°22'N,87o30'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 29c04'N,88o44'W 1 114 USNM 218169 29°18'N,8805rW 1 108 USNM 218192 28C58,N,84°44'W 1 109 USNM uncat. 24°32'N,83C'36'W 1 197 USNM uncat. 28059'N,88048'W 1 198 USNM uncat. 28:35'N,91012'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 0 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 •a8> 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 el 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 (L050) was calculated to be 154mm. L050 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 x2 = 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 0 30 20 - 10 - 0 -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 0 10 5 - 0 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=ths 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 x2=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 .r2=130, df 7, p<0.001; late-stage likelihood ratio .r2=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 |« 0 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 0 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 0 600- E 3 1 400 " CO 2 o I : : ■ 1 ! 200- 0 4:00 8:00 12:00 16:00 20:00 24:00 Time of day Figure 1 0 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: St = Sjl-beKt) where St= size at time t after birth or settlement S„ = asymptotic size K = growth rate coefficient , S„-SB (1) (2) SR = size at t=0 when organisms begin to grow according to Eq. 1. Sometimes Eq. 1 is written St = Sjl-e-Bt -V) where t0 = time at which size would be 0 b = eKt°. (3) (4) Cohort survival was modeled so that the mortality rate was constant: Nt = N0e"zt (5) where Nt = number remaining in a cohort at time t Ntl = initial number in a cohort Z = mortality rate coefficient. ~ 1.0 CO 2> 0.8- 5 °6 E CD £ 0.4 0 2 o B 0.0 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 - 0 4 8 12 0 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,t2 + a,t3) + e(x) (7) er* 1+Px (8) (9) 110) with al = 0.4361836, a, = -0.1201676, a3 = 0.9372980, p = 0.33267, and efxklO5. 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 - 0 0 12 0 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 0 18.5 17.3 1n. ^3 « ^ ^ 1 M „ ^ ^ 93.3 ?S=9 0 -,¥£ 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 x2 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 (X2 =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 - 0 - Female / \ 23 23 / t 4 5 22 ,r^ /n\ - u ir 68 24 \38 1 V" 27 20 i 7 i2^rtMa,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 0 -12 -10 -8-6-4-2 0 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 7.4xl05 hydrated oocytes for a 4.3kg female, to 3.8X106 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 - 0 °-^" 0 o o o o (J m c 2 1000 - 0 - -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 95r/r confidence i terval for hydrated female black drum Pogonias cromis, sampled February and March 1987. fecundity was 1.6xl06 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- A2681 (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 cnr2 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 ~1013 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=aLb), the slope of the regression (b) equaled 5.31 (r2=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 • I010 10" 10* I013 io14 I015 Light Intensity (photon cm"2s"') 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> | 0 7^ 1 Vo V S 0.5- o E * 0.3- (3) (8) 0.1- (3) io10 io" io12 io'3 io14 io15 Light Intensity (photon cnrf2s~') 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- 0s- (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,146056'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 (19o15'S,146o50'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 + , , ,,+ + BD + 0 ) 20 40 60 80 100 120 140 160 180 0 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+ + + + 0 , , %D,+,+ + ) 20 40 60 80 100 120 140 160 180 0 ) 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 0 4 0 2 - 0 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 1oo- _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 ■ + +0 + 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 0 ■ v 0 20 40 60 80 100 120 140 180 180 0 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' — 0 10 20 30 40 60 60 70 0 20 40 60 80 AGE (days) AGE (days) MALES C 7f 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' — 0 2 4 6 8 10 0 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 maturity in the fe- male winter population was out of phase with the males. This situation has been noted with Todarodes pacificus in which males reach maturity 3-6 mo ear- lier than females (Okutani 1983). This may not be an ecological constraint, considering the reproductive tac- tics of cephalopods. In many species of cephalopods, males mate with immature females (Mangold 1987) with the females retaining the spermatophores until they spawn. An extended interval between mating and spawning has also been noted in octopods, with up to a 114 d interval between spermatophore transfer and spawning (Boyle 1990). Alternatively, the apparent absence of mature fe- males in July may have been due to inadequate sam- pling, since the oldest-aged female captured in winter was 134 d whereas males as old as 173 d were cap- tured. Since maturation was rapid (e.g., females ma- tured in <80d in summer), it is possible that older mature females existed in winter but that none were captured. Based on previous work on the effect of temperature on growth rate (e.g., Forsythe & Hanlon 1988 and 1989), seasonal differences in water temperature may also account for the lack of gonad development in L. chinensis during winter. As age of field specimens in relation to maturity processes has not previously been considered, only aquarium culture experiments are available for comparison. Richard (1966, cited in Mangold 1987) has shown that males and females of the cuttlefish Sepia officinalis from the English Chan- nel, which were raised at 20°C, attained sexual matu- rity at 7 mo and 140 mmDML, while conspecifics raised at 10°C were totally immature at the same age and were only 50 mm in length. Moreover, Richard (1970, cited in Mangold 1987) also found that females of S. officinalis raised at 18°C had a gonad index of 8% at 270 d, while it took 480 d for females raised at 13°C to reach the same value. Other factors, such as seasonal variability in food supply and day length may also influence seasonal differences in maturation rates. The fluctuations of the GSI and NGL/ML index in- dicated that ambient environmental conditions (e.g., temperature and food availability) were influencing both the maturity process and the energy L. chinensis partitioned into reproduction. Although individuals were often mature (e.g., the majority of males ana- lyzed) there were time-periods when the ovary and testes accounted for a greater percentage of total body weight. This was consistently recorded for both sexes 268 Fishery Bulletin 91(2). 1993 in the month of October, and would coincide with the spring warming of water in Cleveland Bay after low winter temperatures which begin to rise in August/ September (Kenny 1974, Walker 1981). Highest GSI values were also reported for male and female L. vulgaris during the spring period in the Mediterra- nean (Worms 1983). The interpretation of fluctuations in the GSI and NGL/DML index is complex due to a variety of fac- tors: (1) L. chinensis exhibits fast growth and has a short lifespan, therefore annual data sets reflect a num- ber of different generations of squids; (2) because of rapid growth and the observed variability in gonad growth (especially in females), greater sample num- bers would be needed to adequately describe gonad growth fluctuations on a smaller scale (e.g., intra- monthly variability); (3) due to the tropical nature of the environment, squids (especially males) are mature throughout most months of the year and, therefore, gonad indices are generally reflecting periods of greater investment in reproductive structures rather than pe- riods of immaturity vs. maturity. The data peaks in October may have been due to the increasing day length in spring, stimulating the optic gland to produce increased hormonal levels which accelerates gonad growth (see Mangold 1987). Since day length would be shortened over the winter period (June- August), the increasing day length during the spring period, along with increasing water tempera- tures, may produce physiological responses leading to maximal gonad growth. More intensive sampling (e.g., fortnightly) over the late-spring and early-summer pe- riod may provide a clearer picture of relative gonad growth over this period. The relationship between mantle length and NGL could constitute a good matu- rity index for females, since the relationship between NGL and DML is closer than the relationship between DML and gonad weight (Worms 1983). Data for fe- males of L. chinensis from this study suggest that the NGL/DML index provides results very similar to GSI values. Maturity parameters based on NGL alone could constitute a more convenient means to determine ma- turity for tropical squids. Nidamental gland length The fact that length of the nidamental gland of L. chinensis bears a very close resemblance to growth of the gonad highlights the close association this organ has with maturation and egg development. The nidamental gland serves the func- tion of producing a gelatinous matrix which encases the cephalopod egg (Roper et al. 1984). Previous stud- ies have used the NGL/DML ratio as a convenient means to assess maturity. Temperate loliginids have been shown to possess mature oocytes when this ratio is >0.2 (Yang et al. 1986, Hanlon et al. 1989). This relationship generally holds well for females of L. chinensis. Out of 112 females (captured throughout the 2yr sampling period) analyzed for this ratio, all the mature females had an NGL/DML ratio >0.2. Of the immature females analyzed, six individuals (5.4%) had a ratio >0.2 (highest value=0.26). This parameter therefore appears to be useful for tropical loliginids as well. The 0.2 NGL/DML ratio does appear to be the minimum parameter for mature females. Nidamental gland measurements are thus useful for providing a rapid and convenient means of assessing the level of maturity in tropical loliginids. The nida- mental gland can also provide useful information about the past history of a female squid. The fact that some of the larger, older, immature females captured during January had small underdeveloped nidamental glands was one means to ascertain that these females had actually not yet matured and had not regressed from a previously mature condition. Hectocotylus length The hectocotylus is an impor- tant reproductive structure employed by the male to pass spermatophores to the female during copulation. It is also extremely useful for quick sexual identifica- tion of preserved specimens. Coelho et al. (1985) car- ried out a detailed study of the growth of the hecto- cotylus of Illex illecebrosus to determine if the degree of hectocotylization of the fourth arm could be related to maturity. However, no close relationship was found between maturity and the degree of hectocotylization. A partial explanation for this lack of relationship be- tween these parameters was attributed to the fact that the squids examined might have included individuals which hatched at different localities and which had developed under different temperature regimes (e.g., some squids could have been immigrants from a more southerly population exposed to warmer temperatures). This could well account for considerable confusion in the Illex data, as individuals of L. chinensis from dif- ferent seasons showed a different relationship between hectocotylus length and mantle length, and hectocotylus length and age. These seasonal growth patterns which were temperature-related did have an influence on the development rate of the hectocotylus, with slower- growing individuals eventually possessing a larger hectocotylus than their faster-growing counterparts. Different-sized hectocotili on similar-sized squids could therefore indicate that the squids had developed un- der different growth rates. This also suggests that the size ratio of many other structures to body size may also be influenced by the ambient temperature in which the squid develops. Parameter ratios and indices used for taxonomic purposes should therefore be cautiously employed, at least for nearshore squids, and, when- ever possible, parameters should be measured in Jackson: Cephalopod reproductive investment 269 individuals captured under similar environmental conditions. Idiosepius pygmaeus The pattern of gonad maturation in tropical squids and sepioids becomes more complex when the pattern of seasonal maturation of /. pygmaeus is also considered. As with L. chinensis, individuals of /. pygmaeus which grew during a cooler time of the year also reached a greater age. However, gonad maturation was modified in a different way compared with L. chinensis. For ex- ample, in both seasons males (including mature indi- viduals) had similar-sized testes at an age of 30^40 d suggesting that a minimal gonad size can be reached in just over a month. Nevertheless, as the cooler-season males continued to grow, the gonad continued to reach an appreciably greater size, for the same-sized indi- vidual. Thus, environmental constraints (such as tem- perature, food availability, or light levels) produced a different allometric gonad-soma relationship in this spe- cies which was not apparent in L. chinensis. The fact that /. pygmaeus has a short lifespan and rapid growth may account for the lack of females with mature ova. It is possible that ovum maturation could take place very rapidly just before egg deposition. Therefore, it could be difficult to capture females with mature oocytes unless it was just prior to egg deposi- tion. Alternatively, females with ripe ova may not have been near the water surface and therefore not avail- able for sampling. Idiosepius pygmaeus appears to be employing a 'trade-off in its seasonal reproductive tactics, that is, benefiting from one process bought at the expense of another (Begon et al. 1986). During the cooler period of the year, growth is slowed (presumably as a result of metabolic responses to temperature), therefore lifespan is increased as a necessity, since individuals take longer to reach adult size. As a result, /. pygmaeus appears to change its tactics by partitioning a greater amount of energy into gonads over the longer lifespan. Although a longer time-period is taken to reach matu- rity, there would be a reproductive advantage in that possessing larger gonads would increase reproductive output. A similar situation has been shown to exist with teleost fishes. Stearns (1976) has provided evi- dence from several species of teleosts, showing that the ratio of ovary weight to body weight (as a measure of reproductive effort ) increased with age. The fact that this phenomenon occurs in /. pygmaeus and not the larger L. chinensis may be due to the need for /. pygmaeus to maximize its reproductive chances because of the greater nearshore habitat variability, or because of the greater constraints this species faces due to its small body size. Acknowledgments I would like to thank J.H. Choat for comments through- out this project, C.H. Jackson for critically reading the manuscript and assisting with figures, and the crew of the research vessel James Kirby for specimen collec- tion. Manuscript comments by two anonymous review- ers and L.L. Jones were greatly appreciated. This research was supported by grants from the James Cook University of North Queensland Research Funding Panel. Citations Begon, M., J. L. Harper, & C. R. Townsend 1986 Ecology. Blackwell Sci. Publ., Oxford, 876 p. Boletzky, S. V. 1983 Sepia officinalis. In Boyle, P.R. (ed.), Cephalopod life cycles, vol. I, p. 31-52. Academic Press, London. Boyle, P. R. 1990 Cephalopod biology in the fisheries context. Fish. Res. (Amst.) 8:303-321. Coelho, M. L., M. D. Mallet, & R. K. O'Dor 1985. Evaluation of male reproductive features as ma- turity indices for short-finned squid {Hex illece- brosus). NAFO Sci. Counc. Stud. 9:107-115. Forsythe, J. W., & R. T. Hanlon 1988 Effect of temperature on laboratory growth, repro- duction and life span of Octopus bimaculoides. Mar. Biol. 98:369-379. 1989 Growth of the eastern Atlantic squid, Loligo forbesi Steenstrup (Mollusca: Cephalopoda). Aquacult. Fish. Manage. 20:1-14. Hanlon, R. T„ W. T. Yang, P. E. Turk, P. G. Lee, & R. F. Hixon 1989 Laboratory culture and estimated life span of the eastern Atlantic squid, Loligo forbesi Steenstrup, 1856 (Mollusca: Cephalopoda). Aquacult. Fish. Manage. 20:15-34. Hixon, R. F. 1980 Growth, reproductive biology, distribution and abundance of three species of loliginid squid (My- opsida. Cephalopoda) in the northwest Gulf of Mexico. Ph.D. thesis, Univ. Miami, Coral Gables FL, 233 p. 1983 Loligo opalescens. In Boyle, P.R. (ed), Cephalo- pod life cycles, vol. I, p. 95-114. Academic Press, London. Jackson, G. D. 1989 The use of statolith microstructures to analyze life-history events in the small tropical cephalopod Idiosepius pygmaeus. Fish. Bull, U.S. 87:265-272. 1990a The use of tetracycline staining techniques to determine statolith growth ring periodicity in the tropical loliginid squids Loliolus noctiluca and Loligo chinensis. Veliger 33:389-393. 1990b Age and growth of the tropical nearshore loliginid squid Sepioteuthis lessoniana determined from stato- 270 Fishery Bulletin 91(2). 1993 lith growth-ring analysis. Fish. Bull., U.S. 88:113- 118 1992 Seasonal abundance of the small tropical sepioid Idiosepius pygmaeus. Veliger 35:396-397. Jackson, G. D., & J. H. Choat 1992 Growth in tropical cephalopods: an analysis based on statolith microstructure. Can J. Fish. Aquat. Sci. 49:218-228. Kenny, R. 1974 Inshore surface sea temperatures at Townsville. Aust. J. Mar. Freshwater Res. 25:1-5. Mangold, K. 1983 Food, feeding and growth in cephalopods. Mem. Nat. Mus. Vic. 44:81-93. 1987 Reproduction. In Boyle, PR. (ed.), Cephalopod life cycles, vol. II, p. 157-200. Academic Press, London. Nesis, K. N. 1983 Dosidicus gigas. In Boyle, PR. (ed.), Cephalopod life cycles, vol. I, p. 215-231. Academic Press, London. Okutani, T. 1983 Todarodes pacificus. In Boyle, PR. (ed.), Cepha- lopod life cycles, Vol. I, p. 201-214. Academic Press, London. Rodhouse, P. G., & E. M. C. Hatfield 1990 Age determination in squid using statolith growth increments. Fish. Res. (Amst.) 8:323-334. Roper, C. F. E., M. J. Sweeney, & C. E. Nauen 1984 Cephalopods of the world. An annotated and illus- trated catalogue of species of interest to fisheries. FAO Fish. Synop. 125(3), 277 p. Stearns, S. C. 1976 Life-history tactics: a review of the ideas. Q. Rev. Biol. 51:3-47. Walker, T. A. 1981 Annual temperature cycle in Cleveland Bay Great Barrier Reef province. Aust. J. Mar. Freshwater Res. 32:987-991. Worms, J. 1983 Loligo vulgaris. In Boyle, PR. (ed.), Cephalopod life cycles, vol. I, p. 143-157. Academic Press, Lon- don. Yang, W. T., R. F. Hixon, P. E. Turk, M. E. Krejci, W. H. Hulet, & R. T. Hanlon 1986 Growth, behavior, and sexual maturation of the market squid, Loligo opalescens, cultured through the life cycle. Fish. Bull, U.S. 84:771-798. Abstract.— Recent papers have provided new insights into the prob- lem of estimating von Bertalanffy growth parameters from tag- recapture data. In particular, the in- consistency and bias of Fabens' (1965) estimates appear to have been addressed by James (1991). Using simulation, we examine the pattern of bias associated with different er- ror assumptions for Fabens' esti- mates, weighted Fabens' estimates proposed by James, and a robust method also proposed by James. Our results corroborate James' finding that his robust estimates can be sig- nificantly less biased than other methods. We then apply these esti- mators to tag-recapture data ob- tained for sablefish Anoplopoma fim- bria found in the Gulf of Alaska and off the U.S. west coast, and Pacific cod Gadus macrocephalus found in the eastern Bering Sea. These spe- cies are difficult to directly age, so tag-recapture data provide welcomed independent estimates of growth pa- rameters and an indirect method of validating age-determination crite- ria. The von Bertalanffy parameter estimates using tag-recapture data and James' method were most simi- lar to estimates calculated directly from length-at-age data. Estimating von Bertalanffy growth parameters of sablefish Anoplopoma fimbria and Pacific cod Gadus macrocephaius using tag-recapture data Daniel K. Kimura Allen M. Shimada Sandra A. Lowe Alaska Fisheries Science Center National Marine Fisheries Service. NOAA 7600 Sand Point Way NE. Seattle. Washington 98 1 1 5-0070 Manuscript accepted 22 January 1993. Fishery Bulletin, U.S. 91:271-280 ( 1993). In its most common form, the von Bertalanffy growth curve has three parameters (L,,, K, t0). Conventional interpretation of these parameters is that L, is asymptotic size, K de- scribes the growth rate, and t0 de- scribes the age at length-0. Fabens (1965) was apparently the first to show how least-squares could be used to estimate two of these parameters (L_, K) from tag-recapture data. If at least one additional value of length- at-age was known, then t0 could be estimated, and hence all values of length-at-age could be estimated di- rectly from recapture data without recourse to direct ages from indi- vidual specimens. Besides providing growth-curve estimates, such a pro- cedure would seem to provide an in- direct validation of ageing criteria, since estimated growth parameters from length-at-age data and tag- recapture data could be compared. However, comparisons of growth curves estimated using Fabens' method and ordinary length-at-age data seemed to provide evidence that Fabens' method provided biased pa- rameter estimates. Following the sug- gestion of Chapman (1961) and oth- ers, Sainsbury (1980) showed how individual variability [i.e., the possi- bility that each fish has different val- ues of (L., K)], could lead to bias in population parameter estimates. Francis (1988a) argued from a sampling point of view that von Bertalanffy parameters calculated from length-at-age and tag-recapture data could be different. He also ar- gued that average growth parameters for individual fish may not describe the population growth curve (Francis 1988b). Mailer & deBoer (1988) showed mathematically and with simulation that Fabens' estimates, and related estimates of Kirkwood & Somers (1984), could be inconsistent. Kirkwood (1983) suggested combin- ing length-at-age and tag-recapture data into a single likelihood model, but did not address the problem of bias caused by the interpretation of tag-recapture data. A recent paper by James (1991) suggests that bias in Fabens' esti- mates could arise from bias in the estimating functions. James shows how weighted least-squares can be used to derive improved estimates for Fabens' method under the usual "observational error" assumption. James also gives distribution-free es- timates for a more general model in which, in addition to the usual ob- servational error, L . itself is allowed to vary among individual fish. Al- though James' estimates appear to be consistent and less biased than the least-squares estimates, they also appear to be less efficient, usually 271 272 Fishery Bulletin 91 12), 1993 having larger standard errors. Nevertheless, James' method provides us with more robust parameter esti- mates. This paper illustrates how critical is statistical meth- odology to the estimation of growth parameters from tag-recapture data. We show, by simulation and through analysis of actual data, that different estima- tion methods can easily lead to very different param- eter estimates. With the help of simulation, we feel that realistic choices can be made among these esti- mators. For the actual datasets chosen for this paper, we are able to compare growth-curve parameter esti- mates from tag-recapture data with parameter esti- mates calculated from length-at-age data, using ages estimated from directly counting annuli. This provides another criterion for choosing among estimators. Methods We apply estimators considered by James (1991). The "unweighted Fabens" estimates (Fabens 1965) is the widely used historical method. What we call the "weighted Fabens" estimates, and James' estimates, were derived by James (1991). The weighted Fabens' method weights the residuals by an inverse variance estimate and appears to work well for the observa- tional error model. James' estimators appear to be less prone to bias than the unweighted Fabens' estimators, and more robust to the presence of variability in L„ than the weighted Fabens' estimators. A more theo- retical description of these estimators can be found in James (1991). First we present a simulation study that corrobo- rates the main findings of James ( 1991). Our study is more systematic than that of James ( 1991 ), using three estimation methods and error structures and calculat- ing mean square error. Our simulation is based on parameters estimated for a marine teleost, Pacific whit- ing Merluccius productus, while James' simulations dealt with two shellfish species. Nevertheless, these simulations should be considered an addendum to James' original work. Using notation similar to James (1991), consider a population having von Bertalanffy parameters (L0, K,,) which we wish to estimate. L0 refers to the average value of L. in the population, where L. might vary among individual fish. We assume K=K,> does not vary among fish. Also, recall that t„ cannot be estimated from tag-recapture data alone. To determine t„ we need the average length of at least one age-group which can be supplied by direct age data (i.e., ages obtained from directly counting annuli), or the modal length-frequency of a dominant year-class of known age. Suppose we observe the size at release (y,,) and re- capture (y2l) of i=l n fish, and the ages at release and recapture are (t„ t,+d,). Suppose also that inde- pendent normally-distributed errors (e,„ e2i, e3l) enter into these observations in three possible ways: yu = (L„ + £.,,)[ 1.0 - exp(-K,,t,)] + e,, y2, = (L„ + e3l)[1.0 - exp(-Ko(t, + d,))] + e2i. We refer to e,, and e2i as "observational errors" and e:!, as "variability in L_." The three estimators of (L0, Kq) considered in this paper are all based on the "residuals": n, = y2, - y., - (L0 - y„)[1.0 - exp(-Kodi)]. These estimators are: 1 Unweighted Fabens, estimated by minimizing In,,2 2 Weighted Fabens, estimated by minimizing Ilri.VU.O+exp^Kod,))]. 3 James, estimated by solving the simultaneous equations £n =0 and ld,n,=0. Unweighted and weighted Fabens' estimates, and their standard errors, were calculated using nonlinear least-squares methods. Define g^Iq, and g2=Id,q,. James' estimates were calculated as suggested by James ( 1991) by solving for "L(l in terms of K,, from gl followed by substitution into g2" and then applying the bisection method to estimate K,, (see Press et al. 1986). This method appeared computationally robust and suit- able for simulation studies. The variance estimates of these parameter estimates were calculated according to James (1991): cov(L,K) = G'DfG')-1, where G dg-2 dg2 , dKn dL(l , £ = Ir|2 1.0 d,s \ d, d2ij and the right-hand expressions are evaluated at the estimated parameters. Our simulation considered three different error as- sumptions. Simulation 1 The observational error model where all fish are assumed to have common growth param- eters, with variation due wholly to observational er- rors e,„ e2l ~ N(0,o2). Here, all e3l are assumed to be zero. Simulation 2 The variation in L. model, where all variation in the observations are due to e3l ~ N(0,o2), Kimura et al. Tag-recapture data analysis of Pacific cod growth 273 and all observational errors e,, and e2, are assumed to be zero. Simulation 3 Both observational errors and varia- tion in L„„ are present in this model, with e^, £2l, £3, ~ N(0,o2). For these simulations, we assume that the true popu- lation parameters are (L0= 61.23, K,,=0.296), param- eters previously estimated for female Pacific whiting (Kimura 1980). Measured in years, t,, and d, were uni- formly and independently distributed over the inter- val [1, 5], assuming 365d/yr. New values for tb and d, were generated for every replication of the simulation. Analytic standard errors of parameter estimates were estimated using nonlinear least-squares meth- ods, and the variance estimate described above for James' method. In addition, the sample standard er- rors (i.e., the sample standard deviation of the repli- cate parameter estimates) are also provided. The rea- son for this is that the simulations included variation in th and d, that would not be included in the analytic estimates of standard error, and the probability that the analytic standard errors are themselves biased due to failure of assumptions. For the actual datasets, ana- lytic standard errors were used. For simultaneously considering bias and variance, we include estimates of mean square error (MSE) calculated in the usual way: I (estimate - true fin. Following a presentation of simulation results, we analyze tag- recapture data collected for sablefish Anop- lopoma fimbria from the Gulf of Alaska and off the U.S. west coast, and Pacific cod Gadus macrocephalus from the eastern Bering Sea (Fig. 1). Because sable- fish and Pacific cod are difficult species to directly age, ^^^V*^*** Figure 1 Tag and recovery areas for Pacific cod in eastern Bering Sea (BS). sablefish in Gulf of Alaska (GOA). and sablefish off U.S. west coast (WC). their growth-curve parameters estimated from tag-re- capture data are of special interest. The von Bertalanffy parameters estimated from tag-recapture data were compared with parameters estimated from length-at- age data based on direct ages (i.e., ages from directly counting annuli). For sablefish, we use length-at-age datasets based on break and burn ages (Chilton & Beamish 1982) and parameters estimated using nonlinear least-squares (Kimura 1980). For Pacific cod, we compare tag-recapture results with published von Bertalanffy parameter estimates (Thompson & Bakkala 1990). Results Simulation results For the simulation results given here, the tag-recap- ture sample size was n=300, simulations were repli- cated 200 times, and the normal errors (£h, e2l, e3l) were all independently distributed with o2=25. There- fore, the sample size was realistically small and the error variances were substantial. For each estimation method and parameter, simulation results (Table 1) were summarized by four entries: sample mean of the estimated parameter, mean analytic standard er- ror, sample standard error (i.e., the standard devia- tion of estimated parameters), and mean square error. The unweighted Fabens' estimate of L0 is biased low in Simulation 1, biased high in Simulation 2, with biases apparently canceling each other in Simulation 3. For the weighted Fabens' estimates, the L0 estimate is unbiased for Simulation 1, but biased high for Simulations 2 and 3. James' estimate of L0 is unbi- ased in Simulation 2, and only modestly biased in Simulations 1 and 3. Biases in L0 and K, esti- mates are in the opposite direc- tion, as might be expected from the negative correlation between these parameter estimates. Standard errors are uniformly smallest for the unweighted Fabens' method, in the middle for the weighted Fabens' method, and largest for James' method. Performance as measured by mean square error was entirely dependent on the assumptions of the simulation. For Simulation 1 (i.e., observational errors) the weighted Fabens' method MSE was smallest; for Simulation 2 274 Fishery Bulletin 91(2). 1993 Table 1 Simulation results using three different methods to estimate von Bertalanffy parameters (L0=61.23, K„=0.296> under .hree different error assump- tions: Simulation 1 (observational error), Simulation 2 (variation in L_), and Simulation 3 (both observational error and variability n L„). For each estimation method and parameter, simulation results are summarized by four entries sample mean jf the estimated parameter, mean analytic standard error, sample standard error (i.e., standard deviation of esti- mated parameters), and mean square error. Method Parameter Sim. 1 Sim. 2 Sim. 3 Unweighted mean(L„) 58.00 65.01 60.39 Fabens mean(K„) 0.379 0.238 0.315 analytic SE(L„) 1.042 0.981 1.423 analytic SElK,,) 0.031 0.011 0.028 sample SE(L„) 0.987 0.991 1.370 sample SElK,,) 0.030 0.012 0.026 MSE(L0) 11.426 15.272 2.576 MSElK,,) 0.00775 0.00347 0.00103 Weighted mean(L0) 61.24 67.05 67.21 Fabens mean(Ko) 0.298 0.215 0.216 analytic SE(L(1) 1.355 1.154 2.347 analytic SE(Kq) 0.021 0.009 0.018 sample SE(L„) 1.542 1.205 2.921 sample SElK,,) 0.030 0.012 0.027 MSElL,,) 2.365 35.347 44.197 MSE(K„) 0.00091 0.00671 0.00714 James mean (Lo) 62.22 61.23 62.06 mean (K,,) 0.294 0.299 0.300 analytic SE(L„) 4.099 1.588 4.625 analytic SElK,,) 0.064 0.028 0.071 sample SE(L0) 4.241 1.521 5.332 sample SE(K„) 0.066 0.027 0.069 MSE(L„) 18.871 2.303 28.986 MSE(K„) 0.00432 0.00072 0.00471 (variability in LJ the James' method MSE was the smallest; for Simulation 3 (i.e., both observational error and variability in LJ the unweighted Fabens' method MSE was the smallest. It is clear from these simulation results that unweighted or weighted Fabens' estimates of von Bertalanffy parameters can be significantly biased. It is also clear that standard errors for James' estimators tend to be larger than standard errors for the unweighted and weighted Fabens' estimators. Estimates from sablefish tag-recapture data Sablefish is a species characterized by rapid growth at young ages followed by extremely slow growth at older ages. In tag- recapture data, it is common for sablefish to show negative growth increments (Figs. 2,3,5,6). The exact reason for this ap- parent shrinkage is uncertain. However, Sasaki (1985) noted that shrinkage was usually reported for larger male (>60cm) and female (>70cm) sablefish. This suggests that such occur- rences are due to the combined influence of slow growth and measurement error. Furthermore, Beamish & Chilton (1982) noted that the freezing of commercially-caught fish could cause actual shrinkage. Interestingly, Beamish & Chilton (1982) also noted that data from freshly-caught fish measured by trained bi- ologists occasionally showed shrinkage, indi- cating that shrinkage can occur in the ocean environment. Considering these factors, all recoveries, including those having negative growth increments, were used in the sable- fish data analyses. Although this species is extremely difficult to age (Kimura & Lyons 1991), age-determi- nation criteria have been validated using oxytetracycline tags "(Beamish & Chilton 1982) and radioisotope methods (Kastelle 1991). Because of the difficulty of directly de- termining ages for sablefish, estimating growth using tag-recapture data is of par- ticular interest. 10 30 50 70 90 LENGTH AT RELEASE (CM] o O o o o o o i , . .- ''■' ' •■' o c9 r o SO 1 0 2 4 6 8 10 12 14 TIME TO RECAPTURE (YR) Figure 2 Length-frequency at time of release (top) and re- lationship between time at liberty and growth (bottom) for male sablefish in Gulf of Alaska. Kimura et al.: Tag-recapture data analysis of Pacific cod growth 275 x i— o >- o z UJ 3 o ■SI- ■ 1 LT' III 1 ui 1 1 m 11 III ii i nl II tu i i 10 30 50 70 90 LENGTH AT RELEASE (CM) o UJ en u j_ o rr O <*■ 1 ' 1 ' 1 1 1 1 o o o O O Q °Q° <§> °°0 o 8~ SOo° - CM o o o o - *— ffijsg - ro u :' r~) UJ CM _> O i. O O l„l 111 III ill L, . 10 30 50 70 90 LENGTH AT RELEASE (CM) CM - in o ° o <*> O 0° o° o oo o 0 3y~-r^ o 6«o o c cP O I 0?° 3 - I!) ' a UJ rr o o cc o I 2 4 6 8 12 16 TIME TO RECAPTURE (YR) Figure 5 Length-frequency at time of release (top) and relationship between time at liberty and growth (bottom) for male sablefish off U.S. west coast. O or 0 30 50 70 90 LENGTH AT RELEASE (CM] o uj o 2 CM O un O o DC I o , °0 ° o - 0 O- SKf1' ~ ' ■ • BPSjo ~o " o u - Pqc^1 u O P ,n i.i.i 1 0 2 4 6 8 10 14 TIME TO RECAPTURE (YR) Figure 6 Length-frequency at time of release (top) and relationship between time at liberty and growth (bottom) for female sablefish off U.S. west coast. Table 4 Comparison of von Bertalanffy growth parameters for sab efish off the U.S. west coast, esti- mated from tag-recap ure and length-at-age data. Tag-reca pture sample sizes: mal es rc=467, females n=616. Direct- ages sampl 3 sizes: males n =714, females n=8U. L, K,, to SE(L0) SE(Ko) SE(t0) Males Unweighted Fabens 65.4 0.081 2.08 0.014 Weighted Fabens 144.2 0.009 100.2 0.010 James 56.6 0.556 0.65 0.108 Length-at-age 54.7 0.472 -1.82 0.23 0.055 0.51 Females Unweighted Fabens 71.5 0.110 1.79 0.013 Weighted Fabens 130.1 0.020 33.2 0.008 James 61.4 0.481 0.92 0.085 Length-at-age 61.0 0.499 -0.81 0.35 0.047 0.32 278 Fishery Bulletin 91(2). 1993 9> ° u<® ^0 O (2D, O 0 10 20 30 40 50 60 AGE (YR) 0 10 20 30 40 50 60 AGE (YR) Figure 7 von Bertalanffy growth curves for male (top) and female (bottom) sablefish off U.S. west coast, calculated from length-at-age data. 10 30 50 70 90 LENGTH AT RELEASE (CM) o CO O rr o o cr o 2 3 4 5 6 TIME TO RECAPTURE (YR) Figure 8 Length-frequency at time of release (top) and relationship between time at liberty and growth (bottom) for Pacific cod in eastern Bering Sea. Table 5 Comparison of von Bertalanffy growth parameters for Pacific cod in eastern Bering Sea, esti- mated from tag-recapture data using all growth ncrements in =284), anly positive growth incre- ments (n=252), and length-at-age data. Sexes were combined and von Bertalanffy parameters estimated from length-at-age data tvere taken from Thomp son & Bakkala (1990). NS = no solution, NA = not available. L,i K, to SE(L„ SE(Ko) SE(t„i All growth increments Unweighted Fabens 117.0 0.148 9.69 0.029 Weighted Fabens 271.0 0.035 147.81 0.024 James NS NS NS NS Positive growth increments Unweighted Fabens 126.3 0.133 8.35 0.020 Weighted Fabens 175.1 0.070 28.19 0.017 James 104.3 0.222 10.55 0.065 Length-at-age 105.4 0.237 1.06 NA NA NA Kimura et al.: Tag-recapture data analysis of Pacific cod growth 279 large biases. However, the standard errors for James' estimators tended to be larger, so in terms of mean square error the superior estimation method depended on the assumptions we made concerning variability. Weighted Fabens' estimators performed well when error was restricted to observational error, but were badly biased when variability in L_ was introduced into the simulation. In contrast, James' estimators per- formed well when there was no observational error and error was due only to variability in L,. The unweighted Fabens' estimators were biased when er- ror was solely observational, and biased when vari- ability was due solely to variability in L,, but biases apparently canceled when both errors were present. It is difficult to recommend the unweighted or weighted Fabens' estimators because they are both sub- ject to large biases. It seems difficult to assume biases will cancel, or that there will be no variability in the data due to variability in L,„. On the other hand, the James' estimators appeared unbiased but tended to have larger standard errors. Our experience with analyzing the small sample-size dataset available for Pacific cod indicates that James' estimators may require at least moderate sample sizes, the required size depending on variability in the data. If sample sizes are sufficiently large, reducing bias can be far more important than increasing efficiency. Therefore, it seems important that James' estimators be used to at least guard against possible biases. More experience with James' method will probably be needed before researchers can finally decide the correct scope of application for this new method. In this paper we have used von Bertalanffy param- eter estimates based on direct ages as the yardstick by which to measure bias for the three different estima- tors. Of course, von Bertalanffy parameter estimates based on length-at-age data might themselves be bi- ased. Except for the small sample-size dataset avail- able for Pacific cod, James' estimators compared well with estimates based on direct age data. Unweighted Fabens' estimators compared well with estimates based on direct ages for one dataset. Weighted Fabens' esti- mators of L0 appeared to be biased high for all three datasets. This pattern of bias suggests that the tag-recapture data in these actual datasets contain significant amounts of variability due to variability in L„. This is also evidenced by the length-at-age data (see Figs. 4,7). For the three stocks of fish studied in this paper, we conclude that there are no inconsistencies in growth described by direct ages and growth described by tag- recapture data. Without James' estimators, this con- clusion could not be made. We feel that James' method represents a significant step forward, and that datasets previously analyzed using Fabens' method might ben- efit by being reexamined using James' method. How- ever, results may be the same as Fabens' estimates, or quite different, depending upon the nature of variabil- ity in the data. Still, problems remain when evaluating growth from tag- recapture data. McFarlane & Beamish (1990) con- cluded that external tags could markedly affect the growth of sablefish. We have shown that different fishing gear types may select for slower- or faster- growing individuals. For the west coast sablefish tag- recapture datasets, growth increments appeared larger for fish recovered with pot gear compared with trawl gear. Therefore, sampling gears always play a role in describing the growth parameters we estimate. Acknowledgments We thank R.A. Mailer, an anonymous referee, and the editorial staff of the journal for comments that greatly facilitated our revision. We thank FR. Shaw for provid- ing access to the sablefish tag-recovery database. This study was motivated by original insights into tag-re- capture growth parameter estimation from I.R. James. Citations Beamish, R. J., & D. E. Chilton 1982 Preliminary evaluation of a method to determine the age of sablefish (Anoplopoma fimbria). Can. J. Fish. Aquat. Sci. 39:277-287. Chapman, D. G. 1961 Statistical problems in dynamics of exploited fish populations. Proc. Berkeley Symp. Math. Stat. Probab. 4:153-168. Chilton, D. E., & R. J. Beamish 1982 Age determination methods for fishes studied by the Groundfish Program at the Pacific Biolog- ical Station. Can. Spec. Publ. Fish. Aquat. Sci. 60:1- 102. Fabens, A. J. 1965 Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. Francis, R. I. C. C. 1988a Are growth parameters estimated from tagging and age-length data comparable? Can. J. Fish. Aquat. Sci. 45:936-942. 1988b Maximum likelihood estimation of growth and growth \ ariability from tagging data. N.Z. J. Mar. Freshwater Res. 22:42-51. James, I. R. 1991 Estimation of von Bertalanffy growth curve pa- rameters from recapture data. Biometrics 47:1519- 1530. 280 Fishery Bulletin 91(2). 1993 Kastelle, C. R. 1991 Radioisotope age validation and an estimate of natural mortality from the gonad to somatic weight index for sablefish (Anoplopoma fimbria). M.S. the- sis, Univ. Wash., Seattle, 87 p. Kimura, D. K. 1980 Likelihood estimates for the von Bertalanffy growth curve. Fish. Bull., U.S. 77:765-776. Kimura, D. K., & J. J. Lyons 1990 Choosing a structure for the production ageing of Pacific cod [Gadus macrocephalus). Int. N. Pac. Fish. Comm. Bull. 50:9-23. 1991 Between-reader bias and variability in the age-determination process. Fish. Bull., U.S. 89:53- 60. Kirkwood, G. P. 1983 Estimation of von Bertalanffy growth curve parameters using both length increment and age- length data. Can. J. Fish. Aquat. Sci. 40:1405- 1411. Kirkwood, G. P., & I. F. Somers 1984 Growth of two species of tiger prawn, Penaeus esculentus and P. semisulcatus, in the western Gulf of Carpentaria. Aust. J. Mar. Freshwater Res. 35:703- 712. Mailer, R. A. & E. S. deBoer 1988 An analysis of two methods of fitting the von Bertalanffy curve to capture-recapture data. Aust. J. Mar. Freshwater Res. 39:459-466. McFarlane, G. A., & R. J. Beamish 1990 Effect of an external tag on growth of sablefish (Anoplopoma fimbria) and consequences to mortality and age at maturity. Can. J. Fish. Aquat. Sci. 47:1551-1557. Press, W. H., B. P. Flannery, S. A. Teukolsky, & W. T. Vetterling 1986 Numerical recipes (the art of scientific com- puting). Cambridge Univ. Press, 818 p. Sainsbury, K. J. 1980 Effects of individual variability on the von Berta- lanffy growth equation. Can. J. Fish. Aquat. Sci. 37:241-247. Sasaki, T. 1985 Studies on the sablefish resources in the North Pa- cific Ocean. Bull. Far Seas Fish. Res. Lab. 22:1-108. Thompson G. G., & R. G. Bakkala 1990 Assessment of the eastern Bering Sea Pacific cod stock using a catch-at-age model and trawl sur- vey. Int. N. Pac. Fish. Comm. Bull. 50:215-236. Abstract. -Vertical distribution patterns of haddock Melanogram- mus aeglefinus and Atlantic cod Gadus morhua are summarized from eight research cruises on Georges Bank, during spring and summer 1981-86. Eggs and larvae were sampled at discrete depths with a lm2 MOCNESS, pelagic juveniles with a 10 m2 MOCNESS, and re- cently-settled juveniles with a bot- tom trawl supplemented with sub- mersible observations. In well-mixed waters during early spring, eggs and larvae were distributed throughout the water column. Upon stratifica- tion of waters over the outer margin of Georges Bank in mid-May, the lar- val population was concentrated in the thermocline; the stronger the stratification, the more larvae were confined to this depth zone. At well- mixed shoal sites, pelagic larvae re- mained distributed throughout the water column. Differences in day and night dis- tribution patterns provided evidence to indicate that larvae began migrat- ing as small as 6-8 mm; however, clear sampling differences were evi- dent for larvae only at body lengths of >9-13mm. Larvae tended to be deeper during the day and shoaler by night; larger larvae had a greater vertical range. Pelagic juveniles (>20 mm) occurred deeper in the wa- ter column as they grew. By mid- July most juveniles (-40 mm) were associated with the bottom. The transition from pelagic to demersal life i -40-100 mm) probably occurs over a period of 1-2 mo. Recently- settled juveniles remained demersal by day and migrated upwards 3-5 m at night. Their excursions off-bottom decreased in amplitude with larger sizes. Haddock remained closer to bottom both day and night at a smaller size than cod. The diel ver- tical behavior of cod may be strongly related to the light-dark cycle, whereas behavior of haddock was more variable. Some differences in vertical distributions between had- dock and cod can be attributed to morphology and feeding behavior. Vertical distribution patterns and diel migrations of larval and juvenile haddock Melanogrammus aeglefinus and Atlantic cod Gadus morhua on Georges Bank* R. Gregory Lough David C. Potter Northeast Fisheries Science Center. National Marine Fisheries Service, NOAA Woods Hole. Massachusetts 02543 Manuscript accepted 9 February 1993. Fishery Bulletin, U.S. 91:281-303 1 1993). On Georges Bank off the northeast coast of the United States (Fig. 1, page 293), stocks of haddock Melano- grammus aeglefinus L. and Atlantic cod Gadus morhua L. historically have been important components of the American and Canadian demer- sal fishery (Murawski et al. 1983, Overholtz & Tyler 1985). Eggs and larvae of both species are dominant in the spring ichthyoplankton com- munity (O'Boyle et al. 1984, Sherman et al. 1984). Research survey data have shown that the size of a year- class is usually established by the end of the first year of life (Sissen- wine 1984). At the Northeast Fisher- ies Science Center (NEFSC) since 1976, considerable research effort has been focused toward describing the place and time of spawning of had- dock and cod, dispersal of eggs and larvae, and survival and recruitment of juveniles on Georges Bank. Field, laboratory, and modeling results were integrated to evaluate important mortality processes during the first year of life. Beginning in spring 1980, a series of interdisciplinary, process- oriented cruises used discrete depth samplers to investigate mortality due to starvation by describing the fine- scale spatio-temporal variability of gadid larvae in relation to their prey during the first few weeks post- hatching (Lough 1984, Laurence & Lough 1985). In 1984, a series of summer and early-fall surveys of ju- venile gadids was initiated to docu- ment the distribution and abundance of the 0-group fish and their transi- tion to the demersal phase when pre- dation may be a dominant mortality factor (Cohen et al. 1988). Based on these studies, we now have sufficient information to de- scribe the life history of haddock and cod during the first 6-9 mo. Both spe- cies spawn on the northeast part of Georges Bank (Fig. 1). Peak spawn- ing time for cod is late February- early March, and early April for had- dock. The egg patches drift southwest a few kilometers per day in the clock- wise residual drift and hatch within 2-3 wk (Lough 1984). By May, larval concentrations can be found along the southern edge of Georges Bank be- tween the 60 m and 100 m isobath. In late-spring to fall, a seasonal ther- mocline develops to a depth of 10- 30 m in waters deeper than -60 m. Water shoaler than 60 m is vertically well-mixed throughout the year by the strong ( >60 cm/s ) rotary tidal cur- rents. Larvae >30 d can be found shoaler than the 60 m isobath as the patch moves southwest, and by late June some fraction of the population has moved onto the shoals of west- ern Georges Bank (Walford 1938, *MARMAP Contribution FED/NEFC 90-01. Woods Hole Laboratory, Northeast Fisheries Science Center, National Marine Fisheries Ser- vice, NOAA. 28 I 282 Fishery Bulletin 9 1(2), 1993 Figure 1 Location of vertical distribution studies, 1981-86, on Georges Bank, designated by year and sampled sites. Sites (Roman numerals) within a cruise have same symbols. Lough 1984, Smith & Morse 1985, Lough & Bolz 1989). By late July-early August, the highest catches of juve- nile gadids are usually found on eastern Georges Bank (Lough etal. 1989). Haddock and cod undergo typical development and morphological transformation at sizes >20mm stan- dard length (SL, -2-3 mo post-hatching) (Fahay 1983, Auditore et al. 1993). Their transition from pelagic to demersal life occurs when they are >25mm in mid- summer (Lough et al. 1989). Information on larval and juvenile vertical distribu- tions and diel migrations are needed to (a) study their transport by currents, (b) compare vertical distribu- tions of their prey and predators, and (c) make accu- rate abundance estimates of various stages caught by different sampling gear. In this paper we summarize data on the vertical distribution of 0-group haddock and cod during spring and summer on Georges Bank from 1981 to 1987. The following questions on the early life of haddock and cod were addressed. ( 1 ) What is the vertical distribution pattern of larvae and juve- niles on Georges Bank? (2) What is the timing and range of diel migratory behavior in relation to devel- opment? (3) What is the transition period from pelagic to demersal life for an individual fish? (4) How do light, temperature, salinity, and currents influence ver- tical distribution and migratory behavior? (5) How does gear selectivity bias vertical distribution patterns? Methods Sampling protocol General cruise objectives and sampling strategy for the larval-juvenile fish studies are described by Lough ( 1984). A concentration of larvae on Georges Bank was located by a fine-scale grid of stations using standard bongo-net gear. Based upon real-time sample analyses made during the grid survey, sites were selected within well-mixed and stratified waters for the comparative vertical time-series observations. Haddock and cod were collected by three gear types on eight cruises (Table 1). Pelagic larvae and juveniles were sampled with an electronically-controlled, multiple opening/ closing net and environmental sensing system (MOCNESS), with a lnr or 10 m2 mouth opening (see Wiebe et al. 1976, 1982, 1985). A Yankee 36 bottom trawl was used to sample the recently-settled juvenile fish. The numbers of day, night, and twilight hauls at each site are listed in Table 1. The MOCNESS used to sample the larval stage has an effective mouth opening of lm2 when towed at a 45° net angle near 3.7km/h (2kn). The unit consists of nine 0.333 mm mesh nets that open and close sequen- tially on command. Temperature (and salinity begin- ning in 1985) was obtained simultaneously with stan- dard net-sampling parameters such as depth and volume of water filtered. When a concentration of lar- vae was located, a drogue was deployed at a depth of 15 m to mark the water mass for further sampling. Thereafter, the lm2 MOCNESS was used to sample every 3-6 h at the drogue for a period of 24-48 h. Dis- crete depths were sampled at 10 m intervals from the surface to within 5 m of the bottom within a 5min period. Each net filtered -250 m3 of water. The 10 m2 MOCNESS, used to collect the pelagic juveniles, has five 3.0 mm mesh nets. Because juve- niles tend to remain localized for at least 2-4 d, sam- pling was conducted at fixed sites that were deter- mined to contain high concentrations of juveniles. The vertical distribution of pelagic juveniles was sampled every 6 h. Discrete depth strata ranged 10-30 m from the surface to within 5 m of the bottom; each net filtered between 15,000 and 20,000 m3 of water during a 30 min tow. A Yankee 36 bottom trawl was used to catch re- cently-settled juveniles. Trawl specifications are found in Azarovitz (1981). The opening height and width of the trawl was 3.2 X 10.4 m. The trawl had graded stretch mesh of 127-13 mm (mouth to cod-end). Standard roller gear (40.6cm diameter) was used on the Albatross IV 86-03 June 1986 cruise; however, rollers were replaced with a sweep chain of rubber disks (11.4cm diameter) on the Delaware II cruises 85-05 of July 1985 and 85- 06 of August 1985, to collect fish close to bottom. The standard bottom-trawl set was made at 6.5km/h (3.5 kn) for 30 min on the bottom. Trawl sampling was conducted for 24 h. All specimens were initially fixed in either 4% for- maldehyde-seawater solution or 90% ethyl alcohol and Lough and Potter Vertical distribution of Melanogrammus aeglefinus and Gadus morhua 283 then subsequently preserved in 70% ethyl alcohol and water. Temperature and salinity profiles were obtained from the MOCNESS, XBT drop, or CTD cast. Data analysis Larvae and pelagic juveniles were measured to the nearest O.lmmSL. Demersal juveniles from the bot- tom trawl were measured to the nearest mm. Length of haddock and cod were corrected for shrinkage to live-length using the method of Bolz & Lough (1983). Catches from the lm2 MOCNESS were initially stan- dardized to n/100 m3 of water, and 10 m2 MOCNESS catches to n/10,000 m3. The fish were grouped into size- classes based on developmental characters (Auditore et al. 1993): 2-5, 6-8, 9-13, 14-19, 20-29, 30-39 40-49, 50-59, and 60-69 mm. As an initial step, the percent of a size-class at each depth was plotted with the temperature and salinity profile for each tow (figures not shown). Tows were classified as day, night, or (in a few cases) twilight, i.e., lh before and after sunrise and sunset. For each pelagic tow, the mean number of fish within a time- block and size-class was calculated for each depth by summing all hauls at that site. Day and night vertical profiles are plotted from these arithmetic means. Also, abundance of larvae was integrated over the water column, in terms of n/10 m2 for the 1 m2 MOCNESS catches and h/1000 m2 for the 10 m2 MOCNESS catches. Abundance-weighted mean depth or center of density distributions of size-classes were calculated using the method of Miller et al. ( 1963). To examine trends, mean depths of selected sites were plotted in time-series with water column thermoclines. Since salinity changes throughout the water column are generally small (9-13mm. 284 Fishery Bulletin 91(2), 1993 A. Haddock 2 -5mm 6 • 8 mm 30 20 10 0 tO 20 30 03 02 0.1 0 01 0.2 03 r] 66/IOm2 108/ 10m2 06/tOm2 I8/I0m2 2- 5mm 6 ■ 8 mm 9 • 13 mm 15 10 5 0 5 10 15 15 10 5 0 5 10 15 60 40 20 0 20 40 60 : ' I lV "" 2C E DAY r NIGHT r ^ '. 1 50 r L-i 60 m L J 28/)0m2 37/IOm2 I4/I0m2 30/I0m 98/IOm2 213/IOm2 SALINITY (psu) 32TJ0 2 46 8 3300 2 14 -19 mm 20 -29 mm TEMPERATURES) 15 K) 5 0 5 10 15 6420246 0 5 10 IS 10 ^1 20 E £30 &40 o : I 50 60 u — i 70 L TEMP/ | SAL, lO/IOm2 62/IOm2 3/IOrn2 14/IOfn2 Figure 2 Mean day and night vertical distribution (i n/lOOnv1) of (A) haddock Melanogrammus aeglefinus and (B) cod Gadus morhua size-classes of larvae collected by the lm- MOCNESS on well-mixed Site 81-1 during 27-29 April 1981. Estimated water column abundances (rc/lOnrl are included for day-night comparisons at bottom of plots. Composite temperature and salinity profiles are plotted as depth stratum mean (•) and 95^ confidence limits ( ). Mean depths of the four smallest cod size-classes are plotted for each tow within a 24 h period over the 2.5 d of sampling at Site 81-1, and a trend line con- nects the mean depths of abundance (Fig. 3). There appears to be a daily pattern in vertical distribution. The smaller size-classes, 2-5 mm and 6-8 mm, re- mained at a depth of -35 m day and night. In contrast, larger size-classes of cod, 9-13 mm and 14-19 mm, showed a consistent diel migration from a mean depth of 40 m at noon to 20-25 m after sunset. At the well-mixed shoal (35-58 m) Site 81-111 on 27 May 1981, 2-13 mm haddock and 6-13 mm cod lar- vae were distributed through the water column (Fig. 4 1 with mean depths of abundance generally at 20 m and 40 m. Because of the limited number of tows, a significant difference in day-night patterns could not be determined. TIMEth ) TIME(h) 1200 1800 2400 0600 0600 1200 1600 2400 0600 u;40 i| . i | I . T I I | I i 6-8mm 0600 1200 1800 2400 0600 0600 1200 1800 2400 0600 Figure 3 Abundance weighted-mean depths (•) of larval cod Gadus morhua size-classes plotted over a 24 h period on well-mixed Site 81-1 during 27-29 April 1981. Line connecting replicate points represents the average population trend. 3/IOm2 04/IOm2 ■ Single night haul B Cod 6 B mm 9 ■ 13 mm 150 100 50 0 50 100 150 150 100 50 0 50 100 150 203/IOm2 73/ 10m2 235/IOm2 21 1/ 10m2 SALINITY(psu) 3200 2 4 6 8 3300 2 0 I 20 £30 ;' V P, 1 1 DAY | ] NIGHT* 1 9 - 13 mm 4 0 4 TEMPERATURE !°C) 5 10 15 TEMpj 5/IOm* 5/IOm2 ■ SINGLE NIGHT HAUL I7/I0mz 29/IOm2 Figure 4 Mean day and night vertical distribution (.v n/100m') of (A) haddock Melanogrammus aeglefinus and (B) cod Gadus morhua size-classes of larvae collected by the lm- MOCNESS on well-mixed Site 81-111 during 27 May 1981. Estimated water column abundances (n/10m-) are included at bottom of plots. Temperature and salinity profiles are plotted as depth stratum means (•) and 95% confidence limits ( ). At the well-mixed shoal (34-54 m) Site 83-111 on 16-17 May 1983, three size-classes of haddock and cod larvae (6-8, 9-13, 14-19 mm) were distributed through Lough and Potter: Vertical distribution of Melanogrammus aeglefinus and Oadus morhua 285 A Haddock 6-Bmm 2 3 30 20 10 0 10 2030 6420246 :v 4/lOm2 3/l0m2 • SINGLE HAUL 53/l0m2 53/l0mZ 5/l0m2 9/l0m2 03/l0m2 08/K)m2 • SINGLE HAUL I/I0m2 I5/I0m2 SALINITY (p.su.) 3200 2 4 6 B 3300 2 3/IOm2 5/l0m2 TEMPERATURE ("CI 0 10 l» £30 40- SALINITY Figure 5 Mean day and night vertical distribution (:f n/100m3) of (A) haddock Melanogrammus aeglefinus and (B) cod Gadus morhua size-classes of larvae collected by the lm2 MOCNESS on well-mixed Site 83-111 during 16-17 May 1983. Estimated water column abundances (rc/10nr) are included at bottom of plots. Temperature and salinity profiles are plotted as depth stratum means (•) and 95f;r confidence limits ( ). the water column (Fig. 5). The single night haul and low densities preclude firm conclusions. Stratified water On 21 May 1981, when the water column at Site 81-11 ( 75-83 m) was strongly stratified, a single 1 rrr MOCNESS night tow found haddock and cod larvae 2-13 mm to be confined almost exclusively to the upper 20 m of the water column, and maximum densities usually were within the strong thermocline gradient at 10-20 m depth (Fig. 6). No recently-hatched cod larvae 2-5 mm were caught in this tow. The sur- face temperature approached 10°C with a strong ther- mal gradient at 11-21 m, where the temperature de- creased to 5.9°C. An intense storm on 21-24 May 1981, with winds up to 18-21 m/s (35-40 kn) and sea wave heights of 3-5 m ( 10-15 ft), destroyed the strong ther- mocline shown in Fig. 6. Following the storm, the wa- ter column was well mixed and the larvae were dis- A Haddock 2 -5mm 6 ■ 8 mm 9- 13mm 0 2 4 6 0 50 100 150 0 50 100 150 JMj-J 54ol ZJ B.Cod 14 - 19 mm SALINITY (psuJ 3200 2 4 6 8 3300 2 TEMPERATURE TO o° z i 6 10 20 . L -J 30 Q50 NIGHT 60 70 K Figure 6 Night vertical distribution ix /i/100m:,i of (A) haddock Melanogrammus aeglefinus and (B) cod Gadus morhua size-classes of larvae collected by a single lm- MOCNESS tow on well- stratified Site 81-11 before the storm on 21 May 1981. Estimated water column abundances (fi/lOm-l are included at bottom of plots. Tem- perature and salinity profiles are plotted as depth stratum mean l#) and 959r confidence limits ( 1. tributed broadly (Fig. 7), but about an order of magni- tude more abundant relative to the collections made on 21 May. Differences in mean day and night vertical distributions of larvae during the well-mixed period of 24-26 May 1981 indicate a diel shift in distribution patterns of the larger larvae (Fig. 7), with population centers located deeper in the water column by day (40-60 m) and shoaler by night ( 10-40 m). This diel shift in vertical distribution is supported by the ANOVA (Table 2B), where both haddock and cod have signifi- cant (p<0.05) depth x time interaction effects. By 28 May 1981, near-surface insolation warming of the up- per 20 m occurred, and larvae reaggregated above 20 m near a weak thermocline (Fig. 8). Mean depths of larvae captured in tows before and after the storm are plotted in Fig. 9 along with depth bounds of the thermocline region. The weak thermo- cline deepened after 25 May 1981. Mean depths of larval abundance remained below the weak thermo- cline until 28 May Tows made on 26 May showed that the larger larvae were mainly located deeper in the 286 Fishery Bulletin 91(2). 1993 2 - 5 mm 6 ■ 8 mm 9 - 13 mm 75 50 25 0 25 50 75 150 lOO 50 0 50 100 1 50 40 30 20 iQ 0 10 ffl J30 £40 -MT | ' I NIGHT 143/IOm2 250/IOm2 473/IOm2 483/IOm2 9/HW 66/IOm2 2 ■ 5 mm 6 8 mm 9 13 mm .,6420246 30 20 IP 0 10 20 30 3 2 I 0 12 3 7/IOm2 16/IOm2 35/IOm2 58/IOm* SAUNITt(psu) 3200 2 4 6 8 3300 2 /10m2 7/IOm2 TEMPERATURES) 20- l30' ^ 40 - TEMPI Q. ] ° 50- 60- Figure 7 Mean day and night vertical distribution tx 'i/lOOm'l of (A) haddock Melanogrammus aeglefinus and (B) cod Gadus morhua size-classes of larvae collected by the 1 nr MOCNESS on well-mixed Site 81-11 following the storm on 24-26 May 1981. Estimated water column abundances (»i/10m2) are in- cluded at bottom of plots. Temperature and salinity profiles are plotted as depth stratum means (•) and 95% confidence limits ( 1. water column than smaller larvae. However, when the water column was well-stratified on 21 May and weakly-stratified on 28 May, in both day and night all sizes of larvae were located in or above the thermo- cline. The single day and night tows were not suffi- cient to base firm conclusions on diel differences. In mid-May 1983, two sites were sampled in weakly- stratified waters across the southern margin of Georges Bank. At Site 83-1 (Fig. 10), mean day and night verti- cal distributions show haddock and cod larvae distrib- uted broadly through the water column on 13-14 May, but each of the three size-classes had a different depth 2 -5 mm 6 8 mm 9- 13 mm 1 0 5 10 5 0 60 40 20 3 30 20 10 0 :h_ :L. 2C 1 K . - 40 J SO L 60 . " ». B Cod SALINITY (psu) 3200 2 4 6 8 3300 2 TEMPERATURECC) 5 10 KH 10 20 J30' t° = 50 60 70 80- a 2-5mm I6/I0m2 6 6-Smn 2 0/I0m2 c 9-l3mm 04/IOm2 Figure 8 Day vertical distribution (.? 'i/lOOm') of size-classes of (A) haddock Melano- grammus aeglefinus and (B) cod Gadus morhus larvae collected by a single 1 nr MOCNESS tow on weakly-stratified Site 81-11 following the storm on 28 May 1981. Estimated water column abundances (n/ 10m-'l are included at bottom of plots. Temperature and salinity profiles are plotted as depth stratum means (•) and 95% confidence limits I ). distribution. Recently-hatched larvae (2-5 mm) were most abundant at 10-30 m both day and night. The 6- 8 mm size-classes of both species were more broadly distributed and most abundant at 10-50 m depths, with haddock tending to be deeper in the water column by day than cod. The 9-13 mm size-class for both had- dock and cod had a pronounced diel shift in day-night distribution patterns: most larvae were in the lower half of the water column by day but in the upper half at night. The maximum density for both species was at 10-20 m. There were significant (p<0.05) interac- tions for each combination of depth, time, and size (ANOVA; Table 2C). A time-series of larval catches in relation to the ther- mocline region at Site 83-1 is shown in Fig. 11. Al- though the mean depth of the thermocline centered at -20 m, its bounds varied between 10 and 40 m, per- haps due to the tides. While the population centers of cod and haddock larvae tended to be associated with Lough and Potter Vertical distribution of Melanogrammus aeglefinus and Gadus morhua 287 21 MAY 1981 24 MAY 1981 25 MAY 1981 26 MAY 1981 28 MAY 1981 0 10 20 2400 TIME(h) 1800 2400 0600 1200 1800 2400 0600 1200 1800 „96°C. - * - 60° - 1 1 1 1 1 -6 9° -7.0° -69° 1 S73«' 1 '-179*1 1 J70° Jg9° HADDOCK ' -17 8° ' ys»" J7i° hfi * -168° 68°- * o X 84° 72° E 30 " * o -69° O * * X 8 X * * ' ° 50 - * 0 o X x : - 60 70 0 10 20 NIGHT . DAY -6 5° NIGHT DAY NIGHT OAT _ DAY !n)GHT_ ° 1 36D/pfi trt - -69° -70° -6 9° ]" COD 1 07 jot 8 ]., ] 04 ~ OIK 07 - 130 £40 °50 60 70 o -.9- * -6 5° 2 o * « o X x 1 ■ s * o x x 1 , . 1 ■ sizi :ias' 1 2-5mm # 6-8 o 9-13 » , Figure 9 Time-series of abundance weighted-mean depths of larval haddock Melanogrammus aeglefinus and cod Gadus morhua size-classes on Site 81-11 before and after the storm on 21-28 May 1981. Brackets at tow time record upper and lower depth bounds of the thermocline region. In the haddock time-series, boundary temperatures are recorded beside the brackets; for cod, the rate of temperature change (8T/8Z) over the thermocline region is specified. Where no thermocline was evident, temperatures at selected depths are inserted. the thermocline region, they were not confined exclu- sively to it, as in May 1981 when the water column was strongly stratified. Larger larvae of both species tended to be found at greater depths during the day, especially the 9-13 mm larvae. At Site 83-1, >64% of the 9-13 mm size-class for both haddock and cod was found below the thermocline by day, but <28% by night (Table 3A). A significantly (p<0.05) greater percentage of the 9-13 mm size-class was found within and above the thermocline at night. For the 2-5 mm size-classes, haddock and cod larvae were found mostly within or above the thermocline; <15% were located below the thermocline, and there was no significant day-night difference in the profile percentages. For the 6-8 mm size-classes of haddock and cod, the highest percentage of larvae were located within the thermocline. Of the haddock larvae, 40% were found below the thermocline by day, but only 18% at night (p<0.005). There was no significant change in the day-night percentages for 6-8 mm cod. On 15-16 May 1983, at Site 83-11 (88-93 m) near the shelf/slopewater front, thermocline depth was some- what deeper with a smaller gradient than at Site 83-1. Most haddock and cod larvae were confined to the up- per 40 m, the lower bound of the thermocline (Fig. 12). A small percentage of larger larvae (9-13 mm) were located below the thermocline by day ( 13% for had- dock and 20% for cod) but moved nearer the surface by night (Table 3B). All size-classes of larvae were broadly distributed within the upper part of the water column. There were significant (ANOVA; p<0.001) effects of depth, time, and size for haddock; only depth x size interaction was significant (p<0.01) for cod (Table 2C). Mean depths of haddock and cod were greater by day for the larger larvae, especially for the 9-13 mm size- classes (Fig. 13), and in all but one night tow the mean depths were located within the confines of the thermo- cline region. Vertical distribution of pelagic juveniles In June 1984, pelagic juvenile haddock and cod were collected at a well-mixed shoal ( 40-49 m) Site 84-1 (Fig. 14). Cod were more abundant than haddock at Site 83-1, and the larger fish (14-29 mm) were caught more abundantly in night tows than day. The three size-classes of haddock were similarly distributed throughout the water column. Few were caught in the upper 10 m; abundance usually peaked near 10-20 m both day and night. Pelagic cod also were caught throughout the water column during the day, with abundance peaks near 10-20 m for the 14-19 mm and 20-29 mm size-classes. They were more abundant in 288 Fishery Bulletin 9 1 (2), 1993 - 5 mm 0 25 50 150 100 50 0 50 100 150 43/lOm2 93/IOm* 46B/I0m2 354/IOm2 I89/I0m2 123/IOm2 75/IOm2 74/lOm2 47/IOm2 37710m2 SALINlTY(psu) 32 00 2 4 6 6 3300 2 I2/I0m2 I6/I0m2 TEMPERATURE CO n 5 10 15 1 \ K> rtMpJ ;SALINITT 20 1 '• v L / . i 40 1, 1 60 • : 70 m ■ : Figure 1 0 Mean day and night vertical distribution Li «/100m') of (A) haddock Melanogrammus aeglefinus and (B) cod Gadus morhua size-classes of larvae collected by the 1 m- MOCNESS on weakly-stratified Site 83-1 during 13-14 May 1983. Estimated water column abundances (rc/lOnr) are included at bottom of plots. Temperature and salinity profiles are plotted as depth stratum means (•) and 95r7r confidence limits ( 1. the upper half of the water column at night, increas- ing somewhat towards the surface. There were no sig- nificant interaction effects in the ANOVA for either cod or haddock (Table 4). On 18-2 2 June 1986 at a well-mixed deeper (65- 78 m) Site 86-1, cod juveniles 14-49 mm were caught (Fig. 15). Mean day and night vertical distributions show a clear pattern of larger size-classes of cod located progressively deeper in the water column, with the 40-49 mm cod located almost entirely be- tween 50 m and the bottom. Night catches were greater than day catches for all size-classes of cod. Night mean abundance was significantly greater (ANOVA; p<0.01) than day mean abundance of the 20-29 mm and 30-39 mm size-classes (Table 4). There was no detectable evidence of diel vertical migrations. By July, relatively few pelagic haddock and cod ju- veniles were caught. Highest densities of juvenile gadids were located on eastern Georges Bank (Lough et al. 1989). Three sites were sampled in mid-July 1985 on eastern Georges Bank: Site 85-1 ( 82-87 m), stratified; Site 85-11 (64-72 m), weakly stratified; Site 85-111 (78-83m), stratified (Fig. 16). The deeper Sites 85-1 & 85-111 both had a gradual thermocline at 15- 30 m depths. Only a few pelagic cod juveniles were caught. Pelagic cod ranging 30-59 mm were caught mostly at night, although a day tow was not made at Site III. Relatively few fish were caught in the upper 30 m, and there was a tendency for the highest densi- ties to be located in the bottom third of the water column. Cod were caught in day tows at Site 85-1 in the bottom 60-80 m stratum. Diel migration of recently-settled juveniles In 17 bottom trawls (with rollers) at Site 86-1, 25- 26 June 1986, demersal cod juveniles with a mean length of 4-5 cm were collected (Table 5). Few were caught in day trawls, but at least an order of mag- nitude more cod (x 103/30 min trawl) were caught in night trawls. During 17-18 July 1985 at Site 85- IV, nine bottom trawls (with rubber disk sweep chain) were made (Table 6; Fig. 17). Both haddock and cod appeared to be caught more abundantly in day trawls than at twilight. Only one haddock ju- venile was caught in the single night trawl. The mean length of haddock was 7.1cm (range 4-10 cm), and 5.5 cm for cod (range 2-12 cm). Mean lengths of both species were the same in day and twilight trawls. The high daylight catches of cod juveniles in this series are in contrast to the high night catches of cod of the same average size, 4-5 cm, in June 1986 using a bottom trawl with rollers. Gear difference (rubber disks vs. rollers) may have been the cause of the different day-night catches, but conclusions cannot be made because in July 1985 only one night trawl was made. During 16-17 August 1985, 12 bottom trawls (with rubber disk sweep chain) on southeastern Georges Bank (63-71 m), Site 85- V, caught relatively high num- bers (210-257/30 min trawl I of 9-19 cm haddock juveniles (Table 7; Fig. 18). The mean number and lengths of fish per trawl were not significantly differ- ent (p>0.05) by paired /-tests for the night, day, and twilight catches. Lough and Potter Vertical distribution of Melanogrammus aeglefinus and Oadus morhua 289 13 MAY 1983 14 MAY 1983 TIME (hi 0000 0600 1200 Oi 1 . . , , , _ isoo 2400 0600 1200 1800 10 20 HADDOCK cj -|8l° J .* 8i° • * 8 4° ' 84° 8 5° 1 * -i8 4°~ 30 -160" # 1 J64° *! 63° 9 1 O o 40 ' o J65° \ ° JS5° 6 3° X -Ui° 50 X 1 X - 60 . 70 _ NIGHT I DAY NIGHT DAY - 10 COD 20 30 40 -S 0 X i d .1 .]. ; 10 S 13 o i0 • i ', ' X 1 ' - 60 70 II SIZE CLASS 2-5mm * 6-8 O 9-13 x no 1 1 ■ 1 , i I , 1 1 1 i > 1 1 ' Figure 1 1 Time-series of weighted-mean depths of (A) larval haddock Melanogrammus aeglefinus and (B) cod Gadus morhua size-classes on weakly-stratified Site 83-1 during the period 13-14 May 1983. In the haddock time-series, boundary tem- peratures are recorded beside the brackets; for cod, the rate of temperature change (8T/8ZI over the thermocline region is specified. cially after reaching 9-13 mm. Day abundances of haddock are greater than night for the 6-8 mm and 9- 13 mm size-classes, which may indi- cate a different avoidance behavior at this size. The 10 m2 MOCNESS night-day abundance ratios (zz/lOOOm2) are shown in Table 9 for size-classes of cod and haddock from study sites in June 1984, 1986, and July 1985. A greater size-range of cod is available than for haddock. For cod, the small- est (9-3 mm) and the largest (50- 59 mm) size-classes are not fully vul- nerable to the net, due to extrusion of the smallest larvae through the net mesh and distribution of the larg- est juveniles below the sampling depth. The total night-day ratios for cod are -2-3 (range 1.49-3.05) for size-classes 14-19 mm through 40- 49 mm. Haddock juveniles have night-day ratios of -3-4 for the 14- 19 mm and 20-29 mm size-classes. Sampler retention and avoidance The 1 m2 MOCNESS sampled effectively on a hori- zontal scale of 0.2-0.3 km; the 10 m2 MOCNESS, 1- 2 km; and the bottom trawl, 3-4 km. Minimum length of fish fully retained and not extruded by the vari- ous nets was estimated using a model based on the stretch mesh size and a body height-to-length re- gression (Potter et al. 1990). These estimates indi- cated that almost all haddock and cod larvae >~4mmSL are retained by the 1 m2 MOCNESS 0.333 mm mesh nets. For the 10 m2 MOCNESS, 3.0 mm mesh nets, haddock <17mm and cod <15- 16 mm probably are not fully retained. Minimum size retention of juvenile fish collected by the research bottom trawl is 31-37 mm. The lm- MOCNESS night-day abundance ratios (n/10m2) are shown in Table 8, page 307 for size- classes of cod and haddock collected in April and May 1981 and May 1983. The total night-day ratios for cod are -1.2 for the 2-5 mm and 6-8 mm size-classes, near 2.0 for the 9-13 mm size-class, and to -4.0-6.0 for the 14-19 mm and 20-29 mm size-classes (Fig. 19). An exponential curve fit appears to adequately de- scribe the increase of night-to-day catch ratios with increasing size of fish. Cod appear to visually avoid the 1 m2 MOCNESS net by day at all lengths, espe- Discussion The difficulty of detecting vertical migration in plank- tonic organisms, given the high sampling variability and limited resources for replicate time-series of ob- servations as well as gear limitations, has been dis- cussed by Pearre (1979), Shulenberger (1978), Denman & Piatt (1978), and others. We are forced to look for repeatable patterns, often in the absence of statistical proof. Because of these difficulties, diverse observa- tions have been reported in the literature. In this study, the vertical distribution patterns and diel migrations of haddock and cod larvae and juveniles have been documented in stratified and well-mixed waters of Georges Bank in greater detail than previously seen because of electronic MOCNESS samplers and the use of submersibles. Sampler avoidance Many investigators have reported night-day catch ra- tios >1. Net avoidance usually is greater for day than night tows because of visual detection of the net (Morse 1989). In another Georges Bank study using the 10 m2 MOCNESS, Potter et al. (1990) reported night catches of pelagic cod juveniles averaging 30 mm to be signifi- 290 Fishery Bulletin 9 1(2). 1993 55/IOmz 10/lOm2 236/lcW 168/IOm2 59/l0mz 67/lOmz 13/lOm2 7/lOmz I4/I0mz 13/lOm? SALINITYtp.su) 3200 Z A 6 6 3300 2 TEMPERATURE I*C1 0 5 10 c -J. "■ 10 1 'SALINITY 20 Ttiip/ • i 30 7 1 1 - ° 50 - 60 - 70 - 80 90 - Figure 1 2 Mean day and night vertical distribution ix n/100m') of (A) haddock Melanogrammus aeglefinus and (B) cod Gadus morhua size-classes of larvae collected by the 1 m- MOCNESS on weakly-stratified Site 83-11 during 15-16 May 1983. Estimated water column abundances In/lOnr) are included at bottom of plots. Temperature and salinity profiles are plot- ted as depth stratum means (•) and 95% confidence limits cantly greater than day catches. Their average night- day catch ratio for 16-45 mm cod was 3.0 (range 1.6- 4.7), which is similar to the ratios estimated in this study of cod and haddock pelagic juveniles. In an extensive study of the northeast U.S. conti- nental shelf ichthyoplankton (an 8yr time-series of data using a 61cm diameter bongo net, standard MARMAP double-oblique tows), Morse (1989) found 15 MAY 1983 16 MAY 1983 TIME(h) 0000 0600 1200 1800 2400 10 80°C -|6 3° -,84° 1 8 5° 5 * * V * 20 O o X 0 -i K- 30 - - 6 3° X 5 40 S 50 o 6 2° " 6J *■ 60 HADDOCK 70 - 80 . NIGHT DAY NIGHT 0 10 -| * -i 20 8 X OTVm 8 0 07 OB 8 10 8 - 30 - - E * " .09 " 40 - - m 50 _ - o 60 . a . 70 SIZE CLASS COD 2-5mm # 6-8 o 80 3, 9-13 « i ' ■ 90 1 , 1 i Figure 13 Time-series of abundance weighted-mean depths of larval haddock Melanogrammus aeglefinus and cod Gadus morhua size-classes on weakly-stratified Site 83-11 during the period 15-16 May 1983. In the had- dock time-series, boundary temperatures are re- corded beside the brackets; for cod, the rate of temperature change (dt/dz) over the thermocline region is specified. larval haddock to be the exception among most taxa in that day catches exceeded night catches for all lengths of larvae in the range 4-15 mm. Day catches were some- what greater than night catches for 4-12 mm cod, but for the larger fish ( 13-20 mm) night catches were equal to or greater than day catches. Haddock larvae, as well as cod, may remain still during the day to evade predator detection by visual or mechano-reception (Zaret 1980). When both had- dock and cod reach a larger size and their sensory systems become more developed, they exhibit a startle response and evade the attacking predator (in this case a net) (Blaxter & Fuiman 1990). Flexion and fin ray development is complete for both species at 15-20 mm (Auditore et al. 1993) and they would be more capable of a powerful darting speed. Perhaps at night without visual stimuli, only the larger juveniles exhibit a startle response to the net-mouth pressure stimulus. Light- aided avoidance is expected to decrease with water depth as a function of light attenuation and the fishes' visual threshold (Blaxter & Fuiman 1990). Pressure Lough and Potter Vertical distribution of Melanogrammus aeglefinus and Gadus morhua 291 A Haddock 9 -13 mm " 19 mm 6 .12 0 2 4 6 9 6 3 0 3 6 9 9/l,OOOm2 4/l,O0Om2 16/I.OOOm2 20/i,0O0m2 SALINiTV(iis.u J 3200 2-4 6 6 3300 2 TEMPtRATURECC) 0 5 10 15 I ' I t ' ' ' I 09/I.OOOm2 4/l,000m2 B Cod 9.13mm 14- 19mm „3 2 I 0 12 3 30 20 10 0 10 20 30 2 5/I.OOOm2 03/l,OOOm2 29/l.OOOm2 54/l£00m2 75 50250 25 50 75 3210123 45/I.OOOm2 l70/t,OOOm2 05/l,000m2 5/I.OOOm2 Figure 1 4 Mean day and night vertical distribution (.r /i/10,000m3) of (A) haddock Melano- grammus aeglefinus and (B) cod Gadus morhua size-classes of pelagic juveniles col- lected by the 10 m1' MOCNESS on well- mixed Site 84-1 during 17-19 June 1984. Estimated water column abundances l/i/1000m-') are included at bottom of plots. Temperature and salinity profiles are plot- ted as depth stratum means (•) and 95% confidence limits ( ). or sound avoidance are expected to be the same over depth, however. Therefore, for the three smallest size-classes of haddock and cod <13mm, light-aided avoidance of the lm2 MOCNESS sampler may be rela- tively small (< factor of 2) as a con- tributing factor in observed day-night vertical distribution patterns. How- ever, in fish >14 mm, night-day catch ratios are expected to increase to >3 because of light-aided avoidance response. Haddock larvae appear to respond differently to light than cod in their avoidance behavior. A correction factor for cod was not applied from the wa- ter-column total abundance ratios. The night-day catch ratios are not suitable for correcting depth strata densities, since net avoid- ance may be a function of light level which decreases with depth. That is, avoidance may be depth-dependent. We do not have the data suitable for making an appropriate depth correction. Applying a constant correction factor to all depth levels of a given size-class would not change the vertical distribution pattern. We need to know the fishes' reaction distance to the net at the different depth light levels. During the fishes' transition from a pelagic to demersal life at -30-40 mm, they are not fully vulnerable to either pelagic or demer- sal sampling gear and this contributes to the sampling variability. Scott (1984) studied diel variations in juvenile haddock catch rates from a 24 h bottom-trawl experiment near Sable Island on the Scotian Shelf in July-August. Haddock >6 cm had a marked vertical migra- tion pattern, moving off the bottom at night and returning by day- light. Based on changes in mean lengths, larger fish moved off the bottom in greater proportions than smaller fish. Colton (1965) exam- ined data from bottom-trawl surveys conducted on Georges Bank and found catch-per-tow of 0-group and 1+yr haddock to be mark- edly higher during the night than day. He believed the night-day variability was due to escapement from the net during the day and not to avoidance. Lough et al. (1989) conducted submersible studies on eastern Georges Bank in July 1987 and August 1986 and found recently-settled juvenile gadids (mostly cod) 4-12 cm in length close to the bottom (<0.5m) during the day and a portion of the popula- tion rising 1-5 m off the bottom at sunset, drifting with the current. A comparison of standardized research bottom-trawl (with rollers) catches and submersible transect counts showed the extent of undersampling by the bottom trawl. For 4 cm modal-size cod (1987), daylight trawl abundance estimates were at least an order of magni- tude lower than the submersible transects, but not significantly dif- ferent than the night estimates. For demersally-oriented larger fish (7 cm modal length), few fish were caught by daylight trawls, and night trawl catches were still at least an order of magnitude lower than night submersible counts. The larger fish appeared to stay closer to the bottom both day and night, and consequently they are less vulnerable to the bottom trawl with roller gear which may pass over them. Some preliminary results by Cohen et al. (1985) showed that more cod juveniles of 4-5 cm were caught using the bottom trawl with the smaller rubber disc-covered chain sweep than with the larger roller gear. Thus, the bottom trawl catches of recently- settled cod are known to have a significant sampling bias with more cod collected in night trawls than day, and especially for the larger fish escaping the roller gear. Vertical distribution patterns The initial depth distribution of recently-hatched larvae is dependent upon the late-stage egg distribution. Some limited egg profile data 292 Fishery Bulletin 9 1 (2). 1993 14 19 mm 20 - 29 mm 30 • 39 mm 40 49 mm „3 2 I 0 I 2 3 60 40 20 0 20 40 60 60 40 20 0 20 40 60 6 4 2 0 2 4 6 10 - 20 ! | 1 K i -4. OAT j NIGHT " 5C I 60 70 IL J SALlNITrlpsu I 3200 2 4 6 8 3300 2 TEMPERATURE (*C) 5 10 lb TEMPI SAL 2 7/l,000m2 6 4 /1, 000m2 79/l,O0Om2 I34/I.000m2 UO/I.OOOm2 l54/l,000m2 6.l/l,000mz 8 5/l,0O0m2 Figure 15 Mean day and night vertical distribution (x /z/10,000m:j) of cod Gadus morhua size-classes of pelagic juveniles collected by the 10 nr MOCNESS on well-mixed Site 86-1 during 18-22 June 1986. Water column abundances (/i/1000nr) are included at bottom of figure. Temperature and salinity profiles are plotted as depth stratum means (•) and 95% confidence limits ( ). A. Site 85- 1 12-14 July 1985 salinity 60 L , ™ I6/I,000m* 'T 1 \ TO 2 0/1, 000 m* 4 2 /1,000m2 I \ _ SALINITY (psul C. Site 85 -III 16- 17 July 1985 3200 2 4 6 B 3300 2 30-39mm 40 - 49 mm 50 - 59 mm TEMPERATURE (°C] 0 I 2 3 0123 0123 0 5 10 15 J : SALINITt'' y/ . ' j 20 30 X NIGHT 40 TCWPERATUREJ . L L, / '■. f- 1 i 80 70 L 1 1 63/1,00 m* BO 52/1,000 9mm) tend to be found at greater depths during the day. Day-night shifts in their distri- bution patterns indicate that diel Lough and Potter: Vertical distribution of Melanogrammus aeglefinus and Gadus morhua 293 50 40 30 20 > l0 i 0 S40 30 20- 10 0 HADDOCK 17-18 JULY 1985 pP DAY n = 496 X TWILIGHT n=40 n 10 12 14 0 2 4 6 LENGTH I cm) 10 12 14 16 Figure 1 7 Day and twilight length-frequency distribution of haddock Melanogrammus aeglefinus and cod Gadus morhua juveniles collected by Yankee 36 bottom trawl on Site 85-IV during a 24 h study, 17-18 July 1985. No fish were caught in night trawls. 6 - R-oeeiTe0910" * / r=0 88 / 5 / • o / I4 / / ^F3 / Z / 2 ^■" 0 i ) 5 10 15 20 25 30 LENGTH (mm) Figure 19 Night-to-day catch ratios for cod, Gadus morhua, in tows with 1 m- M0CNESS during April and May 1981 and May 1983. An exponential curve is fitted to the ratios. 30 20 10 HADDOCK " 16-17 AUG 1985 rJll DAY 11 = 1,365 nn_ 20 10 Tr.n l NIGHT . n = 842 riru _ 20 10 0 rnnn n TWILIGHT, n = 514 iIUttT i , 6 8 10 12 14 16 18 20 22 LENGTH (cm) Figure 1 8 Day, night, and twilight length- frequency distribution of haddock Melanogrammus aeglefinus juve- niles collected by Yankee 36 bot- tom trawl on Site 85-V during a 24 h study, 16-17 August 1985. migrations are fairly well estab- lished when larvae reach 9- 13 mm length and may be initi- ated at a smaller size, 6-8 mm, depending on the physical struc- ture of the water column. This pattern of residing deeper by day and shoaler at night continues into the older pelagic juvenile and recently-settled juvenile period. When larvae reach a length of ~9 mm, the swim bladder may be function- ally important in the vertical migration of larvae in both species (Schwartz 1971, Ellertsen et al. 1980, Howell 1984). Haddock and cod have similar developmental patterns; however, there are some important differences (Auditore et al. 1993). At 8-9 mm, haddock have larger and more developed pectoral and pelvic fins than cod. The difference in development is pro- nounced after 20 mm in length; at 23 mm, the pectoral fins of cod are V-2 to 3A the size of haddock, while the pelvic fins of cod are l/s those of haddock. Haddock also possess more total fin rays in the caudal, dorsal, and anal fins than cod. The total number of caudal-fin rays usually determines the overall size of the fin, which is an important means of thrust in gadiform fishes (Cohen 1984). The higher number of caudal- and dorsal-fin rays in haddock, in combination with early paired-fin development, may provide greater ma- neuverability for haddock. The earlier fin development of haddock larvae compared with cod may allow them to more actively locate high concentra- tions of prey and stay within these patches. Miller et al. (1963) studied the vertical distribution of larval haddock at three sites on Georges Bank in May 1958. Haddock larvae were found throughout the water column, with maxi- mum concentrations centered at 20-30 m. At all three sites they found peri- odic changes in the depth distribution of larvae that were related to similar changes in depth of the thermocline. Smaller larvae, 4-8 mm, occurred below the thermocline at each site while >80% of the larger larvae, 9-19 mm, were found in the thermocline. They also found no evidence of diurnal change in the weighted mean depth distribution of larvae, 24 m during the day and 22 m by night. Their observations were in general agreement with those of Frank et al. (1989) for haddock larvae on the Southwestern Scotian Shelf. Frank et al. (1989), using a discrete depth sampler, BIONESS, found recently-hatched haddock larvae in May 1985 and 1986 were concentrated at mid-depth (~30m) in day and night, and their population centers remained at this level until 20 d post-hatching. The hydrographic conditions were char- acterized by a density gradient (sigma-<) that increased linearly with depth both years. Whereas Miller et al. (1963) and Frank et al. (1989) found had- 294 Fishery Bulletin 91(2). 1993 dock larvae (<9mml below the thermocline, we found both haddock and cod larvae of this size tended to reside mostly within or above the base of the thermo- cline in May on Georges Bank. Miller et al.'s (1963) ability to discriminate depth changes in larvae was limited because of the small, non-opening/closing sam- plers ( 14 cm mouth diameter) that were towed at fixed depth intervals (minimum 9-10 m apart). Ellertsen et al. (1984) reported on the vertical distribution of cod larvae (4-5 mm) at a strongly- stratified station off the Lofoten Islands, Norway, in May 1982. First-feeding cod larvae were concentrated at 10-20 m depth near the thermocline throughout a 24 h period, but dispersed with the slightest wind mix- ing. Only under extremely calm conditions were the larvae able to control their vertical distribution and show diurnal migration. Pelagic juveniles, however, are capable of making extensive migrations through the water column. In our Georges Bank studies, at a well-mixed shoal site in June 1984, pelagic cod juveniles appeared to migrate into the upper half of the water column at night and to remain in the lower water column during the day. Juvenile haddock were much less abundant at this site, and their center of abundance was located at mid- depth with no evidence of day-night migrations like those of cod. The few pelagic cod collected at two strati- fied sites on eastern Georges Bank in July 1985 were generally found in the lower third of the water col- umn, and few were caught in the thermocline at 15- 30 m depth. We have no data on pelagic juvenile had- dock at a stratified site for comparison. Perry & Neilson ( 1988) studied the vertical distribu- tion of cod and haddock pelagic juveniles in June 1985 on eastern Georges Bank and found a diel migration pattern for both species at a well-mixed shoal site (66 m). At a nearby stratified site in deeper water (80 m), Perry & Neilson (1988) found that juvenile had- dock generally remained above 40 m depth with their weighted-mean depth in the thermocline at 20-30 m. In contrast, cod juveniles generally remained deeper than 40 m, below the thermocline, yet still exhibited diel migrations. Colton ( 1965) investigated vertical distribution of ju- venile haddock using an Isaacs-Kidd midwater trawl in late summer of 1957-58. Over 75% of the pelagic haddock of 27-124 mm occurred between 10 and 40 m depth, with greatest abundance at 20 m at the depth of the thermocline. There was some diel variation in distribution: the weighted-mean abundance was lo- cated at 40 m during the day and 30 m at night. Results from our studies indicate that juvenile had- dock and cod change from a pelagic to a demersal existence at a body length of 4-10 cm, and once they reach a demersal stage they stay close to bottom by day and move off bottom into the water column at night. During the transition period the vertical ampli- tude of these night excursions decreases with size of fish. The data suggest that cod <6cm make more ex- tensive off-bottom migrations at night than haddock, with some individuals ranging up to the surface. Had- dock appear to assume a more complete demersal life at a smaller size (4-6 cm) than cod, which is consis- tent with older juvenile and adult behavior. Bailey (1975) found that 6.5-13 cm haddock migrated through the water column at night, whereas larger fish were more confined to the seabed in the northern North Sea during August. Beamish (1966) and Woodhead (1966) reported that adult haddock tend to remain associated with the seabed at night, whereas adult cod make ex- tensive vertical migrations. In contrast, Bailey (1975) concluded that cod juveniles remained on or close to the seabed day and night, because fish caught in midwater were never >7cm, while those on the seabed ranged up to 13 cm. Juvenile cod and haddock at -70-90 mm begin feed- ing on benthic prey such as polychaetes and crusta- ceans along with planktonic prey, which is consistent with changes in their mouth (Auditore et al. 1993). The time for juvenile cod and haddock to grow from 40 mm to 70 mm is -30 d, and from 40 to 90 mm is -48 d (Bolz & Lough 1988). Therefore, in terms of de- fining the functional transition period for an individual fish, we consider the transition from a wholly pelagic to demersal existence to take -1-1.5 mo on average. Other factors influencing vertical distributions Neilson & Perry (1990) did a recent review of the lit- erature on diel vertical migrations of marine fishes and concluded that the diversity of vertical migration patterns indicated a facultative process significantly influenced by multiple environmental factors. Prey dis- tributions probably play an important role in the ver- tical distributions of larvae. Based on Georges Bank data (Buckley & Lough 1987) and simulation models (Laurence 1985), haddock larvae are considerably more food-limited than cod. Stable stratified conditions in the spring, which result in the availability of high con- centrations of zooplankton prey, favor growth and sur- vival of haddock larvae. Lough (1984) and Buckley & Lough (1987) found that haddock and cod larvae <~13mm (mode at 7mm) were generally associated with the planktonic prey biomass that had a maxima in the thermocline during May 1981 and 1983 along the southern margin of Georges Bank. The stronger the water column stratification, the more closely lar- vae and prey coincided. At the well-mixed sites where larvae were broadly distributed, prey biomass was Lough and Potter Vertical distribution of Melanogrammus aeglefinus and Gadus morhua 295 lower and more uniformly distributed through the wa- ter column. Frank & McRuer ( 1989) also found the condition of haddock larvae off Southwestern Nova Scotia in May 1986 consistent with the well-mixed/stratified regime findings of Buckley & Lough (1987). Larvae were in good condition at the deeper stratified sites and in poor condition at the shallow, well-mixed sites. They also found evidence that the buoyancy of starved lar- vae (4— 7 mm) increased, which frequently resulted in a bimodal depth profile, with larvae in poorer condi- tion nearer the surface. Ellertsen et al.'s (1984) study in the Vesteralsfjord of Norway showed that under calm conditions larval cod appeared to concentrate in the surface layers during the night coinciding with the highest density of copepod nauplii, their preferred prey in the study. They also noted that turbulent mixing dispersed the larvae and prey evenly throughout the water column. Therefore, the vertical distribution of larvae and their prey needs to be evaluated in relation to larval buoyancy and water-column turbulence. Light plays an important role in the diel migrations of cod and haddock larvae, but it is not known whether larvae follow a specific isolume. There may be a criti- cal light intensity at which visual feeding ceases and vertical migration begins. Minimum light-intensity threshold for larval cod feeding was in the range 0.1- 0.4 lux, and maximum feeding incidence at 1.4 lux (Ellertsen et al. 1980). Light regime also can affect the swimming activity and buoyancy of yolksac cod larvae by changing their metabolism and utilization of yolk (Solberg&Tilseth 1987). In Perry & Neilson's (1988) study of pelagic juve- niles, the migration pattern of cod appeared to be di- rectly related to the light-dark cycle, whereas that of haddock was more complicated and may have been confounded by periodic changes in the rotary tidal cur- rent speed and/or the migratory movements of their preferred mysid prey Neomysis americana. Their ex- planations for the different depth distribution may be related to (1) cod's acclimation to colder waters and haddock's preference for warmer waters, and (2) a pos- sible mechanism for reducing interspecific competition, since their vertical distribution patterns were similar when prey biomass was high and different when prey biomass was low. In our opinion, the differences in vertical distribution between cod and haddock at the stratified site noted by Perry & Neilson (1988) may be attributed to species differences in feeding behavior. On Georges Bank, planktonic prey selected by both species during settlement at 40-60 mm include copep- ods, amphipods, euphausiids, and cumaceans (Bow- man 1981ab, Koeller et al. 1986, Mahon & Neilson 1987, Lough et al. 1989). Robb (1981) showed that pelagic 0-group gadids in the northern North Sea have a different feeding behavior reflective of the adult lifestyle, i.e., cod fed on larger active pelagic prey, while haddock concentrated on smaller slower-moving or sed- entary organisms. Both Robb's (1981) and Perry & Neilson's (1988) feeding results support the view that haddock feed continuously when prey is available. A complicating factor in understanding vertical dis- tribution patterns in haddock is their commensal as- sociation with the large coelenterate Cyanea sp. This association has been reported in the literature for fish up to a length of ~10cm (Colton & Temple 1961, Mansueti 1963, Rees 1966, Bailey 1975, Koeller et al. 1986). Mansueti (1963) stated that jellyfish serve as a source of food and shelter for young fish between the pelagic and demersal stages. The change from a sur- face to deeper-water habitat by the fish is believed by Mansueti (1963) to be a factor causing the breakup of the commensal relationship. Koeller et al. (1986) showed Cyanea sp. to have been most abundant at the surface, decreasing to near zero below 80 m off Nova Scotia in June 1983. Pelagic juvenile haddock, more so than cod, may require contact with some objects such as jellyfish. Whether jellyfish tentacles also serve as a source of food is not known. The substrate on which the juveniles settle, (i.e., an unsuitable habitat such as sand) may prolong the tran- sition to demersal life. Lough et al. (1989) hypoth- esized that a large gravel habitat on northeastern Georges Bank favors survival of recently-settled juve- nile fish because of predator avoidance through cam- ouflage and/or increased prey abundance. Pelagic ju- venile cod were widespread over eastern Georges Bank in June; and in mid-July when they became demersally oriented, they were present on bottom types ranging from sand to gravel. By contrast, in late July to early August, recently-settled juveniles were abundant only on the large gravel deposit on northeastern Georges Bank and were sparse or absent from the large sand and gravelly sand that cover most of eastern Georges Bank (Lough et al. 1989). To summarize the vertical distribution patterns ob- served, we developed a generalized day-night mean depth distribution of cod by size-class, from recently- hatched larvae through recently-settled juveniles (Fig. 20). Cod vertical migrations appear to respond primarily to a day-night cycle, whereas haddock mi- grations respond to other factors such as prey distri- bution. Depth centers of the smallest larvae (2-8 mm) appear at two levels-15 and 35 m-depend- ing on whether the water column is stratified or well- mixed (Fig. 20). In order for the water column to be considered sufficiently stratified for larvae to initially reside at 15 m, a change of >0.3 sigma-< units would have to occur from surface to the base of a pycnocline. Waters around the flank of Georges Bank are typically 296 Fishery Bulletin 91(2), 1993 o. a a Stratified water 1 8 ■ 1 l T f Night T 0 e 8 8 f Well-mixed water 8 Day t 8 8 ! i 2-5 6-8 9-13 14-19 20-29 30-39 40-49 50-59 60-69 7 24 39 56 73 66 97 106 115 Size Class (SL mm) Age (days post hatch) Figure 20 Generalized day-night depth distribution of cod Gadus morhua by size-class (mmSL) and age (days post-hatching) from hatched larva through recently-settled juvenile in a 70m water column. Mean population depth by day I ) and by night (•); vertical range between day and night circles is represented by double arrow. Initial depth centers for the first two size-classes (2-5. 6-8 mm I is largely determined by larval buoyancy and. therefore, whether the water column is well-mixed or stratified. well-mixed in winter, with a seasonal thermocline de- veloping by mid-May subject to perturbations. When larvae reach a length of 9-13 mm, they appear to make day-night migrations regardless of the water-column structure. There is some uncertainty in the day-night migration range of the pelagic juveniles; however, the 20 m range designated is within their swimming capa- bility at this size. Furthermore, we don't know pre- cisely the transition period from a day to night depth. It probably occurs over a 3-4 h period, but it could be shorter. For example, a 10 mm fish swimming at 1 body length/s could transverse 20 m in 1 h. The verti- cal distribution pattern presented is sufficiently gen- eral to be used as a starting point for both cod and haddock at any bottom depth, even though all situa- tions can be interpreted from the data. These studies highlight the need for investigations of behavior- related sampler avoidance and further time-series of observation on fish distributions in different environ- mental conditions. Citations Auditore, P. J., R. G. Lough, & E. A. Broughton 1993 A review of the comparative development of Atlan- tic cod and haddock based on an illustrated series of larvae and juveniles from Georges Bank. NAFO Sci. Counc. Stud. (In press). Azarovitz, T. R. 1981 A brief historical review of the Woods Hole lab- oratory trawl survey time series. In Doubleday, W.G., & D. Rivard (eds.). Bottom trawl surveys. Spec. Publ. Fish. Aquat. Sci. 58:62-67. Bailey, R. S 1975 Observations on diel behaviour patterns of North Sea gadoids in the pelagic phase. J. Mar. Biol. Assoc. U.K. 55:133-142. Beamish, F. W. H. 1966 Vertical migration by demersal fish in the North- west Atlantic. J. Fish. Res. Board Can. 23:109-139. Blaxter, J. H. S., & L. A. Fuiman 1990 The role of the sensory systems of herring larvae in evading predatory fishes. J. Mar. Biol. Assoc. U.K. 70:413-427. Bolz, G. R., & R. G. Lough 1983 Growth of larval Atlantic cod, Gadus morhua, and haddock, Melanogrammus aeglefinus, on Georges Bank, spring 1981. Fish. Bull., U.S. 81:827-836. 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. Bowman, R. E. 1981a Food and feeding of 0-group haddock in the Northwest Atlantic. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 178:322-323. 1981b Food of 10 species of northwest Atlantic juve- nile groundfish. Fish. Bull., U.S. 79:200-206. Lough and Potter. Vertical distribution of Melanogrammus aeglefmus and Oadus morhua 297 Buckley, L. J., & R. G. Lough 1987 Recent growth, biochemical composition, and prey field of larval haddock (Melanogrammus aeglefinus) and Atlantic cod (Gadus morhua) on Georges Bank. Can. J. Fish. Aquat. Sci. 44:14-25. Cohen, D. M. 1984 Gadidae: Development and relationships. In Moser, H.G., et al. (eds.), Ontogeny and systematics of fishes, p. 259-265. Spec. Pub. 1, Am. Soc. Ichthyol. Herpetol. Allen Press, Lawrence KS. Cohen, E. B., J. R. Green, D. C. Potter, & B. P. Hayden 1985 Some preliminary results of juvenile cod and had- dock studies on Georges Bank in 1984 and 1985. Int. Counc. Explor. Sea Comm. Meet. G74:23. Cohen, E. B., M. P. Sissenwine, & G. C. Laurence 1988 The "recruitment problem" for marine fish popula- tions with emphasis on Georges Bank. In Rothschild, B.J. (ed.), Towards a theory on biological-physical in- teractions in the world ocean, p. 373-392. Kluwer Acad. Publ., Dordrecht. Colton, J. B. Jr. 1965 The distribution and behavior of pelagic and early demersal stages of haddock in relation to sampling techniques. Int. Comm. Northwest Atl. Fish. Spec. Publ. 6:318-333. Colton, J. B. Jr., & R. F. Temple 1961 The enigma of Georges Bank spawning. Limnol. Oceanogr. 6:280-291. Denman, K., & T. Piatt 1978 Time series analysis in marine ecosystems. In Shugart, H.H. Jr. (ed.), Time series and ecological pro- cesses, p. 227-242. SIAM (Soc. Ind. Appl. Math) — SIMS (Siam Inst. Math. Soc.) Conf. Ser., Philadelphia. Dunn, O. J., & V. A. Clark 1974 Applied statistics: Analysis of variance and regression. John Wiley, NY. Ellertsen, B., P. Solemdal, T. Stromme, S. Tilseth, & T. Westgard 1980 Some biological aspects of cod larvae. Fiskeridir. Dir. Skr. Ser. Havunders. 17:29-47. Ellertsen, B., P. Fossum, P. Solemdal, S. Sundby, & S. Tilseth 1984 A case study of the distribution of cod larvae and availability of prey organisms in relation to physical processes in Lofoten. In Dahl, E., et al. (eds.). The propagation of cod Gadus morhua L., p. 453- 477. Flodevigen rapp. 1. 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. Flagg, C. N. 1987 Hydrographic structure and variability. In Back- us, R.H. (ed.), Georges Bank, p. 108-124. The MIT Press, Cambridge. Frank, K. T., & J. K. McRuer 1989 Nutritional status of field-collected haddock (Mel- anogrammus aeglefinus) larvae from Southwestern Nova Scotia: an assessment based on morphometric and vertical distribution data. Can. J. Fish. Aquat. Sci. 46(Suppl. 1): 125-133). Frank, K. T., F. H. Page, & J. K. McRuer 1989 Hydrographic effects on the vertical distribution of haddock (Melanogrammus aeglefinus) eggs and lar- vae on the Southwestern Scotian Shelf. Can. J. Fish. Aquat. Aci. 46(Suppl. l):82-92. Howell, B. R. 1984 The intensive rearing of juvenile cod, Gadus morhua L. la Dahl, E., et al. (eds.). The propagation of cod Gadus morhua L., p. 657-675. Flodevigen rapp. 1. Koeller, P. A., P. C. F. Hurley, P. Perley, & J. D. Neilson 1986 Juvenile fish surveys on the Scotian Shelf: impli- cations for year-class assessments. J. Cons. Int. Explor. Mer 43:59-76. Laurence, G. C. 1985 A report on the development of stochastic models of food limited growth and survival of cod and had- dock larvae. In Laurence, G.C., & R.G Lough (eds.), Growth and survival of larval fishes in relation to the trophodynamics of Georges Bank cod and haddock, p. 83-150. NOAA Tech. Memo. NMFS-F/NEC-36, Woods Hole MA. Laurence, G. C, & R. G. Lough 1985 Growth and survival of larval fishes in relation to the trophodynamics of Georges Bank cod and haddock. NOAA Tech. Memo. NMFS-F/NEC-36, Woods Hole MA. Lough, R. G. 1984 Larval fish trophodynamic studies on Georges Bank: sampling strategy and initial results. In Dahl, E., et al. (eds.), The propagation of cod Gadus morhua L., p. 395-434. Flodevigen rapp. 1. Lough, R. G., & G. R. Bolz 1989 The movement of cod and haddock larvae onto the shoals of Georges Bank. J. Fish. Biol. 35:7 1-79. Lough, R. G., & D. C. Potter 1983 Rapid shipboard identification and enumeration of zooplankton samples. J. Plankton Res. 5:775-782. Lough, R. G., P. C. Valentine, D. C. Potter, P. J. Auditore, G. R. Bolz, J.D. Neilson, & R. I. Perry 1989 Ecology and distribution of juvenile cod and had- dock in relation to sediment type and bottom cur- rents on eastern Georges Bank. Mar. Ecol. Prog. Ser. 56:1-12. Mahon, R., & J. D. Neilson 1987 Diet changes in Scotian Shelf haddock during pe- lagic and demersal phases of the first year of life. Mar. Ecol. Prog. Ser. 37:123-130. Mansueti, R. 1963 Symbiotic behaviour between small fishes and jelly fishes, with new data on that between the Stromateid Pepnlus alepidotus, and the Scyphomedusa, Chrys- aora quinquecirrtia. Copeia 1963:40-80. Miller, D., J. B. Colton Jr., & R. R. Marak 1963 A study of the vertical distribution of larval haddock. J. Cons. Int. Explor. Mer 28:37-49. 298 Fishery Bulletin 91(2). 1993 Morse, W. W. 1989 Catchability, growth, and mortality of larval fishes. Fish. Bull., U.S. 87:417-446. Murawski, S. A., A. M. Lange, M. P. Sissenwine, & R. K. Mayo 1983 Definition and analysis of multispecies otter- trawl fisheries off the northeast coast of the United States. J. Cons. Int. Explor. Mer 41:13-27. Neilson, J. D., & R. I. Perry 1990 Diel vertical migrations of marine fishes: an obli- gate or facultative process? Adv. Mar. Biol. 26:115- 168. O'Boyle, R. N., M. Sinclair, R. J. Conover, K. H. Mann, & A. C. Kohler 1984 Temporal and spatial distributions of ichthyo- plankton communities of the Scotian Shelf in relation to biological, hydrological, and physiographic features. Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 183:27-40. Overholtz, W. J., & A. V. Tyler 1985 Long-term responses of the demersal fish assem- blages of Georges Bank. Fish. Bull., U.S. 83:507- 520. Page, F. H., K. T. Frank, & K. R. Thompson 1989 Stage dependent vertical distribution of haddock (Melanogrammus aeglefinus) eggs in a stratified wa- ter column: observations and model. Can. J. Fish. Aquat. Sci. 46(Suppl. 11:55-67. Pearre, S. Jr. 1979 Problems of detection and interpretation of verti- cal migration. J. Plankton Res. 1:29-44. Perry, R. I., & J. D. Neilson 1988 Vertical distributions and trophic interactions of age-0 Atlantic cod and haddock in mixed and strati- fied waters of Georges Bank. Mar. Ecol. Prog. Ser. 49:199-214. Potter, D. C, R. G. Lough, R. I. Perry, & J. D. Neilson 1990 Comparison of the MOCNESS and IYGPT pelagic samplers for the capture of 0-group cod iGadus morhua) on Georges Bank. J. Cons. Int. Explor. Mer 46:121-128. Rees, W. J. 1966 Cyanea lamarckl Peron & Lesuer (Scyphozoa) and its association with young Gaclus merlangus L. (Pisces). Ann. Mag. Nat. Hist. 9:285-287. Robb, A. P. 1981 Observations on the food and diel feeding behaviour of pelagic 0-group gadoids in the northern North Sea. J. Fish. Biol. 18:183-194. Schwartz, A. 1971 Swimbladder development and function in the haddock, Melanogrammus aeglefinus L. Biol. Bull. (Woods Hole) 141:176-188. Scott, J. S. 1984 Short-term changes in distribution, size and avail- ability of juvenile haddock around Sable Island off Nova Scotia. J. Northwest Atl. Fish. Sci. 5:109-112. 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. Ser. 18:1-19. Shulenberger, E. 1978 Vertical distributions, diurnal migrations, and sam- pling problems of hyperiid amphipods in the North Pacific central gyre. Deep-Sea Res. 25:605-623. Sissenwine, M. P. 1984 Why do fish populations vary? /;; May, R. (ed.), Exploitation of marine communities, p. 59-94. Springer- Verlag, NY. Smith, W. G., & W. W. Morse 1985 Retention of larval haddock Melanogrammus aeglefinus in the Georges Bank region, a gyre- influenced spawning area. Mar. Ecol. Prog. Ser. 24:1- 13. Sokal, R. R., & F. J. Rohlf 1969 Biometry. W.H. Freeman, San Francisco, 776 p. Solberg, T. S., & S. Tilseth 1987 Variations in growth patterns among yolk-sac lar- vae of cod iGadus morhua L.) due to differences in rearing temperature and light regime. Sarsia 72:347-349. Walford, L. A. 1938 Effect of currents on distribution and survival of the eggs and larvae of haddock (Melanogrammus aeglefinus) on Georges Bank. Fish. Bull., U.S. 49:1-73. Wiebe, P. H., K. H. Burt, S. H. Boyd, & A. W. Morton 1976 A multiple opening/closing net and environmental sensing system for sampling zooplankton. J. Mar. Res. 34:313-326. Wiebe, P. H., S. H. Boyd, B. M. David, & J. L. Cox 1982 Avoidance of towed nets by the euphausiid Nematoscelis megalops. Fish. Buil, U.S. 80:75-91. Wiebe, P. H., A. W. Morton, A. M. Bradley, R. H. Backus, J. E. Craddock, V. Barber, T. J. Cowles, & G. R. Flierl 1985 New developments in the MOCNESS, an appara- tus for sampling zooplankton and micronekton. Mar. Biol. (Berl. 187:313-323. Woodhead, P. M. J. 1966 The behavior of fish in relation to light in the sea. Oceanogr. Mar. Biol. Annu. Rev. 4:337-403. Zaret, T. M. 1980 The effect of prey motion on planktivore choice. In Kerfoot, W.C. (ed. ), Evaluation and ecology of zoop- lankton communities. Limnol. Oceanogr. Spec. Symp. 3:594-603. Lough and Potter: Vertical distribution of Melanogrammus aeglefinus and Gadus morhua 299 Cruises, sampling sites, Table 1 and gear used for vertical distribution study on Georges Bar k, 1981- -86. Vessel cruise no. Sampling dates Site Bottom depth (m) No. depth Gear strata sampled Tows Night Day Twilight Albatross TV 81-03 27-29 April 1981 81-1 67-70 lnr MOCNESS 6 5 5 0 Albatross TV 81-05 21-27 May 1981 81-11 75-83 lnr MOCNESS 7-8 4 6 0 81-111 35-58 lnr MOCNESS 7-8 1 3 0 Albatross TV 83-03 13-17 May 1983 83-1 66-79 lm- MOCNESS 7 3 5 0 83-11 88-93 lm2 MOCNESS 8 2 3 0 83-111 34-54 lm- MOCNESS 5 1 3 0 Albatross TV 84-05 17-19 June 1984 84-1 40-49 10m- MOCNESS 4 3 3 0 Albatross TV 85-06 12-17 July 1985 85-1 80-87 10m2 MOCNESS 2-3 4 3 0 85-11 64-72 10m- MOCNESS 3 3 3 0 85-111 78-83 10m- MOCNESS 3 2 0 0 Delaware II 85-05 17-18 July 1985 85-IV 80-90 36 Yankee otter trawl with rubber disks Bottom 1 6 2 Delaware II 85-06 16-17 Aug. 1985 85-V 63-71 36 Yankee otter trawl with rubber disks Bottom 4 6 2 Albatross TV 86-03 18-22 June 1986 86-1 65-78 10m2 MOCNESS 4 5 6 0 25-26 June 1986 36 Yankee otter trawl with rollers Bottom 5 11 1 300 Fishery Bulletin 91(2), 1993 Table 2 Summary of GLM ANOVA for 1 arval haddock and cod abundance on Georges Bank. P-values are denoted by their level of signi "icance. A. 27-29 Apri Factor 1 1981 Site 81-1 Haddock" Codb Cod1 Codd Depth ID) 3.15* 8.84*** 6.50 8.52 Time(T) 2.07 16.59 39.37 50.92 Size(S) 24.68 31.10 29.09*** D x T 0.49 2.03 3.53*** 5.31*** D xS 0.28 1.26 0.93 T x S 3.66* 2.33* 0.14 D x T x S 0.68 mm size-class 0.71 0.76 ***P<0.001 a2-5 **P<0.01 b2-5 6- -8, 9-13 mm size-class *P<0.05 '2-5, 6- -8, 9-13, 14- 19, 20-29 mm size-class d6-8 9- -13, 14-19 mm size-class B. 24-26 May Factor 1981 Site 81-11 Haddock'1 Cod Depth ( D) 6.71 3.76 Time(T) 7.72 9.72 Size(S) 18.17 16.36 D ■ T 3.52** 2.90 k* D < S 0.78 0.33 T x S 1.74 0.69 D x T x S 6- 0.20 0.19 -8, 9-13 mm size-class ***P<0.01 "2-5 C. 13-26 May Factor 1983 Site 83-1 Site 83-11 Haddock11 Coda Haddock'1 Cod" Depth (D) 4.74 8.17 31.97*** 23.84 Time(T) 0.35 0.40 12.11*** 0.23 Size(S) 39.67 46.52 15.24*** 1.77 D < T 3.13** 0.99 0.86 0.09 D xS 1.25 1.13 1.59 2.29** T x S 5.33** 0.32 0.13 0.46 D x T x S 1.49 "2-5 6 2.31** -8, 9-13 mm 1.00 size-class 0.93 ***P<0.001 **P<0.01 Lough and Potter Vertical distribution of Melanogrammus aeglefmus and Gadus morhua 301 Table 3 Average percentage distribution of haddock and cod larvae in relation to the thermocline at Sites 83-1 and 83-11 on 1983. Angular transformation was used on the percentage data: angle = arcsin ^proportion. Values are presented as its standard error followed by a reconverted percentage mean in parentheses. Georges Bank May the angular mean 1 Angle iSE(%) Larval size-class 2-5 mm 6-8 nm 9-13 mm Day Night Day Night Day Night A. Site 83-1 13-14 May 1983 Haddock Above thermocline Within thermocline Below thermocline Cod Above thermocline Within thermocline Below thermocline 43.0112.2(46.6) 32.01 9.3(28.1) 21.7 1 5.5(13.6) 40.71 8.1(42.6) 39.4 1 4.5(40.4) 22.81 5.8(15.0) 22.616.4(14.8) 39.5 + 5.4(40.5) 39.416.1(40.3) 27.4 19.5(21.1) 49.716.2(58.1) 24.912.5(17.7) 9.914.61 2.9) 29.914.0(24.9) 56.914.8(70.2) 25.21 9.5(18.1) 53.41 4.2(64.4) 21.0 1 5.1(12.8) 41.0110.5(43.0) 44.1114.1(48.4) 18.7 1 6.4(10.3) 29.3116.9(23.9) 47.5112.5(54.4) 22.01 3.5(14.1) 23.616.3(16.0) 45.716.9(51.2) 32.615.5(29.1) 23.517.5(15.9) 45.614.3(51.1) 39.714.8(40.7) 6.21 2.8 ( 1.2) 35.813.1(34.3) 53.113.2(63.9) 24.91 2.1(17.8) 46.81 3.4(53.2) 32.01 5.2(28.1) B. Site 83-11 15-16 May 1983 Haddock Above thermocline Within thermocline Below thermocline Cod Above thermocline Within thermocline Below thermocline 31.31 2.5(27.4) 58.41 2.4(72.5) 3.51 1.7l 0.1) 48.0114.5(55.1) 42.1114.5(44.9) 1.81 0.0( 0.1) 23.717.3(16.1) 64.816.5(81.9) 6.61 1.7 ( 1.3) 34.119.1(31.4) 55.718.9(68.2) 3.11 1.3 ( 0.3) 12.11 5.2 ( 4.4) 60.512.8(75.7) 21.0 + 8.0(12.8) 31.9112.3(27.9) 56.0111.1(68.7) 9.41 2.4 ( 2.6) 35.61 6.7(33.9) 54.51 6.7(66.1) 8.11 0.0 ( 0.1) 33.61 4.7(30.5) 45.0110.1(50.0) 1.81 0.0 ( 0.1) 20.911.9(12.7) 69.211.9(87.4) 1.810.01 0.1) 27.213.1(20.9) 62.813.1(79.1) 1.810.01 0.1) 1.810.01 0.1) 63.615.5(80.3) 26.415.5(19.7) 18.11 16.3 ( 9.6) 72.9117.1(91.3) 1.81 0.0 1 0.1) Table 4 Summary of GLM ANOVA for juvenile haddock and cod abundance on Georges Bank. P-values are denoted by their level of significance. A. 17-29 J Factor line 1984 Site 84-1 Haddock'1 Haddockb Cod' Codd Depth (D) 1.12 0.59 0.24 0.04 Time(T) 0.48 0.53 0.12 0.06 Size (S) 11.60*** 11.46** 0.34 D x T 1.23 0.46 0.34 0.11 D x S 0.37 0.10 0.03 T x S 0.26 0.49 0.02 DxTxS 0.48 0.09 0.03 ***P<0.001 "9-13, 14-19, 20-29 mm size-class **P<0.01 14-19 mm size-class *P<0.05 c14-19, 20-29, 30-39 mm size-class 14-19, 20-29 mm size-class B. 18-22 J Factor Line 1986 Site 86-1 Cod" Depth (D) 41.78*** TimeiT) 9.98* Size (S) 4.76* D x T 0.65 D x S 1.09 T x S 1.91 D x T x S 0.30 ***P<0.001 '20-29, 30-39 mm size-class **P<0.01 *P<0.05 302 Fishery Bulletin 91 [2). 1993 Table 5 Number and size of juvenile cod (3-9cm) collected by time of day in bottom trawls. Site 86-1, 25- 26 June 1986, ALBATROSS IV 86-03. No. Total no. trawls .v n/trawl SD CV% x length (cm SD fish Night 5 102.8 52.7 51 4.2 0.45 513 Day 11 2.9 1.3 45 4.8 1.48 32 Twilight 1 6.0 — — 4.4 1.27 6 Table 6 Number and size of juvenile haddock (4-10 cm) and cod (2-12 cm) collected bv t ime of day in bottom trawls. Site 85-IV, 17- 18 July 1985, DELAWARE II 85-05. No. Total no. trawls x rc/trawl SD CV% x length (cm) SD fish Haddock Night 1 1.0 — — 9.0 — 1 Day 6 84.8 64.6 76 7.1 1.1 496 Twilight 2 20.0 9.9 50 7.1 1.0 40 Cod Night 1 — — — — — — Day 6 305.0 391.3 128 5.4 1.6 1,830 Twilight 2 256.6 244.0 95 5.5 1.1 513 Table 7 Number and size of juvenile haddock (3- 9 cm) collected bv time of dav in bott om trawls. Site 85-V, 16-17 August 1985, DELAWARE II 85-06. No. Total no. trawls x ri/trawl SD CV% x length (cm) SD fish Night 4 210.5 143.6 68 13.9 1.68 842 Day 6 227.5 114.4 50 13.5 1.76 1,365 Twilight 2 257.0 4.2 2 14.0 1.68 514 Lough and Potter: Vertical distribution of Melanogrammus aeglefinus and Oadus morhua 303 Table 8 Night/day abundance ratios (n/lOm-) for cod and haddock by length size-classes using the 1 m2 MOCNESS, April and May 1981 and May 1983 Size-class April 1981 May 1981 May 1983 May 1983 (mm) Site 81-1 Site 81-11 Site 83-1 Site 83-11 Total Ratio Cod 2-5 37/28 16/7 7.4/7.5 7/13 67.4/55.5 1.21 6-8 30/14 58/35 37/47 13/14 138/110 1.26 9-13 213/98 7/8 16/12 5/4 241/122 1.98 14-19 62/10 62/10 6.20 20-29 14/3 14/3 4.67 Haddock 2-5 108/66 250/143 93/43 10/55 461/307 1.50 6-8 0.6/1.8 483/473 354/468 168/236 1006/1179 0.85 9-13 66/119 123/189 67/59 256/367 0.70 Table 9 Night/day abundance ratios (rc/lOOOnr) for cod ar d haddock bv ength size-cl jsses using the 10 nr MOCNESS, June 1984 June 1986. and July 1985. Size-class June 1984 June 1986 July 1985 (mm) Site 84-1 Site 86-1 Site 85-1 Total Ratio Cod 9-13 0.3/2.5 0.3/2.5 0.12 14-19 54/29 6.4/2.7 60.4/31.7 1.91 20-29 170/45 134/79 304/124 2.45 30-39 5/0.5 154/110 9.5/3.7 168.5/113.5 1.49 40-49 8.5/6.12 14.4/1.4 22.9/7.5 3.05 50-59 1.4/3.1 1.4/3.1 0.45 Haddock 9-13 4/9 4/9 0.44 14-19 20/16 20/16 3.33 20-29 4/0.9 4/0.9 4.44 Abstract.— Line transects run from a manned submersible were used to estimate the current den- sity of yelloweye rockfish in two ar- eas of the eastern Gulf of Alaska. Yelloweye rockfish were seen in cobble, continuous rock, broken rock, and boulder habitats but were most abundant in broken rock and boul- der habitats. The presence of refuge spaces appears to be an important factor affecting occurrence of yelloweye rockfish. Boulder areas in deep water (>108m) were the most densely-populated habitat, with an estimated density of 9135 adult yelloweye/km2. Overall density by area and year ranged from 1954 to 2217 adult yelloweye rockfish/knr. Habitat-specific density estimates were less precise than general area estimates because of smaller sample sizes. For fisheries management, density estimates may be extrapo- lated to larger areas based on areal estimation of yelloweye habitat from available bathymetric data. Habitat-specific density of adult yelloweye rockfish Sebastes ruberrimus in the eastern Gulf of Alaska* Victoria M. O'Connell Alaska Department of Fish and Game. 304 Lake Street, Room 1 03 Sitka. Alaska 99835 David W. Carlile Alaska Department of Fish and Game. 802 3rd Street, Box 240020 Douglas, Alaska 99824-0020 Manuscript accepted 28 January 1993. Fishery- Bulletin, U.S. 91:304-309 1 19931. The yelloweye rockfish Sebastes ruberrimus is the target species of the commercial longline fishery for Demersal Shelf Rockfishes (DSR) in the eastern Gulf of Alaska (O'Connell & Fujioka 1991). Rockfishes {Sebastes spp.) are managed on an assemblage basis in the Gulf of Alaska under the advice of the North Pacific Fishery Management Council (NPFMC). De- mersal Shelf Rockfishes comprise eight species of bottom-dwelling rock- fishes inhabiting rocky areas of the continental shelf; yelloweye rockfish account for 96% of the landed catch of targeted DSR. Traditional stock-assessment meth- ods are difficult to apply to DSR be- cause of a combination of behavioral and physiological factors. The close association of DSR with rugged bot- tom precludes the use of bottom-trawl surveys used for assessing other groundfish in the Gulf of Alaska. Mark- recapture studies are also ineffective because rockfishes incur high embo- lism mortality when brought to the surface from depth (O'Connell 1991). Consequently, prior to our research, DSR was one of only two assemblages managed under the Gulf of Alaska Fisheries Management Plan for which no biomass estimates were available. It has been well documented that rockfish tend to be habitat-specific in their distribution (Love & Ebeling 1978, Larson 1980, Richards 1986, Matthews 1991, Love et al. 1991, Matthews & Richards 1991, Rosen- thal et al. 1982). Therefore, to esti- mate their abundance, we initiated a project designed to take advantage of the preference by DSR for rough, rocky habitat. Our approach was based on the assumption that DSR abundance increases with structural habitat complexity (i.e., increased to- pographic relief and more interstitial space in and between rocks). Our ob- jective was to estimate density of yelloweye rockfish in the Gulf of Alaska for selected habitat and depth categories. We hope to eventually de- velop a model predicting the relation- ship between DSR abundance and habitat complexity and to use this model to indirectly estimate the abundance of DSR. If successful, this approach would allow for expansion of abundance estimates to other areas in the eastern Gulf of Alaska without replicating costly surveys. Methods Using the submersible Delta, we made 20 dives and covered 47 transects dur- ing 17-25 August 1990 in two areas off southeastern Alaska (Fig. 1). "Contribution PP-058 of the Alaska Depart- ment of Fish and Game. Division of Commer- cial Fisheries, Juneau. 304 O'Connell and Carlile: Density of adult Sebastes ruberriums in eastern Gulf of Alaska 305 Fairweather Ground Study Area Figure 1 Study sites for submersible survey of yelloweye rockfish, eastern Gulf of Alaska, 1990 1991. Eleven dives were made on the Fairweather Ground and nine dives were made off Sitka Sound. Eighteen dives and 30 transects were completed off Sitka Sound during 27 May-3 June 1991. One transect was repli- cated at night to compare day and night effects. Transect locations were chosen systematically to include sites with a range of topographical relief. Bathymetric data from the National Ocean Service Hydrographic and Ma- rine Geophysical databases were used as an aid in site location. In a typical dive in 1990, three transects were run per dive with each transect lasting 30min. In 1991, transect durations were extended to 1 h with two transects run per dive. Deltas pilot attempted to main- tain a constant speed of 0.5 kn and to remain within lm of the bottom, terrain permitting. A predetermined compass heading was used to maintain position along a transect line. Periodic locational fixes of Delta along each transect were recorded using a TRAK-POINT navigation system aboard the support vessel, in com- bination with a Loran (19901 or a global positioning system (1991). The length of each transect (1, ) was measured as the sum of distances between locational fixes. The usual procedure for line- transect sampling entails count- ing objects on both sides of a transect line. Due to the configu- ration of the submersible, we only counted fish on the right side of each line. Horizontal vis- ibility was usually good, 5-15 m. All fish observed from the star- board port were individually counted and their perpendicular distance from the line recorded (Buckland 1985). The observer used the lower and middle star- board ports for viewing. An ex- ternally-mounted video camera was used to record both habitat and audio observations of species encountered and perpendicular distances to fish. Yelloweye rock- fish have distinct coloration dif- ferences between juveniles and adults, so observations of the two were recorded separately, and also on a back-up tape re- corder. A Pisces Box data-logger recorded depth of the submersible and its distance from the bottom, time of day, and temperature onto the videotape at Is intervals. These data were later downloaded onto a microcom- puter spreadsheet. In addition to the video system, we used a Photosea 35 mm camera with strobe to photo- graph habitat and fish. Two lamps were mounted exter- nally to provide lighting for the camera systems. Hand-held sonars were modified to obtain perpen- dicular distance recordings. Two sonar models were used: Manta and Scubapro. The end of each gun was fit with a tight (rubber) reservoir cinched to the sonar barrel. The reservoir was filled with water, and a sy- ringe used to remove air bubbles. The end of the reser- voir was kept damp by resting it on a wet sponge. A digital read-out of the distance from the submersible to its target was obtained by pressing the reservoir end of the gun against the port, aiming the gun, and pressing the trigger. To verify the accuracy of this method, we confirmed readings by positioning a scuba diver at intervals along a marked transect line. Six habitat categories were used: soft, gravel, cobble, continuous rock, boulder, and broken rock. Other de- scriptions of habitat were also recorded, including rock type (e.g., basalt), invertebrate cover, and vertical re- lief. To analyze depth differences, two depth intervals were defined: shallow <108m, and deep >108m. 306 Fishery Bulletin 91(2), 1993 Density estimates were obtained for combinations of species, area, depth, and habitat type using the den- sity estimator advanced by Burnham et al. (1980), ex- cept that length of transect (1,) as used in the denomi- nator was not doubled, since we were able to count fishes only on one side of each line transect. Data from all transects were combined, and yelloweye rockfish density was estimated as DyE nf(0) where n = total yelloweye rockfish adults observed (from all transects), and L = total line length (all transects combined) in meters. We used the indirect method of estimating variance in density (Burnham et al. 1980:54), and f(0) was esti- mated from detection functions based on the hazard- rate perpendicular distance model of Hayes & Buckland (1983). Results Density estimates No discernable difference was noted in the occurrence of yelloweye rockfish between day and night replicate transects, and all density estimates are calculated from daytime dives. Estimated densities of yelloweye rock- fish in 1990 and 1991 varied from 1954/knr to 2531/ km2 (Table 1). Estimated probability density functions (pdf) generally exhibited the "shoulder" (i.e., an inflec- tion and asymptote in the pdf for perpendicular dis- tances near 0) that Burnham et al. ( 1980) discuss as a desirable attribute of the pdf for estimation of f(0) (Fig. 2). Habitat effects Boulder fields were the most densely populated habitat type, followed by broken rock (Table 2). Although density was greater in boulder cover regions, Table 1 Density estimates, CV(Dl, and 95% CL for adult by area and year. yelloweye rockfish Density Year/Area no./km- n CV(D) 95% CL Lower Upper 1990 Fairweather 2217 1990 Sitka 1954 1991 Sitka 2065 221 236 352 19.5 25.3 14.9 1184 813 1347 3250 3054 3165 the occurrence of this habitat type was relatively in- frequent, accounting for only 16% of the total bottom surveyed. Surveyed habitat ranged from low-relief mud to high-relief pinnacles and cliff faces. The predominant rock type was volcanic, including folded lava flats and columnar basalt. Adult yelloweye rockfish were en- countered over all habitats surveyed, but occurred most frequently over broken rock and boulder fields, often observed resting in refuge spaces such as cracks, caves, or overhangs (Fig. 3). They were also observed hovering off cliff faces above cobble or pavement bot- toms. Generally, only one yelloweye rockfish was ob- served per refuge space, but often the hole was co- occupied by a tiger rockfish Sebastes nigrocinctus. Frequency-of-occurrence of yelloweye rockfish over soft, cobble, and continuous rock bottom was too low to provide acceptable line-transect density estimates. In both the Sitka Sound and Fairweather areas, there are infrequent abrupt pinnacles overlaid with mas- sive boulders and overhangs. Density of yelloweye and other species was extremely high in these habi- tats. In the Sitka area we surveyed two adjacent pinnacles: one pinnacle highly dissected with many boulders and overhangs, the other comparatively smooth. Occurrence of yelloweye rockfish on the com- plex habitat pinnacle was clearly higher than on the smoother pinnacle. Alhough gross vertical relief is similar between the two pinnacles, the number of refuge spaces differed and appeared to be important in the occurrence of yelloweye rockfish. Depth effects Depth effects on abundance were difficult to interpret because of the confounding influence of habitat. Depth- related density differences were evaluated by combin- ing the 1990 and 1991 Sitka data and examining two depth zones within the two preferred habitat types: broken rock and boulder. In both habitats, deep zones had higher abundances than shallow zones within the same habitat type. Boulder areas differed the most between shallow and deep zones: The high- est estimated density in boulder areas was in the deep zone where 9135 yelloweye rockfish/ km2 were counted, compared with 6122 yelloweye rockfish/km- in the shallow zone. The deep-zone broken-rock habitat had 2831 yelloweye rockfish/ km2- compared with 2748/km2 for the shallow zone. Juvenile yelloweye Abundance of juvenile yelloweye rockfish was too low to obtain reliable habitat-specific density es- O'Connell and Carlile. Density of adult Sebastes rubernums in eastern Gulf of Alaska 307 0 25 ■ „ f(0) = 0.228 probability density function 0.2 - n = 67 0.15 ■ 0.1 ■ n = 45 0.05 ■ n = 25 n = 24 n- n = 8 n = 7 0 1.67 3.34 5.01 6. 48 8.35 perpendicular distance (m) Figure 2 Representative probability density function Ipdfl and histogram of relative frequencies (n,/(A,n)) of yelloweye rockfish observed for 5 distance intervals over boulder habitat in the Sitka study area. 1990. For relative frequencies, n, = number of fish observed in interval i, A, = size of class interval (in this case. 1.67ml, and n = total number of yelloweye observed I in this case, 176). timates; hence they varied considerably (Table 3). Data on frequency-of-occurrence/meter traversed suggested preference by juvenile yelloweye rockfish for the shal- low-zone broken-rock habitat. Discussion In addition to confirming the strong association of yelloweye rockfish with rocky habitat noted by Richards (1986), we were able to estimate overall and habitat- specific densities for yelloweye rockfish. The only pre- viously published estimates of yelloweye rockfish den- sities are from the Strait of Georgia, British Columbia. Richards (1986) used strip transects conducted from the submersible Pisces VI to describe spatial distribu- tion patterns of rockfish. Although our results were Table 2 Density estimates for adult yelloweye rockfish by habitat category. area, and year. 95% CL Density Habitat Year/Area n no./km- CVlDl Lower Upper Boulder 1990 Fairweather 59 3660 36.60 792.2 6527.1 1990 Sitka 52 3712 53.70 0 8015.8 1991 Sitka 82 5026 30.99 1693.0 8358.9 Broken rock 1990 Fairweather 143 3312 24.50 1575.1 5048.8 1990 Sitka 175 2515 20.80 1393.2 3637.6 1991 Sitka 253 2405 40.81 304.84 4505.16 not directly comparable to hers because of differences in habitat categorization and because juve- nile yelloweye rockfish were in- cluded in her density estimates, our results were similar. For two depth categories, 21-80 m and 81-140 m, Richards (1986) estimated yelloweye rockfish densities of -10,000 and 14,000 yelloweye/km2 in her complex habitat category, including high- relief areas of cobble with large rock, broken rock, and boulder, a category that may encompass our broken-rock and boulder habitat types. In our study, estimated densities in boulder and broken- rock habitats varied from 2405 to 9135 yelloweye rockfish/km- depending on depth. Comparison of estimated densities from roughly similar habitat types in the two studies suggests that yelloweye densities in the Strait of Georgia, BC were greater than we estimated for the coast of southeast Alaska. Use of the submersible allowed collection of qualita- tive data for assessment of factors contributing to the distribution and abundance of fishes. Occurrence of refuge spaces may be one key to the presence of yelloweye rockfish, which were normally in areas where refuge spaces were available, even if the surrounding habitat was not the preferred habitat of boulder or broken rock. For example, we often encountered yelloweye rockfish under overhangs of large, solitary boulders in cobble flats. Continuous rock bottom was not particularly good habitat in terms of yelloweye density. Additional transects will be needed to increase the precision of habitat-specific estimates and narrow the associated confidence limits. Further refinement of habitat categories (e.g., subcat- egories of boulders based on av- erage size of boulders and/or in- terstitial spaces) may also yield more precise habitat-specific es- timates for yelloweye. Differences in density esti- mates for juvenile yelloweye be- tween the Fairweather and Sitka sites are interesting. The higher density on the Fairweather site may be due to terrain: the Fair- weather Ground is a well-defined bank surrounded by large ex- 308 Fishery Bulletin 91(2), 1993 Figure 3 An adult yelloweye rockfish in a typical "refuge space.' Table 3 Density estimates for juvenile yellowe Gulf of Alaska, 1990. ye rockfish by area, eastern Density Area n no./km- CV(D) 95% CL Lower Upper Fairweather 140 Sitka 48 1766 497 21.34 40.77 959.2 2572.7 63.3 929.7 of area of rocky habitat inside the 180 m deep edge of the conti- nental shelf. Allowable biological catch (ABC) and TAC can then be set, using known life-history information including estimates of natural mortality. In 1991, the Gulf of Alaska Plan Team of the NPFMC used this approach in recommending an ABC for the demersal shelf rockfish assem- blage in the East Yakutat Dis- trict (O'Connell et al. 1991). Acknowledgments Funding for the submersible and support vessel was provided by NOAAs West Coast National Un- dersea Research Center, Fair- banks, Alaska. Barry Bracken, Dave Gordon, and Beverly Rich- ardson helped with the field work. Rich Slater and the crews of Delta Ocean- ographies, the M/V Wm. A. McGaw, and the M/V Privateer provided invaluable assistance. Thanks also to Ken Krieger and Paul Skvorc for techni- cal advice, and Gordon Kruse and Terry Quinn II for their input regarding line-transect methods and statistical analysis. Citations panses of soft bottom. This topography may cause ju- veniles to be more closely aggregated than in the Sitka area, where reefs and pinnacles are often linked by hard bottom, e.g., continuous lava flats. Although hard bottom is not ideal habitat, it may promote movement offish between reefs. Our ultimate goal is to develop a quantitative pre- dictive model to estimate density of yelloweye rockfish and other DSR species based on one or more param- eters reflective of structural habitat complexity. In the interim, the use of line transect-derived density esti- mates can be directly applied in fisheries management. In the absence of a fishery-independent biomass esti- mate, the total allowable catch (TAC) currently in use for demersal shelf rockfish had previously been based on historical catch information. In contrast, the transect data allowed us to estimate biomass/km2 and to ex- pand this estimate to a larger area using an estimate Buckland, S. J. 1985 Perpendicular distance models for line transect sampling. Biometrics 41:177-195. Burnham, K. P, D. R. Anderson, & J. L. Laake 1980 Estimation of density from line transect sampling of biological populations. Wildl. Monogr. 72, 202 p. Hayes, R. J., & S. J. Buckland 1983 Radial distance models for the line transect method. Biometrics 39:29-42. Larson, R. J. 1980 Competition, habitat selection, and the bathymetric segregation of two rockfish (Sebastcs) species. Ecol. Monogr. 50:221-239. Love, M. S. & A. W. Ebeling 1978 Food and habitat of three switch-feeding fishes in the kelp I'oR'sts i iff Santa Barbara. California Fish. Bull., U.S. 76:257-271. Love, M. S., M. H. Carr, & L. J. Haldorson 1991 The ecology of substrate-associated juveniles of the genus Sebastes. Environ. Biol. Fishes 30:225-243. O'Connell and Carlile Density of adult Sebastes rubemums in eastern Gulf of Alaska 309 Matthews, K. R. 1991 An experimental study of the habitat preferences and movement patterns of copper, quillback, and brown rockfishes iSebastes spp). Environ. Biol. Fishes 29:161-178. Matthews, K. R., & L. R. Richards 1991 Rockfish (Scorpaenidae) assemblages of trawlable and untrawlable habitats off Vancouver Island, British Columbia. N. Am. J. Fish. Manage. 11:312- 318. O'Connell, V. M. 1991 A preliminary examination of breakaway tagging for demersal rockfishes. Fish. Res. Bull. 91-06, Alaska Dep. Fish & Game, Div. Commer. Fish., Juneau. O'Connell, V. M., & J. T. Fujioka 1991 Demersal shelf rockfish. In Loh-Lee Low (ed.), Status of living marine resources off Alaska as as- sessed in 1991, p. 46-47. NOAA Tech. Memo. NMFS F/NWC-211, Northwest Fish. Sci. Cent., Auke Bay AK, 95 p. O'Connell, V. M., B. E. Bracken, & D. W. Carlile 1991 Demersal shelf rockfish. In Stock assessment and fishery evaluation report for the 1992 Gulf of Alaska groundfish fishery. N. Pac. Fish. Manage. Counc, Anchorage AK. Richards, L. J. 1986 Depth and habitat distributions of three species of rockfish (Sebasles) in British Columbia: observations from the submersible PISCES IV. Environ. Biol. Fishes 17:13-21. Rosenthal, R. J., L. J. Halderson, L. J. Field, V. M. O'Connell, M. J. LaRiviere, J. Underwood, & M. C. Murphy 1982 Inshore and shallow offshore bottom fishery sources in the southeastern Gulf of Alaska. Unpubl. rep., Div. Commer. Fish., Alaska Dep. Fish & Game, Junea 99801. Abstract.— We describe a simple biostatistical model of reproductive success (logarithm of recruits/ spawner) applied to three coastal pe- lagic fish stocks off southern Cali- fornia: northern anchovy Engraulis mordax, Pacific sardine Sardinops sagax, and chub mackerel Scomber japonicus. We used the model to de- tect possible influences of gross cli- matic conditions and contaminant loadings (particularly of metals and organochlorines) on reproduction in these three stocks. Data included several decades of annual estimates of recruitment and stock size, monthly measures of climate, and annual estimates of contaminant loadings; the model included a com- pensatory stock-size component be- fore adding environmental effects. The study was meant to generate, rather than test, hypotheses. For the chub mackerel stock, we detected cli- mate influences, but no contaminant influences, on reproductive success, which was usually high during con- ditions typical of El Nino-Southern Oscillation events. For the northern anchovy stock, we detected no cli- mate or contaminant influences on spawning success; however, the negative results may reflect low sta- tistical power, rather than absence of contaminant influences. For spawning success in the Pacific sar- dine stock, we detected no consis- tent climate influences, but we found a strong negative correlation with contaminant loadings. This result is consistent with the hypothesis that contaminant loadings accelerated the collapse of the Pacific sardine stock while it was under stress from severe overfishing. Although many scientific questions about validation of models, mechanisms of action, and identity of specific deleterious con- taminants remain to be answered, the observed data are well described by the hypothesis of contaminant- mediated decline. 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* Michael H. Prager Southeast Fisheries Science Center, National Marine Fisheries Service, NOAA 75 Virginia Beach Drive, Miami, Florida 33 149 Alec D. MacCall Southwest Fisheries Science Center, National Marine Fisheries Service, NOAA 3 I 50 Paradise Drive, Tiburon, California 94920 Manuscript accepted 29 January 1993. Fishery Bulletin, U.S. 91:310-327 1 19931. In recent years, there has been in- creasing concern about possible ef- fects of contaminants on the nation's living resources, particularly ex- ploited fish and shellfish populations. Although contaminants such as heavy metals and synthetic organo- chlorines are commonly detectable in coastal sediments, waters, and organ- isms (Mearns et al. 1991), it has been difficult, if not impossible, to deter- mine whether sublethal levels of con- taminants have influenced the pro- ductivity of coastal fish stocks. Mass mortalities and deformities of fish and shellfish have occurred after ex- posure to high levels of contaminants, but mortality and morbidity in natu- ral populations attributed to chronic low-level exposures is difficult to prove conclusively, even though it may be strongly supported by circum- stantial evidence (Sindermann 1978). In addition, the question of popu- lation-level effects has been largely unexplored. For individuals of most species, the physiological effects (if any) of chronic sublethal exposure to contaminants are poorly understood. For populations, the effects of such exposure are not known. Nonetheless, evidence from many studies of indi- viduals suggests that populations should be affected also. This paper describes exploratory models of the effects of contaminants, climate, and spawning-stock size on three exploited fish populations off southern California. With a rich body of data, this area is unusually well suited to such a study. The waters off southern California contain el- evated levels of numerous contami- nants, including PCBs, DDT, and sev- eral toxic metals (McCain et al. 1988). Estimates of historical contaminant loadings have been made by Sum- mers et al. (1988), and data on cli- matic conditions are available from public records. Major coastal pelagic fisheries have been monitored by the California Department of Fish and Game since about 1930. Two main conceptual approaches might be used to model the effects of contaminant exposure on pop- ulations. The first approach is a "bottom-up" one: constructing a "Contribution MIA-91/92-26 of the Miami Laboratory, Southeast Fisheries Science Cen- ter, National Marine Fisheries Service, NOAA. 310 Prager and MacCall: Contaminant and climate effects on spawning of three pelagic fishes mechanistic population model that somehow extrapo- lates known effects on individual organisms to effects at the population level. The second approach is en- tirely at the population level: construction of statisti- cal models that correlate historical changes in popula- tion properties (especially properties thought to be sensitive to environmental fluctuation) to changes in contaminant loadings. Neither of these approaches is completely satisfac- tory. Weighing against the first approach is the fact that the existing theory of population biology does not allow extrapolation from physiology and behavior of individuals to the net productivity of the popula- tion. Indeed, biological populations seem to have emergent properties unknowable from the observable properties of individuals (Mayr 1982). A similar prob- lem is that combinations of contaminants may work in unforeseen ways; one might say that such combina- tions have their own emergent properties. Thus the biological activity and availability of mixtures of con- taminants in the coastal environment are difficult, if not impossible, to extrapolate from laboratory studies of exposure to individual contaminants. Given the seemingly insurmountable barriers to constructing mechanistic bottom-up models of populations chroni- cally exposed to low levels of contaminants, we used the alternative approach: we constructed empirical statistical models of contaminant effects on popula- tions. The empirical approach, of course, also has many limitations; these are discussed at length in later sections. A common interpretation of Eq. 1 is that a represents the net fecundity of a unit of spawning biomass and P reflects the degree of density-dependence in the stock's recruitment. The Ricker model is often expressed in terms of the natural logarithm of recruits per spawner: log(f?/P) = a + 6P + e, (2) where a = log(a), b = -p\ and £ = log(£). (The sign change in b is made merely to simplify the notation; b is negative in a compensatory stock.) Assuming that the total number of eggs spawned is proportional to the spawning biomass P, the quantity RIP is an index of an egg's probability of survival to recruitment. This seems to be a more appropriate quantity to use for detecting exogenous effects than recruitment itself, which is usually held to depend in the first order upon stock size. We refer to the quantity RIP as "spawning success" and to its natural logarithm as "log spawning success." The model of log spawning success (Eq. 2) is linear in the parameters and contains an additive er- ror structure, and thus can be fit by ordinary least- squares (OLS) regression. Eq. 2 can easily be modified to incorporate external variables, such as climate or contaminant effects, that might affect spawning success. Suppose we have m such variables, \xu x2, ..., xj, for which we wish to estimate parameters 19,, 92, ..., 9,„|. Then a model in- cluding these variables is Model structure \og(R/P) = a + bP + I e,x, + e. (3) i=l Evidence suggests that the youngest stages of fishes should be most sensitive to the effects of contaminants (Weis & Weis 1989). For that reason, our model of contaminant effects addresses survival from the egg stage to age of recruitment (roughly the first year of life). The model is based on the widely-used Ricker ( 1954) model offish recruitment: In this expanded model, m explanatory variables and population size affect log spawning success in an addi- tive manner. The application of this linearized model is described in a later section. Other adaptations of the Ricker model to represent contaminant effects were made by Goodyear (1983) and Vaughan et al. (1984). R = aPe^pC,, (1) Data sources and processing where R = recruitment in number of fish or biomass, P = parent spawning-stock size (usually spawning biomass), a, (3 = estimated parameters of the Ricker model, e = base of natural logarithms, and C, = a lognormally-distributed stochastic component with zero mean. Two main categories of data were used in this study: data on fish abundances, and explanatory data on the environment. Data on fish populations com- prised time-series of age-structured abundance esti- mates from virtual population analyses (VPA). We ex- amined three coastal pelagic stocks off southern California: northern anchovy Engraulis mordax, Pa- cific sardine Sardinops sagax, and chub mackerel Scomber japonicus (known locally as Pacific mackerel). 312 Fishery Bulletin 91(2), 1993 These stocks were chosen because they have each sup- ported major fisheries and have been studied suffi- ciently well so that long time-series of stock and re- cruitment data are available. These coastal pelagic species also form a conspicuous part of the California Current ecosystem. Explanatory data (data on the environment) included time-series of two types: variables related to climate, and variables reflecting contaminant loadings or other potential human impacts. These data were subjected to a number of stages of preliminary processing, in- cluding two principal component analyses, to compute two sets of derived variables (principal components) from a larger number of potential explanatory variables. Data on fish abundances Stock and recruitment data on northern anchovy (Table 1, Fig. 1) were obtained from Table 8 of Methot ( 1989), who derived his estimates from fishery statistics and research survey data. For spawning-stock size, we used Table 1 Stock and recruitment data used for modeling effects of contaminants and climate on spawning success of three pelagic fish stocks off southern California. Data on chub mackerel have tradi- tionally been reported. maintained, and analyzed in the English system, and are so reported here. Northern anchovy Pacific sardine Chub mackerel Engraulis mordax Sardinax sagax Scorn ber japonicus Spawning Spawning Spawning biomass Recruitment biomass Recruitment biomass Recruitment Year 10"t 10s fish 103t 10" fish 10" lbs 10fi lbs 1946 — 566 3875 97.87 20.11 1947 — — 405 4261 57.91 139.20 1948 — — 740 3690 28.90 70.92 1949 — — 793 290 72.43 11.21 1950 — — 780 397 89.46 3.59 1951 — — 277 972 72.05 3.80 1952 — — 136 1197 32.89 55.35 1953 — — 202 382 10.62 106.48 1954 — — 239 264 38.58 45.90 1955 — — 170 588 80.69 91.41 1956 — — 108 1586 95.00 20.05 1957 — — 90 905 77.30 28.00 1958 — — 177 288 36.98 91.81 1959 — — 122 111 33.59 70.18 1960 — — 88 74 63.02 128.67 1961 — — 54 56 81.32 85.60 1962 — — 27 11 133.10 18.18 1963 — — — — 168.65 9.58 1964 639,210 38.6 — — 112.05 2.67 1965 531,520 13.4 — — 44.27 4.21 1966 541,880 13.3 — — 17.63 5.75 1967 409.170 27.9 — — 4.11 1.14 1968 396,260 19.8 — — 2.57 4.08 1969 374,300 33.1 — — — — 1970 321,880 73.1 — — — — 1971 333,600 123.3 — — — — 1972 588,460 154.2 — — — — 1973 1,812,900 114.9 — — — — 1974 1,481,160 32.6 — — — 10.15 1975 1,458,430 14.3 — — 0.63 2.53 1976 1,050,180 57.2 — — 5.00 138.43 1977 868,890 6.5 — — 22.68 89.87 1978 435,370 114.3 — — 95.06 252.20 1979 582,890 82.0 — — 166.36 35.40 1980 798,280 74.0 — — 255.27 185.41 1981 782,770 19.7 — — 332.28 158.58 Prager and MacCall. Contaminant and climate effects on spawning of three pelagic fishes 313 0.5 1.0 1.5 Spawning biomass 2.0 Figure 1 Recruitment of the northern anchovy Engraulis mordax stock off southern California, (a) Time trajectories of recruitment R ( 10s fish I and spawning biomass P (10fit); (b) time trajectory of log lR/P); (c) stock-recruitment scatterplot, in same units as la). Methot's estimates of spawning biomass on 15 Feb- ruary of each year. For recruitment, we used Methot's estimates of the number of recruits on 1 July of the same year. Methot stated that his esti- mates of recruitments before 1964 are much less precise; we used his estimates of 1964 through 81 (the end of the contaminant data series). Stock and recruitment data on Pacific sardine (Table 1, Fig. 2) are the estimates of MacCall (1979, Table 3), and were derived from fishery statistics. Although the estimates extend through 1964, MacCall (1979) discounted the 1963 and 1964 estimates of recruitment as not sufficiently precise. We used the esti- mates from 1946 (the beginning of the contaminant series) through 1962. Stock and recruitment data on chub mackerel (Table 1, Fig. 3) are from Table 1 of Prager & Hoenig (1989), who compiled data from MacCall et al. ( 1985), Parrish & MacCall (1978), and Prager & MacCall (1988a). The estimates, all derived from fishery statistics, were adjusted by Prager & Hoenig (1989) to a common reporting date. The temporal limits of our analysis, 1946-81, were determined by the lim- its of the contaminant series. The chub mackerel data span a stock collapse and subsequent closure of the fishery and thus have 6yr (1969-74) of data missing. Nonetheless, the chub mackerel series (n=30) is the longest of the three stocks. The data encompass at least two distinct epochs in the ex- istence of the population: the pre-collapse epoch of steadily increasing exploitation rate, and the post-collapse epoch of more strictly regulated fishing and the associated popula- tion increase. Explanatory data: Contaminants Historical contaminant-loadings data were compiled and re- constructed by Summers et al. (1988) according to a mass-balance approach based on manufacturing volume, land use practices, and mobility of particular contaminants. Sum- mers et al. (1988) also compiled more general measures (e.g., wastewater flow) from governmental records. The resulting annual time-series include five major categories: gross in- dicators, physical-biological factors, nutrient loadings, or- ganochlorine loadings, and metal loadings (Table 2). Prager & MacCall (1990) discussed the loading patterns of indi- vidual contaminants and provided time-series plots of each; here, we review major characteristics abstracted from Sum- mers et al. (1988) and Prager & MacCall (1990). The human population of southern California has increased sharply over the last 50 yr, and this is reflected in many contaminant-loading patterns, particularly in the first three categories (Table 2). For example, loadings of nitrogen, phos- phorus, and total organic carbon arise principally from waste- water (sewage) flow, which has increased with population size; power-plant cooling flows, for the most part, also re- flect the growing population's use of electricity. While size of the annual kelp harvest is strongly correlated with human population size, the relationship may be noncausal; this is impossible to ascertain statistically. Nonetheless, size of the kelp harvest was included among other potential indicators of environmental stress, as many species are present in kelp- forest habitat during early stages of life history. With the exception of PCBs, the organochlorines listed in Table 2 were all introduced in the late 1940s. Most fluc- tuations in individual loadings have followed legal restric- tions on use (e.g., of DDT) and subsequent increased use of 314 Fishery Bulletin 91(2), 1993 (a) Spawning biomass (+) o o o o o o o o ' 1 " 4000 .x c 0 3000 % o 2000 o) c 3 «i 1000 « 0 x-x-xt* 0 65 1945 1950 1955 1960 19 Year 2.5 (b) S3 2.0 §1.5. f - | 0.5. 8. o.o CO \ /VV f-05 1 i \ -1.0 v + -1.5. 5 19.= 15 1950 1955 1960 19€ Year (c) „ 4000 c 1 3000 A A A o C 2000 3 W o 1 1000 A a" A A 0 A AA A A . 0 200 400 600 8 00 Spawning biomass Figure 2 Recruitment of the Pacific sardine Sardinops sagax stock off southern California, (a) Time trajectories of recruitment R (106 fish) and spawning biomass P llO'tl; (b) time trajectory of log {RIP); (c) stock-recruitment scatterplot, in same units as (a). 300 1945 1950 1955 1960 1965 1970 1975 1980 1985 Year ' 1945 1950 1955 1960 1965 1970 1975 1980 1985 Year 100 200 300 Spawning biomass Figure 3 Recruitment of the chub mackerel Scomber japonicus stock off southern California, (a) Time trajectories of recruitment R ( 10* lb) and spawning biomass P (10* lbi; (b) time trajectory of log (RIP); (c) stock-recruitment scatterplot, in same units as (a). other pesticides. The sharp rise in pesticide use in the 1950s and the subsequent decline in the 1960s and 1970s are similar for many of these compounds, which leads to difficulty in distinguishing their effects statis- tically. The PCBs, a group of non-pesticide toxic or- ganochlorine compounds, have been used increasingly throughout most of this century, but little is known about the volume of their use and discharge before 1960. Loadings of metals have varied with growth of the human population and also with fluctuations in pat- terns of industrial use. Loadings of some particularly toxic metals, such as lead and cadmium, have declined recently because of environmental regulations. Most of the historical reconstructions of Summers et al. (1988) included values for all years from 1946 through 1981, but some contaminant variables lacked values for 1981. We extrapolated from immediately- Prager and MacCall: Contaminant and climate effects on spawning of three pelagic fishes 315 Table 2 Data available on contaminant loadings and other human impacts to the coastal waters of southern California, used for modeling spawning success of three :oastal pelagic fish species. Headings indicate major categories. Variables were also assigned to groups (in parentheses for principal-com- ponents analyses, used to reduce the number of variables before modeling. Asterisks I*) indicate variable not used for modeling. Type of data Years available Gross indicators (Indicators group) Human population 1890-1980" Municipal wastewater flow* 1890-1985 Physical-biological factors (Indicators group! Kelp harvest 1916-1985 Dredging volume 1900-1983 Power-plant cooling flow 1928-1985 Nutrient loadings (Indicators group 1 Nitrogen 1910-1986 Phosphorus 1910-1986 Total organic carbon 1910-1986 Organochlorine loadings (Organochlorines group) Aldrin 1945-1982 BHC or Lindane 1945-1982 Chlordane 1945-1982 DDT 1945-1982 Dieldrin 1945-1982 Endrin 1945-1982 Heptachlor 1945-1982 Polvchlorinated biphenyls 1929-1986 Toxaphene 1945-1982 Metal loadings (Metals group) Cadmium 1920-1980" Chromium 1929-1980" Copper 1929-1980" Mercury 1929-1980" Nickel* 1945-1980' Lead 1929-1980" Zinc 1929- 1980" d 1980 values. " Value for 1981 extrapolated from 1979 ar b Value for 1981 set equal to 1980 value. ' Time-series contains missing values with in this period. preceding values as necessary (Table 2). The extrapo- lated 1981 values, although necessary for the princi- pal-component analyses, had very little influence on later calculations. The reconstructions of Summers et al. (1988) com- prise the best available dataset, and the only one on this scale, describing contaminant loadings off south- ern California in recent decades. Because of this, veri- fication is difficult. Prager & MacCall (1990) found that the reconstructions agreed fairly well with some recent estimates of metals loadings made by the South- ern California Coastal Water Resources Project (SCCWRP), which has monitored fluxes of metals from southern California sewage outfalls since 1971 (Konrad 1989). However, only a small portion of total metals loadings are in sewage; this may have caused the dis- crepancies noted for some metals by Prager & MacCall ( 1990). We also compared reconstructions with a study of metals deposition in anaerobic sediments in the Santa Barbara basin (H. Schmidt & C. Reimers, Scripps Inst. Oceanogr., La Jolla CA, pers. commun.). How- ever, the sedimentary record, much less precise than the reconstructions, was unable to discern any clear trends or patterns in metals since 1932. Explanatory data: Climate A large suite of climate data (Table 3) for the years 1920-84 was compiled by Prager & MacCall ( 1987a,b,c), who described the data sources and processing in de- tail. Most of the raw data were provided by Dr. An- drew Bakun (Pac. Fish. Environ. Group, NMFS South- west Fish. Sci. Cent., Monterey CA). We detrended most of the sea level (SL) and sea- surface temperature (SST) series to remove rising trends with time. Values for 1983 were not used in computing trend lines, to prevent this extreme El Nino year near the end of the series from distorting the results. Only the series of SST in San Diego did not exhibit a significant trend, and accordingly that series was not detrended. Seasonal effects were removed by standardizing each variable to zero mean and unit standard deviation (SD) by month of year. For example, the January values (1920-84) of Los Angeles rainfall were standardized as a group to mean 0.0 and SD 1.0. Standardization was performed separately for each month and time- series. The climate dataset had a few missing values that we replaced with estimates. We used the BMDP proce- dure AM (Dixon et al. 1983) to estimate these values by stepwise regression on the available data. Reason- ableness of the estimates was verified by simulation and by comparison with the nonmissing data (Prager & MacCall 1990). Principal-component analyses Many of the explanatory variables contained redun- dant information. For example, sea-surface tempera- ture and sea level exhibit very similar patterns over time, as both are related to El Nino conditions and the flow of the California Current. Another example, noted above, is that many contaminant time-series are tightly 316 Fishery Bulletin 91(2). 1993 Table 3 Summary of climate data used for mode] ing spawning success of three coastal pelagic fish stocks off southern California. Head- ings indicate major categories. Variables were also assigned to groups (in parentheses) for principal-corn ponent analyses used to reduce the number of variab es before modeling. Type of data and Years of monthly station location! s) data available Rainfall (Rainfall group) 1920- -1984 San Francisco, Oakland Los Angeles, Long Beach San Diego Seawater salinity (Rainfall group) 1920- -1984 Scripps Pier: surface Scripps Pier: bottom Sea-surface temperature (El Nino group) 1920- -1984 Farralon Islands Pacific Grove San Luis Wharf San Diego Sea level (El Nino group) 1920- -1984 San Francisco Los Angeles San Diego Bakun's upwelling index (Upwelling group) 1946- -1984 30°N, 119°W 33°N, 119°W 36°N, 122°W 39:N, 122:W coupled to human population size, which generates them. This multicollinearity causes problems in fitting and interpreting statistical models such as ours. In extreme cases of multicollinearity, numerical algorithms used for fitting can fail. A more likely problem is that interpretation is not straightforward, because param- eter values depend upon the other variables in the model. When sea-surface temperature is included in a model, for example, it may explain a large amount of the variance associated with sea level, so there is little to be gained from including sea level itself in the model. Yet an effect may truly be due to sea level (or to the strength of the California Current), and this relation- ship may thus be overlooked. Because of this difficulty, any biostatistical study of this nature is inherently unable to isolate the effects of individual causes. An important related practical problem is that a large number of potential explanatory variables makes any fitting procedure unwieldy. We addressed some of these statistical problems by combining the explanatory data into new composite variables that did not contain duplicate informa- tion. These new variables, constructed by principal- component analysis, are linear combinations of the original variables. To reduce the number of variables (initially -216) to a number more easily analyzed by standard statistical software, data on sea level, sea- surface temperature, and salinity were converted to bimonthly means; also, surface and bottom salinities at Scripps Pier were averaged. Monthly rainfall val- ues were transformed into an annual value of total rainfall (preceding 1 July to current 30 June) and a value representing the median date of the season's rainfall. The explanatory dataset with these changes contained 94 variables before reduction by principal- component analysis. We constructed two separate sets of principal com- ponents. The first set was constructed from physical and climate data (Table 3) but not data on contami- nants or general stressors (Table 2). A recruitment model based on this analysis would show how much of the variability in spawning success could be attributed to stock size and climatic variation alone. The second principal-component analysis included all explanatory variables, and was used to reveal how much more vari- ability could be explained by adding contaminant in- formation to the analysis. Weighting was used in the principal-component analyses to avoid giving undue emphasis to variables (e.g., SST) measured at many locations. To determine weights, each variable was assigned to one of six groups, as shown in Tables 2 and 3. Each group re- ceived 1/6 of the total weighting, which was divided equally among the variables within the group. The results of principal-component analyses are fre- quently difficult to interpret, as the components are formed on purely statistical grounds. Extensive graphi- cal analysis (presented in Prager & MacCall 1987c, 1990) allowed attaching an interpretation to some, but not all, of the components used in these analyses (Table 4). In interpreting results of the recruitment models, we used a different approach, that of measur- ing the correlation of the model's explanatory effect with the individual variables, as explained below. Application of model to three fish stocks Overview For each stock, we developed two alternative models: one using the principal components of climate data only, and the second using the principal components of cli- mate and contaminant data. We then examined corre- lations of each model's estimated explanatory time- series (i.e., the summation on the right side of Eq. 3) to the original climate and contaminant variables. Be- Prager and MacCall: Contaminant and climate effects on spawning of three pelagic fishes 317 Table 4 Interpretation of principal components (PCs) used as explanatory vari- ables in models of spawning success in three coastal pelagic fish stocks off southern California. Variables Wj— W10 are PCs of data on climate; vari- ables C,-Cl0 are PCs of data on climate and on contaminant (e.g., metals, organochlorines) loadings. Because interpretation of PCs is difficult, other methods were used to interpret model results (Fig. 4-6; text). SST = sea- surface temperature, SL = sea level. Variable Interpretation W, Heavy rainfall; elevated SL and SST; IV, has peaks in 1958 and 1983 corresponding to El Nino events. w2 Very low rainfall, high SST and SL in second half of year. w3 Increased upwelling, especially in mid-year. wt Late rainy season with lower-than-average rainfall, decreased SST at northern stations. ws Reduced upwelling at 30°N station, late rainy season. w6 Late rainy season in San Francisco. W-, Heavy rainfall in San Francisco, high early-year salinity in La Jolla. wt No clear interpretation. w9 No clear interpretation. wu, No clear interpretation. c, Increased upwelling midyear; high loadings of metals, nutrients, and most organochlorines; C, is in part a trend component, increasing from 1945 to 1978 and declining slightly thereafter; c2 High loadings of organochlorines, Cd, Cr, and Pb; reduced rainfall. c3 Elevated SL and SST, heavy rainfall; highest in years of El Nino events. c, Elevated upwelling, especially at 30°N station; high loadings of Cr and BHC; C, is high around 1970, lower in other years. cs Early, wet, rainy season, especially in San Francisco. c6 High loadings of Hg, early rainy season in southern California. c7 No clear interpretation. cs No clear interpretation. c9 No clear interpretation. c,„ No clear interpretation. cause one of the models contained no contaminant data, agreement between the two models on the importance of contaminant variates was considered an indication of severe and troublesome collinearity between relevant climate and contaminant variables. Alternatively, it might simply indicate no contaminant effects. In the presence of strong contaminant effects, and lacking such severe collinearity, one would expect the ex- planatory effect from the model including contaminants to be more highly correlated with the contaminant variables. Selection of explanatory variables It is widely recognized by modelers and stat- isticians (e.g., Gilchrist 1984:11) that a pro- found source of uncertainty in statistical modeling is the possibility of error in speci- fying the model's structure. Such specifica- tion error biases parameter estimates and renders most confidence intervals and hy- pothesis tests invalid (Kennedy 1979, Gil- christ 1984). Unfortunately, the possibilities of specification error and its ramifications are often overlooked when statistical mod- els are used in ecology. The models chosen and presented here were undoubtedly mis- specified; they may have included unimpor- tant effects or omitted important ones, and they were limited to a linear functional re- lationship, which is unlikely to be the true one. After specifying the linear structure of Eq. 3, choice of variables was the main con- cern. When no theoretical basis exists to guide it, choice of variables in a regression model must be regarded as heuristic. Nei- ther the theory of the underlying disci- pline— ecology — nor that of statistics can answer this question unequivocally, so we were forced to use an empirical approach. We started by retaining only the first 10 components from each of the principal-com- ponent analyses. (In each case, this retained -80% of the weighted variance. ) To arrive at a parsimonious model, we fit regression models to all combinations of 10 or fewer variables and ranked the many models by Cp, a goodness-of-fit statistic from Mallows (1973). For each combination of stock and data type (climate or combined), we accepted the model with the lowest Cp. However, al- ternative models of similar fit were similar in structure. Any method of variable selection in which many candidate explanatory variables and combinations of variables are examined may lead the investigator to accept models which fit well solely through chance. This point has recently been empha- sized by Flack & Chang ( 1987). Although the use of Cp is a relatively conservative approach, the true statisti- cal significance of our results is unknown. Thus we view the modeling exercise as one of hypothesis gen- eration, rather than hypothesis testing. The limita- tions of inference that come from an empirical choice of model structure are not unique to this study; they pertain to all modeling exercises except those in which 318 Fishery Bulletin 91(2). 1993 the final model structure and set of explanatory vari- ables are correctly chosen before fitting. Here, we use elements of hypothesis testing, but acknowledge the impossibility of testing the hypotheses rigorously. Detecting the influence of contaminants We started with the following null hypothesis: "any apparent influence of contaminants can be explained by random variability alone, including chance correla- tions of contaminant loadings with climate variabil- ity." We imposed three statistical criteria to be met before we would consider rejecting this null hypoth- esis on the basis of a model including contaminant- loadings data. The first criterion was qualitative: the correlations of the model's explanatory effect with most contaminant variables had to be negative. The other criteria were quantitative: the regression coefficients of the model had to be statistically significant at oc=0.05, and the inclusion of contaminant information had to provide more explanatory power than use of climate data alone. If all three criteria were met, we would admit the possibility of contaminant influence on spawning success. The three criteria are described in more detail immediately below. Sign test Of the variables in Table 3, the metals and organochlorines have been found by bioassay to have deleterious effects on fish (Weis & Weis 1989). (We omitted nickel because of an incomplete time-series and, more importantly, its relatively low toxicity.) Power-plant cooling flow also is generally considered deleterious, but the remaining variables in the indica- tors group could be favorable as well as unfavorable. The signs of the correlations with the 16 "deleterious" variables were subjected to a (one-tailed) statistical test of the following null hypothesis: "The observed number of negative correlations could arise by chance alone." Under this null hypothesis, the probability of a negative correlation between a contaminant variable and the model's estimated explanatory time-series is 0.5. Assuming independence among the 16 contami- nant variables (an assumption that may be violated; see below), the probability of obtaining x negative cor- relations by chance is given by the binomial probabil- ity fix; « = 16, p=0.5). The probability of observing 11 or more negative correlations is 0.105, and of observing >12 is 0.038. Thus, the conventional error rate of a<0.05 requires >12 negative correlations between the 16 del- eterious contaminant variables and the explanatory time-series from a model that uses the combined ex- planatory data. Such a test concludes that apparent contaminant effects are qualitatively significant if 12 or more nega- tive correlations are found. However, the true prob- ability of Type-I error is >0.038, because many of the contaminant variables are positively correlated with one another, violating the assumption of independence in the binomial probability model and increasing the tail probabilities. While we cannot calculate an exact critical value of x in view of lack of independence, the value would be larger than the nominal value of 12 required for a model to pass this test. Test of coefficients This test was used to assess whether the coefficients of the chosen model were sig- nificantly different from zero. Although we always chose the model with lowest Cp, that criterion does not de- pend directly on how precisely the regression coeffi- cients are estimated. To test the significance of regres- sion coefficients, we used the standard <-tests (at a=0.05) provided by the statistical software, and re- quired all coefficients to be significant for the model to pass the test. Improvement-in-fit test Our third criterion required that the model that incorporated contaminant infor- mation (i.e., which used the combined explanatory data) had to provide a substantially better fit to the spawn- ing success data than the model using the climate data alone. To judge this, we used the Schwarz criterion (Smith 1988), a statistic for comparing non-nested mod- els. The Schwarz criterion requires an estimate of o, the true model variance; we used the MSE of the bet- ter-fitting model. To pass the test, the model with com- bined data was required to have a higher value of this statistic than the model with climate data. Results Northern anchovy Not unexpectedly for a short-lived species, the spawn- ing biomass of this northern anchovy stock appears to depend strongly upon the preceding few years' recruit- ments. A peak in spawning biomass generally follows a corresponding peak in recruitment by 1 or 2yr (Fig. la). The stock-recruitment relationship, although noisy, appears density-dependent, in that spawning success improves at lower stock sizes (Fig. lc), and each model includes a significant compensatory term (the nega- tive stock-size parameter; Table 5). Ricker (1954) and MacCall (1980) have pointed out that cannibalism of eggs or larvae by the adults could cause such compen- sation. Neither the climate model nor the combined model explained even half of the variability in log spawning success of northern anchovy (Table 5). Although the model incorporating contaminants had somewhat Prager and MacCall. Contaminant and climate effects on spawning of three pelagic fishes 319 Table 5 Summary results of regression models of logarithm of recruits per spawner for three coastal pelagic fish stocks off southern California. Predictor variables were parent stock size IP) and either principal components I W„ ; = 1 10 1 of climate data I models denoted ) or principal components ( C,, i = l 10) of climate data and data on contaminant loadings i models denoted V). Each model also included an intercept. Variable selection used an all- subsets algorithm with the C. statistic; thus models may include variables not significant at P<0.05. and all P values are considered nominal. Variables included in model (with sign of df of coefficient! and nominal probabilities oft- model, statistics for H„: coefficient = 0. error F Nominal statistic Prob > F R- Northern anchovy Engraulis mordax -P. 0.025; -W„ 0.13 2,15 • -P, 0.17; -C5, 0.087; -C;, 0.13 3,14 Pacific sardine Sardinops sagax -W-„ 0.030 1, 15 • -P. 0.0004; -C„ 0.0001; -C5, 0.002; +C10, 0.002 4, 12 Chub mackerel Scomber japonicus -P. 0.0001; -W, 0.015; +W„ 0.021; +W6, 0.008; -VV7, 0.023; -W„„ 0.09 6, 23 • -P. 0.0003; +C„ 0.094; -C4, 0.004 3,26 4.20 4.14 5.77 14.4 0.036 0.027 0.36 0.47 0.030 0.28 0.0002 0.47 6.74 0.0003 0.64 8.11 0.0006 0.48 higher R2 and nominal significance level* than the model including only climate effects, it is impossible to attribute this difference to the effects of contaminants on recruitment. Given the small sample size and rela- tively small difference in fit, it seems more logical to attribute it to chance. The two models of anchovy spawning success can be compared by plotting the correlations of the two corre- sponding explanatory effects with the original explana- tory variables (Fig. 4). For some variables (e.g., those related to rainfall), correlations are similar between models; however, many other variables fall into the second and fourth quadrants of the plane, meaning that they are positively associated with spawning suc- cess in one model and negatively associated in the other. For example, the two models assign opposite signs to the influence of most upwelling variables. Based on this inconsistent pattern of correlations (Fig. 4) and the lack of nominally significant param- eters (except for the stock-size parameters) in Table 5, we conclude that these models identify neither climate- nor contaminant-related variability in the spawning success of this stock. Indeed, the major determinants of anchovy recruit- ment strength remain to be discovered. Smith (1985) speculated that recruitment might be controlled in late- larval stages through "plasticity of the interaction be- tween growth rate and survival." Peterman & Bradford (1987) detected a decrease in larval survival associ- ated with episodes of high wind speed; however, Peterman et al. ( 1988 ) did not detect significant corre- *To emphasize the impossibility of determining true significance levels, reported significance levels are denoted "nominal." R Rainfall O Sea Level O Sea Surface Temperature U Upwelling • Contaminants 08 - u 0.4 - U uu ! u « f • • •u* - u u u • u ! u R "GO o 0.4 - U R u # R u • • R Ou R • ! „ a L° ® JW On • Vq, or " o ° *p <*P o o • i o ° „i u • 6 -0.8 -0.4 0 0 0 4 0 8 Correlation of "Combined" Effect with Variables Figure 6 Comparison of two regression models of logarithm of spawn- ing success (recruits/spawner) of chub mackerel Scomber japonicus stock off southern California. Models include ef- fects of stock size and environment. Coordinates of a point are the correlations of the models' explanatory effects (see textl with an explanatory variable. For legibility, only cat- egory of variable (point) is indicated. Vertical axis: correla- tions with model estimated on climate data. Horizontal axis: correlations with model estimated on combined climate and contaminant data. Points in first and third quadrants of plane indicate agreement between models as to a variable's effect; points in other quadrants indicate disagreement. Strong agree- ment between models suggests that addition of contaminant data adds little to this model. occurred in the late 1960s and beyond, when the chub mackerel stock was declining from overfishing and then recovering (MacCall et al. 1985); this coincidence could have induced noncausal statistical correlations. Corre- lations with organochlorines were negative. In inter- preting this, it is relevant that spawning success in chub mackerel is highly autocorrelated and appears periodic (Fig. 3b). The organochlorine abundances are also very highly autocorrelated, usually consisting of a single peak. This could easily lead to correlations like those observed, even if the observed contaminant lev- els did not affect spawning success. Although several contaminants correlate very highly with spawning success, the correlation patterns are nearly identical for the two models, one of which con- tains no information on contaminants. We interpret this to indicate that any contaminant effects are indistinguishable from effects of climate variables that fluctuate similarly over time. The strong agree- ment between the two models also suggests that recruitment-based contaminant analyses of this stock may be exceptionally prone to Type-II error, i.e., fail- ure to detect any contaminant effects that might actu- ally be present. Parrish & MacCall (1978) also explored the effects of climate on chub mackerel spawning success, but comparison of this work and theirs is difficult for two reasons. Parrish & MacCall (1978) examined many explanatory variables not included in this study In addition, changes in the estimated maturity schedule and the VPA-based stock and recruitment information (Prager & MacCall 1988) have resulted in a markedly different, as well as considerably extended, stock and recruitment series. Nonetheless, many of our estimated correlations agree with those of Parrish & MacCall (e.g., the positive effects of sea-surface temperature in the south, and the upwelling index at 30°N). Whereas Parrish & MacCall found that spawning success had a negative correlation to sea level at La Jolla, we esti- mated positive correlations with sea level at La Jolla and Los Angeles from January to June and indepen- dence in the later months. Statistical evaluation of contaminant effects In a preceding section we developed three statistical criteria for admission of possible contaminant effects. The sign test required 12 negative correlations between 16 "deleterious" contaminant variables and the explana- tory effect of a model using the combined explanatory data. Correlations for models of all three stocks were mostly negative. Models of northern anchovy (Fig. 5) and chub mackerel (Fig. 6) each contained 11 negative correlations, but neither result was sufficient to reject 322 Fishery Bulletin 9 1(2). 1993 the null hypothesis. However, all 16 correlations were negative for the Pacific sardine (Table 61. Regression coefficients were significant in models of Pacific sardine and chub mackerel. The model of north- ern anchovy spawning success fit relatively poorly (Table 5), and although the overall model was statisti- cally significant, the individual regression coefficients were not. According to the Schwarz criterion S, the models using combined data were marginally preferable to the models using climate data for chub mackerel and north- ern anchovy (S increased in the third significant figure). For sardine, the combined model was strongly prefer- able (S in ratio 2:1). Even allowing that the chub mack- erel and northern anchovy models passed this test, the only stock to pass all three tests for suspicion of contaminant effects was the Pacific sardine. Discussion A somewhat similar biostatistical study was conducted by Martin Marietta Environmental Systems to exam- ine possible effects of contaminants on fishes in five estuaries of the U.S. Atlantic coast, with emphasis on the Hudson-Raritan basin (Summers et al. 1985, 1987). The study used a categorical regression model of his- torical landings (Rose et al. 1986); the autoregressive model did not attempt biological or demographic real- ism. After accounting for the effects of hydrographic conditions and previous stock sizes, Summers et al. (1987) detected correlations related to contaminant variables including dissolved oxygen, dredging, and bio- chemical oxygen demand. The study concluded that there were consistent patterns of anthropogenic influ- Table 6 Statistical criteria used to screen models of spawning success of three coastal pelagic fish stocks off southern California. As predictors, models used parent stock-size and principal components derived from measures of climate and of contaminant loadings. Models were required to meet all three criteria in this table before being considered suggestive of contaminant influences on spawning success. Statistical criterion Northern anchovy Number of significant negative correlations with contaminants thought to be deleterious 1 12 required!. Significance of coefficients in model that included contaminants. Improved fit of model including contaminants, compared with model including only climate. (Improvement judged by Schwarz criterion. I 11 ences among similar stocks across different estuaries. In a second categorical regression study, Summers et al. (1990), while expressing some reservations about ultimate causes, concluded that sewage loadings have had a negative effect on white bass Morone americana production in the Choptank River, Maryland. The present study was not undertaken with the ex- pectation of finding clear evidence that contaminants affected fish stocks, nor clear evidence to the contrary. We knew of many reasons that contaminant effects might be difficult to detect, including major theoreti- cal and practical difficulties in the data, shortcomings in biological knowledge, and the limitations of avail- able statistical methodology (for details, see Appen- dix). As expected, we failed to detect contaminant ef- fects on anchovy or chub mackerel, although we are unable to state whether this reflects a lack of such effects or merely low statistical power. In contrast, the results for Pacific sardine suggest that contaminant stress, at a time of severe overfishing, contributed to the decline and collapse of this stock. Partly because we searched for the best among many possible models, we cannot attach any level of statistical certainty to this suggestion. Our search selected a model of high explanatory power (R2=83%) in which contaminants were strongly represented, and that, we believe, makes a strong case for further research into the question. Important areas of research would include investigat- ing how such a model of historical events, not formu- lated a priori, can be statistically validated (or invali- dated); which individual contaminants may have contributed to the stock collapse; and by what mecha- nisms contaminants may have played a role in the decline of Pacific sardine. An alternative hypothesis about Pacific sardine is that the true spawning-success relation- ship is depensatory rather than linear, and thus depensation, rather than exogenous influ- ences, mediated the decline of the stock. Consistent with this hypothesis, the stock-size com- ponent of our model (Fig. 7a) fits the observed spawning-suc- cess data quite well until the mid-1950s (years of very low population size) when the stock-size component predicts higher spawning success than was observed. To examine this alternative hypothesis in more detail, we fit spawning-success models using the following variant of the gamma function: Fish stock Pacific sardine Chub mackerel 16 yes yes yes yes Prager and MacCall. Contaminant and climate effects on spawning of three pelagic fishes 323 \og(R/P) = a + ylnP + 6P+Ie,.v, + 8. (4) This differs from our original model, Eq. 3, only in including the term containing the estimated parameter y. When y = 0, this term drops out and the model is equivalent to Eq. 3. When y < 0, the model is more strongly compensatory than Eq. 3; when y > 0, the model is depensatory. In results of fitting this model, the depensatory effect was estimated as significant only in the absence of climate or combined vari- ables stags to 60 cm and age 20 yr (Lavenda 1949) with greater variability in size-at-age than females. Conse- quently, a black sea bass has three possible growth rates: female, male, and transitional. These dif- ferences in growth rates and vari- ability in size-at-age between the sexes will influence the potential yield from a cohort. An alternative approach for modeling changes in a cohort over time is use of delay models. Delay models were first devel- oped in the field of industrial dy- namics as a technique to moni- tor movement of individuals through a system of substages (Forrester 1961), and have since been modified to incorporate ef- fects of attrition (mortality) and variability in timing of move- ments (Manetsch 1976, Vansickle 1977). The subclass of distributed delay models can be a useful tool to describe the movement of any item through a process or, in a biological context, through developmental stages (Manetsch 1976). Extending the model to include a series of consecutive processes, it has been used to simulate growth dynamics in marine crustaceans (Idoine & Finn 1984), insects (Ravlin et al. 1978, Schaub & Baumgartner 1989), and agricultural crops (Gutierrez et al. 1984, 1988). The major advantage of this model type is its focus on the aggregated behavior of individuals rather than a representative mean. To estimate Y/R and SSB7R in black sea bass, we modeled the growth and mortality of a cohort as a se- ries of distributed delays with an associated mortality, using length categories as individual developmental stages. This approach allowed us to simulate the de- cline of the cohort while retaining information about variation in size composition and size-specific mortality in the cohort. In addition, we were able to evaluate the influence of additional mortality on the population dy- namics of a species with an hermaphroditic life history. Methods The model structure Time-invariant distributed delay models with a mor- tality term (Vansickle 1977) can be characterized as a sequence of stages, with flow through each stage i rep- resented by a series of differential equations (Fig. 1): TRANSITION Cf the size of recruitment to the fishery x the ratio of fishing mortality/total mortality (F/Z). The biomass offish removed via fishing mortality was calculated using the appropriate sex- specific length-weight equations. Spawning stock was calculated as the sum of the number of females per length remaining in the system at the time of spawn- ing (1 June) x the probability of female maturity- at-length. The number of mature females-at-length was converted to biomass using a length-weight equation. With this model, hermaphroditism was included by splitting the cohort into three growth regimes and us- ing a size-specific probability of sex transformation. At the initial stage, the cohort was divided into an appro- priate number of males and females, then passed through length stages at sex-specific rates. At a desig- nated length-interval, females began transformation by passing through an intermediate transitional stage prior to entering the male growth sequence (Fig. 1). Input parameters Input parameters required in the model were mean transit time (D) and its associated variance (S2) per 1 cm length-category in days and days2 respectively, by sex; maximum potential length for each sex in the population; probability of sex transformation-at-length for females; a range of lengths-at-recruitment (Lc) to the fishery; instantaneous natural mortality rate (spe- cific to length groups and sex if appropriate); a range of instantaneous fishing mortality rates; and the per- centage of females mature-at-length. In addition, length-weight equations by sex were necessary for con- version of numbers-at-length to biomass. Estimates of mean transit times and their associ- ated variances for black sea bass were determined from back-calculation of scale data. Sea bass scales were collected in coastal Long Island during 1979-80, aged, and length-at-age back-calculated (Mark Alexander, Conn. Dep. Environ. Prot, Old Lyme, CT, pers. commun.). Daily time-increments of growth were chosen to pro- vide estimates of k. Transit time (D) by cm-intervals was calculated as D = 365/[(SI,-S„.1)*SL1 S where Sn = scale annulus n, Sn.! = scale annulus n-1, S = total scale size, SL = standard length offish, with the assumption of linear growth between annuli (i.e., D equal between cm-intervals (S^/S^SL and (S„/S)*SL). Thus, each individual scale provided esti- mates of transit times (D) by cm, up to the maximum length of the fish (SL). The means of D and S2 were Shepherd and Idoine Yield- and spawning biomass-per-recruit for Centropnstis striate 331 B FEMALE TRANSITIONAL MALE TRANSITIONAL FEMALE 10 20 30 40 50 STANDARD LENGTH (cm) 10 20 30 40 50 STANDARD LENGTH (cm) Figure 3 (A) Calculated values of number of substages, k, and (B) mean transit times, D, for female, male, and transitional stages of black sea bass Centropnstis striata. 0 16 £° < DC °-o z 1° 3° cc 10 15 20 25 30 35 40 STANDARD LENGTH Figure 4 Transformation rates-at-length of female black sea bass Centropristis striata used in the delay model. estimated per length category for each sex, with sex determined at the time of capture. Delay and variance estimates were extrapolated with a linear regression between the largest length categories in the dataset and maximum potential length (Fig. 3). Maximum lengths were 49cmSL for females, 39 cm SL for transitionals, and 60 cm SL for males. D and S2 for each length-interval in the transitional category were estimated by using the mean values of females and males. Transformation rates-at-length were estimated from a composite of the frequency of transitional-stage fish recorded in field observations (Fig. 4) (Mercer 1978, Low 1981, Wenner et al. 1986). Transformations be- gan at 8 cm, and individual fish were allowed to exist in a transi- tional state for 1 cm. Sex-specific length- weight equations (Mercer 1978) used to convert standard length (cm) to weight in grams were males: wt=0.01773 SL 3 1525 females: wt=0.02810 SL 30104 transitional: wt=0.02120 SL 30991 Annual instantaneous natural mortality (M) for both the Y/R and SSB/R calculations was mod- eled as a length-dependent rate equal to 0.3 at lengths 1-10 cm and 0.2 at 11 cm-max. length. In- stantaneous fishing mortality (F) was varied between 0.0 and 1.50 in both calculations. Mortality rates were applied over time- intervals of Id (t= 1/365). Length-at-first-capture (Lc), which was modeled as knife-edge recruitment, varied between 16 and 32 cm SL. Increments were equivalent to annual mean lengths at successive ages, as determined from the delay model. The effect of harvesting was examined by incorporating an additional mortality term at each length category beyond the size-at-recruitment. Maturity-at-length data were collected during NEFSC bottom-trawl surveys between 1982 and 1990 (O'Brien et al. 1993). The initial cohort in the model consisted of 2000 fish divided according to a sex ratio of 99:1 female to male (Mercer 1978, Wenner et al. 1986). Pre- cision of rounded values in the computer program re- sulted in the net loss of some individuals; therefore, the number of recruits used in the per-recruit calcula- tions was the sum of individuals accounted for at the end of the run rather than the initial input value. The yield model was run for 25 yr or until the number of remaining fish in the cohort was <1; and the spawning biomass model (females only) was run for 11 yr or un- til the cohort was reduced to <1 fish. The Y/R and SSB/R models were developed as separate computer programs written in ANSI standard FORTRAN. The Y/R estimates from the distributed delay model were compared with results from a Thompson-Bell (T-B) Y/R model. Lengths-at-annual-intervals for males and females were derived from execution of the delay model without a transitional phase and converted to grams using the appropriate length-weight equations. The mean of the male and female weights-at-age served as input to the T-B Y/R model. M was set equal to 0.3 for age 1, and 0.2 thereafter. The traditional SSB/R 332 Fishery Bulletin 91(2), 1993 model is a simple extension of the T-B Y/R model (Gabriel et al. 1989) and, consequently, the relation- ship to the delay models was similar. To avoid redun- dancy, only the relationship between the T-B Y/R model and the delay models was represented. The sensitivity of the delay model to changes in the transition rates was examined. The transitional size- range was divided into 8 cm length-intervals and the transition rates within each interval were doubled. Spawning stock biomass-per-recruit was estimated for each interval of increased transition rates over a range of fishing mortalities and a constant size-at-first-capture of 16 cm. Table 1 Mean length -at-age (cm) and variance from distributed delay model and mean back-calcu ated lengths-at- age (Mercer 1978 for black sea bass Cen ropristis striata. Age (yr) Distributed delay Back-calculated Male Female Male Female 1 9.6 (0.56) 9.5 (0.45) 8.7 9.0 2 17.7 (0.89) 17.2 (0.96) 16.5 16.3 3 22.5 (1.25) 21.7 (1.20) 21.1 20.4 4 26.5 (1.65) 25.4 (1.64) 24.4 23.6 5 30.1 (1.85) 29.0 (2.20) 27.6 26.1 6 33.5 (1.92) 32.8 (2.35) 31.4 27.9 7 36.6 (1.93) 36.4 (2.17) 34.6 33.6 8 39.5 (1.91) 39.9 (2.11) 36.5 9 42.3 (1.86) 43.2 (2.02) 38.4 10 44.9 (1.82) 46.5 (1.84) 11 47.4 (1.79) 48.4 (0.65) 12 49.8 (1.76) 13 52.1 (1.72) 14 54.3 (1.68) 15 56.4 (1.64) 16 58.4 (1.31) 17 59.5 (0.54) 18 59.8 (0.18) Results To confirm that the model accurately represented growth in the absence of any harvest mortality, we calculated the mean lengths of the cohort (by sex) at successive 365d periods as generated by the model, and compared the results with independent back- calculated mean lengths-at-age (Mercer 1978) (Table 1). Growth simulated by the delay model was com- parable to growth observed from back-calculations, although sizes-at-age tended to diverge in older ages, with slightly higher lengths-at-age produced by the delay model. Under the influence of natural mortal- ity only, a cohort introduced into the delay model was reduced to 0.1% of its initial number by the time the maximum length was attained. The reduc- tion of the cohort using only natural mortality indi- cated that the model provided an accurate portrayal of the black sea bass growth rate for the stated set of mortality parameters. To evaluate the effect of harvest mortality in the absence of sex transformation, we calculated Y/R as generated by the delay model without transition and compared the results with those obtained using the T-B model. Under various Lc values, yields-per-recruit from the two models were similar, indicating the basic Table 2 Estimates of F ,,„ for black sea bass Centropristis striata from distributed delay model with and without transitional phase and using Thompson-Be 1 model. Age-at-entrv in yr. Size (age) at entry Distributed delay Thompson-Bell w/ transition w/o transition 16 cm (2) 0.16 0.17 0.17 21 cm (3) 0.20 0.21 0.21 25 cm (4) 0.25 0.27 0.26 28 cm (5) 0.31 0.33 0.33 32cm(6i 0.45 0.47 0.45 35 cm (7) 0.63 0.64 0.63 Shepherd and Idoine Yield- and spawning biomass-per-recruit for Centropristis striata 333 350 300 CT250 1- EC 200 III DC //-. Lc = 32 if ^<<^ YIELD per o o 1 Lc=16 ^-^^ 50 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 FISHING MORTALITY Figure 5 Yield-per-recruit for black sea bass Centropristis striata calculated by the distrib- uted delay model without transitions (solid) and with a transitional phase (dash I at vari- ous alternative lengths-at-first-capture (Lc). M = 0.3 ( l-10cm) and 0.2 (11-max. length). 120 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 FISHING MORTALITY Figure 6 Sex composition of total yield for black sea bass Centropristis striata estimated from a dis- tributed delay model with sexual transition (dashed) and without a transitional phase (solid) for varying sizes-at-first-capture (Lr). delay model was comparable to traditional methods of estimating yield-per-recruit. The calculated biological reference point of Fmax from both models was nearly identical (Table 2). The addition of the transitional phase into the model had little effect on the estimates of Y/R. The calculated values of Y/R and Fmaj( were similar for each recruitment size and fishing mortality (Fig. 5, Table 2). Although the effect of including the transition phase on the total yield was negligible, the effect on sex composi- tion of that total yield was substantial. With no transition phase, the proportion of females in the yield was dependent upon the initial sex ratio, in this case 99% female (Fig. 6). If the sexual transformation of females to males was included, the percentage of females to the total yield increased as a function of F and de- creased as a function of size-at-first-capture (Fig. 6). At a recruit- ment size of 16 cm SL (age 2), the percentage of female biomass in the yield went from 21.3% at an F=0.05 to 70.4% at F=1.5, whereas a change in recruitment size to 32cm (age 6) decreased the effect of F, resulting in a percentage range of 17.3-20.3% females for F=0.05 and 1.5 respectively. The addition of a transitional phase had a significant impact on the estimates of SSB/R from the distributed delay model and sub- sequent estimates of a total female spawning biomass. In the ab- sence of any harvest mortality, maximum SSB/R for a cohort that undergoes a transition was 600.8 g/recruit as compared with 2373.3 g/recruit without transition (Fig. 7). Over the life of the cohort, the spawning biomass-at-age increases similarly for the first 2 yr regardless of the form of the delay model (Fig. 8 for size- at-recruitment of 25 cm), at which point the effect of transition begins. At age 2.5, a cohort undergoing transition approaches the maximum contribution to spawning biomass-at-age, while a cohort that does not undergo transition makes its maximum contribution at age 10. The inclusion of the transitional term in the model led to lower proportional reductions in SSB/R with increasing F In a model with- out the transitional form (e.g., size-at-first-capture of 25cm [age 4] Fig. 9) at fishing mortality of 1.5, the spawning potential was reduced to 15.2% of maximum spawning potential (%MSP). In a model with the transition term, however, 28.7% of maximum spawning potential was obtained (Fig. 9). The pattern is also obtained at other sizes-at- recruitment. The transitional phase also reduced the sensitivity of SSB/R to changes in F and size-at-recruitment (Lc). Decreasing F from 0.6 to 0.2 (Lc=25) increased SSB/R by 107% in the non- transitional model, but only 60.5% in the delay model with transitions. Similarly, an increase in size-at-first-recruitment from 16 to 32 cm (with F=0.6) increased SSB/R by 201.8% in the transitional model but 368.9% in the non-transitional version. The inclusion of a transitional phase increased the natural reduction in the number of females in the system, which consequently reduced the maximum SSB/R for the cohort and the relative influence of fishing mortality on SSB/R. The sensitivity of estimates of SSB/R to changes in transforma- tion rates was examined. Doubling the transition rate across all lengths decreased the SSB/R by 71%- from 600.8 to 174.6 g/recruit. When the transition rates were doubled over 8 cm length-increments, the impact in SSB/R varied by F and the size-range over which the transformation rate was changed (Fig. 10). Doubling the rate of 334 Fishery Bulletin 91(2), 1993 2500 2000 1\ — 1500 Ol \ \ \ X CD U W 1000 16\ X5 " 500 V\^5^-~_ 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 FISHING MORTALITY Figure 7 SSB/R for black sea bass Centropristis striata estimated from the distributed delay model without transition (solid) and with the transi- tional phase (dashed). M = 0.3 (1-10 cm) and 0.2 (11-max. length) with varying lengths-at- first-capture (Lc). 700 10 11 Figure 8 Spawning stock biomass-at-age for black sea bass Centropristis striata estimated with a dis- tributed delay model with transitions (dashed) and without transitions (solid) for length-at- recruitment of 25cm. M = 0.3 (1-10 cm) and 0.2 ( 11-max. length). transformation in the 16-23 cm range had the greatest effect, decreasing SSB/R by 39.1% with F=0 (Table 3). Changes in transformation rates in the tails of the trans- formation probability distribution had little impact. In all cases, increases in F decreased the relative impact of transformation loss and, consequently, the sensitivity to changes in transformation rates decreased. Discussion Ignoring the unique population dynamics that may oc- cur in hemaphroditic fishes will increase the risk of incorrectly estimating optimal exploitation levels from yield-per-recruit models (Bannerot et al. 1987). The problems can be compounded by disregarding the in- creased size variability within the system resulting from sexual transformations. Distributed delay mod- els provide a mechanism for handling length-specific life-history changes, and results can be summarized as an aggregate of individuals and the associated vari- ability rather than a simple mean estimate without error. The consequence is greater flexibility in dealing with changes between length categories without ignor- ing the size variability inherent within the system. In the black sea bass example presented, the incor- poration of a transitional term in the delay model had little effect on the estimate of yield-per-recruit. Since the sexes were combined to produce total yield, the growth differences between sexes were not large enough to produce significant changes in overall yield. Also, the inclusion of fishing mortality quickly removed any added yield contributed by transformed females. It is likely that the result would diverge further from a traditional model estimate as the differences in lon- gevity or growth rate between the sexes increased. The effect of modeling length- and sex-specific characteristics with a delay model was more appar- ent in spawning stock biomass-per-recruit estimates. The removal of female biomass via transformations was similar to increasing natural mortality at each length step, and consequently the spawning poten- tial was reduced more quickly than under typical effects of fishing mortality and a constant M of 0.2. The system was more sensitive to changes in size- at-recruitment to the fishery than reduction in F at a constant size-at-recruitment. The differences in spawning biomass-per-recruit es- timates resulting from the distributed delay model have important management implications. If management Shepherd and Idome Yield- and spawning biomass-per-recruit for Centropnstis striata 335 120 100 Q. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 FISHING MORTALITY Figure 9 Percentage of maximum potential spawning biomass-per-recruit for black sea bass Centropnstis striata under various fishing mor- talities and lengths-at-first-capture (L(l using the distributed delay model with transitions (dashed) and without transitions ( solid K Table 3 SSB/R for black sea bass Centropristis striata as percentage of original values (a nd L=16cm). wh ere transformation rates have been doubled in differing 8 cm ntervals. F 8-15 cm 16— 23 cm 24-31 cm 32-39 cm 0 99.7 39.1 69.5 96.3 0.1 99.7 43.3 74.0 97.2 0.2 99.7 47.9 78.6 98.0 0.3 99.7 52.8 ^82.9 98.6 0.4 99.7 57.6 86.6 99.1 0.5 99.7 62.1 89.6 99.4 0.6 99.8 66.1 92.0 99.6 0.7 99.8 69.7 93.9 99.8 0.8 99.8 72.9 95.3 99.9 0.9 99.8 75.7 96.4 99.9 1.0 99.8 78.2 97.2 99.9 1.1 99.8 80.4 97.9 100.0 1.2 99.8 82.4 98.4 100.0 1.3 99.8 84.2 98.7 100.0 1.4 99.9 85.8 99.0 100.0 1.5 99.9 87.2 99.2 100.0 600 500 0)400 \ \ 24-31 cm DC m W 300 \\ original \\ 32-39 cm 200 \ >\ 8-15 cm 100 16-23 cm 0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 FISHING MORTALITY Figure 10 SSB/R for black sea bass Centropristis striata estimated under varying transition rates-at- length (L, = 16 cm). Rates were increased 209r within each length-group indicated. were based on a target %MSP, the presence or absence of transition terms in the model would provide dra- matically different interpretations. For instance, if the target % MSP were 30% with an F=0.8, and Lc=25 cm, the non-transitional model would indicate the situa- tion was below the target at 22%, whereas the transi- tional model would place the current conditions above the target at 34%. It would be critical that calculation of SSB in the stock-recruitment relationship used for development of a target SSB/R (and associated %MSP) be made using a framework similar to calculation of SSB/R. Beyond the issue of the model structure is the appro- priateness of using spawning biomass models in general for management of hermaphroditic fishes. The complex social hierarchy of reef fishes (black sea bass can be considered a temperate reef fish ) during spawning im- plies that the number of males may be an important factor limiting reproductive potential (Smith 1982). Re- cent theoretical studies suggest that males are not limit- ing in hermaphroditic sea bass populations to the degree that non-dominant males participate in spawning ("streakers") (Peterson 1991). If streakers are abundant, there would be no benefit for a female to transform only to engage in sperm competition with other spawning males. This suggests that under limited exploitation, the reproductive potential of the population is restricted in 336 Fishery Bulletin 91(2), 1993 terms of egg production. The possibility exists that ex- ploitation could reach a level at which males are elimi- nated at such a rate that females are forced to transform faster or at an earlier size. The optimal sex ratio, from a behavioral perspective, has not yet been determined for C. straita. As suggested by the sensitivity analysis, sub- stantial changes in the transformation rate could signifi- cantly alter the contributions of a cohort to the spawning biomass. There is no information currently available to determine the controlling mechanism or degree of plas- ticity for transformation rates within a population. Future refinements of this distributed delay model may provide a more detailed picture of the effects of harvest on species such as black sea bass. Density- dependent feedback mechanisms influencing transi- tion rates and age-at-first-transformation will ulti- mately be incorporated into the model to examine theoretical implications. A time-varying distributed delay model is also possible to account for seasonal changes in growth and mortality. With a minimal amount of growth information, either from back-cal- culations or tagging data, and some reasonable esti- mates of other important life-history events, we be- lieve the distributed delay model can provide an effective length-based alternative to traditional dy- namic pool models. Citations Alexander, M. S. 1981 Population response of the sequential hermaphro- dite black sea bass, Centropristis striata, to fishing. M.S. thesis, State Univ. New York, Stony Brook, 104 p. Bannerot, S., W. W. Fox Jr., & J. E. Powers 1987 Reproductive strategies and the management of snappers and groupers in the Gulf of Mexico and Caribbean. In Polovina, J. J., & S. Ralston (eds.), Tropical snappers and groupers, biology and fisher- ies management, p. 561-603. Westview Press, Boulder. Beverton, R. J. H., & S. H. Holt 1957 On the dynamics of exploited fish populations. Fish. Invest, Ser. II, Mar. Fish., G. B. Minist. Agric. Fish. Food 19, 533 p. Forrester, J. W. 1961 Industrial dynamics. MIT Press, Cambridge, 464 p. Gabriel, W. L., M. P. Sissenwine, & W. J. Overholtz 1989 Analysis of spawning on biomass per recruit: an example for Georges Bank haddock. N. Am. J. Fish. Manage. 9:383-391. Gulland, J. 1977 The analysis of data and development of models. In Gulland, J. (ed.), Fish population dynam- ics, p. 67-95. John Wiley, NY Gutierrez, A. P., M. A. Pizzamiglio, W. J. Dos Santos, R. Tennyson, & A. M. Villacorta 1984 A general distributed delay time varying life table plant population model: cotton {Gossypium hirsutum L. ) growth and development as an example. Ecolog. Model. 26(1984): 231-249. Gutierrez, A. P., B. Wermelinger, F. Schulthess, J.U. Baumgaertner, H. R. Herren, C. K. Ellis, & J. S. Yaninek 1988 Analysis of biological control of cassava pests in Africa. I. Simulation of carbon, nitrogen and water dynamics in cassava. J. Appl.Ecol. 25:901-920. Idoine, J. S., & J. T. Finn 1984 A model of lobster growth. In Wade, W. D. (ed.), Proc, 1984 Summer Comput. Simul. Conf., p. 887- 891. Soc. Comput. Simul., Boston. Kirkpatrick, M. 1984 Demographic models based on size, not age, for organisms with indeterminate growth. Ecology 65(61:1874-1884. Lavenda, N. 1949 Sexual differences and normal protogynous her- maphroditism in the Atlantic sea bass, Centropristes striatus. Copeia 3:185-194. Low, R. A. Jr. 1981 Mortality rates and management strategies for black sea bass off the southeast coast of the United States. N. Am. J. Fish. Manage. 1:93-103. Manetsch, T. J. 1976 Time-varying distributed delays and their use in aggregative models of large systems. IEEE (Inst. Electr. Electron. Eng.) Trans. Syst. Man Cybern. 6:547-553. Mercer, L. P. 1978 The reproductive biology and population dybnamics of black sea bass, Centropristis straita. Ph.D. thesis, Coll. William & Mary, Williamsburg VA. O'Brien, L., J. Burnett, & R. K. Mayo 1993 Maturation of nineteen species of finfish off the northeast coast of the United States, 1985- 1990. NOAA Techn. Rep. NMFS 113, 67 p. Petersen, C. W. 1991 Sex allocation in hermaphroditic sea basses. Am. Nat. 138(31:650-667. Ravlin, F. W., R.I. Carruthers, V. Varadarajan, D. L. Haynes, & R. L. Tummala 1978 Simulation of a natural biological system in which two insect populations are attacked by a common parasite. Proc, Ninth. Annu. Pittsburgh Symp. Model. Simul., p. 117-123. Univ. Pittsburgh, Pitts- burgh PA. Ricker, W. E. 1954 Stock and recruitment. J. Fish. Res. Board Can. 11:559-623. Russell, E. S. 1931 Some theoretical considerations on the "overfish- ing problem". J. Cons. Explor. Mer 6:3-27. Shepherd and Idoine: Yield- and spawning biomass-per-recruit for Centropnstis striata 337 Schaub, L. P., & J. U. Baumgartner 1989 Significance of mortality and temperature on the phenology of Orthotylus marginalis (Heteroptera: Miridae). Mitt. Schweitz. Entomol. Ges. 62:235-245. Smith, C. L. 1982 Patterns of reproduction in coral reef fishes. In Huntsman, G. R., W.R. Nicholson, & W. W. Fox Jr. (eds.), The biological bases for reef fishery manage- ment, p. 49-66. NOAA Tech. Memo. NMFS-SEC-80, NMFS Southeast Fish. Sci. Cent., Miami FL. Thompson, W. F., & F. H. Bell 1934 Biological statistics of the Pacific halibut fish- ery. 2. Effects of changes in intensity upon total yield and yield per unit gear. Rep. Int. Halibut Comm., 49 p. Vansickle, J. 1977 Attribution in distributed delay models. IEEE (Inst. Electr. Electron. Eng. ) Trans. Syst. Man/Cybern. 7(9):635-638 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. 84:723-741. Abstract.— Changing temporal and spatial distribution patterns of Atlantic herring Clupea harengus larvae collected off southern New England over two decades provided early signals of large-scale changes in adult spawning biomass that are now a matter of record. Four con- trasting spawning patterns were evi- dent during the 20 yr period. Each pattern covered successive multi- year intervals and reflected the cor- responding status of the adult popu- lation. In 1971, spawning occurred throughout the Georges Bank/Nan- tucket Shoals/Massachusetts Bay study area. The principal spawning grounds of herring in the Gulf of Maine region were located on the Northeast Peak of Georges Bank. With the collapse of the Georges Bank fishery in 1976, spawning re- ceded westward to Nantucket Shoals. By 1979, larvae occurred only in the Stellwagen Banks area of Massachusetts Bay, the smallest of the three subareas. After a 6yr hiatus, spawning beds on Nantucket Shoals were reoccupied in 1985. By 1988 spawning had advanced east- ward to Cultivator Shoals on Georges Bank, but through 1990 we found no evidence of renewed spawn- ing activity on the historically-promi- nent spawning beds on Northeast Peak. The rebuilding process was at- tributed to recolonization rather than resurgence. Larval distribution patterns: Early signals for the collapse/recovery of Atlantic herring Clupea harengus in the Georges Bank area Wallace G. Smith Wallace W. Morse Sandy Hook Laboratory. Northeast Fisheries Science Center National Marine Fisheries Service, NOAA Highlands. New Jersey 07732 Manuscript accepted 8 February 1993. Fishery Bulletin, U.S. 91:338-347(1993). 338 The Atlantic herring Clupea harengus has been a target species of broad-scale early-life-history research in the Gulf of Maine since 1956. The initial larval surveys were part of a collaborative study by U.S. and Canadian biologists to define the stock structure of her- ring in the Northwest Atlantic (Tibbo et al. 1958). This effort continued over two spawning seasons and was fol- lowed by annual observations on dis- tribution patterns of herring larvae during 1962-70 (Boyar et al. 1973b). Thereafter, U.S. biologists participated in a multinational field program that represented the starting point of the Northeast Fisheries Science Center's (NEFSC) current database for larval herring. This study, sponsored by the International Commission for the Northwest Atlantic Fisheries (ICNAF), was initiated in 1971 to identify mecha- nisms that influence survival of larvae (Lough etal. 1985). With the termination of ICNAF lar- val herring surveys in 1977, NEFSC began a comprehensive fisheries eco- system study of the Northeast Shelf Ecosystem (Sherman 1986, Smith 1988). This program, known as Ma- rine Resources Monitoring, Assessment and Prediction (MARMAP), included standardized year-round surveys of fish eggs and larvae to provide baselines against which shifts in species compo- sition and diversity within the finfish community could be observed and evaluated. These surveys continued for 11 years, ending in December 1987. In 1988, ecosystems research em- phasis at NEFSC shifted away from year-round surveys of the entire Northeast Shelf Ecosystem and fo- cused on the Georges Bank area dur- ing autumn and winter. The new ini- tiative was designed to document the changing status of herring and in- vestigate density-dependent popula- tion regulation between herring and sand lance Ammodytes spp., impor- tant coastal pelagic species in the western North Atlantic. Wlien com- bined, the three multi-year research endeavors conducted since 1971 pro- vide an uninterrupted 20 yr database for larval herring in the Georges Bank/Nantucket Shoals/Massachu- setts Bay area. During this period, dramatic changes occurred in the her- ring population. Spawning biomass was relatively high at the onset of the ICNAF lar- val herring surveys in 1971, although in decline due to intense foreign fishing pressure that began in th.e mid 1960s. During the 1960s and con- tinuing through the mid 1970s, bio- mass of herring on Georges Bank was estimated at 400,000-600,000 1. An- nual catches ranged from 150,000 to 374,000 1. Dominating the population during the 1960s were three year- classes, 1956, 1960, 1961, and to a lesser extent 1966, although spawn- ing success was poor in the closing years of the decade. Despite a strong 1970 year-class that entered the Georges Bank fishery in 1973, fishing Smith and Morse Larval distribution patterns of C/upea harengus "on the Georges Bank 339 pressure by the large international fleet was so in- tense that landings declined after 1971. From 1973 to 1975 the principal herring landings shifted from the bank to the vicinity of Great South Channel, just east of Nantucket Shoals. After 15yr of moder- ate-to-heavy exploitation, and several years of quota management under ICNAF, the Georges Bank her- ring fishery collapsed in 1976 (Anthony & Waring 1980). This paper examines the 20 yr larval database to look for early signals of the dramatic changes in adult biomass that have occurred since 1971. We evaluate the influence of transport mechanisms on the spatial and temporal distribution patterns of lar- vae to corroborate the herring stock hypothesis of lies & Sinclair (1982) and Sinclair & lies (1985). Finally, we search for supporting evidence that the late-1980s recovery of herring in the Georges Bank area resulted from population resurgence rather than recolonization, a conclusion of Stephenson & Kornfield(1990). Methods The 20 yr time-series began during the final years of the international fishery. It included a 10 yr period when essentially no fishing took place in the study area, and ended with the 1990 spawning season when a limited fishery was again underway. Without a di- rected fishery after 1976, this dataset, along with NEFSC semi-annual trawl surveys, provided the only source of information to evaluate the status of herring in the Georges Bank area for more than a decade. Sampling designs All three larval surveys used the 61cm bongo, fitted with 0.333 and 0.505 mm mesh nets. Larval herring catches from the 0.505 mm net were analyzed. Net mesh selection was based on recommendations in Smith & Richardson (1977) and on Colton et al. (1980) who found no significant difference in retention rates of herring larvae captured in the two mesh sizes. ^K M ee E^ • • _ _• * ' 'x ,f'' ft:*..% MB. • > • "„ 42 § ■■ m. .i ' ; gb; • • * • •' >> • ' ' NS. " • • • ■ • • • • • • * * . r " • ••••• • ••••• 0 60 , ,''" v — -;*-\ #i ,»•.'""" a 0 i. n w 88 "sx 08 L ' , • ,-*'"-- « 1 K- "*\ , % m • *i BJ^V** • • • ' • ', WT& -a . • . • ? > • • • • • • • • • • ^/ •' l< Hamsters 41' • , p=p^ -■' •" *- - i. i b Figure 1 Station plan for (a) ICNAF larval herring surveys in the Georges Bank area, 1971-76, (b) MARMAP surveys in Georges Bank, 1977- 87. and (c) studies of herring/sand lance interactions, 1988-90. (d) General surface circulation in the Georges Bank area (after Ingham 1982). 340 Fishery Bulletin 9 1(2). 1993 The ICNAF larval herring program included the entire Gulf of Maine region but concentrated on the Nantucket Shoals/Georges Bank area (Lough & Bolz 1979a, Lough et al. 1985). The 33 surveys conducted between 1971 and 1976 were used in our analysis. ICNAF participants deployed the bongo at 50m/min to a maximum depth of 100 m and retrieved it in a smooth oblique profile at lOm/min. Towing speed was 3.5 kn at stations spaced at 28-37 km intervals on a standard grid pattern (Fig. la). Additional stations in areas of high larval abundance were sampled after 1974. Lough & Bolz (1979b) provided a detailed ac- count of sampling operations and cruises for the ICNAF time-series. MARMAP surveys, conducted at monthly to bi- monthly intervals, covered the continental shelf from Cape Hatteras, North Carolina to Cape Sable, Nova Scotia. Cruise activities were described by Sibunka & Silverman (1984, 1989). MARMAP surveys fished the 61cm bongo to a maximum depth of 200 m with payout at 50m/min and retrieval at 20m/min. Ship speed var- ied between 1 and 2 kn to maintain a 45° angle in the towing wire. Station intervals ranged from 15 to 45 km (Fig. lb). The herring recovery study initiated in 1988 em- ployed MARMAP sampling methods. Cruises were con- ducted at about monthly intervals and occupied 151 stations in the Nantucket Shoals/Massachusetts Bay/ Georges Bank study area (Fig. lc). Data analysis We adjusted the number of larvae caught at each sta- tion in the time-series to reflect day/night/twilight catch differences (see Morse 1989). Larvae were partitioned into size-intervals to depict changes in distribution pat- terns over time. When the partitioned size-intervals of 4.0-7.9 mm, 8.0-12.9 mm, 13.0-17.9 mm, and >17.9mm were corrected for shrinkage (see Theilacker 1980), they approximated age-groups of <2 wk, 2-5 wk, 5-8 wk, and >8wk, respectively (Lough et al. 1982). Surveys were grouped by multi-year periods that represented time-intervals between the most apparent changes in spawning patterns during the 20 yr period. Figures showing mean numbers of larvae per 10 m2 surface area were based on methods recommended for the delta distribution (Pennington 1983). Distributions of lar- vae grouped by age over time are discussed in relation to circulation patterns shown in Fig. 1 d. Results Changes in distribution and abundance of herring lar- vae during the 20 yr time-series reflected four contrast- ing spawning patterns that spanned successive multi- year intervals. Larvae were abundant on both Nan- tucket Shoals and Georges Bank, but not so in Massa- chusetts Bay, during 1971-75. Conversely, we found low to no measurable numbers of larvae on Nantucket Shoals and Georges Bank during 1976-84, a period when spawning activity increased in Massachusetts Bay. Spawning resumed on Nantucket Shoals in 1985. During 1985-87, larval abundance peaked in Massa- chusetts Bay but the center of spawning activity shifted to Nantucket Shoals. Herring were again spawning on the western half of Georges Bank in 1988. The dramatic increase in spawn- ing activity on Nantucket Shoals and renewed spawn- ing on western Georges Bank during 1988-90 elevated larval abundance estimates to their highest levels dur- ing the closing years of the time-series (Fig. 2). When the time-series began in 1971, Georges Bank had been the principal spawning grounds for herring INT. 1 INT. 2 INT. 3 INT 4 CO 3 j- c\j co *r in o -— ryco^mcof^-coojo .-^^cocpcococooococococoo) aia>o>oioo)OiO)0)ooCT)a>o)cj>o> YEAR Figure 2 Changes in abundance of Atlantic herring Clupea haivni;us larvae by subarea during the 20yr time-series. Intervals i Int i represent multi-year periods that reflect the most apparent changes in spawning patterns. Smith and Morse Larval distribution patterns of Oupea harengus on the Georges Bank 341 <2 weeks 2 to 5 weeks 5 to 8 weeks """I ?o Figure 3 Composite representation of the distribution of Atlantic herring Clupea harengus larvae by age in the Georges Bank area, 1971-75. in the western North Atlantic for at least 15yr (Tibbo et al. 1958, Boyar et al. 1973b, Anthony & Waring 1980, Lough et al. 1985, Grosslein 1987). The magni- tude of spawning activity, as measured by the abun- dance of larvae, was low in 1971 and 1972 but in- creased sharply in 1973 and continued upward in 1974, the initial spawning years of the strong 1970 year- class. In 1975, larval abundance on Georges Bank de- clined dramatically and spawning activity on Nantucket Shoals equaled that on the bank. By 1976, when sur- veys showed only negligible spawning activity east of the 69° meridian, Georges Bank had lost its longstanding status as the principal spawning ground for herring in the western North Atlantic (Fig. 2). ICNAF surveys of 1971-75 identified two geographi- cally-separate spawning areas: one centered along the eastern edge of Nantucket Shoals, the other on north- eastern Georges Bank (Fig. 3). Within 2-5 wk of hatch- ing, larvae began to disperse and the two spawning centers were no longer discrete. At 5-8 wk after hatch- ing, some larvae originating on the Northeast Peak were transported westward across the southern half of Georges Bank by anticyclonic currents. Together with larvae from Nantucket Shoals, their distribution ex- tended over shelf waters from Cape Cod eastward and continued into adjacent slope waters. Although distri- butions of both Nantucket Shoals and Georges Bank larvae exhibited the influence of known circulation pat- terns within 5 wk of hatching (Fig. Id), centers of abun- dance remained near their points of origin for at least 8wk. Thereafter, a single center of abundance for lar- vae >8 wk old emerged over the shallow central part of Georges Bank (Fig. 3). Herring larvae in all four age-groups occurred largely in coastal waters immediately adjacent to Cape Cod during the 1976-84 time-interval, a distribution pat- tern that differed significantly from those observed dur- ing the first 5yr of the time-series (Fig. 4). We found no larvae <2wk old on Georges Bank in 1977. In 1978, only a few larvae in the youngest grouping were caught over the once-productive Northeast Peak. For the next decade, we caught no recently-hatched larvae anywhere on Georges Bank. Larvae <2wk old were essentially absent on Nantucket Shoals during the 1976-84 time- interval as well. The center of larval abundance in each of the four groupings occurred in Massachusetts 342 Fishery Bulletin 91(2), 1993 £" <2 weeks fc. ---"I 5 to 8 weeks 2 to 5 weeks k^ &- >8 weeks Figure 4 Composite representation of the distribution of Atlantic herring Clupea harengus larvae by age in the Georges Bank area, 1976-84. and Cape Cod bays. As during the 1971-75 interval, larvae partitioned by age exhibited the influence of transport, especially in 1981 and 1982 when distribu- tion patterns expanded from Massachusetts Bay to Nantucket Shoals. Averaging catches over a 9yr period did not mask the effects of larval transport dur- ing the 1981 and 1982 spawning seasons (Fig. 4). Further change marked the ensuing 3yr period (1985-87), providing the first evidence that the her- ring population in the study area was beginning to recover. Spawning beds on Nantucket Shoals were re- activated in 1985, and the center of larval abundance shifted from Massachusetts Bay to Nantucket Shoals (Fig. 5). The influences of advective processes on the distribution of larvae were evident within a month after hatching. Nantucket Shoals larvae 2-5 wk old were transported by the Georges Bank gyre eastward across the northern part of the bank as far as the 68° meridian. The distribution pattern of 5-8 wk old lar- vae resembled that of 2-5 wk old fish, but exhibited the influence of further drift away from the Nantucket Shoals spawning beds. Larvae >8 wk old were caught in Massachusetts Bay and from Nantucket Shoals east- ward onto Georges Bank. As in the 1976-84 period, no recently-hatched larvae occurred over the traditional spawning beds on eastern Georges Bank through 1987 (Fig. 5). The reoccupation of spawning beds on the western half of Georges Bank and the increasing abundance of larvae on Nantucket Shoals during 1988-90 provided further signals of the changing status of herring in the study area (Fig. 6). Although the principal spawning grounds remained on Nantucket Shoals, the appear- ance of recently-hatched larvae on Georges and Culti- vator shoals in all 3yr provided the first evidence in a decade of spawning east of the 68° meridian on Georges Bank. Within 5 wk of hatching, larvae were dispersed over all but the eastern tip of the bank. By the time larvae reached 8 wk of age, they occurred throughout the study area (Fig. 6). As with the reactivation of Nantucket Shoals spawning beds in 1985, 3yr lapsed from the time we first observed larval transport onto Georges Bank until we found evidence of spawning on the bank. Although the recovery of herring in the Georges Bank region was clearly underway during the closing years of the 1980s, we found no evidence in Smith and Morse Larval distribution patterns of Oupea harengus or\ the Georges Bank 343 <2 weeks 5 to 8 weeks ■€■ 2 to 5 weeks ■8 weeks Figure 5 Composite representation of the distribution of Atlantic herring Clupea harengus larvae by age in the Georges Bank area, 1985-87. larval distribution patterns through the 1990 spawn- ing season that the historically prominent spawning beds on eastern Georges Bank had been reoccupied. Discussion The literature contains conflicting descriptions of the herring population structure in the western North At- lantic, lies & Sinclair (1982) hypothesized that the number of Atlantic herring spawning stocks in the Gulf of Maine region is determined by the number of geographically-stable larval retention areas. They rec- ognized four stocks: southwest Nova Scotia, Grand Manan, Georges Bank, and Nantucket Shoals. Sinclair & lies (1985) revised the stock structure to include Jeffreys Ledge, coastal Gulf of Maine, and Scots Bay Both lies & Sinclair (1982) and Sinclair & lies (1985) further hypothesized that (1) stocks are segregated only during autumn when herring return to their respec- tive spawning grounds, (2) larvae of different stocks do not intermix for several months after hatching, the period when imprinting occurs, and (3) hydrographic conditions create retention areas by isolating larvae in vertically-mixed areas surrounded by stratified waters. Their retention hypothesis recognized that the geo- graphic extent of the retention areas could vary annu- ally and that some horizontal displacement could oc- cur during larval development. Grosslein ( 1987) reviewed the stock structure of her- ring in the Gulf of Maine region using the collective evidence from prior studies. He concluded that the population was composed of three major stocks: Georges Bank, western Nova Scotia, and the Gulf of Maine, with lesser spawning occurring elsewhere around the gulf. Grosslein indicated, however, that this breakdown was speculative and that direct attempts to differenti- ate between stocks through larval surveys, tagging studies, meristic and morphometric characters, bio- chemical methods, and parasitology produced incon- clusive results. Smith & Jamieson (1986) argued that no scientific basis existed for the stock concepts of Atlantic herring in either the eastern or western North Atlantic. They viewed the breakdown of herring stocks as transient subdivisions having no taxonomic or evolutionary 344 Fishery Bulletin 91(2). 1993 <2 weeks 5 to B weeks ■€■ n -4 70 Ea \ jdlj|§jjjL -,. EE <- ■*2 : fT - , m jfflmMliiii'l " ;' 'n 4? ^pl^* C\i -, ' &> mm* i rf*lil?r 40 kO.vJpP^'" 40- ,72 2 to 5 weeks 70 66 SB Figure 6 Composite representation of the distribution of Atlantic herring Clupea harengus larvae by age in the Georges Bank area, 1988-90. status. Cushing (1986) also took exception to the re- tention hypothesis as applied by lies & Sinclair (1982) to the North Sea area. Cushing ( 1986) noted that her- ring do not always spawn in transition zones around the British Isles and, where they do, the fronts brake down within a month after hatching. He further dis- missed the retention concept by demonstrating that postlarval metamorphosis takes place near the nurs- ery grounds after prolonged larval drift. Our analysis of the 20 yr larval time-series supports that segment of the literature showing herring larvae vulnerable to advection after hatching (e.g., Cushing 1986, Bartsch et al. 1989, Heath & Walker 1987). Al- though transport in our dataset is not extensive, it is not restricted to the retention areas described by Sinclair & lies (1985). Larvae drift from Massachu- setts Bay southward to Nantucket Shoals, and the an- ticyclonic gyre on Georges Bank mixes larvae originat- ing there with those spawned on the shoals. The direction of transport generally corresponds with re- ported circulation, a pattern consistent with that of Parrish et al. (1981). Mixing of larvae from the three subareas occurs within weeks after hatching, or ear- lier in the life cycle than proposed by Sinclair & Tremblay ( 1984) and Sinclair & lies (1985). Drift trajectories of herring larvae from other stud- ies in the Gulf of Maine region agree with our findings in that they correspond to dominant circulation pat- terns described by Ingham (1982) and Butman & Beardsley (1987). Graham et al. (1972) and Graham (1982) recognized that some herring larvae spawned off eastern Maine drifted westward to mix with larvae originating in central and western Maine waters. Townsend et al. (1986) provided further evidence that herring larvae off Maine are advected from east to west. Although Chenoweth et al. (1989) noted that re- sidual currents do not demonstrate a persistent flow in any direction off eastern Maine, they too observed a westward displacement of herring larvae. Finally, Boyar et al. (1973a) showed herring larvae spawned on Jeffreys Ledge, another of the retention areas of Sinclair & lies (1985), drifting into Massachusetts Bay where they could mix with larvae originating on and around Stellwagen Bank. Given these reports of drift and mixing in western Gulf of Maine waters near our study area and the demonstrated dispersion and Smith and Morse Larval distribution patterns of Oupea harengus "on the Georges Bank 345 mixing of early-stage larvae reported here, we see no compelling argument for the separation of herring stocks through larval retention in the western Gulf of Maine and Georges Bank areas. We consistently found sufficient larval drift and mix- ing to question the stock structure of Sinclair & lies ( 1985) in the Georges Bank region. However, Figs. 3-6 lend some support to the retention hypothesis by show- ing that a significant portion of the larvae up to 8wk old remained near their points of origin. Campana et al. (1989a, 1989b) reported similar results for haddock Melanogrammus aeglefinus and Atlantic cod Gadus morhua larvae, respectively, on Browns Bank. Some larvae of the two species were transported shoreward while others remained near the spawning grounds longer than known water residence time. Both of the 1989 studies concluded that the instability of the Browns Bank gyre accounted for the dichotomous dis- tribution patterns. They interpreted the onshore drift as an all-or-none situation for individual larvae. Some readily exited the influence of a weakly established gyre, while most were retained during periods of pro- nounced gyral circulation. In our view, their theory provides a more plausible explanation for the shifting distribution patterns of herring larvae around Georges Bank, an area also influenced by gyral circulation, than the stock hypothesis proposed by lies & Sinclair ( 1982). Stephenson & Kornfield (1990) attributed the in- creasing biomass of herring on Georges Bank during the late 1980s to the resurgence of residual fish rather than to recolonization by fish from surrounding areas. Based largely on a 4yr dataset that excluded Nan- tucket Shoals and Massachusetts Bay, they concluded that (1) adult herring caught on Georges Bank dif- fered in age composition from herring on Jeffreys Ledge and southwestern Nova Scotia, (2) some of the Georges Bank herring exhibited different isoenzyme character- istics than neighboring herring, and (3) the return of herring to Georges Bank took longer than they ex- pected through recolonization. In our view, the change in spawning sites on Georges Bank during the 20 yr time-series, coupled with the recession and subsequent expansion of the spawning range over time, makes a convincing case for recoloniz- ation rather than resurgence. Larval distribution pat- terns on Georges Bank during the closing years of the time-series differed markedly from those observed at the outset of the program. Whereas recently-hatched larvae were concentrated on Northeast Peak during the initial survey years, their distribution on the bank progressed eastward only as far as Georges Shoals during the final years of survey activity. If, as pro- posed by Stephenson & Kornfield (1990), resurgence of residual fish accounted for the recovery of spawning herring on the bank, we would expect little or no change in the location of spawning centers over time. Instead, our 20 yr time-series shows dramatic change. When the herring population in the study area de- clined in the early 1970s, its spawning range receded from Georges Bank to Massachusetts Bay. As the popu- lation began to grow in the 1980s, the spawning range expanded, initially from Massachusetts Bay to Nan- tucket Shoals, then from the shoals to Georges Bank. Whether the recovery resulted from an expanding adult population, from fortuitous transport and survival of larvae, or from a combination of the two, is not clear. However, it is worth noting that the reoccupation of spawning beds on both Nantucket Shoals and Georges Bank occurred 3-4 yr after we first observed evidence of larval drift towards these two subareas. This time period, perhaps coincidently, represents the age-at- maturity for Atlantic herring. Acknowledgments Although too numerous to mention individually, we herewith acknowledge everyone involved in the collec- tion of the 8590 plankton samples that form the basis of this document; our colleagues at the Polish Sorting Center in Szczecin for their diligence and persever- ance in sorting the samples; the technicians and biolo- gists responsible for implementing quality control, data entry, and database management procedures; and those who reviewed earlier drafts of the manuscript and pro- vided constructive comment. Citations Anthony V., & G. Waring 1980 The assessment and management of the Georges Bank herring fishery. Rapp. P.-V. Reun. Cons. int. Explor. Mer 177:72-ill. Bartsch, J., K. Brander, M. Heath, P. Munk, K. Richardson, & E. Svendsen 1989 Modelling the advection of herring larvae in the North Sea. Nature ( Lond. ) 340:632-636. Boyar, H., R. Cooper, & R. Clifford 1973a A study of the spawning and early life history of herring iClupea harengus harengus L.) on Jeffreys Ledge in 1972. Int. Comm. Northwest Atl. Fish. Res Doc. 73/96, 27p. Boyar, H., R. Marak, F. Perkins, & R. Clifford 1973b Seasonal distribution and growth of larval her- ring iClupea harengus L.) in the Georges Bank-Gulf of Maine area from 1962-1970. J. Cons. Cons. Int. Explor Mer 35:36-51. Butman, B., & R. Beardsley 1987 Physical oceanography. In Backus, R., & D. Borne (eds.), Georges Bank, p.88-98. MIT Press, Cambridge. 346 Fishery Bulletin 91(2). 1993 Campana, S., S. Smith, & P. Hurley 1989a A drift retention dichotomy for larval haddock (Melanogrammus aeglefinus) spawned on Browns Bank. Can. J. Fish. Aquat. Sci. 46 (Suppl. 1): 93-102. 1989b An age-structured index of cod larval drift and re- tention in the waters of southwest Nova Scotia. Rapp.- V Reun. Cons. Int. Explor. Mer 191:50-62. Chenoweth, S., D. Libby, R. Stephenson, & M. Powers 1989 Origin and dispersion of larval herring (Clupea harengus) in coastal waters of eastern Maine and southwestern New Brunswick. Can. J. Fish. Aquat. Sci. 46:624-632. Cushing, D. 1986 The migration of larval and juvenile fish from spawning grounds to nursery grounds. J. Cons. Cons. Int. Explor. Mer 43:43-49. Colton, J., J. Green, R. Byron, & J. Frisella 1980 Bongo net retention rates as effected by towing speed and mesh size. Can. J. Fish. Aquat. Sci. 37:606-623. Graham, J. 1982 Production of larval herring, Clupea harengus, along the Maine coast. J. Northwest Atl. Fish. Sci. 3:63-85. Graham, J., S. Chenoweth, & C. Davis 1972 Abundance, distribution, movements and lengths of larval herring along the western coast of the Gulf of Maine. Fish. Bull., U.S. 70:307-32 1 . Grosslein, M. 1987 Synopsis of knowledge of the recruitment process for Atlantic herring, Clupea harengus, with special reference to Georges Bank. NAFO (Northwest Atl. Fish. Organ.) Sci. Counc. Stud. 11:91-108. Heath, M., & J. Walker 1987 A preliminary study of the drift of larval herring (Clupea harengus L.) using gene-frequency data. J. Cons. Cons. Int. Explor. Mer 43:139-145. lies, T., & M. Sinclair 1982 Atlantic herring: stock discreetness and abundance. Science ( Wash. DC ) 215:627-633. Ingham, M. (editor) 1982 Summary of the physical oceanographic processes and features pertinent to pollution distribution in the coastal and offshore waters of the northeastern United States, Virginia to Maine. NOAA Tech. Memo. NMFS-F/NEC-17, NMFS Northeast Fish. Sci. Cent, Woods Hole MA, 166 p. Lough, R., & G. Bolz 1979a Abundance of sea herring (Clupea harengus L.) larvae in relation to spawning stock size and re- cruitment for the Gulf of Maine and Georges Bank, 1968-1978. Ref. Doc. 79-50, NMFS Woods Hole Lab, 13 p. 1979b A description of the sampling methods, and lar- val herring (Clupea harengus L. ) data report for sur- veys conducted from 1968-1978 in the Georges Bank and Gulf of Maine areas. Ref. Doc. 79-60, NMFS Woods Hole Lab., 23 p. Lough, R., M. Pennington, G. Bolz, & A. A. Rosenberg 1982 Age and growth of larval Atlantic herring, Clupea harengus L., in the Gulf of Maine-Georges Bank re- gion based on otolith growth increments. Fish. Bull, U.S. 80:187-199. Lough, R., G. Bolz, M. Pennington, & M. Grosslein 1985 Larval abundance and mortality of Atlantic her- ring (Clupea harengus L.) spawned in the Georges Bank and Nantucket Shoals areas, 1971-78 seasons, in relation to spawning stock size. J. Northwest Atl. Fish. Sci. 6:21-35. Morse, W. 1989 Catchability, growth and mortality of larval fishes. Fish. Bull., U.S. 87:417-446. Parrish, R., C. Nelson, & A. Bakun 1981 Transport mechanisms and reproductive success of fishes in the California Current. Biol. Oceanogr. K2C175-203. Pennington, M. 1983 Efficient estimators of abundance for fish and plankton surveys. Biometrics 39:281-286. Sherman, K. 1986 Measurement strategies for monitoring and fore- casting variability in large marine ecosystems. In Sherman, K., & L. Alexander (eds.), Variability and management of large marine ecosystems, p. 203- 236. AAAS Sel. Symp. 99. Sibunka, J., & M. Silverman 1984MARMAP surveys of the continental shelf from Cape Hatteras, North Carolina to Cape Sable, Nova Scotia (1977-1983). Atlas 1. Summary of operations. NOAA Tech. Memo. NMFS-F/NEC-33, NMFS Northeast Fish. Sci. Cent., Woods Hole MA, 306 p. 1989MARMAP surveys of the continental shelf from Cape Hatteras, North Carolina to Cape Sable, Nova Scotia ( 1984-1987 ). Atlas 3. Summary of operations. NOAA Tech. Memo. NMFS-F/NEC-68, NMFS Northeast Fish. Sci. Cent., Woods Hole MA, 197 p. Sinclair, M., & T. lies 1985 Atlantic herring (Clupea harengus) distribution in the Gulf of Maine - Scotian Shelf area in relation to hydrographic features. Can. J. Fish. Aquat. Sci. 42:880-887. Sinclair, M., & M. Tremblay 1984 Timing of spawning of Atlantic herring (Clupea harengus harengus) populations and the match-mis- match theory. Can. J. Fish. Aquat. Sci. 41:1055-1065. Smith, P., & A. Jamieson 1986 Stock discreetness in herring: a conceptual revolution. Fish. Res. (Amst.) 4:223-234. Smith, P., & S. Richardson 1977 Standard techniques for pelagic fish egg and lar- val surveys. FAO Fish. Tech. Pap. 175, 100 p. Smith, W. (editor) 1988 An analysis and evaluation of ichthyoplankton sur- vey data from the Northeast Continental Shelf ecosystems. NOAA Tech. Memo. NMFS-F/NEC-57, NMFS Northeast Fish. Sci. Cent., Woods Hole MA, 132 p. Smith and Morse. Larval distribution patterns of Oupea harengus "on the Georges Bank 347 Stephenson, K.. & I. Korn field 1990 Reappearance of spawning Atlantic herring iClupea harengus harengus) on Georges Bank: population re- surgence not recolonization. Can. J. Fish. Aquat. Sci. 47:1060-1064. Theilacker, G. 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. Tibbo, S., J. Legare, L. Scattergood, & R. Temple 1958 On the occurrence and distribution of larval her- ring {Clupea harengus L. ) in the Bay of Fundy and Gulf of Maine. J. Fish. Res. Board Can. 15:1451-1469. Townsend, D., J. Graham, & D. Stevenson 1986 Dynamics of larval herring (Clupea harengus L.) production in tidally mixed waters of the eastern coastal Gulf of Maine. In Bowman, J., C. Yentsch, & W.T. Peterson (eds.), Tidal mixing plankton dynam- ics, p. 253-277. Springer- Verlag, Berlin. Abstract. -During the period 1962-90, an aerial monitoring pro- gram conducted in cooperation with aerial spotters searching for pelagic fishes off California and Baja Califor- nia, Mexico resulted in the develop- ment of an index of apparent abun- dance for six species of fishes: the northern anchovy Engraulis mordax, Pacific sardine Sardinops sagax, Pa- cific bonito Sarda chiliensis. Chub mackerel Scomber japonicus, jack mackerel Trachurus symmetricus, and bluefin tuna Thunnus thynnus. North- ern anchovy was the dominant spe- cies observed, accounting for 89.7% of tonnage recorded during 1962-90. Chub mackerel comprised 6.1% of the total tonnage, jack mackerel 2.1%, Pa- cific sardine 1.0%, Pacific bonito 0.6%, and bluefin tuna 0.5%. Apparent abundance indices were computed by dividing the tonnage ob- served by the number of "block areas" (10' lat. x 10" long.) searched, ex- pressed as tons/block area flown (T/ BAF). Indices were calculated for the total area and for the core area of dis- tribution and abundance for each spe- cies. All species exhibited large fluctuations in apparent abundance over time. The apparent abundance index for Pacific sardine declined to a very low level during 1966-83. A sub- stantial increase in abundance oc- curred during the mid- to late-1980s, with the 1990 index value 58 times that observed during the early 1960s. The chub mackerel abundance index declined to a very low level during 1966-76, then increased to a record high value in 1978 and has since de- clined. The northern anchovy abun- dance index increased in the early 1970s, with high abundance levels re- corded during 1972-81, and then de- clined sharply to very low levels in the 1980s. The Pacific bonito abundance index was high during 1965-67 and 1983-85, with current levels well be- low the long-term mean. Very high abundance index levels for jack mack- erel were recorded during 1975-79, then declining to very low levels in the 1980s. The bluefin tuna abundance in- dex increased during 1972-80, declin- ing to very low index values since. Based on these data, several species appear to be fluctuating unpredictably with respect to species abundance over time. Aerial indices were compared with available total, spawning, and larval biomass estimates developed for sev- eral species. The most significant cor- relation of the aerial index was be- tween the northern anchovy and Pacific sardine larval indices. 348 Relative abundance of pelagic resources utilized by the California purse-seine fishery: Results of an airborne monitoring program, 1962-90 James L. Squire, Jr. Southwest Fisheries Science Center National Marine Fisheries Service. NOAA 8604 La Jolla Shores Drive, La Jolla, California 92038 The rapid decline of the Pacific sar- dine Sardinops sagax fishery in the late 1940s resulted in increased re- search by the State of California Department of Fish and Game (CDF&G), Federal Government, and other institutions to determine the underlying principles that govern the sardine's behavior, availability to the fishery, and total abundance (Clark & Marr 1955, Sette 1969, Radovich 1982). Catch-per-unit-effort (CPUE) studies of the sardine purse- seine fishery off California did not indicate a significant decline in abundance until just prior to the fishery collapse (Fox 1974). Fisher- ies tend to target their fishing effort in areas of high densities of fishes (Radovich 1982, Squire & Au 1990), hence CPUE may not accurately re- flect resource abundance (de La Mare 1984; J.B. Phillips, CDF&G, pers. commun., 1954). Clark & Marr ( 1955) concluded that, in addition to other information needs, better in- formation is needed on changes in availability of sardine to the fishery. Related to this is the need for better real-time measures of trends in ap- parent abundance for successful management of pelagic schooling re- sources. In the mid-1950s, the use of air- craft in searching and fishing opera- Manuscript accepted 28 January 1993. Fishery Bulletin, U.S. 93:348-361 ( 1993). tions of the California purse-seine fleet had become well established, as was the case in many other U.S. fisheries (Squire 1961). Having had experience in conducting commercial aerial fish spotting and several aerial fish-resource surveys, I initiated a pe- lagic monitoring program in the fall of 1962, with the cooperation of aerial fish-spotter pilots searching for fish for the commercial purse-seine fishery off central and southern Cali- fornia. The purpose of the program was to develop an effective method of measuring apparent abundance and monitoring changes in abun- dance. In the case of aerial observa- tions, the concentrations of fish ob- served may or may not be subjected to fishing. This paper updates the analysis of apparent abundance data col- lected by the aerial monitoring pro- gram from the first full year of data, 1963-90, using different analytical methods than used in earlier papers (Squire 1972, 1983). It also reviews the trends in apparent abundance during the period 1963-90 for six pelagic species by geographical area and compares the relation of the aerial abundance estimates to other independent biomass estimates made for the same species. The term "apparent abundance" (Marr 1951) refers to that portion of total abundance that is available to the fishery. Squire. Aerial monitoring of abundance of pelagic fishes 349 Species included in this study are the northern an- chovy Engraulis mordax, Pacific sardine Sardinops sagax, Pacific bonito Sarda chiliensis, chub mackerel Scomber japonicus, jack mackerel Trachurus symmetricus, and bluefin tuna Thunnus thynnus. These species, along with market squid Loligo opalescens, are the primary targets of the commercial purse-seine fleet based in San Pedro, California. Most are also fished by the southern California sportfishing fleet. The species are members of a complex interacting coastal and oceanic ecosystem that crosses interna- tional boundaries (Sette 1969). Methods Procedures for collection and tabulation of data Since the fall of 1962, fish-spotter pilots have been directed to complete a flight logchart at the termina- tion of each spotting flight. These pilots were instructed to draw their flight track on the chart and record the locations of all schools of fish species observed, along with their best estimates of tonnage observed. These data were recorded either as total tonnage of fish in 123° 122* 121* SAM FRANCISCO Ba M 118° 116* 34° 33° 32' 117' 116' 115' Figure 1 Block areas designated for sampling from off central California to Mexico, grouped in 18 zones, A to T. (Zones Q and S were originally used on flight logs; however, no flights were recorded for these zones. ) 350 Fishery Bulletin 91(2). 1993 the area or numbers of schools with an estimate of the tonnage of each school. Procedures for recording spe- cies observations and survey effort have not changed since 1962. Other information such as time of flight (day or night) and flight duration also are recorded on the flight log. Flight logs are submitted quarterly and data from each log are entered into the computer after editing. Spotter pilots' income is derived from the commer- cial fishery at a current rate of -6% of the vessel's gross income. Spotter pilots are paid a nominal fee by the National Marine Fisheries Service (NMFS) to maintain a flight log record of their observations. Pilots in the program averaged ~600h flight time/yr and, with 4-6 pilots, the total number of hours flown was ~3000/yr. From 1963, the first full year of operation, to 1990, annual indices of abundance for each species were calculated for the total area (from northern Baja Cali- fornia, Mexico to San Francisco, California) and for selected groups of 10" lat. x 10" long, block areas, similar to the "block area" statistical system used by CDF&G (Anon. 1935). These groups of block areas are referred to in this paper as "zones" and were designed to encompass important commercial fishing areas (Fig. 1). Flight effort for the total area and each zone was determined by counting the number of times the aerial spotter's flight track entered a block area, termed a "block area flight" (BAF). Species abundance was re- corded as the tonnage estimate (and hence approxi- mates biomass) for the area observed. Also recorded on the flight log was time of the observation (day or night), pilot, and, since 1988, the target species of the commercial operation. The procedure to measure effort has not changed from previous studies; however, the computation of spe- cies abundance has been modified. In previous analy- ses (Squire 1972, 1983), abundance was estimated as a rank index according to discrete incremental values of tonnage (tonnage range values). For example, for the northern anchovy, an observation within a block area estimated to be 0-100t was recorded as 1, 100- 200t was recorded as 2, etc. Each species had different tonnage range values, since the amount of tonnage observed varied greatly between species. Therefore, the abundance index values for each species are not di- rectly comparable. For this analysis, actual tonnage values as recorded by the spotter pilot and a modifica- tion of the index used by Squire ( 1972) are used, which allows for direct comparison of abundance index val- ues between species. The apparent abundance index is T/BAF, where T = tonnage estimate, and BAF = total number of block area flights (day plus night, all zones). Results and discussion Observation effort and sightings During the period 1963-90, a total of 24 aerial fish- spotter pilots participated in the program and reported flight operations totaling over 67,000h, searching a total of 376,446 block areas. Table 1 gives the annual search effort (BAF/yr), day and night, during the pe- riod 1962-90. From 1963 to 1990, a total of 253,239 BAF were made during day operations (67%) and 123,207 BAF during the night. Pilots have searched an average of 13,444 BAF/yr; however, the variation in annual spotting effort is considerable. There was reduced effort in early 1970. Abundance data for 1970 Table 1 Search e ffort in block area flights (BAF), day and night, 1963-90, from off Baja California Mexico to cen tral Califor- nia. Asterisk ndicates <^/>0 Yk- . ^*£-*\ Figure 4 Abundance index levels in T/BAF for Pacific bonito Sarda chiliensis core area, day index, weighted x, and running aver- age (X3>. Chub mackerel (Fig. 5) Declining abundance levels noted in the early to mid-1960s matched the collapse of the fishery, with extremely low index levels (>5 T/ BAF) noted for the core-area index (night index) in 1967-75. After 1976, the abundance index rose rapidly to a peak of 170 T/BAF in 1978, then declined to a range of 15-55 T/BAF in 1981-89. Jack mackerel (Fig. 6) Abundance was high during the mid 1970s. The core-area index (night index) in- creased to 234 T/BAF in 1976 and declined to <10 T/ BAF after 1979. Bluefin tuna (Fig. 7) For the core area, there were three peaks in the index (day index) of abun- dance: 1967, 1973, and 1978. The index for recent years is well below the 1963-88 average of 10.8 T/ BAF. 356 Fishery Bulletin 91 [2). 1993 300 r 250 - O Pacific Mackerel A aerial index < 2°° Core Area Zones \ , - <»«'9|>'Mi < O 150 g CO or £ 100 f \ AVG WEIGHTED X / I. 1976-88 -65 6 T.'BAF AERIAL / \\ 1 INDEX 1 U / A RUNNING ,7 \\ f / \ AVG WEIGHTED X AVERAGE // 1 '. / -M 1963-66 - 1 52 T7BAF OF 3 .1 l, / "\. A o so t=*x^ , .. / / ^ Vv-»V> 62 64 66 68 70 72 74 76 78 80 82 B4 86 88 90 YEAR Figure 5 Abundance index levels in T/BAF for chub mackerel Scomber japonicus core area, night index, weighted x, and running average ( x 3 ). 80 15S?i Jack Mackerel 2» lit 70 Core Area Zones O -i 60 < CC 50 < o O 40 CD £ 30 0_ (f) O 20 AERIAL INDEX I ltvr.ighle.3i 1 1/ AERIAL J 1 INDEX 1 \ HUNNING 1 I AVERAGE , f -r. OF 3 \ ' / / \ V' f AVG WEIGHTED x 1963 88 = 39 7 T/BAF I / 10 V v' " A' " V ^%<^V^-*-4-^^ 1 }62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 YEAR Figure 6 Abundance index levels in T/BAF for jack mackerel D'achurus symmetricus core area, night index, weighted .v, and running average f x3). 70 Bluefin Tuna 60 Core Area Zones ft a Ij SO Li. < LU ? 40 U O ^ 30 AERIAL INDEX / (weighted) 1 Ki rr Q. z o in 1 AFRIAL /»--,l ;/ \ A IN0EX /' \ \ l\ HUNNING / 1*. .' / ft ^ / \ AVERAGE •/ 1 '. •' / \ / \ °F 3 *t 1 '■ «' 1 AVG WEIGHTED X 1963-98. 10 8 T/BAF i *>. -•/ ^ \r^* ^^"5*^-*-*. 1962 64 66 68 70 72 74 76 78 80 82 84 86 88 90 YEAR Figure 7 Abundance index levels in T/BAF for bluefish tuna Thunnus thynnus cure area, night index, weighted .v, and running av- eragei 3) Interaction between species To compare periods of high and low abundance of the six species, the logarithm of the moving weighted averages of core-area abundance indices (day or night, dependent upon which time the species was observed more frequently and in greater quantity) were plotted (Fig. 8). The dominant feature of Fig. 8 is the abun- dance index for the northern anchovy. High index lev- els of 200-500 T/BAF were recorded during 1962-82, a 21 yr period. The index declined sharply in 1983 and 1984, increasing since. The northern anchovy is the more stable species in terms of less annual percentage change in abundance than any of the other five spe- cies. During 1966 and 1967, jack mackerel, bluefin tuna, and Pacific bonito were at higher levels of abun- dance than those observed in the early 1970s. The Pacific sardine and chub mackerel resources were low in abundance by 1965, and declined to extremely low levels by 1970. The northern anchovy, jack mackerel, chub mackerel, and bluefin tuna increased in abun- dance in 1974-79. Exceptions were the Pacific bonito and sardine. The very strong El Nino condition of 1982-83 (Quinn et al. 1987) had a significant effect on the distribution of coastal pelagic species (Fluharty 1984, Squire 1987). This period saw declines in abundance for all species, except Pacific bonito and chub mackerel. These two species showed significant declines in 1986-87. A rapid increase in Pacific sardine abundance was evident in 1984, immediately following the El Nino event and continuing into the late 1980s. 1,000 r 500 / \S ^~~A Northern -•'N/ \ Anchovy S 200 ^^-s ..y\ ^ \ t 100 Jack • t^ . S \ Mackerel J / ', ^ \ pacif.c < 50 ' \ \ \y yc O m 20 "* / • i "'i| *" \ j' ** * Pacilc CC V V^ , . . V"^ Mackerel CO / i ' | i , o 5 J 1 Bluefin ', rf"~"! B°ni'° c\ -^~., ; iun* .\ A ..y 2 * ,*- ..-■'• i . \j 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 60 81 82 83 84 85 86 87 88 89 YEAR Figure 8 Comparison of core-area fluctuations in apparent abundance index levels for northern anchovy Engraulis mordax. Pacific sardine Sardinops sagax, chub mackerel Scomber japonicus. Pacific bonito Sarda chiliensis, jack mackerel Trachurus symmetricus, and bluefin tuna Thunnus thynnus. Data plot- ted on a semi-logarithmic scale, in weighted X, T/BAF, by moving average ( x3). Squire Aerial monitoring of abundance of pelagic fishes 357 From inspection of core-area fluctuations (Fig. 8), it is difficult to draw any conclusions regarding interac- tion among most of the species included in this study. The annual abundance values that appear to be more closely related are those of the northern anchovy and bluefin tuna which were significantly correlated (y=0.76,P<0.1). Pilot estimation of tonnage Differences between pilots' ability to estimate school or group size are difficult to evaluate using the type of data recorded in flight logs. Computer recording of flight track and species tonnage observed did not in- clude details of what portion of the block was searched, or where the species were sighted. Aerial spotters en- tering the block area (an 8xl0nm area) may have observed fish schools of the same species in two geo- graphically different locations. Only the sum of the school sizes was coded. The time of observation is probably the most critical factor in comparisons of pilot sightings because near- surface fish abundance may be highly variable over a short period of time. Therefore, unless pilots are search- ing the same area at the same time, different levels of abundance are likely to be observed. Williams (1981) in surveys of pelagic fish resources off southeast Australia indicated that the professional fish spotters were extremely accurate in tonnage esti- mates of individual schools. In tests conducted by NMFS and CDF&G to determine northern anchovy school size using acoustic gear, a purse seiner and aerial spotter were used. The aerial spotter had a more accu- rate estimate of school tonnage than obtained from acoustic gear (P. Mardesich, Flying Fishermen, Inc., pers. commun. 1978). MacCall ( 1975), using the NMFS computer database, investigated variation in anchovy abundance estimates (comparison of data collected during the same lOd pe- riod in 1973 and 1975) and found considerable differ- ences among spotters. The differences were consistent for individual pilots for day and night observations. Variability in reported tonnage was large, but it could be reduced by long-term averaging. MacCall ( 1975) fur- ther suggested that corrections for pilots estimating efficiency be incorporated in calculations of the abun- dance index to avoid biasing the result toward the spotter with the highest tonnage. The accuracy of fish- spotter estimates of tonnage was approached statisti- cally by Lo et al. (In press) using a delta-lognormal model. All aerial spotter data were used in this analy- sis. However, until spotter-pilot estimating efficiency is measured in field experiments, correction factors are unknown. It is possible that these differences be- tween spotters may be due more to differences in num- bers of schools seen than to differences in size esti- mates. Comparison of aerial index to other abundance indices The problem of variability among spotter pilots com- plicates comparisons; however, pilots are paid relative to the amount of tonnage caught. Therefore, the aerial spotter needs to have a reasonable estimate of school size to compare with the amount of fish caught by the seiner. The aerial spotter is continually obtaining "ground truth" from the purse-seine vessel relative to species composition and size of the school caught. Since the decline of the Pacific sardine resource in 1984, considerable effort has been expended on sur- veys, using various techniques to collect apparent abun- dance data that could be modeled to produce an esti- mate of biomass. Acoustic, egg and larva, and aerial-spotter surveys have been conducted. Each sur- vey measures a different component of the population and, therefore, comparisons of abundance estimates from different methods and models must be viewed with caution. The aerial-spotter program measures apparent abun- dance of adults and subadults of epi-pelagic schooling fishes, and assumes that most adults in the popula- tion are available to the spotters at some time during the year. The aerial index is a direct calculation (T/ BAF). It is a non-random type survey, but is not time- restricted. The operational procedures used by aerial spotters affect the results. Although they do work closely with the fleet in locating and sometimes directing the set- ting of the purse seine, much of their time is spent searching large areas of the ocean for fishable resources. Many times, the aerial spotter will conduct flight op- erations in a general survey mode for the fleet, but does not work directly with the fleet. Even though aerial monitoring is conducted throughout the year, most observation effort is in the summer and early fall (49% of BAF during July, August, and September) when concentrations of commercial species are more com- mon in the nearshore areas. Quantitative apparent-abundance data derived from field surveys (egg and larvae, and acoustic) are statis- tically treated under various biological assumptions to ultimately obtain an estimate of biomass or tonnage: spawning biomass and total biomass in the case of egg and larva surveys (Lo 1985), and schooled biomass in the case of acoustic surveys (Mais 1974). Although some biomass estimates have been made for other species, such as the chub mackerel and Pacific 358 Fishery Bulletin 9 1 12). 1993 sardine, most biomass studies and modeling for the California region have involved the northern anchovy (Methot 1989). Over the past 35 yr, studies by fishery biologists have produced a wide range of population estimates for a number of the pelagic species, in particular, the north- ern anchovy, Pacific sardine, and chub mackerel. The following describes the results of these and others in relation to aerial abundance indices. Northern anchovy A review of the many abundance estimates for the northern anchovy over the years 1940-66 was given by Messersmith et al. (1969). For the early years of the aerial program, MacCall et al. (1974) compared anchovy biomass estimates from egg and larva surveys with the night aerial index, result- ing in a correlation of +0.30 ( 1962-69). They also com- pared the aerial night index with the acoustic esti- mate (1966-72) producing a negative correlation (r=-0.47). In a review of egg production of the northern an- chovy central stock, Lo (1985) presented a correlation between various indices of anchovy spawning biomass. The biomass index from the aerial surveys had a high correlation (r=0.818) with egg production, and there was a similar correlation (r=0.807) between biomass egg production and acoustic surveys. Correlation plots of normalized data (standard devi- ates; Snedecor 1959) were used to compare northern anchovy abundance data presented by Lo (1988) with aerial index values. The aerial index (day plus night) for the core area was compared with Lo's (1988) esti- mate of spawner biomass, determined from a combi- nation of various estimates from the egg production method and the results of the 1969-85 sonar or acous- tic surveys (Lo 1988), giving a correlation of r=0.61, df=14, P<0.057c The aerial index was also compared with the estimate of schooled biomass based on results of acoustic surveys presented by Lo (1988), giving a correlation of r=0.51, df=13, P<0.05%. MacCall & Prager (1988) presented anchovy larval- abundance indices for 1963-85. These indices, which are a measure of spawning population, were compared with the core-area aerial index for the same set of years. There was statistically significant correlation between the two indices (aerial/larva) (r=0.82, df=13, P<0.1). Pacific sardine MacCall & Prager ( 1988) also reviewed the historical trend of sardine larval abundance, 1951- 85. A comparison was made between their arithmetic scale of larval abundance (standard index) and the aerial index during the period 1963-85. Indications of a resurgence of the Pacific sardine resource began in 1978, and for the 1963-85 period MacCall & Prager 's ( 1988) estimate of larval abundance correlates well with the aerial index (/-=0.98, df=8, P<0.01). Pacific bonito Pacific bonito, a species whose sea- sonal abundance off California is more closely associ- ated with environmental conditions (El Nino effects) than the other species (Squire 1987), has not been subjected to detailed assessment. MacCall et al. (1974) used recreational partyboat CPUE data in yearly time- lag periods and compared this with the aerial index. The recreational partyboat fleet generally fishes nearshore and catches younger fish than the offshore commercial purse-seine fleet. The highest correlation (r=0.69) between the recreational fleet CPUE and the aerial index was for a 3yr lag period (aerial index lags the recreational CPUE). No other independent esti- mates of Pacific bonito biomass are available for com- parison. Chub mackerel Assessment of the chub mackerel re- source has used both standard population-dynamics techniques and methods that utilize egg and larva sur- veys. For comparison of changes in population over time, Squire (1983) plotted the trend of the aerial in- dex against the spawning biomass index and the larva index of Smith & Richardson (1977). The limited data (1963-67) precluded any long-term comparison; how- ever, all three indices showed a similar decrease. The aerial index was significantly correlated (P<0.05) with both spawning biomass (r=0.99) and the larva index (r=0.94). The chub mackerel biomass declined to very low lev- els in the mid-1960s and remained very low until the mid-1970s. A sharp increase in biomass was noted starting in 1976. MacCall & Prager (1988) presented data on chub mackerel larval abundance through 1985. A correlation of the aerial index (core area) with the larva index for 1963-81 was statistically insignificant (r=0.205). A comparison made between the aerial in- dex (core area) and estimated biomass of MacCall et al. ( 1985) for 1963-84 gave what would appear to be a good fit for the early years, but comparison of later years' estimates was not significant (r=0.26). Jack mackerel Various indices of abundance for jack mackerel, including the aerial index, show that the resource tends to be highly variable in apparent abun- dance. Egg and larva surveys have shown the stock to be extremely widespread. Based on tagging data, Knaggs (1973) estimated between 700,000 and 1,500,000 1 were available to the fishery in 1973. MacCall et al. (1974) stated that these estimates are in agreement with trends in aerial survey index (Squire Squire Aerial monitoring of abundance of pelagic fishes 359 1972). MacCall & Prager ( 1988) presented graphics on a time-series of abundance indices for jack mackerel larva off southern California. They noted that the in- dex did not correspond well with other larval indices that included samples from central California. An ex- amination of the MacCall & Prager (1988) larva index vs. the aerial index indicated a poor correlation (r=0.16, not significant), particularly for the years 1977-78, a period of marked increase in jack mackerel abundance off southern California. Bluefin tuna No biomass estimates are available for bluefin tuna in the northeast Pacific, which is part of a Pacific-wide resource. Young year-classes (l-3yr) are common in the northeast Pacific with some catches of fish up to 6yr of age (Bayliff 1980). CPUE estimates presented by Calkins (1982) of catch/boat-day (C/BD) for the U.S. purse-seine fishery operating off Califor- nia and Baja California, Mexico for the years 1963-80 in the area north of 29° N lat. were compared with the aerial core-area mean index. Although both indices (C/ BD and T/BAF) show a decline prior to 1980, the cor- relation was not significant (r=0.033). Calkins (1982) stated that "total catch may be the best indicator of bluefin abundance." Total catch (north of 29° N lat.) as given by Calkins (1982) was compared with the core- area index, and the correlation was again not signifi- cant (r=0.147). Summary and conclusions The aerial monitoring program has produced informa- tion on the trend of apparent abundance for six spe- cies of pelagic near-surface schooling species. The pro- gram in operation since 1962 utilizes information provided by commercial aerial fish spotters. Spotter pilots operate under strong incentive to estimate ton- nage accurately, since their compensation is based on actual tonnage caught. The cost of the program is a small fraction of the cost of other field assessment methods currently used to develop apparent abundance data, such as egg and larva surveys and acoustic sur- veys. Data are presented in a relatively simple and un- derstandable index format of tons observed/block area searched (T/BAF). Annual indices (T/BAF) of apparent abundance are calculated for the major species of im- portance to the commercial purse-seine fleet operating off southern California and northwestern Baja Califor- nia, Mexico. All species have shown a fluctuating pattern of ap- parent abundance over the 29 yr survey period ( 1962- 90). The northern anchovy is the dominant species observed, accounting for 89.7% of all tonnage observed; chub mackerel is second with 6.1%; followed by jack mackerel with 2.1%, and Pacific sardine, Pacific bo- nito, and bluefin tuna with 2.1%. From inspection of abundance data, peak periods of higher abundance for four of the six species (chub mack- erel, jack mackerel, northern anchovy, and bluefin tuna) occurred in 1977-79. Analysis indicated that species more closely related in abundance over time are the northern anchovy and bluefin tuna, both being at higher levels of abundance during 1972-80. A regres- sion was calculated for 1963-88, giving a significant correlation (r=0.76, P<0.1). A number of comparisons were made between esti- mates of biomass (spawning or total biomass) and the aerial abundance indices for the northern anchovy. Pa- cific sardine, and chub mackerel, and the apparent abundance of bluefin tuna based on CPUE data. There is a significant correlation with the aerial index, the anchovy total biomass estimate, and the historical egg- production index. Also, there was a slightly less but significant correlation with the acoustic survey esti- mates. The more significant correlations were found in comparing the aerial index with the northern anchovy larva index and the Pacific sardine larva index. For the chub mackerel, a species recognized as having a highly fluctuating record of abundance, the correla- tions during the early years (1963-66) were very good (aerial index vs. larva index and spawning biomass index); in later years, correlations between the larva index and the estimates of total biomass and the aerial index were not significant. There is a lack of good data relative to the abundance for other species of interest (Pacific bonito, jack mackerel, bluefin tuna) although some comparisons with the aerial data have been made for Pacific bonito, giving a reasonable correlation us- ing a 3yr lag period. Bluefin tuna C/BD (catch/boat- day) and total catch for the northwest Mexico-southern California fishery were compared with the aerial core index and gave no significant correlation. Trends of aerial indices indicate that the pelagic resources studied can be described in a state of chaotic fluctuations (Jensen 1987) relative to apparent abun- dance. The fluctuating record indicates that even short- term predictions of abundance appear not to be feasible. Management must rely on near real-time estimates based on a measurement of a small fraction of the resource, and what is important is the development and use of more than one method or approach in evalu- ating population status. None of the survey results appear in complete agreement; however, utilizing sev- eral approaches to provide "triangles of agreement" (from a method used in navigation) can provide a con- sensus for management action. 360 Fishery Bulletin 91(2). 1993 Acknowledgments The assistance of Dr. David Au in making the statis- tical comparisons between various methods of measur- ing apparent abundance and other editorial sugges- tions is greatly appreciated. The assistance of Dr. Nancy Lo, Dr. Larry Jacobsen, Cindy Meyer, Richard Tauber, Lorraine Prescott, and Karen Handschuh of NMFS, and of Gordon Broadhead and Tom Barnes of Living Marine Resources, Inc., for their review and contribu- tion to the development of this paper is appreciated. To the aerial fish spotters who have provided at least 67,000 flight hours of observation to the aerial pro- gram, I express my thanks. Citations Anonymous 1935 Commercial fish catch of California for the years 1930-1934 inclusive. Calif. Div. Fish & Game, Bur. Commer. Fish., Fish. Bull. 44, 126 p. Bayliff, W.H. 1980 Synopsis of biological data on the northern bluefin tuna, Thunnus thynnus (Linneaus 1758) in the Pa- cific Ocean. Inter-Am. Trop. Tuna Comm. Spec. Rep. 2:261-294. Calkins, T.P. 1982 Observations on the purse-seine fishery for north- ern bluefin tuna (Thunnus thynnus) in the eastern Pacific Ocean. Inter-Am. Trop. Tuna Comm. Bull. 18(2), 225 p. Clark, F.N., & J.C. Marr 1955 Population dynamics of the Pacific sardine. Calif. Coop. Oceanic Fish. Invest. Prog. Rep., 1 July 1953- 30 March 1955, p. 11-48. Cochran, W.G. 1966 Sampling techniques. John Wiley, NY, 413 p. de la Mare, W.K. 1984 On the power of catch per unit effort to detect declines in whale stocks. Rep. Int. Whaling Comm. 34:655-661. Fluharty, D. (coordinator) 1984 1982-1983 El Nino task force summary. Inst. Mar. Stud., Univ. Wash., Seattle, 25 p. Fox, W.W. 1974 An overview of production modeling. Admin. Rep. LJ-74-10, NMFS Southwest Fish. Sci. Cent., La Jolla CA, 20 p. Jensen, R.V. 1987 Classical chaos. Am. Sci. 95:168-181. Knaggs, E.H. 1973 Status of jack mackerel resource and its management. In Proceedings of the State-Federal marine fisheries research program planning workshop, March 12-15, 1973, San Clemente, p. 150-156. Ca- lif. Dep. Fish Game, NMFS Southwest Fish. Sci. Cent., La Jolla CA. Lo, N.C.H. 1985 Egg production of the central stock of northern anchovy, Engraidis mordax, 1951-82. Fish. Bull., U.S. 83:137-150. 1988 Preliminary spawning biomass estimate of the northern anchovy in 1988. Admin. Rep. LJ-88-17, NMFS Southwest Fish. Sci. Cent., La Jolla CA, 15 p. Lo, N.C.H. , L. Jacobsen, & J. Squire In Press Indices of relative abundance from fish spot- ter data. Can. J. Fish. Aquat. Sci. MacCall, A. 1975 Investigation of aerial spotter variation in reported abundance of anchovies. Contrib. 19, Anchovy Work- shop Meet., 21-22 July 1975. NMFS Southwest Fish. Sci. Cent., La Jolla CA, 3 p. MacCall, A., & M. Prager 1988 Historical changes in abundance of six fish spe- cies off southern California, based on CalCOFI egg and larva samples. Calif. Coop. Oceanic Fish. In- vest. Rep. 29:91-101. MacCall, A., G. Stauffer, & J. Troadec 1974 Stock assessment, fishery evaluation, and fishery management of southern California recreational and commercial fisheries. Admin Rep. LJ-74-24, NMFS Southwest Fish. Sci. Cent., La Jolla CA, 144 p. MacCall, A., R. Klingbeil, & R. Methot 1985 Recent increased abundance and potential produc- tivity of Pacific mackerel (Scomber japonicus). Calif. Coop. Oceanic Fish. Invest. Rep. 16:119-129. Mais, K. 1974 Pelagic fish surveys in the California Current. Calif. Dep. Fish Game, Fish. Bull. 162, 79 p. Marr, J. 1951 On the use of the terms abundance, availability and apparent abundance in fishery biology. Copeia 1951:163-169. Messersmith, J.D., J. Baxter, & P. Roedel 1969 The anchovy resources of the California Current region of California and Baja California. Calif. Coop. Oceanic Fish. Invest. Rep. 12:32-38. Methot, R.D. 1989 Synthetic estimates of historical abundance and mortality for the northern anchovy. In Edwards, E.F., & B.A. Megrey (eds.), Mathematical analysis offish stock dynamics, p. 66-82. Am. Fish. Soc. Symp. 6, Bethesda. Quinn, W.H., V.T. Neal, & S.E. Antunez de Mayola 1987 El Nino occurrences over the past four and a half centuries. J. Geophys. Res. 92(C13):14,449-14,461. Radovich, J. 1960 Redistribution in the eastern north Pacific Ocean, 1957and 1985. Calif. Coop. Oceanic Fish. Invest. Rep. 7:163-171. 1982 Collapse of the California sardine fishery. Calif. Coop. Oceanic Fish. Invest. Rep. 23:56-78. Sette, O.E. 1969 A perspective of a multi-species fishery. Calif. Coop Oceanic Fish. Invest. Rep. 13:81-87. Squire: Aerial monitoring of abundance of pelagic fishes 361 Smith, P., & S. Richardson 1977 Standard techniques for pelagic fish egg and lar- vae surveys. FAO Fish. Tech. Pap. 175, 100 p. Snedecor, G.W. 1959 Statistical methods. Iowa State Coll. Press, 334 P. Squire, J.L. 1961 Aerial fish spotting in the United States commer- cial fisheries. U.S. Fish. Wildl. Serv., Commer. Fish. Rev. 23(12):l-7. 1972 Apparent abundance of some pelagic marine fishes off the southern and central California coast as sur- veyed by an airborne monitoring program. Fish. Bull, U.S. 70:1005-1019. 1983 Abundance of pelagic resources off California, 1963-78 as measured by an airborne fish monitoring program. NOAA Tech. Rep. NMFS SSRF-762, 75 p. 1987 Relation of sea surface temperature changes dur- ing the 1983 El Nino to the geographical distribution of some important recreational pelagic species and their catch temperature parameters. Mar. Fish. Rev. 49(2):44-57. Squire, J.L., & D.W.K. Au 1990 Striped marlin in the northeast Pacific - A case for local depletion and core area management. In Plan- ning the future of billfishes; Proc. 2d Int. Billfish Symp., Pt. 2 Marine recreational fisheries, p. 199- 214. Natl. Coalition for Mar. Conserv., Savannah GA. Williams, K. 1981 Aerial surveys of pelagic fish resources off south- east Australia 1973-1977. CSIRO, Div. Fish. Ocean., Rep. 130. Abstract. —Analysis of restriction- site polymorphisms of mtDNA and Monte Carlo statistics were used to test the hypothesis that at least two genetic stocks of Spanish sardine Sardinella aurita are present in the eastern Gulf of Mexico, with one oc- curring at each of the two main fishery locations off Florida. Also in- cluded for comparison were speci- mens of Spanish sardine collected off southern Brazil. No significant het- erogeneity of mtDNA haplotype fre- quencies was detected among speci- mens from the two locations within the Gulf of Mexico after analysis of 57 individuals (28 from Tampa Bay and 29 from the Florida Panhandle) using 9 informative restriction en- zymes. However, highly-significant differences were observed between the specimens from the Gulf of Mexico and 16 specimens from southern Brazil. Based on counts of gill rakers of all specimens from both regions and the results of the mtDNA analysis, it is suggested that Sardinella brasiliensis, the Brazil- ian sardine, is conspecific with S. aurita, and that S. aurita is prob- ably represented by genetically- identifiable populations in the west- ern North and South Atlantic. Differences in haplotype frequencies of mtDNA of the Spanish sardine Sardinella aurita between specimens from the eastern Gulf of Mexico and southern Brazil Michael D. Tringali Department of Marine Science, University of South Florida 140 Seventh Avenue South. St Petersburg, Florida 33701 Present address. Florida Marine Research Institute Department of Natural Resources 100 Eighth Avenue SE, St. Petersburg, Florida 33701-5095 Raymond R. Wilson, Jr. Department of Marine Science, University of South Florida 140 Seventh Avenue South, St. Petersburg, Florida 33701 Manuscript accepted 17 February 1993. Fishery Bulletin. U.S. 91:362-370(1993). Spanish sardine Sardinella aurita (Valenciennes) ranges in the western Atlantic from Cape Cod, throughout the Gulf of Mexico and Caribbean Sea, to southern Brazil (Whitehead 1973). It is also present in the Medi- terranean Sea and in the eastern At- lantic off the African coast (Fisher 1978, Whitehead 1985). A second sar- dine species, Sardinella brasiliensis (Steindachner), reportedly occurs within the range of S. aurita between the Gulf of Mexico and southern Bra- zil (Whitehead 1973 and 1985, John- son & Vaught 1986), but there have been no studies successful in fully differentiating the two species within the eastern Gulf of Mexico (Wilson & Alberdi 1991). There has tradition- ally been a small but important baitfish fishery for Spanish sardine off Florida with an annual landing of over 1500 1 as recently as 1988. However, this fishery closed abruptly in 1989 after landings dropped sharply between 1988 and 1989. The closure was partly a response to un- certainty about the number of spe- cies or stocks being fished (Sutter & Mahmoudi 1992). Management concerns about the sardine fishery in the eastern Gulf of Mexico involve the possibility of lo- cal stock depletion or collapse from increased fishing pressure at the two primary fishing areas in the eastern Gulf of Mexico, the Florida Pan- handle and Tampa Bay. There is also concern over the possibility of over- fishing one of the two species sup- posedly present. The first concern has arisen from the observation that Spanish sardines caught by the fisheries operating at these two loca- tions appear to consistently differ morphometrically (Johnson & Vaught 1986), suggesting the possibility of separate genetic stocks or species. In response to these concerns, Wilson & Alberdi (1991) electrophoretically compared more than 300 specimens from the two areas at 41 presump- tive genetic loci and found no impor- tant differences in allele frequencies, nor any fixed differences. They con- cluded that stock differences were only weakly indicated, and that no genetic evidence of two species of Spanish sardine in the eastern Gulf of Mexico had appeared. 362 Trmgali and Wilson mtDNA analysis of two stocks of Sardinella aurita 363 Analysis of restriction-site polymorphisms of mito- chondrial DNA (mtDNA) is generally regarded as of- fering a higher degree of resolution than protein elec- trophoresis in population genetic studies (e.g., Ferris & Berg 1987). and has recently become important in studies of potential fishery stocks. The most direct ana- lytical approach is to compare the frequency distribu- tions of composite mtDNA haplotypes among the sample populations, testing (x2) whether the distribu- tions are homogeneous and, therefore, indicative of stock mixing (Bentzen et al. 1989). However, the usu- ally skewed distribution toward a common haplotype(s) coupled with the common occurrence of several rare haplotypes (Billington & Hebert 1991) in a population sample usually results in low sample numbers per cell in contingency tables, and tabulated values of .r2 can- not be used properly. This problem can be avoided, however, by using a Monte Carlo method of computing the exact x2 expected from the observed haplotype fre- quencies among sample populations (Roff & Bentzen 1989, Bernatchez & Dodson 1990). In this study we employ analysis of restriction-site polymorphisms of mtDNA and Monte Carlo statistics to test the hypothesis that the two main fisheries in the eastern Gulf of Mexico are comprised of at least two genetic stocks of S. aurita. We have included speci- mens of S. aurita collected off southern Brazil under an a priori expectation that stronger differences in mtDNA haplotype frequencies should exist between specimens from the Gulf of Mexico and southern Bra- zil. Considering the geographic proximity of our sample populations in the eastern Gulf of Mexico, and recognizing the fact that inadequate sampling of the mitochondrial genome (as evidenced by the num- ber of restriction enzymes used) can bias the outcome in the direction of "no significant differences," inclu- sion of the Brazilian specimens allows us to objec- tively gauge the power of our methods to address the stated hypothesis. Materials and methods Specimens of S. aurita from the eastern Gulf of Mexico were obtained from the two main commercial fisheries along Florida's west coast, one near Tampa Bay and the other off the Panhandle. Individuals were collected fresh from both sites during the summer of 1989 (Fig. 1). Ripe ovaries were excised soon after capture and held on ice or in refrigerated buffer (MSB-Ca" of Lansman et al. 1981) for up to lOd prior to purifica- tion of mtDNA; some ovaries from the Tampa Bay sample were stored frozen at -86 C until processing. The timing of specimen collections was structured to avoid resampling a single migrating school. Two samples were taken concurrently from Port St. Joe, Florida (re=10) and Destin, Florida (n=13) in May 1989. and a third sample was taken from Destin (n=6) in September 1989, for a total of 29 specimens (Fig. 1). In the Tampa Bay region a total of 28 individuals were taken during three collection times: February (n-12). May (n = 10), and August (ra=6) of 1989. Whole fish specimens obtained from off Santos, Bra- zil were collected from the fishery and hand-carried frozen on dry ice to St. Petersburg, Florida within 2d and were stored at -86 C for several months without removing the ovaries. Because these specimens were A. 7 %- GULF OF MEXICO - >\ SOUTHERN BRAZIL i 1 40°N 20CN 20°S 40°S 40°W 100°W 60°W 20°W B. GULF OF MEXICO _l_ 30°N 25°N 90°W 85°W 80°W Figure 1 Collection sites of Spanish sardines Sardinella aurita in the eastern Gulf of Mexico and southern Brazil: (A) The two major sites, and (B) two area sites in the eastern Gulf of Mexico. 364 Fishery Bulletin 91(2). 1993 also used in the electrophoretic study of Wilson & Alberdi (1991), one thaw/refreeze cycle occurred prior to use in the present study. All specimens were identi- fied to nominal species using the criteria of Whitehead (1973, 1985). That is, we counted the number of gill rakers on the lower limb of the first gill arch and related the count to standard length (SL), and we ex- amined the condition (curled or flattened) of the ante- rior rakers of the lower limb of the first arch. Mitochondrial DNA was purified from the ripe ova- ries of the specimens taken from the Gulf of Mexico as in Wilson & Tringali (1990), except that the mtDNA collected after density-gradient ultracentrifugation was extracted in n-butanol which was not saturated with NaCl. This removed most of the salts, allowing greater precision over the molar concentration of the DNA- containing solution during the ethanol precipitation. The lower yield of mtDNA caused by thawing and re- freezing of the Brazilian specimens necessitated use of a modification of the Chapman & Powers ( 1984) method of extraction (Tringali 1991). Restriction digests were carried out as specified by the manufacturers using the following enzymes: Apo-I, BamU-l, Dra-l, EcoR-l, Hind-UI, Pst-1, Pvu-ll, Sac-l, Xba-l, and Xho-l. Fragments were separated by hori- zontal agarose gel electrophoresis and visualized in most cases using the gel-incorporated Hoechst 33258 (CalBiochem #SR5A03-0388) fluorochrome dye as in Deflaun & Paul (1986). However, some poor yields of mtDNA during purification made visualization of the smaller restriction fragments difficult using the fluorochrome, thus requiring application of an mtDNA hybridization probe to southern blots of the agarose gels (Tringali 1991). The non-radioactive hybridization probe was made from cesium-purified mtDNA of S. aurita using a commercial kit ( Genius Kit of Boehringer Mannheim, Cat. #1093657). Hybridization and immu- nological detection of the probe was performed as speci- fied by the manufacturer with the following modifica- tions. The concentration of the blocking reagent in the prehybridization and hybridization solutions was doubled to 1.0% weight/volume to reduce background coloring of the hybridization filter (Zeta-Probe). The prehybridization period was extended to a minimum of 2h and the volume of the probe-containing hybrid- ization solution was increased to 10 m L/100 cm2 of filter. The ELISA reaction period was increased to 2 h. Under these modifications, the mtDNA restriction frag- ments were usually visible on the filter ~15min into the color reaction. Molecular weights (in base pairs) of the restriction fragments were estimated by the global form of the reciprocal method as described in Elder & Southern (1987). Restriction-fragment length polymorphisms (RFLPs) were expressed as two data types: restriction- site data and composite mtDNA haplotypes. The restriction-site composition of each individual was in- ferred from fragment profiles and intraspecific sequence diversity (Nei & Miller 1990) was calculated using the RESTSITE vl.2 algorithm. Composite mtDNA haplotypes were constructed for each individual from their restriction profiles with haplotypes grouped by sample location and time of collection. The frequency distribution of mtDNA haplotypes was tested for het- erogeneity, with respect to location and time, by the log-likelihood G-test for independence (Sokal & Rohlf 1981). G-tests were also performed on selected samples following generation of 100 randomized data sets from the observed data, using the Monte Carlo technique described by Roff & Bentzen (1989) where the prob- ability, P, of obtaining a randomized GH value greater than the original observed value is given by P = n / N„ (1) where n equals the number of outcomes in which the randomized GH value is greater than the actual GH value, and Nr represents the number of randomiza- tions performed. The standard error of P was taken as [P(l-P)/Nr]1/2. (2) The frequency of the most common haplotype was tested, after arcsine square-root transformation (Free- man & Tukey 1950), using the "V" statistic (Desalle et al. 1987) for heterogeneity between sampling locations. Results Counts of gill rakers identified 72 of 73 sample speci- mens as S. aurita; one specimen from Brazil had gill- raker counts expected for S. brasiliensis. Counts were lowest for the Panhandle specimens while specimens from Tampa Bay and southern Brazil were the most similar, having higher counts (Table 1). Specimens from the Florida Panhandle were of larger average size than those from Tampa Bay and southern Brazil, which were more similar in size (Table 1). A one-way analysis of variance revealed significant heterogeneity (P<0.05) in gill-raker counts between the specimens taken from the Panhandle and Tampa Bay, but not between those taken from Tampa Bay and southern Brazil. One Bra- zilian specimen had a gill-raker count of 154 at 150 mm SL, nearly identical to that reported for the lectotype of S. brasiliensis (155 at 148.3 mm SL; Whitehead 1973). However, we could detect no differ- ences in the appearance of the anterior rakers (curled vs. flattened) for this or other specimens. The overall appearance of the rakers of all specimens most closely resembled Whitehead's (1985) diagrammatic drawing Tnngali and Wilson: mtDNA analysis of two stocks of Sardmella aunta 365 Table 1 Data for Spanish sardines Sardmella aunta used in mtDNA analysis. Size gill-raker count range1 (mmSLl Location counted Mean- Range Port St. Joe 10 162-189 113111SD 102-127 i Panhandle I Destin 19 159-189 11419 99-128 l Panhandle i Tampa Bay 28 149-171 134±7 121-145 Santos, Brazil 16 149-181 132±9 122-154 1 Mean length of pooled specimens from the Panhandle is significantly larger than that of specimens from the other two regions (ANOVA, P<0.05). ' Mean number of gill rakers of pooled specimens from the Panhandle is significantly lower than that of specimens from the other two regions (ANOVA, P<0.05). of S. brasiliensis, curling anteriorly downward. How- ever, use of this character was extremely problematic for us because, other than the drawing, we found no published taxonomic account of the discovery of this character, no photographic evidence nor written descrip- tion detailing the character, nor reference of this char- acter to [any] catalogued or type material. The combination of mtDNA purification and visual- ization techniques used here provided well-resolved digestion patterns that could be consistently and un- ambiguously scored for all individuals. Of the 10 re- striction endonucleases used, only BamH-l was not informative. This enzyme had but one restriction site in all samples and was used only in probe construction to linearize template mtDNA prior to labeling. Poly- morphic digestion patterns occurred for the remaining nine enzymes, with Pst-l and Pvu-ll highly variable (Table 2). The mean mtDNA genome size, calculated from the sums of all digestion patterns, was 16,304 base pairs (bp) ± 71 bp (SD). No size variation was detected between specimens or within the mtDNAs of individual specimens. The nine informative enzymes sampled 52 unique restriction sites over all individuals (Table 3) with a mean of 43 sites per individual. All RFLPs were con- sistent with the assumption of a single nucleotide sub- stitution between variant digestion patterns. Intraspe- cific sequence diversity (Nei & Miller 1990) over all individuals was 0.00525±0.00204 (bootstrapped SE) substitutions per nucleotide, a value consistent among all three sampling locations (x=0.00509±0. 00054, range 0.00452-0.00561). Twenty-four composite mtDNA haplotypes were generated from polymorphic restriction- fragment profiles of the 73 completely characterized Spanish sardines (Table 4). The composite mtDNA haplotype (BAABAAAAA) of the one putative S. brasiliensis specimen with a gill raker count close to the lectotype's (Whitehead 1970, 1973) was identical to that of another Brazilian specimen having the low gill raker count indicative of S. aurita (126 at 169 mm SL), as well as to that of seven specimens from the Gulf of Mexico (Table 4). The mtDNA haplotype most common in the Gulf of Mexico, AAAAAAAAA, occurred in very low frequency among the Brazilian specimens. The fre- quency of this haplotype was significantly different be- tween specimens of the Gulf of Mexico and Brazil (V= 10.90, P<0.001), but not among specimens from the different regions within the Gulf of Mexico (Table 5). Ten composite haplotypes were unique to the Pan- handle region, four to the Tampa Bay region, and five to Brazil. Four were observed in all three locations. G- tests of mtDNA haplotype frequencies for temporal and geographical heterogeneity among specimens from within the Gulf of Mexico did not allow rejection of the null hypothesis of population homogeneity (P>0.05). Pooling the Gulf specimens by location and testing for heterogeneity between locations using Monte Carlo sta- tistics (100 randomizations) generated 35 GH values exceeding the initial value of 25.774, again indicating that the null hypothesis of homogeneity for specimens taken within the Gulf of Mexico cannot be rejected (P>0.35, 2SE=0.09). Pooling all Gulf specimens and testing against the Brazilian specimens using Monte Carlo statistics resulted in a GH value of 44.282 which was not exceeded by any of the 100 randomizations, indicating significant heterogeneity between specimens from those locations (P<0.01 ). Discussion Our results indicate an evident lack of genetic stock structuring between the commercially-fished popula- tions of Spanish sardine in the eastern Gulf of Mexico, even though there is a significant difference in the average number of gill rakers between locations, in addition to the morphometric differences reported by Johnson & Vaught ( 1986). This significant variation in gill-raker counts relative to SL appears to confirm Whitehead's ( 1973 ) reference to two rates of increase in the number of gill rakers with increasing standard length among Spanish sardines of the western Atlan- tic, although the lowest counts (Panhandle) were lower than those Whitehead (1973) mentioned. Whitehead ( 1973 ) thought this variation in gill-raker counts in- dicative of the presence of two valid species of Spanish sardines in the western Atlantic, but our mtDNA analy- sis does not support his suggestion. 366 Fishery Bulletin 91(2). 1993 Table 2 Approximate sizes (base pairs) of restriction fragments from mtDNA of Sardinella aurita. Letters designate genotypes observed with each enzyme. "A" genotype was the most common "or each enzyme. Apa -I Bam H-I Dra-\ EcoR-l Hind-m A B A A B A B A B C 6457 6457 16356 10233 8434 5621 5621 4169 3342 6607 7405 7405 5623 5623 4069 4069 3342 3296 4169 1948 1948 438 1798 3436 3436 3296 2661 2661 360 460 438 1318 1318 2661 2251 2104 107 665 620 525 1147 665 2104 436 331 2104 1929 436 331 436 331 16277 16270 16356 16294 Pst-l 16293 16254 16256 16339 16350 Xho-l 16308 A B C D E F G A B C 4880 7100 7100 4360 4880 8184 4880 10965 9593 16274 4360 4360 4880 4200 4360 4880 4360 5309 5309 3850 4200 4360 3850 3800 3150 3850 1372 3150 564 3150 564 3000 150 2350 750 16249 16224 16340 16124 16190 16214 16190 16274 16274 16274 Pvu-B Sac-I A B C D E F G H A B 6461 6461 10133 10133 6461 6461 6461 10753 7786 15230 4265 3672 2238 2238 3672 4770 3672 2238 7444 912 2238 2238 1779 1779 2238 2238 1878 1778 912 1778 1778 1188 912 1779 1779 1778 912 912 1188 912 690 912 912 1188 505 505 912 230 505 690 230 912 230 230 230 230 505 230 360 230 16389 16479 16480 16487 16487 16390 16479 16416 16142 16142 Xba-1 A B 8317 6905 4809 4809 2248 2248 1402 685 685 335 16394 16384 The intraspecific sequence variability of mtDNA of S. aurita is comparable to that observed in most other marine fishes. With a mean sequence divergence of about p=0.51 (Nei & Li 1979), the mtDNAs of Spanish sardines of the Gulf of Mexico fall within the range of mean values observed among other clupeids, i.e., be- tween 0.37% for shad (Bentzen et al. 1989) and 2.4% for menhaden (Avise et al. 1989). In both shad (Roff & Bentzen 1989) and menhaden (Bowen & Avise 1990), population-level mtDNA variability was observed Tringali and Wilson mtDNA analysis of two stocks of Sardmella aurita 367 Table 3 Site matrices inferred from RFLPs of Sardinella aurita. Letters denote restriction patt :rns Column numbers are the restriction sites observed over all individual s for the gi ven enzyme. I aside the matrices: l=site presence 0=site absence. Pst-l Xba-l Hit d-III A 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 0 1 1 0 0 1 0 1 1 1 1 10 11 1 1 1 1 B 1 1 0 0 0 1 1 1 1 1 1 1111 1 1 1 1 C 0 1 0 0 0 10 10 1 1 1 1 D 1 1 1 0 0 E 0 1 1 0 1 F 0 0 1 0 0 G 0 1 1 1 0 Puu-Il Dra-l Apa-l A 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 1 0 1 0 1 0 1 1 1111 1 B 1 1 0 0 1 1 1 1 1111 0 C 0 1 0 0 D 0 1 1 0 E 1 1 1 0 F 1 0 0 0 G 1 1 0 1 H 0 0 1 0 £coR-I Xho-I Sac-l A 40 41 42 43 44 45 46 47 48 49 50 51 52 1 1 1 1 1 1 1 1 0 1 1 1 1 B 1 1 1 1 1 0 1 1 1 1 1 0 1 C 1 0 0 within the geographic regions sampled. Thus, the evi- dent lack of significant heterogeneity in mtDNA haplotype frequencies among the Gulf specimens of S. aurita is probably not due to an inadequate level of mtDNA genome sampling relative to the sequence vari- ability present. Accepting that the mtDNA variability among sample locations of S. aurita was sufficiently characterized, at least two possibilities could account for the observed genetic homogeneity between our sample locations in eastern Gulf of Mexico. This lack of genetic structur- ing is the result of recent or ongoing genetic exchange between the sampled locations, or there may have been insufficient time for discrete Gulf populations, if they do exist, to diverge through mtDNA lineage sorting. The first explanation seems the more likely one, con- sidering the life history of the species. S. aurita is known to be highly migratory (Hildebrand 1963, Simpson & Gonzalez 1967) and to spawn repeatedly throughout much of the year (Grail 1984). Although the pattern of adult migration of S. aurita in the east- ern Gulf is still poorly understood, spawning occurs continuously between the Panhandle and Tampa Bay with no apparent geographic partitioning in either the distribution or abundance of eggs or larvae (Houde 1976). There are no known physical barriers to dis- persal or migration in this region. Significant geographic variation in mtDNA haplotype frequencies does exist between Spanish sardines in the eastern Gulf of Mexico and waters off southern Brazil, based upon the results of goodness-of-fit test- ing. This variation is due to two related factors. One is that the haplotype most common among the pooled specimens from the eastern Gulf of Mexico, AAAAAAAAA, possessed by 46% of the specimens, has a much lower occurrence (6%) among the specimens from southern Brazil. The other is the difference in the distribution of the Pst-l digestion pattern. Whereas 5 of 16 (31%) of the Brazilian specimens possessed the G-type pattern (Table 2), this pattern was absent in the Gulf specimens (re=57). Relative to the common A- type pattern of Pst-l present among the Gulf speci- mens, this G-type pattern represents a site gain at Pst-I. 368 Fishery Bulletin 91 12). 1993 Table 4 Frequency distribution of the 24 composite haplotypes from 73 Sardinella aurita from the western Atlantic. Restriction enzymes are in the following order: Pst-l, Xba-\, Hind-Ill, Pvu-ll, Dra-l, Apa-l, £coR-I, Xho-l, and Sac-I. Frequency dis- tribution of pooled haplotypes from the Gulf of Mexico differ from the Brazilian at P<0.01 (GH=44.282). Composite Haplotype digestion Panhandle Tampa Bay Santos number profile Florida Florida Brazil 1 AAAAAAAAA 14 12 1 2 BAABAAAAA 6 2' 3 AAADABAAA 4 AABAAAAAA 5 BAACBAABA 6 BBABAAAAA 7 BAACAAAAA 1 8 CAAAAAAAA 9 DAAAAABAA 10 BAAFAAAAA 11 CAABAAAAA 4 3 12 AAAAAAACA 13 AAAEAAAAA 1 1 14 EAAAAAABA 15 AAAEABBAA 16 BAABAAAAB 1 17 FAAAAAAAA 1 18 BAAGAAAAA 1 19 AAAFAAAAA 1 20 AAAHAAAAA 2 21 GACAAAAAA 2 22 GAAAAAAAA 3 23 BAAAAAAAA 1 24 AAAAABAAA 1 29 as S. brasil 28 16 ensis by gill-raker including specimen identifiec count. We interpret the significant heterogeneity in haplotype frequencies between specimens from the east- ern Gulf of Mexico and southern Brazil as genetic stock differences between Spanish sardines of the western North and South Atlantic, rather than as species dif- ferences as, for example, between S. aurita and S. brasiliensis. The one specimen of this study referable to S. brasiliensis, with a gill-raker count of 154 at 150mmSL, came from southern Brazil, but it had one of the mtDNA composite haplotypes (haplotype 2; Table 4) present among specimens from the eastern Gulf of Mexico. All other specimens from both Gulf and Bra- zilian waters possessed the lower gill-raker counts typi- cal of S. aurita. This interpretation also agrees with results of the electrophoretic study of Wilson & Alberdi ( 1991 ). After comparing 350 specimens from the eastern Gulf of Mexico with 41 specimens of Spanish sardine from off Table 5 Tests for temporal an d geographic heterogeneity of Sardinella aurita mtDNA haplotypes. GH values were generated by com- paring the complete distribution of mtDNA haplotypes for each group. "V" statistics were calculated by comparing the frequency of angul ar-transformed Haplotype 1 between regions. Sample G„ df V df Tampa Bay vs. Panhandle 24.278 17 0.01 1 Tampa Bay vs. Brazil 29.302' 12 9.68- 1 Panhandle vs. Brazil 39.386' 19 11.49" 1 Pooled Gulf of Mexico vs. Brazil 44.282' 23 10.90" 1 (from tabulated critical values of the x2 * P<0.05, ** P<0.001 distribution). southern Brazil at 36 presumptive gene loci, they found no genetic evidence of more than one species of Span- ish sardine in those two areas. That is, there were no fixed allelic differences that would suggest the pres- ence of fully segregated gene pools, as in two species. Neither in the present study nor in that of Wilson & Alberdi (1991), which used the same Brazilian speci- mens as here, was our nominal S. brasiliensis distin- guishable from S. aurita by electrophoretic genotype or mtDNA haplotype. We think it likely that S. brasiliensis is conspecific with S. aurita. Whitehead (1970) removed S. brasiliensis from the synonymy of S. aurita on the observation that the syntypes of S. brasiliensis from off Brazil had both high and low gill-raker counts at similar lengths. That is, he concluded that the syntypes with low gill-raker counts were S. aurita and that the others with high gill-raker counts were S. brasiliensis of Steindachner. Whitehead (1973) again proposed the specificity of S. brasiliensis on this character based on his review of published data on gill-raker counts of western Atlantic Sardinella, but he did not report data from any additional specimens. Whitehead (1985) pro- posed that S. brasiliensis might be distinguished from S. aurita based on the shape of the gill rakers forming the branchial basket, i.e., flattened in S. aurita vs. curled in S. brasiliensis, but he did not specifically connect this character to catalogued specimens which he examined. No mention was made of this character in his prior discussion of the syntypes (Whitehead 1970). To our knowledge, no thorough quantitative study of variation in gill-raker counts with SL among Brazil- ian sardines has been published. Montero & Perez Tringali and Wilson mtDNA analysis of two stocks of Sardinella aurita 369 (1981) conducted a protein electrophoretic comparison at 24 presumptive loci of Brazilian Sardinella which had been first identified as either S. brasiliensis or S. aurita by high or low gill-raker counts, respectively. They reported no electrophoretic differences between the two putative species, which is consistent with our finding of significant variation in gill-raker counts be- tween specimens from the Florida Panhandle and Tampa Bay in the apparent absence of any genetic differences (Table 1). Thus, the higher gill-raker counts which supposedly distinguish S. brasiliensis from S. aurita are probably due to non-genetic causes, and are of no greater taxonomic importance than the intraspe- cific variation in gill-raker counts observed between S. aurita inhabiting the Panhandle and Tampa Bay loca- tions of the eastern Gulf of Mexico (Table 1). Until the present classification of Western Atlantic Sardinella is supported by a thorough study of variation in the char- acters on which it is based, the validity of S. brasiliensis as a species is in question. Given the non-reticulate nature of mtDNA trans- mission, it is not difficult to understand intraspecific divergence when there is a mechanism between popu- lations affecting the balance between gene flow and genetic drift. There are a number of potential barriers to gene flow within the trans-equatorial range of S. aurita, including diverging oceanic currents (e.g., the North Equatorial current) and the major river plume of the Orinoco which can extend far out into the Carib- bean. Even without discrete physical barriers, the sto- chastic process of mtDNA lineage survival or extinc- tion might in itself be enough to alter haplotype frequencies in this continuously-distributed species if the effective gene flow is low relative to the sizes of the populations or their geographic ranges (Neigel & Avise 1986). However, the mtDNA variability among S. aurita might be a product of differential selective forces because the species is distributed over so great a latitudinal (environmental) gradient. For the Span- ish sardine of the western Atlantic, there are not yet enough data to choose between the alternative hypoth- eses of drift vs. selection in accounting for the evident genetic structuring. To our knowledge, this report represents the first substantiated case of a holopelagic teleost with a sup- posedly continuous distribution between sample regions (Gulf of Mexico to southern Brazil) exhibiting signifi- cant heterogeneity in mtDNA haplotype frequencies. Whereas other studies (e.g.. Graves & Dizon 1989) have compared sample populations of holopelagic fishes re- moved by great longitudinal distances, or on opposite sides of land masses (Ovenden et al. 1989), ours is the first to compare specimens at near-opposite ends of a trans-equatorial distribution spanning many degrees of latitude (30°N to 26°S). Not only do these results further confirm the potential usefulness of mtDNA in studies of hypothesized fishery stocks, but also sug- gest the need for additional studies on fishes having similar distributions. Acknowledgments We thank Dan Woodsen, Harry Mofield, and all those who assisted in the collection of S. aurita specimens from the Gulf of Mexico. We also thank Drs. Y. Matsuura (Instituto Oceanographi, Sao Paulo, Brazil) and Frank Miiller-Karger (University of South Florida) for help in obtaining the Brazilian specimens. Joyce C. Miller kindly provided the Restsite v. 1.2 computer pro- gram. Partial support for this research came from an Aylesworth Marine Scholarship (Florida Sea Grant) and from grants-in-aid-of-research to MDT from Sigma- Xi, Cortez Chapter of the Organized Fishermen of Florida, the USF Department of Marine Science, and by NOAA Grant #NA89WC-H-MF028 (MARFIN) to RRW. Citations Avise, J. C, B. W. Bowen, & T. Lamb 1989 DNA fingerprints from hypervariable mitochon- drial genotypes. Mol. Biol. Evol. 6(3):258-269. Bentzen, P., W. C. Leggett, & G.C. Brown 1989 Mitochondrial DNA polymorphism, population structure, and life history variation in American shad Alosa sapidissima. Can. J. Fish. Aquat. Sci. 46:1446- 1454. Bernatchez, L., & J. J. Dodson 1990 Mitochondrial DNA variation among anadromous populations of cisco, Coregonus artedii, as revealed by restriction analysis. Can. J. Fish. Aquat. Sci. 47:533-543. Billington, N., & P. D. Hebert 1991 Mitochondrial DNA diversity in fishes and its im- plications for introductions. Can. J. Fish. Aquat. Sci. 48(Suppl. 11:80-94. Bowen, B. W., & J. C. Avise 1990 Genetic structure of Atlantic and Gulf of Mexico populations of sea bass, menhaden, and sturgeon: Influence of zoogeographic factors and life- history patterns. Mar. Biol. (Berl.) 107:371-381. Chapman, R. W., & D. A. Powers 1984 A method for the rapid isolation of mitochondrial DNA from fishes. Tech. Rep. UM-SG-TS-84-05, Mary- land Sea Grant Coll., Univ. Maryland, College Park, lip. DeFlaun, M. F., & J. H. Paul 1986 Hoechst 33258 staining of DNA in agarose gel electrophoresis. J. Microbiol. Methods 5:265-270. 370 Fishery Bulletin 9 1(2). 1993 Desalle, R., A. Templeton, I. Mori, S. Pletscher, &d J. S. Johnston 1987 Temporal and spatial heterogeneity of mtDNA poly- morphisms in natural populations of Drosophila mercatorum. Genetics 116:215-223. Elder, J. K., & E. M. Southern 1987 Computer-aided analysis of one dimensional re- striction fragment gels. In Bishop, M.J., & C.J. Rawlings (eds.), Nucleic acid and protein sequence analysis: A practical approach, p. 165-172. IRL Press Ltd., Oxford, Wash DC. Ferris, S. D., & W. J. Berg 1987 The utility of mitochondrial DNA in fish genetics and fishery management. In Ryman, N., & F. J. Ut- ter (eds.). Population genetics and fishery manage- ment, p. 277-299. Univ. Wash. Press, Seattle. Fisher, W. 1978 FAO species identification sheets for fishery pur- poses (Fishing area 31), vol. II. FAO, Rome. Freeman, M. F., & J.W. Tukey 1950 Transformations related to the angular and the square root. Ann. Math. Stat. 21:607-611. Grail, C. 1984 A study of the biology of the Spanish sardine, S. aurita, in Florida waters. Master's thesis, Univ. Mi- ami, Miami FL, 107 p. Graves, J. E., & A. E. Dizon 1989 Mitochondrial DNA sequence similarity of Atlan- tic and Pacific albacore tuna, Thunnus alalunga. Can. J. Fish. Aquat. Sci. 46:870-873. Hildebrand, S. F. 1963 Genus Sardinella Cuvier and Valenciennes 1847, Spanish sardines. In Bigelow, H.B. (ed.), Fishes of the western North Atlantic. Mem. Sears Found. Mar. Res., Yale Univ. 1 (3):397-410. Houde, E. D. 1976Abundance and potential for fisheries development of some sardine-like fishes in the eastern Gulf of Mexico. Proc. Gulf Caribb. Fish. Inst. 28:73-82. Johnson, A. G., & R. N. Vaught 1986 Species profile of Spanish Sardine (Sardinella aurita). NOAA Tech. Memo. NMFS-SEFC-187, NMFS Southeast Fish. Sci. Cent., Miami, p. 142-153. Lansman, R. A., R. O. Shade, J. F. Shapina, & J. C. Avise 1981 The use of restriction endonucleases to measure mitochondrial DNA sequence relatedness in natural populations. III. Techniques and potential applications. J. Mol. Evol. 17:214-226. Montero, G. V., & J. E. Perez 1981 Relaciones taxonomicas entre algunas especies de la familia Clupeidae (Pisces). Biol. Inst. Oceanogr. Venez. Univ. Oriente 20 ( l&2):79-84. Nei, M., & W. H. Li 1979 Mathematical model for studying genetic varia- tion in terms of restriction endonucleases. Proc. Natl. Acad. Sci. USA 76:5269-5273. Nei, M., & J. C. Miller 1990 A simple method for estimating average number of nucleotide substitutions within and between popu- lations from restriction data. Genetics 125:873-879. Neigel, J. E., & J. C. Avise 1986 Phylogenetic relationships of mitochondrial DNA under various demographic models of speciation. In Karlin, S., & E. Nevo (eds.), Evolutionary processes and theory, p. 515-534. Academic Press, NY. Ovenden, J. R., A. J. Smolenski, & R.G. White 1989 Mitochondrial DNA restriction site variation in Tasmanian populations of Orange Roughy (Hoplostethus atlanticus), a deep-water marine teleost. Aust. J. Mar. Freshwater Res. 40:1-9. Roff, D. A., & P. Bentzen 1989 The statistical analysis of mitochondrial DNA polymorphisms: Chi-square and the problem of small samples. Mol. Biol. Evol. 6:539-545. Simpson, J. G., & G. G. Gonzalez 1967 Some aspects of the early life history and environ- ment of the sardine, Sardinella anchovia, in eastern Venezuela. Minist. Agric. Cria, Invest. Pesq. Serv. Rec. Explor. Pesq. 1 (21:38-93. Sokal, R. R., & F. J. Rohlf 1981 Biometry, p. 691-767. W.H. Freeman, NY. Sutter, F. C, & B. M. Mahmoudi 1992 Investigations of inshore and offshore population dynamics of Spanish sardines along the central west coast of Florida. Unpubl. final rep., NMFS/MARFIN, Award NA90AA-H-MF749, NMFS Southeast Fish. Sci. Cent., Miami FL, 59 p. Tringali, M. D. 1991 Intraspecific divergence of mitochondrial DNA within the continuous western Atlantic range of the Spanish sardine, Sardinella aurita. Master's thesis, Univ. South Florida, Tampa, 43 p. Whitehead, P. J. 1970 The clupeoid fishes described by Steindachner. Bull. Br. Mus. Nat. Hist. (Zool.) 20 (I):l-46. 1973 The clupeoid fishes of the Guianas. Bull. Br. Mus. Nat. Hist. (Zool), Suppl. 5:1-219. 1985 Clupeoid fishes of the world (suborder clupeoidei). FAO Fish. Synop. 125 (7):l-303. Wilson, R. R. Jr., & M. D. Tringali 1990 Improved methods for isolation of fish mtDNA by ultracentrifugation and visualization of restriction fragments using fluorochrome dye: Results from Gulf of Mexico clupeids. Fish. Bull., US 88:611-615. Wilson, R. R. Jr., & P. D. Alberdi Jr. 1991 An electrophoretic study of Spanish sardine sug- gests a single predominant species in the eastern Gulf of Mexico, Sardinella aurita. Can. J. Fish. Aquat. Sci. 48(51:792-798. Estimated drift gillnet selectivity for albacore Thunnus alalunga Norman Bartoo David Holts La Jolla Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service. NO/V\ P.O. Box 271, La Jolla, California 92038-0271 Drift gillnet (DGN) fisheries, some directed toward albacore Thunnus alalunga and some toward other species, have operated in the north and south Pacific Ocean. These fisheries, some of which had their beginnings decades ago, expanded in the early 1980s and exerted a substantial, but as yet unknown, mortality on Pacific albacore popu- lations. Knowledge of the selectiv- ity of DGNs is useful for applying complex population-dynamics tech- niques to the albacore population. Several DGN experimental sur- veys conducted in the north and south Pacific Ocean have produced data on albacore distribution and gear selectivity. In 1984-85 the Southwest Fisheries Science Cen- ter of the National Marine Fisher- ies Service (NMFS) undertook an experiment to quantify DGN mesh- size selectivity for albacore. The ex- periment produced size-selectivity data for a single mesh size used in the U.S. DGN commercial fishery for albacore ( 184 mm; all mesh sizes referred to are stretch mesh mea- surements), in the eastern North Pacific Ocean. Other experiments designed to sample albacore were conducted by the Japan Marine Fishery Resource Center in 1980 (JAMARC 1983), 1982 (JAMARC 1985), and 1983 (JAMARC 1986). These experiments used other mesh sizes which produced complimen- tary data used in this study. The objective of this analysis was to quantify drift gillnet mesh-size selectivity for albacore, using exist- ing data collected from various ex- periments and surveys in the Pa- cific Ocean. Methods Various experiments providing data have been conducted independently at different times and in different areas (Fig. 1). Consequently, some assumptions had to be made con- cerning comparability of nets used and availability and vulnerability of fish sizes to be sampled. However, there was enough information to provide insights into net selectivity for albacore. The nets used in all the experiments were similar to typical Japanese commercial drift nets. All nets had similar mesh- hanging ratios, twine/mesh-size ratios, and were of multifilament construction (Table 1). These are the most important factors affecting the selectivity curve for a given mesh size(Hamley 1975). Detailed descriptions for the JAMARC experiments, including set locations, timing, mesh size placement, etc, are contained in the JAMARC references. These experi- ments were designed to obtain in- formation on resource distribution and catch rates for possible future commercial fisheries. In brief, the JAMARC North Pacific experiment was a lyr survey beginning in the spring of 1980 (Fig. 1). It targeted albacore using nets with mesh sizes of 130-250mm. The JAMARC 1982 and 1983 experiments in the South Pacific were conducted in the area shown in Fig. 1 and targeted all tuna-like species, using 180 mm mesh nets in 1982, and 118 mm and 160-180 mm mesh nets in 1983. The NMFS experiment targeted albacore. A total of 27 night sets was made: 12 sets in waters south of Point Conception and within 300 mi of shore; 7 sets in the vicin- ity of the Guide Seamount south of the Farallon Islands; and 8 sets approximately 1000 mi west of Or- egon near 142° W long, and 45° N lat. (Fig. 1). All sets used 184mm mesh nets. In all the experiments using mesh sizes of >130mm, the amount of gear fished during each set ranged from 2760 m to 10,800 m, and the number of sets ranged from 11 to 176 (Table 1). Nets in each experi- ment were deployed and retrieved as in typical commercial operations. Nets were deployed at dusk, and retrieval began at first light often extending into midmorning. Soak times were -10-16 h/set. Data collected Data published by JAMARC are limited to fork length (FL) frequen- cies by mesh size by set. The NMFS data include fork length-frequencies, maximum or largest body girth, and opercular girths for each fish, and notation on the mode of capture for each fish (tangled, wedged, or gilled). Fish were described as wedged or gilled when inserted into a mesh so that the entire perimeter of the mesh held the fish. This oc- curred at locations on the body from the snout to the point of maximum body girth. Fish were described as tangled when snared or tangled by fins or tail. Length-frequencies for U.S. troll-caught fish from the same time and areas are available for Manuscript accepted 9 February 1993. Fishery Bulletin, U.S. 92:371-378 ( 1993). 37) 372 Fishery Bulletin 91(2). 1993 aO° 100° 120° 140° 160° 180° '60° .140° 120° 100° 80" 60° 80° 100° 120° 140° 160° 180° 160° 140° 120° 100° 80° Figure 1 Areas sampled for albacore Thunnus alalunga by drift gillnets. (A) JAMARC 1980 survey (JAMARC 1983); (B) JAMARC 1982 survey (JAMARC 1985); (C) JAMARC 1983 survey (JAMARC 1986); (D) NMFS 1985 experiment. Table 1 Characteristics of drift gillnets used, amount of gear fished, and alalunga :aught for selected stretch mesh sizes. Twine Hang-in-ratio size (% stretch mesh Mesh (dinier/ Net Distance/ No. Survey size strand ) material hanging) set JAMARC 1980 130 210/21 Nvlon 56.1 (JAMARC 1983) 150 210/24 Nylon 56.1 North Pacific 160 210/27 Nylon 56.1 170 210/30 Nylon 55.5 180 210/33 Nylon 56.1 200 210/36 Nylon 56.1 All 178 JAMARC 1982 180 210/30 Nylon 55.3 38 (JAMARC 1985) South Pacific JAMARC 1983 118 210/10 Nylon 11 (JAMARC 1986) South Pacific NMFS 1985 184 210/18 Polyester 50.0 27 North Pacific comparison. Sample sizes of al- bacore caught by mesh size for each experiment, as well as the quantity of gear fished, are shown in Table 1. Results JAMARC 1 980 data Gear selectivity can only be esti- mated because there is no infor- mation on the actual size struc- ture of the North Pacific albacore population (Hamley 1975). Data from several mesh sizes fished simultaneously and assumptions about shape, variance, and effi- ciency of the selectivity curves are required. The JAMARC 1980 experiment provided the only published dataset containing this information. Length-frequencies obtained in the 1980 experiment for mesh sizes 130, 150, 160, 170, 180, and 200 mm are shown in Fig. 2 (A-F). In all these mesh sizes, albacore captured had size modes at 53, 62, and 78cmFL. Two options are available for calculating an indirect selectiv- ity curve: (1) Fitting a pre- determined distribution function to data, using a method similar to that of Ishida (1962); or (2) estimating the selectivity of dif- ferent meshes to one common size-class (mode) of fish and ex- trapolating to other mesh sizes (Regier & Robson 1966). Lack of an objective mathematical func- tion in determining the shape of the selectivity curve in the first method, and the need for fitting the selectivity curve "by eye," re- quired that a combination ap- proach was employed. A basic assumption of gillnet selectivity analysis is that selec- tivity is the same for those com- binations of length interval L(i) and mesh size M(j ) where the ra- tios of L( i l/M(j ) are equal ( Regier NOTE Bartoo and Holts Drift gillnet selectivity for Thunnus alalunga 373 A: 130 mm 14 - N = 438 12 - 10- 8 - 6 - / 4 - I l\ 2 - J y vw. 7 60 80 FORK LENGTH (CM) 60 80 FORK LENGTH (CM) C: 160 mm »T 10 - N = 672 8 - - 6 - 4 - 2 - 0 < -t 1 U-)^^^^^m^B— 60 80 FORK LENGTH (CM) 60 80 FORK LENGTH (CM) 60 80 FORK LENGTH (CM) 60 80 FORK LENGTH (CM) Figure 2 Percent frequency-of-occurrence of albacore Thunnus alalunga by fork-length interval by stretch-mesh size. Data from the 1980 JAMARC survey (JAMARC 1983). (A) 130mm mesh; (B) 150mm mesh; (C) 160mm mesh; (D) 170mm mesh; (E) 180mm mesh; (F) 200mm mesh. 374 Fishery Bulletin 91(2), 1993 & Robson 1966). This assumes that the length:girth ratio is the same for all fish. Each fork-length interval was divided by the appropriate mesh size of capture, and then multiplied by the appropriate observed fre- quency. This produced "generic" size-frequencies by mesh size. Each resulting size-frequency interval, by mesh size, was normalized to the greatest frequency. The resulting frequencies, one for each mesh size, were then scaled to an assumed 100% efficiency at the peak frequency sampled (the mid-60 cm mode); thus, differ- ential sample sizes by mesh are taken into account. If all mesh sizes have the same selectivity curve, and no sampling variance, they would all be the same. Rather than using the method of Ishida to describe a common selectivity curve (Regier & Robson 1966), the "mean" shape of the selectivity curve was esti- mated from the selectivity of the different mesh sizes for the largest (78cmFL) and smallest (53 cmFL) modes captured by all mesh sizes. The curve obtained for the 53 cm mode described the left or ascending limb of the selectivity curve, and the curve for the 78 cm mode described the right or descending limb. Lengths between these two limbs were assumed to be sampled at 100%. This included some lengths with very low percent-frequencies (i.e., we assumed few fish were available) as well as the size mode near 62 cm. There was no assumption that the left and right limbs have the same shape. For the 78 cm mode, the cumulative percent- frequency of capture vs. length of fish caught for all mesh sizes are shown in Fig. 3. When plotted on nor- mal probability scales, cumulative percent-frequencies 09 - 0.8 - o o C 0.7 - >- O 2 °-6- 2 05- OC U- fy/f/ ■ 130 mm yy If/' + 1 50 mm £ 0.4 - UJ «* v 160 mm S 0.3- UJ a 0.2 - a a 170 mm J£jy/ 180 mm 01 - Xy/jS// 200 mm 72 74 76 78 80 82 84 86 88 91 3 FORK LENGTH (CM) Figure 3 Cumulative percent frequency-of-occurrence by fork-length interval for 130mm, 150mm, 160mm, 170mm, 180mm, and 200 mm stretch mesh sizes for the 78cmFL mode of albacore Thunnus alalunga sampled in the 1980 JAMARC survey (JAMARC 1983). PERCENT FREQUENCY \ 69 71 73 75 77 79 61 83 85 87 89 FORK LENGTH (CM) Figure 4 Mean selectivity-curve right limb for albacore Thunnus alalunga from the 78cmFL mode sampled in the 1980 JAMARC survey (JAMARC 1983). 48 50 52 FORK LENGTH (CM) Figure 5 Mean selectivity-curve left limb for albacore Thunnus alalunga from the 53cmFL mode sampled in the 1980 JAMARC sur- vey (JAMARC 1983). formed straight lines with virtually the same slopes. An average cumulative percent-frequency (subtracted from 100) provided the shape of the right limb (Fig. 4). This introduced the assumption of a constant variance and normal distribution (Hamley 1975). The same pro- cedure provided the left limb of the selectivity curve which increased more rapidly or had a greater slope (Fig. 5). These two limbs must now be placed appropriately on the X axis for each mesh size. This is accomplished separately for each limb by mesh size. The sampling efficiency at 53.5 cmFL for each mesh size was esti- mated by linear regression. The observed maximum NOTE Bartoo and Holts Drift gillnet selectivity for Thunnus alalunga 375 FORK LENGTH (CM) Figure 6 Linear regressions of observed maximum percent sampling efficiency at 53 cm and 78cmFL against stretch-mesh size for albacore Thunnus alalunga. Table 2 Estimated percent sam pling efficiency for albacore Thunnus alalunga at fork lengths of 53 cm and 78 cm for various stretch mesh sizes. Mesh size Fork length (% efficiency) (mm) 53cm 78cm 130 53.3 21.7 150 42.3 31.2 160 36.8 35.9 170 31.3 40.7 180 25.7 45.4 184 23.5 47.3 200 14.7 54.8 100 - 90 - 1 30 mm mesh 150 mm mesh > 80 - hjj U\\\ 1 60 mm mesh S 70 - o & W' 170 mm mesh £ 60 - O 50 - 5? 40 - 2 < "> 30 - ►- S 20 DC £ 10 - IS- \La- 'SOmmmesn 200 mm mesh ,i" "%*,.. 30 40 50 60 70 80 90 100 FORK LENGTH (CM) Figure 7 Selectivity curves (9c) for albacore Thunnus alalunga for 130 mm, 150 mm, 160 mm, 170 mm, 180 mm, and 200 mm stretch-mesh sizes percent efficiency of sampling at the 53 cmFL peak was regressed against mesh size for all mesh sizes (Fig. 6). The estimated percent sampling efficiency at 53cmFL for each mesh size is shown in Table 2. Using the same method, the percent sampling effi- ciency was estimated for 78cmFL for each mesh size (Fig. 6, Table 2). The family of selectivity curves by mesh size is shown in Fig. 7. These were determined by shifting the mean selectivity-curve limbs to the left or right so that sam- pling efficiencies at 53 and 78cmFL matched the val- ues in Table 2 for each mesh size. Fork lengths between the two limbs were assumed to be sampled with maximum efficiency, consistent with theory asso- ciated with unimodal selectivity curves (Hamley 1975). 18 - T + 184 mm GILLNET 16 - N =892 FREOUENCY ■ TROLL N = 1.144 PERCEWT 1 4 - J | 2 - J^J V^vl. 40 60 80 100 FORK LENGTH (CM) Figure 8 Percent frequency-of-occurrence by fork-length interval for albacore Thunnus alalunga sampled in the 1985 NMFS ex- periment by 184mm stretch mesh and by trolling. FORK LENGTH (CM) Figure 9 Selectivity curve for albacore Thunnus alalunga for 184mm stretch-mesh size. 376 Fishery Bulletin 91(2). 1993 NMFS 1985 experiment Length-frequency data for albacore collected in the 184 mm mesh during the 1985 NMFS experiment are shown in Fig. 8. The estimated selectivity curve is shown in Fig. 9 and, as expected, is almost identical to the 180mm mesh selectivity curve (Fig. 7). For com- parison, length-frequencies from the NMFS experiment taken by both trolling jigs and drift netting in the same area are shown in Fig. 8. As predicted by the selectivity curve (Fig. 9), few fish <50cm length were sampled by the 184 mm mesh even though the fish were assumed available to the net, as indicated by the FORK LENGTH (CM) Figure 10 Percent frequency-of-occurrence by fork-length interval for albacore Thunnus alalunga for 180mm stretch mesh sampled by the 1982 JAMARC survey (JAMARC 1985). 118 mm N =87 go FORK LENGTH (CM) Figure 1 1 Percent frequency-of-occurrence by fork-length interval for albacore Thunnus alalunga for 118mm stretch mesh sampled by the 1983 JAMARC survey (JAMARC 19861. size-frequency sampled by the trolling gear fished in the same area and time. JAMARC 1 982 experiment The 1982 JAMARC South Pacific experiment col- lected data for a 180 mm mesh size. Length- frequency data obtained are shown in Fig. 10. The estimated selectivity curve for this mesh is the same as shown for the 180 mm mesh in Fig. 7. From in- spection of the length-frequencies for 180 mm and 184 mm mesh sizes from the North Pacific, it was apparent that the South Pacific experiment sampled an albacore population with a different size struc- ture. The results show a much greater proportion of fish in the 70-80 cmFL range than in the North Pa- cific sampling. This is also seen in samples taken from the U.S. troll fishery that operated in the South Pacific (Rensink 1991) JAMARC 1 983 experiment Length-frequency data obtained from the 1983 JAMARC South Pacific experiment using 118 mm mesh are shown in Fig. 11. The selectivity relationships de- veloped from the North Pacific data do not accurately predict the expected selectivity curve for the 118 mm mesh. This is a result of small sample sizes and an observed peak size mode well below the sizes of alba- core captured in the mesh sizes used to derive the selectivity curves. Discussion Selectivity curves derived from the JAMARC 1980 ex- perimental data appear similar in shape to selectivity curves presented for salmon Onchorynchus nerka and O. keta, slender tuna Allothunnus fallai, and mackerel Scomberomorus niphonius, and other fusiform fishes (Taguchi 1961, Ishida 1964, Shimazaki et al. 1984). The asymmetry of each selectivity curve (i.e., steeper on the left limb and flatter on the right limb) is simi- lar to that observed for selectivity curves of other fusi- form fishes (Hamley 1975) and is explained by the mechanics of capture. The smallest fish, retained by the net at some particular girth/mesh-perimeter ratio, tend to be caught near maximum girth. Also, the range of sizes providing the ratio is relatively narrow. For larger fish, the same ratio can be found over a larger range offish sizes because the point of capture may be anywhere from the maximum girth forward to the snout tip. For example, maximum girth of albacore increases more rapidly than opercular girth as a func- tion of length (Fig. 12) which extends or skews the NOTE Bartoo and Holts Drift gillnet selectivity for Thunnus ala/unga 377 + 75 - MEAN MAXIMUM GIRTH 70 - ■ MEAN GILL GIRTH N = 1,144 S* 65 - J/* * 2 60 - O s* ^? Z jT^ ^S^^ I 55- i O 50 - y\^ 45 - y^\^^ 40 - y^^ 35 - ^T 40 60 60 100 FORK LENGTH (CM) Figure 12 Relationship of albacore Thunnus alalunga maximum and opercular girths to fork length for 1144 fish sampled by the 1985 NMFS experiment. right limb of the selectivity curve. This skewing can be troublesome when symmetrical or pre-determined selectivity curves are fit to data. The method used in this study minimizes this problem. Assumptions that selectivity curves estimated from the 1980 JAMARC experiment applied to similar nets, such as the NMFS 184 mm mesh and the JAMARC- 1982 180 mm mesh, appear acceptable from the data available. The principal factor limiting the application of these selectivity curves to other albacore popula- tions is the girth/length relationships discussed above (Fig. 12). Significant changes in this relationship will change the steepness of the selectivity-curve limbs and the skewness. Extrapolation of the mean selectivity curve to smaller mesh sizes than those providing the initial data may be improper based on the data observed for the 118 mm mesh. Although the sample size was relatively small, the apparent increase in tangling of larger fish caught in smaller mesh sizes compared with larger mesh sizes suggests that the right portion of selectivity curves for meshes < 130 mm are more skewed. Tangled fish did not appear to be a problem in the NMFS or JAMARC 1980 datasets, presumably because all mesh sizes ei- ther gilled or wedged fish in the largest mode. Addi- tionally, there may have been fewer larger fish avail- able in the survey areas, as shown by troll-caught samples. The absence of larger, tangled fish in both net-caught and troll-caught samples could result from reduced availability or vulnerability of larger fish in the areas sampled. Surface fisheries (i.e., drift net, troll, and pole- and-line) for albacore in the North Pacific produce catches with similar length-frequencies in the same areas (Bartoo & Foreman In press). However, albacore taken by subsurface longline gear at the same time and area are generally larger in length than those taken in the surface fisheries. This indicates albacore show differential availability or vulnerability between surface and subsurface gears. This could be due to spatial separation or behavior which may change with fish size. The results presented here assume that vul- nerability to DGNs is relatively static over time. Acknowledgments We acknowledge contributions to this work by several individuals. Thanks to Tony West and the crew of the F/V Steelfin II for their assistance during the cruises. Thanks to Earl Weber for helping conduct the NMFS drift gillnet experiment, Sonee Sonu for providing Japa- nese-English translations, and to the many reviewers for constructive comments on the manuscript. Finally, thanks to Karen Handschuh for production of the manuscript. Citations Bartoo, N., & T. J. Foreman In Press A synopsis of the biology and fisheries for north Pacific albacore tuna. In Shomura, R. (ed). Report of the FAO Expert Consultation on Interac- tions of Pacific Ocean Tuna Fisheries. FAO Fish. Tech Pap. Hamley, J. M. 1975 Review of gillnet selectivity. J. Fish. Res. Board Can. 32(11):1943-1969. Ishida, T. 1962 On the gill-net mesh selectivity curve. Bull. Hokkaido Reg. Fish. Res. Lab. 25l/4 cara- pace length) posterolateral spines, and pronounced dorsal ridge. No su- praorbital spines. Eyes sessile. Antennule (Fig. 1C) First an- tenna (antennule) with unsegment- ed tubular portion (peduncle) and distal conical projection. Peduncle with ventral plumose seta. Conical projection with 5-6 aesthetascs and a simple seta terminally and two aesthetascs subterminally. Antenna (Fig. ID) Antenna with endopod and scale. Endopod slightly shorter than scale, and tipped with two hook-like projections. Antennal scale unsegmented, with fringe of 9 heavily plumose setae along termi- nal inner margin and prominent spine on distal outer margin. Ventral surface of protopodite with spinulose spine at base of endopod and naked spine at base of antennal scale. Mandibles (Fig. IE) Incisor pro- cess of right and left mandibles tooth-like with minute serrations. Right mandible: anterior margin with small teeth and denticulated projection; dorsal margin of molar with two denticular ridges; posterior margin with two denticulated pro- jections. Left mandible: anterior margin with premolar denticles. Mandibles without subterminal pro- cesses, movable premolar denticle, palp, or palp bud. Maxillule (Fig. IF) Endopod 3- segmented, with 3 setae terminally, long distal seta on second segment, and short distal seta on 1st seg- ment; coxal endite unsegmented with 4 plumodenticulate setae and 3 setae subterminally that some- times have either a few minute spinules or setules; basal endite with 2 elongate, spinelike teeth armed with small denticles and 2 naked setae subterminally; no fine hairs on maxillule. Maxilla (Fig. 1G) Endopod bi- lobed, setation formula 3,1,3; basal and coxal endites bilobed; coxal endite with 7 (sometimes 6) termi- nal and 1 subterminal setae on proximal lobe, 3 terminal and 1 sub- terminal setae on distal lobe, basal endite with 4 terminal and 1 sub- terminal setae on proximal lobe and 3 terminal and 1 subterminal setae on distal lobe; scaphognathite with 4 long, marginal plumose setae; fine Manuscript accepted 2 February 1993. Fishery Bulletin, U.S. 91:379-381 (1993). 379 380 Fishery Bulletin 91(2), 1993 hairs on outer margin of endopod and proximal lobe of coxal endite. Maxilliped 1 (Fig. 1H) No coxal setae, setation formula of basipod 2,2,3,3; endopod 5-segmented, setation formula 3,2,1,2,4+1 (Ro- man numeral denoting subtermi- nal seta); exopod partially seg- mented with 4 plumose natatory setae; endopod barely longer than exopod. Maxilliped 2 (Fig. 1 1) No coxal setae; basipod with 1 distal thin spine, armed with marginal spinules plus 1 naked seta and a naked seta in proximal half; endopod 4-segmented, first 3 seg- ments each with distal thin spine armed with marginal spinules plus 1 plumose seta, fourth with 4+1 plumose setae; exopod incom- pletely 2-segmented, 4 plumose natatory setae. Maxilliped 3 (Fig. 1J) Exopod and endopod undeveloped; exo- pod partially segmented, with 3 undeveloped setae terminally; endopod with undeveloped seta terminally. Figure 1 Stage-I zoea of Lopholithodes mandtii: (A) whole animal, right side; (B) carapace, dorsal; (C) antennule, ventral; (D) antenna, ventral; (E) mandibles (left and right), posterior; Pereopods (Fig. IK) Poorly de- veloped, without exopods; 1st pereopod bilobed; 5th pereopod arises medially between 1st and 2nd pereopods. Abdomen and telson (Fig. 1L) Abdomen with 5 somites and telson (somite 6 fused with telson); somites 2-4 each with 3 pairs of spines, 2 pairs posterodorsal, 1 pair lateral, lateral pair longest; posterodorsal spines of somite 2 generally equal-sized; median pair of somites 2-4 strongest; lateral pair on somite 5 long (-1.4 times somite width), pointed, somewhat sinuate; telson margin convex with median cleft and 8+8 pro- cesses (l,i,3-8), 1st an articulated simple spine, 2nd an anomuran hair (i), 3rd-8th denticulate spines, 4th pair longest, about equal to maximum telson width; all articulated with telson; no uropods or anal spine. Distinction between L mandtii and other lithodid zoeae Stage-I zoeae of L. mandtii are typical of Stage-I lithodid larvae of the northern North Pacific Ocean as characterized by Haynes (1984). The morphological characteristics of Stage-I lithodid larvae are: sessile eyes and 4 natatory setae each on maxillipeds 1 and 2, maxilliped 3 undeveloped and without natatory setae, pleopods and uropods absent, telson and 6th abdomi- nal somite fused. Lithodid species with typical Stage-I zoeae have 4 zoeal stages. Thus, L. mandtii probably has 4 zoeal stages also. Since the review of lithodid larvae of the northern North Pacific Ocean by Haynes (1984), larvae of Hapalogaster dentata have been described by Konishi (1986) who noted that zoeae of H. dentata are most similar to those of Dermaturus mandtii and can be distinguished from the latter by setation of the anten- nal scale (7 vs. 10 setae) and the absence of a minute subterminal spine on the antennal endopod in D. mandtii. Stage-I zoeae of H. dentata are readily dis- tinguished from Stage-I L. mandtii by the short ( N (2) The terms in the right-hand side of Eq. 2 represent the portion of the total variance due to variation within length-strata and that due to variation between strata, respectively. The covariance of p and p\ is derived using the method of Kimura ( 1977): A , . A . A A A9A A , A A A A A AA L-l (I- l)q n , l -l-q q , l I q q p p p,p.)=£ — i — ' V|"'* + X-—^- + X -H '->;'* -qpi Cov(p/Pk =i n y, a a y A A A A l -I- a a , l I q q pp. <=i rc <=i N N The approximate form of covariance omits the first term because this term is small compared with the sum of the other two terms. A quadratic loss function (Jinn et al. 1987) is used to infer the precision of the estimated age composition p'=(pi,p2,...,pA): Ap,p)= (p-p) 'W(fi-p) = X w Var (p)+I«) Cov ( p , p, r a A = E I w (p-p + E JW.ipr-pp^-p,) .(4) The loss function presented in Eq. 4 is identical to Kimura's Vartot provided that W is an identity ma- trix, i.e., wn =1 and wlk-0 for j*k. Substituting Eq. 2 and the approximate form of Eq. 3 into Eq. 4 and collecting terms, we obtain: i / a,-u,\ b+v-m i=i \ n. I N (5) LA A A A v = 51 S^jk l.qij qik- i=l j *k A A . m ^u^PjPk, and where a, u„ b, m, and v are all positive. A linear cost function is used for the optimal sam- pling design: C - c,N + Y.c2lnit (6) where C is total cost, q is per-unit cost of collecting a random length sample, and c2, is per-unit cost for age- ing a fish in the ith length-stratum. Survey designs generally are based on two constraints: (i) a fixed total cost, i.e., minimize the loss function in Eq. 4 at a fixed cost; or (ii) a desired precision of the estimators, i.e., minimize the total cost at a given level of the loss function. Therefore, the problem for optimal allocation becomes one of deter- mining the optimal set of N' and n,"'s which minimizes L at a given total cost or which minimizes total cost at a desired precision level of L (N* and n,* are the opti- mal sample sizes of length and age samples, respec- tively). Kendall & Stuart (1977, Sect. 39.20) and Cochran (1977, Sect. 5.5) show that choosing the opti- mal set of N* and n,"s to minimize L for a fixed C or to minimize C for a fixed L are both equivalent to mini- mizing the product of L and C: 1C l a,-u, b + v-m I— - + ,=1 n, N C;N + X c2,n, (7) Applying the Cauchy-Schwarz inequality to Eq. 7, the product LC is J1C 1 1 | a.-u, V ' I b+v-m \* Al—U.. N L 5 (Vc,A02+X(Vc2a)2 i=l XVc2,-(a,-u,) + ^lb+v- »i Kendall & Stuart (1977) showed that the minimum value of the product LC occurs when Vc21ra! a,-«, ^c2ln, a-u, Vc2LnL Vc~^V~ a,-u, b+v-m N = constant>0 (8) where a, = Xm^ Pq,j (l-q\j), A A> A A L A b = XS^l/q.j-p,,)2, 1=1 j =i Use the terms of the equality between the ith and the (L+l)th terms and rearrange the variables to obtain r^n^/N*, the optimal subsampling ratio between age subsamples and length samples in the ith length stra- tum. The solution of r<" is: 384 Fishery Bulletin 91(2), 1993 cja-u,) c9,(b+v-m) . (9) For a survey design subject to a given total cost C, r,* is the optimal subsampling ratio required to reach the minimum loss function value (min. £). Thus, sub- stitute Eq. 9 into Eq. 6 and solve for the optimal set of N* and n,*: AT: cx+%c2lr\ n, = r\ AT, l /a-u, \ b+v-m min. £ = X — 1 + '=' n, AT (10) For a survey design subject to a desired precision level of £, r,' is the optimal subsampling ratio to reach the minimum total cost (min. C). The optimal set of N* and n,* can be obtained by substituting Eq. 9 into Eq. 5: 1 AT = — l 1 at-u, \ X + b+v-m £ [•A r\ I J n\ = r, AT, L min. C = c,AT + Xc2, n' . (11) Similarly, the above derivation can be extended to the traditional fixed- and proportional-age subsampling schemes. For these age subsampling schemes, the per- unit cost for ageing a fish is not length-specified (i.e., c2i =c2 for all i's). The loss and cost functions in Eq. 5 and 6 are modified according to the definition of the two age subsampling schemes: (1) n=n/L for fixed- age subsampling, and (2) n, =n 1, for proportional-age subsampling, where n = Sn,. The loss function for fixed- age subsampling is X L (a ,-«,) b+v-m £ - H + n N and that for a proportional-age subsampling is (12) £= t X L (aru,)l lt b+v-m (13) n N C = ClAT+ c,n (14) Using the Cauchy-Schwarz inequality, the optimal subsampling ratio (r*) for either minimizing J^'at a fixed total cost C or minimizing total cost C at a desired precision level of £ for a fixed-age subsampling is: r = AT c,IL (aru,) c, (b+v-m) (15) The optimal set of N* and n* and min. £ subject to fixed cost is: AT = cl + c2r , n =r'AT, (16) L X L (a,-u,) b+v-m min. £ = nl + . n N and the optimal set of N* and n' and min. C subject to a desired precision level of ^'is AT Z(a— ut) M + b+v-m n'= r'N", min. 0 = 0,^+ c2n' (17) For proportional-age subsampling, the optimal subsampling ratio (r) for either minimizing £ at a given total cost C, or minimizing total cost C at a desired precision level of £, is: cx X (a-u,) I lt Co (b+v-m) (18) The optimal set of N' and n' and min. £ subject to a given total cost C is (19) AT = ct + c2r ' n' = r'N*, m.£ L X (a,-u ri A )//, + b+v-m AT The cost function for both age subsampling schemes is: and the optimal set of N* and n' and min. C subject to a desired precision level of ^'is: NOTE Lai: Age-length key to estimate age composition of fish population 385 1 L A r X(a,-w,)// n — — 1=1 ' £ f ri = r N~, + (b+v-m. (20) min. C = CjAT + c2n" The solutions given in Eq. 16-20 are similar to that of Lai (1987), provided that the matrix W is an identity matrix. Example The lemon sole (= English sole Pleuronectes uetulus) example of Jinn et al. (1987) is used to illustrate the length-based optimal sampling design. The ALK and per-unit costs are summarized in Table 1. The total cost is C=$229.15, and the per-unit cost for collecting a length offish is c,=$0.15. For simplicity without loss of generality, consider the special case: wjk=0 for j*k. Three different sets of wn are used to reflect different aspects of interest: Table 1 Age-length key and length-frequency distribution of male lemon sole ( = English sole, Pleuronectes uetulus) collected from Strait of Georgi a, British Columbia Original dataset is from Ricker 1 1975:68), and per- unit cost for ageing is rom Jinn et al. 11987). Length Number of Age Number of length Per- age =uu- unit cost (cm) samples 4 5 6 7 8 9 samples of ageing 27 6 5 1 6 1.0 28 9 3 4 2 9 1.2 29 10 4 4 1 1 30 1.4 30 10 1 5 4 51 1.6 31 10 8 2 54 1.8 32 10 1 7 1 1 48 2.0 33 10 1 3 3 2 1 41 2.2 34 10 2 6 1 1 27 2.4 35 10 1 4 3 2 13 2.6 36 6 1 3 2 6 2.8 37 3 1 1 1 3 3.0 38 1 1 1 3.2 Age proportion 0.12 0.48 0.26 0.09 0.04 0.01 Variance ( x 10 3 1.09 2.98 2.46 0.88 0.33 0.04 Case 1 11,1,1,1,1,11 A , P,s; equal interest in estimating all Case 2 (10,30,30,10,1,1): increase precision of four major age-classes with larger Var(p,); Case 3 11,1,1,1,10,601: interest in older but rare age- classes. The results obtained from length-based design are compared with those from fixed- and proportional-age subsampling schemes. The average per-unit cost for ageing a fish (c,) is calculated as the weighted mean of c,„ which is c2=$1.96. The optimal set of IN', n,'| and min. X subject to the given total cost of $229.15 are computed using Eq. 9 and 10 for length-based age subsampling, Eq. 15 and 16 for fixed-age subsampling, and Eq. 18 and 19 for proportional-age subsampling. Precision improves substantially when length-based age subsampling rather than fixed-age subsampling is used in all three cases (Table 2). In the first two cases, however, precision improves marginally by using length-based age rather than proportional-age subsamplings. When rare and older fish (ages 8 and 9, Case 3) are of interest, precision is substantially im- proved by using length-based age instead of propor- tional-age subsampling. This is due to the fact that proportional-age subsampling is not designed to in- crease age subsamples from length-strata consisting of older age-classes. The optimal set of (N\ n,*) and min. C subject to a desired precision level of -Z*=0.01 are computed from Eq. 9 and 11 for length-based age subsampling, Eq. 15 and 17 for fixed-age subsampling, and Eq. 18 and 20 for proportional-age subsampling. Tables 2 and 3 show similar trends. The cost efficiency of length-based age subsampling is superior to fixed-age subsampling in all cases. However, cost efficiency is only marginally 386 Fishery Bulletin 9 1 (2). 1993 Table 2 Optimal sample sizes of length and age samples and min. £ (£. loss function) subject to fixed total cost, C=$229.15, for three different age subsampling schemes. Age-length key dataset is listed in Table 1. Per-unit cost of observing a length sample, c,=$0.15. Case 1 |u>,|=|l.U,l.l,H; Case 2 (H>i)=(10,30,30,10,l,l); and Case 3 |u>,|=| 1,1,1,1,10,601. Length (cm) Case 1 Case 2 Case 3 27 2 2 2 28 4 5 3 29 14 13 10 30 21 23 15 31 15 18 11 32 16 16 12 33 17 16 18 34 9 9 11 35 5 4 13 36 2 1 3 37 1 1 1 38 1 1 1 Length-based n' 107 109 100 sampling NT 183 182 204 design min. £ 0.0058 0.1312 0.0110 Improvement of precision f'i I" vs. Fixed-age subsampling 36.26 Proportional-age subsampling 4.92 39.15 4.23 Fixed-age subsampling n N' ..£ Proportional-age n' subsampling N* min. £ 103 179 0.0061 28.57 19.12 106 106 104 146 141 172 0.0091 0.2156 0.0154 103 103 177 183 0.1370 0.0136 " % = percent difference of min. ./'between length-based sam- pling design and fixed or proportional-age subsampling. Table 3 Optimal sample sizes of length and age samples and minimum total cost (min. C) subject to a desired precision level of loss function, .2=0.01 for the three different age suk sampling schemes. Age-length key dataset is listed in Table 1 . Per-unit cost of observing a length samp e, c,=$0.1£ . |if ,1 for each case are the same as that in Table 2. Length (cm) Case 1 Case 2 Case 3 27 1 3 2 28 3 6 4 29 8 17 11 30 12 30 17 31 9 23 12 32 9 21 13 33 10 21 20 34 5 12 12 35 3 5 14 36 1 2 3 37 1 1 2 38 1 1 1 Length-based n' 63 142 111 sampling N" 106 236 223 design min. C 135.98 300.6 257.25 Cost efficiency (%) " vs. Fixed-age subsampling 35.14 39.09 27.06 Proportional-age subsampling 1.08 3.68 17.33 Fixed-age n' 97 229 160 subsampling N* 133 304 265 min. C 209.66 493.48 352.68 Proportional-age n' 62 141 140 subsampling N* 108 242 249 min. C erence of min 137.46 312.07 311.16 C between length-based sam- " % = percent diff pling design and fixed- or proportional-age subsampling. different between length-based age and proportional age subsamplings in Cases 1 and 2. In Case 3, the cost efficiency of length-based age subsampling increases subtantially over proportional-age subsampling. Discussion To draw a general conclusion, many different sets of w^'s and full matrices of W also were investigated. The results from these additional analyses were similar to that of Tables 2 and 3. In general, the length-based age subsampling is superior to either fixed- or propor- tional-age subsampling. However, precision improve- ment and cost efficiency depend on the weights placed on particular age-classes. Total cost will change in ac- cord with the different weights and desired precision. A higher total cost should be allowed for cases where sampling is designed to improve the precision of highly variable estimates, usually young and old age-classes. A larger budget will increase precision of the esti- mates, especially for highly variable age-classes; how- ever, as Lai (1987) showed, there is a point of dimin- ishing returns as the budget increases (Fig. 1). For the examples used in this paper, precision improvement is marginal when total cost (C) increased beyond $40 for Cases 1 and 3, and beyond $120 for Case 2. Kimura (1989) showed that satisfactory results from cohort analysis can be obtained at low sampling levels (i.e., total cost) provided that the representativeness of the samples can be maintained. It is difficult to compare the methods of Jinn et al. (1987) with those of this paper because the Bayesian approach and classic sampling techniques are derived from different theoretical backgrounds. Nontheless, the results obtained from this paper are similar to that of NOTE Lai: Age-length key to estimate age composition of fish population 387 60 80 100 Total Cost 160 Figure 1 Relationship between minimum loss function value and total cost for the lemon sole (= English sole Pleuronectes vetulus) example. The relationship can be obtained from substituting N' and n,' in Eq. 10 into the third equation in Eq. 10. Jinn et al. (1987) although the values of N* and n,* are different. An advantage of the classic double-sampling technique is the explicit solutions of the optimal set of N" and n,\ which reduces computational effort. In this paper, the optimal sampling design is for stratified-age subsampling. For random-age subsampl- ing in which the number of age subsamples (n) is ran- domly taken from the entire length sample of size N, the estimated variance and covariance (Kimura 1977) are Var(^)=l 'l A /I A , A ,A A N and Cou(prpk A A A AA A — +1 — - — N A PjPk N These two equations are similar to that of proportional- age subsampling in which n,-nl, is substituted into Eq. 2 and 3. However, the estimated variance and co- variance for proportional-age subsampling are approxi- mate, and those for random-age subsampling are not. The similarities of random- and proportional-age subsamplings can be anticipated because N is a ran- dom sample from a population so that E(1,)=E(N,/N)=1„ and n is randomly taken from N so that E(n/n)=N,/N. This indicates that the size of each n, will be approxi- mately proportional to 1„ i.e., n,/n=N,/N=l, and n=n«l, (Kutkuhn 1963). An ALK requires a large random sample of fish from which a length-stratified subsample is collected for ageing. Most fishery data are collected either from surveys in which fish from different tows are sampled or from commercial catches in which fish from differ- ent vessel-trips are sampled. Pooling data over such clusters is necessary because of the cost of data gath- ering. In addition to cluster sampling, fisheries data are frequently stratified into time-area, fisheries (or gears), and sex strata (Kimura 1989). The question is how to make the optimal sampling design of ALK generally applicable. To address this, the following factors must be considered: (1) Need of stratifica- tion, (2) ALK sampling within stratum, and (3) com- bined-strata estimation. Westrheim & Ricker (1978) showed the need for stratification. An ALK obtained from a population at a time-interval should not be universally applied to length-frequency datasets from other populations or other time-intervals if growth and survival rates are different among the populations and time-intervals. Therefore, the factors that may result in differences in growth and survival rates should be evaluated, and stratification should account for these factors. Current sampling programs (e.g., Doubleday & Rivard 1983, Quinn et al. 1983, Kimura 1989) adopted the strategy in which length-frequency data collected from clustered sampling units (e.g., tows or vessel- trips) within a stratum are pooled, from which a length- stratified subsample is collected for ageing. Southward (1963) evaluated an old method (Southward 1963:12) in which a set of length and age data is collected from each landing of a vessel-trip. Because this old method is not developed from a probability sampling design, the within-vessel variability in fish lengths is assumed to be less than between-vessel variability. Southward (1963) showed that this assumption is not valid and the estimated variances of age composition from this old method are so large that little confidence can be placed in it. The length-frequency data pooled over clusters should be a representative sampling of that stratum. Therefore, the weighting factor of each sample should be included in the pooling. Ignoring the weighting fac- tor will bias the estimated age composition (Kimura 1989). Quinn et al. (1983) described a sampling-rate method in which a fixed proportion of halibut were sampled from landings >1000 lbs for length data, and then age data were subsampled from the pooled length samples from these landings. All length samples are self-weighted and can be pooled directly. Quinn et al. (1983) evaluated the methods of com- bined-strata estimation and found that the "project- and-add" method (total catch-at-age is estimated for each stratum and then the estimates are added over 388 Fishery Bulletin 91(2). 1993 strata; see Quinn et al. 1983) produces unbiased esti- mators if all strata are sampled. The project-and-add method uses the concept of a stratified, random sam- pling technique (Cochran 1977). Therefore, Cochran's rules (Cochran 1977:98) of optimal allocation for strati- fied random sampling can be applied. In a given stra- tum, take larger length and age samples if (1) the stratum is larger, (2) the stratum is more variable internally, and (3) sampling is cheaper in the stratum. The first two rules are the basis of the Neyman alloca- tion (Cochran 1977:99). The sampling-rate method pro- posed by Quinn et al. (1983) built upon the first rule and can easily incorporate other rules in the sampling program. It is clear that the number of strata and variability (vartot or loss function) of each stratum should be evaluated first for designing ALK sampling. Then, the total cost is allocated into various strata according to Cochran's rules. Once the total cost for each stratum is determined, an optimal sampling design for ALK can be applied. The sampling rate method of Quinn et al. (1983) can be used to collect the optimal length sample size from clusters. After pooling the length samples, a length-stratified subsample is collected for ageing. Acknowledgments I thank D.K. Kimura for his help with derivation of the covariance. Thanks to M. Sigler for his construc- tive comments on an earlier draft. This paper is par- tially supported by the U.S. Agency for International Development, Fisheries Stock Assessment CRSP (DAN- 4146-G-SS-5071-00). Citations Cochran, W.G. 1977 Sampling techniques, 3rd ed. John Wiley, NY. Doubleday, W. G., & D. Rivard (editors) 1983 Sampling commercial catches of marine fish and invertebrates. Can. Spec. Publ. Fish. Aquat. Sci. 66, 290 p. Jinn, J. H., J. Sedransk, & P. Smith 1987 Optimum two-phase stratified sampling for esti- mation of the age composition of a fish population. Biometrics 43:343-353. Kendall, M., & A. Stuart 1977 The advanced theory of statistics, vol. 3, 3rd ed. Griffin & Co., London, 585 p. Ketchen, K. S. 1949 Stratified subsampling for determining age distributions. Trans. Am. Fish. Soc. 79:205-212. Kimura, D. K. 1977 Statistical assessment of age-length key. J. Fish. Res. Board Can. 31:317-324. 1989 Variability in estimating catch-in-numbers-at-age and its impact on cohort analysis. In Beamish, J. R., & G. A. McFarlane (eds.), Effects of ocean vari- ability on recruitment and an evaluation of param- eters used in stock assessment models, p. 57-66. Can. Spec. Publ. Fish. Aquat. Sci. 108. Kutkuhn, J. H. 1963 Estimating absolute age composition of California salmon landings. Calif. Fish Game Fish. Bull. 120, 29 p. Lai, H. L. 1987 Optimum allocation for estimation age composi- tion using age-length key. Fish. Bull., U.S. 85:179- 185. Quinn, T. J. II, E. A. Best, L. Bijsterveld, & I. R. McGregor 1983 Port sampling for age composition of Pacific hali- but landings. In Doubleday, W. G, & D. Rivard (eds.), Sampling commercial catches of marine fish and in- vertebrates, p. 194-205. Can. Spec. Publ. Fish. Aquat. Sci. 66. Rao, J. N. K., & P. D. Ghangurde 1972 Bayesian optimization in sampling finite popula- tions. J. Am. Stat. Assoc. 67:439-443. Ricker, W. E. 1975 Computation and interpretation of biological sta- tistics offish populations. Fish. Res. Board Can. Bull. 191, 382 p. Southward, G. M. 1963 Sampling landings of halibut for age composition. Int. Pac. Halibut Comm. Sci. Rep. 58, 31 p. Tanaka, S. 1953 Precision of age-composition of fish estimated by double sampling method using the length for stratification. Bull. Jpn. Soc. Sci. Fish. 19:657-670. Westrheim, S. J., & W. E. Ricker 1978 Bias in using an age-length key to estimate age- frequency distributions. J. Fish. Res. Board Can. 35:184-189. Recruitment of bluefish Pomatomus saltatrix to estuaries of the U.S. South Atlantic Bight* Richard S. McBride Rutgers University Marine Field Station Institute of Marine and Coastal Sciences PO Box 278, Tuckerton. New Jersey 08087 Jeffrey L. Ross North Carolina Division of Marine Fisheries PO Box I 550. Manteo, North Carolina 27954 David O. Conover Marine Sciences Research Center State University of New York Stony Brook. New York 1 I 794-5000 The bluefish Pomatomus saltatrix is a pelagic species that is distrib- uted circumtropically (Briggs 1960, Richards 1965, van der Elst 1976, Fable et al. 1981, Pollock 1984, Lenanton & Potter 1987). Ichthyo- plankton surveys off the U.S. east coast indicate three separate spawn- ing concentrations of bluefish, namely (Da spring-spawned cohort produced between March and May in the South Atlantic Bight (SAB), (2) a summer-spawned cohort origi- nating between June and August in the Middle Atlantic Bight (MAB), and (3) a fall-spawned cohort pro- duced in the SAB between Septem- ber and January (Norcross et al. 1974, Kendall & Walford 1979, Powles 1981, Collins & Stender 1987). The abundance of bluefish has fluctuated widely along the Atlan- tic coast during the past century (Baird 1873, Bigelow & Schroeder 1953). Recently, population abun- dance increased steadily from the early 1960s to the late 1970s (Gilmore 1985) and remained rela- tively high during the 1980s (NMFS 1987, 1988). Chiarella & Conover (1990) speculated that increased recruitment success of the spring- spawned cohort is largely respon- sible for this recent increase in overall bluefish abundance. They demonstrated that bluefish off the New York coast during the 1980s were primarily spring-spawned fish, and this contrasts with a more equal proportion of spring-spawned and summer-spawned fish observed by Lassiter ( 1962) in the late 1950s, when overall bluefish abundance was lower. Our understanding of how these three intra-annual cohorts contrib- ute to the overall year-class strength of bluefish along the U.S. Atlantic coast, however, is still un- clear. For example, Kendall & Walford (1979) hypothesized that spring-spawned bluefish were trans- ported from the SAB northward into the MAB estuaries, but Collins & Stender (1987) postulated that spring-spawned fish were trans- ported directly inshore to estuaries of the SAB. Otolith analyses of young recruits have confirmed the northerly dispersal of spring- spawned fish as proposed by Kendall & Walford (Nyman & Conover 1988, McBride & Conover 1991), but similar studies of recruit- ment to SAB estuaries are nonex- istent. Length-frequency data for bluefish from SAB estuaries is also sparse, thus it has been difficult to evaluate if bluefish use estuaries in the SAB as nursery grounds. Here we present evidence of recruitment by three cohorts of young-of-the- year (YOY) bluefish to estuaries and nearshore habitats of the SAB. Materials and methods Field sampling Bluefish <360mm fork length (FL) were collected from estuarine pound- net (a summer fishery) and oceanic trawl (a winter fishery) samples taken in North Carolina (Table 1). The upper limit of 360mmFL repre- sents the largest size attained by young bluefish in their first year of life (Lassiter 1962). Additional data were also obtained from a variety of fishery-independent sources of sam- pling from North Carolina to Florida (Table 1). Otolith aging To avoid dissolution, sagittal otoliths were extracted directly in the field, or from frozen or preserved (95% ETOH) fish. Nyman & Conover (1988) validated that daily incre- ments are present in otolith micro- structure, and we followed their methods of preparation. Sagittae were chosen randomly from YOY specimens of the large fishery col- lections. Eight of the total 51 sagittae prepared were determined to be yearling fish and were excluded from further analysis. Daily sagittal in- crements were counted indepen- dently three times, and a mean count was considered valid if the range was <10% of the mean count. If the range was greater (11% of the time ), then a fourth count was made and the outlier was discarded. Manuscript accepted 22 January 1993. Fishery Bulletin, U.S. 91:389-395 (19931 * Contribution 867 of the Marine Sciences Research Center, State University of New York at Stony Brook. 389 390 Fishery Bulletin 91(2). 1993 Table 1 Sources of bluefish Pomatomus saltatrix samples examined in this study. Size range is given as mmFL. Larger datasets (n>500) were used to compile length-frequency distributions, and some fish from 1987-88 were used for otolith analyses. Other samples ln<13) were obtained specifically for otolith analyses. Bluefish Sampling location Gear (size & mesh) Sampling period Sample size Length range Source' N.C. estuaries Otter trawls March-November 547 21-320 NCDMF Pamlico and Core Sounds (3.2 & 6.1m headropel 1979-90 (Phalen& Stephan 1984) N.C. estuaries Pound-net fishery May-November 4484 100-360" NCDMF Pamlico Sound (Lead 150mm mesh, 1982-85 Oliver's Reef Tunnel 50mm mesh) N.C. coast Trawl fishery November-April 4466 150-360" NCDMF Oregon — Beaufort Inlets (27-33m headrope 1982-85 4.5-18.5m depths 400-1400mm wing mesh 50mm cod end mesh I S.C. coast Otter trawl 2 November 1 101 NCDMF Stono Inlet (6.1m headrope) 1988 S.C. coast 1x2m neuston net April-June 9 39-68 SCWMRD Breech Inlet (2 mm mesh) 1988 S.C. Surface trawl 26 July 13 101-123 use North Inlet (Town Creek) 1988 Nearshore continental shelf Falcon trawl May-November 2276 80-360" SEAMAP Capes Hatteras — Canaveral (9.1m Mongoose type, 1987-88 (Wenner 1989) 4.9-9. lm depths 39-45mm stretch mesh) Most data and specimens supplied by the N.C. Dep. Mar. Fish. (NCDMF); additional material from S.C. Wildl. Mar. Resour. Dep. (SCWMRD), Univ. South Carolina (USC) Belle Baruch Lab., and the Southeast Area Monitoring and Assessment Program (SEAMAP). Data from these sources were truncated at 360mm FL. Mean ring count (= daily age) was subtracted from date of capture to estimate birthdate, assuming no lag between birthdate and date of first ring deposition. Average growth rate was calculated as ([FL-2mm]/daily age); a 2mm constant represents the size-at-hatching (Deuel et al. 1966). Growth rate was also calculated using least-squares regression of FL on daily age. Results Evidence of spring-spawned juveniles Small blue- fish (21-60mmFL) entered North Carolina estuaries in March and April (Fig. 1A). These fish were similar in size to spring-spawned bluefish collected by neus- ton nets on incoming tides at Breech Inlet (South Caro- lina) as early as 29 April 1988, and as late as 16 June 1988 (Table 2). This spring-spawned cohort grew rapidly and re- cruited to the North Carolina pound-net fishery in July (Fig. IB). These fish became progressively larger until the fishery ended in October. A July recruitment of spring-spawned fish was also evident in SEAMAP trawling (Fig. 1C). Bluefish <175mmFL in July were YOY that had been spawned primarily in April (Table 2, Fig. 2). For ex- ample, a sample from 29 July 1987 had a mean birthdate of 19 April ± 13d (±1SD), and a sample for 26 July 1988 had a mean birthdate of 26 April ± 12d (Table 2). Bluefish observed in July, but >200mmFL, showed evidence of an overwinter growth mark on their otoliths. Very few specimens of 175-200mm were avail- able for otolith analysis, which precluded a more precise separation of age from size modes. Bluefish juveniles were larger and older in August samples (Table 2). Spring-spawned YOY bluefish grew 1.2- 1.9mm/d during this time-period (Fig. 3). Evidence of summer-spawned juveniles Small blue- fish (<150-170mmFL) appeared in North Carolina es- tuaries and in nearshore continental shelf waters of the SAB in October (Fig. 1). Small bluefish (< 150mm) collected near Bogue and Stono Inlets in October and November 1988 had a mid-July birthdate (Table 2, NOTE McBnde et al. : Recruitment of Pomatomus saltatrix. to estuaries 391 20. 10 lf ,M| J^HUlM. f J. September n-23 £-3 ,■1.1 ill ll ■■ October n-20 ju juju November n-33 i r"'"i 'i I 10 100 ISO 260 250 300 FORK LENGTH (mm) B May n-49 JufllL June n-238 r>Ml)Ullpil,ll,n. 100 190 200 230 300 350 FORK LENGTH (mm) 100 130 200 250 FORK LENGTH (mm) 100 150 200 250 300 150 FORK LENGTH (mm) Figure 1 Length-frequency histograms of bluefish Pomatomus saltatrix <360mmFL for (A) monthly data pooled over the years 1979-90 from a N.C. Dep. Mar. Fish, trawl survey; (B) 1984 monthly data from the N.C. summer estuarine pound-net fishery; (C) 1988 monthly data from the SEAMAP nearshore trawl survey between N.C. and Florida; and (D) 1984-85 monthly data from the N.C. winter oceanic trawl fishery. Years plotted for the last three data sets are representative of data from other years. Ordinates are either 20 or 30%. Fig. 2). These summer-spawned fish appeared to reach a size similar to that of spring-spawned fish at a com- mon age (Fig. 3), but because they were spawned 2mo later (on average), they were smaller on common sam- pling dates (Fig. 1A,B,C). Evidence of fall-spawned juveniles Winter trawl col- lections did not contain small juveniles between October and January, despite the expectation of an additional length-mode representing fall-spawned fish (Fig. ID). There was very little evidence of fall-spawned fish in any of the length-frequency data, except possibly two fish (~40-60mmFL) collected in October and Novem- ber in North Carolina estuaries (Fig. 1A). Based on otolith ageing, a single specimen collected in August was determined to be 206d old (i.e., with a correspond- ing birthdate of 9 January) which was unique in rela- tion to all other estimated birthdates in this study 392 Fishery Bulletin 91(2). 1993 Table 2 Descriptive statistics for 43 aged specimens of juvenile bluefish Pomatomus saltatrix collected in North and South Carolina (NC and SO. Fork length (mmFL), age (d), and growth rate (GR; mm/d) are given is means. One standard deviation (SD) is provided where more than one specimen was measured. Collection Mean Mean Mean date Location FL SD Age SD GR SD n 1987 15 Jun Pamlico Sound NC 75 74.0 0.986 1 29 Jul Pamlico River NC 119.6 14.5 101.3 13.1 1.169 0.134 5 13 Aug Pamlico Sound NC 208 142.0 1.456 1 17 Aug Pamlico Sound NC 192 1988 125.3 1.516 1 29 Apr Breech Inlet SC 50.7 3.05 62.4 1.21 0.779 1.147 3 17 May Breech Inlet SC 40 51.3 0.740 1 16 Jun Breech Inlet SC 57.8 11.9 51.4 7.53 1.079 0.128 5 26 Jul North Inlet SC 112.7 6.74 91.8 12.0 1.216 0.102 10 2 Aug Pamlico Sound NC 159.5 37.0 1436 8.94 1.088 0.206 4 2 Aug Pamlico Sound NC 220 206 1.058 1 25 Oct Bogue Inlet NC 118.2 31.3 103.2 12.3 1.128 0.064 10 2 Nov Stono Inlet SC 101 110.0 0.900 1 I (/) E tu O cr. LU m 5 z BIRTHDATE Figure 2 Back-calculated birthdate for 43 YOY bluefish Pomatomus saltatrix examined in this study. Date of collection is also shown by month. continental shelf in July. Spring-spawned bluefish ar- rived in SAB estuaries at about the same age (i.e., 40- 60d old) but at an earlier first date of estuarine re- cruitment than in northern estuaries such as New York and New Jersey. Spring-spawned bluefish arrive to the New York Bight no earlier than late May and approxi- mately 60d after hatching (Nyman & Conover 1988, McBride & Conover 1991). Our findings support Collins & Stender's (1987) hypothesis that at least some spring-spawned YOY bluefish are transported directly inshore and enter es- tuaries south of Cape Fear. Combined with previous evidence of dispersal north of Cape Hatteras (Nyman & Conover 1988, McBride & Conover 1991), spring- spawned bluefish appear to recruit to estuaries within a broad geographic region encompassing both the SAB and the MAB. (Table 2, Fig. 2) and may have been produced at the tail end of the "fall-spawning" period. Discussion Spring-spawned YOY bluefish This cohort recruited to the estuaries of North and South Carolina as early as April and as young as 40d old, and continued to ingress as late as mid-June. Spring-spawned bluefish grew rapidly and appeared in the North Carolina pound-net fishery and in trawls over the nearshore Summer-spawned YOY bluefish This cohort also ap- peared in estuaries and in nearshore habitats within the SAB. Summer-spawned fish arrived in October at ~100-150mmFL and were generally older than lOOd. The presence of these summer-spawned fish in the SAB demonstrates that this cohort migrates from their MAB spawning/nursery grounds into the SAB in the fall. Lund and Maltezos (1970) demonstrated that YOY bluefish tagged in New York migrated to the SAB, but these tagged fish were the larger spring-spawned fish. Fall-spawned YOY bluefish This cohort was not clearly observed in length-frequency data and, at most, NOTE McBnde et a\ Recruitment of Pomatomus saltatrix to estuaries 393 • Spring-spawned fish (1987) • If 175- E O Spring-spawned fish (1988) 0 + Summer-spawned fish (1988) r>o £ • + § 125- _l £P+ + W + O u. + 75- 0O # 0 0 25 50 75 100 125 150 175 Age (Days) Figure 3 Fork length (FL) as related to daily age of bluefish Pomatomus saltatrix, based on samples collected near inlets and within estuaries. Linear regression equations for spring-spawned YOY bluefish were FL= -62.8+ 1.86I Age) [r'=0.85] for 1987, and FL= -6.67+1.20(Age) [r2=0.87] for 1988. one specimen appeared in our otolith analysis. The winter data used in our study may be inappropriate for sampling fall-spawned bluefish because the trawl fishery uses a large mesh size and fishes in deep water (Table 1). Presumptive fall-spawned YOY fish were uncommon in previous surveys between North Carolina and Florida (Table 3). Wenner & Sedberry (1989) speculated that they had col- lected fall-spawned fish during Janu- ary and May trawl collections in nearshore habitat of the SAB, but their interpretation may have been con- founded by use of age-length relation- ships from Wilk (1977). Wilk's growth estimates do not account for inter- cohort differences in size-at-first-annu- lus or in growth rates (Lassiter 1962, McBride & Conover 1991). Future analysis of age structure of young blue- fish in the SAB should focus on this cohort. Implications for relative contribution of each cohort We observed polymodal length distributions that represented the size difference between spring- and summer-spawned YOY bluefish cohorts during the fall. Barger (1990) reported that bluefish collected in the SAB had a backcalculated mean FL at Annulus-I of 290mm. This size matches the mean FL at Annulus-I reported for spring-spawned fish from North Carolina in the late 1950s (Lassiter 1962). This observation supports the conclusion that the spring-spawned cohort is of domi- nant abundance relative to the other cohorts, as inter- preted by Chiarella & Conover (1990). Table 3 Total number of young-of-the-year bluefish Pomatomus saltatrix collected in various seining studies from North Carolina to Florida. Number of bluefish is given by month, "Years" refers to the duration of each study, and size range is given in parentheses as mmFL (either as reported or as converted using regressions from McBride 1989). Asterisk indicates that length-frequencies were given as 5mm class-intervals, nd = no data. Sampling frequency ranged from biweekly to bimonthly; seine sizes 12. 2-21. 3m in length. References: (1) Tagatz & Dudley 1961, (2) Anderson et al. 1977, (3) Cupka 1972, (4) Miller & Jorgenson 1969, (5) Futch & Dwinell 1977. Location ( Ref. ) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Years Beaufort NCI 1) Ocean Beach 0 0 0 0 1 (46) 0 1 (100) 0 0 7 12 (65-79) (40-55) 0 1957-60 Beaufort NC ( 1 ) Salt Marsh 0 0 0 0 0 0 1 (136) 0 0 0 0 0 1957-60 Beaufort NC(l) River 0 0 1 (120) 0 4 (66-77) 1 (86) 1 (123) 0 0 0 0 0 1957-60 Folly Beach SC (2) Ocean Beach 0 0 0 0 0 1 (104) 1 (137) 0 0 0 0 0 1969-71 Murrells Inlet — Edisto beaches SC (3) 0 0 0 0 0 44 69-124) 0 0 0 0 0 0 1971 St. Simons I. GA Ocean Beach (4) ' 0 3 2 17 9 6 3 (66-851(96-125) (46-50) (36-90) (36-95) (46-70)196-125) 0 2 (51-60) 0 0 1953-61 Hutchinson I. FL 2 nd Ocean Beach ( 5 ) (325-366 ) 0 nd 4 (273-317 nd 3 42-51) nd 0 nd (70 2 -367) nd 1971-73 394 Fishery Bulletin 9 1 (2), 1993 Fall-spawned fish are the least represented cohort among juveniles and are known mostly from larval surveys (e.g., Kendall & Walford 1979, Collins & Stender 1987). Collins & Stender ( 1987) reported blue- fish larvae <4mm in the SAB to be more abundant in the spring compared with the fall. Other estuarine surveys have collected small bluefish (<100mm) from October-March in the SAB (Table 3) which strongly suggests that at least some fall-spawned fish do re- cruit to estuaries. We suggest that the fall-spawned cohort contributes the least, the summer-spawned cohort contributes somewhat more, and the spring-spawned cohort con- tributes the most to the Atlantic coast bluefish popula- tion, at least in recent years. It is not clear why such differences in abundance among cohorts occur nor for how long such differences persist. The spring-spawned cohort may be more abundant because it is spawned at a place and time enabling invasion of estuaries within the SAB, as well as to estuaries as far north as Maine (Targett & McCleave 1974). The fall-spawned cohort may be less abundant, on average, as it is likely to experience colder temperatures and lower food sup- plies relative to the other cohorts. Estuaries of the MAB may be more important as nursery grounds for bluefish than are those of the SAB. Summer abundances of spring- and summer-spawned YOY bluefish in New York Bight estuaries are appar- ently much higher (0.2-6.3 fish/30m seine haul; McBride & Conover 1991) than in the SAB, based on small numbers of bluefish collected during year-long studies in the south (Table 3). Rountree & Able (1992) also found YOY bluefish to be very common (sixth most abundant fish of 63 spp.) in New Jersey marsh creeks, in contrast to much lower relative abundances in simi- lar polyhaline creeks in the SAB (e.g., Cain & Dean 1976). The wide latitudinal distribution and movements of young bluefish, within and between the Middle and South Atlantic Bights, demonstrate the need to moni- tor and manage bluefish on a large geographic scale (i.e., nearly the entire U.S. Atlantic coast). Acknowledgments We thank E. Day, P. Kenny, V. Ogburn-Matthews, and B. Stender for their help in collecting samples and arranging collecting trips in South Carolina. J. Hare, W. Morse, and R. Rountree offered helpful suggestions on the manuscript. This project was partially funded by grants to D.O.C. from the Office of Sea Grant, Na- tional Oceanic and Atmospheric Administration, un- der grant NA86AA-D-SG045 to the New York Sea Grant Institute and from the New York State Department of Environmental Conservation. Citations Anderson, W. D. Jr., J. K. Dias, R. K. Dias, D. M. Cupka, & N. A. Chamberlain 1977 The macrofauna of the surf zone off Folly Beach, South Carolina. U.S. Fish. Wild! Serv., Spec. Sci. Rep. Fish., 23 p. Baird, S. F. 1873 Natural history of some of the more important food fishes of the south shore of New England. Rep. U.S. Comm. Fish Fish, for 1871-1872, p. 228-252. Barger, L. E. 1990 Age and growth of bluefish, Pomatomus saltatrix, from the northern Gulf of Mexico and U.S. South Atlantic coast. Fish. Bull., U.S. 88:805-810. Bigelow, H. B., & W. C. Schroeder 1953 Fishes of the Gulf of Maine. U.S. Fish. Wildl. Serv., Fish. Bull. U.S. 53:383-389. Briggs, J. C. 1960 Fishes of worldwide (Circumtropical) distribution. Copeia 1960(3):171-180. Cain, R. L., & J. M. Dean 1976 Annual occurrence, abundance and diversity offish in a South Carolina intertidal creek. Mar. Biol. (Berl.) 36:369-379. Chiarella, L. A., and D. O. Conover 1990 Spawning season and first-year growth of adult bluefish from the New York Bight. Trans. Am. Fish. Soc. 119:455-462. Collins, M. R, & B. W. Stender 1987 Larval king mackerel (Scomberomorus cavalla), Spanish mackerel (S. maculatus), and bluefish (Pomatomus saltatrix) off the southeast coast of the United States, 1973-1980. Bull. Mar. Sci. 41:822-834. Cupka, D. M. 1972 A survey of the ichthyofauna of the surf zone in South Carolina. Tech. Rep. 4, S.C. Wildl. Mar. Resour. Dep., Charleston. Deuel, D. G., J. R. Clark, & A. J. Mansuetti 1966 Description of embryonic and early larval stages of bluefish, Pomatomus saltatrix. Trans. Am. Fish. Soc. 95:264-271. Fable, W. A. Jr., H. A. Brasher, L. Trent, & J. Finnegan Jr. 1981 Possible temperature effects on charter boat catches of king mackerel and other coastal pelagic species in Northwest Florida. Mar. Fish. Rev. 43(8):21-26. Futch, C. R., & S. E. Dwinell 1977 Nearshore marine ecology at Hutchinson Island, Florida: 1971-1974. rV. Lancelets and fishes. Fla. Mar. Res. Publ. 24, 23 p. Gilmore, J. J. 1985 Oceanic distribution, abundance and migration of bluefish {Pomatomus saltatrix) along the east coast of the United States. M.S. thesis, State Univ. New York, Stony Brook, 60 p. Kendall, A. W. Jr., & L. A. Walford 1979 Sources and distribution of bluefish, Pomatomus saltatrix, larvae and juveniles off the east coast of the United States. Fish. Bull., U.S. 77:213-227. NOTE McBnde et al. : Recruitment of Pomatomus saltatrix to estuaries 395 Lassiter, R. R. 1962 Life history aspects of the bluefish, Pomatomus saltatrix (Linnaeus), from the coast of North Carolina. M.S. thesis, N. Carolina State Coll., Ra- leigh, 103 p. Lenanton, R. C. J., & I. C. Potter 1987 Contribution of estuaries to commercial fisheries in temperate western Australia and the concept of estuarine dependence. Estuaries 10(11:28-35. Lund, W. A., & G. C. Maltezos 1970 Movements and migrations of the bluefish, Pomatomus saltatrix, tagged in waters of New York and southern New England. Trans. Am. Fish. Soc. 99:719-725. McBride, R. S. 1989 Comparative growth and abundance of spring- versus summer-spawned young-of-year bluefish, Pomatomus saltatrix, recruiting to the New York Bight. M.S. thesis, State Univ. New York, Stony Brook, 92 p. McBride, R. S., & D. O. Conover 1991 Recruitment of young-of-the-year bluefish (Pomatomus saltatrix) to the New York Bight: Variation in abundance and growth of spring and summer-spawned cohorts. Mar. Ecol. Prog. Ser. 78:205-216. Miller, G. L., & S. C. Jorgenson 1969 Seasonal abundance and length frequency distri- bution of some marine fishes in coastal Georgia. Data Rep. 35, U.S. Fish Wildl. Serv., 103 p. [microfiche]. NMFS (National Marine Fisheries Service) 1987 Status of the fishery resources off the northeast- ern United States for 1987. NOAA Tech. Memo. NMFS-F/NEC-50, NMFS Northeast Fish. Sci. Cent., Woods Hole MA. 1988 Status of the fishery resources off the northeast- ern United States for 1988. NOAA Tech. Memo. NMFS-F/NEC-63, NMFS Northeast Fish. Sci. Cent, Woods Hole MA. Norcross, J. J., S. L. Richardson, W. H. Massmann & E. B. Joseph 1974 Development of young bluefish {Pomatomus saltatrix) and distribution of eggs and young in Vir- ginian coastal waters. Trans. Am. Fish. Soc. 103:477- 497. Nyman, R. M., & D. O. Conover 1988 The relation between spawning season and the recruitment of young-of-the-year bluefish, Pomatomus saltatrix, to New York. Fish. Bull, U.S. 66:237-250. Phalen, P. S., & D. C. Stephan 1989 North Carolina marine fisheries trawl surveys. In Azarovitz, T.R., J. McGurrin, & R. Seagraves (eds.). Proceedings of a workshop on bottom trawl surveys, p. 59-62. Spec. Rep. 17, Atl. States Mar. Fish. Comm. Pollock, B. R. 1984 The tailor (Pomatomus saltatrix) fishery at Fraser Island and its relation to the life-history of the fish. Proc. R. Soc. Queensl. 95:23-28. Powles, H. 1981 Distribution and movements of neustonic young of estuarine dependent (Mugil spp., Pomatomus saltatrix) and estuarine independent (Coryphaena spp.) fishes off the southeastern United States. Rapp. P.-V. Reun. Cons. int. Explor. Mer 178:207-209. Richards, C. E. 1965 Availability patterns of marine fishes caught by charter boats operating off Virginia's Eastern Shore, 1955-1962. Chesapeake Sci. 6(2):96-108. Rountree, R. A., & K. W. Able 1992 Fauna of high salinity subtidal marsh creeks in southern New Jersey: composition, abundance and biomass. Estuaries 15(21:171-185. Tagatz, M. E., & D. L. Dudley 1961 Seasonal occurrence of marine fishes in four shore habitats near Beaufort, N.C., 1957-60. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 390, 19 p. Targett, T. E., & J. D. McCleave 1974 Summer abundance of fishes in a Maine tidal cove with special reference to temperature. Trans. Am. Fish. Soc. 103:325-330. van der Elst, R. 1976 Game fish of the east coast of southern Africa. I. The biology of the elf, Pomatomus saltatrix (Linnaeus), in the coastal waters of Natal. Invest. Rep. 44, Oceanogr. Res. Inst., S. Afr. Wenner, C. A., & G. R. Sedberry 1989 Species composition, distribution, and relative abundance of fishes in the coastal habitat off the southeastern United States. NOAA Tech. Rep. NMFS 79, 49 p. Wenner, E. 1989 Southeast Area Monitoring and Assessment Pro- gram (SEAMAP). In Azarovitz, T. R., J. McGurrin, & R. Seagraves (eds.), Proceedings of a workshop on bottom trawl surveys, p. 23-26. Spec. Rep. 17, Atl. States Mar. Fish. Comm. Wilk, S. J. 1977 Biological and fisheries data on bluefish Pomatomus saltatrix (Linnaeus). Tech. Ser. Rep. 11, Sandy Hook Lab., NMFS Northeast Fish. Sci. Cent., Highlands NJ, 56 p. Order Processing Code * wsr Superintendent of Documents Subscriptions Order Form I — I YES, enter mv subscription(s) as follows: v To fax your orders (202) 512-2233 subscriptions to FISHERY BULLETIN (FB) for $24.00 per year ($30.00 foreign). Charge your order. It's Easy! The total cost of my order is $. change. (Company or Personal Name) (Additional address/attention line) Prices include regular domestic postage and handling and are subject to For privacy protection, check the box below: | Do not make my name available to other mailers Please choose method of payment: I | Check Payable to the Superintendent of Documents ] GPO Deposit Account (Please type or print) (Street address) (City, State. ZIP Code) (Daytime phone including area code) (Purchase Order No. □ VISA or MasterCard Account Thank you for (Credit card expiration date) . your order! (Authorizing Signature) Mail To: 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. ±33& O V 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 17 (for decapod crustaceans). Text footnotes should be numbered in Arabic numerals and typed on a sep- arate sheet from the text. Footnotes are not used for reference material or per- sonal communications, but rather to explain or define terms in the text and for contribution numbers on the title page. Informal sources Personal communi- cations, unpublished data, and untitled manuscripts in preparation are noted parenthetically in the text (full name, affiliation, brief address including zip code, and month and year when appro- priate). Acknowledgments Gather all ac- knowledgments into a brief statement at the end of the text. Give credit only for exceptional contributions and not to those whose contributions are part of their normal duties. Citations All titled sources should be listed in the Citations section, including unpublished and processed material. In text, cite as Smith and Jones (1977) or (Smith and Jones 1977); if more than one citation, list chronologically (Smith 1936, Jones 1975, Doe In press). All sources cited in the text should be listed alphabetically by the senior authors' surnames under the heading CITA- TIONS. Abbreviations of periodicals and serials should conform to Serial Sources for the BIOSIS Data Base™. Indicate whether sources are in a langu- age other than English. For informal literature, include address of author or publisher. Authors are responsible for the accuracy of all citations. Tables should supplement, not dupli- cate, the text. Each table should be numbered and cited consecutively, with headings short but amply descriptive so that the reader need not refer to the text. For values less than one, zeros should precede all decimal points. In- dicate units of measure in column head- ings; do not deviate from the unit of measure within a column. Table foot- notes should be noted consecutively in Roman letters across the page from left to right and then down. Since all tables are typeset, they need not be submitted camera-ready. Figures Photographs and line draw- ings should be of professional quality- clear and concise— and reducible to 42 picas for full-page width or to 20 picas for a single-column width, and to a max- imum 55 picas high. All graphic ele- ments in illustrations must be propor- tioned to insure legibility when reduced to fit the page format. Line weight and lettering should be sharp and even. Let- tering should be upper and lower case, and vertical lettering should be avoided whenever possible (except for vertical, y, axis). Zeros should precede all deci- mal points for values less than one. Re- productions of line artwork are accepted in the form of high-quality photographic prints from negatives or photomechan- ical transfer (PMT). Halftones should be sharply focused with good contrast. Micron rules should be inserted on elec- tron micrographs, even when magnifi- cation is included in the figure legend. There should be clear distinction be- tween identifying letters (press-on or overlay) and background of photograph. Label each figure in pencil on the back. Send only xerox copies of figures to the Scientific Editor; originals or photo- graphic prints will be requested later when the manuscript is accepted for publication. Copyright Government publications are in the public domain, i.e., they are not protected by copyright. Submission of manuscript Disks Authors are encouraged to re- tain manuscripts on word-processing storage media (diskettes, floppy disks) and submit a double-spaced hardcopy run from the storage media. Submit disks as MS-DOS "print" or "non- document" files (often called "ASCII files"). If a disk cannot be converted to an ASCII file, the author should indicate on the disk the source computer and software language along with the file name. Either 5V4-inch or 3V2-inch disks from IBM-compatible or Apple/Macin- tosh systems (non-graphics only) can be submitted, double-sided/double-density or high-density, limiting each file to 300 kilobytes. All 8-inch word-processing disks (e.g., Wang or NBI) must be con- verted onto 5Vi- 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. 'wmI ?H01 LIBRARY UH ITblfl V 9VIUU U.S. Department of Commerce Volume^y Numbers July 1993 \v Fishery Bulleti Marin gtoiy OCT 2 6 1933 '/ 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 , !□ h 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 SS0M, Washington, DC 20402-9373. Although the contents have not been copyrighted and may be reprinted entire- ly, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this period- ical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $24.00 domestic and $30.00 foreign. Cost per single issue: $12.00 domestic and $15.00 foreign. See back page for order form. Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE. BIN C 1 5700 Seattle. Washington 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 3 July 1993 Fishery Bulletin rine Biological Laboratory LIBRARY OCT 2 6 1993 Contents Woods Hole, Mass. 397 Busby, Morgan S., and David A. Ambrose Development of larval and early juvenile pygmy poacher, Odontopyxis tr/sp/nosa, and blacktlp poacher. Xeneretmus latifrons (Scorpaeniformes: Agonidae) 414 DeMartini, Edward E. Modeling the potential of fishery reserves for managing Pacific coral reef fishes 428 Edwards, Elizabeth F. Allometry of energetics parameters in spotted Dolphin (Stenella attenuataj from the eastern tropical Pacific Ocean 440 Gaskin, David E., Satoru Yamamoto, and Akito Kawamura Harbor Porpoise, Phocoena phocoena (L). in the coastal waters of northern Japan 455 Humphreys, Robert L., Jr., Mark A. Crossler, and Craig M. Rowland Use of a monogenean gill parasite and feasibility of condition indices for identifying new recruits to a seamount population of armorhead Pseudopentaceros wheeleri (Pentacerotidae) 464 Kane, Joseph Variability of zooplankton biomass and dominant species abundance on Georges Bank, I 977-1 986 475 McKenna, James E., Jr. Spatial structure and temporal continuity of the South Georgian Antarctic fish community 491 Olesiuk, Peter F. Annual prey consumption by harbor seals (Phoca vitulinaj in the Strait of Georgia, British Columbia Fishery Bulletin 9 1(3), 1993 5 1 6 Sandt, Veronique J., and Allan W. Stoner Ontogenetic shift in habitat by early juvenile queen conch, Strombus gigas: patterns and potential mechanisms 526 Schmidt, David J., Mark R. Collins, and David M. Wyanski Age, growth, maturity, and spawning of Spanish mackerel, Scomberomorus maculatus (Mitchill), from the Atlantic Coast of the southeastern United States 534 Terceiro, Mark, and Jeffrey L. Ross A comparison of alternative methods for the estimation of age from length data for Atlantic coast bluefish (Pomatomus saltatrixj Notes 550 Hayes, Daniel B. A statistical method for evaluating differences between age-length keys with application to Georges Bank haddock, Melanogrammus aeglefinus 558 Matlock, Gary C, Robert L. Colura, and Lawrence W. McEachron Direct validation of black drum (Pogonias cromisj ages determined from scales 564 Millar, Russell B. Incorporation of between-haul variation using bootstrapping and nonparametric estimation of selection curves 573 Pearson, Donald E., David A. Douglas, and Bill Barss Biological observations from the Cobb Seamount rockfish fishery 577 Witting, David A., and Kenneth W. Able Effects of body size on probability of predation for juvenile summer and winter flounder based on laboratory experiments 582 Yamanaka, Lynne K., and Laura J. Richards Movements of transplanted lingcod, Ophiodon elongatus, determined by ultrasonic telemetry Abstract.— Developmental stages of pygmy poacher, Odontopyxis trispinosa, and blacktip poacher, Xeneretmus latifrons, are described and illustrated from specimens col- lected from the northeastern Pacific Ocean. External morphology, pig- mentation, and meristic features are described which distinguish larvae of these species from other agonids occurring in these waters. Postanal pigment patterns, par- ticularly on the caudal finfold, dis- tinguish preflexion larvae. Odon- topyxis trispinosa larvae possess a semicircular patch of melanophores that covers nearly the entire caudal finfold. The caudal finfold of pre- flexion X. latifrons larvae are void of pigment with the exception of a small patch of melanophores near the ventral margin of the notochord tip. Flexion and postflexion larvae can be distinguished by caudal and anal fin pigmentation, head spina- tion, and adult meristic features. Development of larval and early juvenile pygmy poacher, Odontopyxis trispinosa, and blacktip poacher, Xeneretmus latifrons (Scorpaeniformes: Agonidae) Morgan S. Busby Resource Assessment and Conservation Engineering Division Alaska Fisheries Science Center National Marine Fisheries Service. NOAA 7600 Sand Point Way NE, Seattle, Washington 98 I I 5-0070 David A. Ambrose Southwest Fisheries Science Center National Marine Fisheries Service, NOAA PO. Box 271, La Jolla. California 92038 Manuscript accepted 26 February 1993. Fishery Bulletin, 91:397-413 ( 1993). The family Agonidae is a morphologi- cally diverse group of relatively small, benthic marine fishes. Agonids, com- monly called poachers, are character- ized by the presence of fused or over- lapping bony plates that encase the body. Fifteen genera represented by 25 species occur in the northeastern Pacific (Matarese et al., 1989). The pygmy poacher Odontopyxis trispinosa is a small (to 8.1cm SL) subtidal agonid distinguished by an elongate body, small vertical spine at the snout tip, and a moderately de- veloped occipital pit divided by a lon- gitudinal ridge. Fin-element counts of O. trispinosa are D III-VI, 5-7; A 5-7; P 13-15; V 1,2 (Matarese et al, 1989). The blacktip poacher Xener- etmus latifrons is a larger (to 19cm SL) subtidal agonid distinguished by black margins on the dorsal fins and a weakly developed occipital depres- sion (Miller and Lea, 1972; Hart, 1973). Fin-element counts of X. latifrons are D VI-VIII, 6-8; A 6-9; P 13-15; V 1,2 (Matarese et al., 1989). The presence of spiny scales on the eyeballs distinguishes X. latifrons from X. leiops. The absence of cheekplates distinguishes X latifrons from X. triacanthus (Miller and Lea, 1972; Eschmeyer et al., 1983). Freeman (1951) hypothesized that these two taxa are closely related phylogenetically and placed them in his subfamily Xeneretminae. The cla- distic analysis of Kanayama (1991) demonstrated close relationships be- tween the two taxa which he placed in the subfamily Anoplagoninae. The geographic range of adult O. trispinosa extends from Southeast Alaska to central Baja California. Adult X. latifrons occur over a slightly narrower range from Van- couver Island to northern Baja Cali- fornia (Eschmeyer et al., 1983). Both taxa occur at depths of 18-370 m (Miller and Lea, 1972; Hart, 1973). Larval O. trispinosa and X. latifrons are the most commonly occurring agonids in the California Cooperative Oceanic Fisheries Investigations (CalCOFI) ichthyoplankton collection1. 'Moser, H. G., ed. Guide to the early stages of fishes from the California Current region. South- west Fisheries Science Center, NMFS, P.O. Box 271, La Jolla, CA 92038. In preparation. 397 398 Fishery Bulletin 91(3), 1993 Larval development of agonids is poorly known. Washington et al. (1984) and Maeda and Amaoka ( 1988) reported that for- mation of most external body parts, including dermal bony plates and spines, begins early in larval development. Development of the pygmy poacher (0. trispinosa) is de- scribed for the first time here. Development of X. latifrons is clarified based on new material. A partial developmental series (7, 10, 16 mm SL) of X. latifrons was previously described by Marliave (1975). Washington et al. (1984) included a single illustration of X. latifrons (9.6mm SL) which differed markedly in appearance from those in Marliave (1975). Matarese et al. (1989) combined the illustrations from both sources to create a complete de- velopmental series. As a result, some confusion exists as to which illustrations in this series are ac- tually those of X. latifrons lar- vae. The following descriptions are the first complete accounts of larval development in the family Agonidae. Characters are presented that permit identifica- tion of these larvae from field collections. Methods Specimens Agonid larvae examined in this study were obtained from collec- tions made off the coasts of Baja California, California, Oregon, Washington (including Puget Sound), Southeast Alaska, and the inside passage waters of Brit- ish Columbia, Canada (Fig. 1). One hundred and thirty-six O. trispinosa (4.3-56.0 mm SL) and 77 X. latifrons (4.8-79.0 mm SL) were examined. The largest specimens of each taxa were determined to be adults and are not included in the descriptions. Larvae and 56° 00N Odontopyx-i3 trispinosa Xeneretmi 135°00'W 131° 127° 115° 123° 119° Figure 1 Collection locations of Odontopyxis trispinosa and Xeneretmus latifrons larvae and early juveniles used in this study. juveniles of both taxa were collected with dipnets, 60- and 70-cm bongo nets, Isaacs-Kidd midwater trawls, Busby and Ambrose Development of larval and juvenile Odontopyxis tnspmosa and Xemeretmus latifrons 399 Tucker trawls, sled trawls, and bottom trawls. Collec- tions were made by the Vancouver Public Aquarium (VPA), Oregon State University (OSU), University of Washington (UW), NOAA's Southwest Fisheries Sci- ence Center (SWFSC-CalCOFI program), and the Alaska Fisheries Science Center (AFSC) from the years 1932 to 1991. Specimens are currently housed in the larval fish collections of these institutions. Larvae and juveniles were originally fixed in 3.5 or 5.0% formalin and subsequently transferred and stored in 3.5% buff- ered formalin or 70% ethanol. Larvae of 0. trispinosa and X. latifrons were identi- fied using the serial approach. This method uses adult characters to progressively link juveniles to smaller specimens through a continuous sequence of shared simi- larities. Pigmentation, head spination, body morphol- ogy, dermal plates, and meristic features were all used as diagnostic characters. Identification of adult and ju- venile specimens was accomplished by using methods of Miller and Lea (1972), Hart (1973), and Kanayama (1991). Nomenclature and taxonomic classification of the family Agonidae follow Kanayama (1991). Develop- mental series were illustrated by using a camera lucida attached to a dissecting stereomicroscope. Only melanistic pigmentation is described because formalin fails to preserve color pigments. In the de- scription of pigmentation, the term "band" refers to any aggregation of melanophores that approximates a vertically oriented rectangle. A "bar" also approximates a rectangle but is horizontally oriented. A "patch" is any other distinguishable aggregation of melanophores. Measurements The following measurements were made on 45 larvae and early juveniles of 0. trispinosa and 45 larvae of X. latifrons by using an ocular micrometer in a stereo- microscope: Standard length (SL) — Snout tip to notochord tip prior to development of caudal fin, then to posterior margin of hypural element. (All body lengths in this study are standard lengths.) Body depth — Vertical distance from dorsal to ventral body margin at pectoral-fin base. Snout to anus length — Distance along body midline from snout tip to a vertical line through center of anal opening. Head length (HL) — Snout tip to posterior edge of opercle (to pectoral-fin base in small larvae before opercular margin is visible). Head width — Distance across head between dorsal margins of orbits. Snout length — Snout tip to anterior margin of orbit of left eye. Eye diameter — Greatest diameter of left orbit. Pectoral fin length — Distance from pectoral-fin base to tip of the longest ray. Osteology Selected specimens were cleared and differentially stained to identify cartilage and bone with alcian blue and alizarin red-S (Pothoff, 1984). Skeletal ele- ments and dermal plates were recognized as ossified upon initial uptake of alizarin red-S. Twenty-seven 0. trispinosa (5.3^41 mm) and 12 X. latifrons (7.4- 39.2 mm) were cleared and stained for study. Counts of meristic features were made on stained specimens only. Not all stages of development were stained for X. latifrons because specimens were limited. Preflexion, flexion, and postflexion stage larvae were stained (Kendall et al., 1984). Nomenclature of skeletal ele- ments follows that used by Leipertz (1985) for X. triacanthus. Plate nomenclature follows that of Gruchy (1969) and is described in Figure 2. Terminology of larval head spination follows that proposed for adult DLP LLP VLP Figure 2 Terminology of bony plates in agonid larvae: DLP=dorsolateral; MDP=mid-dorsal; SLP=supralateral; LLP=lateral line; ILP=infralateral; VLP=ventrolateral; MVP=mid-ventral (after Gruchy, 1969). 400 Fishery Bulletin 91(3), 1993 agonids by Laroche (1986). For larvae with spines and no named analogous adult spine, terminology follows Moser and Ahlstrom ( 1978) or Richardson and Laroche (1979) for larvae of rockfish Sebastes spp. (another Scorpaeniform) (Table 1, Figs. 3 and 4). Results Development of Odontopyxis trispinoa Morphology Larvae of O. trispinosa are elongate and slender with mean body depth at the pectoral fin ori- gin of 11.7% SL in preflexion specimens decreasing to 10.7% SL in flexion and postflexion larvae (Tables 2 and 3). Mean head length is 18.4% SL in preflexion larvae and increases to 23.8%< SL in juveniles (Table 3). Mean head width is approximately 50.0% HL throughout larval development and decreases to 37.1% HL in juveniles (Table 3). Mean snout length increases from 20.4% HL in preflexion larvae to 27.0% HL in postflexion larvae and eye diameter decreases from 33.3% HL in preflexion larvae to 23.3% HL in juve- niles. Mean pectoral-fin length increases from 7.0% SL in preflexion larvae to 19.5% SL in postflexion larvae (Table 3). The gut is moderately long: mean snout to anus distance is 47.9% SL in preflexion larvae and decreases to 33.0% SL in juveniles. Pigmentation Pigmentation in O. trispinosa larvae was relatively consistent among specimens and is a useful distinguishing character (Fig. 5). Head region Pigmentation on the head of preflexion larvae is present as rows of melanophores on the up- per and lower jaws. A few additional melanophores are present on the snout. Melanophores on the opercu- lar and hyoid regions join the upper and lower jaw pigmentation to form a continuous swath giving lar- vae a "bearded" appearance (Fig. 5A). In flexion lar- vae, additional melanophores appear posterior to the eye. Lateral body and gut region The dorsolateral sur- face of the body above the gut is covered with melano- phores with the exception of a patch along the dorsal midline above the pectoral fin. The dorsolateral pig- mentation recedes ventrolaterally toward the gut with development and is completely absent in postflexion larvae. Melanophores cover nearly the entire caudal portion of the body in preflexion larvae. In some speci- mens, the notochord tip is unpigmented. In late flexion and postflexion larvae, pigmentation on the lateral body surface begins to separate gradually into seven bands (Fig. 5D). The first band extends from the posterior region of the first dorsal fin to the ventral body mid- line immediately posterior to the anus. The second band extends between the first two soft rays of the Table 1 Head spine terminology in agonid larvae listed by seque nee of development in Odontopyxit trispinosa. (P) designates dermal plate. Abbreviation Spine/Plate name Bone of origin Adult spine/plate PA Parietal Parietal Parietal SPO Supraocular Frontal Supraocular SC Supracleithral Supracleithrum Supracleithral CO Coronal Frontal (Overgrown) PT Pterotic Pterotic Pterotic NA Nasal Nasal Nasal APO-4 4th Anterior Preopercular Preopercle (Overgrown) PPO-1,2 1st, 2nd Posterior Preopercular Preopercle Preopercular PPO-3,4 3rd, 4th Posterior Preopercular Preopercle Preopercular SIO-5,6 5th, 6th Superior Infraorbital Infraorbital 3 Posterior infraorbitals CL Cleithral Cleithrum (Overgrown) TM Tympanic Frontal ( Overgrown 1 OP Opercular Opercle Opercular FR Frontal Frontal (Overgrown) SOP Subopercular Subopercle Gill Cover Spine SIO-1,2 1st, 2nd Superior Infraorbital Infraorbital 1 ( Lachrymal 1 Anterior Infraorbital PSO-1 1st Postocular Frontal Postocular PSO-2.3 2nd, 3rdPostocular(P) (Dermal) Postocular Plates RO Rostral Rostral Plate Rostral PST Posttemporal (P) (Dermal) (Overgrown l 110-1,2 1st, 2nd Inferior Infraorbital (P) (Dermal) Anterior. Medial Cheek Plates SIO-3,4 3rd, 4th Superior Infraorbital Infraorbital 2 ( Jugal 1 Medial Infraorbital IIO-3 3rd Inferior Infraorbital (Pi ( Dermal ) Posterior Cheek Plate SCL Sclerotics (P) Sclerotic Eyeball Plates Busby and Ambrose Development of larval and juvenile Odontopyxis trispinosa and Xemeretmus latifrons 401 PSO-1 APO-4 PPO-2 PPO-3 IIO-3 PPO-4 Figure 3 Positions and abbreviations of larval head spines and plates in agonids. Based on a 17.2-mm stained larva of Odontopyxis trispinosa (modified from Moser and Ahlstrom (1978), Richardson and Laroche (1979), and Laroche (1986)). Figure 4 Top view above eye showing superior infraorbital and frontal spines of a 17.2-mm SL, cleared and stained Odontopyxis trispinosa. dorsal and anal fins and is connected ventrally to the first band by a bar. The third band extends from the posterior half of the second dorsal fin to the posterior half of the anal fin. The remaining four bands are evenly spaced on the body between the posterior edges of the second dorsal and anal fins and the caudal fin. The third, fifth, and seventh bands are the widest. The posteriormost band of pigmentation in postflexion larvae and early juveniles is located at the hypural margin and is continuous with the caudal-fin pigment. The entire ventral and lateral surfaces of the gut in preflexion larvae are covered with pigmentation. A dense row of melanophores is present along the ven- tral midline from the isthmus to the end of the pre- anal finfold. The ventral midline pigmentation can be distinguished easily in lateral view through the postflexion stage. In some specimens, a small, circular, unpigmented area is present on the lateral surface of the gut posterior to the pectoral-fin rays. Fins The base of the pectoral fin is completely cov- ered with melanophores throughout development. Pig- mentation is absent from the pectoral-fin blade, rays, and membrane. Most preflexion and flexion specimens possess a small group of melanophores on the anterior portion of the dorsal finfold near the body margin over the g it (Fig. 5A). This patch migrates posteriorly with dev< i opment and becomes the pigmentation seen on the first dorsal fin in postflexion larvae and juveniles (Fig. 5, C and D). A larger patch of melanophores is present at approximately midbody which usually ex- tends to the dorsal edge of the finfold in preflexion larvae. This dorsal midbody patch recedes toward the body and splits into two somewhat triangular-shaped patches of pigmentation in flexion larvae. This larger patch of pigmentation is retained on the second dorsal fin in postflexion larvae and juveniles. The finfold patches roughly correspond to body bands on postflexion and juvenile specimens. Three additional triangular- shaped patches of melanophores are present approxi- mately two-thirds of the distance between the anus and notochord tip on the dorsal finfold and form a continuous region of pigment extending to the caudal finfold. The triangular-shaped patches of pigmentation separate in flexion larvae and disappear as the finfold recedes in postflexion larvae. 402 Fishery Bulletin 91(3). 1993 Table 2 Morphometric measurements in millimeters) of 45 Odontopyxis trispinosa larvae and early juveniles. Specimens between dashed lines ( ) were underg oing notochord flexion. Standard Body Snout to Head Head Snout Eye Pectoral length depth anus length length width length diameter fin length 4.3 0.44 2.02 0.78 0.40 0.14 0.28 0.30 4.5 0.60 2.06 0.72 0.52 0.12 0.30 0.30 4.6 0.60 2.24 0.80 0.44 0.18 0.30 0.34 5.2 0.62 2.48 1.02 0.48 0.22 0.32 0.34 5.3 0.60 2.52 0.96 0.46 0.20 0.32 0.44 5.3 0.64 2.60 1.02 0.46 0.18 0.36 0.36 5.9 0.62 2.72 1.04 0.42 0.20 0.34 0.38 6.4 0.74 3.24 1.14 0.44 0.26 0.36 0.34 6.7 0.70 3.24 1.20 0.56 0.22 0.40 0.46 6.8 0.76 3.20 1.26 0.56 0.30 0.40 0.50 6.8 0.82 3.12 1.40 0.62 0.30 0.36 0.50 7.2 0.86 3.48 1.36 0.70 0.30 0.40 0.52 7.1 0.88 3.60 1.38 0.70 0.30 0.40 0.50 7.9 0.66 3.40 1.34 0.66 0.32 0.36 0.60 8.5 0.92 4.12 1.76 0.80 0.50 0.44 0.70 9.1 0.96 4.12 1.50 0.66 0.38 0.42 1.00 9.8 1.00 4.60 2.06 0.92 0.60 0.52 1.40 10.0 1.06 4.40 1.70 1.25 0.41 0.45 1.05 10.8 1.20 5.00 2.36 1.08 0.52 0.60 1.46 11.7 1.36 5.25 2.52 1.20 0.60 0.58 1.64 11.8 1.42 5.50 2.26 1.24 0.74 0.56 2.46 11.8 1.44 5.42 2.80 1.24 0.76 0.60 2.26 12.5 1.22 5.67 2.70 1.20 0.74 0.60 2.08 12.6 1.32 5.83 2.80 1.44 0.86 0.70 2.48 12.7 1.26 5.42 2.76 1.28 0.76 0.64 2.42 12.8 1.30 5.83 2.70 1.20 0.68 0.60 2.42 13.7 1.64 6.33 2.92 1.50 0.90 0.72 2.24 13.8 1.58 6.08 2.92 1.30 0.80 0.66 2.35 13.9 1.38 6.08 3.00 1.46 0.74 0.70 2.80 14.3 1.50 6.17 3.25 1.50 0.80 0.74 2.96 14.7 1.32 5.87 3.00 1.40 0.72 0.76 2.64 14.8 1.52 6.08 3.00 1.40 0.76 0.66 2.68 15.0 1.60 5.83 3.60 1.54 0.90 0.84 3.20 15.2 1.70 5.92 3.84 1.80 1.00 0.84 3.20 15.4 1.66 6.17 3.20 1.60 0.84 0.72 3.20 15.8 1.62 6.63 3.50 1.82 0.91 0.86 3.16 16.7 1.70 6.50 4.24 2.04 1.12 0.98 3.68 17.5 1.87 6.84 4.26 1.98 1.13 0.90 3.08 17.7 2.20 7.58 4.32 1.86 1.24 0.84 3.72 17.7 1.76 7.00 4.48 2.00 1.20 1.06 3.68 28.0 2.75 9.58 7.40 2.88 1.56 1.64 4.68 32.1 3.33 10.50 7.58 2.88 1.84 1.75 5.58 34.8 3.42 11.30 8.08 3.12 1.92 2.00 6.00 36.4 3.92 11.70 8.25 2.92 2.17 1.92 6.33 41.6 4.83 14.50 9.67 3.33 2.50 2.25 6.92 A small patch of melano- phores is present on the anal finfold immediately posterior to the anus. This small patch of pigmentation is a useful char- acter which persists until postflexion when it migrates dor- sally with the receding finfold to form the bar between the first and second body bands (Fig. 5D). A large triangular-shaped patch of pigment is present at midbody and extends nearly to the finfold margin in preflexion and early flexion larvae. The large trian- gular-shaped patch expands ven- trally to the finfold margin and becomes more rectangular shaped in late flexion larvae. This is the only pigmentation present on the anal fin in postflexion larvae and juveniles (Fig. 5, C and D). Two additional patches of melanophores are present on the anal finfold in preflexion larvae. These form a nearly continuous region of pig- mentation which begins at ap- proximately two-thirds of the distance between the anus and notochord tip which extends to the caudal finfold. This pigmen- tation breaks apart in late flexion larvae and disappears as the finfold recedes in postflexion larvae. The posteriormost dorsal and anal finfold pigmentation con- nect and are continuous with a large semicircular patch of mel- anophores which covers nearly Table 3 Body proportions of Odontopyxis trispinosa larvae and early juveniles. Values given for each body proportion are expressed as percentage of standard length (SL) or head length (HL): mean, standard deviation. and range. Body proportion Preflexion Flexion Postflexion Juvenile Sample size 12 8 20 5 Standard length 5.711.0 (4.3-7.2) 9.411.5 (7.1-11.7) 14.511.9 (11.8-17.7) 34.615.1 (28.0-41.6) Bodydepth/SL 11.711.0 (10.2-13.4) 10.711.2 (8.4-12.4) 10.710.9 (9.0-12.4) 10.510.7 (9.8-11.61 Snout to anus length/SL 47.911.4 (45.7-50.3) 46.212.5 (43.2-50.8) 42.512.8 (38.9-46.6) 33.311.2 (32.1-34.9) Head length/SL 18.4+1.2 (16.1-20.5) 19.412.2 (16.5-21.8) 22.4+1.9 (19.2-25.4) 23.811.5 (22.7-26.4) Headwidth/HL 48.7+8.7 (40.4-72.2) 50.119.7 (44.0-73.5) 47.213.3 (43.1-54.9) 37.112.0 (34.4-38.9) Snout length/HL 20.412.3 (16.7-23.8) 24.812.7 (21.7-29.1) 27.012.3 (24.0-32.7) 24.3+2.1 (21.1-26.3) Eye diameter/HL 33.314.0 (25.7-41.7) 26.111.9 (23.0-29.0) 23.011.5 (19.4-25.3) 23.3+0.9 (22.2-24.8) Pectoral fin length/SL 7.0+0.7 (5.3-8.4) 10.812.9 (7.1-14.2) 19.511.7 (16.3-21.3) 17.110.4 (16.6-17.4) Busby and Ambrose. Development of larval and juvenile Odontopyxis trispmosa and Xemeretmus latifrons 403 4.3 mm B 7.9 mm 10.3 mm 14.7 mm Figure 5 Larval stages of Odontopyxis tnspinosa. (A) Preflexion larva 4.3mm SL, CalCOFI 5405-73.50. (B) Early flexion larva 7.9mm SL, CalCOFl 6304-110.32. (Cl Flexion larva 10.3mm SL, VPA, Lambert Channel 4/26/91 No. 7. (Dl Postflexion larva 14.7mm SL, CalCOFI 7412-83.44. 404 Fishery Bulletin 91(3), 1993 the entire caudal finfold. Caudal pigmentation persists throughout development. Osteology Although precursors of some bony struc- tures such as dermal plates and fin rays are discernable as early as 8.0 mm, actual ossification of skeletal ele- ments in 0. trispinosa does not begin until approxi- mately 12.0 mm. Cranium The parasphenoid and basioccipital bones of the cranium begin to ossify at 11.7 mm. At 13.2 mm, several bones, including the nasal, frontal, parietal, parasphenoid, basioccipital, and exoccipital are com- pletely ossified. The rostral plate, nasal, lateral eth- moid, supraethmoid, vomer, exoccipital, lachrymal, and remainder of the circumorbital series ossify by 14.8 mm. The sphenotic, prootic, epiotic, tabular, and ptero- sphenoid are ossified by 27.0 mm. Spines Numerous head spines are ossified at 12.6 mm including the nasal, supraocular, the large bilobed parietal, coronal, pterotics, supracleithral, an- terior preoperculars and posterior preoperculars 1 and 2. The tympanic, cleithral, opercular, fifth and sixth superior infraorbital spines, and the third and fourth posterior preopercular spines are also ossified at this stage. The frontal spine forms by 13.8 mm but is not visible in lateral view because it is small and located behind the anterior margin of the supraocular spine. The frontal spine becomes overgrown with bone and is difficult to distinguish at 17.2 mm. Interopercular and superior infraorbital spines 1 and 2 begin forming at about 14.2 mm. The rostral and postocular spines and the postocular and posttemporal plates ossify at 14.8 mm. Superior infraorbital spines 3 and 4 and in- ferior infraorbital plates 1 and 2 are completely ossi- fied at this size. The third inferior infraorbital and sclerotic plates are ossified by 17.2 mm. All spines described here are paired with the exception of the rostral. Mandibular region The dentary, angular, and ar- ticular bones of the lower jaw are the first to ossify in postflexion larvae of 11.7 mm. At 12.6 mm, the pre- maxilla and maxilla are ossified. Palatine region Palatine, quadrate, metapterygoid, mesopterygoid, ectopterygoid, and symplectic bones are ossified at 12.6 mm. Opercular region The preopercle, opercle, inter- opercle, and subopercle are ossified by 12.6 mm. Hyoid region The basihyal, hypohyal, urohyal, ceratohyal, epihyal, interhyal, glossohyal, and bran- chiostegal rays are also ossified by 12.6 mm. The hyomandibula is ossified by 27.0 mm. Branchial region The pharyngobranchial teeth are the first structures to ossify by 11.7 mm. Ossification of the pharyngobranchials occurs at about 17.2 mm. The pharyngobranchials begin as four pieces of carti- lage that apparently fuse before ossification of the pharyngobranchial teeth. This process, however, was not observed. The epibranchials (n=A) and cerato- branchials (n=5) also ossify by 17.2mm. The remain- ing branchial support structures, including the hypobranchials (n=3) and basibranchials (n=3), ossify by 27.0 mm. Appendicular region The cleithrum, postcleithrum, and coracoid are ossified by 12.6 mm. The pelvic-fin spine and all pectoral-fin rays, with the exception of the ventralmost, ossify by 12.8 mm. The two pelvic-fin rays and the final (14th) pectoral-fin ray are ossified by 13.2 mm. The supracleithrum ossifies at about 13.8 mm. At 14.8 mm, ossification of the posttemporal is completed. Ossification of the basipterygium occurs at 17.2 mm. The scapula and three radials supporting the pectoral-fin rays are ossified by 27.0 mm. Median fins All dorsal- and anal-fin spines and soft rays are ossified by 12.6mm (Table 4). Six supe- rior principal rays and five inferior principal rays in the caudal fin are ossified at this size (total=ll). One inferior principal (12), and two superior procurrent rays (13, 14) are ossified by 13.2 mm. By 17.2 mm one addi- tional superior (15) procurrent caudal ray is formed which completes the adult count (3+6+6+0=15). Ossifi- cation of the epural, hypural plate, and pterygiophores supporting the dorsal- and anal-fin rays is completed by 27.0mm. Vertebral column Ossification of vertebral elements progresses from anterior to posterior. Notochord flexion begins at approximately 7.0 mm and is completed by 11.8 mm. All, except the posteriormost caudal verte- bral centrum, are ossified by 12.6 mm. The urostyle and the anteriormost neural and haemal spines are also complete by 12.6 mm. All abdominal neural spines are complete and three additional haemal spines form anteriorly by 13.2 mm. All vertebral centra are ossified by 13.8 mm and two additional haemal spines have formed. By 14.2 mm, 22 additional haemal spines de- velop. All neural spines are ossified by 14.8 mm and haemal spines are complete at 15.4 mm. Dermal plates Precursors of the supra (SLP) and infralateral (ILP) plates first appear as small singular spines which are distinguishable at about 8.0 mm. Os- sification, however, does not begin until approximately 12.0 mm and progresses from anterior to posterior. Bone develops radially from the base of the spine to eventu- ally form the juvenile/adult dermal plate. By 12.6 mm, ossification of the SLP and ILP series is complete. The dorsolateral and mid-dorsal (DLP+MDP) and the ven- trolateral and mid-ventral ( VLP+MVP) plate series are complete by 14.8mm. The lateral line (LLP) plate se- ries begins to form at about 14.0 mm. Lateral line plates are the only plates that begin development as bifur- cate spines. Ossification of the LLP series is complete Busby and Ambrose: Development of larval and juvenile Odontopyxis tnspinosa and Xemeretmus latifrons 405 "8 d 3 -d C B 3 X 1 CO cu CO c* •8 Ph O CO -a 7^ | c ofl pu, X CO 00 H lO lO CO CO C^CXJCOCOCOCOCOCOCOCOCO O OOlQOiOlCOOlOlOlOOCO CO COCOCOCOCOCOCOCOCOCOCO O Tf O t^ n o co q ^ c^ i> co od aj --i h CM CM^CMCMCMCMCNCMCMCMCNl CO COCNCOCOCOCOCOCOCOCOCOCOCO rX> ^CD^DCOCDtDCDCOCOCDCDiXiiXi •— ' CM CNCl(NO)CN(NC>l(NCNO)(N CO CM Tj" cOTj-'^'CO'^'^J'^t^M'^CO CD (N CO t^COCOCDCOt-CDt^iOCOCCi c£)CTiCX>COtDC^t-CD'^CC>^^DtD(XicX5 io n w ^mmw't'tmcD^'io^' r-WC£)H(NQ0C0N00lNCN^t-(NqqO II CU II a. cu B n Q T3 II J Q 406 Fishery Bulletin 91(3), 1993 by 17.2 mm. Breast plates on the ventral surface of the abdomen and pelvic region are complete by 12.6 mm. The number of pec- toral-fin base plates in 0. trispinosa is variable ranging from 3 to 8 among specimens ex- amined. Ossification of pectoral- fin base plates is complete by about 13.8 mm. Development of Xeneretmus latifrons Morphology Larvae of X. latifrons are deeper bodied than 0. trispinosa; mean body depth at the pectoral-fin origin ranges from 13.3% SL in preflexion specimens to 14.6% SL post- flexion (Tables 5 and 6; Fig. 6). Head length increases from 20.3% SL in preflexion larvae to 26.6%' SL in postflexion larvae, and snout length as a propor- tion of head length increases from 18.8% HL in preflexion lar- vae to 24.4% HL in postflexion larvae (Table 6). Eye diameter decreases from about 37.5% HL in preflexion larvae to 28.4% HL in postflexion larvae, and head width remains approximately 54-55% HL throughout develop- ment (Table 6). Pectoral-fin length increases from 8.4% SL in preflexion lar- vae to 21.8% SL in postflexion larvae. In preflexion larvae, snout to anus distance is 51.7% Table 5 Morphometric measurements (in millimeters) of 45 Xeneretmus latifrons larvae. Speci- mens between dashed lines ( ) were undergoing notoehord flexion. Standard Body length depth Snout to anus length Head length Head width Snout length Eye diameter Pectoral fin length 4.9 5.1 5.1 5.3 5.5 5.6 6.0 6.0 6.8 6.8 7.0 7.2 7.8 7.9 8.3 8.9 9.3 9.8 9.9 10.0 10.0 10.5 10.8 0.72 0.66 0.70 0.66 0.76 0.68 0.78 0.72 0.82 0.80 1.10 0.88 1.16 0.96 1.30 1.14 1.30 1.18 1.60 1.20 1.48 1.50 1.72 2.20 2.56 2.48 2.40 2.88 2.92 2.92 2.98 3.40 3.36 3.80 4.04 4.80 4.12 4.78 0.84 0.96 1.06 0.90 1.26 1.00 1.14 1.15 1.32 1.20 1.66 1.80 2.08 1.54 1.54 0.48 0.50 0.54 0.44 0.66 0.50 0.64 0.60 0.66 0.66 0.80 1.00 1.08 0.84 1.13 0.06 0.16 0.20 0.16 0.26 0.20 0.23 0.22 0.24 0.20 0.28 0.42 0.56 0.30 0.32 4.78 5.67 5.33 5.50 6.08 5.67 5.75 6.17 1.94 2.68 2.44 2.56 2.44 2.36 2.96 2.68 1.04 1.52 1.24 1.44 1.36 1.30 1.64 1.42 0.44 0.50 0.56 0.64 0.60 0.54 0.62 0.62 0.38 0.38 0.38 0.38 0.40 0.40 0.44 0.40 0.46 0.42 0.46 0.52 0.64 0.48 0.61 0.62 0.82 0.68 0.74 0.74 0.70 0.86 0.80 0.34 0.34 0.38 0.32 0.44 0.40 0.50 0.40 0.42 0.50 0.52 1.08 1.14 0.62 0.81 0.88 1.80 1.50 1.80 1.62 1.42 2.48 2.14 10.8 1.72 6.25 2.80 1.80 0.64 0.88 2.32 11.4 1.68 6.50 2.88 1.86 0.72 0.88 2.30 11.8 1.88 6.50 2.88 1.70 0.74 0.78 2.20 11.8 1.64 6.08 2.88 1.90 0.70 0.76 3.00 12.5 1.96 6.67 3.28 1.94 0.78 1.02 2.64 12.7 1.97 7.00 3.44 1.88 0.94 0.86 2.80 12.7 1.96 6.67 3.36 2.06 0.66 0.96 3.00 13.2 2.06 6.70 3.52 2.00 0.90 0.92 3.12 13.9 2.30 7.08 3.92 2.30 0.98 1.02 3.52 14.2 2.10 7.12 4.04 2.10 1.10 1.08 3.72 14.7 2.22 7.58 3.84 2.12 0.94 1.02 3.52 14.7 2.08 7.33 3.56 2.10 0.78 0.96 3.76 15.5 2.58 7.14 3.80 2.58 0.91 1.09 3.73 16.6 2.74 8.06 4.86 2.30 1.06 1.22 4.40 16.8 2.52 7.92 4.60 2.40 1.16 1.26 3.84 18.3 2.58 8.06 4.86 2.74 1.37 1.22 4.10 22.0 3.04 9.40 5.17 3.04 1.67 1.52 4.41 30.0 3.67 11.50 8.67 3.60 2.25 2.25 5.33 30.5 4.00 11.30 8.67 4.33 1.92 2.92 5.50 31.8 3.92 11.80 9.58 4.42 2.00 3.08 5.58 39.2 4.33 12.70 10.70 4.75 2.33 3.42 6.83 42.0 5.12 14.20 11.20 5.42 2.50 3.67 7.08 Table 6 Body proportions of Xeneretmus latifrons larvae. Values given for each body proportion are expressed as percentage of standard length (SL) or head length (HL): mean, standard deviation. and range. Body proportion Preflexion Flexion Postflexion Sample size 15 8 22 Standard length 6.3±1.1 (4.9-8.3) 9.9±0.6 (8.9-10.8) 19.0+9.4 (10.8-42.0) Body depth/SL 13.3±1.4 (11.9-15.7) 14.2+1.6 (12.1-15.9) 14.6±1.6 (11.0-16.6) Snout to anus length/SL 51.7±4.5 (44.7-61.9) 56.7±2.7 (53.6-60.8) 47.4±7.6 (32.4-57.9) Head length/SL 20.3±3.0 (17.1-26.8) 25.3±2.3 (21.7-28.7) 26.6±1.8 (23.5-30.1) Head width/HL 53.9±6.0 (48.2-73.4) 54.8±2.0 (50.8-56.7 1 55.617.2 (41.5-67.9) Snout length/HL 18.8±4.2 (7.1-26.9) 22.612.0 (18.7-25.0l 24.412.8 (19.6-32.3) Eye diameter/HL 37.5±5.1 (27.7-45.2l 29.8+1.2 (27.9-32.0) 28.412.8 (25.0-33.7) Pectoral fin length/SL 8.4±2.8 (6.1-15.1) 17.1±4.1 (9.9-23.6) 21.813.1 (16.9-26.5) Busby and Ambrose: Development of larval and juvenile Odontopyxis trispinosa and Xemeretmus latifrons 407 3.50 - 3.00 - A 2.50 - A A A A 2 00 A * • A * * AA -C Q. 0) Q >. ■o 0 m 1 50 1.00 - 0.50 - ' .' ."■ • • Odontopyxis trispinosa * Xeneretmus latifrons • 0 5 10 15 20 Standard Length (mm) Figure 6 25 Plot of body depth vs standard length (mm) for larval Odontopyxis trispinosa and Xeneretmus latifrons. SL, increases to 56.7% SL during flexion, and decreases to 47.4% SL during postflexion. Pigmentation Pigmentation in X. latifrons is consis- tent among specimens and is a useful distinguishing character (Fig. 7). Because pigmentation on the head and gut regions of X. latifrons is so similar to that described previously for O. trispinosa, discussion here will be limited to areas of the lateral body and fins that are diagnostically important. Lateral body The lateral surface of the body above the gut in preflexion larvae is covered with melano- phores, with the exception of an elongate, unpigmented area along the dorsal midline above the pectoral fin. Melanophores cover nearly the entire caudal region of the body, with the exception of the notochord tip. Mel- anophores extend along the dorsal and ventral mar- gins of the unpigmented area around the notochord tip (Fig. 7A). By early flexion, melanophores completely cover the notochord tip (Fig. 7B). Three patches of melanophores appear along the dor- sal midline near the notochord tip by about 10.5 mm SL (Fig. 7C). The anterior patch forms the first band which extends from the dorsal margin of the body, posterior to the pectoral-fin base, to the anterior lat- eral gut area. The second band extends dorsally to connect with pigmentation seen on the first dorsal fin. The third band transverses the body between the posteriormost dorsal spine and the gut. Four or five distinct bands of pigmentation form between the ante- rior edges of the second dorsal and anal fins and the caudal fin. The first two bands transverse the body between the anterior and posterior margins of the sec- ond dorsal and anal fins. One or two bands are present between the posteriormost dorsal- and anal-fin rays and the hypural margin. The last band is the widest and covers the hypural region, extending into the cau- dal fin approximately 15-20% of its length (Fig. 7, C andD). Fins Melanophores begin to appear on the ventral margin of pectoral-fin base at about 5.0 mm. The pec- toral-fin base is completely covered with pigment by 6.0 mm. Pigmentation is always absent from the pectoral-fin blade, rays, and membrane (Fig. 7, A and B). A small patch of melanophores appears in the dor- sal finfold above the constriction of the gut at about 6.0mm (Fig. 7B). This pigmentation later expands and is that seen on the spinous (first) dorsal fin in late flexion and postflexion larvae (Fig. 7, C and D). Poste- riorly, an additional patch of pigmentation begins about one-third of the distance between the anus and noto- chord tip and exhibits an irregular edge, appearing somewhat serrated, below the dorsal margin. In preflexion larvae, a large patch of melanophores covers most of the anal finfold beginning at approxi- mately one-fourth of the distance between the anus and notochord tip and continuing its entire length. The rela- 408 Fishery Bulletin 9 1 (3), 1993 :'«.■■'» mU-1—* ill » » -" "i» ■' ' J " "* /el)} r* *' ■« 2 mm B 8.7 mm 10.8 mm 18.3 mm Figure 7 Larval stages ofXeneretmus latifrons. (A) Preflexion larva 5.2mm SL, CalCOFI 6501BD-83.43. (B) Early flexion larva 8.7mm SL, VPA, Lambert Channel 4/22/90 No. 5. (C) Flexion larva 10.8mm SL, CalCOFI 5403CR-83.48. (D) Postflexion larva 18.3mm SL, VPA, Lambert Channel 5/3/88 No. 4. Busby and Ambrose. Development of larval and juvenile Odontopyxis trispinosa and Xemeretmus latifrons 409 tively large unpigmented area anterior of this patch is an important diagnostic feature (Fig. 7, A and B). Both the dorsal and ventral finfold melanophore patches are constricted nearly to the body margins at a point about 80%-90% of the distance between the anus and notochord tip (Fig. 7B). As notochord flexion progresses, the ventral patch of pigmentation poste- rior to the constriction on the anal finfold is seen as the post-hypural pigment present on the caudal fin of postflexion specimens (8.7-22.0 mmKFig. 7C). This cau- dal-fin pigmentation extends posteriorly to about 15- 20% of the caudal-fin length (Fig. 7, C and D). The remaining melanophores on the dorsal and anal finfolds migrate toward the body margins as the finfold recedes. Discrete melanophore patches remain on the anterior and posterior margins of the second dorsal and anal fins in postflexion larvae (Fig. 7D). Osteology Although precursors of some bony struc- tures such as dermal plates and fin rays are discernable as early as 8.0 mm, ossification in X. latifrons does not begin until approximately 13.0 mm. The sequence of ossification of bony structures in X. latifrons is nearly identical to that described previously for O. trispinosa . The most important difference to note is that ossifica- tion in X. latifrons begins later (13.0 mm), progresses more rapidly for lateral body plates, and is slower for most skeletal elements than O. trispinosa. The intent here is to highlight important differences between the two taxa. Cranium All parts of the cranium, with the excep- tion of the sphenotic, prootic, epiotic, tabular, and pterosphenoid are ossified at 13.8 mm. These remain- ing cranial elements ossify at about 30.5 mm. Mandibular region Ossification of all mandibular structures is complete at 13.8 mm. Spines Head spination is generally reduced in X. latifrons larvae. Xeneretmus latifrons larvae have no inferior infraorbital, postocular, or posttemporal plates and have only four superior infraorbital spines. The rostral spine, which is weaker than that of O. trispinosa, and sclerotic plates do not develop until the late postflexion stage (about 25-30 mm). The tym- panic and frontal spines are more pronounced in X. latifrons larvae than in O. trispinosa. Palatine region The palatine, quadrate, meta- pterygoid, mesopterygoid, ectopterygoid, and symplectic are ossified at 13.8 mm. Opercular region The preopercle, opercle, sub- opercle, and interopercle are ossified by 13.8 mm. Hyoid region The basihyal, hypohyal, urohyal, ceratohyal, epihyal, interhyal, glossohyal, and bran- chiostegal rays are ossified by 13.8 mm. The hyo- mandibula is ossified at 30.5 mm. Branchial region The pharyngobranchial teeth, pharyngobranchials (n=A, fused as in O. trispinosa), ceratobranchials (n=5), and epibranchials (rc=4) ossify at about 21.0 mm. The remainder of the branchial ap- paratus including the basibranchials (n=3) and hypobranchials (n=3), are ossified by 30.5 mm. Appendicular region The cleithrum, postcleithrum, supracleithrum, and coracoid are ossified by 13.8 mm. Pelvic-fin spines and rays and all pectoral-fin rays are complete at 13.8mm (Table 7). The basipterygium and posttemporal ossify by 21.0 mm. The scapula and three radials supporting the pectoral fin are ossified by 30.5 mm. Median fins All dorsal-, anal-, and caudal-fin spines and soft rays are ossified by 13.8mm (Table 7). The caudal fin has 6 superior principal, 2 superior pro- current, 6 inferior principal, and 1 inferior procurrent rays (2+6+6+1=15 total). Ossification of the hypural plate and the pterygiophores supporting the dorsal- and anal-fin rays was not complete in the largest speci- men examined (39.2 mm). Vertebral column Notochord flexion begins at ap- proximately 8.5 mm and is completed by 11.0 mm. All vertebral centra and the urostyle are ossified by 13.8 mm. All except the two posteriormost neural and three haemal spines are also ossified at 13.8 mm. Ossi- fication of all neural and haemal spines is complete by 14.5 mm (Table 7). Dermal plates All dermal plates are ossified by 13.8 mm. Sequence and direction of formation and os- sification is the same as previously described for O. trispinosa. Xeneretmus latifrons, however, has higher DLP+MDP, ILP, and VLP+MVP lateral body plate se- ries counts than O. trispinosa (Tables 4 and 7). A maxi- mum of five pectoral-fin plates was counted in X. latifrons. Discussion Summary comparison of O. trispinosa *M/1 X. latifrons larvae Larvae of O. trispinosa and X. latifrons can be distin- guished by pigmentation, morphological, and meristic characters. Larvae of O. trispinosa possess a semicircular patch of melanophores that nearly covers the entire caudal finfold. This character is diagnostic and present throughout development. The caudal finfold of pre- flexion X. latifrons lacks pigmentation except for a patch located near the ventral margin of the noto- chord tip. This patch becomes elongate as notochord flexion progresses and becomes a band at the hypural margin. This band extends onto the caudal fin and may cover as much as 20% of its anterior surface. Preflexion larvae of O. trispinosa possess a small patch of melanophores on the anal finfold immediately posterior to the origin, X. latifrons larvae have a large 410 Fishery Bulletin 91(3), 1993 + 1 1 O0 CO H CO GO CO 0h CO CO Tf CO CO CO £ * OJ Oh CO CO O X CO CO CO CO ^ CO CO CO J3 & Oh ►J 1 1 CO CO H O IN O >> J CO CO ^ Tf t}" tJ- c T3 o O £ PQ OJ Oh 1 1 CO CO i-t CO CO CO cc CO CO Tj* CO CO CO -a CO i~ o Oh X a o S o + 1 1 CO CO •— < Q0 co co c 0h CO CO -^ CO CO CO bo c hJ a '5 V n3 03 CU HJ CM CM CM ■— ' ^h <-< -o O ,*J* M* Tf ^ "^ ^f c 3 > 03 Ft -6 0) II CO C CO CO CO CO CO CO PL, > - E «' o -S C3 — cj (m S * 03 OS rH i-l O O iH Qj co -*r Tt -^ - GO c 'S. T3 U0 t> CD U0 U0 CD c "li* cu CO o "5 3 to 2 3 03 "o3 a (M CN CM CM CM .5 M CM CM CN CM CM CM 15 T3 c -? o. *- Q- cC a; m 3 Oh CO -a II C3 0h T3 3 h ii PL en Q c 03 t- t> CO t> C- CO s y= 03 co co 03 Eh QJ O 0J -2 P C t- CD N f- t- CD "5. CO CO I- -o c t- -a si! cm' aicocoicoiooocM n cu ^ 1 ^ ^ o 1 ^h'cn'co'^^o^ctj t^| COCTi'— 2 wk-long) planktonic larval stages, MFRs can function as sources of benthic recruits to exploited, often distant, sink populations. The sub- sequent contribution of these recruits to the fishery might counteract (or forestall) the effects of recruitment overfishing (Carr and Reed, in press; Russ et al., in press). Through the directed movements of settled stages, MFRs might also augment the stand- ing biomass of stocks, including spawning adults, in exploited areas adjacent to MFRs as well as within the MFRs themselves (Polacheck, 1990; Roberts and Polunin, 1991). MFRs thus might provide a manage- ment tool that addresses growth over- fishing, particularly for multispecies fisheries on tropical coral reefs where conventional management is ineffec- tive (Roberts and Polunin, 1991). Despite the recognized importance of MFRs as sanctuaries for conserv- ing biomass, their contribution to fishery stocks and yields in adjacent exploited areas is poorly understood (Davis, 1989; Polacheck, 1990; Rob- erts and Polunin, 1991; Russ et al., in press). An introduction to the prob- lem is provided by Polacheck (1990), who expands the Beverton-Holt equa- tion (Beverton and Holt, 1957) de- scribing the effects of a year-round refuge on the spawning stock biomass (SSB) and yield of a cohort in a sur- rounding exploited area. Although ar- eal closures on coral reefs are one identified application, Polacheck's (1990, Table 1) simulations use empirical data on the growth, matu- 414 DeMartini. Potential of fishery reserves for managing Pacific coral reef fishes 415 rity, and harvesting schedules of Georges Bank cod Gadus morhua and haddock Melanogrammus aeglefinus. The growth and mortality dynamics of tropical coral reef fishes, however, may be quite unlike those of higher- latitude species (Munro and Williams, 1985; Longhurst and Pauly, 1987). Further simulations using growth and mortality data of other fishes are needed. Observations of traditional practices by marine is- landers (Johannes, 1978) suggest that reserves can augment the SSB, and perhaps the yield of tropical reef fishes. To date, controlled empirical measurements of fishery yields in an area adjacent to a MFR have been published for only one study site (Sumilon Is- land, central Philippines; Alcala and Russ, 1990). The changes in catches that Alcala and Russ (1990) ob- served, however, were based on yield, not yield per recruit (Y/R), over a 1-year period and therefore may not have represented equilibrium conditions. The present paper evaluates the effects of perma- nently closed MFRs of different sizes on net changes in SSB and yield for several types of tropical Pacific reef fishes. The author simulated various combinations of fishing mortality and emigration-immigration ("transfer") rates for fishes having different but typical natural growth and mortality schedules. Because tropi- cal reef fishes have higher natural growth and death rates than do temperate zone fishes, the focus is on the relative sensitivity of spawning stock biomass per recruit (SSB/R) and Y/R to MFR size, transfer rates, variations in natural mortality and growth, fishing ef- fort, and age at first capture. Another objective is to compare the management potential of "single large or several small" (SLOSS) (Simberloff, 1988) reserves of equal total area. Finally, the potential of MFRs on island reefs is discussed, with particular reference to reef areas surrounding the island of Oahu. Hawaii. Methods Parameter values for von Bertalanffy growth rates, derived mortality rates, and maturity and harvesting schedules were chosen to bracket the spectrum of life histories and exploitation characteristics of tropical Pacific reef fishes. At one extreme, values were as- sembled to describe a reef transient (e.g., a jack of the family Carangidae) that is relatively slow-growing but long-lived and large-bodied is likely to travel rapidly over relatively large distances. Such species often sup- port valuable commercial fisheries. Values used for growth and mortality rates of the jack resemble those 2Ralston, S., and H. A. Williams. 1988. Age and growth of Lutjanus kasmira, Lethrinus rubrioperculatus, Acanthurus lineatus, and Ctenochaetus striatus from American Samoa. of many species of commercially important snappers, groupers, and jacks from the South Pacific and other tropical seas (Munro, 1983; Munro and Williams, 1985). A short-lived, fast-growing but small-bodied reef damselfish (family Pomacentridae), with limited move- ments after settlement, was used to represent the op- posite extreme. Such small tropical reef species are collected for the aquarium fish trade. In between these two extremes lies a broad con- tinuum of fishes with intermediate longevities, body sizes, and movement rates. These fishes probably rep- resent most species targeted by recreational and artisanal fisheries on tropical reefs. The few data avail- able (Galzin, 1987; Ralston and Williams2; Russ and St. John, 1988; Dalzell, 1989) suggested a range of moderate growth and mortality rates for a number of Pacific parrotfishes (family Scaridae) and surgeonfishes (family Acanthuridae). For convenience, these fishes were labeled the "surgeonfish" type. The growth parameters used in this study were based partly on published values for a particular species popu- lation, complemented by data for other Pacific popula- tions of the same species. For the pomacentrid, the au- thor used a Moorean population of Stegastes nigricans (Galzin, 1987). He selected Ctenochaetus striatus as the surgeonfish; the length-weight relation was based on a Moorean population (Galzin, 1987), and a Samoan popu- lation provided the VBGF parameters (Ralston and Wil- liams, 1988). Sudekum et al.'s (1991) data for NWHI Caranx ignobilis were used to represent the jack. Thus the degree to which values were population-specific var- ied among the three fish types. Natural mortality rates iM) were estimated using Pauly 's (1980) multiple re- gression of M on maximum size, growth coefficient, and mean water temperature (Pauly and Ingles, 1981); 25°C was chosen as representative for shallow, tropical Pa- cific waters. All parameter values used are listed by fish type in Tables 1 and 2. Modeling closure effects Polacheck's (1990) model of the effects of closure size (1-50%), rates offish transfer between closed and ex- ploited areas, and fishing mortality rate on the bio- mass and production of fishes in an adjacent exploited area was used with one small but important differ- ence: the use of a three- rather than a four-sided closed area. (Closures on island reefs usually extend seaward from the shoreline, so fishes can move across upcoast, downcoast, and offshore boundaries only. Other fac- tors being equal, dispersion rates out of and into shore- Dep. of Commer., Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. H-88-18, lip. 416 Fishery Bulletin 91 [3), 1993 Table 1 Length-weight relations, von Bertalanffy growth function parameters, and natural mortality rates used to simulate spawning stock biomass per recruit and yield per recruit contour surfaces for damselfish (based on Moorean Stegastes nigricans from Galzin, 1987); surgeonfish (length-weight relation: Mnorean Ctenochaetus striatus from Galzin, 1987; VBGF parameters: Sa- moan C. striatus from Ralston and Williams2); and jack (NWHI Caranx ignobilis from Sudekum et al., 1991). M values were estimated after Pauly ( 1980) and Pauly and Ingles ( 1981; see Methods). Variable Damselfish Surgeonfish Jack Length-weight relation (Total Length/g) L. (Total Length, cm) »' (g) KtVyear) t,Jyear) Mil/year) W = 0.0195Li 17.5 128.9 0.374 -0.042 1.5 W = 0.0111L:U" 28.2 348.6 0.447 -0.760 1.0 W = 0.0072L298 217 120,139 0.111 -0.097 0.2 Table 2 Age-specific parameters used to simulate the standing stock and catch (yield) values for each of the reef fish types described in Table 1. Mean Partial Age weight' Percent recruit- Fish type (yr) (gl mature- ment' Damselfish 1.0 4.77 100 1.00 2.0 32.40 100 1.00 3.0 72.20 100 1.00 Surgeonfish 1.0 40.81 0 0 2.0 101.87 100 1.00 3.0 164.43 100 1.00 4.0 217.18 100 1.00 5.0 257.24 100 1.00 Jack 1.0 110.0 0 0 2.0 861.7 0 0 3.0 2587.0 0 0 4.0 5348.6 100 1.00 5.0 9061.8 100 1.00 6.0 13,569.6 100 1.00 7.0 18,688.4 100 1.00 8.0 24,235.1 100 1.00 9.0 30,042.0 100 1.00 10.0 35,963.5 100 1.00 11.0 41,878.8 100 1.00 12.0 47,691.4 100 1.00 13.0 53,327.1 100 1.00 14.0 58,731.1 100 1.00 15.0 63,865.3 100 1.00 'Based on mean weight-at-age estimated from weight-specific VBGF; see Table 1 for sources. -Sexual maturity assumed 100'/r complete at age corresponding to the asymptote of the Von Bertalanffy growth function curve; all fish were assumed immature prior to this age. 'All species were defined as fully recruited to the fishery at the age of 100% sexual maturity, no fish entering the fishery prior to that age: Age-at-first-capture, A, = 1.0, 2.0, and 4.0 years for the dam- selfish, surgeonfish, and jack, respectively. line-bounded reef closures will be one- fourth slower than the respective rates for completely nested closures of equal areal extent. ) Polacheck's modified Equation 8 was T12 = 3-T,s-(R1/Rs)-\ (1) where Tv, is the instantaneous emigration rate from closed area 1 to exploited area 2; Tls is the emigration rate from a closure of Rs standard size; and i?, is the frac- tional closure size evaluated. As in Polacheck (1990), the initial fish densities were assumed homogeneous in both areas, and thus the number of fish initially present in area 1 at time / was defined as N,, = R,-N,„ (2) where NtotaU = number of individuals in the co- hort entering areas 1 and 2 at time /. Tls was allowed to vary among fish types, as types surely differ in their fundamental emigration and im- migration rates. Selected values of Tu spanned realistic values for each fish type (damselfish: 0.001, 0.01, and 0.1; surgeonfish: 0.1, 0.25, and 0.5; jack: 0.1, 0.5, and 1.0). T21, the rate of immi- gration into area 1 from area 2, given the type- specific value of T|,, was as defined by Polacheck (1990, Equation 6). A basic assumption of the model is that fishing does not affect fish distributions by altering habi- tat or by promoting density gradients between areas 1 and 2, i.e., dispersal is random and uni- form throughout both areas and across their boundaries (Polacheck, 1990). This is realistic for fishes like jacks that range widely or that have very large home ranges. However, the assump- tion of uniform dispersal throughout both areas is debatable for fishes with home ranges that are small relative to the size of MFRs, particularly for strongly site-attached, territorial species like damselfishes whose ambits are trivial compared to the area of a reserve. Transfer rates that func- tionally approximate those based on an as- sumption of uniform, random dispersal might be tenable for home-ranging (or even territorial) or- ganisms, if emigration out of MFRs into adjacent fished areas followed a "stepping stone" pattern with minimal time lags. Limited field data at present both support (Walsh, 1984) and counter (Wellington and Victor, 1988) such a model. The key factors here are the magnitude of time lags and the relative sizes of individual home ranges within particular MFRs. DeMartini Potential of fishery reserves for managing Pacific coral reef fishes 417 An Rs value of 0.10 was used to represent the "stan- dard-size" closure (Polacheck, 1990). Fishing mortality rates (F,) were input over the range from 0.1 to >3M, with special evaluation of Fc = 0.5 M, M, and 2M, based on the best estimate of natural mortality for each fish type. Total fishing effort (F2) was homogeneously re- distributed throughout the exploited area and fixed in magnitude (at a given Fc), regardless of the size of the closed area (Equation 7 in Polacheck, 1990). SSB/R and Y/R are used as the primary bases of evaluation. Although measures of biomass are not usu- ally applicable to species like the damselfish that are harvested on a numerical basis (see Ingles and Pauly ( 1984] for examples of consumptive exploitation of small- bodied fishes in artisanal fisheries), these measures were evaluated in the same way to maintain consistency. First, the SSB and yield of a cohort were calculated, by using Polacheck's (1990) Equations 2 and 3 for the numerical standing stock and the numerical catch of each cohort comprising that stock: n I uvM+1 (*=1) n (t=i) WM ■ %mat,tl), Wl+l ■ %mat(+1), and n X (c2,( (t=\) W;); (3) (4) (5) where ATU+1 and N2M are the numbers of a cohort surviving in the respective area at time (t+1); C2I is the numerical catch of a cohort in area 2 at time t\ W, and W,+1 represent the mean weights of individual fish of the cohort at times t and {t+1), respectively; and %mat,+1 is the percentage of the cohort that is sexually mature at time (t+1). SSB/R and Y/R were then calculated by standard- izing the total spawning biomass and yield, respec- tively, of the cohort over its life span by the total num- ber of recruits potentially (area 1) and directly (area 2) entering the fishery from that cohort (Gabriel et al., 1989). SSB/R was evaluated in terms of percentage of the virgin stock biomass possible if fishing was disal- lowed in both areas (Polacheck, 1990); 20% of virgin biomass was considered the threshold for recruitment overfishing (Beddington and Cooke, 1983). Two series of simulations were run. One series treated the T12 and T21 values as fixed, assuming that, analogous to Polacheck's (1990) analyses for Georges Bank cod, transfer rates would remain constant and independent of relative fish densities in the two areas. Additional simulations, using the same initial Tu val- ues, were run for selected cases; in these runs, subse- quent values of the transfer rates were treated as a density-dependent function of the changing, relative fish densities in the two areas. Subsequent values of Tr2 and T.n were adjusted as follows: T12,M = TVLI ■ I {NJNJI (NJN,) h and (6) T2i,M = T,h, • I (N2JN2) / (NJNJ |"; (7) where N} is defined in Equation 2 and N2 is the initial number of the cohort present in area 2; iVu and N2l are the numbers of fish surviving in the respective area at subsequent age t; and x is the power used to scale the ratio of fish densities. In Equations 6 and 7, the ratio of the numbers offish surviving at time t was further adjusted by the ratio of initial densities in the two areas in order to scale for the propensity to emi- grate at the onset. Two values of .r were evaluated: 0.125 (eighth root) and 0.5 (square root). (Note: When x equals 0, the Tl2 and T.n are fixed; when .r equals 1.0, these rates are continually readjusted by the changing ratio of relative densities.) Exponents of 0.125 and 0.5 were chosen because they bracketed rate changes of reasonable magnitude. In the surgeonfish, for example, an eighth-root adjustment would initially accelerate a median Tu of 0.25 by about 20% for a median closure size of 25%, at an F, of 1.0. The corresponding rate increase due to a square-root adjustment would be 60%. Inclusion of a term for the density-dependent ad- justment of transfer rates, as stocks are fished-down in the non-closed area, extends Polacheck's (1990) evaluation. This is perhaps an unnecessary refinement for stocks such as Georges Bank cod for which har- vesting by trawl might reduce habitat quality in the non-closed area (Polacheck, 1990). However, non- destructive (e.g., hook and line) methods of artisanal fisheries on coral reefs do not reduce habitat quality. Furthermore, fishes may emigrate at an accelerated rate from a closure into the surrounding non-closed area where densities continue to decrease (tantamount to improving habitat quality). Compensatory emigra- tion resulting from a density gradient is recognized as potentially important in the siting and design of na- ture reserves (Schonewald-Cox and Bavless, 1986). Complementary simulations In addition, the effects of varying M and the age-at- first capture (A,) on total biomass per recruit (B/R) were simulated with the conventional Y/R model of Beverton-Holt (Sparre et al, 1989). Fishing mortality was evaluated at 0.5M, M, and 1.5M or 2M. Ac was 418 Fishery Bulletin 91(3). 1993 examined with the most reasonable estimate of A, ± 507 for each of the three fish types (see footnote C of Table 2 for the type-specific values of A, used in modeling closure effects). An age-at-recruitment to fishery habitat (A, ) value of 1.0 year was assumed for the surgeonfish and jack, and an A, of 0.5 year for the damselfish. Results General Results corroborate several of Polacheck's (1990) major results: SSB/R increased, while Y/R generally decreased, with increasing refuge size (Fig. 1). Even in cases where a closure positively affected yield, Y/R increased little. Y/R often was greater at higher levels of fishing effort (Fig. 1 ) and usually decreased at larger refuge sizes; the rate of decrease was less at higher fundamental transfer rates (Fig. 2, A-C). SSB/R was generally greater for closures of larger size at any particular transfer rate (Figs. 1 and 2). The positive effects of a closure of a given size on SSB/R were greater at lower rates of exploita- tion and diminished at higher transfer rates (Fig. 3). The rate of increase in SSB/R with refuge size was greater at lower transfer rates (Fig. 1, A-C). The progressive gain in SSB /R usually was greater than the progressive loss in Y/R for refuges of increasing size (Fig. 2). Another general pattern was the relative importance of the interactive effect of natural and fishing mortal- ity rates, compared to age at recruitment to the fishery. For most combinations of M and F,, B/R was more strongly influenced by these two rates than by age at recruitment (Fig. 4, Table 3). Depending on fish type, B/R values ranged two- to five-fold as M and Fr values varied ±507 of their midpoint estimates, but ranged only twofold or less as A, varied ±50% (Table 3). By comparison, transfer rates had relatively little effect on SSB/R, compared with the effects of natural and fishing mortality rates. For a median-sized clo- sure of 257, changes in percentage of virgin SSB/R differed <107 within fish type for transfer rates that varied as much as two orders of magnitude (Fig. 3). Another generality, not previously made explicit, is that ages at maturity and at first capture can more strongly influence spawning stock than can the pres- ence and size of a refuge. The presence of a refuge will have a relatively weak effect on spawning biomass if the resource begins to be heavily exploited well before sexual maturity. In extreme cases, SSB/R might not be appreciably enhanced by large closures, despite rela- tively low transfer rates. Fish types Perhaps the most significant, specific result of the simulations was that fish types differed in how ref- uge size affected SSB/R levels. Large, apparent gains in percent virgin SSB/R of the damselfish occurred only at the expense of large losses in yield (Figs. 1A, 2A, 3A). Nontrivial (>5%) gains in percentage SSB/R of surgeonfish could occur at small (0.1) refuge sizes, although the overall rate of increase in SSB/R would be greater for larger closures (Figs. IB, 3B). For the jack, however, gains in SSB/R of magnitude similar to those of the surgeonfish could be realized with in- creasing refuge size only for large (fl, > 0.3) closures (Figs. 1C, 3C). Simulation results appeared to vary among fish types in response to their differing growth rates and natural mortality and exploitation schedules. For ex- ample, in comparing the results for the high-mortal- ity surgeonfish with those for the low-mortality jack, substantive gains in SSB/R of the jack (restricted to large closures) were further restricted to low (F,. < 0.4) levels of fishing effort (Fig. 1C). SSB/R dropped below 207 of the virgin stock at F, = 0.3, the fishing effort that was optimal for maximizing Y/R (Fig. 1C). For the surgeonfish, appreciable (>10%) gains in SSB/R could occur under heavy (F,. > 1) exploita- tion rates for refuges as small as 10-207 (Fig. 3B), and >207 of virgin SSB/R can be sustained at F < 1.15 (Fig. IB). The observed differences among fish types in the potential for MFRs to increase SSB/R also were strongly influenced by the ages at sexual maturity and recruitment to the fishery. Ages at first maturity and at first capture were attributes that differed greatly among the three fish types (damselfish: age 1; surgeonfish: age 2; jack: age 4; Table 2). Median B/R (equal to SSB/R for all except the damselfish at A, = 0.5 year) varied by 307 for the damselfish, 1007 for surgeonfish, and 357 for the jack among ages at first capture that varied ±507 of their respective midpoints (Table 3). Compared with age at first catch, transfer rates had less effect; SSB/R consistently varied <107 for the damselfish (with twofold variation in very low transfer rates; Fig. 3A), for the surgeonfish (with fivefold variation in moderate transfer rates; Fig. 3B), and for the jack (with tenfold variation in rapid trans- fer rates; Fig. 3C). At a reference F, of 1.0 and a me- dian closure size of 257, SSB/R varied from <6 to 357 of virgin stock biomass, depending on fish type (Fig. 1,A-C). Compensatory transfer rates Even small (eighth root of density ratio) compensatory increases in transfer rates could have negated poten- tial increases in SSB/R that might result from the presence of a closure. A larger adjustment with the square root of the density ratio, of course, can have an DeMartmi: Potential of fishery reserves for managing Pacific coral reef fishes 419 . 1/ , k ' DAMSHHSH ' ' ' 4.0 v/ A / / / , 3.5 / \ ' \ / / f / / j 1 1 3 0 \ ' \ / / / / / / 2.5 / 7 n / // / V / / / / // /' 28 / / /*/ / / 1.5 / ^/ />' / V' V / / / / ^^-^^ / ^^^ , / ■»/ 1 i-l — — — — ~~~^^ ' — ' — ^ ' ./ / 44 y^ % ~~~" *s^-^-- — " — """ -•' '_-— -~~~^ / ^^" / ^^ "~~ — ~- "~ ^.^J___.— - — 52^^^ 0.5 ___ =^- o,J 0.3 0.4 0.5 1.25 Figure 1 Overlay of a yield-per-recruit surface (various dashed lines! on a percent- age spawning stock biomass-per- recruit surface (solid lines I as a func- tion of refuge size !/?,> and fishing mortality rate (F, ) when transfer rate i T, i is constant: (A) damselfish ( 7",. = 0.01 1; (B) surgeonfish iT,, = 0.251; and (C) the jack (7,, = 0.5). Spawning bio- mass values are expressed as a per- centage of virgin biomass that would accrue if no fishing were allowed in areas 1 and 2. 420 Fishery Bulletin 91(3). 1993 o k»22 8 £ ° *" o o o o OO O O o o ° Q OO o U/BSS «B«lu«xj»d U/A«ta|iM3j*d i/i o «» o J? O ul O 1/1 o I 10 c z s o S 10 CO gggg Sis°8a8S8S8g°8s§a8ag P» f» O* CM "■ ^ «1 !■» «*«*-. ,* W (* Ol - j= 3 % CO '^ T-1 a: 0) •?, N T3 t/J C CO bO CO C cd ro /. trt in a- C c — II a &-< >, fj N z U U -a * — ' fc_ - ■1-. -J cu ~ m cu c c _ . ._ Q OJ ' ' 1> T1 — *- b OS W i— I ^ — cu 03 - is s ^~ 0) ^£ g en O E I i £ -j £ ... < cu t5 - is C- cu "*^ g co ■- 5 t. « « >> g c S co CO O Jl a» c 3 DC C g 5 w> S 2 c ^ ox « CU ^ -J= — 15 C cu *> -5 &. \ s N \ s s ^S' vv^ "\ ** . -* .•-• """"-- ** '' .^^ ." O O o ° o o 1 1 f 1 II / / s > / 1 / 1 / ' / / / / / / "7 / / / ' / / / ' i "/ t i > / , / i ■' >' / i ' ' / t / > / / i / • s / 1 r / / / / .< 9 I I I '; i : : i • i • / i ' : III, I ' ! I • ' : : i i : ! ///// / ; / ' i 1 ! i i > i : I I ,' i I i I ' i N ?N ■ II N E S N\ \ N. X N M 5 N. N , r--- _. o "- "-•--^ - "* *-- -- — — — - — — — „ -__. .___ __ ■7 /' / .' / i I i i ! : ' / / / y A s CO bo u 3 i a r u w a c M C5 DeMartini: Potential of fishery reserves for managing Pacific coral reef fishes 423 Table 3 Equilibrium B/R (total biomass per recruit) values without any refuge at select combina- tions of A, (age-at-first-capture), M (natural mortality rate), and F, (fish ng mortality rate) for each of the three fish types. See text Methods, Table 1, and footnotes to Table 2 for basis of chose n A,, and M values. F was evaluated at va ues equal to 0.5 M. M. and 1.5 Af. B/R at the most reasonable ( midpoi nt) values of A, and M art emph isized in bold and underlined t) 'pe, res pectively, and the relative influences of A , Af, and Fc on B/R are noted. Ar = 0.5 year (damselfish) and 1.0 year surgeonfish and jack). AtAf Damselfish B/filgi A, = 0.5 Ac = 1.0 A, = 1.5 = 075 L5 2.25 0.75 15 2.25 0.75 15 2.25 IfF, = 0.75 5.4 2.1 1.1 6.7 2.3 1.0 7.4 1.9 0.6 1.50 2.1 1.1 0.7 3.3 1.4 0.7 4.0 1.2 0.4 2.25 1.1 0.7 0.5 2.0 M = F »A, 0.5 2.6 0.9 0.3 AtAf Surgeonfish B /ft (g) A, = 1.0 Ac = 2.0 A, = 3.0 = 0j5 10 15 05 Lfi 15 oj5 10 15 IfF = 0.50 119.9 66.8 44.8 110.0 40.5 17.4 85.4 19.9 5.3 1.00 66.8 44.8 33.2 66.7 28.7 13.4 54.0 14.4 4.2 1.50 44.8 33.2 26.2 47.3 M>F 22.1 = Ar 10.9 39.2 11.3 3.4 AtAf JackB/ffikgi A, = 2.0 A, = 4.0 A. = 6.0 = 0.1 0.2 0.3 0.1 0.2 0.3 0.1 0.2 0.3 IfF = 0.1 96.4 36.6 17.6 112.3 40.3 18.0 123.0 39.8 15.6 0.2 40.4 19.4 10.8 54.5 24.3 12.2 65.6 25.8 11.3 0.3 21.5 11.9 7.3 32.8 M>F 16.5 >A 9.0 42.5 18.6 8.7 inconclusive: Alcala and Russ's (1990) observations of an effect on reef fish catches adjacent to the Sumilon Island reserve were from a series of years in which the reserve operated and from only one year of reduced catches that began 18 months after the reserve's protected status ended. Catches were dominated by one taxon (fusiliers, family Caesi- onidae; 659c of total) that used the reef for nocturnal shelter but whose zooplankton prey may have been unrelated to reef area (Alcala and Russ, 1990). Brief changes in yield, rather than Y/R, particularly on a small (0.5 km2) spatial scale, may rep- resent nonequilibrium phenom- ena (e.g., lagged effects on adult abundance resulting from a lo- calized change in recruitment). Alcala and Russ's (1990) obser- vations therefore are not neces- sarily inconsistent with simula- tion results (see Russ et al, in press). Clearly, additional empirical measures are needed for Y/R as well as SSB/R of fishery re- sources in exploited regions ad- jacent to MFRs. For the present discussion, however, this paper will focus on simulated SSB/R results. even larger countering effect (Fig. 5). The effect of com- pensatory rate adjustments apparently varied with fish type in a manner not directly related to fundamental transfer rate. When the surgeonfish and jack were each evaluated using Tu - 0.1 as the fundamental rate, compensatory dampening in the rate of SSB/R gain at larger refuge sizes was less for the surgeonfish ( Fig. 5A) than for the jack (Fig. 5B). Discussion MFRs and enhanced Y/R Simulations to date suggest that MFRs have the poten- tial to augment SSB/R, but enhance Y/R little, if at all, in adjacent non-closed areas (Polacheck, 1990; the present study). The only empirical study to date was Relative influences of input parameters Growth rate is a major influence of biomass accrual within refuges (Polacheck, 1990), as fast-growing fish can elaborate more surplus production (yield) per unit time and unit area of refuge than can slow-growing fish. And since growth and mortality rates are linked, it is not surprising that mortality rate is important, as fishes with a higher natural mortality can support fisheries in which they are harvested more heavily and earlier in life than can fishes with a lower mortal- ity rate. In the basic yield model, the effects of natural mor- tality and level of exploitation overwhelm those of age at first capture. Proportional changes in rates obvi- ously have larger effects than equal-sized changes in the time period over which the rates apply. Of greater interest here are the relative magnitudes of the effects 424 Fishery Bulletin 91(3). 1993 70 A SURGEONFISH 60 E ^* a in so __-— — " a 40 a c 8 30 8 £ _...-— " ,...;;-::;'" 20 ...-■•' 10 0 0.0 0.1 0.2 0.3 0.4 0.6 Fc 0.5 i.o 2.0 70 g JACK 60 B » 50 CO §> 40 1 g 30 3 °" 20 =====EE^= --«====«=-------'-:-"--— 10 0 0.0 0.1 0.2 0.3 0.4 0.5 Fc 0.1 0.2 0.4 Figure 5 Percentage changes in spawning stock biomass per recruit (SSBI R) as a function of refuge size (fl,), fundamental transfer rate (TV,), and two levels of density-dependent, compensatory adjust- ment of this fundamental rate: (A) surgeonfish (7V = 0.1): (B) jack (TV = 0.1). Within each triplet of lines, the upper, middle, and lower line depicts the unadjusted, adjusted 0.125-power, and adjusted 0.5-power rate, respectively. of ages at sexual maturity and first capture, versus those of refuge size and transfer rates. I suggest that the former can have large influences for relatively short-lived, fast-growing species like the surgeonfish for which changes in age at first capture of only ±1 year represent a large fraction of its life span. Circumstantial evidence suggests that the relative ages at sexual maturity and first capture can be im- portant. Prior evaluations of the potential for MFRs to conserve the SSB of overfished commercial stocks (e.g., red snapper, Lutjanus campechanus) have concurrently considered other management safeguards, such as size limits in the non-closed area (Plan Development Team, 1990). However, nearshore reef fishes subject to recre- ational harvest are often fished at sizes and ages con- siderably less than those at sexual maturity, even on moderately populated islands (e.g., the island of Ha- waii; Hayes et al., 1982). For reef fishes near densely populated areas where exploitation is likely to be intense on pre-reproductive fish, the potential for MFRs to enhance SSB/R may be severely compro- mised, even at closure sizes that are as large as is practical (e.g., 25%). Compensatory emigration Density-dependent increases in the fundamental transfer rate can in essence depress potential gains in SSB/R in a manner analogous to that of a higher but constant fundamental transfer rate. The obser- vation that gains in SSB/R at increasingly large refuge sizes were strongly offset by compensatory movements in the jack, but not in the surgeonfish, perhaps reflects the jack's slower growth rate, greater longevity, and lower mortality rate. Longev- ity, per se, may be important, because the model that was used to describe compensatory emigration is dependent upon time as well as age. SLOSS effects on MFR function Simberloff (1988) reviewed the SLOSS concept and the meager results to date regarding whether mul- tiple, small reserves function the same as single reserves of equal total size. Historically the issue of SLOSS has been applied to the preservation of threatened and endangered species or the conserva- tion of biotic diversity (Bell and Boecklen, 1990, but see Goeden, 1979). It is increasingly apparent that meaningful stewardship of the environment and its biota extends beyond these simple (although of- ten difficult to implement) criteria. Overexploiting an ecosystem's productivity by overfishing (Russ, 1991) is just as detrimental as recruitment or growth overfishing. Bohnsack (1991) recently applied the SLOSS con- cept in a review of the function of artificial reefs. Obvi- ously MFRs also can be evaluated in terms of SLOSS — the question then may be, "Do several small reserves potentially enhance SSB/R to the same extent as one reserve of equivalent total size?" A realistic example may be the relative fishery enhancement potential of ten 1% closures versus one 10% closure. Establishing a MFR of 10% may be impossible on a heavily popu- lated island (like Oahu, Hawaii) where shoreline de- velopment is near saturation. The siting of multiple, smaller refuges, each about 1% of the total area, may be feasible, however. But can a total closure of realis- tic size (10%) be beneficial, and can the potential of ten 1% closures approximate it? The simulation results suggest that a total closure of 10% may enhance the spawning stock of a fast- DeMartini: Potential of fishery reserves for managing Pacific coral reef fishes 425 growing, moderately vagile reef species such as surgeonfish: at a fishing mortality rate (Fc) of 1.0 (equal to M = 1.0), SSB/R can be increased by about 24% over the no-closure case for an MFR of 10%, if the lower bound of Tu (0.10) is used as the fundamental emigration rate. By using the higher bound for Tu (0.5) and an Fc still at 1.0, SSB/R is increased by 12% for the single 10% closure. The overall contribution of multiple, small closures will be less, however, than that of one closure of equal total size, in inverse proportion to the increase in the total perimeter of (hence dispersal from) the multiple closures. (This follows from the general rule that smaller reserves have greater perimeter-to-area ratios than larger reserves of equivalent shape [Schonewald- Cox and Bayless, 1986].) In our example, the total perimeter often 1% closures is about three times (VlO) that of one 10% closure. Hence, for a given fundamen- tal transfer rate, the actual transfer rate will be ap- preciably greater, and the additive contribution of ten 1% closures to SSB/R will be several times less, than that of one 10% closure. Multiple 1% closures might translate to an actual emigration rate approaching 0.5 for a surgeonfish that is moderately vagile (Tls = 0.1- 0.25) relative to an MFR of 10%. For the preceding arguments to hold, we must fur- ther assume that the multiple closures will not inter- act spatially. It seems reasonable though that for post-settlement stages of species with low fundamen- tal transfer rates, ten 1% reserves may function inde- pendently if well-spaced, perhaps even on a relatively small (e.g., 100-km perimeter) island such as Oahu, Hawaii. For simplicity, SLOSS arguments generally ignore the interactive effects of refuge shape and size. How- ever, habitat geometry (shape as well as size) impor- tantly influences dispersal into and from a habitat, regardless of whether an organism's movements are home-ranging or free-ranging (Stamps et al., 1987). Edge "permeability" (i.e., whether a reserve is sited within a larger area of homogeneous "soft-edged" habi- tat or is a functional island surrounded by discontinu- ous "hard-edged" habitat) also significantly affects dis- persal (Buechner, 1987; Stamps et al., 1987). Small increases in edge permeability have disproportionately large effects on dispersal when boundaries are hard, whereas habitat shape exerts a controlling influence on dispersal when boundaries are soft (Buechner, 1987; Stamps et al., 1987). For shoreline-bounded reef clo- sures, the shape of the MFR is clearly more important to residential reef fishes if the MFR is sited within a larger region of similar habitat than if different habi- tats (e.g., extensive sand channels) are used to provide its physical boundaries. Future research The overall net effect of the diverse compensatory and depensatory factors that may influence the movement patterns of fishes into and out of MFRs is beyond even semiquantitative appraisal at present. Major advances in our understanding of the function of MFRs await development of techniques that describe the funda- mental transfer rates of fishes (Polacheck, 1990) and that further estimate the changes in movement rates that may occur as densities change over time. The results of my preliminary analyses suggest that future studies should focus on fast-growing, moderately vag- ile species such as many surgeonfishes, rather than reef transients or philopatriots. My preliminary con- clusions should be reevaluated as more data become available for these and other types of reef fishes. Management potential MFRs have the potential to enhance the biomass and spawning stock of species with rapid growth and mod- erate, fundamental transfer rates (Polacheck, 1990; the present paper). Many factors, however, contribute to whether this potential is likely to be realized in prac- tice: the age schedule of harvest and the magnitude of fishing effort in the non-closed area; whether one ver- sus a few or many MFRs of a given total area are used, thereby promoting progressively larger increases in fundamental transfer rates; and whether compen- satory emigration acts to further inflate vulnerability and deflate SSB/R. Decisions as to the number versus size of MFRs that can be sited, given the total shore- line extent of reef available, are social issues. Size limits and fishing effort (bag limits) also are subject to their own management politics, and MFRs will not make them obsolete. Although the fast growth and moderate movement rates of certain tropical reef fishes predispose them to benefit from refuges, it is unlikely that the establishment of MFRs alone, without comple- mentary regulation of effort and the size composition of catch in non-closed areas, can augment the SSB/R of fishes on heavily exploited island reefs. Acknowledgments I thank M. Yong and especially D. Tagami for help with programming code. J. Polovina and D. Somerton offered stimulating discussion, and J. Bohnsack, J. Polovina, T. Ragen, C. Roberts, G. Russ, D. Somerton, and an anonymous reviewer provided constructive criti- cisms of draft manuscripts. T Polacheck graciously provided a copy of his computer program and assisted 426 Fishery Bulletin 91 (3), 1993 with program re-specifications. I accept full responsi- bility for any remaining errors. Literature cited Alcala, A. C. 1988. Effects of marine reserves on coral fish abun- dances and yields of Philippine coral reefs. Ambio 17(3):194-199. Alcala, A. C, and G. R. Russ. 1990. A direct test of the effects of protective manage- ment on abundance and yield of tropical marine re- sources. J. Cons. Int. Explor. Mer 46:40-47. Beddington, J. R., and J. G. Cooke. 1983. The potential yield of fish stocks. FAO Fish. Tech. Pap. 242, 47 p. Bell, G. W., and W. J. Boecklen. 1990. On island biogeography theory and nature re- serve design. J. Biogeogr. 17:97-101. Bell, J. D. 1983. Effects of depth and marine reserve fishing re- strictions on the structure of a rocky reef fish assem- blage in the north-western Mediterranean Sea. J. Appl. Ecol. 20:357-369. Bennett, B. A., and C. G. Attwood. 1991. Evidence for recovery of a surf-zone fish assem- blage following the establishment of a marine reserve on the southern coast of South Africa. Mar. Ecol. Prog. Ser. 75:173-181. Beverton, R. J. H., and S. J. Holt. 1957. On the dynamics of exploited fish popula- tions. Fish. Invest. (Series II) 19:1-533. Bohnsack, J. A. 1991. Habitat structure and the design of artificial reefs. In S. S. Bell, E. D. McCoy, and H. R. Mush- insky (eds.l. Habitat structure: the physical arrange- ment of objects in space, p. 412-426. Chapman and Hall., London. Bohnsack, J. A. In press. How marine fishery reserves can improve reef fisheries. Proc. Gulf Caribb. Fish. Inst. 43. Buechner, M. 1987. Conservation in insular parks: simulation mod- els of factors affecting the movement of animals across park boundaries. Biol. Conserv. 41:57-76. Buxton, C. D., and M. J. Smale. 1989. Abundance and distribution patterns of three temperate marine reef fish (Teleostei: Sparidae) in exploited and unexploited areas off the southern Cape coast. J. Appl. Ecol. 26:441-451. Carr, M. H., and D. C. Reed. In press. Conceptual issues relevant to marine harvest refuges: examples from temperate reef fishes. Can. J. Fish. Aquat. Sci. Clark, J. R., B. Causey, and J. A. Bohnsack. 1989. Benefits from coral reef protection: Looe Key Reef, Florida. Coastal Zone '89: proc. sixth symp. Coastal and Ocean Management 4:3076-3086. Cole, R. G., T. M. Ayling, and R. G. Creese. 1990. Effects of marine reserve protection at Goat Is- land, northern New Zealand. N.Z.J. Mar. Freshwa- ter Res. 24:197-210. Craik, G. J. S. 1981. Underwater survey of coral trout Plectropomus leopardus (Serranidae) populations in the Capricorn section of the Great Barrier Reef Marine Park. Proc. Fourth Int. Coral Reef Symp., Manila, 1:53-58. Dalzell, P. 1989. The biology of surgeon fishes (Family: Acan- thuridae), with particular emphasis on Acanthurus nigricauda and A. xanthopterus from northern Papua New Guinea. M.S. thesis, Univ. Newcastle upon Tyne, 285 p. Davis, G. E. 1989. Designated harvest refugia: the next stage of marine fishery management in California. CalCOFI Rep. 30:53-58. Gabriel, W. L., M. P. Sissenwine, and W. J. Overholtz. 1989. Analysis of spawning stock biomass per recruit: an example for Georges Bank haddock. N. Am. J. Fish. Manage. 9:383-391. Galzin, R. 1987. Potential fisheries yield of a Moorea fringing reef (French Polynesia) by the analysis of three dominant fishes. Atoll Res. Bull. No. 305, 21 p. Goeden, G. B. 1979. Biogeographic theory as a management tool. Environ. Conserv. 6:27-32. Hayes, T. A., T. F. Hourigan, S. C. Jazwinski, S. R. Johnson, J. D. Parrish, and D. J. Walsh. 1982. The coastal resources, fisheries and fishery ecol- ogy of Puako, west Hawaii. Hawaii Coop. Fish. Res. Unit, Tech. Rep. 82-1, 159 p. Ingles, J., and D. Pauly. 1984. An atlas of the growth, mortality and recruit- ment of Philippine fishes. Int. Center Liv. Aquat. Res. (ICLARM), Manila, Philippines, 13 p. Johannes, R. E. 1978. Traditional marine conservation methods in Oceania and their demise. Ann. Rev. Ecol. Syst. 9:349-364. Katnik, S. 1982. The effects of fishing pressures on some economi- cally important fishes on Guam's reef flats. Proc. Fourth Int. Coral Reef Symp., Manila, 1:111. Koslow, J. A., R. Hanley, and R. Wicklund. 1988. Effects of fishing on coral reef fish communities at Pedro Bank and Port Royal Cays, Jamaica. Mar. Ecol. Prog. Ser. 43:201-212. Longhurst, A. R., and D. Pauly. 1987. Ecology of tropical oceans. Acad. Press, NY, 407 p. McCormick, M. I., and J. H. Choat. 1987. Estimating total abundance of a large temper- ate reef fish using visual strip-transects. Mar. Biol. 96:469-478. DeMartini: Potential of fishery reserves for managing Pacific coral reef fishes 427 Munro, J. L. 1983. Biological and ecological characteristics of Car- ibbean reef fishes. In J.L. Munro (ed.), Caribbean coral reef fishery resources, p. 223-231. ICLARM Studies and Reviews, 7. Int. Cent. Living Aquat. Res. Manage., Manila. Munro, J. L., and D. McB. Williams. 1985. Assessment and management of coral reef fisheries: biological, environmental and socio-economic aspects. Proc. Fifth Int. Coral Reef Congr., Tahiti 4:545-578. Pauly, D. 1980. On the interrelationships between natural mor- tality, growth parameters, and mean environmental temperature in 175 fish stocks. J. Cons. Int. Explor. Mer 39:175-192. Pauly, D., and J. Ingles. 1981. Aspects of the growth and natural mortality of exploited coral reef fishes. Proc. Fifth Int. Coral Reef Symp., Manila 1:89-98. Plan Development Team. 1990. The potential of marine fishery reserves for reef fish management in the U.S. southern Atlan- tic. NOAA Tech. Memo. NMFS-SEFC-261, 40 p. Polaeheck, T. 1990. Year around closed areas as a management tool. Nat. Res. Model. 4:327-354. Roberts, C. M., and N. V. C. Polunin. 1991. Are marine reserves effective in management of reef fisheries? Rev. Fish Biol. Fish. 1:65-91. Russ, G. R. 1985. Effects of protective management on coral reef fishes in the central Philippines. Proc. Fifth Int. Coral Reef Congr., Tahiti 4:219-224. 1989. Distribution and abundance of coral reef fishes in the Sumilon Island reserve, central Philippines, after nine years of protection from fishing. Asian Mar. Biol. 6:59-71. 1991. Coral reef fisheries: effects and yields. /;; PF. Sale led.). The ecology of fishes on coral reefs, p. 601- 635. Acad. Press, NY. Russ, G. R., and J. St. John. 1988. Diets, growth rates and secondary production of herbivorous coral reef fishes. Proc. Sixth Int. Coral Reef Symp., Australia 2:37-43. Russ, G. R., and A. C. Alcala. 1989. Effects of intense fishing pressure on an assem- blage of coral reef fishes. Mar. Ecol. Prog. Ser. 56:13- 27. Russ, G. R., A. C. Alcala, and A. S. Cabanban. In press. Marine reserves and fisheries management on coral reefs with preliminary modelling of the ef- fects of yield per recruit. Proc. Seventh Int. Coral Reef Symp., Guam. Schonewald-Cox, C. M., and J. W. Bayless. 1986. The boundary model: a geographic analysis of design and conservation of nature reserves. Biol. Conserv. 38:305-322. Simberloff, D. 1988. The contribution of population and community biology to conservation science. Annu. Rev. Ecol. Syst. 19:473-511. Sparre, P., E. Ursin, and S. C. Venema. 1989. Introduction to tropical fish stock assessment, part 1 - manual. FAO Fish. Tech. Pap. 306/1. Rome, 337 p. Stamps, J. A., M. Buechner, and V. V. Krishnan. 1987. The effects of edge permeability and habitat ge- ometry on emigration from patches of habitat. Am. Nat. 129:533-552. Sudekum, A. E., J. D. Parrish, R. L. Radtke, and S. Ralston. 1991. Life history and ecology of large jacks in undis- turbed, shallow, oceanic communities. Fish. Bull. 89:493-513. Walsh, W. J., III. 1984. Aspects of nocturnal shelter, habitat space, and juvenile recruitment in Hawaiian coral reef fishes. Ph.D. diss., Univ. Hawaii, Honolulu, 475 p. Wellington, G. M., and B. C. Victor. 1988. Variation in components of reproductive success in an undersaturated population of coral-reef dam- selfish: a field perspective. Am. Nat. 131:588-601. Abstract.— Interest in estimating energy fluxes through populations of marine mammals has been increas- ing as these populations are more frequently recognized to compete with fishermen for commercially ex- ploited fish stocks. Testing for the presence of allometric trends with size in the parameters used to esti- mate these energy fluxes is impor- tant because if such trends exist and are large, it will be inappropriate to apply measurements derived from one size of animal to animals of other sizes, in any given population. To test for (and to measure, if present) allometric effects in a population of small cetaceans, morphological mea- surements (energetics parameters) related to estimating energy flux were taken from 35 spotted dolphins iStenella attenuate!) ranging in size from 77 to 210 cm total length, cap- tured incidental to fishing operations in the eastern tropical Pacific Ocean. Significant allometric (nonlinear) changes with size were observed in 23 of the 25 parameters measured; no trends were isometric (linear). Most of the significant trends were expressed primarily during the first two or three years of life (5-30 kg wet weight). Thus, parameter esti- mates for small dolphins (less than about 30 kg wet weight) should be derived from measurements on ani- mals near the specific size of inter- est. Estimates for larger animals, with the exception of several surface area measurements, could be esti- mated reasonably well from any specimens greater than about 30 kg wet. Although the energy fluxes and standing stock of energy represented by animals younger than 2-3 years are relatively small compared with the total population, constraints resulting from the energy character- istics of the smaller animals may ex- ert significant control over popula- tion energy flux, implying that the rapid changes in energetics param- eters of the younger animals should not be ignored. Allometry of energetics parameters in spotted Dolphin (Stenella attenuata) from the eastern tropical Pacific Ocean Elizabeth F. Edwards Southwest Fisheries Science Center RO.Box 27 1 , La Jolla. CA 92038 Manuscript accepted 14 May 1993. Fishery Bulletin 91:428-439 ( 1993). Estimating cetacean population en- ergy fluxes, in particular energy con- sumption in the form of commercially valuable fishes, is important because many commercially exploited fish populations are decreasing in abun- dance yet must be shared by both human and cetacean predators. The existence in cetacean populations of allometric trends (nonlinear changes with size) in morphological charac- teristics related to energy processing complicates energy flux estimation for these animals because the size range of cetacean specimens avail- able for study tends to be very nar- row. If allometric trends are present and large, measurements from that narrow size range cannot be assumed to apply equally to all other sizes in the population. Neither commercial nor noncom- mercial sources of cetacean speci- mens generally provide an adequate range of sizes from which to deter- mine whether allometric trends ex- ist. Specimens available from com- mercial sources will include only those sizes sought by the fishery. For example, morphological samples from large cetaceans (whales) tend to be dominated by adult (and when preg- nant females are captured, by fetal) specimens because the larger indi- viduals have more commercial value. Juveniles are rarely captured. Mor- phological samples from small ceta- ceans (dolphins and porpoises) are rare for all sizes because these ani- mals are rarely the subject of directed fisheries. Specimens from noncom- mercial sources are generally avail- able only as beached or stranded individuals, or where observer pro- grams are employed to monitor kill rates of both target and nontarget species. These noncommercial speci- mens tend to be too rare to provide an adequate sample of sizes. In addition to these sampling bi- ases in age and species composition, the data collected are generally un- suitable for deriving estimates of en- ergy flux. Although considerable mor- phological data exist for commercially exploited large whales (e.g., Fujino 1954, 1956; Omura and Fujino, 1954; Omura and Sakiura, 1956; Ohsumi, 1960; Nishiwaki et al., 1963; Lockyer, 1981, a and b, and references there- in) and a few similar reports exist for small cetaceans (e.g., Sergeant, 1962;, Perrin, 1975; Yasui, 1980; Perrin et al., 1987), these data gen- erally include only a standardized set of external morphological measure- ments used primarily for taxonomic purposes. Taxonomic data tend to be inappropriate for deriving estimates of energy flux because many size classes and most of the morphologi- cal measurements required specifi- cally for energetics estimates are missing. Although some estimates have been developed for energy process- ing in neonate through adult large whales (Lockyer, 1981, a and b), al- lometric trends in energetics param- eters have never been examined in 428 Edwards. Allometry of energetics parameters in Stenella attenuata 429 small cetaceans. Existing studies of small cetaceans include only Yasui and Gaskin's (1986) energy budget for adult harbor porpoise (Phocoena), in addition to a few estimates for various aspects of energy processing by juvenile (e.g., Anderson, 1981) or adult (e.g., Lang, 1966; Lockyer, 1981, a and b; Yasui and Gaskin, 1986; Hui, 1987; Bose and Lien, 1989; Bose et al, 1990) small cetaceans. Incidental mortality of spotted dolphins (Stenella attenuata), a small cetacean that is inadvertently killed during some tuna purse-seining operations in the east- ern tropical Pacific Ocean (ETP), presents an excep- tion to this problem of specimen unavailability. Speci- mens of all sizes die in purse-seines (Barlow and Hohn, 1984; Hohn and Hammond, 1985), making it possible to collect appropriate data for the entire size range occurring in the natural population. Taking advantage of this availability of specimens, spotted dolphins ranging in size from fetal through mature adult were collected from the fishery and ana- lyzed for morphological characteristics related to en- ergy processing. I report here regression equations re- lating changes in 25 energetics parameters to changes in total wet weight for these specimens. The param- eters can be used in estimation of three components contributing to cetacean energy budgets: active me- tabolism, passive heat loss, and energy density. Esti- mated energy density is derived from a subset of the measured parameters and is reported as the 25th pa- rameter. Total energy budgets, active metabolism, and passive heat loss are not estimated here because such estimates depend on various assumptions about other factors (e.g., environmental conditions, activity levels) and other energy budget components (e.g., standard metabolism, heat of digestion) not included in this study. The work reported here was conducted to determine the extent to which allometric trends, if any, could be determined from simple morphological measure- ments on deceased cetaceans. the eastern tropical Pacific Ocean (spotted dolphins are about 80cm total length at birth (Hohn and Hammond, 1985)). Specimens included 22 females and 13 males (Table 1). Males specimens included 3 fe- tuses, 6 immature, and 4 mature. Females specimens included 3 fetuses, 6 immature, 1 mature resting, 1 mature lactating, and 11 mature pregnant. Not all mea- surements were made on all specimens. All specimens were collected incidental to tuna fishing operations in the eastern tropical Pacific Ocean. Seventeen of the specimens were collected on 28 December 1985; nine were collected during July and two during August 1985 (Table 1). Specimens were kept frozen after collection and were thawed in fresh water (about 27° C) just prior to sampling for energetics parameters. Parameter measurements Active metabolism Active metabolism in cetaceans is the total energy cost of swimming, i.e., energy (me- chanical plus waste heat production heat) expended to overcome hydrodynamic drag. Cost of steady submerged swimming by cetaceans can be estimated by using Magnuson's (1978) procedure for estimating cost of swimming by tuna (see also Webb, 1975), including drag on both body and fins. Eleven morphological mea- surements are required to estimate hydrodynamic drag: maximum diameter of the body CDmaJ; wetted surface areas of the body (WSAh), flippers iWSAfp), dorsal fin (WSAj), and flukes iWSA,,); mid-chord depths of flippers (MCDfp), dorsal fin (MCDlt), and flukes (MCDfl); and characteristic length of flippers (CLlp), dorsal fin (CLd), and flukes (CL„) (Table 2). Z)max was derived from measurements of body cir- cumference immediately anterior to the dorsal fin (C3; Methods and materials Specimen collection Measurements were taken from 35 spotted dolphins ranging in size from 71 to 210 cm total length (TL [tip of rostrum to fluke notch] Fig. 1). This repre- sents the entire size range ( near- term fetus through mature adult) of the spotted dolphins found in C2 X I C3/ ( ■ — 7^ ^^ CA c 1 J' "' \ J} II I ' ==£5 A (TX. \ <- -il CjJL Locations o (metabolic CI: at eye peduncle. Figure 1 f the 5 circumference measurements on dolphin specimens. surface area) extends from eye to mid-caudal peduncle (sensu, C2: at axilla, C3: just anterior to dorsal, C4: at anus, C5: Thermal core Brodie, 1975). at mid-caudal 430 Fishery Bulletin 91(3), 1993 Table 1 Collection data for dolphin specimens. wet total Specimen Spec. weight length Reprod.1 Date Location Captured om„ BD„ no. Code (gm) (cm) Sex condition captured (lat.) (long.) (cm) (cm) 1 BXR252 27013 132 F I 12/28/85 17°46'N 111°37'W 71.8 0.75 2 BXR253 68100 194 F M-R 12/28/85 17°46'N 111°37"W 95.0 0.97 3 BXR254 73775 192 F M-P 12/28/85 17°46'N 111°37'W 103.2 0.86 4 BXR254F 4983 78 M F 12/28/85 17°46'N 111°37'W 35.4 0.45 5 BXR255 64540 186 F M-P 12/28/85 17°46'N 111°37"W 107.5 1.08 6 BXR255F 5800 78 F F 12/28/85 17°46'N 111°37'W 40.8 0.68 7 BXR264 62425 183 F M-P 12/28/85 18°09'W 111°17'W 97.8 0.82 8 BXR264F 3760 71 F F 12/28/85 18°09'W lll017'W 34.0 0.51 9 BXR271 24516 136 M I 12/28/85 18°09'W 111°17'W 66.9 0.83 10 BXR280 19976 127 M I 12/28/85 18°09'W 111°17'W 61.0 0.68 11 BXR282 76272 193 F M-P 12/28/85 18'09'W 111°17'W 103.0 1.00 12 BXR295F 4200 74 F F 12/28/85 18°09'W 111°17'W 34.6 0.52 13 BXR306 37228 163 F I 12/28/85 18°09'W 111°17'W 75.0 0.90 14 BXR312 35639 152 F I 12/28/85 18°09'W 111°17'W 75.0 0.88 15 BXR313 63787 175 F M-P 12/28/85 18°09'W 111°17'W 97.6 0.92 16 BXRIA284F 3410 71 M F 12/28/85 18°09W 111°17'W 34.8 0.45 17 BXR(A)284 69689 191 F M-P 12/28/85 18°09'W 111°17'W 100.4 NR 18 SRM044(F) 5400 81 M F 2/14/80 06°02S 85°46'W 40.0 0.73 19 WFP660 16400 114 F I NR NR NR 55.0 NR 20 WFP680 30900 136 M I 7/7/83 NR NR 74.0 NR 21 SRM044 73100 189 F M-P 2/14/80 06°02'S 85°46'W 110.0 0.97 22 WFP700 48124 170 M M NR NR NR 84.6 0.61 23 PEL307 42903 167 F I 7/10/85 8°55'N 129°58W 80.0 0.70 24 PEL308 41314 151 M I 7/10/85 8°55'N 129°58'W 81.0 0.72 25 PEL309 74002 189 F M-P 7/13/85 8°13'N 129°40'W 99.0 0.95 26 PEL310 71732 202 F M-P 7/13/85 8°13'N 129°40'W 97.5 0.85 27 PEL311 55388 176 F M-P 7/14/85 10°21'N 129°35W 91.0 NR 28 PEL313 72640 202 M M 7/14/85 10°21'N 129°35'W 94.2 0.66 29 PEL314 44265 160 M I 7/14/85 10°21'N 129°35W 85.4 0.69 30 PEL315 69008 188 M M 7/14/85 10°21'N 129°35'W 95.5 0.82 31 PEL316 58566 175 F M-P 7/14/85 10°21'N 129°35"W 94.0 0.87 32 SD1 11350 100 M I NR NR NR 53.5 0.71 33 SD2 14301 113 F I NR NR NR 56.0 0.61 34 STB122 84444 206 M M 8/30/85 10°14'N 125°03'W 98.0 0.86 35 STB 126 74202 188 F M-L 8/30/85 10°14'N 125°03'W 102.0 0.64 'F = fetus; I = immature; M = mature; M-R = mature resting M-P = mature pregnant; M-L = mature lactating; NR = no record. Fig. 1). WSAb was measured as the sum of surface areas of right circular cones. Conic surfaces were cal- culated from measurements of body circumference at five locations and distances from tip of rostrum to each circumference and to fluke notch (Fig. 1). WSAlp, WSA,, and WSA,, were estimated by tracing the perimeter of each fin onto white paper, measuring with a digitizer the surface enclosed, correcting for curvature, and multiplying by 2. Both flukes were traced as a unit. Each dorsal fin and flipper was traced separately. The correction for curvature along fin surfaces was derived from one small (132 cm TL) and one large (193cm TL) dolphin. Flippers, flukes, and dorsal fin from each specimen were sliced laterally into 4 or 5 sections. The cross-section of each piece was then xeroxed onto white paper. Distances straight across and around the perimeter of each cross-section were then measured twice for each cross-sectioned piece of fin. There were no significant differences between fins or sizes of dolphin in the ratio of curved to flat mea- surements so a single correction was used for all fins. The curvature correction was an increase of 6% over the estimated flat area of each fin (3% per side; SE = 0.4; n = 23 sections). MCDlp, MCD,h and MCDP were measured to the near- est millimeter by using calipers at the thickest part of each fin at the mid-point of the characteristic length. Edwards: Allometry of energetics parameters in Stenella attenuate 431 Table 2 Symbols and definitions. ACT active metabolism (calories active metabolism calorie animal ' -hr~') BD,, average blubber depth (cm) cf coefficient of friction drag CT coefficient of total drag CDd caloric density of dolphin (calg wet weight1) CLh, characteristic length of the flippers (cm ) CLb characteristic length of the dorsal fin (cm) CLf characteristic length of the flukes (cm) D , maximum diameter of the body (cm ) DT total drag(dyn) EDan energy density of total animal (mJ/kg wet weight 1 EDU energy density of blubber (mJ/kg wet weight) ED„, energy density of muscle (mJ/kg wet weight) EDb energy density of bone ImJ/kg wet weight) Em fraction of body wet mass due to muscle F„, fraction of body wet mass due to blubber F„ fraction of body wet mass due to bone F, fraction of body wet mass due to viscera E, fraction of body wet mass due to fins FID fin-induced drag (fraction of total drag) Hu minimum unavoidable heat loss (calories lost as heatcalorie animals-day1) MCDfp mid-chord depth of the flippers (cm) MCD„ mid-chord depth of the dorsal fin (cm) MCD„ mid-chord depth of the flukes (cm) ME mechanical efficiency MP mechanical power (erg/s) MSAh metabolic surface area of the body (cm2) N density of sea water (g/cm3) PE propeller efficiency Ri Reynolds number based on length L (dyn/cm2) TL total length (cm) V kinematic viscosity (stoke) VL velocity (cm. sec') WSA„ wetted surface area of the body (cm2) WSAfl, wetted surface area of the flippers (cm-'i WSAcl wetted surface area of the dorsal fin (cnvi WSA„ wetted surface area of the flukes (cm2 1 WWK total body wet weight (g) VBfiu water content of blubber (percent of wet weight) %H„0m water contents of muscle ( percent of wet weight ) %H20b water contents of bone (percent of wet weight) CLlp, CLd, and CLt, were estimated from the traced figures, by measuring the length of the fin parallel to the main axis of the body. Unavoidable passive heat loss [HLU] Unavoidable passive heat loss was denned as heat loss due to con- duction through blubber, if one assumes no heat loss from appendages or from the head anterior to the eyes (e.g., Brodie, 1975). Two morphological measurements contribute to this estimate of HLU; metabolic surface area of the body (MSAb) and average blubber depth (BD,). MSAh was estimated by the same method as wetted surface area, except that the first anterior and last posterior sections of the body (Fig. 1) were omit- ted from the assumed thermal core and the conic ra- dius used in the estimate was the radius of the body core beneath the blubber. This metabolic radius was estimated by determining the radius to the outer sur- face at each circumference and by subtracting the av- erage blubber depth measured at that circumference. Two or three measurements of blubber depth were made along each circumference at the dorsal midline, ventral midline, and mid-way between these two lines. BDa for the entire animal was estimated as the weighted sum of average blubber depths at each cir- cumference. Weightings were the circumferences them- selves, which gives more weight to the relatively sym- metrical mid-body areas that comprise the majority of the insulating area, and less weight to the thick aver- age blubber depths related to the hydrodynamic keel in the tail region. Energy density Eleven morphological measurements contribute to estimating the overall energy density of an individual spotted dolphin (EDan): fractions of wet mass due to muscle (F,„), blubber (Fbl), bone (Fh), vis- cera (F„), and fins CF}); energy densities of blubber (EDhl), muscle (ED,„), and bone (EDbn); and water con- tents of blubber (%H20W), muscle (%H20J, and bone (%H206n). Energy content of body fluids (blood, inter- stitial fluids) were ignored. Fluid losses accounted for about 10% of the difference between total weights of undissected specimens and the sum of dissected body fractions. Fm, Fbl, Fh„, Fv and Ff were determined by measur- ing the total wet weight of each specimen, and by dissecting the specimen into component parts and weighing each component. Skeletal weight was deter- mined after carefully flensing and scraping all tissue from each bone, including tissue between ribs, be- tween spinal column processes, and within the skull and jaw structures. EDhh EDm, and EDh„ were deter- mined by bomb calorimetry (Cummins and Wuycheck, 1971). Data used in regressions are means of two or four replicate energy density determinations. Dry weights were determined after freeze-drying samples to constant weight (48 hours) and storing in a desic- cator for 24 hours. Ash-free dry weights were deter- mined after ashing samples at 450° C for 4 hours and cooling for 24 hours in a desiccator. Dry and ash-free dry weight determinations were made on samples of 1-20 g wet weight. %H20W, %H,0„, and calculated as fHA,, were ( 1.0 - (dry weight/wet weight)) * 100. Energy density (ED„n) of entire dolphins was esti- mated as the weighted sum of predicted energy densi- 432 Fishery Bulletin 91(3). 1993 ties of blubber, muscle, viscera, bone, and fins, where weightings were the predicted fractions of total wet body mass for each tissue type. Energy density of vis- cera was assumed to be the same as energy density of muscle, based on the relatively muscular nature of these organs (compared with bone or blubber) and the absence of any visible fat deposits upon or within or- gans. Errors were assumed negligible because the vis- cera comprised a relatively small fraction of total body weight and it appeared unlikely that the viscera rep- resented a major lipid depot in these tropical ceta- ceans. Energy density of fins was assumed to be the same as energy density of bone, based on the assump- tion that caloric content of cartilage (the actual com- position of fin material) is closer to bone than to either muscle or blubber. Again, errors due to this assump- tion should be relatively small as the fins contribute very little to overall body weight. Estimated total en- ergy density is presented as a function of the regres- sion-predicted values of each parameter, where param- eter variances are ignored. Data analysis Relationships between morphological measurements and total wet weight were determined by using linear regression analysis of log transformed variables, where morphological measurements and wet weights were converted to logarithms (base 10) prior to analysis. Results Significant allometric trends (P < 0.05) occurred in 23 of the 25 parameters (Table 3; Figs. 2-11). No signifi- Table 3 Regression equation parameters relating energetics parameters to total wet weight of spotted dolphins tStenella attenuata) from the eastern tropical Pacific Ocean Regressions Derformed on log 0-transformed data in both X and Y. Re-transformed equation of the form Y = aXh. X units are wet weight in kg. y units are as noted. Symbols are defined in Table 1 n1 b2 se(b)3 a' Ps Active metabolism A™, (cm) 35 0.294 0.0133 3.17 0.0001 WSA,, (cm2) body 35 0.682 0.006 5.65 0.0001 WSA,„ (cnr) flippers 26 0.488 0.022 1.13 0.0001 WSA,, (cm-) dorsal 26 0.543 0.018 1.84 0.0001 WSA„ (cm2) flukes 26 0.598 0.200 0.46 0.0001 MCDr„ (cm) flippers 26 0.211 0.015 0.14 0.0001 MCDd (cm) dorsal 26 0.299 0.145 0.06 0.0001 MCD„ (cm) flukes 26 0.224 0.021 0.14 0.0001 CLfp (cm) flippers 26 0.209 0.030 2.08 0.0001 CLd (cm) dorsal 26 0.285 0.015 1.00 0.0001 CL„ (cm) flukes 26 0.194 0.019 2.00 0.0001 Passive heat loss (HLJ MSAk (cm2) body 35 0.729 0.008 2.76 0.0001 BD„ (cm) body 29 0.161 0.026 0.14 0.0001 Energy density Fractions of body wet weight Fu Blubber 24 -0.247 0.021 0.32 0.0001 K Muscle 24 0.159 0.012 0.32 0.0001 F„ Viscera 24 -0.153 0.038 0.24 0.0005 n„ Bone 24 -0.141 0.022 0.12 0.0001 F, Fins 22 -0.239 0.034 0.05 0.0001 Energy densities ImJ/kg dry weight) EDH Blubber 11 0.021 0.007 33.27 0.0178 EDm Muscle 9 -0.041 0.013 25.00 0.0132 EDbn Bone 5 -0.059 0.050 9.12 0.3248 Water content (% of sample wet weight) %H.A, Blubber 11 -0.289 0.085 316.23 0.0067 %H,0,„ Muscle 10 -0.004 0.007 75.68 0.5956 %H20in bone 5 -0.191 0.023 331.13 0.0141 Estimated Energy Density EDa„ (mJ/kg wet weight) 35 -0.063 0.000 10.21 0.0001 'Sample size. ^Estimated slope coefficient. 'Estimated standard error of slope coefficient. 'Estimated intercept for fitted regression. Significance level (probability ). Edwards: Allometry of energetics parameters in Stenella attenuata 433 140 - AMETER (cm) 8 § MAXIMUM BODY D o o o m* — ■■ 20 - 0 20 40 60 80 100 WET WEIGHT (kg) Figure 2 Relationship between maximum body diameter (just anterior to dorsal fin; C3 in Fig. 1) and total wet weight of body in kilo- grams for 35 specimens of spotted dolphin tStenella attenuate! from the eastern tropical Pacific Ocean, both sexes and all ages (sizes) represented. Solid line is fitted regression. METABOLIC 0 20 40 60 80 100 WET WEIGHT (kg) Figure 3 Relationship between wetted and metabolic surface areas, and total wet weight of body in kilograms for 35 specimens of spotted dolphin (.Stenella attenuata) from the eastern tropical Pacific Ocean, both sexes and all ages (sizes) represented. Lines through points are fitted regressions. WSAh, WSAfp, WSAd, WSA/i, Figs. 3 and 4). Weak trends were found for energy densities of blubber and muscle (EDhh EDm; Fig. 9). Intermediate trends represented the majority of the relationships ( 16 of 25) and included Dmas (Fig. 2), all measures of MCD (Fig. 5), CL (Fig. 6), BD„ (Fig. 7), F (Fig. 8, A and B), %H20„ and %H206„ (Fig. 10), and EDan (Fig. 11). Parameters that exhibited strong trends dem- onstrated large changes in parameter values throughout the weight range studied. Although the allometric effect decreased somewhat with increas- ing size, parameter estimates were 300-400% larger for 30-kg than for 5-kg dolphins (compared with a difference of 600% (30 kg/5 kg* 100) in wet weight), and 50%-85% larger for 70-kg than for 30-kg dolphins (compared with a difference of 233% in wet weight; Table 4). Parameters showing intermediate trends changed rapidly only from birth through the first 2-3 years of life (5-30 kg; Perrin et al., 1976; Hohn and Hammond, 1985). Parameter values changed much more slowly with size in older (larger) dol- phins (30-70 kg) than in smaller dolphins (e.g., less than 30 kg). For example, estimated muscle fraction of body wet mass increased by 33% from 5-kg to 30-kg dolphins, but only by 14% from 30 to 70kg wet weight (Table 4). Parameter estimates for these intermediate effects differed from 10% to 70% between 5- and 30-kg dolphins, but only from 5% to 30% between 30- versus 70-kg dolphins (Table 4). Parameter values showing weak trends changed little with increasing size. For example, estimated ED„, increased only 8% in animals 5 kg to 30 kg in weight, and only an additional 4% from animals 30 kg to 70 kg in weight (Table 4). The lack of any significant trend in EDh„ may be due to small sample size (n = 5 animals) but the relatively small scatter of the existing points indicates that the apparent absence of trend is probably real (Fig. 9). The lack of trend in H20„, also appears real, as sample size was reasonably large (n = 10) and scatter about the regression line relatively small (Fig. 10). cant trend was found for %H20„, (Fig, 10) or EDhn (Fig. 9). No trends were isometric (linear) in the untransformed variables. Trends fell into three gen- eral groups based on the strength of the relationship between parameter value and wet weight (Table 4). Expressed in terms of the regression coefficient (6), these relationship groups were strong (0.49 < b < 0.73), intermediate (-0.29 < b < -0.14 or 0.16 < b < 0.30), and weak (-0.14 < b < 0.16). Strong trends were found for all five of the surface area measurements (MSAh, Discussion Estimates for individual dolphins The practical importance of any allometric trend in any particular parameter will depend not only on the strength of the trend, but also on that parameter's relative contribution to the energy flux being estimated. This contribution is affected by the algebraic place- ment of the parameter in the energy flux calculation, 434 Fishery Bulletin 91(3). 1993 1.1 - 1.0 - 0.9 - X X X 0.8 - X **»** 0 7 - FLUKES X ^***^ X ■""*^ X 0.6 - x X x 0.5 - X DORSAL 0.4 - 0.3 - 0.2- 0 1 - ** •3 X ■ *o„. - — -*•* — ■■ ■ ■ FLIPPERS o u.u -i 1 ■ i 1 i ■ i 1 WET WEIGHT (kg) Figure 4 Relationship between wetted surface areas of fins and total wet weight of body in kilograms for 26 specimens of spotted dol- phin tStenella attenuata) from the eastern tropical Pacific Ocean, both sexes and all ages (sizes) represented. Areas are total for both sides of dorsal, both sides of both flippers, and both sides of both flukes. Lines through points are fitted regressions. ACT = 20650 * MP/(ME*PE) 2.4 - 2.2 - 2.0 - X o FLUKES X X o o 1 6 - x K X x BL.-y ■ X 1 4 - .-■■ ■ 1.0 - X *£< ^r DORSAL FLIPPERS 0.8 - 0.6 - 0.4 - 02 - 0.0 ~l ■ 0 20 40 60 80 100 WET WEIGHT (kg) Figure 5 Relationship between mid-chord depths of fins and total wet weight of body in kilograms for 26 specimens of spotted dolphin {Stenella attenuata ) from the eastern tropical Pacific Ocean, both sexes and all ages (sizes) represented. Lines through points are fitted regressions. by the actual values taken on by the parameter, and by the relative importance of the energy characteristic in the overall energy budget. For example, the relative importance of the param- eters WSA,, and Dmta in the equation for active me- tabolism is highly dependent on both parameter place- ment and relationship with wet weight (trend). Active metabolism (calhr ') can be estimated for steady sub- merged swimming by torpedo-shaped bodies as where ME = 0.20 PE = 0.85 MP = ((Zyie7) * VL) DT = (0.5 * N *VL2*WSAb*CT) (1.0 -FID) N = (1.025 g/cm3) FID = 0.21 \^"p = C/*[l + <1.5*((Dmax)/TL))3/2 (7*((Z)max)/TL))3] cf = 0.072 RL'yB R, = (TL*VL)/v V = 0.01 Stokes VL is velocity (cm. sec1) and TL is total length in centimeters (Edwards, 1992, following Webb, 1975). Collecting and assuming constant all terms ex- cept the energetics parameters WSAh and Dmax, the expression for active metabolism can be sim- plified to ACT = C * WSAh * 11 + ( 1.5*((Dn (7*((Dmax)/TL)):1]. ,)/TL))3/2 + where C represents the collected terms. The effects of increasing or decreasing the values used for WSAh and DmM can be seen more readily in this formulation. Changes in the value of WSAh lead directly to equivalent changes in the estimate of ACT (e.g., increas- ing WSAh by 50% will increase the estimated cost of activity by 50% ). WSAh not only has a direct effect on estimates of activity cost, but has also a strong relationship with wet weight, leading to differences of up to 80% in estimates of activity costs for 30-kg versus 70-kg spotted dolphins (WSAh for a 70-kg dolphin is 78% larger than for a 30-kg dol- phin. Table 4). Differences are near 350% for 5-kg versus 30-kg dolphins (WSAh for a 30- kg dolphin is 339% greater than for a 5-kg dolphin, Table 4). Conversely, changes in D,„.1X have little ac- tual effect on activity estimates, both because changes with size are smaller overall and because of the parameter's algebraic placement in the equa- tion for activity costs. The ratio Dmax/TL will al- ways be quite small (e.g., 5/80 = 0.065 in an 80-cm dolphin (Dmas = 5 cm) weighing 5 kg; 11/210 = 0.052 in a 210-cm dolphin (D„iav = 11cm) weighing 70kg). Because this small term is made even smaller by raising it to higher powers, changes in Dmax will have little effect on estimates of ACT (e.g., dou- bling Dmm from 5 to 10cm for an 80-cm dolphin Edwards: Allometry of energetics parameters in Stenella attenuata 435 changes the value of the term in brackets only from 1.03 to 1.08). This lack of effect occurs de- spite the fact that Dmax has an intermediate-level relationship with wet weight (Table 4). Examining the placement of MSAb and BD„ in the equation for passive heat loss also illustrates that their relative importance is highly dependent on both parameter placement and relationship with wet weight. Passive heat loss can be estimated as (Edwards, 1992, following Brodie, 1975). H„ (21.18/BDJ * (37.0-Tj * MSAf/10000.0) * 24 mvg*(czy iooo.O) Collecting and assuming constant all terms ex- cept the energetics parameters BDn and MSAh, Hu = MSAJBDJ * C, where C represents the collected terms. Once again, changes in the surface area mea- surement (MSAh) will lead directly to equivalent changes in estimates of HL,„ and the strong rela- tionship between MSAh and wet weight will be readily reflected in the estimates. However, unlike the case for Dmax, changes in values used for BDa will have a significant (reciprocal) effect on esti- mates of HLU (e.g., increasing blubber depth by 30% [i.e., by a factor of 1.30] will decrease the estimate for Hu by about 23% [1/1.30 = 0.77]) ow- ing to the difference in the way the parameter is expressed in the equation. In this case, the inter- mediate-level relationship of BDa with wet weight will be reflected in the estimate of energy flux. Allometric effects in parameters related to bone provide an example of largely irrelevant, though statistically significant, trends. The bone fraction of body weight is small (about 10%: Fig. 8A), and its energy density low (less than 1/4 the energy den- sity of blubber, 1/3 the energy density of muscle). Given no indication from the energy densities in the small sample studied that spotted dolphins store lipid within the skeleton, even relatively large changes with size will have little effect on whole- animal energy calculations. Even though the frac- tion of body weight due to bone has an intermediate- level relationship with wet weight, decreasing 30% from 5-kg to 30-kg dolphins, the decrease in terms of total body weight is only from 10% to 7% (Table 4). Thus, the practical importance of any allometric trend depends on other factors in addition to actual strength of the trend with wet weight, but in general for spotted dolphins, these trends are expressed pri- marily during the first two or three years of life (5-30 kg wet weight). Parameter estimates for small DORSAL 26- o • ».v 24 - ■ ■ 0 ■ ..-o-'l*'.*0 *•» 22 - ?n - FLIPPERS **■ © ■ X 18 - 16 - 14 - ■ ■r .■<■■* •'o.« ^^_- X XX x X - 7T^ ■< X 12 - sit x X FLUKES 11) - * 8 - 6 - 4 - 2 - 0 - — i 1 r i 1 r ■i r i 0 20 40 60 80 100 WET WEIGHT (kg) Figure 6 Relationship between characteristic lengths of fins and total wet weight of body in kilograms for 26 specimens of spotted dol- phin [Stenella attenuata) from the eastern tropical Pacific Ocean, both sexes and all ages (sizes) represented. Lines through points are fitted regressions. 40 60 WET WEIGHT (kg) Figure 7 Relationship between average blubber depth and total wet weight of body in kilograms for 29 specimens of spotted dolphin (Stenella attenuata) from the eastern tropical Pacific Ocean, both sexes and all ages (sizes) represented. Lines through points are fitted regressions. dolphins (less than about 30kg wet weight) should be derived from measurements on animals near the spe- cific size of interest. Estimates for larger animals, with the exception of the surface area measurements, could be estimated reasonably well from any specimens greater than about 30 kg wet weight. Population energetics The examples presented above pertain to estimates for individual spotted dolphins. Implications of alio- 436 Fishery Bulletin 91(3). 1993 1- 0-6 I a LU 5 0.5 o < -V * ■ •■••'••■-^ . / ■ **"r- * r „k" X -#*- 40 60 WET WEIGHT (kg) B WET WEIGHT (kg) Figure 8 (A) Relationship between fractions of total body wet weight com- prising muscle, blubber, and bone, and total wet weight of body in kilograms for 24 specimens of spotted dolphin {Stenella attenuata) from the eastern tropical Pacific Ocean, both sexes and all ages (sizes) represented. Lines through points are fitted regres- sions. (B) Relationship between fractions of total body wet weight comprising viscera and fins, and total wet weight of body in kilograms for spotted dolphins (Stenella attenuata) from the east- ern tropical Pacific Ocean, both sexes and all ages (sizes) repre- sented. Lines through points are fitted regressions. metric trends for estimates of population energy flux are not necessarily straightforward. Dolphins less than about 3 years old represent only about 10% of the population in these relatively long-lived and slowly re- producing mammals; dolphins age 3-12 compose about 40% and sexually mature adults about 50% of the population (Barlow and Hohn, 1984; Hohn and Hammond, 1985). Because they are so few and be- cause their biomass is so small, ignoring allometric effects when estimating energy fluxes for the younger animals, or ignoring the younger animals altogether, will have little direct effect on population energy budgets. However, the indirect effects may be consider- able. To the extent that activities of older animals are constrained by energy-related characteristics of young dolphins, indirect effects of size-related differences in estimated energy fluxes may be more important than the absolute fluxes themselves, through regulating behavior or ecological relation- ships. For example, muscle fraction of body mass in neonate dolphins (5kg) is about 30% less than the muscle fraction in adults (70 kg; Figure 8A, Table 4). Thus the power available for swimming (as a function of muscle mass) is 30%> less than would be estimated based on measurements from adult animals. The estimated speeds that small dolphins, with their smaller muscle mass, can maintain will be slower than speeds estimated by simply applying to neonates parameter values de- rived from adults. Differences in estimated swimming speed may be important because dolphins are schooling mammals with apparently strong social ties and prolonged periods of parental care for nursing offspring. In order to remain within the same school, the average speed for all individuals will be constrained to the slower speeds that can be maintained by smaller dolphins. Estimates from tagging studies of spotted dolphins in the ETP indicate that in fact the average cruising speed of dolphin schools is the optimum for neonates rather than for adults, despite the fact that adults represent the majority of individuals in the schools (Edwards, 1992, and references therein). Using parameters appropriate for adult dolphins in estimates for the smaller animals would produce unreasonably high estimates of sustainable power output (and therefore food con- sumption) by the smaller animals (Edwards, 1992). School speeds, and therefore energy re- quirements to maintain these speeds, may thus be constrained by energetics characteristics of small animals that cannot be extrapolated sim- ply from measurements on adults. As cost of transport is a large fraction of the total energy re- quirements of a swimming mammal (active metabo- lism in swimming homeotherms [including dolphins, sea otters, and penguins] is generally 2 to 3 times that of resting metabolic rates [Hui, 1987]), main- taining reduced speeds should reduce energy costs and therefore forage requirements for the school mov- ing as a unit. This is particularly significant because forage requirements are the most commonly esti- mated energy flux used to estimate the impact of a Edwards: Allometry of energetics parameters in Stenella attenuate 437 ■ ■ \ BLUBBER ■ o O o ^ MUSCLE * BONE / X 1 — X 1 — r X ■ 1 WET WEIGHT (kg) Figure 9 Relationship between energy density (per gram dry weight) of blubber (11 specimens), muscle (9 specimens) and bone (5 speci- mens), and total wet weight of body in kilograms for spotted dol- phins {Stenella attenuata) from the eastern tropical Pacific Ocean, both sexes and all ages (sizes) represented. Lines through points are fitted regressions. Table 4 Predicted values and fractional differences (ratios) between those values for energetics parameters showing strong, intermediate, and weak allometric effects, for spotted dolphins weighing 5, 30, and 70 kg wet weight. Regression coefficient indicated by (b). Predicted va ues Ratios 5 kg 30 kg 70 kg 30 kg/5 kg (6.0) 70 kg/30 kg (2.3) Strong relationship (6>0.49) MSAh 8.92 32.94 61.09 3.69 1.85 WSA„ 16.93 57.47 102.42 3.39 1.78 WSAfc 2.48 5.94 8.98 2.40 1.51 WSAtl 4.41 11.67 18.48 2.65 1.58 WSA„ 1.20 3.52 5.84 2.92 1.66 Intermediate relationship (-0.14'"' __—WEP E £140i ^t>^>-'"" len Jy^,/' s12°- o W' m I'fr/ •' 100 §7 80- $' ■ 0 5 , , 10 Age (yrs) Figure 2 Age-length records for harbor porpoises in Japanese waters determined by Miyazaki et al. (1987) and the authors. Plotted against average growth curves for male and female harbor porpoises from the eastern North Pacific by Stuart and Morejohn (1980) (solid lines), eastern Canada by Gaskin and Blair (1977) (broken lines 1 and the North Sea by Van Utrecht (1978) (broken and dotted lines). Solid circles indicate male and open circles female porpoises from Japan. Atlantic pattern of births in late spring or early sum- mer and to mating between early and late summer. These populations are distributed in the same latitu- dinal zone and hence the same photoperiod regime. Sea surface temperature regimes are also similar, eg., Winn6 and the Japan Meterological Agency (1986). Harbor porpoises have testes that are large in proportion to body size (Fisher and Harrison, 1970; Gaskin et al., 1984). Miyazaki et al. (1987), comparing their material to data of Fisher and Harrison (1970), noted that adult harbor porpoises from Japanese waters had smaller testes than those from the North At- lantic and had weight ranges that did not over- lap. They concluded that the difference in sam- pling periods in the two areas was not a factor (April-May off Japan v. May-September in the eastern North Atlantic, respectively). However, Gaskin et al. (1984) demonstrated a seasonal cycle of spermatogenesis, including a dramatic change in testis volume/size in adult harbor porpoises during June-August in the Bay of Fundy region. Read (1990a) concluded that the active mating period extended only from late June into July or occasionally early August. The specimens examined by Miyazaki et al. (1987) in April-May appear to have been collected be- fore adult males had begun to exhibit seasonal spermatogenesis and testis enlargement. P. phocoena appear to be a relatively asocial odontocete which has also undergone selection for sperm competition and for consequent de- velopment of proportionately large testes like some other cetaceans (Brownell and Ralls, 1986). In fiWinn, H. E. 1982. A characterization of marine mammals and turtles in the Mid- and North Atlantic areas of the U.S. outer continental shelf. U.S. Dep. Interior, Bureau of Land Management, Washington D.C., Final Rep. of the Cetacean and Turtle Assessment Program, 437 p. 448 Fishery Bulletin 91(3), 1993 the case of small marine mammals there is energetic advantage in evolving a limited mating period, together with a reduction in testis size out of the breeding sea- son (Gaskin et al., 1984). Field observations of Phocoena phocoena populations suggest that there is little co- herent school structure (eg., Amundin and Amundin, 1971; Watson, 1975; Watts and Gaskin, 1989); there is a mean group size of only two or three individuals, often mother, calf and yearling. Larger groups are gen- erally temporary feeding aggregations. Similar results were obtained in studies of Neophocoena phocoenoides by Kasuya and Kureha ( 1979). Distributional limits The most southerly record of harbor porpoise from Japanese waters (Taiji, Wakayama, 34°15'N) quite closely parallels that for the oceanographically similar western North Atlantic (Cape Fear, southern N. Caro- lina) (Gaskin, 1984). The confirmed northernmost limit for the western North Pacific in recent years is repre- sented by a number of sightings on the eastern side of the entrance of Shelikhova Bay, between 57°-58°N and 157°-159°E on 10-11 August 1989 (Miyashita and Doroshenko, 1990). On the eastern side of the North Pacific there are verified records in the Beaufort Sea as far as the MacKenzie Delta at about 70°N (Van Bree et al., 1977). Tomilin (1957), citing Sleptsov (un- published data in manuscript report) who made obser- vations in 1947-48, noted that the harbor porpoise was known in the USSR as far south as Peter the Great Bay, through the Sea of Okhotsk, around Kamchatka (where it was sometimes trapped in fishing nets), and northwards at least to Olyutorskii Bay at 60°N. A similar distribution was indicated by Klumov (1959). Tomilin (1957) thought its range might extend into the western part of the Chukchi Sea but gave no records. Sleptsov had also recorded porpoises in sum- mer months around the Komandorski Islands and on both the east and west sides of the Kurile Islands. Miyashita and Berzin (1991) observed harbour por- poises just west of the southwestern Kuriles in early August 1990 and to the northwest of these islands and close to southeastern Sakhalin in mid-late August of the same year. Recent records from Alaska and the eastern Aleutian Islands are summarized by Gaskin (1984); and from the Attu island group in the western Aleutians, by Jones (1984). Postulated seasonal changes in distribution in the N.W. Pacific Given the similarities of life history parameters of har- bor porpoise populations in both major oceans of the northern hemisphere, one way to estimate the likely range of the species in the western North Pacific and Bering Sea is to plot surface isotherms for values which generally appear to be limiting in regions where dis- tribution is better known (Figs. 3 and 4). Winn6, work- ing off the eastern United States, found the species were confined almost entirely to coastal shelf waters and 90% of sightings were concentrated over water depths of 18-224 m off the east coast, in sea surface temperatures of 6.5-17.0°C. Barlow (1988) found that sightings off the west coast of the United States were most abundant over depths of 18-37 m and there were no sightings at depths greater than 110 m. Surface temperatures were not specified in this study. The ar- eas of the North Pacific and Bering Sea which ap- proximate these average conditions are shown for win- ter and summer in Figures 3 and 4. There is no doubt, however, that this species must sometimes follow productive convergence zones away from the coastal shelf, or it could not have attained its present distri- bution (Gaskin, 1992). Following the important off- shore North Pacific record of Jones (1984) (see Intro- duction), Stenson and Reddin (1990) reported harbor porpoises over deep water in the Labrador Sea. Given the timing of seasonal occurrence in this region (Gaskin, 1984), these animals may have been travelling between the coastal shelves of Canada and West Greenland. The surface temperature regime in the western North Atlantic indicated by Winn6 more or less defines the productive, well-mixed boundary interaction zone be- tween the cold, southward-flowing Labrador Current and its lesser branches and the warm northeastward- flowing North Atlantic Drift; but the temperature re- gime also defines the general range of the harbor por- poise. The range of the species in the northwestern Pacific, therefore, may be similarly defined by the in- teractions of the cold Oyashio water masses coming out of the western Bering Sea and Sea of Okhotsk with the northward and northeastward flow of the Kuroshio current and its subsidiaries around Japan (Hikosaka and Watanabe, 1957; Fukuoka, 1962). The Sea of Okhotsk is significantly colder in winter than the surrounding ocean regions, and the northern and western zones of it are characteristically ice-covered from late December to April, as is the Tartary Strait between Sakhalin Island and Siberia from early De- cember to May (eg., Japan Meteorological Agency, 1986). While a small number of harbour porpoises ei- ther stay in or visit the western Bay of Fundy in win- ter (November-April) when water temperatures range from 1° to 4°C (Gaskin, 1984), any P. phocoena that stay around northern Japan during the winter months can be no further north in coastal waters than the extreme southeastern Sea of Okhotsk adjacent to the central and northern Kuriles and southwest of Sak- halin. The situation in most of the Sea of Okhotsk Gaskm et al.: Phocoena phocoena in the coastal waters of northern Japan 449 Figure 3 The western North Pacific and western Bering Sea, showing average summer surface isotherms (July-September) (Japanese Meteorological Agency 1986; Miyashita and Kasuya 1988; and Nasu 1963, 1966). Diagonal hatching indicates postulated summer range of about 90<7r of the harbor porpoise population based on criteria of Winn*. Solid circles indicate confirmed summer records. Solid squares are confirmed spring I April-June) records. Figure 4 The western North Pacific and western Bering Sea, showing average winter surface isotherms (December-March) (Japanese Meteorological Agency 1986: Miyashita and Kasuya 1988; and Nasu 1963, 1966). Horizontal hatching indicates postulated winter range of about 90<7r of the harbor porpoise population based on criteria of Winn6. Solid circles indicate confirmed winter records. 450 Fishery Bulletin 91(3), 1993 must be analogous to that of the outer Estuary and inner Gulf of the St. Lawrence in eastern Canada (Laurin, 1976) and the Baltic Sea (M0hl-Hansen, 1954), where harbor porpoises leave ahead of the ice forma- tion. The situation is likely to be different in the Sea of Japan, which in general has a much less severe marine climate than the Sea of Okhotsk (Hirano, 1957). The records from Wakkanai show that harbor por- poises occur off northern Hokkaido during early win- ter in some years. In December 1986, a small zone of inshore surface water off Wakkanai was still at 9°C (Japan Meterological Agency, 1986). Some specimens have been taken incidentally by the gill-net fishery around the Shakotan Peninsula in January and Feb- ruary7 where sea surface temperatures are about 4-5°C. The seasonal isotherm distributions suggest that in mid-winter P. phocoena might be expected to be found sometimes as far south as eastern Shikoku, to the southern tip of western Honshu and also along the east coast of the Korean peninsula (Fig. 3). Neverthe- less, it has yet to be confirmed further south than Nishiyama (about 39°N) on the Japan Sea coast of Honshu (Institute of Cetacean Research, 1989). On the east coast of Honshu, the known southern limit of range at Taiji also coincides with the intrusion of deep water associated with the Fossa Magna geological disconti- nuity, which is a feature of the northern extremity of the Philippine tectonic plate margin (Pinet, 1992). The west coast manifestation of this structure occurs in Toyama Bay, at about 37°N. In the summer months, harbor porpoises probably do not occur much further south on the west coast of Japan than the Tsugaru Strait and the western ex- tremity of Hokkaido (Kawamura et al. 1983; Kawamura 1986) (Fig. 4), where surface temperatures of 12 to 16°C often persist through summer (Miyashita and Kasuya, 1988). The progression of seasonal reoccu- pation of northern waters in the western North Pacific is not well understood. The changes in distribution predicted here are based on the limits for the occur- rence of about 90% of the population in the 6.5°-17.0°C range in the western North Atlantic (Winn6). Further- more, Yurick (1977) noted that the greatest concentra- tion of harbor porpoises in the lower Bay of Fundy was associated with the 8°-15°C surface isotherm zone. Sea ice leaves the northern coast of Hokkaido usu- ally in late April-early May and warmer surface wa- ters of up to 8°C intrude into this region in May. By June surface temperatures of 5.0°-7.0°C occur as far north as the Komandorski Islands in average years, and by mid-July the same temperatures can be found off Cape Navarin at 63°N (Nasu, 1963, 1966). There is Hiroshi Nitto, Otaru Aquarium. Otaru, Shakotan, Hokkaido. Japan, pets, commun. much year-to-year variation; in seasons with late de- velopment of summer conditions, the same regions can have surface temperatures as low as 2° to 4°C (Nasu, 1966). During August, surface waters reach 10°C both in this region and off Unalaska Island in the Aleutians (Nasu, 1966). Surface temperatures in the Sea of Okhotsk during June are often only 4°-5°C; but, by August-September 12°-15°C (Japan Meterological Agency, 1986). Although cooler intermediate water up- welling in summer around the central Kuriles is still only at 4°-6°C, highly productive boundary conditions for marine life are developed here (Hikosaka and Watanabe, 1957). The summer (August and early Sep- tember) sightings of harbor porpoises in the Sea of Okhotsk during the extensive cruises of 1989 and 1990 (Miyashita and Doroshenko, 1990; Miyashita and Berzin, 1991) were almost all made over shelf regions; in contrast the pelagic central region was dominated by Phocoenoides dalli. On the basis of distributional and energetic data from the western North Atlantic (Yasui and Gaskin, 1986), P. phocoena should tolerate a 6.5°-12°C surface regime with relative ease. Nevertheless, recent results of mitochondrial DNA comparisons of harbor porpoises in eastern Canadian waters (Wang et al., 1991) sup- port the concept of distinct genetic demes coexisting within the regional range of the western North Atlan- tic population. We should not discount the possibility that some of these demes, either through differing di- etary requirements or variation in thermoregulatory capacity, may be adapted to somewhat different sur- face temperature regimes. A specific area could be oc- cupied by porpoises from one deme in summer and another in the winter, giving the appearance of con- tinuous occupation by a single population. Migrations may exist but would be difficult to define without ex- tensive tagging experiments. In other areas of the northern hemisphere where the diet of harbor porpoises has been investigated (sum- marized by Gaskin, 1985), the main prey are schooling fishes of the coastal shelf, particularly herring Clupea sp., mackerel Scomber spp., squid, and small gadoids. Too little is known of the harbor porpoise in the west- ern North Pacific to assess its dietary spectrum; in other areas, prey switches occur when herring avail- ability is low (Recchia and Read, 1989). It seems, from the few data currently available, that harbor porpoises in Japanese waters eat sardines, anchovy, small hake, squid, and herring when encountered. The intensity of inshore fishing activity in Japanese coastal waters is such that the harbor porpoise cer- tainly has been subjected to some level of incidental catch historically, especially by set or drift nets of vari- ous kinds. Entrapments still occur; total annual in- cidental catches around the Shakotan Peninsula, Gaskin et al.: Phocoena phocoena in the coastal waters of northern Japan 451 Nemuro, and northwestern Honshu are possibly in the low hundreds8. Tobayama et al. (1991) noted that about 1,700 large trap nets and about 10 times that number of smaller-scale nets are in operation at some time of year in Japan. About 4,800 nets operate in Hokkaido waters; 2,500 on the Pacific coast of northern and west- ern Honshu and perhaps about 1,500 on the northern half of the Sea of Japan coast of Honshu. All these regions are coincident with the distribution of the har- bor porpoise in Japanese waters. Tobayama et al. con- cluded that where minke whale entrapments were con- cerned, the catches were largely unrecorded through the coastal fisheries statistics. Presumably this is also true of harbor porpoise entrapments. Given that we noted deaths of P. phocoena in gill nets not recorded in the annual statistics, there is significant under-report- ing of incidental kills by gill netting too. We lacked the resources to conduct a survey to determine how large the incidental catch of harbor porpoises by fixed gear in Japanese waters might be. The status of the harbor porpoise in Japanese wa- ters is basically unknown. Anecdotal accounts from fishermen in the Usujiri region indicate that it is not particularly abundant in comparison to other dolphins (species not specified) and seasonal in occurrence (spring and fall). It is present in waters off Otaru in all but the months of July-September (Hiroshi Nitto, personal communication). A detailed assessment of its status in the western North Pacific is required. The species now appears to be threatened in some eastern and western North Atlantic waters (Gaskin, 1992). The limitations imposed by the fixed litter size of one, the relatively late age at first maturity and an annual reproductive cycle all point to little reproductive flex- ibility with which to respond to significant additional mortality from hunting or incidental capture (Woodley and Read, 1991; Gaskin, 1992). Acknowledgments The authors thank the following for their cooperation during the collection and study of specimens for this study: Nobuyuki Miyazaki of the National Museum of Science, Tokyo; Seiji Ohsumi (retired) and Toshio Kasuya of the Far Seas Fisheries Research Labora- tory, Shimizu; Hideo Omura, former Director of the Whales Research Institute, Tokyo; Hiroshi Nitto of the Otaru Aquarium; Yasunori Sakurai and Takahiro Koga of the Asamushi Aquarium, Aomori; Masaaki Mori, Act- ing Director of the Sunshine City Aquarium, Tokyo, 'A. Kawamura, Faculty of Bioresources, Mie University, Tsu 514, Japan, unpubl. data. "Hiroshi Nitto, pers. commun. and Kiyoji Miyagaki, President of Sunshine Kogyo Cor- poration, Tokyo; Junroku Ogasawara and Syuichi Nishiuchi of the Hokkaido Fisheries Experimental Sta- tion, Wakkanai; Kiyoshi Nomura and all the fishermen of the Nomura Fishing Company in Usujiri, Hokkaido; and the University of Hokkaido Faculty of Fisheries at Hakodate which provided laboratory and field facili- ties and subsidized accommodation. Special thanks are extended to our good friend Yoetsu Arashida, Manager of the University of Hokkaido Marine Biology Station at Usujiri who looked after our field accommodation and gave many personal kindnesses during our two lengthy stays. Funding for this work was provided by two travel fellowships (1985 and 1986), under the Bi- lateral Agreement between the Natural Sciences and Engineering Research Council of Canada and the Japan Society for the Promotion of Science (special thanks to Hiroshi Kida and A. Kugler), and with the further support of NSERC Operating Grant A5683. Literature Cited Amano, M., and N. Miyazaki. 1992. Geographic variation in skulls of the harbor por- poise, Phocoena phocoena. Mammalia 56: 133-144. Amu iid in, M., and B. Amundin. 1974. On the behavior and study of the harbour por- poise, Phocoena phocoena, in the wild. In G. Pilleri (editor), Investigations on Cetacea, Vol. V, p. 317- 328. Berne, Switzerland. Andersen, L. W. 1990. The population structure of Phocoena phocoena in Danish waters. Annual meeting of the Scientific Committee of the International Whaling Commission, Noordwijkerhout, The Netherlands, 10-23 June 1990, document SC/42/SM50. Barlow, J. 1988. Harbor porpoise, Phocoena phocoena, abundance estimation for California, Oregon and Washington: I. Ship Surveys. Fish. Bull. 86: 417-432. Barlow, J., and D. A. Hanan. 1990. An assessment of the status of harbor porpoise populations in California. Annual Meeting of the Sci- entific Committee of the International Whaling Com- mission, Noordwijkerhout, The Netherlands, 10-23 June 1990, document SC/42/SM6. Barlow, J., C. W. Oliver, T. D. Jackson and B. L. Taylor. 1988. Harbor porpoise, Phocoena phocoena, abundance estimation for California, Oregon, and Washington: II. Aerial Surveys. Fish. Bull. 86: 433-444. Bjorge, A., and S. Kaarstad. 1990. Preliminary analysis of growth and reproduc- tion of harbour porpoises Phocoena phocoena in Nor- wegian waters. Annual Meeting of the Scientific Committee of the International Whaling Commission, Noordwijkerhout, The Netherlands, 10-23 June, docu- ment SC/42/SM2. 452 Fishery Bulletin 91(3), 1993 Bjorge, A., and N. Oien. 1990. Distribution and abundance of harbour porpoise Phocoena phocoena in Norwegian waters. Annual Meeting of the Scientific Committee of the Interna- tional Whaling Commission, Noordwijkerhout, The Netherlands, 10-23 June, document SC/42/SM3. Brownell, R., and K. Ralls. 1986. Potential for sperm competition in baleen whales. Rep. Int. Whal. Comm. (Spec. Issue 8): 97-112. Celikkale, M. S. 1990. The fishery in the Black Sea. Annual Meeting of the Scientific Committee of the International Whal- ing Commission, Noordwijkerhout, The Netherlands, 10-23 June, document SC/42/SM40. Dalheim, M., A. York, J. Waite and C. Goebel— Diaz. 1991. Alaskan harbor porpoise {Phocoena phocoena) population studies. Ninth Biennial Conference on the Biology of Marine Mammals, Chicago, Illinois, 5-9 December 1991. Abstracts: 16. Evans, P. G. H. 1990. Harbour porpoises {Phocoena phocoena) in Brit- ish and Irish waters. Annual Meeting of the Scien- tific Committee of the International Whaling Commission, Noordwijkerhout, The Netherlands, 10- 23 June, document SC/42/SM49. Fisher, H. D. and R. J. Harrison. 1970. Reproduction in the common porpoise {Phocoena phocoena) of the North Atlantic. J. Zool. Lond. 16: 471-486. Fontaine, P. M., C. Barrette, and M. O. Hammill. 1991. Testis size and reproduction of male harbour por- poise in the Estuary and the Gulf of St. -Lawrence, Quebec, Canada. Ninth Biennial Conference on the Biology of Marine Mammals, Chicago, Illinois, 5-9 December 1991. Abstracts: 23. Forney, K. A., and J. Barlow. 1991. Detecting trends in abundance of harbor por- poise, Phocoena phocoena , in central California. Ninth Biennial Conference on the Biology of Marine Mam- mals, Chicago, Illinois, 5-9 December 1991. Abstracts: 23. Fraser, F. C. 1974. Report on Cetacea stranded on the British coasts from 1948 to 1966. 14. British Museum (Natural History) London. Fukuoka, J. 1962. Characteristics of hydrography of the Japan Sea in 1962 — comparison with hydrography of the North Pacific. J. Oceanogr. Soc. Japan, 20th Anniversary Vol. 8: 180-188. Gaskin, D. E. 1977. Harbour porpoise, Phocoena phocoena (L), in the western approaches to the Bay of Fundy 1969- 75. Rep. Int. Whal. Commn. 27: 487-492. 1984. The harbour porpoise Phocoena phocoena (L.) : Regional populations, status and information on di- rect and indirect catches. Rep. Int. Whal. Comm. 34: 569-586. 1985. The ecology of whales and dolphins. Heinemann Educational Books, London and Exeter, UK, 459 p. 1992. The status of the harbour porpoise, Phocoena phocoena in Canada. Can. Field-Nat. 106: 36-54. Gaskin, D. E., and B. A. Blair. 1977. Age determination of harbour porpoise, Phocoena phocoena (L.), in the western North Atlantic. Can J. Zool. 55: 18-30. Gaskin, D. E. and A. P. Watson. 1985. The harbor porpoise, Phocoena phocoena, in Fish Harbour, New Brunswick, Canada: Occupancy, distri- bution and movements. Fish. Bull. 83: 427-442. Gaskin, D. E., G. J. D. Smith, A. P. Watson, W. Y. Yasui and D. B. Yurick. 1984. Reproduction in the porpoises {Phocoenidae): Im- plications for management. Rep. Int. Whal. Comm. (Spec. Issue 6): 135-148. Hawley, F. 1960(1958). Whales and whaling in Japan. Kawakita Publ. Co., Miscellanea Japonica Vol. 1, 354 p. Hikosaka, S., and R. Watanabe. 1957. Areas of divergence and convergence of surface currents in the Northwestern Pacific. Proc. UNESCO Symp. Phys. Oceanogr. (1955): 101-103. Hirano, T. 1957. The oceanographic study on the subarctic region of the Northwestern Pacific Ocean, 2. On the forma- tion of the subarctic water system. Bull. Tokai Reg. Fish. Res. Lab. (15): 57-69. Hohn, A. A., and R. L. Brownell Jr. 1990. Harbour porpoise in central California waters: Life history and incidental catches. Annual Meeting of the Scientific Committee of the International Whal- ing Commission, Noordwijkerhout, The Netherlands, 10-23 June, document SC/42/SM47. Hohn, A. A., and K. Peltier. 1990. An annotated bibliography of harbour porpoises, Phocoena phocoena, life history and exploitation. An- nual Meeting of the Scientific Committee of the In- ternational Whaling Commission, Noordwijkerhout, The Netherlands, 10-23 June, document SC/42/SM19. Institute of Cetacean Research. 1989. [Stranding Records] (In Japanese.) Geiken Tushin [Newsletter of the Institute of Cetacean Re- search] No. 376. International Whaling Commission. 1984. Report of the sub-committee on small cetaceans. (Annex H). Rep. Int. Whal. Comm. 34: 144-160. 1985. Japan. Progress report on cetacean research, June 1983 to April 1984. Rep. Int. Whal. Comm. 35: 168-172. 1986. Japan. Progress report on cetacean research, May 1984 to May 1985. Rep. Int. Whal. Comm. 36: 158- 161. 1987. Japan. Progress report on cetacean research, June 1985 to April 1986. Rep. Int. Whal. Comm. 37: 172-175. 1988. Japan. Progress report on cetacean research, May 1986 to May 1987. Rep. Int. Whal. Comm. 38: 194-198. 1989. Japan. Progress report on cetacean research, June 1987 to March 1988. Rep. Int. Whal. Comm. 39: 201-204. Gaskin et al.: Phocoena phocoena in the coastal waters of northern Japan 453 1990. Japan. Progress report on cetacean research, May 1988 to April 1989. Rep. Int. Whal. Comm. 40: 198- 201. Japan Meteorological Agency. 1986. The ten day marine report: mean sea surface temperatures. Nos. 1413 to 1448 [Jan. 1-10 to Dec. 1-10, 1986]. IGOSS Specialized Oceanographic Cen- ter for the Pacific Ocean, Marine Department, Tokyo. Jones, L. L. 1984. Incidental take of the Dall's porpoise and the harbor porpoise by Japanese salmon driftnet fisheries in the western North Pacific. Rept.-Int. Whal. Comm. 34: 531-538. Kasuya, T., and K. Kureha. 1979. The population of finless porpoise in the Inland Sea of Japan. Sci. Rep. Whales Res. Inst. 31: 1-44. Kawamura, A. 1986. [Seasonal distribution of dolphins and porpoises in the fishing grounds around Aomori Prefecture and Hokkaido — Results and some considerations of the questionaire to the fishermen's union in 1984.] (In Japanese.) Geiken Tushin [Newsletter of the Whales Research Institute] No. 362. Kawamura, A., H. Nakano, H. Tanaka, S. Sato, Y. Fujise and K. Nishida. 1983. [Small toothed whales in Tsugaru Straits as ob- served by the Japanese National Railways ferry boats] (In Japanese.) Geiken Tushin [Newsletter of the Whales Research Institute] No. 351-352. Kinze, C. C. 1985. Intraspecific variation in Baltic and North Sea harbour porpoises (Phocoena phocoena L., 1758). Vidensk. Meddr dansk naturh. Foren 146: 63- 74. 1990a. The distribution of the harbour porpoise (Phocoena phocoena) in Danish waters 1983- 1989. Annual Meeting of the Scientific Committee of the International Whaling Commission, Noord- wijkerhout. The Netherlands, 10-23 June, document SC/42/SM34. 1990b. Non-metric analysis of harbour porpoise (Phocoena phocoena ) from the North and Baltic Seas: implications for stock identity. Annual Meeting of the Scientific Committee of the International Whal- ing Commission, Noordwijkerhout, The Netherlands, 10-23 June, document SC/42/SM35. Klumov, S. K. 1959. Commercial dolphins of the Far East. (Transl. by L.V. Sagen). U.S. Fish. Wildl. Service Marine Mam- mal Biology Laboratory, Seattle, mimeo, 14 p. Laurin, J. 1976. Preliminary study of the distribution, hunting and incidental catch of harbour porpoise, Phocoena phocoena L. in the Gulf and estuary of the St. Lawrence. FAO Scientific Consultation on Marine Mammals, Bergen, document ACMRR/MM/SC/93. Miyashita, T., and A. A. Berzin. 1991. Report of the whale sighting survey in the Okhotsk Sea and adjacent waters in 1990. Annual Meeting of the Scientific Committee of the Interna- tional Whaling Commission, Reykjavik, Iceland, 27- 31 May 1991, document SC/43/05. Miyashita, T., and N. V. Doroskenko. 1990. Report of the whale sightings survey in the Okhotsk Sea, August 1989. Annual Meeting of the Scientific Committee of the International Whaling Commission, Noordwijkerhout, The Netherlands, 10- 23 June, document SC/42/018. Miyashita, T., and T. Kasuya. 1988. Distribution and abundance of Dall's porpoises off Japan. Sci. Rep. Whales Res. Inst. 39: 121-150. Miyazaki, N. 1983. Catch statistics of small cetaceans taken in Japa- nese waters. Rep. Int. Whal. Comm. 33: 621-631. Miyazaki, N., M. Amano and Y. Fujise. 1987. Growth and skull morphology of the harbor por- poises in the Japanese waters. Mem. Natl. Sci. Mus. (Tokyo) 20: 137-146. MohhHansen, U. 1954. Investigations on reproduction and growth of the porpoise (Phocoena phocoena (L)) from the Bal- tic. Vidensk. Meddr dansk naturh. Foren. 116: 369- 396. Nasu, K. 1963. Oceanography and whaling ground in the sub- arctic region of the Pacific ocean. Sci. Rep. Whales Res. Inst. 17: 105-155. 1966. Fishery oceanographic study on the baleen whal- ing grounds. Sci. Rep. Whales. Res. Inst. 20: 157-210. Nishiwaki, M. 1966. Distribution and migration of marine mammals in the North Pacific area. Proc. 11th Pacific Sci. Congr. Symp. No. 4: 1-49. Noldus, L. P. J. J., and R. J. J. de Klerk. 1984. Growth of the skull of the harbour porpoise, Phocoena phocoena (Linnaeus, 1758), in the North Sea, after age determination based on dentinal growth layer groups. Zool. Meded., Leiden. 58: 213-239. Ohsumi, S. 1972. Catch of marine mammals, mainly of small cetaceans, by local fisheries along the coast of Japan. Bull. Fish. Res. Lab. Shimizu 7: 137-166. Pinet, P. R. 1992. Oceanography. West Publishing Company, St. Paul, MN, and New York, NY, 571 p. Polacheck, T., and F. W. Wenzel. 1990. What do stranding data say about harbor por- poise (Phocoena phocoena)? Annual Meeting of the Scientific Committee of the International Whaling Commission, Noordwijkerhout, The Netherlands, 10- 23 June, document SC/42/SM39. Prescott, J. H., and P. M. Fiorelli. 1980. Review of the harbor porpoise (Phocoena pho- coena) in the U.S. Northwest Atlantic. National Tech- nical Information Service U.S. Dep. of Commer. PB-80-1 76928, 64 p. Read, A. J. 1990a. Reproductive seasonality in harbour porpoises, Phocoena phocoena, from the Bay of Fundy. Can. J. Zool. 68: 284-288. 454 Fishery Bulletin 91(3), 1993 1990b. Age at sexual maturity and pregnancy rates of harbour porpoises Phocoena phocoena from the Bay of Fundy. Can. J. Fish. Aquat. Sci. 47: 561-565. Read, A. J., and D. E. Gaskin. 1990. Changes in growth and reproduction of harbour porpoises, Phocoena phocoena, from the Bay of Fundy. Can. J. Fish. Aquat. Sci. 47:2158-2163. Recchia, C. A., and A. J. Read. 1989. Stomach contents of harbor porpoises, Phocoena phocoena (L.), from the Bay of Fundy. Can. J. Zool. 67: 2140-2146. Rosel, P. E., and M. G. Haygood. 1991. Relative discreteness of harbor porpoise stocks in the northeast Pacific from mDNA se- quences. Ninth Biennial Conference on the Biology of Marine Mammals, Chicago, Illinois, 5-9 December 1991. Abstracts: 59. Sequiera, M. L. 1990. Gillnets and cetacean mortality in Portu- gal. IWC Symposium on mortality of cetaceans in passive fishing nets and traps. La Jolla, California, October 20-21, 1990. Abstract, p. 42. Stacey, P. J., R. W. Baird and D. A. Duffus. 1990. A preliminary evaluation of incidental mortality of small cetaceans, primarily Dall's porpoise (Pho- coenoides dalli), harbour porpoise {Phocoena phocoena), and Pacific white-sided dolphins (Lagen- orhynchus obliquidens), in inshore fisheries in Brit- ish Columbia, Canada. Annual Meeting of the Scientific Committee of the International Whaling Commission, Noordwijkerhout, The Netherlands, 10- 23 June, document SC/42/SM20. Stenson, G. B., and D. G. Reddin. 1990. Incidental catches of small cetaceans in drift nets during salmon tagging experiments in the Northwest Atlantic. IWC Symposium on mortality of cetaceans in passive fishing nets and traps. La Jolla, Califor- nia, October 20-21, 1990. Abstract, p. 46. Stuart, L. J., and G. V. Morejohn. 1980. Developmental patterns in osteology and exter- nal morphology in Phocoena phocoena. Rep. Int. Whal. Comm. (Spec. Issue 3): 133-142. Taylor, B. L., and P. K. Dawson. 1984. Seasonal changes in density and behaviour of harbor porpoises (Phocoena phocoena) affecting censusing methodology in Glacier Bay National Park, Alaska. Rep. Int. Whal. Comm. 34: 479-483. Tobayama, T., F. Yanagisawa and T. Kasuya. 1991. Incidental take of minke whales in Japanese trap nets. Annual Meeting of the Scientific Committee of the International Whaling Commission, Reykjavik, Iceland, 27-31 May 1991, document SC/43/MI29. Tomilin, A. G. 1957. Mammals of the U.S.S.R. and Adjacent Coun- tries. Vol. IX Cetacea. Izkatel'stvo Akademi Nauk U.S.S.R., Moscow. [Translated and published in 1967 in agreement with the Smithsonian Institution and the National Science Foundation, Washington, D.C., U.S.A., 717 p.] Van Bree, P. J. H., D. E. Sergeant and W. Hoek. 1977. A harbour porpoise Phocoena phocoena (Linnaeus, 1758) from the MacKenzie River delta, Northwest Territories, Canada (Notes on Cetacea, Delphinoidea VIII) Beaufortia 26: 99-105. Van Utrecht, W. L. 1978. Age and growth in Phocoena phocoena Linnaeus 1758 (Cetacea, Odontoceti) from the North Sea. Bijdr. Dierk. 48: 16-28. Wang, J. Y., D. E. Gaskin and B. N. White. 1991. Visible band variation in the DNA of harbour porpoises (Phocoena phocoena [L.]) from the Bay of Fundy. Ninth Biennial Conference on the Biology of Marine Mammals, 5-9 December 1991, Chicago, Illi- nois, USA. Watts, P., and D. E. Gaskin. 1989. A comparison of age determination techniques for the harbour porpoise, Phocoena phocoena ( L. ). Can. J. Zool. 67. 1832-1836. Woodley, T. H. and A. J. Read. 1991. Potential growth of a harbor porpoise (Phocoena phocoena) population subjected to incidental mortality. Can. J. Fish. Aquat. Sci. 48: 2429-2435. Yasui, W. Y. and D. E. Gaskin. 1986. Energy budget of a small cetacean, the harbour porpoise, Phocoena phocoena (L.). Ophelia 25(3): 183-197. Yurick, D. B. 1977. Populations, subpopulations, and zoogeography of the harbour porpoise, Phocoena phocoena (L.). M.Sc. thesis, Univ. Guelph, Guelph, Ontario. Yurick, D. B., and D. E. Gaskin. 1987. Morphometric and meristic comparisons of skulls of the harbour porpoise Phocoena phocoena (L. ) from the North Atlantic and North Pacific. Ophelia 27: 53- 75. Zalkin, V. I. 1940. Materaily K biologii morski svin'i (Phocoena phocoena) Azovskogo i chernogo morei. Bull. Soc. Nat. Moscow 49: 61-68. Abstract. -Recruitment of indi- viduals from the epipelagic phase to the demersal armorhead Pseudo- pentaceros wheeleri population at Southeast (SE) Hancock Seamount has typically been recognized by the influx of fish with a high fatness in- dex (FI; body depth relative to fork length). Reliance on this index was dictated by the peculiarity of the sea- mount life history stage in which armorhead cease somatic growth soon after recruitment to the sea- mount and FI declines during sea- mount residence until death. Lim- ited sampling opportunities and variability in FI at the time of re- cruitment preclude the exclusive use of FI as a means of identifying newly arrived recruits among recruits-of- the-year and hence the timing of an- nual recruitment. Settlement and recruitment to the seamount popu- lation are synonymous. Efforts to de- velop a method of identifying new recruits were initiated by an exami- nation of epipelagic and seamount (SE Hancock) armorhead for differ- ences in macroparasites as well as hepatosomatic and visceral fat-so- matic indices. Of the two condition indices, the hepatosomatic index held the most potential but was con- sidered too labile because pre- recruits probably experience de- creases that vary according to ener- getic demands during seamount migration. The monogenean gill parasite Microcotyle macropharynx was highly prevalent among sea- mount individuals in all sampling periods but absent from epipelagics. Based on probable rapid infection and maturation, identification of new recruits was based on the absence of mature M. macrophaynx. Results indicate that significant new recruit- ment occurred only during the late spring to midsummer sampling pe- riods, and the mean FI of new sea- mount recruits was lower than that of comparable size epipelagics. Use of a monogenean gill parasite and feasibility of condition indices for identifying new recruits to a seamount population of armorhead Pseudopentaceros wheeleri (Pentacerotidae) Robert L. Humphreys Jr. Honolulu Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service. NOAA 2570 Dole Street, Honolulu, Hawaii 96822-2396 Mark A. Crossler Honolulu Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service, NOAA 2570 Dole Street. Honolulu, Hawaii 96822-2396 Present address: School of Business Administration University of Oregon, Eugene, Oregon 97402 Craig M. Rowland Honolulu Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service, NOAA 2570 Dole Street, Honolulu, Hawaii 96822-2396 Present address: U.S. Fish and Wildlife Service, PO Box 50167, Honolulu. Hawaii 96850 Manuscript accepted: 19 March 1993 Fishery Bulletin 91:455-463 ( 1993) The armorhead Pseudopentaceros wheeleri (Hardy) in the North Pacific Ocean initially occupies an epipelagic and then demersal habitat during its life history. Progeny originate from winter spawning of the demersal armorhead population associated with the summit and upper slopes of the southern Emperor-northern Ha- waiian Ridge (SE-NHR) seamounts (Fig. 1) (Bilim et al., 1978; Borets, 1979). Larvae are initially found in surface waters around the seamounts (Fedosova and Komrakov, 1975; Borets, 1979) but soon move or are advected into subarctic waters of the central and northeastern Pacific (Boehlert and Sasaki, 1988). Here they feed in oceanic surface waters and undergo somatic growth, accu- mulate energy reserves, but experi- ence no reproductive development. These epipelagic individuals appear deep-bodied and display a striking pattern of blue and silver blotches along their dorsal and lateral regions. The process and timing by which epipelagic individuals recruit back to the SE-NHR seamounts is not well understood (Boehlert and Sasaki, 1988). Individuals are primarily 28- 33 cm fork length (FL) upon recruit- ment to the SE-NHR seamounts (Humphreys et al., 1989). Since re- cruits are similar in length and co- occur among the resident seamount population, settlement and recruit- ment to the seamounts are consid- ered synonymous events ( Humphreys et al, 1989). Unlike the epipelagics, the seamount population is demer- sal and virtually uniform in length, but morphologically variable. Three morphological variants (the "fat," "in- termediate," and "lean" morphotypes) were previously recognized and found 455 456 Fishery Bulletin 91(3), 1993 to represent different stages of the demersal seamount phase (Humphreys et al., 1989). This variation inter- grades from the deep-bodied, overall bluish-gray col- oration of presumably recent recruits (fat morphotype) through progressively leaner forms of uniformly brown- ish coloration (intermediate and lean morphotypes). Relative body depth refers to the ratio of body depth at first anal spine to FL and hereafter is called the fatness index (FI; Somerton and Kikkawa, 1992). Us- age of FI values recognizes that these variations form a continuum rather than discrete changes and has re- placed classification by morphotype (Somerton and Kikkawa, 1992). Declining FI reflects the transforma- tion by individuals as the elapsed time of seamount residence increases. During this phase, somatic growth ceases upon recruitment, and stored energy reserves are eventually mobilized for reproductive development and spawning. These processes contribute to the de- cline in FI and ultimately lead to death (Humphreys etal., 1989). The discovery of large aggregations of armorhead over the summits and upper slopes of SE-NHR sea- mounts (Fig. 1) during exploratory work by the former Soviet Union in 1967-1969 (Komrakov, 1970) led to a full-scale bottom trawl fishery in this region by both Japanese and the former Soviet trawl fleet (Uchida and Tagami, 1984). By the end of 1975, the cumulative combined fleet catch of armorhead was some 880,000 metric tons (t). Total annual catch of armorhead by the Japanese trawl fleet declined dramatically after 1976 (Humphreys et al., 1984) and continues to re- main low throughout the SE-NHR region (R. Humphreys, unpubl. data). In 1977 the Hancock Sea- mounts, located in the southermost portion of the SE- NHR trawl fishery, came under U.S. jurisdiction with the implementation of the 200-mile fishery conserva- tion zone (Humphreys et al., 1984). During 1978-1984, several Japanese trawlers were allowed to fish the Hancock Seamounts but were regulated by permit and an annual catch quota (Uchida and Tagami, 1984; -UJt 1 \ Koko Seamount Kimmei Seamount _ Oaikakuji Seamount -^ Yuryaku Seamount MILWAUKEE SEAMOUNT . GROUP J2.0, they found recruitment highest during April-May 1973 and secondary peaks in August 1972 and July 1973. An- nual recruitment to Southeast (SE) Hancock Seamount, one of two guyots composing the Hancock Seamounts, was sporadic between 1978 to 1990. Strong recruit- ment (>50% of population biomass) at SE Hancock Sea- mount occurred only in 1980 and 1986, and moderate recruitment (ca. 20-30% of population biomass) in 1988 and 1990, based on catch per unit of effort (CPUE) and FI distributions derived from Japanese trawl data and NMFS stock assessment surveys (Somerton and Kikkawa, 1992). Little further progress, however, has been made in discerning the timing of recruitment and biological characteristics of newly arrived recruits. The apparent temporal variability in recruitment, the change in FI during time elapsed between recruitment and sam- pling, coupled with limited sampling opportunities, make the identification and study of new recruits and their temporal recruitment pattern problematic. A po- tential alternative for identifying newly recruited armorhead uses biological indicators to distinguish a new seamount arrival from longer resident armorhead. Parasites have been frequently used as biological indi- cators in fish and their benefit to studies of fish stock separation, movement, and diet are numerous (Wil- liams et al., 1992). In relation to seamount studies, differences in the trematode parasite fauna of sable- fish off Canada's west coast led Kabata et al. (1988) to surmise that seamount populations of sablefish are distinct and separate from those sablefish inhabiting the continental slope. Since seamount armorhead popu- lations are derived from epipelagic individuals, we ex- amined specimens of both life history phases for para- site differences and other distinguishing features. We report on the feasibility of using a monogenean gill parasite and two condition indices as potential indica- tors of new seamount recruits. Methods Epipelagic specimens of P. wheeleri (n = 53) were inci- dentally captured from the central North Pacific (around lat. 45°N, long. 155°W) in July 1984 and 1985 and from the eastern North Pacific (around lat. 46°N, long. 129°W) in July 1985 by the Japanese fishing vessels Oshoro Maru and Tomi Maru No. 88, respec- tively, during salmon longline and surface drift gill- net operations. These specimens were saved intact and stored frozen prior to examination. Specimens of sea- mount armorhead (n = 1,220) were collected from the southern end of the SE-NHR at SE Hancock Seamount (29°48'N, 179°04'E; see Fig. 1) by bottom longline gear during research cruises of the NOAA ship Townsend Cromwell in June 1985, August and October 1986, April and August 1987, and January, July, August, and No- vember 1988. All seamount specimens were derived from efforts beginning in 1985 to estimate armorhead relative abundance (catch per unit of effort) at SE Hancock Seamount by using standardized assessment techniques described in Somerton and Kikkawa (1992). Seamount specimens were either saved intact and stored frozen, or gills and viscera were removed and preserved at sea in either alcohol-formalin-acetic acid (AFA) solution or placed in a dilute (0.0004%) formalin solution for several hours and then preserved in 10% formalin. Collection data for armorhead specimens in- cluded body weight, FL, body depth at first anal spine, sex, and date and location of capture. Measurements were made to the nearest millimeter, and body weight to nearest gram. Under a dissecting microscope, the gill arches and visceral organs of armorhead specimens were exam- ined for macroparasites. Guidelines proposed by Sindermann (1983) were followed to determine which species have potential as biological tags. Ideal para- site characteristics include ease of detection and iden- tification, a single host-life cycle, and temporal stabil- ity in prevalence. Furthermore, the parasite should persist in the host during the study, and prevalence should differ significantly between study areas. These criteria were best fulfilled by the monogenean Microcotyle macropharynx (Mamaev), based on exami- nations of the seven parasite taxa (monogenean Trochopus sp., digenean Bivesicula sp., larvae of the nematode Anisakis type I, nematode Hysterothylacium 458 Fishery Bulletin 9 1(3), 1993 sp., an unidentified caligoid copepod, and juveniles of an unidentified gnathiid isopod) found in the epipe- lagic and seamount armorhead collected in June 1985 and August 1986. Subsequent examinations of armorhead were restricted to the gills where M. macropharynx were exclusively found. Specimens of M. macropharynx were readily distinguished from Trochopus sp. by the disk-shaped opisthaptor of the latter monogenean. All specimens of M. macropharynx were initially identified as that of M. sebastis though the pharynx of these specimens appeared to be unusu- ally large1. During the review process, a new species description (Mamaev, 1989) of the monogenean M. macropharynx collected in 1969-70 from P. richardsoni (=P. wheeleri) in the Hawaiian Islands region was brought to the attention of the senior author. Subse- quently specimens initially identified as M. sebastis were re-examined and all found to be M. macro- pharynx2. The large pharynx (in relation to the anterior suckers) of M. macropharynx distinguishes this species from all other species of Micocotyle (Mamaev, 1989). Each M. macropharynx was staged as either immature or potentially mature, based on absence or development of the paired yolk ducts, re- spectively. These organs are the last to develop before egg development in the related species M. sebastis (Thoney, 1986). All parasite taxa were saved and stored in 70% ethyl alcohol. Prevalence and mean intensity of M. macropharynx in seamount armorhead were com- puted for each sampling period. Following the termi- nology defined in Margolis et al. (1982), prevalence (expressed as a percentage) refers to the number of host species infected with a particular parasite spe- cies, divided by the number of hosts examined; and mean intensity is derived by dividing the total number of a particular parasite species by the number of indi- vidual hosts infected by that parasite. Prevalence and mean intensity data from seamount armorhead were partitioned into two FI groups of <0.26 and >0.26. This grouping is based on the observation that 0.26 was the lowest FI found for epipelagic armorhead in this study and thus a possible minimum FI for newly arrived seamount recruits. Liver and visceral fat weights were also determined for the 27 epipelagic armorhead >27.5cm FL (of com- parable length to seamount individuals) and 410 sea- mount armorhead collected during October 1986 (n = 115) and January (n = 195), July (n = 18), and Novem- ber 1988 (n = 82). Liver and visceral fat deposits were R. R- Payne, Department of Biological Sciences, Biola Univ., La Mirada, CA, personal commun. 1988. -R. R. Payne, Department of Biological Sciences, Biola University, La Mirada. CA, pers. commun. 1992. removed from the body cavity, blotted dry, and weighed to the nearest 0.1 g. Liver and visceral fat weights are expressed in relation to total body weight and referred to as the hepatosomatic index (HSI) and visceral fat- somatic index (VFSI), respectively. Data was arcsine transformed and tested for normality with the Lilliefor's method of the Kolmogorov-Smirnov test (Wilkinson, 1990). Sample means were compared by either t-test, or by Mann-Whitney test if samples remained non- normal (Wilkinson, 1990). Estimates of the proportion of recruits-of-the-year to the total SE Hancock armorhead population during each sampling period were based on stock-assesssment- derived FI distributions of the armorhead population. That proportion of the FI distribution >0.26 were con- sidered recruits-of-the-year. Data and procedures used to calculate the population FI distribution are presented in Somerton and Kikkawa (1992). The proportion of recruit-of-the-year armorhead in each sampling period un-infected by mature stages of M. macropharynx was estimated from results of the prevalence data for those armorhead (from samples examined for M. macro- pharynx) with a FI >0.26. Results Microcotyle macropharynx were not present in any of the 53 epipelagic armorhead examined but were present in all 775 seamount armorhead with FI <0.26 and in 417 (93.7%) of 445 specimens with FI >0.26. Among seamount armorhead with FI <0.26, both the combined (irrespective of maturity stage) category and the im- mature and mature stage of M. macropharynx were highly prevalent (>75%) in all nine sample periods surveyed (Table 1). A similar pattern of high M. macropharynx prevalence was found among the sea- mount armorhead with FI >0.26 although instances of lower prevalence appear during the two earliest sample periods (Table 1). Mean intensity of immature M. macropharynx in both FI groups was considerably lower in each sample period than the corresponding mean intensities of the mature stage and combined category (Table 2). Mean intensity of mature stage M. macro- pharynx in each sample period and FI group was suffi- ciently high to ensure easy detection in armorhead infected with mature stage individuals of this monoge- nean. A wide range in sizes of M. macropharynx, in- cluding mature stages with eggs, was present among the gills of armorhead, indicating that this parasite has a monoxenous life cycle. Seamount armorhead without mature M. macro- pharynx represented 5.2% (64) of the 1,220 fish exam- ined. Among those seamount armorhead found un- Humphreys et al.: Identification of Pseudopentaceros wheelen recruits 459 Fable 1 Prevalence of the monoge nean Mierocotyle macropharynx in the gills of armorhead Pseudopentaceros wheeleri by fatness ndex (FI) groups of <0.26 and >0.26, collected from Southeast Hancock Seamount during 1985-88. Within each FI group prevalence is further categorized by maturity stage of monogenean or regardless of (combined). Preva- lence expressed as a percentage of total number 177 ) of armorhead examine d. Differences in (71) between m aturity stages indicates missing data. Occurrences of combined (71 1 exceeding In) of both maturity stages represents instances where monogeneans from an additional sample were not staged. Sample FI <0.26 FI >0.26 Immature Mature Immature Mature period stage stage Combined stage stage Combined Jun 1985 98.0 95.9 100.0 70.0 45.0 80.0 72=49 77=49 71=50 77=20 77=20 77=20 Aug 1986 93.0 100.0 100.0 68.8 66.7 75.0 7i=57 77=57 71=58 77=48 77=48 77=48 Oct 1986 92.5 100.0 100.0 84.9 96.2 99.1 7! =80 77=80 7i=80 71 = 106 77 = 106 71 = 107 Apr 1987 80.6 100.0 100.0 85.2 96.3 96.3 77 = 134 77 = 137 « = 137 77=27 77=27 77=27 Aug 1987 88.2 100.0 100.0 95.8 89.6 95.8 77=68 77=68 71=68 7i=48 7i=48 7i=48 Jan 1988 94.5 100.0 100.0 100.0 100.0 100.0 77 = 181 77 = 183 71=183 77 = 12 77 = 12 77 = 12 Jul 1988 93.3 100.0 100.0 78.8 80.8 90.4 77=45 77=45 7i=45 77=52 7i=52 77=52 Aug 1988 93.6 100.0 100.0 90.4 78.8 94.2 77=47 77=47 77=47 77=52 77=52 7i=52 Oct-Nov 1988 96.6 97.7 100.0 98.7 97.5 100.0 77=87 77=87 77=87 77=79 7i=79 7i=79 Table 2 Mean intensity of the monogenean Mierocotyle mac ropharynx ir the gills of armorhead Pseudopentaceros wheeleri bv fatness 1 ndex (FI) groups of <0.26 and >0.26, collected from Southeast Hancock Seamount during 1985-88 Within each FI group. mean inten- sity is further categorized by maturity stage of monogenean or regardless of (combined). Sample size from which mean intensity is based on is denoted bv (77). Differences in (n) between maturity stages indicates missing data. Occurrences of combined 77 ) exceeding (n) of both maturity stages represents instances where monogenean was present but stage of maturity not determ ned. Smal er sizes of (n) in this table compared to the (77) of corresponding eel s in Table 1 indicates instances where data on prevalence, but not mean intensity, is available. Sample FI <0.26 ^ FI >0.26 Immature Mature Immature Mature period stage stage Combined stage stage Combined June 1985 12.23 35.72 52.06 4.64 95.22 57.63 n=48 71=47 77=50 71 = 14 77=9 77 = 16 Aug 1986 9.09 63.84 72.12 6.91 30.41 34.03 77=53 7i=57 77=58 77=33 77=32 77=36 Oct 1986 9.74 88.01 97.03 5.42 65.77 68.18 77=74 7i=80 7i=80 77=90 7i=102 7i = 106 Apr 1987 16.21 65.17 79.92 19.23 162.86 181.14 77=79 77=24 77=24 77 = 13 71 = 14 71 = 14 Aug 1987 17.97 58.48 75.48 14.42 58.21 72.68 77=31 77=33 77=33 77 = 19 77 = 19 77 = 19 Jan 1988 15.84 56.66 71.11 20.44 110.00 130.44 77=31 7i=35 71=35 77=9 71=9 71=9 Jul 1988 14.47 27.69 41.25 7.67 36.47 37.80 77=30 7i=32 7i=32 77 = 18 71 = 17 77=20 infected by mature stage M. macropharynx, virtually all were of a FI >0.26 (Table 3). The four armorhead with FI <0.26 and un- infected by mature M. macro- pharynx had FI values disjunct from similarly un-infected armor- head with FI >0.26 (Fig. 2). Mean FI's of the seamount armorhead group (FI >0.26) without mature M. macropharynx (x = 0.291) and the epipelagic group (x = 0.307) were significantly different (i-test, P < 0.001). From samples of sea- mount armorhead (FI >0.26) ex- amined for prevalence of mature stage M. macropharynx, the per- centage of those found without mature M. macropharynx was higher in the June-August samples (Table 3). However, these latter results did not always cor- respond with peaks in the pro- portion of the seamount armorhead population (with FI >0.26) obtained from stock assess- Table 3 Percent absence of mature stages of the monogenean Mierocotyle macropharynx from the gills of ar- morhead Pseudopentaceros wheeleri by fatness index ( FI ) groups of <0.26 and >0.26, sampled from Southeast Hancock Seamount during 1985-88. Sample size examined for each cat- egory denoted by ( 71 1. Sample period FI <0.26 FI >0.26 Jun 1985 4.1 55.0 77=49 77=20 Aug 1986 0 33.3 7i=57 7i=48 Oct 1986 0 3.8 77=80 77 = 106 Apr 1987 0 3.7 77 = 137 7i=27 Aug 1987 0 10.4 71=68 77=48 Jan 1988 0 0 n = 183 71 = 12 Jul 1988 0 19.2 71=45 77=52 Aug 1988 0 21.2 u=47 77=52 Nov 1988 2.3 2.5 7i=87 ii=79 460 Fishery Bulletin 91(3). 1993 c 0.26 (x = 1.358, Mann-Whitney test, P < 0.001) and those with FI <0.26 (x = 1.028, t- test, P < 0.001). Furthermore, none of the seamount armorhead except for two individuals in the FI >0.26 group had HSI values exceeding the lowest value (2.154) of the epipelagic group. Differences between the mean HSI of the two seamount FI groups were 100 c u CL 40 20 | Stock assessment estimate of percent of SE Hancock armorhead population with Fl>0.26 ^\ Percent of armorhead sample (Fl>0.26) found un-infected with mature stage Microcotyle macro-pharynx I l^n. fir JUN AUG OCT APR AUG JAN JUL AUG NOV 1985 19B6 1986 1987 1987 1988 1988 1988 1988 Sampling period Figure 3 Percent of armorhead Pseudopentaceros wheeleri samples (fat- ness index (FI) >0.26> without mature Microcotyle macropharynx (from Table 3), and the percent of SE Hancock armorhead population with FI >0.26; the latter estimated from stock assessment surveys during each of the periods (except November 1988 1 where armorhead were sampled for M. macropharynx infection. (D XJ c o V) o D Q_ CO I 1 -Epipelagic 2-Seamount. without mature Microcotyle macropharynx 3-Seamount, Fl>0.26 4 — Seamount. with mature Microcotyle macropharynx 5-Seamount. FK0 26 = 27 T _L n=1( :167 n = 243 12 3 4 5 Armorhead groupings Figure 4 Hepatosomatic index by each grouping of armorhead Pseudopentaceros wheeleri. The mean is indicated by the mid- point horizontal line within each box, one standard deviation about each mean is denoted by the box, and the range is displayed as the horizontal lines above and below the box. also highly significant (Mann-Whitney test, P < 0.001). Regardless of FI value, the difference in mean HSI between epipelagics and seamount armorhead without (.r = 1.744, t-test, P < 0.001) and with (x = 1.148, Mest, P < 0.001) mature M. macropharynx remained significantly different, as were differences between the latter two groups (/-test, P < 0.001). The seamount armorhead group without mature M. macropharynx exhibited the least overlap in HSI values with other seamount groups (Fig. 4). For the epipelagic group, mean VFSI (x = 2.152; Fig. 5) was significantly different from seamount armorhead with FI >0.26 (x = 3.580, Mest, P < 0.001) and those with FI <0.26 (x = 2.933, Mann-Whitney test, P = 0.007). Mean VFSI differences between the latter two seamount groups were also significant (Mann-Whitney test, P < 0.001). The mean VFSI of the epipelagic group was also significantly different from that of seamount armorhead with (x = 3.185, Mann-Whitney test, P < 0.001) and without (x = 3.663, t-test, P = 0.009) mature M. macropharynx, whereas the latter two seamount groups did not significantly differ in mean VFSI (Mann-Whitney, P = 0.364). Over- lap in VFSI among the various groups (Fig. 5) was substantially greater than exhibited for HSI. Discussion The complete absence of the monogenean M. macropharynx from all epipelagic armorhead exam- Humphreys et al.: Identification of Pseudopentaceros wheelen "recruits 461 c □ E o (0 I 0) o en > 1 -Epipelagic 2-Seamount, without mature Microcotyle Tnacropharynx 3-Seamount. Fl^j0.26 4-Seamount, with mature Aficrocotyle macropharynx 5-Seamount. FK0.26 n=400 n=243 = 167 T JL 12 3 4 5 Armorhead groupings Figure 5 Visceral fat-somatic index by each grouping of armorhead Pseudopentaceros wheeleri. The mean is indicated by the mid- point horizontal line within each box, one standard deviation about each mean is denoted by the box, and the range is displayed as the horizontal lines above and below the box. ined, and its high prevalence among seamount armorhead during all sample periods, support the no- tion that infestation of this parasite originates at the seamount. The prevalence of immature M. macro- pharynx, particularly for those armorhead with FI <0.26, indicates that infestation at the seamount is probably a continuing process. The presence of M. macropharynx in all 775 seamount armorhead with FI of <0.26 also negates the likelihood of a sizable propor- tion of seamount armorhead becoming uninfested over time. The maturation rate of M. macropharynx on armorhead remains undetermined, but Thoney (1986) has found the oncomiracidia of the morphologically similar M. sebastis capable of maturing within 12 days on the gills of black rockhsh Sebastes melanops (Girard) off northern California. Although armorhead are dif- ferent hosts, inhabit greater depths, and are geographi- cally distant from black rockfish, the temperature range (12°-17°C) used by Thoney (1986) to determine the maturation time of M. sebastis is similar to that found on the summit and upper slope of SE Hancock Sea- mount. Assuming this rapid maturation also occurs among M. macropharynx infecting armorhead, we pro- pose that armorhead with exclusively immature M. macropharyx have probably only recently become in- fested and have seamount residence times that differ little from those of uninfested armorhead. The 0.26 FI value may represent a lower FI limit for potential new recruits, for only 6% of seamount armorhead without mature M. macropharynx had FI values <0.26 and the lowest FI value of the epipelagics examined was 0.26. However, an unknown proportion of armorhead with FI >0.26 perhaps recruit during the previous year at sufficiently high FI values that these FI values still remain above 0.26 when sampled the following year. Even in this situation, such individuals would be in- fected with mature M. macropharynx by the time they are sampled the following year and therefore not mis- taken for new recruits. The presence of M. macro- pharynx probably cannot identify recruits that arrived one or several months before sampling if initial infec- tion occurs soon after recruitment and M. macro- pharynx mature some two weeks thereafter. Hence, our technique can detect only new recruits among the recruit-of-the-year population and is therefore limited to determining whether recruitment is a continuous process. Of the two condition indices, VFSI shows little prom- ise as an indicator of new recruits, for the VFSI values in epipelagic and seamount individuals broadly over- lap, regardless of FI and the presence or absence of mature M. macropharynx. Conversely, virtually no over- lap in HSI values occurs between epipelagic individu- als and any of the seamount armorhead groupings. Preliminary results of sagittal otolith increment enu- meration on epipelagic individuals (>27.5cm FL) and SE Hancock armorhead without mature M. macro- pharynx indicates a common age around 2.5 years (R. Humphreys, unpubl. data). This appears to sug- gest that those epipelagic individuals examined either remain pelagic or represent "strays" who will eventu- ally or have begun a delayed movement that will re- sult in recruitment to a SE-NHR seamount but a pro- tracted epipelagic phase. The higher mean HSI value of these epipelagics versus seamount armorhead (FI >0.26) un-infected by mature M. macropharynx sug- gests that HSI decreases somewhere during the period between movement to the seamounts and time of cap- ture after seamount recruitment. To explain the higher mean HSI of epipelagics and evaluate its efficacy as a recruitment indicator, this forementioned period must be examined in relation to both time and distance. Of the 27 epipelagic individuals with FLs com- parable to seamount populations (>27.5cm FL) and least distant from the SE-NHR seamounts, 22 were collected around lat. 45°N, long. 155°W; some 2,860 km northeast of the SE-NHR seamounts. Summertime cap- tures of smaller (22-26 cm FL) epipelagics are also frequent from the above general location and indicate an age of some 1.5 years (Boehlert and Sasaki, 1988; R. Humphreys, unpubl. data). Since recruit-of-the-year seamount armorhead are age 2.5 years (R. Humphreys, unpubl. data), these smaller individuals apparently rep- resent the year class which recruit to the SE-NHR 462 Fishery Bulletin 91 [3). 1993 seamounts in the following year. Allowing a one-year period (recruitment by the following summer) for indi- viduals of the 1+ epipelagic year-class to move from the open ocean site (45°N 155°W) to the SE-NHR sea- mounts and also that the total distance traversed is twice the straight-line distance between these sites (assuming this is comparable to results of open-sea tracking of coho salmon Oncorhynchus kisutch reported in Ogura and Ishida (1992)), a rough estimate of the average ground speed required is 18 cm/s if currents are disregarded. Based on a length (27.1cm FL) mid- way between the average FL of 1+ (x = 24.49 cm FL) and 2+ year-old (x = 29.64cm FL) epipelagic speci- mens collected during summers at the open ocean site, individuals would need to swim 0.66 body length/s to arrive at an SE-NHR seamount one year later. This speed is close to the mean sustained cruising speeds (0.2-0.5 body length/second) typical of horizontally mi- grating fish (Beamish, 1978). Since no evidence exists of gonadal maturation in the epipelagic stage (Humphreys et al., 1989), the major energy demand during seamount migration is likely locomotion, in ad- dition to basal metabolism. It is unknown whether feeding occurs during this migration; however, McKeown (1984) suggests that even if fish species feed during migration, the added energy intake may be ne- gated by the additional energy required for feeding activity. Regardless of feeding, lipid reserves are typi- cally used during migration (McKeown, 1984), and their storage in the liver and muscle makes these prime sites from which energy reserves can be drawn (Woodhead, 1975). As such, the liver is susceptible to net decreases in weight by the end of migration; an example of this occurs among migrating Barents Sea cod Gadus morhua (Woodhead, 1975). Evidence of fewer prey contents among higher FI armorhead examined at SE Hancock3 indicates that recent recruits initially may feed little at the seamounts. This suggests that the liver remains susceptible to continued energy deple- tion and therefore continued decrease in liver weight. Assuming that liver weight is declining relatively faster than body weight, the HSI will not only decrease by the time of initial recruitment to the seamount but perhaps also during the interval between initial re- cruitment and sampling. Therefore, an annual or sea- sonal change in mean HSI among new recruits may be caused by inter- or intra-annual differences experienced by epipelagics in terms of environmental conditions and distance traversed during the seamount migra- tion, along with an unknown time gap between initial recruitment and sampling. Hence, using HSI as a re- cruitment indicator appears to be inherently more 'M. P. Seki, NMFS Honolulu Laboratory, pers. commun. January 1992. labile and thereby less feasible and reliable than the method based on the absence of mature M. macropharynx. The parasite approach revealed that new recruit- ment coincided primarily with the late spring and mid- summer sampling periods. Somerton and Kikkawa (1992) have analyzed the recruitment patterns of armorhead to SE Hancock, using the modal progres- sion of FI in which separate modes are considered an- nual cohorts. If a modal influx of high FI fish repre- sents a "recruitment cohort," estimated recruitment was highest in terms of biomass (58 1) and percentage (79%) of the SE Hancock population during August and October 1986, whereas the next highest recruit- ment (8t and 21% of the population, respectively) in July and August 1988 was comparatively modest. In- terestingly, however, the level of new recruitment, as determined by the percentage of sampled armorhead (FI >0.26) un-infected with mature M. macropharynx, was highest in June 1985 and very low in October 1986. These results appear contrary to stock assess- ment based estimates of modest and high proportions of recruit-of-the-year armorhead (FI >0.26) in the sea- mount population during June 1985 and October 1986, respectively (see Fig. 2). These results can be recon- ciled if we consider that the rate of post-recruitment decline in FI (estimated at 0.00169/month; Somerton and Kikkawa (1992)) does not allow one to decipher between recruits which may have just arrived and those which arrived months prior to sampling. The absence of mature stage M. macropharynx appears to be the only criteria examined in this study capable of detect- ing whether a recruit-of-the-year (FI >0.26) is a new arrival to the seamount. Based on this criteria, the level of new recruitment during the months sampled was highest in June, less so in July and August, and very low during October, November, January, and April. These results tend to corroborate evidence of similar seasonal armorhead recruitment at other SE-NHR sea- mounts (Boehlert and Sasaki, 1988). Recruits identified via the parasite approach typi- cally have FI values >0.26 (x = 0.291). However, a high FI value alone provides no assurance that an indi- vidual is a new recruit. The difference in mean FI between epipelagic individuals and new recruits (un- infected with mature M. macropharynx) may indicate that FI decreases sometime before epipelagics of re- cruitment size reach the seamounts. This can be ad- equately examined only after both groups are further sampled. Acknowledgments We are indebted to the officers and crew of the NOAA ship Townsend Cromwell, the RV Oshoro Maru, and Humphreys et al.: Identification of Pseudopentaceros whee/eri recruits 463 the FV Tomi Maru No. 88 for their assistance in col- lecting armorhead samples at sea. We particularly thank B. Kikkawa for collecting armorhead gill samples during Townsend Cromwell cruises to SE Hancock Sea- mount. Identification of parasites and general guid- ance on their handling and preservation were kindly provided by R. Payne, M. Pritchard, and D. Thoney. G. Boehlert, M. Love, and D. Thoney reviewed the manu- script, and we thank them for their helpful comments and suggestions. Literature cited Beamish, F. W. H. 1978. Swimming capacity. In W. S. Hoar and D. J. Randall (eds.), Fish physiology, vol. VII (Locomotion), p. 101-187. Academic Press, New York. Bilim, L. A., L. A. Borets, and L. K. Platosina. 1978. Characteristics of ovogenesis and spawning of the boarfish in the region of the Hawaiian Islands. In Fisheries oceanography, hydrobiology, biology of fishes and other denizens of the Pacific Ocean, p. 51-57. Izv. Pac. Ocean Sci. Res. Inst. Fish. Oceanogr. (TINROi, Vladivostok, 102. (Engl, transl. by W. G. Van Campen, 1986, 9p., Transl. No. 106; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396. ) Boehlert, G. W., and T. Sasaki. 1988. Pelagic biogeography of the armorhead, Pseudo- pentaceros wheeleri, and recruitment to isolated sea- mounts in the North Pacific Ocean. Fish. Bull. 86:453-465. Borets, L. A. 1979. The population structure of the boarfish, Pen- taceros richardsoni, from the Emperor Seamounts and the Hawaiian Ridge. J. Ichthyol. 19:15-20. Fedosova, R. A. and O. E. Komrakov. 1975. Feeding of Pentaceros richardsoni frys in the Hawaiian region. [In Russ.] Invest. Biol. Fishes Fish. Oceanogr., TINRO, Vladivostok 6:52-55. (Engl. transl. by W. G. Van Campen, 1985, 4 p., Transl. No. 100; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv, NOAA, Honolulu, HI 96822- 2396. ) Humphreys, R. L. Jr., D. T. Tagami, and M. P. Seki. 1984. Seamount fishery resources within the southern Emperor-northern Hawaiian Ridge area. In R. W. Grigg and K. T. Tanoue (eds.), Proceedings of the Sym- posium on Resource Investigations in the Northwest- ern Hawaiian Islands, Vol. 1, May 25-27, 1983, University of Hawaii, Honolulu, HI, p. 283-327. UNIHI-SE AGRANT-MR-84-0 1 . Humphreys, R. L. Jr., G. Winans, and D. T. Tagami. 1989. Synonymy and life history of the North Pacific pelagic armorhead, Pseudopentaceros wheeleri Hardy (Pisces: Pentacerotidae). Copeia 1:142-153. Kabata, Z., G. A. McFarlane, and D. J. Whitaker. 1988. Trematoda of sablefish, Anoplopoma fimbria (Pallas, 1811), as possible biological tags for stock identification. Can. J. Zool. 66( 1 ): 195-200. Komrakov, O. E. 1970. Distribution and fishery of the boarfish (Pen- taceros richardsoni Smith) in the Hawaiian region. In the collection: The present state of biological pro- ductivity and the volume of biological resources of the world ocean and prospects for their utilization, p. 155-163. Kaliningrad. (Engl, transl. by W. G. Van Campen, 1987, 9 p., Transl. No. 117; available South- west Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396.) Mamaev, Yu. L. 1989. On species composition and morphological fea- tures of the Microcotyle genus (Microcotylidae, Monogenoideai. [In Russ.] Investigations in Parasi- tology, Collection of Papers, p. 32-38. Margolis, L., G. W. Esch, J. C. Holmes, A. M. Kuris, and G. A. Schad. 1982. The use of ecological terms in parasitology (re- port of an ad hoc committee of the American Society of Parasitologists ). J. Parasitol. 68: 13 1-133. McKeown, B. A. 1984. Fish migration, Chapter 4 (Bioenergetics). Croom Helm, London. Ogura, M., and Y. Ishida. 1992. Swimming behavior of coho salmon, Oncor- hynchus kisutch, in the open sea as determined by ultrasonic telemetry. Can. J. Fish. Aquat. Sci. 49:453- 457. Sindermann, C. J. 1983. Parasites as natural tags for marine fish: a review. Northwest Atl. Fish. Organ. Sci. Counc. Stud. 6:63-71. Somerton, D. A., and B. S. Kikkawa. 1992. Population dynamics of pelagic armorhead Pseudopentaceros wheeleri on the Hancock Sea- mounts. Fish. Bull. 90:756-769. Thoney, D. A. 1986. Post-larval growth of Mwrocotyle sebastis (Platy- helminthes: Monogenea), a gill parasite of the black rockfish. Trans. Am. Microsc. Soc. 105:170-181. Uchida, R. N., and D. T. Tagami. 1984. Groundfish fisheries and research in the vicin- ity of seamounts in the North Pacific Ocean. Mar. Fish. Rev. 46(2):1-17. Wilkinson, L. 1990. SYSTAT: The system for statistics. Evanston. IL. SYSTAT, Inc. (Version 5.01). Williams, H. H., K. MacKenzie, and A. M. McCarthy. 1992. Parasites as biological indicators of the popula- tion biology, migrations, diet, and phylogenetics offish. Reviews in Fish Biology and Fisheries, 2:144-176. Woodhead, A. D. 1975. Endocrine physiology offish migration. Ocean- ogr. Mar. Biol. Annua. Rev. 13:287-382. Abstract.— Annual and seasonal variability of Georges Bank zoo- plankton biomass and dominant spe- cies abundance are described and related to variations in mean sur- face temperature and average depth distribution. Data were obtained from plankton samples collected bi- monthly with a 0.333-mm mesh net throughout a ten-year period: 1977- 86. Biomass was measured by dis- placement volume and the dominant species analyzed were the copepods Calanus finmarchicus, Pseudo- calanus minutus, Centropages typi- cus, Centropages hamatus, and Metridia lucens. Biomass levels were high in 1977 through 1979, low in 1982 through 1984. Biomass and copepod abun- dance in the spring of 1977 were ex- traordinary. Measurements over the entire bank were two to three times above a ten-year median. Unlike the first five years of monitoring, the av- erage seasonal biomass cycle was not coherent from 1982 through 1986. Departures from the average sea- sonal cycle occurred several times during the second half of the time series. Calanus finmarchicus and Pseudo- calanus minutus abundance trends were nearly identical, suggesting that their populations may be af- fected by similar factors. Centro- pages hamatus abundance in the central shoal depth zone (<61m) was related to surface temperature vari- ability and its spring abundance es- timates were indirectly proportional to the abundance of other dominant copepods. Centropages typicus counts in autumn 1985 were nearly double all other years, and Metridia lucens abundance surged in late spring 1979 but was low from 1983 through 1986. Variability of zooplankton biomass and dominant species abundance on Georges Bank, 1977-1986 Joseph Kane National Marine Fisheries Service. NOAA 28 Tarzwell Drive Narragansett. Rhode Island. 02882 Manuscript accepted 16 April 1993. Fishery Bulletin 91:464-474 1 1993). Zooplankton biomass has long been recognized as an important index for estimating the seasonal and annual variability of secondary production in marine ecosystems. Zooplankton play a key role in pelagic food chains, serv- ing as the connecting link between primary producers and secondary consumers. The availability of zoo- plankton as food for larval fish is thought to be one of the key factors determining year class strength of commercial fish species (Cushing, 1978). The rich fishing grounds of Georges Bank in the northwest Atlantic have been the focus of zooplankton studies since the turn of the century. The Ma- rine Resources Monitoring, Assess- ment, and Prediction (MARMAP) pro- gram (Sherman, 1980) has monitored the U.S. Northeastern continental shelf marine ecosystem from 1977 through 1987 with bimonthly surveys, measuring a variety of biological and physical parameters. During the first five years of MARMAP monitoring, zooplankton biomass on Georges Bank formed a coherent seasonal pattern, not changing significantly from year to year (Sherman et al., 1983). Com- parison of this data to that collected by Bigelow (1926) from 1912 to 1920 showed that biomass levels, species composition, and abundance estimates of dominant species were essentially the same in both studies. Sherman et al. (1987) described in greater detail the seasonal cycle of Georges Bank zooplankton and how it relates to ichthyoplankton life histories. Addi- tional studies (Davis, 1984; Meise- Munns et al., 1990) of Georges Bank zooplankton have used subsets of this large data base to help define and simulate seasonal cycles of dominant species in relation to environmental parameters. The purpose of this paper is to fur- ther describe the Georges Bank zoo- plankton community by utilizing data collected during MARMAP surveys from 1977 to 1986. The annual and seasonal variability of zooplankton biomass captured with 0.333-mm mesh nets is reported and related to changes in the abundance of the five dominant zooplankton species (Sher- man et al., 1987). The average depth distribution of biomass and dominant species abundance is described and departures from it are compared to overall population variability. The sensitivity of the above param- eters to surface water temperature readings is also examined to consider the potential effects of climatic change on Georges Bank zooplank- ton populations. This study is part of a continuing long-term investigation by the National Marine Fisheries Service (NMFS), which monitors the zooplankton component of the U.S. Northeast shelf ecosystem. Methods Plankton samples were collected at monthly to bi-monthly intervals at 32 station locations during MARMAP surveys on Georges Bank (Fig. 1). 464 Kane. Zooplankton biomass and species abundance on Georges Bank 465 r~*l- i / Depth area <61m 61-100m >100m « _ / ♦ ^ /y- Figure 1 Location of standard MARMAP (Marine Resources Monitoring, Assessment, and Prediction program) stations on Georges Bank off the U.S. Northeast coast, 1977-86. Plankton samples were also collected on trawl and dredge surveys at randomly selected locations that changed yearly. Areal coverage and sampling spacing on these surveys were similar to plankton cruises. Samples from different surveys, closely overlapping in time and space, were sometimes combined to ensure adequate coverage of the survey area. Zooplankton was collected with a 61-cm bongo fitted with a 0.333-mm mesh net towed obliquely to a maxi- mum depth of 200 m or 5 m from the bottom and back to the surface. Ship speed varied between 1 and 2 knots to maintain a 45 degree wire angle. Winter sur- veys in 1977 and 1978 towed bongos at 3.5 knots. A flowmeter was positioned in the center of the bongo frame to measure volume of water filtered during the tow. Samples were preserved in 5% formalin. At all stations, sea-surface temperature was measured with a stem thermometer to the nearest 0.1°C. Detailed sampling procedures, cruise tracks, and survey logis- tics are summarized by Sibunka and Silverman ( 1984, 1989). Biomass was measured by displacement volume in the laboratory. Initially, organisms larger than 2.5 cm were removed. The plankton sample with preserving liquid was then measured in a graduated cylinder, poured through a mesh cone into a second cylinder, and drained until the interval between drops from the cone increased to 15 seconds. The liquid in the second cylinder was measured and the displacement volume of the sample was the difference between readings. Samples with high concentrations of gelatinous organ- isms were eliminated because the interstitial water retained by these animals leads to gross overestimates of zooplankton biomass. Samples were later subsampled by aliquoting to about 500 organisms and identified to species. Volumes («=1937) are expressed as cc/100m:) of water filtered, and abundance (rc=1839) as number/ 100 m3. The adults and late stage copepodites of the cope- pods Calanus finmarchicus, Pseiiducalanus minutus, Centropages typicus, Centropages hamatus, and Metridia lucens were the dominant species analyzed. Depth distribution of biomass and species abundance was examined by subsetting the data into three geo- graphic subareas according to bottom depth: 1) central shoal (<61m); 2) intermediate (61-100 m); and 3) deep (>100m). Seasonal shifts in biomass and community structure were investigated by grouping the data into the six seasons defined in Table 1. Extenuating cir- cumstances prevented adequate areal coverage in only one season: winter 1979. The Shapiro- Wilk test for normality was applied to each seasonal biomass and species data set. The null hypothesis that the data values were a random sample from a normal distribution was rejected (P<0.01). Zoo- plankton data are often log transformed to normalize zooplankton distributions. However, Roesler and Chelton ( 1987 ) found that transformation of zooplank- 466 Fishery Bulletin 91(3). 1993 Table 1 Seasonal median values of biomass an d dominant species abundance by year and for all years combined during A) Winter, B) Early Spring, C ) Late Spring, D) £ ummer. E) Early Autumn, and F) Late Aut umn. Cer tropages hamatus abundance numbers are only from the central shoal depth area and Metridia lucens ones are restricted to the deep water depth area. Midpoint refers to the median day number of that years sampling. A ' — ' indicates no data were collected or processed. A: Winter (1 Jan.- 23 Mar.) Year 77 78 79 80 81 82 83 84 85 86 All data Midpoint 51 61 ' — 62 64 66 20 18 17 34 48 Biomass 7 20 9 10 9 10 12 8 9 9 C. finmarchicus 1170 1523 — 2814 5539 1531 508 1191 639 1023 1258 P. minutus 1419 8678 — 1746 4814 3157 1131 1685 1701 954 2120 C. typicus 10 263 — 1253 145 2648 2988 1735 3720 2057 976 C. hamatus 0 77 — 1626 0 16 5709 1770 3021 10379 188 M. lucens 121 884 — 154 4334 1422 766 1112 340 49 444 B: Early spring (24 Mar.-4 May) Year 77 78 79 80 81 82 83 84 85 86 All data Midpoint 115 105 101 102 112 116 103 95 92 104 101 Biomass 86 50 66 43 43 29 36 14 54 76 42 C. finmarchicus 44912 — 22680 15659 36957 27511 17600 5192 6729 19261 18100 P. minutus 19299 — 4483 7217 17032 17098 10654 3339 2449 4600 7217 C. typicus 0 — 1031 353 261 1300 89 34 108 182 207 C. hamatus 306 — 65 3889 549 115 4538 16 1702 10722 569 M. lucens 1666 — 5117 5210 4164 4373 213 1645 567 468 1621 C: Late spring (5 May-22 June) Year 77 78 79 80 81 82 83 84 85 86 All data Midpoint 157 140 144 168 152 141 168 149 133 153 147 Biomass 137 89 95 42 50 37 15 38 46 21 50 C. finmarchicus 89437 59030 22748 15895 12429 20887 4590 15622 11221 9088 18255 P. minutus 43677 19034 19467 13727 19595 10765 2742 6853 4597 2833 9313 C. typicus 0 0 511 829 356 1274 0 0 0 9 51 C. hamatus 453 0 201 21188 21065 697 26190 13910 34157 7476 2657 M. lucens 8612 9147 24391 9963 2728 13486 117 1145 823 409 2568 D: Summer (23 June-12 Sept.) Year 77 78 79 80 81 82 83 84 85 86 All data Midpoint 230 206 217 213 198 206 233 233 230 238 202 Biomass 43 37 43 33 46 15 25 25 16 25 31 C. finmarchicus 21856 5346 4737 4556 4667 875 5029 4693 2913 3473 4042 P. minutus 12205 8109 4655 5661 12908 5175 6328 1980 4354 7704 5776 C. typicus 7561 6739 11201 3416 4264 1678 22659 8954 6113 11225 6456 C. hamatus 9958 66547 89476 33095 97265 39490 27922 63458 67507 40335 46140 M. lucens 4040 635 12195 1879 1315 6280 1089 390 9 78 1597 E: Early Autumn Year (13 Sept- 77 -9 Nov.) 78 79 80 81 82 83 84 85 86 All data Midpoint 308 289 297 293 295 302 292 289 297 289 294 Biomass 33 30 30 33 22 19 15 22 32 9 24 C. finmarchicus 638 912 591 282 679 1364 429 454 971 347 587 P. minutus 5986 3472 2284 39 1285 3089 291 220 475 536 898 C. typicus 35103 33342 39431 39256 36453 11784 14322 18970 71184 5124 28431 C. hamatus 25608 17018 49178 10833 13524 14138 2024 17480 15068 5765 14138 M. lucens 2180 1020 4332 1895 3227 49 257 71 27 28 540 F: Late Autumn (10 Nov.-31 Dec.) Year 77 78 79 80 81 82 83 84 85 86 All data Midpoint 327 330 337 350 337 325 344 333 338 339 334 Biomass 22 23 24 11 13 27 7 12 18 6 16 C. finmarchicus 1186 573 133 683 419 1137 812 711 538 78 539 P. minutus 13752 7922 942 1826 799 2169 540 187 198 388 1030 < ' t xpicus 17960 33013 24117 9200 21241 15835 9209 15646 50081 2594 16378 ( hamatus 4739 1691 23626 345 3999 21416 6057 24439 45741 10759 5523 M. lucens 1742 1570 7058 4773 4503 870 243 385 263 12 735 Kane: Zooplankton biomass and species abundance on Georges Bank 467 ton biomass can screen important biological events by repressing anomalous values. Therefore, nonparamet- ric statistical techniques were employed to test for dif- ferences between years within the defined seasons. The Kruskal-Wallis ANOVA was used to determine if sig- nificant (P<0.05) differences existed between years within each season. The test showed that all biomass and species seasonal groupings had at least one pair of years different from each other. The Dunns multiple comparison procedure was applied to pinpoint the anomalous year(s). Results Annual cycle Median biomass and dominant copepod abundance val- ues for each of the defined seasons by year and for the ten-year study period are given in Table 1, A-F. Biomass on average increased fourfold in early spring from its winter low. It peaked in late spring and then gradually declined through the summer and autumn seasons (Fig. 2). Overall, zooplankton standing stock CO E E o CD A o / \ -o- 77 / \ -+" 78 / A \ -4- 79 '/ 1 - \ \ -a- 77-86 *'" \\ \ r °^^1A 50 100 150 200 250 300 350 Survey Mid Point (jday) CO E B -O- 82 - +- 83 ^S*\ -a- 84 **~~3 / » \ ~°~ 77-86 / // \ ^ — _ -/ // \ \ >^rr-n p / "* - ^ ° 0 50 100 150 200 250 300 350 Survey Mid Point (jday) Figure 2 (A) Annual cycle of seasonal median biomass values in 1977 through 1979 and the ten-year median. (B) Annual cycle of seasonal median biomass values in 1982 through 1984 and the ten-year median. was high from 1977 through 1979. Seasonal medians from early spring through late autumn were above time-series median values (Fig. 2A). The opposite pat- tern was evident from 1982 through 1984. Zooplank- ton standing stock was below average throughout these years, except for late autumn 1982 (Fig. 2B). Winter biomass showed little variation throughout the study period (Table 1A). The seasonal biomass cycle on Georges Bank from 1977 through 1981 was coherent; annual departures from the mean were insignificant (Sherman et al., 1983). However, substantial departures from the aver- age seasonal cycle occurred on several occasions after 1981. Surveys in 1983, 1985, and 1986 recorded peak biomass in early spring instead of late spring and in two of those years, 1983 and 1986, there was an anoma- lous increase in biomass during the summer months (Fig. 3). However, because only below average summer levels were reached during these years, the latter in- creases appear to represent a recovery from the mini- mal levels measured in late spring, rather than a sum- mer bloom. Zooplankton standing stock increased between summer and early autumn in only two years: 1982 and 1985 (Table 1, D-E). The usual decline in biomass between early and late autumn was not ob- served in 1982 (Table 1, E-F). The average ten-year seasonal distribution of each variable as a function of bottom depth is depicted in Figure 4. There were no apparent long-term distribu- tion shifts in biomass or species abundance during the ten-year period. Calanus finmarchicus and P. minutus both domi- nated the zooplankton community in early and late spring. During the summer, their populations declined and C. typicus began to increase in abundance until it peaked in early autumn (Table 1). Centropages hamatus and M. lucens were only prevalent in specific depth strata (Fig. 4). Centropages hamatus was almost en- tirely restricted to the central shoal depth region, peak- ing there during summer months. Metridia lucens was most numerous in deep water where its large size caused it to be a major contributor to spring biomass. Overall, it was the only dominant species that exhib- ited a long-term abundance trend. Population estimates for M. lucens were low in 1983 through 1986 (Table 1). All the above species showed departures from their average seasonal cycles during their periods of peak abundance. Calanus finmarchicus, M. lucens, and P. minutus all declined between early and late spring in 1983 and 1986 (Table 1, B-C). Centropages typicus departed from its typical annual abundance pattern during three years; population estimates in 1983 and 1986 declined from summer to early autumn and in 1982, C. typicus did not reach peak abundance until late autumn. The C. hamatus population was more vari- able than the other species. From 1977 to 1981, abun- 468 Fishery Bulletin 9 1(3). 1993 CO E o o 100 80 Early Spring □ Surr 1977-86 1983 1985 Year(s) Figure 3 Departures from the ten-year (1977-86) median biomass cycle occurred between early spring and summer in 1983, 1985, and 1986. out the winter time series, C. hamatus was virtually absent at stations where surface temperature was below 5°C (Fig. 6). Median surface temperatures for the years 1983-86 were all above 5°C (Table 2) and significantly higher (P<0.05) than the earlier years, which were all sampled later in the season when the annual minimum tempera- ture on Georges Bank is normally reached. Biomass and the abundance of the other four copepod species were not affected by winter temperature regimes. dance decreased from early to late autumn. During the next five years, increases were recorded between seasons (Table 1, E-F). After an unusual pulse in early spring 1986 (Table IB), C. hamatus declined in late spring and then rebounded to their average summer abundance. In one year, 1977, its annual high was delayed until early autumn (Table 1, D-E). Surface temperature variability was examined by fitting seasonal medians (Table 2) to a harmonic re- gression (Fig. 5). The model depicted a strong annual cycle; r = 0.91. Certain seasons in specific years had above or below average temperatures, but there were no prolonged warm or cool periods. Interannual variability by season Winter Zooplankton standing stock reaches its an- nual low and exhibits little interannual variation dur- ing the cold winter months (Table 1A). The only re- markable year was 1978, when median biomass was nearly double all other years. However, this may have been an artifact of survey logistics, rather than en- hanced winter productivity. None of the usually low biomass stations along the northern and southern pe- rimeter of Georges Bank were sampled that year. Centropages hamatus was the only dominant species that demonstrated substantial interannual variation during the winter season. Population estimates in the central shoal depth zone from 1983 to 1986 were well above the ten-year median (Table 1A). These depar- tures can be directly related to surface-temperature variability caused by dates of survey coverage. Through- Early spring Zooplankton biomass in early spring was high at the begin- ning and end of the time series; 1977 and 1986 (Table IB). Standing stock in 1984 was very low; significantly less (P<0.10) than all other years except 1982. C. finmarchicus and P. minutus abundance (Table IB) was similar to biomass trends in some years. Like biomass, abundance estimates for both species were highest in 1977 and low in 1984. In 1986, an unusual early spring pulse of C. hamatus (Table IB) in the central shoals depth area elevated standing stock levels. The above average biomass lev- els recorded in 1985 could not be related to the low abundance estimates of the dominant species in that year (Table IB). Notations made by shipboard person- nel indicated that nets were frequently clogged with dense concentrations of phytoplankton. High biomass in 1985 was likely elevated by phytoplankton and en- trapped organisms not usually captured with 0.333- mm mesh nets. As in winter, the C. hamatus population was de- pressed by cold temperatures in early spring. Their highest early spring abundance occurred in 1986 when surface waters in the central shoals depth area were warmest (median=6.5°C) in this season. Calanus finmarchicus were also more abundant at stations where surface temperatures were 6°C or more. An- nual changes in biomass and the abundance of other dominant species could not be related to surface tem- perature variability in early spring (Table 2). Late spring Zooplankton biomass surged in the late spring of 1977 (Table 1C). The median volume was nearly three times higher than the seasonal 10-year median, and significantly different (P<0.05) from all years except 1978 and 1979. Though biomass declined from this peak in 1978 and 1979, median estimates in these two years were nearly double those recorded in Kane: Zooplankton biomass and species abundance on Georges Bank 469 <61m Biomass P. minutus M. lucens BO BO 40 20 - ■M_- .•^^flH 61 -100m >101m C. finmarchicus L>. C. typicus C. hamatus Season Season Figure 4 The average seasonal distribution of biomass and the five dominant species as a function of depth during the ten year study period. Season abbreviation key: W = Winter, ES = Early Spring, LS = Late Spring, S = Summer. EA = Early Autumn, and LA = Late Autumn. the 1980's. Extremely low estimates were recorded in 1983. Interannual fluctuations in late spring biomass were closely related to C. finmarchicus and P. minutus abun- dance (Table 1C). The similarities between abundance plots of both species during spring seasons (Fig. 7) suggest that their population dynamics may be con- trolled by similar factors. 470 Fishery Bulletin 9 1(3), 1993 Table 2 Seasonal median sea surface temperature (°C) by year and for all years combined. Mo data were collected in Winter 1979. Season Yeai All data 77 78 79 80 81 82 83 84 85 86 Winter 4.9 4.6 — 4.8 4.2 3.4 7 5.7 6.9 5.7 5.2 Early spring 6.6 4.7 4.7 5.7 5.6 5.2 6.2 5.2 5.4 6 5.5 Late spring 9.6 7.4 8.5 11.6 10.2 7.2 10.4 8.9 7.9 9.2 9.1 Summer 16.4 14.7 16.7 18.1 15.0 15.6 16.1 18.2 16.8 15.3 16.1 Early autumn 14.0 13.5 14.4 14.9 12.5 12.9 14.3 14.1 13.8 13.6 13.6 Late autumn 10.6 11.6 10.4 7.4 9.9 11.5 9.2 10.8 9.8 10.6 10.2 There were two notable annual distribution shifts in late spring. Calanus finmarchicus abundance was usually sparse within the 60-m contour and evenly distributed offshore of it (Fig. 4). However, in the high biomass years of the late 1970s, the population thrived in central shoal waters. From 1977 to 1979, C. fin- marchicus medians were 48,366, 33,397,and 15,364/ 100 m3 respectively, while other years were all below 5000/100 m3. Metridia lucens abundance in deep water peaked in 1979 (Table 3C) and they extended their range of dominance inshore across the 100-m contour line. Its high abundance (median = 22,015/100 m3) in the intermediate depth zone, where its 10-year me- dian was 365/100 m3, elevated biomass there to nearly double the time-series median, despite only average C. finmarchicus abundance. The M. lucens abundance peak in 1979 accounts for the high biomass measured in that year. Centropages hamatus was the dominant copepod spe- cies in the central shoals area during 1980 and 1981 observed predicted o Q. E a> h- c : SA = Sim0.0172*sampling midpoint(jday)!. and from 1983 to 1986. In other years, their numbers were sparse there; median values fell below 1000/100 m3 (Table 1C). Their high years correlated with all but one (1981 — P. minutus) of the low abundance years for C. finmarchicus and P. minutus (Fig. 8). The strongest relationship between abundance and surface temperature variability in late spring (Table 2) was for C. hamatus numbers in the central shoals area. Five of the six years, during which they domi- nated those waters, temperatures were warmer than average, and three of the four years in which they were sparse were below the ten-year median tempera- ture. Biomass and other dominant species abundance could not be related to interannual differences in me- dian surface temperatures. Summer The spring zooplankton biomass surge of the late 1970's continued through the summer months. Bio- mass levels in those years and in 1981 were signifi- cantly different (P<0.05) from the low measures ob- tained in 1982 and 1985 (Table ID). Centropages hamatus was prevalent in the well mixed central shoals, but Georges Bank as a whole was not dominated by one copepod species during summer months (Table ID). Biomass in 1977 was elevated when C. finmarchicus and P. minutus abun- dance reached near seasonal highs during the ten-year period. Notable in 1977 was that C. hamatus numbers in shallow water were minimal and C. finmarchicus abun- dance there (8386\100m3) was at its summer high, reinforcing the inverse abun- dance relationship observed between them in late spring. As it did in late spring, high M. lucens abundance outside the 60-m isobath raised biomass in 1979. Its abun- dance within the intermediate depth zone (4193\100m3) was well above its 10-year median there of 167M00 m3. Biomass in 1981 was raised when both P. minutus and C. hamatus abundance estimates reached Kane: Zooplarrkton biomass and species abundance on Georges Bank 471 8000 6000 4000 2000 4-4.9 5-5.9 Degrees (°C) 6-6.9 Figure 6 Ten-year median abundance of Centropages hamatus during the winter season at 1°C surface temperature intervals. Popu- lation estimates are from the central shoal depth area only. seasonal highs for the ten-year period. Calanus pin- marehicus, P. minutus, and C. typicus were all below average abundance in the low biomass years of 1982 and 1985. No strong correlation between median surface tem- peratures and biomass was evident during the sum- mer. Calanus finmarchicus abundance estimates were highest at stations where surface temperatures had warmed above 17°C, reflecting their shift to warmer waters outside the 100-m isobath in summer (Fig. 4). This relationship is probably related to the cooler water found below the seasonal thermocline (Manning and Holzwarth, 1990), rather than to the warm surface layer where this cold water species is unlikely to concen- trate. Abundance of other species was vari- able over the range of summer temperatures. Early autumn Four high biomass years, from 1977 to 1980, were followed by dimin- ishing ones through 1983 (Table 5E). Biomass began to climb in 1984 and again reached high levels in 1985. Standing stock fell to its lowest level in 1986. The depth distribution of the high biomass of 1980 differed substan- tially from average conditions (Fig. 4). Me- dian biomass from the intermediate (39cc/ 100 m;i) and deep (30.5cc/100m3) water zones P. minutus ■ C. finmarchicus Figure 7 Median abundance of Calanus finmarchicus and Pseudo- calanus minutus in 1977 through 1986 during (A) early spring and (B) late spring. E o s •D Figure 8 Median late spring abundance of Centropages hamatus, Calanus fin- marchicus, and Pseudocalanus minutus in the central shoal depth region in 1977 through 1986. 472 Fishery Bulletin 91(3), 1993 were both higher than those from the central shoals area (27cc/100ma), the region where biomass is usu- ally concentrated. Centropages typicus dominates the zooplankton popu- lation in early autumn; it makes up on the average 41% of total zooplankton abundance. Consequently, variation in its interannual abundance pattern nearly mirrored that of early autumn biomass. The only sub- stantial deviation between patterns occurred in 1985 when C. typicus density soared to its ten-year high (Table IE). Biomass was high in early autumn 1985, but not in proportion to this copepod's abundance. Centropages typicus abundance usually declined sub- stantially offshore of the 100-m contour (Fig. 4). How- ever, its abundance ( 10,805/100 m3) in the deep-water depth area in 1980 was substantially higher than the time series median ( 1319/100 m'). Consequently, off- shore biomass in 1980 (30.5cc/100m:i) was also well above the deep-water 10-year median (7cc/100m3). Though their numbers decline from summer, C. hamatus abundance continues to be a large compo- nent of zooplankton biomass in central shoal waters. Population estimates peaked in 1979 and, like bio- mass, were low in 1986 (Table IE). Early autumn de- partures from average annual or seasonal cycles of biomass and species abundance and distribution could not be related to variations in surface water tempera- ture (Table 2). Late autumn The high zooplankton biomass of the late 1970s continued through late autumn (Table IF). The seasonal peak recorded in 1982 was unexpected. Biomass increased from early autumn and pushed 1982 measurements above average for the first time. These high years were all significantly different (P<0.05) from the seasons lowest biomass measured in 1983 and 1986, the same years that were low in early autumn. Centropages typicus continue to dominate the zoo- plankton as biomass declined towards its winter low. Its abundance was above average in the high biomass years of the late 1970's (Table IF). The high biomass of 1982 was not related to total copepod or zooplank- ton numbers. Median counts of the five dominant cope- pod species and total zooplankton were not significantly different (P>0.05) from any of the other years. Cursory examination of other species abundance indicated that chaetognaths were prevalent and may have increased the biomass. The reverse occurred in 1985 when slightly above average biomass did not correlate to high zoo- plankton abundance. C. typicus and C. hamatus abun- dance in late autumn 1985 were three or more times above the ten-year median (Table IF). Warm late autumn temperatures appear to slow the decline of zooplankton biomass and abundance to the annual lows found in winter. High biomass years in late autumn all had above average surface tempera- tures (Table 2). The only high biomass levels of early autumn that were not sustained through late autumn were those collected in 1985, when surface tempera- ture had fallen below the ten-year median. Conse- quently, it is not surprising that biomass and domi- nant copepod abundance were highest at stations where surface temperature was warm. The bimodal annual cycle displayed by C. hamatus in the mixed depth zone, that is to say the decline in numbers between early and late autumn from 1977 to 1981 and the increase in subsequent years (Table 1, E-F), cannot be explained by variability in autumn surface temperatures. Discussion The zooplankton population on Georges Bank exhib- ited considerable interannual and seasonal variability during the period from 1977 to 1986. Overall, biomass was above average from 1977 through 1979 and low from 1982 through 1984. Unique to the late seventies was the high late spring abundance of C. finmarchicus in the central shoal depth zone. The species was only a minor component of the zooplankton community there in later years. Compared with other years, 1977 bio- mass and copepod abundance levels in spring were extraordinary: two to three times above the ten-year median. Calanus finmarchicus and P. minutus abundance fluctuations were nearly identical throughout the ten- year time series. This suggests that the annual abun- dance of these two species is regulated by similar pro- cesses and events, despite differences in their life cycles (Davis, 1987). This is in contrast to the results derived by Davis (1984) from model simulations of their sea- sonal cycles on Georges Bank. He concluded that pre- dation pressure alone controls P. minutus population levels and that C. finmarchicus is regulated by both predation and food availability. The data presented here indicates that its unlikely that these species have different factors limiting their annual abundance. Fur- thermore, preliminary studies indicate that abundance estimates of both species on Georges Bank are corre- lated to chlorophyll levels in the water column1. Inves- tigations on P. minutus population dynamics should not exclude food supply as a potential limiting factor. Centropages hamatus abundance estimates were more variable and temperature sensitive than those for other dominant copepod species. Of special interest is that C. hamatus spring abundance pulses were in- versely related to both C. finmarchicus and P. minutus population estimates. It is unlikely that temperatures which stimulate C. hamatus production would be det- rimental to C. finmarchicus or P. minutus. Laboratory 'C. Meise, National Marine Fisheries Service, Narragansett, RI 002882, unpubl. data. Kane. Zooplankton biomass and species abundance on Georges Bank 473 studies (Marshall and Orr, 1955; Corkett and McLaren, 1978) have shown that both C. finmarchicus and P. minutus grow and reproduce within the upper range of spring temperatures. There was no evidence in our data that either species had a strong response to spring surface temperature variability. One possibility is that the omnivorous C. hamatus may have depressed production in the other species by preying on their egg and naupliar stages, as Davis (1984) suggests they do in autumn months. Physical or behavioral responses to other changing spring conditions, such as daylight or thermocline formation, cannot be eliminated as poten- tial factors that triggered or limited their production. There is growing evidence that the earth's climate is changing (Mitchell, 1989). The impact of global warm- ing on Georges Bank zooplankton could be substantial because the area is a faunal transition zone between northern boreal species and southern warm water plankton (GLOBEC, 1991). The sensitivity of the Georges Bank C. hamatus population to temperature indicates that the effects of a long-term warming trend might first affect the life cycle of this species. Cen- tropages hamatus virtually disappears from the water column when surface temperature falls below 5°C. Marcus (1989) has shown it produces diapause eggs and reports that nauplii appeared after incubation of Georges Bank sediment. Centropages hamatus appar- ently overwinters as bottom resting eggs that hatch when water column temperatures rise to some thresh- old. If this phase of its life cycle is shortened or elimi- nated by global warming, growth and reproduction in the central shoals could continue year round and po- tentially limit the production of other copepod species. The early to late autumn rise of C. hamatus abun- dance from 1982 to 1986 suggests dormancy was post- poned and an additional generation produced in these years. Since monthly anomalies of sea surface tem- perature for 1981-86 in the area indicate a warming trend (Wood and Tang, 1988), this may be the first signal that climatic change is affecting the marine eco- system. Interannual variability in the amount of food avail- able to larval fishes is believed to be an important determinant of their survival and subsequent recruit- ment to adult populations. Field evidence for the link- age between larval survival and zooplankton prey con- centrations is poor (Laurence and Lough, 1984; Leak and Houde, 1987). Though this report was not designed to examine the relationship between zooplankton vari- ability and its effect on fishery resources on Georges Bank, it should be noted that zooplankton biomass patterns closely resembled those of northern sand lance [Ammodytes dubius) population estimates. Their popu- lation surged in the late seventies, responding to a reduction in predation and competition pressure caused by the fishery-induced collapse of herring and mack- erel populations (Sherman et al., 1981). Relative abun- dance of the zooplanktiverous sand lance in survey trawls increased dramatically between 1977 and 1981, decreasing thereafter through 1986 (Nelson, 1990). The sand lance explosion may have been fueled by the high concentrations of zooplankton food stocks available in the late 1970's. Interannual variations in time of sampling can bias estimates of biomass and abundance for predefined seasons. Early spring on Georges Bank is especially sensitive to this bias because the zooplankton popula- tion is beginning to harvest the late winter phytoplank- ton bloom and is rapidly transferring it to higher lev- els of the food chain. Obviously, the calendar definition of early spring used here to subset data may not be real in nature. The question arises whether the bio- mass estimates recorded in early spring 1984 were truly low or was sampling conducted too early? Me- dian surface temperature in 1984 was only 0.3°C be- low the ten-year median of 5.5°C. There were other years with lower temperatures that had higher esti- mates of biomass. A biological sign of spring's arrival on Georges Bank is the presence of early copepodite stages of C. finmarchicus in the water column. The spring phytoplankton bloom triggers their spawning and, if spawning had not yet occurred, only overwin- tering stage-5 copepodites would be present (Davis, 1987). However, in 1984, 52.7% of the population was stage-2 and stage-3 copepodites, similar to the high biomass year of 1977 where 56.2% of the population were these early developmental stages. Thus, the low zooplankton biomass measured in early spring 1984 was probably real. There is presently only a perceptual understanding of how physical processes affect the abundance and distribution of Georges Bank zooplankton. The success of a population depends not only on food availability and predator abundance but also upon the dynamics of its physical environment, which influence feeding efficiency, susceptibility to predation, transport, and recruitment success. Sea-surface temperature was the only physical parameter discussed in this report and its narrow range of variability could not be correlated to the comparatively large fluctuations of the zooplank- ton population. Monthly mean-derived wind stress com- ponents in the area (see Ingham and Wood, 1987) and anomalies in the volume of Georges Bank shelf water (Mountain, 1991) were also examined from 1977 to 1986 and no persistent correlation to biomass variabil- ity was evident. Ongoing studies are presently analyz- ing historical time series of physical and biological pa- rameters to help direct future research efforts attempting to couple the physics and biology of the marine environment. Future monitoring surveys of the U.S. Northeast shelf ecosystem will continue to mea- sure the variability of zooplankton and gather infor- 474 Fishery Bulletin 91 13). 1993 mation to identify the key environmental and physical factors that drive their population dynamics. Acknowledgments The author wishes to thank his colleagues in the Na- tional Marine Fisheries Service who participated in the collection and processing of the data used in this report. Special thanks go to Kenneth Sherman, Mark Berman, and John Green for carefully reading and editing the early drafts of the manuscript. Literature cited Bigelow, H. B. 1926. Plankton of the offshore waters of the Gulf of Maine. Bull. Bur. Fish., Wash. 40:1-509. Corkett, C. J., and I. A. McLaren. 1978. The biology of Pseudocalanus. Adv. Mar. Biol. 15:1-231. Cushing, D. H. 1978. Biological effects of climatic changes. RAPP P- v Reun. Cons. int. Explor. Mer 173:107-116. Davis, C. S. 1984. Predatory control of copepod seasonal cycles on Georges Bank. Mar. Biol. 82:31-40. 1987. Zooplankton life cycles. In (R. A. Bakus and D. W. Bourne, eds.) Georges Bank, p. 256-267. MIT Press, Cambridge. GLOBEC: Northwest Atlantic Program. 1991. Canada/U.S. meeting on N.W. Atlantic fisheries and climate. Rep. No. 2, Feb. 1991. Ingham, M. C, and G. Wood. 1987. Variations in the spring windfield in the North- western Atlantic, 1946-1985: a progress report. NAFO SCR Doc. 87/15. Serial No. N1295. Northwest Atlantic Fisheries Org., Dartmouth, Nova Scotia, Canada. Laurence, G. C, and R. G. Lough. 1984. Growth and survival of larval fishes in relation to the trophodynamics of Georges Bank cod and haddock. NOAATech. Mem. NMFS-F/NEC-36. Leak, J. C, and E. D. Houde. 1987. Cohort growth and survival of bay anchovy larvae Anchoa mitchilli in Biscayne Bay, Florida. Mar. Ecol. Prog. Ser. 37:109-122. Manning, J., and T. Holzwarth. 1990. Description of the oceanographic conditions on the Northeast Continental Shelf: 1977-1985. North- east Fisheries Center Reference Document 90-04, NMFS Woods Hole Lab., 373 p. Marcus, N. H. 1989. Abundance in bottom sediments and hatching requirements of eggs of Centropages hamatus (Copepoda: Calanoida) from the Alligator Harbor re- gion, Florida. Biol. Bull. 176:142-146. Marshall, S. M., and A. P. Orr. 1955. The biology of a marine copepod. Oliver and Boyd, London, 188 p. Meise-Munns, C, J. Green, M. Ingham, and D. Moun- tain. 1990. Interannual variability in the copepod popula- tions of Georges Bank and the western Gulf of Maine. Mar. Ecol. Prog. Ser. 65:225-232. Mitchell, J. F. B. 1989. The "greenhouse" effect and climate change. Rev. Geophys. 27:115-139. Mountain, D. G. 1991. The volume of shelf water in the Middle Atlantic Bight: seasonal and interannual variability, 1977- 1987. Cont. Shelf Res. 11:251-267. Roesler, C. S., and D. B. Chelton. 1987. Zooplankton variability in the California Cur- rent, 1951-1982. CalCOFI Rep., Vol.28, p. 59-96. Nelson, G. A. 1990. Population biology and dynamics of northern sand lance (Ammodytes dubius) from the Gulf of Maine to Middle Atlantic Bight region. M.S. thesis, Univ. Mass., Amherst, MA. Sherman, K. 1980. MARMAP, a fisheries ecosystem study in the NW Atlantic: fluctuations in the ichthyoplankton- zooplankton components and their potential for impact on the system. In (F. P. Diemer, F. J. Vern-berg, and D. Z. Mirkes, eds.) Advanced concepts on ocean measurements for marine biology, p. 9-37. Belle W. Baruch Institute for Marine Biology and Coastal Re- search, Univ. South Carolina Press, Columbia, SC. Sherman, K., C. Jones, L. Sullivan, W. Smith, P. Berrien, and L. Ejsymont. 1981. Congruent shifts in sand eel abundance in west- ern and eastern North Atlantic ecosystems. Nature 291:486-489. Sherman, K., J. R. Green, J. R. Goulet, and L. Ejsymont. 1983. Coherence in zooplankton of a large Northwest Atlantic ecosystem. Fish. Bull. 81(4):855-862. Sherman, K., W. G. Smith, J. R. Green, E. B. Cohen, M. S. Berman, K. A. Marti, and J. R. Goulet. 1987. Zooplankton production and the fisheries of the Northeastern Shelf. In (R. H. Backus, ed. ), Georges Bank, p. 268-282. MIT Press, Cambridge. Sibunka, J. D., and M. J. Silverman. 1984. MARMAP surveys of the continental shelf from Cape Hatteras, North Carolina, to Cape Sable, Nova Scotia (1977-1983). Atlas No. 1. Summary of opera- tions. NOAA Tech. Mem. NMFS-F/NEC-33. 1989. MARMAP surveys of the continental shelf from Cape Hatteras, North Carolina, to Cape Sable, Nova Scotia (1984-1987). Atlas No. 3. Summary of operations. NOAATech. Mem. NMFS-F/NEC-68. Wood, G. and V. Tang. 1988. Sea-surface temperature anomalies off the north- eastern USA during 1981-1986. NAFO SCR Doc. 88/ 84. Serial No. N1536. Northwest Atlantic Fish Org. Dartmouth, Nova Scotia, Canada. AbStfclCt.— Demersal fish repre- sent one of the most heavily ex- ploited resources in the Antarctic ecosystem. The stocks around South Georgia Island have contributed a substantial portion of the annual catches and have declined over the past decade. Fishing has been im- plicated as the cause of this decline. However, a clear description of the community structure of this system, which is necessary to judge the in- fluence of fishing accurately, has been lacking. The spatial structure of the South Georgian fish community was inves- tigated through the use of survey data collected over a three-year pe- riod. The results clearly indicated the absence of spatial structure in that community. The presence or ab- sence of rare species at various sta- tions was responsible for the weak structure found in the initial analy- sis. The general lack of structure was consistent from year to year. The available data do not provide an explanation for this lack of struc- ture. All surveys were conducted during the austral summer only. Events and community structure at other times of the year remain un- known. Although the data were rep- resentative of the fish community during the austral summer, no com- parable data were available on the abundance and distribution of their prey items, especially krill (JEuphau- sia superba). More extensive sam- pling, expanded to include other seasons, is necessary to properly ad- dress the questions of seasonal change in community structure and the role of competition in this Ant- arctic system. Spatial structure and temporal continuity of the South Georgian Antarctic fish community James E. McKenna Jr. Graduate School of Oceanography. University of Rhode Island. Narragansett. Rl 02882 Present Address. Florida Marine Research Institute. 100 Eighth Ave S.E. St Petersburg, FL3370! Manuscript accepted 11 May 1993. Fishery Bulletin 91:475-490 1 1993). The marine systems of the Antarctic are important both ecologically and economically. Many of the seasonal and long-term events that occur in this region have a significant impact on global environmental conditions (Gordon, 1975; Broecker and Peng, 1982; Kennett 1982, p. 249). The Ant- arctic is a large region physically iso- lated from the rest of the globe by the circumpolar circulation of the West Wind Drift Current (Kennett 1982, p. 725). The organisms of the Southern Ocean are unique and of- ten highly productive. The waters south of the Antarctic convergence (about 5% of the world ocean) con- tribute a total production equivalent to 209c of that produced by all the oceans of the world (El-Sayed, 1968). In fact, it is believed that the evolu- tion of the Mysticeti (baleen whales) depended on the development of the great production of this area (Fordyce, 1977). The environment of Antarctica and its associated islands has been es- tablished for a long time (ca. 37 mil- lion years), allowing a group of perciform fishes to radiate into a va- riety of niches and to dominate the fish fauna of this highly productive region (DeWitt, 1971; Targett, 1981; Eastman, 1985). Like other polar communities it is one of low species richness (Hedgepeth, 1969; Everson, 1984). Over 70% of the species and 90% of the individuals belong to four families in the suborder Nototheni- oidei and 95*7 of the species in this group are endemic to the Antarctic region (DeWitt, 1971). The four domi- nant families are the Antarctic cods (Nototheniidae), dragonfish (Bathyd- raconidae), icefish (Channichthyidea), and plunderfish (Harpagiferidae). These fish are generally sedentary, benthic forms found on the Antarctic continental shelves (Targett, 1981). Many of the species in this group have evolved to fill niches usually oc- cupied by different families or orders offish (Eastman, 1985). Large stocks of demersal fish were discovered around some of the islands of the Scotia Arc in the late 1960s (Kock, 1986). A relatively intense and successful fishery developed around these stocks, especially those found at South Georgia Island (Fig. 1). However, the decline of those stocks over the past two decades has been evident (Kock, 1985a, 1986; Gabriel, 1987; McKenna and Saila, 1989; McKenna1 ). This is especially true for the target species of this fishery, the marbled rockcod iNotothenia rossii) and the mackerel icefish (Champ- socephalus gunnari). There is also evidence that a short-term (two to three years) shift in the species as- semblage inhabiting the continental shelf around South Georgia has oc- curred (McKenna and Saila, 1991). Although overfishing is suspected as the cause of the changes observed 'McKenna, J. 1990. Status of the stocks of Ant- arctic demersal fish in the vicinity of South Georgia Island. January 1989. Antarctic Mar. Living Res. Contract Rep. Available from Dr. R. Holt, NOAA, NMFS, Southwest Fisheries Sci. Center, La Jolla, CA 92038. 475 476 Fishery Bulletin 91(3). 1993 SOUTH GEORGIA Drake' s Passage ra>» 80° U \f ■* itfl^lJT, 50" U 40° U 30" U 50° S 60° S Figure 1 Map showing the location of South Georgia Island in the Southern ocean. Inset map identifies samples sited during each of the three AMLR cruises. Each asterisk (*) represents a sample taken during the 1986-87 survey. Each triangle ( ) represents a sample taken during the 1987-88 survey. Each square ( ) represents a sample taken during the 1988-89 survey. in the demersal fish community (Kock, 1985b, 1986, 1991; Kock and Koster, 1989, 1990), there is a need for more basic information about the ecology of these fishes. The effects of natural (or anthropogenic) events on the populations of these animals cannot be accurately judged without a clear understanding of their basic community organization. Analysis of Antarctic fish communities has generally been limited to descriptions of the species present and some investigations of their diets. Targett (1981) in- cluded some measures of diversity in his examination of the fish community at three different islands along the Scotia Arc. His research emphasized diets and re- source partitioning, and gives some of the first descrip- tions of relationships between the species based on a quantitative analysis. However, his sampling was lim- ited to one or two locations at each island. More ex- tensive work has been conducted at Elephant Island by Tiedtke and Kock (1989). Their examination of the demersal fish assemblage around that Island found evidence of spatial structure associated with depth. A survey program concentrating on the South Geor- gia area was established as part of the U.S. Antarctic Marine Living Resources (AMLR) project. The research described here used the data from that program to examine the spatial structure of the demersal fish com- munity in the vicinity of South Georgia Island and its temporal consistency from 1986 to 1989. Methods Data used in this study consisted of species abundances (and associated length-frequency information) collected during three research survey cruises in the South Geor- gia area. The abundance of each species in the region was measured both numerically (number/standard tow) and on a biomass (kg/standard tow) basis. These surveys were conducted during the austral summers of 1986-87 (29 Nov.-17 Dec. 1986), 1987-88 (19Dec.-10Jan.), and 1988-89 (17-28 Jan. 1989). The 1986-87 and 1987-88 surveys sampled the abundance McKenna Spatial structure and temporal continuity of South Georgian fish community 477 of fish from approximately 100 stations (Fig. 1) by thirty minute tows of a P32/36 otter trawl (mouth open- ing of 17.5 m, 43-52 mm mesh liner). These stations had been randomly located within three depth strata (50-150 m, 150-250 m, 250-500 m) (Gabriel, 1987; McKenna and Saila, 1989). The 1988-89 survey sampled 41 stations from a regular grid between 50-m and 250-m depth around South Georgia Island (Fig. 1). Collections of demersal fish during that sur- vey were made with fifteen minute tows of a Chris- tensen Bottom Trawl (wing spread of 4.6m, 50-mm mesh codend with a 6-mm liner )( McKenna1). Complete discussions of the methods and results of these sur- veys may be found in Gabriel (1987), McKenna and Saila ( 1989), and McKenna1. Species diversity has become a standard tool for describing natural communities. Three measures of diversity were calculated for each station sampled dur- ing the three surveys and for each survey as a whole based on the total catches. The first was species rich- ness, which was simply the number of species caught at each station. The second was the Shannon-Wiener information index (H\ using loge) (Shannon and Weaver, 1949). The third index (V") was a measure of evenness (Pielou, 1977), V = H7\og(s*), where s* is the total number of species in the region and was assumed to be 30. Species associations give a more detailed descrip- tion of a natural community than the simple summary provided by diversity indices. Species associations within each survey were identified by use of Spearman rank correlation (r') analysis (Freund, 1970, p. 311- 313), and were based on the numerical abundance of fish caught during each survey. Species were arbitrarily designated as rare if they occurred at 5% or less of the sampled stations and uncommon if they occurred at 25% or less of the stations sampled. All pairwise com- binations of common species (within each season) were examined. These correlation values were then used to generate Z-scores (Z = r'V(ra-l)) to test the null hy- pothesis that the correlation was not significantly dif- ferent from zero. Significance of associations was determined at the 0.01 level (Z>2.58) (Freund 1970, p. 313). Cluster analysis was used to examine the spatial structure of the community in an effort to identify significant subcommunities. This analysis method re- duces the complex, multivariate data from field stud- ies to a manageable level (Boesch, 1977) and often reveals the presence of important physical or biologi- cal factors affecting the distribution of the various spe- cies assemblages occupying the sampled region (Pielou, 1977). However, one of the major drawbacks of cluster analysis (and many other exploratory techniques) is the subjectivity associated with its application (Boesch, 1977; Pielou, 1984; Jain and Dubes, 1988). The heterogeneity ratio (HR) was used in this analy- sis as the measure of similarity for the normal (R- mode) cluster analyses, because of the objectivity it provides. It is not affected by sample size or group size (the number of samples included in a cluster), mea- sures the beta-diversity (McNaughton and Wolf, 1979a) existing among samples, and can be statistically tested for significance (Kobayashi, 1987). HR = Sq/ECSq), where SQ is the total number of species present in Q samples and E(SW) is the expected number of species in the Q samples. E(S^) is obtained by applying the mean number of species per sample to the logarithmic series distribution (Fisher et al., 1943), which is used as the model describing the relation between the sample size and number of species in the community. The logarith- mic series describes communities which are dominated by one or a few species and contain numerous rare species. It is a common distribution in nature (Shepard, 1984; Dial and Marzluff, 1989) and is the most appro- priate for this application (Kobayashi, 1987). HR is robust to the requirement of good fit to the logarithmic series model (Kobayashi, 1987). However, the fit of the data to that distribution was tested for each station sampled during the three surveys with the aid of the BASIC program LOGSRFIT.BAS (Saila et al., 1991). This program generates values of X, Fisher's a, and K. X and a are the parameters of the logarithmic series model and a is also a diversity in- dex (Fisher et al, 1943; Saila et al., 1991). Values ofK measure the goodness-of-fit of the data to the logarith- mic series model (Fisher et al., 1943). Values of K less than 1.0 indicate a reasonable fit to the model. A normal (i?-mode) clustering of the stations was performed for each survey. The clustering program, HRCLUSTR.BAS (written in Microsoft QuickBasic v. 4.5, Microsoft 1988), was developed for this purpose (available upon request). It was a modification of Kobayashi's2 program and used the HR and an un- weighted paired group method of averaging (UPGMA) linkage method (Sneath and Sokal, 1973). The null hypothesis of non-significant clusters was rejected at the 0.05 level (Kobayashi, 1987). Inverse (Q-mode) clus- tering was also preformed on each survey's data to classify species into groups. This provides insight into -Kobayashi, Faculty of Agriculture, Yamagata Univ., Tsuruoka, 997 Japan. Pers. commun. 1988. 478 Fishery Bulletin 91(3). 1993 their associations and distributions. The species were clustered with the correlation coefficient and an UPGMA linkage method (Jain and Dubes, 1988, p. 16). The results of these analyses were then combined through nodal analysis (Williams and Lambert, 1961; Lambert and Williams, 1962; Noy-Meir, 1971; Boesch, 1977) into two-way tables displaying the constancy and fidelity of the identified groups (Fager, 1963; Westhoff and van der Maarel, 1973). Constancy describes how widespread a species group is within a given habitat and is expressed as C,j = a,/(nlnl), where a„ is the actual number of occurrences of mem- bers of species group ;' in collection (station) group j, and n, and nt are the number of entities in the respec- tive groups. C,j ranges from 0 to 1, where 1 indicates that all species occurred in all collections in the group (Boesch, 1977). Fidelity describes the restriction of a species group to a given habitat and is expressed as F„ = (a,lI.nJ)/(nJ'La,J) often classes (0-0.1, 0.1-0.2, etc.). Each data set was sorted in the order of the gradient being examined and probed with 500 skewers. Twenty random tables were generated during each analysis to provide a test of the null hypothesis that the assemblages were distributed randomly along the gradient at a 0.05 probability level (Pielou, 1984). Although the change in an assemblage along an en- vironmental gradient may be gradual, there may also be cases of rapid change. These ecotones may be due to rapid change in the environmental gradient or pos- sibly a biological change for some other reason (Whittaker, 1960; McNaughton and Wolf, 1979b). Gra- dient analysis (Webster, 1973; Ludwig and Cornelius, 1987) was applied to the data sorted in order of each of the above potential gradients, to search for bound- aries along the length of those gradients. The BASIC program, GRADSECTBAS (written in Microsoft QuickBasic v. 4.5, Microsoft 1988), was developed for that purpose (available upon request). It used a mov- ing split-window distance method of comparing mov- ing averages (Whittaker, 1960). Each window contained nine stations and the multivariate distance measure was based on the squared Euclidean distance. Values of this index greater than 2 suggest a strong preference of species in a group (i) for a collection group (j), and values much less than 1 suggest avoidance of a collection group (J) by a group of species (i) (Boesch, 1977). The size structure of each species' population (1987- 88 and 1988-89 surveys only) was also examined for patterns. Stations were clustered (using HR and UPGMA linkage) based on the abundance of individu- als in each centimeter length class. Natural communities may be aligned along environ- mental gradients, changing continuously from one end to the other (Whittaker, 1960). Skewer analysis uses multivariate correlations to measure the significance of linear trends in natural communities (Pielou, 1984; Saila et al., 1991). It was used in this study to exam- ine the significance of trends in each community along suspected gradients of depth, longitude, latitude, and time (date within each survey). It was performed twice on each data set. The first application used the abso- lute measures of abundance, which detects trends that are a combination of changes in the species assem- blage and the absolute abundances of individuals at each station. The second application used those abun- dance values converted to proportions, which is sensi- tive only to the changes in species composition along the gradient. Kendall's Tau provided the measure of correlation between each skewer and the observed data. The distribution of Tau was observed through the use Results The assemblage of demersal fish occupying the South Georgia region, during the three AMLR surveys, con- sisted of slightly more than two dozen (28) species, with representatives from thirteen families (Table 1). Seven of these species accounted for the vast majority (>85%) offish biomass and individuals in the system (Fig. 2). Five are relatively large (>50cm), commer- cially valuable species. These included the three icefish (Channichthyidae) of the region (Chaenocephalus aceratus, Champsocephalus gunnari, and Pseudo- chaenichthys georgianus) and two species of rock cod (Nototheniidae: Notothenia gibberifrons and Notothenia squamifrons). The other two important species (Nototheniops larseni and Nototheniops nudifrons) are small but abundant members of the Nototheniidae. The distributions of these species with depth was simi- lar to those described by Tiedtke and Kock ( 1989) for the fishes of Elephant Island. Champsocephalus gunnari was the clear dominant both numerically and based on biomass (Fig. 2). One quarter of the biomass and nearly one third of the individuals were members of this species. It feeds al- most strictly on krill (McKenna. 1991) and leads a more pelagic existance than the other icefish of the region (Kock, 1985b). Notothenia gibberifrons was a close second in biomass and was the third most abun- dant species. It is adapted to an epibenthic environ- McKenna: Spatial structure and temporal continuity of South Georgian fish community 479 Table 1 Species codes and rareness designations. R = RARE (occurred at 5% or fewer of the stations sampled). U = UNCOMMON (occurred at 259c or fewer of the stations sampled . C = COMMON (occurred at more than 259r of the stations sampled). ' — ' indicates that that species was not caught during a particular survey. Species Rareness designation code Family and species 86-87 87-88 88-89 Artedidraconidae (Plunderfishes) ARTE Artididraco mints Bathydraconidae (Dragonfishes) C C C BATH Bathydraconidae spp. R — — PARA Parachaemchthys georgianus C C C PSIL Psilodraco breviceps Bothidae (Armless Flounders) U u u MANT Maneopsetta maculata Centrolophidae (Ruffs) U u R CENT Ccntrolophidae spp. Channichthyidae (Icefishes) R ACER Chaenocephalus aceratus C C C GUNN Champsoeephalus gunnari c C C PSEU Pseudochaenwhthys georgianus Gempylidae (Snake Mackerels) c c C DIPL Paradiplospinus antarcticusa Harpagiferidae ( Spiny Plunderfishes R HARP Harpagifer georgianus Liparididae (Snailfishes) R CARE Careproetus georgianus — R — LIPA Paraliparis spp. Muraenolepididae (Moray Cods) u U u MICR Muraenolepis mierops Myctophidae (Lanternfishes) c c c ELEC Eleetrona antarctica R — R MYCT Myctophidae spp. U — — NICH Gymnoscopelus nicholsi Nototheniidae (Antarctic Rock Cods) u R ANGU Notothenia angustifrons R R R ELEG Dissostichus eleginoides C C R GIBB Notothenia gibberifrons C C C GUNT Patagonothen brevicauda R R — HANS Pagothema hansom C C U KEMP Notothenia kempt R R — LARS Notothemops larseni C C C NUDI Nototheniops nudifrons C C c ROSS Notothenia rossii C C u SQUA Notothenia squamifrons Rajidae (Skates) C U — RAJA Raja georgiana Zoarcidae (Eelpouts) C U u MELA Melanostigma gelatinosum — R — ment (Daniels, 1982) and feeds on invertebrate infauna, primarily polychete worms (McKenna, 1991). Nototheniops larseni was the second most abundant species and is the most pelagically adapted of the com- mon species (Targett, 1981). It is a relatively small fish (<25cm) that feeds on krill and other pelagic in- vertebrates (McKenna, 1991). Chaenocephalus aceratus and P. georgianus are the other two icefish found in the vacinity of South Georgia Island. Both species are closely associated with the sea floor and feed heavily on other fish and krill (McKenna, 1991). Notothenia squamifrons is a demersal rock cod that prefers the deeper strata of the region. It feeds on a wide variety of benthic invertebrates but has a preference for tunicates (McKenna, 1991). Nototheniops nudifrons is one of the smallest species in the region (10-5 cm). It lives in and among the sedentary megainver- tebrates growing on the bottom (e.g., sponges). It feeds on benthic epifauna, but most of its diet con- sists of krill (McKenna, 1991). Other species collected by the AMLR sur- veys generally accounted for less than 2% of the fish in the region. For a more detailed description of this community see Gabriel (1987), McKenna and Saila (1989, 1991), and McKenna (1991). Diversity values were moderate throughout the region and relatively consistent from year to year. The overall diversities (H') for each sur- vey (1986-87-1988-89) were 1.51. 1.94, and 1.83 (based on numerical abundances) (Table 2). The value for any given station ranged from 0 to 2.097 (Fig. 3). Richness never ex- ceeded 16 at a single station, but was as low as 1 (which accounts for the zero H' values) in a few cases. Survey-wide richness was nearly the same for 1986-87 and 1987-88 but declined in 1988-89 because of the lack of rare species collected during the survey. Overall evenness for each survey ranged from 1.02 to 1.35 and was greater in the latter two surveys. Significant associations were found between species within each survey. However, all were weakly correlated (Table 3). Only three as- sociations had r' values greater than 60cr and none explained more than 707r of the variability in their distributions. None of the fifteen associations with r' values greater than 50% were consistent from year to year. However, three of these associations (C. aceratus- P. georgianus, Artedidraco mirus-N. nudifrons. Mur- aenolepis microps-N. nudifrons) persisted from 1986- 87 to 1987-88. The inverse (Q-mode) cluster analysis appeared to classify the species into groups based on their relative rareness, but there were exceptions to such a classification in nearly every group. 480 Fishery Bulletin 9 1 (3), 1993 SQUA (10.1%) other (3.7' RAJA (3 2%) BOSS (3.6%) MICR (2-6%) HANS (2.0%) LARS (3.9%) ACER (11.0%) GUNN(23 9%) GIBB (21 9%) PSEU (14 0%) 6 Figure 2 Pie charts illustrating the average composition of the demer- sal fish community around South Georgia Island during the AMLR surveys. (A) Composition based on the biomasses of the fish collected; (B) composition based on the numerical abundance of the fish collected. Table 2 Diversity of South Georgia Island Demersal Fish Commu- nity. R represents species richness. The # symbol indicates that the diversity values were calculated on a numerical abun- dance basis. The wt. symbol indicates that the diversity val- ues were calculated on a biomass basis. H' is the Shannon-Wiener information index. V is an index of species evenness. 1986-87 1987-88 1988 -89 R— 27 R— 25 R- 18 # wt. # wt. # wt. H' V 1.51 1.91 1.02 1.29 1.94 1.94 1.31 1.31 1.83 1.24 1.99 1.35 The distribution of species abundances at most sta- tions was described well by the logarithmic series model (Fig. 4). However, 27% of the stations ( 11) in the 1988- 89 biomass data set fit the logarithmic series model poorly, while those of earlier surveys had less than half that proportion of poor fitting stations. The bio- mass data sets generally had more poor fitting sta- tions than did the numerical data sets. The normal cluster analyses clearly demonstrated that there was little spatial structure in the demersal fish community. Extensive chaining (Boesch, 1977; Jain and Dubes, 1988), frequent reversals (crossovers) (Kobayashi, 1987), and low number of significant clus- ters were indicative of the absence of spatial structure (Fig. 5, page 484). The 1986-87 and 1987-88 commu- nities displayed slightly more significant structure than that in 1988-89, but chaining and reversals were preva- lent there as well. The nodal analyses demonstrated that the species groups occurred with similar constancy throughout the region and fidelity was not strong for any particular group (Fig. 6, page 484). Species composition of significant clusters was basi- cally the same as that for the region-wide species as- semblage (Fig. 2), except for those clusters formed by a small number of samples with rare species (Fig. 7, page 485). Clusters containing samples from Shag Rocks had Patagonothen brevicauda as an additional component. Close examination of the clusterings re- vealed that the presence or absence of the rarest spe- cies was the basis for what little structure was present. When these rare species (those which occurred at 5% or fewer stations) were removed, all or nearly all sig- nificant structure disappeared. The lack of structure was consistent from year to year. Clustering of the length distribution of each species revealed little structure there as well. Notothenia gibberifrons, C. gunnari, andiV. larseni showed a weak separation of large and small fish; large fish associ- ated with deeper water and small fish with shallower stations during 1987-88 (Figures 8, A and B; 9A). Only N. larseni continued to show that trend in 1988- 89 (Fig. 9B). Notothenia larseni also displayed a tem- poral separation in size, with larger fish caught later in the cruise in both surveys (Fig. 9C). Muraenolepis microps showed a similar weak trend of larger fish later in the cruise during the 1988-89 survey (Fig. 8C). The gradient analyses revealed several significant, but weak trends. Skewer analysis detected significant trends in longitude in all three years. Significant trends in latitude and depth were identified in 1987-88 and in 1988-89. In 1986-87 and 1988-89, significant tern- McKenna Spatial structure and temporal continuity of South Georgian fish community 481 50 A45 40 35 J? 30 c 0) D 25 IT 0) S. *> 15 10 5 0 45 40 35- "• 20- 15 10 5 0 E 18 16 14 >. 12 I 10 based on 1986-87 numerical abundance il CD v WZ R/10 it 0 03 0.6 0-9 1.2 1 5 Class of Diversity Measure based on 1987-88 numerical abundance ■■ H' CD V VZ2 R/10 Ui. Mil 0.3 0.6 0.9 1.2 1.5 Class of Diversity Measure based on 1988-89 numerical abundance JPUU wm H' CD V ?77X R/10 111 0.3 06 0.9 1.2 Class of Diversity Measure 44- B >- 12 I 10 based on 1986-87 biomass K H' CD V T77A R/10 0.3 0.6 0.9 1.2 1.5 1 Class of Diversity Measure 0.3 0.6 0.9 1.2 1.5 Class of Diversity Measure based on 1988-89 biomass ■■ H' CD v VZA R/10 Class of Diversity Measure Figure 3 Distributions of diversity values for each AMLR survey. H' (SOLID BAR) represents the Shannon-Wiener Diversity index, and VlOPEN BAR) represents evenness, R (SHADED BAR) represents richness. (Richness was adjusted to the scale of the other indices by dividing by 10.) (A) Diversity of fish community at stations sampled during the 1986-87 survey, based on numerical abundance; (B) diversity of fish community at stations sampled during the 1986-87 survey, based on biomass; (C) diversity offish community at stations sampled during the 1987-88 survey, based on numerical abundance; (D) diversity offish community at stations sampled during the 1987-88 survey, based on biomass; (E) diversity of fish community at stations sampled during the 1988-89 survey, based on numerical abundance; (F) diversity of fish community at stations sampled during the 1988-89 survey, based on biomass. 482 Fishery Bulletin 91(3), 1993 Table 3 Results of spearman rank correlation analyses on common Antarctic demersal fish species associations. All correlation values in this table were significant at the 0.01 level when tested against the null hypothesis of zero correlation 1986/87 1987/88 1988/89 Species Species Species pair r' pair r' pair ;■' ACER-ARTE -0.40 ACER-GUNN +0.30 GIBB-LARS +0.50 ACER-GIBB +0.35 ACER-PSEU +0.55 GIBB-MICR +0.51 ACER-GUNN +0.31 ACER-PSIL +0.29 LARS-ACER +0.54 ACER-NUDI -0.33 ACER-RAJA +0.33 LARS-MICR +0.59 ACER-PSEU +0.52 ARTE-ELEG -0.38 MICR-ACER +0.48 ARTE-MICR -0.64 ARTE-GIBB +0.26 ARTE-NUDI +0.70 ARTE-HANS -0.39 ARTE-PARA +0.36 ARTE-MICR -0.40 ARTE-PSEU -0.47 ARTE-NUDI +0.60 ARTE-RAJA +0.35 ELEG-HANS +0.40 ARTE-SQUA -0.49 ELEG-MANT +0.43 ELEG-ARTE -0.36 ELEG-NUDI -0.38 ELEG-HANS -0.47 GIBB-LARS +0.28 ELEG-NUDI -0.37 GIBB-NUDI +0.36 ELEG-PSEU -0.36 GIBB-PARA +0.30 ELEG-ROSS -0.46 GIBB-PSIL +0.26 GIBB-LARS +0.29 GIBB-ROSS +0.34 GIBB-NUDI +0.33 GUNN-PSEU +0.40 GIBB-PARA +0.30 GUNN-PSIL +0.33 GIBB-PSEU +0.32 HANS-NUDI -0.36 GIBB-PSIL +0.41 HANS-PSEU +0.28 GIBB-ROSS +0.36 LARS-PARA +0.30 GUNN-PARA +0.34 LARS-PSIL +0.41 GUNN-PSEU +0.34 LARS-RAJA +0.40 HANS-PSIL +0.36 MICR-MANT +0.34 LARS-PARA +0.46 MICR-NUDI -0.55 LARS-PSIL +0.36 MICR-PARA -0.32 MICR-MANT +0.56 MICR-PSIL +0.31 MICR-NUDI -0.63 MICR-RAJA +0.35 MICR-PARA -0.39 PSEU-PSIL +0.35 MICR-SQUA +0.59 NUDI-RAJA -0.29 NUDI-SQUA -0.37 PARA-LARS +0.46 PARA-NUDI +0.49 PARA-SQUA +0.35 PSEU-NUDI -0.41 SQUA-MANT +0.53 SQUA-RAJA -0.40 peaks are real boundaries be- tween communities (Fig. 11, page 488). Figure 12, page 488 sum- marizes the most pronounced boundaries from each of the gra- dients examined. poral trends also existed. However, all of the trends revealed by skewer analysis were weak; the primary mode fell into either the smallest or second smallest class interval of Tau (Fig. 10, page 487). Gradient analysis, using GRADSECT.BAS, measures the rate of change of the species assemblage along the gradient of interest. However, it offers no objective de- cision about the value that the rate must reach in order to be considered significant. Thus, it will iden- tify ecotones along any gradient and it is a matter of subjective judgment to determine which (if any) of the Discussion The diversity values for the South Georgian fish community were generally greater than those re- ported by Targett (1981) for the same region. They were within the range of those reported for de- mersal fish in tropical estuarine areas (Yanez-Arancibia et al., 1980). These values place the South Georgia fish community on the relative, global diversity con- tinuum, but the wide range of values demonstrates that diver- sity alone is a poor model of this community. The richness component of di- versity was low and similar to that reported by Targett (1981), but the evenness was consider- ably larger than had been re- ported for this region. This obser- vation is consistent with the finding that a significant shift in the species assemblage from 1986-87 to 1987-88 was associ- ated with an increase in diver- sity because of an increase in the evenness component of that index (McKenna and Saila, 1991). The increasing number of stations that poorly fit the logarithmic se- ries distribution with time also supports the hypothesis that the community was changing toward one which was less dominated by a small number of species. This appar- ent change in the community was compounded in 1988- 89 by the use of different sampling techniques and gear. The sampling during that season was insuffi- cient to accurately represent the abundance of the rar- est species. A bias was also introduced by the small gear used for that survey. In all previous AMLR sur- veys the mackerel icefish (C. gunnari) was the most abundant species. The smaller individuals of this spe- cies tends to be more pelagic than many of the others McKenna: Spatial structure and temporal continuity of South Georgian fish community 483 A based on 1986-87 numerical abundance B based on 1986-87 blomese 25- c D cr £ to- ill. .. >. 5 c cr s- iII..lm . 006 i>3 Q4B ass ass i ot i as 1 46 i as i ss" zos ooe' a a ate' 'ass' bae i os 12s' 1 46' 1 «' its' ic* Class of K (midpoint) Class of K (midpoint) c based on 1987-88 numerical abundance D • based on 1987-88 blomaae m ■<■ o c C7 E lllll O '* c cr 0) J III aw oas n*i bee a as its' i-as" i «*' ic its sc* O 06 Q JE~ 0 44 0 S6 0 B6 1 06' 1 26* 1 46 1 6s' I W 2 06 Class of K (midpoint) Class of K (midpoint) E a based on 1988-89 numerical abundance F„ based on 1988-89 blomasa o C s (1) D cr ,. 1 II s- > c 2 are not shown. (A) K- values offish community at stations sampled during the 1986-87 survey, based on numerical abundance; (B) A'-values offish community at stations sampled during the 1986-87 survey, based on biomass; (C) /^-values offish community at stations sampled 'during the 1987-88 survey, based on numerical abundance; (D) K- values of fish community at stations sampled during the 1987-88 survey, based on biomass; (E) K- values of fish community at stations sampled during the 1988-89 survey, based on numerical abundance; (F) /f-values offish community at stations sampled during the 1988-89 survey, based on biomass. in this system (Slosarczyk, 1983). The smaller gear might have missed some concentrations of this spe- cies, thus underestimating its population and dimin- ishing the apparent dominance of the community. The species associations and distribution of species groups were weak and inconsistent. The association between A. mirus and N. nudifrons was the strongest and still only accounted for, at most, 70% of the vari- ability in their distributions (Table 3). Both of these fish are small (<15 cm), benthic species; however, their diets overlapped only slightly (38%) (McKenna, 1991). They were often removed from within sponges and other large invertebrates that formed a major part of the catch. They were not numerically dominant, but they were ubiquitous in the region. Although their dis- tributions may have been linked, the association is probably more indicative of how widespread their pre- ferred habitat is (i.e. habitat containing an abundance 484 Fishery Bulletin 91(3). 1993 1.1 Bosed on 19SB-ag — Nunericol Rbundance B 8.9 5TRTION Based on 198B-Q9 Blonass n, Figure 5 Results of cluster analysis of the 1988-89 AMLR survey data. Station labels are read vertically under each branch of the dendrogram. The dashed line represents the level at which significantly different clusters begin. (A) Dendrogram of clustering based on numerical abundance; (B) dendrogram of clustering based on biomass. of large, epibenthic invertebrates: branching sponges, corals, tunicates, crinoides, etc.). The C. aceratus-P. georgianus association is most likely a result of the similar feeding behavior and habi- tat preferences of these species. Both are relatively sedentary, benthic fish which feed heavily on krill and other fish. Their diets strongly overlap (64%, McKenna, 1991) and their distribution may well reflect the dis- tribution of their prey. Significant associations suggest some biological in- teractions between the species involved. However, little can be said about the cause of these associations. Two species may display a negative association owing to distinct habitat requirements; they may have similar habitat requirements but experience a niche shift be- cause of competition ( McNaughton and Wolf, 1979b); or they may not coexist because of predation by one on the other. Most of the significant negative associations in this study involved one piscivorous member SPECIES GROUPS STATION GROUPS •O.irj O.10 0.2g| 0.3^ 0.4H 0.5H 0.6|g 0.7| Nodal Constancy Diagram (1986-87 biomass) SPECIES GROUPS STATION GROUPS <'□ '0 =^ m >m Nodal Fidelity Diagram (1986-87 biomass) Figure 6 Results of nodal analysis of the 1986-87 AMLR survey data, based on biomass. The width of each band is proportional to the number of entities in the associated group. (McKenna, 1991), which suggests predator-prey rela- tionships. The strongest of these negative associations were between M. microps and two of the smallest spe- cies in the region, N. nudifrons and A. mirus (Table 3). These three species are found close to the bottom and much of their diet consists of benthic organisms (37% overlap, McKenna, 1991). However, M. microps grows to over twice the size of the other two (Fischer and Hureau, 1985) and will feed on fish (McKenna, 1991) (though its small mouth probably limits any predation on the other two species to juveniles). Information about available habitats and the diets of the individuals within each habitat are necessary before these rela- tionships may be more clearly defined for the South Georgia system. The classification of the species into groups (by in- verse clustering) and their distributions were not McKenna: Spatial structure and temporal continuity of South Georgian fish community 485 CLUSTER 1 other (9.1%) ACER (3.4%) LARS (14 6%) GUNN (16.6%) ACER PSEU (4.8%) CLUSTER 2 other (54%) (4 5%) GIBB (14.1%) NUDI (16.2%) CLUSTER 3 (77) CLUSTER 4 other (6 3%) ACER (3.8%) olhe' (4 7%) GIBB (5" RARE SQUA (9 8%) GIBB (10.5%) - GUNN (22.5%) GUNT -(35.8%) SPECIES — ]ANGU (KEMP ICARE CENT DIPL HARP LIPA MELA NICH MICR (6.7%) LARS (17.2%) (13) (4) CLUSTER 5 other (4 5%) GUNN (6.7%) GIBB (10.5%) PSEU (13.3%) (2) Figure 7 Charts describing the composition of the communities identified by cluster analysis of the 1987-88 AMLR survey data, based on numerical abundance. Pie charts depicting the composition of each community. Stations comprising each cluster are listed horizontally below each chart. The species which accounted for at least 10% of the community are listed to the left of each chart. Bars depicting the presence/absence of the rarest species in each community. Shaded boxes indicate presence. Stations comprising each cluster are listed horizontally below each chart. Species codes are read vertically under each box of a given bar. clearly defined. The weakness of the pairwise associa- tions and the nebulous basis for the classification of species into groups suggests the lack of stronger eco- logical ties between some species than others in the community. Normal (r?-mode) clustering of sample sites based on both species assemblages and size structure of each species' population indicated the lack of spatial struc- ture in the community, as well. The movement of fish towards deeper water as they grow larger offers a pos- 486 Fishery Bulletin 91(3). 1993 Notothenia gibberifrons Length Clustering + + + + + ^^ H/R I 1 1 1 1 1 1 I 1 1 1 1 1 1 1.1 0 10 20 30 40 50 60 50 150 250 350 LENGTH (cm) DEPTH (m) Champsocephalus gunnari Length Clustering B 20 1.7 1.4 0 10 20 30 40 50 60 50 150 250 350 LENGTH (cm) DEPTH (m) Muraenolepis microps Length Clustering CLUSTER ■■>■ K]^H I— »• h^TJ KLZ^ ^nih ^H {±}H 30 H/R I — I — I — I — I — (— 0 10 20 30 40 50 LENGTH (cm) I 1 1 1 1 50 1 50 250 DEPTH (m) Figure 8 Results of cluster analysis of the size structure of three fish populations around South Georgia Island. Each dendrogram depicts the classification of samples according to the size com- position of the associated fish population. Box plots describe the location and spread of the total length of fish in each cluster and the depth or sequential order of the stations that comprise each cluster. The broken verticle lines within each box represent the upper and lower notches and provide a rough measure of the significance of differences between me- dians (Hoaglin et al, 1983). A = N. gibberifrons during 1987- 88 survey compared with depth; B = C. gunnari during 1987-88 survey compared with depth; C = M. microps during 1988-89 survey compared with sample order. Nototheniopslarseni Length Clustering (1987/88) CLUSTER -3- •-* > — CSH I — io-hQh. • ' 8 4Q-I* -EEB— ■ -n+n — 1 H 1 1 1- r-EH + + I 1 1 1 1 1 1.5 1.3 1.1 0 10 20 30 40 50 0 100 200 300 400 H/R Length (cm) Depth (m) Nototheniops larseni Length Clustering (1988/89) -6- ■^h .r-TjH. ^+IH K^D^ K±^ i — EH I 1 h I 1 1 1 1 1 I 1 1 1 1 1 1.6 13 0 10 20 30 40 50 0 50 100 150 200 H/R Length (cm) Depth (m) Nototheniops larseni Length Clustering (1988/89) CLUSTER <^}H "-GD-" -6- h^+ih ^HUSH •i — © 0}-h H/R 10 20 30 40 Length (cm) H h 10 20 30 40 Event Number Figure 9 Results of cluster analysis of the size structure of the N. larseni fish population around South Georgia Island. Each dendrogram depicts the classification of samples according to the size composition of the associated fish population. Box plots describe the location and spread of the total length of fish in each cluster and the depth or sequential order of the stations that comprise each cluster. The broken verticle lines within each box represent the upper and lower notches and provide a rough measure of the significance of differences between medians (Hoaglin et al., 19831. A = Comparison of 1987-88 population with depth; B = comparison of 1988-89 population with depth; C = comparison of 1988-89 popula- tion with sample order. McKenna. Spatial structure and temporal continuity of South Georgian fish community 487 ■o o E 90- A 1986-87 AMLR Survey Data 80- 7Q- 60- random f"l long. bio. prop. Long.#'s prop. 50- 40- 30- 20- 10- 0- | I I ! J i i In i 1 0.05 0.15 0 25 0 35 0.45 0.55 0 65 0.75 0.85 0.95 Class of |Tau| B 1986-87 AMLR Survey Data random st#.#'s prop. sl# bio prop Sl# #S.# k. 005 015 0.25 0 35 0 45 0 55 0.65 0.75 0 85 0 95 Class of )Tau| Figure 10 Results of skewer analysis of the 1986-87 AMLR survey data. LEGEND: random indicates the bar representing the results from the 20 random tables generated by each analysis, long. = longitude; st# = station number; # = analysis based on numerical abundance; bio. = analysis based on biomass; prop. = use of proportion. (A) Histogram indicating the frequency of the location of the primary mode while examining the trend in longitude; (B) histogram indicating the frequency of the location of the primary mode while examining the trend in date (as indicated by station number). sible explanation for changes in size structure of some of the species within a survey. However, those trends were all weak and inconsistent. In general the results indicate that there was little or no structure to the demersal fish community of South Georgia Island. No areas that could be considered nurseries were identified. Nor were there areas of particularly high abundance of only the most valuable commercial spe- cies. Variability in the abundance of each species was evident. However, essentially the same assemblage with similar dominant species was found at nearly every sampled location on the continental shelf. This consistency in the species assemblage has been noted in the South Georgia region before (Targett, 1981). The absence of spatial structure is unusual in natu- ral communities and it is unclear why this situation exists in the South Georgia Island system. The species within this community can be classified into three tro- phic groups: krill-eaters (e.g., C. gunnari and N. larseni), piscivores (e.g., C. aceratus and P. georgianus), and benthic invertebrate feeders (e.g., N. gibberfrons and N. squamifrons) (McKenna, 1991). One would ex- pect these animals to be distributed according to the availability of their prey or appropriate habitat (or both). Although the trophic groups are clear, they are closely linked and the distribution of all these species may be dependent on the abundance and distribution cf krill (Euphausia superba ). If krill was superabun- dant then, even if it were patchy, it might have more than met the demands of the krill-eaters, explaining their uniform distribution. The piscivores will follow their fish prey, which fed either on krill or some benthic resource. The benthic invertebrate feeders were also uniformly distributed. Little is known about the benthic community around South Georgia Island. However, it has been shown that most of the organisms of that habitat are inedible to fish (Belyaev and Ushakov, 1957). It might be expected then that the edible benthic resources are distributed in patches within this habi- tat. That was not apparent from the distribution of the fish and may have been concealed by the fact that even the benthic feeders ate krill (McKenna, 1991), including N. gibberifrons and N. nudifrons. These two species are highly adapted to feeding on benthic or- ganisms and yet their diets included krill. Targett (1981) also observed krill in the diet of these benthic- feeding species and suggested that a shoal of krill had moved into shallow water and was forced close to, or in actual contact with, the bottom. This would seem to be an indication of the great abundance of krill in the vacinity of South Georgia Island. The results of this research imply that during the austral summer there was, in general, a uniform dis- tribution of the necessary resources throughout the sampled region, at scales of 10 km to 100's of km. It must be emphasized that these surveys took place only in the summer and little is known about this commu- nity at other times of the year. The strong seasonality of this environment may present the inhabitants with an annual 'boom' and 'bust' cycle, having abundant resources in the summer and a 'bottleneck' period (when resources are less available than at other time of year) during the winter (Wiens, 1977; DuBowy, 1988). The abundance of available resources during the sum- mer may allow different species to feed on the same 488 Fishery Bulletin 91(3). 1993 50- 45- 40- 120n ft *13 )m 35- of Change (Millions) Oi O 2 20- m DC ,5; , ,, 10- i r k 5- 0- 1 15Pm 250m 40 67 22 97 74 124 75 98 29 80 25 88 125 51 56 110 34 Station at Middle of Window Figure 1 1 Example output from gradient analysis performed by GRADSECT, with the 1986-87 AMLR survey data sorted by depth. the populations of these species than will those of the summer. To better understand this ecosys- tem, information on the seasonal dynamics of the community and its constituent populations (in- cluding krill and benthic inver- tebrates) is needed. Acknowledgments My thanks go to S. Saila, J. Heltshe, and K. Sherman for their technical and editorial as- sistance. Supporting research for this study was funded by grants from the National Oceanic and Atmospheric Administration through the Antarctic Marine Living Resources Project of the National Marine Fisheries Ser- vice. resource (e.g., krill) without experiencing competition (McKenna, 1991). However, during the long winter, resources may be limiting, causing strict resource par- titioning, and more highly defined community struc- ture may develop. If this is the case, the events of the austral winter will be more important to regulation of Literature cited Belyaev, G. M., and P. V. Ushakov. 1957. Certain regularities in the quantitative distri- bution of bottom fauna in Antarctic waters. Doklady Akad. Nauk, SSSR. 112(1): 137. 53.08 53.88 DEGREES SOUTH LATITUDE 54.67 55.58 '88 n '89 Tl»b '88 b -r-r '8b n«b 43.88 42.08 41.88 40.88 39.80 38.88 DEGREES UEST LONGITUDE 37.88 36.08 35.00 Figure 12 Summary of major ecotones identified by GRADSECT analysis. The horizontal lines represent the location of major breaks in the latitudinal gradient. The vertical lines represent the location of major breaks in the longitudinal gradient. The dashed lines represent the major breaks identified along the depth gradient. Abbreviations: '86 = based on 1986-87 survey data: '88 = based on 1988-89 survey data; n = based on numerical abundance; b = based on biomass. McKenna. Spatial structure and temporal continuity of South Georgian fish community 489 Boesch, D. F. 1977. Application of numerical classification in ecologi- cal investigations of water pollution. Virginia Insti- tute of Marine Science, Special Scientific Report No. 77. EPA-600/3-77-033, 113 p. Broecker, W. S., and T-H. Peng. 1982. Tracers in the sea, p. 275-499. Lamont-Doherty Geological Observatory, Columbia Univ., Palisades, NY. Daniels, R. A. 1982. Feeding ecology of some fishes of the Antarctic Peninsula. Fish. Bull. 80:575-588. DeWitt, H. H. 1971. Coastal and deep-water benthic fishes of the Ant- arctic. Antarctic map folio series. Folio 15. New York: American Geographical Society. Dial, K. P., and J. M. Marzluff. 1989. Non-random diversification within taxonomic as- semblages. Systematic Zoology 38:26-37. DuBowy, P. J. 1988. Waterfowl communities and seasonal environ- ments: temporal variability in interspecific competi- tion. Ecology 69:1439-1453. Eastman, J. T. 1985. The evolution of neutrally buoyant Notothenioid fishes: their specializations and potential interactions in the Antarctic marine food web. In W. R. Siegfried, PR. Condy, and R. M. Laws (eds.), Antarctic nutrient cycles and food webs, p. 430-436. Springer-Verlag, NY. El-Sayed, S. Z. 1968. On the productivity of the southwest Atlantic Ocean and the waters west of the Antarctic Penin- sula. In G.A. Llano and W. L. Schmitt (eds.), Biol- ogy of the Antarctic seas III, p. 15-47. Antarctic research series. American Geophysical Union, 11. Everson, I. 1984. Chapter 10: Fish biology. In R. W. Laws (ed.), Antarctic ecology, vol. 2, p. 491-532. Acad. Press, NY. Fager, E. W. 1963. Communities of organisms. In M. N. Hill (ed.), The sea: ideas and observations of progress in the study of the seas, p. 415-433. Wiley-Intersecience, NY. Fisher, R. A., A. S. Corbet, and C. B. Williams. 1943. The relation between the number of species and the number of individuals in a random sample of an animal population. J. Animal Ecology 12:42-58. Fischer, W., and J. C. Hureau. 1985. FAO species identification sheets for fishery pur- poses, p. 261-277. Southern Ocean, fishing areas 48. 58, and 88. CCAMLR Convention Area, FAO, Rome. Fordyce, R E. 1977. The development of the circum-Antarctic Cur- rent and the evolution of the Mysticeti (Mammalia- Cetacea). Paleo Geology, Paleo Climatology, Paleo Ecology. 21:265-271. Freund, J. E. 1970. Statistics: a first course. 2nd ed., p. 311— 313. Prentice-Hall, Inc., Englewood Cliffs, NJ. Gabriel, W. 1987. Results of fish stock assessment survey, South Georgia region, November-December 1986. SC- CAMLR-WBG/12. Gordon, A. L. 1975. Antarctic oceanographic section along 170°E. Deep Sea Research 22:357-377. Hedgepeth, J. W. 1969. Introduction to Antarctic zoogeography. Distri- bution of selected groups of marine invertebrates in waters south of 35°S latitude. Antarctic map folio series; no. 11:1-9. Hoaglin, D .C, F. Mosteller, and J. W. Tukey. 1983. Understanding robust and exploratory data analysis, p. 33-77. John Wiley & Sons, Inc., N Y. Jain, A. K., and R. C. Dubes. 1988. Algorithms for clustering data, p. 16. Prentice- Hall, Englewood Cliffs, NJ. Kennett, J. P. 1982. Marine Geology, p. 725. Prentice-Hall, Inc., Englewood Cliffs, NJ. Kobayashi, S. 1987. Heterogeneity ratio: a measure of beta-diversity and its use in community classification. Ecological Research 2:101-111. Kock, K-H. 1985a. Preliminary results of investigations of the Fed- eral Republic of Germany on Notothenia rossii marmorata (Fischer, 1885) in January/ February 1985. SC-CAMLR-IV/BG/11; Available from CCAMLR, 25 Old Wharf, Hobart, Tasmania 7000 Australia. 1985b. Marine habitats — Antarctic fish. In W. N. Bonner and D. W. H. Walton (eds.) Key environments Antarctica, p. 173-192. Pergamon Press, NY. 1986. The state of exploited Antarctic fish stocks in the Scotia Arc region during SIBEX (1983- 1985). Arch. FischereiWiss. 37:129-186. 1991. The state of exploited Antarctic fish stocks in the Southern Ocean — A review. Arch. FischereiWiss. 41:1-66. Kock, K-H., and F. W. Koster. 1989. The state of exploited fish stocks in the Atlantic sector of the Southern Ocean. Mitt. Inst. Seefisch. Hamb. 46:1-73. 1990. The state of exploited fish stocks in the Atlantic sector of the Southern Ocean. In K. R. Kerry and G. Hempel (eds.), Antarctic ecosystems, p. 308- 322. Springer-Verlag, NY. Lambert, J. M., and W. T. Williams. 1962. Multivariate methods in plant ecology. VI. Nodal analysis. J. Ecology 50:775-802. Ludwig, J. A., and J. M. Cornelius. 1987. Locating discontinuities along ecological gradi- ents. Ecology 68:448-450. McKenna, J. E. Jr. 1991. Trophic relationships within the Antarctic dem- ersal fish community of South Georgia Island. Fish. Bull. 89:643-654. 490 Fishery Bulletin 9 1 (3). 1993 McKenna, J. E. Jr., and S .B. Saila. 1989. Results of fish stock assessment survey. South Georgia, December 1987^January 1988. CCAMLR Selected Scientific Papers, 1988, SC-CAMLR-VII/BG/ 23. Available from: CCAMLR, 25 Old Wharf, Hobart, Tasmania 7000 Australia. 1991. Shifts in the Antarctic demersal fish community of South Georgia Island. Fish. Res. 12:109-124. McNaughton, S. J., and L. F. Wolf. 1979a. General ecology. 2nd ed. Chapter 19: Species diversity and coexistence, p. 376. Holt, Rinehart, and Winston, NY. 1979b. General ecology. 2nd ed. Chapter 18:. Commu- nity organization. Holt. Rinehart, and Winston, NY. Microsoft Corporation. 1988. QuickBasic, version 4.5. Noy-Meir, I. 1971. Multivariate analysis of the semi-arid vegeta- tion in southeastern Australia. I. Nodal ordination by component analysis. Proc. Ecological Society of Aus- tralia 6:159-193. Pielou, E. C. 1977. Mathematical ecology. Chapter 20: The classifi- cation of communities. John Wiley & Sons, NY. 1984. Probing multivariate data with random skewers: a preliminary to direct gradient analysis. Oikos 42:161-165. Saila, S. B., J. E. McKenna, S. Formacion, C. Silvestre, and J. McManus. 1991. Empirical methods and models for multispecies stock assessment. CRISP manual on assessment of tropical multispecies fish stocks. Lewis Publ., Inc., Boca Raton, FL. In press. Shannon, C. E., and W. Weaver. 1949. The mathematical theory of communication. Univ. Illinois Press, Urbana, IL, 117 p. Shepard, R. B. 1984. The log-series distribution and Mountford's simi- larity index as a basis for the study of stream benthic community structure. Freshwater Biol. 14:53-71. Slosarczyk, W. 1983. Preliminary estimation of abundance of juvenile Nototheniidae and Channichthyidae within krill swarms east of South Georgia. Acta Ichthyologica et Piscatoria. 13:1-11. Sneath, P. H. A., and R R. Sokal. 1973. Numerical taxonomy. W.H. Freeman, San Fran- cisco, CA, 573 p. Targett, T. E. 1981. Trophic ecology and structure of coastal Ant- arctic fish communities. Mar. Ecol. Prog. Ser. 4:243- 263. Tiedtke, J. E. and K-H. Kock. 1989. Structure and composition of the demersal fish fauna around Elephant Island. Arch. FischereiWiss. 39:143-169. Webster, R. 1973. Automatic soil-boundary location from transect data. Math. Geology 5:27-37. Westhoff, V., and E. van der Maarel. 1973. The Braun-Blanquet approach. In R. H. Whittaker (ed.), Ordination and classification of com- munities, p. 617-726. Handbook of vegetation sci- ence, no. 5. Junk, The Hague. Whittaker, R. H. 1960. Vegetation of the Siskiyou Mountains, Oregon and California. Ecological Monographs 30:279-338. Wiens, J. A. 1977. On competition and variable environments. Am. Sci. 65:590-597. Williams, W. T., and J. M. Lambert. 1961. Nodal analysis of associated populations. Na- ture 191:202. Yanez-Arancibia, A., F. A. Linares, and J. W. Jr. Day. 1980. Fish community structure and function in Terminos Lagoon, a tropical estuary in the southern Gulf of Mexico. In V. S. Kennedy (ed.), Estuarine perspectives, p. 465-482. Acad. Press, NY. Abstract.-A bioenergetic population model that integrated input on the abundance, distribution, sex- and age-structure, feeding rates, and diet of harbor seals was developed and used to estimate annual prey con- sumption in the Strait of Georgia during 1988. Owing to recruitment and mortality, the size of the Strait of Georgia population fluctuated sea- sonally from a minimum of 12,990 prior to the pupping season to a maximum of 15,810 following pup- ping. The study population repre- sented a population that was in- creasing at an intrinsic rate of 12. 5** per annum and was therefore skewed toward younger age-classes. Mean daily per capita gross energy requirements were estimated at 172 watts, of which 30% was lost in fae- ces, urine, and the heat increment associated with feeding, 42.39c was expended for basal metabolism, 23.4% for activity, 1.29c for body growth, and 3.29c for reproduction. Mean daily per capita food require- ments were estimated to be 1.9 kg, or 4.39c of mean body mass. Diet composition varied seasonally: hake was dominant during April- November and herring during De- cember-March. Combined, hake and herring accounted for 75% of the diet both in terms of energy and biomass. Total annual consumption was esti- mated at 9,892 (range 6,432-13,359) metric tons, which comprised 4,214 1 2.215-6.664 ) t of hake, 3,206 ( 1,679- 5,818) t of herring, 398 (171-846) tons of salmon, 335 (135-745) t of plainfin midshipman, 294 ( 137-556) t of lingcod, and the remaining 1,445 t of a wide variety of different prey. Gross and net population efficiency was estimated to be 3.99c and 2.29c in terms of energy, and 1.69c and 0.9% in terms of biomass. Annual prey consumption by harbor seals [Phoca vitulina) in the Strait of Georgia, British Columbia Peter F. Olesiuk Department of Fisheries and Oceans Biological Sciences Branch. Pacific Biological Station Nanaimo. British Columbia V9R 5K6 The recovery of many formerly depleted pinniped populations and the rapid expansion of fisheries has prompted concern over potential pinniped-fishery conflicts (Mate1; Beverton, 1982; Contos2; Melteff and Rosenberg, 1984; Beddington et al., 1985; Harwood and Croxall, 1988). The nature of these conflicts can be broadly categorized as operational (direct) or as ecological (indirect) in- teractions (e.g., Mate and Harvey, 1987). Operational interactions en- compass those that occur when pin- nipeds and fishing operations come in direct contact. For example, pin- nipeds may be injured or killed in fishing gear, and fishing gear and catches may be damaged by pinni- peds (Beach et al.3; DeMaster et al., 1982; Mate and Harvey, 1987). Since these interactions can be observed, they are relatively straightforward, at least conceptually, to assess through observer programs, logbooks, or in- terview surveys. Manuscript accepted 19 March 1993. Fishery Bulletin 91:491-515 (1993 1. 'Mate, B. R. 1980. Workshop on marine mam- mal fisheries interactions. U.S. Dep. Commer. Rep. PBH80-175144, 48 p. -'Contos, S. M. 1982. Workshop on marine mammal-fisheries interactions. Final report for MMC contract MM2079341-0. NTIS PB82- 189507, 64 p. 'Beach, R. J., A. C. Geiger, S. J. Jefferies, and S. D. Treacy. 1982. Marine mammal fisheries interactions in the Columbia River and adja- cent waters. 2nd Annual Rep., Nov. 1, 1980- Nov. 1, 1981. Natl. Marine Mammal Lab., NWAFC Proc. Rep. 82-04, 186 p. Ecological interactions, in contrast, encompass the indirect effects of pin- nipeds on fisheries and fisheries on pinnipeds. For example, sustainable harvest levels may be reduced as a result of pinniped predation on valu- able prey species, and the carrying capacity of pinniped populations may be reduced by human exploitation of their prey. Because these interactions are temporally or spatially displaced, or both, they are conceptually more difficult to assess than operational interactions. In their comprehensive review, Lavigne et al. ( 1982) outlined how the prey requirements of pinni- peds could be addressed using a bioenergetics approach, and how the approach could be extended to the population level. However, such popu- lation assessments require detailed information on the abundance, dis- tribution, sex- and age-structure, feeding rates, and diet of pinnipeds, as well as the interactions between these variables. For example, feed- ing rates (in biomass) are dependent upon the quality of the diet, and vary both with the sex and age of animals. The distribution and diet of pinni- peds may also be correlated if their seasonal movements are dictated by changes in the local and seasonal abundance of their prey. In British Columbia, recent con- cern over potential pinniped — fishery conflicts has focused mainly on the harbor seal (Phoca vitulina) (Malouf, 1986). Historic and recent manage- 491 492 Fishery Bulletin 91|3). 1993 ment issues concerning harbor seals in British Colum- bia are similar to those facing pinnipeds in other re- gions. Earlier this century ( 1913-64), harbor seal popu- lations throughout the province were maintained below historic levels by government-sponsored predator con- trol hunts and bounty kills (Fisher, 1952; Bigg, 1969). During 1963-68, populations were further depleted by an intense commercial harvest for pelts4. However, in 1970 the species was protected and since then harbor seal populations throughout British Columbia have been increasing exponentially at an annual finite rate of about 12.5% (Olesiuk et al., 1990a). Abundance in British Columbia is estimated to have increased from 9,000-10,500 in 1970 to 75,000-88,000 by 1988 (Olesiuk etal., 1990a). Investigations of harbor seals in British Columbia, particularly the Strait of Georgia, have provided much of the background information necessary to assess predator-prey interactions. Bigg (1969) described life history and population parameters, which were recently combined with information on the status of the popu- lation to construct life tables and a population model4. Aerial harbor seal censuses have been conducted in the Strait of Georgia periodically since 1966 and an- nually since 1982 to determine abundance and moni- tor population trends (Olesiuk et al., 1990a). Olesiuk et al. (1990b) recently described regional and seasonal variations in diet composition based on scat analyses. In addition to these local studies, the energetics of captive harbor seals and related phocids has been in- vestigated in some detail by other researchers, and feeding rates of free-ranging harbor seals have been estimated from the volume of prey in stomachs col- lected on the east coast of Canada (Boulva and McLaren, 1979). In this report, I synthesize these data into a bioen- ergetic population model and use the model to esti- mate annual prey consumption by harbor seals in the Strait of Georgia. The model serves several purposes. First, it provides improved estimates of the annual consumption by harbor seals of commercially and recreationally utilized species, which may be of par- ticular interest to resource managers who must base real-time management decisions on the best informa- tion currently available. Second, the model identifies the relative sensitivity of the prey consumption esti- mates to, and the present level of certainty in, each of its parameters, and is therefore useful for directing future research. Third, the model provides a frame- work for examining interactions between its individual components. For example, Olesiuk4 employed the model to predict the effects of demographic changes in har- bor seal populations on mean per capita energy re- quirements. Finally, the basic model serves as a foun- dation upon which additional refinements can be added as more accurate and detailed information becomes available. For example, the basic model is a prerequi- site for more elaborate models that incorporate the depensatory and compensatory responses of prey to their predators and predators to their prey (Beverton, 1985) or economic parameters (Clark, 1985). Materials and methods Study area Annual prey consumption was estimated for harbor seals inhabiting the Strait of Georgia (Fig. 1) during the 1988 calendar year. The Strait of Georgia was se- lected for intensive study because the area is readily accessible and supports a higher concentration of har- bor seals than other regions of British Columbia. Al- though the study area represents only 12% of the total British Columbia coastline, it supports 18-21% of the province's harbor seal population (Olesiuk et al, 1990a). 401esiuk, P. F. Population dynamics of the harbor seal {Phoca vitulina ) in British Columbia. Manuscr. in prep. Figure 1 Map of southwestern British Columbia showing boundaries of the Strait of Georgia study area (solid line I and scat collection sites i = estuaries; • = non-estuaries I. Olesiuk Prey consumption of Phoca vitulina 493 The Strait of Georgia also supports important com- mercial and sport fisheries. With respect to salmon, the most valuable fishery, the study area accounts for approximately 28% by number and 31% by weight of total escapement in British Columbia5. Ketchen et al. (1983) estimated that the Strait of Georgia accounted for 27% of the total non-salmonid commercial harvest by weight, 34% by landed value, and 26% by whole- sale value. The Strait of Georgia also accounts for about 96% of all recreational angler-days expended in the tidal waters of British Columbia (Sinclair6), which rep- resents about 600,000 boat-trips annually (Shardlow and Collicutt, 1989). Harbor seal-fishery conflicts in the Strait of Georgia would therefore be expected to be more intense than in other regions of the province. Two distinct types of habitat were recognized within the study area: estuaries and non-estuaries. Estuaries were defined as the large, shallow, soft-bottomed areas that occurred at the mouths of some rivers. Twelve estuaries (Fig. 1) within the study area were inhab- ited by appreciable numbers (>10) of seals. In estuar- ies, seals typically hauled out on logbooms or on tidal sandbars. Seals were widely distributed outside of es- tuaries and utilized 285 different haulout sites, most of which consisted of tidal boulders, reefs, islets, and ledges at the base of bluffs (Olesiuk et al., 1990a). The model Input parameters for the model consisted of informa- tion on harbor seal abundance, distribution, the sex- and age-structure of the population, diet composition, and daily energy and food requirements. The first four parameters were derived from local studies (Bigg, 1969; Olesiuk et al., 1990a; Olesiuk et al., 1990b; Olesiuk4), whereas the last parameter was deduced from the metabolic and feeding rates reported for phocids in the literature. The primary output from the model con- sisted of estimates of the annual consumption of each prey species. The annual consumption of a particular prey was estimated as the product of the abundance of seals, feeding rates, and the proportion of that prey in their overall diet. Daily food requirements were estimated separately for each sex-and age-class in the popula- tion, and mean per capita requirements calculated from energetic life tables by weighting the individual esti- mates according to the sex-and age-structure of the population. Feeding rates were assumed to be the same within and outside of estuaries and constant with sea- sEstimated from data provided by G. Hudson, Pacific Biological Sta- tion, Nanaimo, B.C., pers. commun. 1989. "Sinclair, W. F. 1972. The British Columbia sport fisherman. Can. Dep. Environ. Fish. Serv., Pac. Reg., 69 p. son. However, because the abundance and diet compo- sition of seals differed between estuaries and non- estuaries and varied seasonally in both habitat types (Olesiuk et al, 1990b), the abundance and diet data were stratified by habitat type and integrated over time. The biomass of the Mh prey species consumed annually, Bk, was thus estimated as Bt = FR (365 PFkl-dt + FR- 365 N,„ i Pnt,-dt [1] where FR is the mean per capita daily food require- ment, NEl and Not the number of seals within and out- side of estuaries, respectively, on the tth date ( ^Janu- ary 1st; 365=December 31st), and PEkl and Pm the proportion of the diet within and outside estuaries com- prised of the &th prey species on the tth date. The integrals were solved by calculating finite approxima- tions using daily \t increments. Abundance and distribution The abundance and distribution of harbor seals in the study area were determined by aerial censuses. In 1988, the entire study area was surveyed twice; once just prior to the pupping season (31 May-16 June) and again at the end of the pupping season (9-26 August). A third estimate of the population size in 1988 was obtained by projecting the population trends observed during censuses conducted between 1973 and 1988. Aerial censuses were conducted under standardized conditions under which maximum numbers of seals were hauled out: 1) the lower of semi-diurnal low tides; 2) between 08:00-11:30 AM; 3) usually toward the end of the pupping season which extends from early July to mid August (Bigg, 1969); and 4) not during inclem- ent weather such as rough seas or heavy precipitation. Under these conditions, the variability of replicate cen- suses indicated that counts represented, on average, 88.4% of the actual population (Olesiuk et al., 1990a). The confounding effects of seasonal fluctuations in the size of the population due to recruitment during the protracted pupping season were removed by ad- justing all census counts, C„ to post-pupping levels by applying a correction, P„ to account for pups born sub- sequent to the date, t, of the census. Since births are normally distributed over time with a mean Julian birth date, u, of 208 (27 July) and a of 16.1 days (Bigg, 1969; Olesiuk et al., 1990a), P, was calculated as f t- P, = 1+3 (f-u)2 26.0 months that were considered to have been collected in a ran- dom fashion (Bigg, 1969). Following Bigg ( 1969), speci- mens aged 0-5.9 months were excluded from the analy- sis because of potential sampling biases, specimens aged 6.0-17.9 months were tallied as 1-year-olds, speci- mens aged 18.0-29.9 months as 2-year-olds, etc. The exponential rate of decline, r, in the relative number of animals of each sex, S (/^female and w=male) in age- classes, iVsflJ, with age, X, was smoothed by regressing the logarithm of the estimated number of survivors in each age-class on age, such that Ws,x+u=-/Vs*e-". [6] The residuals from the log-linear regressions indicated that the rates of decline differed between sexes and Olesiuk Prey consumption of Phoca vitulma 495 were not constant with age (Bigg, 1969; Olesiuk4). The best fitting series of log-linear segments was therefore obtained iteratively by applying piecewise regressions with varying inflection points. Because pups aged <6 months were likely over-represented in Bigg's (1969) sample, their abundance, Nm, relative to older age- classes, Ns!x„ was calculated as MAf NSl0) = 0.5-l(Nflx:FEC,x [7] which assumes that the sex ratio at birth was equal. The number of seals in each sex- and age-class at the end of the 1988 pupping was estimated by normaliz- ing the relative Ns,x, series to sum to Np. Assuming that births occurred as a pulse at the beginning of the annual cycle and deaths throughout the year, the finite annual birth rate, (3, was estimated as P = (Nfi0l+Nm,0l MAf MA„, A"=l X=l [8] and the mean annual finite death rate, d, as d = Ho/tl+pD, 19] where a is the finite annual population multiplication rate (1.125; Olesiuk et al., 1990a) and MAs denotes the maximum ages attained by each sex (see Results). Rates of growth in body mass were calculated sepa- rately for each sex by fitting specialized von Bertalanffy curves (Zullinger et al., 1984) to the body mass at age data summarized in Bigg ( 1969). Ages were estimated to the nearest month by assuming that all animals were born in June. Owing to a small number of adult males in Bigg's (1969) sample, his data were supple- mented with data for 10 males aged 10-25 years col- lected in the Gulf of Alaska (Bishop, 1967). Param- eters of the specialized von Bertalanffy growth curves M., :A-(l-0.33-e,A'IA'-II,P, [10] where Mslx, represents the body mass of sex S at age X, and A, K, and / are the growth parameters, were esti- mated iteratively by a Quasi-Newton method (Fletcher7) using least squares criteria. Energetics Daily food requirements were estimated from two Tletcher, R. 1972. FORTRAN subroutines for minimization by quasi- Newton methods. AERER 7125. sources of data: 1) energetic parameters reported in the literature for harbor seals and related phocids based on captive studies; and 2) volumes of undigested prey in the stomachs of harbor seals collected on the east coast of Canada (Boulva and McLaren, 1979). In both cases, daily food requirements were estimated sepa- rately for each sex- and age-class in the population based on their mean body masses, and mean per capita requirements calculated from energetic life tables by weighting the individual estimates according to the sex- and age-structure of the study population. Mean body masses of age-classes were obtained by taking the geometric mean of their estimated masses at the beginning and end of the age interval (Eqn. 10), which assumes that growth was uniform throughout the year. Energetic parameters were calculated according to the International System of Units (ASTM, 1982). Where necessary, non-conforming values in the literature were converted by assuming that 1 calorie = 4.184 joule (J), such that lkcal-day-1 = 0.0484 J-sec1 or Watts (W). Where efficiency was not stated, net energy (NE) ex- penditures were transformed to gross energy (GE) re- quirements by assuming that overall efficiency was 70%; i.e. 13% of the GE in the diet, which usually consisted of herring, was not metabolizable (6% lost in feces and 7% in urine; Keiver et al, 1984) and 17% of GE was expended as the heat increment associated with feeding (Gallivan and Ronald, 1981). The basal metabolic rates of adult pinnipeds (Lavigne et al., 1986), like those of other adult mammals, con- form with Kleiber's (1975) relationship. The net basal metabolic rates of adults of aged X, BMRXX, in watts, was therefore estimated as BMR, 3.4-Af, (from Kleiber 1975) [11] where Ms(xl denotes the mean body mass (kg) of sex S at age X. Since other major components of the energetic budget also scale to M" 75 ( Lavigne, 1982 ), it is convenient to consider total energy requirements relative to BMR. A large portion of the overall energy budget of seals is expended on maintenance, which encompasses the energy required for basal metabolism, activity and ther- moregulation (Lavigne et al, 1982). Innes et al. (1987) provided one estimate of the gross maintenance re- quirements of non-growing, adult phocids based on the rates of energy ingestion in captivity, MR1„X, in watts: MRL 7.5-M , "71 (Eqn. 7 in Innes et. al. 1987) [12J The metabolic rates of juveniles, however, are usu- ally elevated relative to those of adults of equivalent mass (Innes et al., 1987). For mammals, the magni- tude by which juvenile metabolic rates are elevated generally declines from a peak at the onset of feeding 496 Fishery Bulletin 91(3), 1993 to adult levels by the onset of maturity (Kleiber, 1975; Brody, 1945). Data given in Worthy (1987a )'s Figure 4, together with his reported efficiency of 68%, indicates that the gross maintenance requirements of neonate harp and grey seals increases to about 1.8 x the pre- dicted adult MR1 (Eqn. 12) at the onset of feeding. In the absence of precise empirical data on the ontogeny of juvenile harbor seal metabolic rates, it was assumed that MR1 converged from a post-weaning maximum of 1.8X the predicted adult MR1 at the onset of feeding to adult MR1 levels at the onset of sexual maturity in an exponential fashion. A multiplier to account for el- evated metabolic rates of juveniles at a given age, JCF,lx„ was thus derived by calculating a series of 3- 4- 5- and 6-year exponential decays and weighting them according to the proportion of animals of each sex that matured at each of these ages. The correction for an age-class was calculated as the geometric mean of JCFs(x) at age X and age X+l, which assumes that the metabolic rate evolved at a constant rate throughout the year. The correction was used to cor- rect both basal (Eqn. 11) and maintenance (Eqn. 12) requirements. One of the potential shortcomings in directly ex- trapolating the maintenance requirements of captive seals, MR1, to free-ranging seals is that normal activ- ity patterns may be disrupted in captivity. For example, Innes et al. (1987) noted that some of the phocids in- cluded in their analysis were quiescent, and would thus be expected to have lower energy requirements than seals in the wild which spend a portion of their time foraging. A second estimate of gross maintenance requirements, MR2slz), was therefore derived by weight- ing the metabolic rates of swimming, SMRXIX„ and rest- ing, RMR,IX„ harbor seals according to a crude activity budget for free-ranging harbor seals: MR2slxl = (PH-SMRslxl) = (P/RMRXIXI) [13] where Ps and Pr denote the proportion of time seals spend swimming and resting. P, and Pr were set at 0.6 and 0.4 respectively based on the mean estimated per- centage of time free-ranging radio-tagged harbor seals spent hauled out on land: 44% (Sullivan, 1979), 35- 60% (Pitcher and McAllister, 1981) and 37% (Yochem et al., 1987). Age-specific swimming metabolic rates, SMRslxl, were inferred (see Results) from the swim- ming metabolic rates of captive harbor seals (Davis et al., 1985; Williams, 1987). Resting metabolic rates, RMRSIX„ were assumed to be equivalent to BMR,(XI (ap- propriately elevated for juveniles). Since the extreme air and sea temperatures in the study area were likely within the thermoneutral zone, thermoregulatory costs were assumed to be negligible (see Results and Gen- eral Discussion I. In addition to maintenance requirements, growing animals require energy for body growth. Daily energy requirements for growth for each sex- and age-class, DGR„U, were calculated as DGRM = CGGIS, [14] where CG is the apparent gross cost of growth, 201 W(kgd')-1, as given in Innes et al. (1987), and GIslxl the daily growth increment of each sex- and age-class (i.e. [Mste+1)-AfsW]/365). Finally, mature seals may invest additional energy in reproduction. For females, the total costs of repro- duction were partitioned into: 1) foetal development; and 2) nursing. The net energy invested in foetal de- velopment was estimated from the mass and energetic density of term fetuses and the placenta. The net en- ergy invested in lactation was estimated indirectly from the amount of energy transferred to nursing pups as reflected by changes in the total mass and body com- position of pups between birth and weaning and their maintenance requirements, MR, while nursing. The energy content of the placenta and the energetic den- sities of neonate carcasses and of the mass gained dur- ing nursing were extrapolated from those reported for harp seals (Worthy and Lavigne, 1983). The MP-of nursing pups was assumed to be the same as that of adults of equivalent mass (2.0 x BMR) and growth of pups was assumed to be linear while nursing. Gross reproductive costs were estimated from net costs by assuming that the net efficiency of mothers was 70% (see above) and that for lactation, 95% of the energy in milk transferred to pups was metabolizable (Oftedal and Iverson, 1987). Since females deplete blubber re- serves accumulated during the non-breeding season to meet these costs (Pitcher, 1986), age-specific daily re- productive requirements, DRRlxl, were estimated by amortizing the annual cost over the entire year and applying it to female age-classes based on their fecun- dity rates, FECtxl. Since harbor seals are promiscuous and males are not known to fast or fiercely compete for breeding rights (Bigg, 1981), reproductive costs for males were assumed to be negligible and absorbed into their daily maintenance requirements. Two estimates of daily food requirements, FRlx, in kg, were derived. The first estimate, FRlslxl, was ob- tained by summing the components of the energetic budget (MRstII, DGRslxl and, for females, DRRlxl) to de- termine the total daily energy requirements, DERslxl. Estimates for each sex- and age-class were derived by taking the geometric mean of the parameters at age X and X+l, which assumes that the parameters changed at a constant rate throughout the year. DERSIX, was subsequently converted to units of biomass, FRlslxl by dividing it by the mean weighted energetic density of Olesiuk Prey consumption of Phoca vituhna 497 the diet. A second estimate, FR2sll), was obtained di- rectly from Boulva and McLaren's (1979) relationship between daily ingestion rates and body mass: FR2^tu = 0.089-Ms, (from Fig. 2 in Boulva and McLaren, 1979) [15] based on the amount of undigested prey found in the stomachs of harbor seals collected on the east coast of Canada. Diet composition The diet of harbor seals in the study area was deter- mined by scat analyses. During 1982-88, a total of 2,841 scats (216 collections) were collected from 58 sites (11 estuaries and 47 non-estuary sites) distrib- uted throughout the Strait of Georgia (Fig. 1). Samples were collected in all months of the year (Fig. 2). Be- cause most major haulout sites were sampled, the sampled sites accounted for about 45% of all seals ob- A NUMBER OF SAMPLES 400 350 300 250- 200- 150 - 100- 50 - | NON-ESTUARIES |3 ESTUARIES ' * B NUMBER OF COLLECTIONS 30- 25 20 - 15 - 10 5- ^] NON- ESTUARIES [^"j ESTUARIES , ; . * Hist colle and J FMAMJJ ASOND MONTH Figure 2 ograms showing (A) number of harbor seal scat sample cted; and (Bl number of scat collections by month withi outside of estuaries. s served during censuses of the study area both in May- June and in August, 1988. The collections thus pro- vided a representative sample from which regional and seasonal variations in diet composition could be assessed. Undigested prey remnants were recovered from the scat samples with an elutriator (Bigg and Olesiuk, 1990). Elutriation recovery rates ranged between 90- 100% (X=98.6%) for various fish structures and 70- 100% (X=85.0%) for cephalopod beaks. In contrast to previous scat studies which have relied almost exclu- sively on otoliths to identify prey, harbor seal prey were identified by using a wide array of different struc- tures including otoliths, teeth, scutes and scales, as well as numerous cranial, appendicular, axial and cau- dal elements (see Appendix I in Olesiuk et al., 1990b). The relative importance of prey in the diet was mea- sured by using a new index, termed split-sample fre- quency of occurrence, designed specifically for scat analyses (Olesiuk et al., 1990b). The index was predi- cated on two assumptions: 1) prey identified in scat samples represented all those consumed in the previ- ous meal (i.e., 24-hour period); and 2) all prey species constituting a meal had been consumed in equal vol- umes. Thus, the proportion of the diet comprised of the /jth prey species in the jth strata (j=E for estuaries and .7=0 for outside estuaries) in the mth month (m=l for January, etc.), Pjkm, was estimated from the i=l,..,n samples collected from that strata in that month by * jkn BO ijkm' ®0„kJn k: ijkm 1,..,N (N=# different prey species) [16] where Ol]km is a binary variate that indicates whether the £th prey species was absent or present (0=absent and l=present) in the z'th sample collected in thej'th strata in the mth month, such that Y.0„km for k=\,..JSl represents the total number of prey species present in the ith sample. Therefore, if only one prey species oc- curred in a sample, its occurrence was scored as 1, if two prey species occurred each occurrence was scored as 0.5, and so forth. The split-sample index is consid- ered an improvement over conventional frequency of occurrence indices in that prey species that comprised an entire meal, which had presumably been consumed in large quantities, weighted the split-sample index more than prey species consumed in a diverse meal comprising many species, each of which had presum- ably been consumed in smaller quantities. Estimates of the diet composition on the tth date, PEkl and P0kl, were derived by linearly interpolating between the monthly estimates, PEkni and Pok„„ plotted at the mid- point of each month. 498 Fishery Bulletin 91(3), 1993 Although it was not possible to assess the assump- tion that all prey in a meal had been consumed in equal volumes (see Results), the maximum potential biases introduced by deviations from the assumption could be determined. An upper limit of the importance of a particular prey species in the overall diet was obtained by assuming that whenever the species oc- curred in a meal, it comprised the entire meal and that all other species in the same meal had been con- sumed in negligible quantities. Conversely, a lower limit was obtained by assuming that whenever a particular species was consumed in the same meal along with other species, it had been consumed in negligible quan- tities. Mathematically, the upper and lower limits for the /th of k=l,...JV species were calculated from Equa- tion 16 by 1) Setting Oijkm=l for k=l and 0„,„ =0 for k±l; 2) Setting Oijk„ =0 for k=l and Oijkm=l for kd when N>1, respectively. Two corrections were applied to the split-sample in- dex to account for suspected biases. First, very small scat samples, which undoubtedly represented only a small fraction of whole scats and probably contained only a portion of all the prey species actually con- sumed in meals, were weighted less than large scat samples. Second, the relative proportions of hake and herring, the two predominant prey species, in samples that contained both species were volumetrically weighted based on the mean relative number of ele- ments of each species in the sample compared with the relative number in samples that were composed exclusively of each of these two species. Both correc- tions, each of which had a relatively minor influence on the results, are described in detail in Olesiuk et al. (1990b). Results Abundance and distribution The two complete censuses of the study area in 1988, when adjusted to post-pupping levels and corrected for missed seals (Eqn. 3), yielded population estimates of 16,531 and 15,091, respectively. Their mean, 15,810, was adopted in subsequent analyses. The mean was considered most appropriate because the proportion of the population counted during any single census may have been lower or greater than the mean estimate of 0.884. The validity of the 1988 census estimate was substantiated by a projected population estimate of 15,050 based on population trends during 1973-88, over which period the population had been steadily increasing (log-linear r2=0.994) at a rate of 12.5% per annum (Olesiuk et al., 1990a). Seasonal fluctuations in the size of population as a result of mortality and recruitment are shown in Fig- ure 3A. The total size of the population ranged from a low of 12,990 just prior to the pupping season (16 June) to a peak of 15,810 at the end of the pupping season (6 September). The mean size of the population was 14,270. Because the population was below carrying ca- pacity and increasing at its intrinsic rate, it was 12.5% larger at the end than at the beginning of the year. Seasonal changes in the distribution of harbor seals between estuaries and non-estuaries are shown in Fig- ure 3B. The overall abundance within estuaries was lowest (4% of the total population) in December and remained low throughout the winter and into early spring. Numbers increased sharply during June- August, mainly due to large influxes into the two larg- est estuaries — Boundary Bay and lower Fraser River (Fig. 1). These influxes coincided with the pupping sea- son (Bigg, 1969; Olesiuk et al, 1990a) which might indicate these two large estuaries were preferred whelp- ing areas, as has been reported for other major es- M J J A MONTH Figure 3 Seasonal fluctuations in (Al the total size of the Strait of Georgia harbor seal population, N„ due to recruitment and mortality; and (B) the total number of seals inhabiting estu- aries, NRv Note that the two parameters are plotted on differ- ent scales. Olesiuk: Prey consumption of Phoca vitulina 499 tuaries along the coast of Washington and Oregon ( Jefferies8). Alternatively, these influxes also coincided with the earliest low tides of the year in which the sandbars utilized as haulouts were exposed during day- light hours and may therefore merely indicate that seals preferred to occupy areas where they could haulout during daylight. Although numbers in Boundary Bay declined after August (i.e., the end of the pupping season), the over- all proportion of the population in estuaries continued to increase and peaked at 18% in September. This was due largely to a migration of seals from Boundary Bay to the Fraser River and an influx of seals into many of the smaller estuaries, where peak abundance gener- ally occurred in September-November coinciding with the return of spawning salmon to their natal rivers (Olesiuk et al., 1990b). Weighted seasonally, 10.3% of the total population inhabited estuaries. Population parameters As expected, the relative number of animals in age- classes decreased with age (Fig. 4). However, piece- wise log-linear regressions indicated that the rate of decline for both sexes changed abruptly at 4 years of age, which roughly coincided with the onset of sexual maturity (see Tables 1 and 2). The regressions for fe- males and males aged 1-4 years were not significantly "Jefferies, S. J. 1986. Seasonal movements and population trends of harbor seals (Phoca vitulina richardsi) in the Columbia River and adjacent waters of Washington and Oregon. Final Rep. Mar. Mam- mal Comm., Contr. No. MM2079357-5, 41 p. A PUPS • JUVENILES 0 ADULT FEMALES £ ADULT MALES AGE (yearsl Figure 4 Exponential rates of decline in the size of age-classes of har- bor seals as a function of age (Olesiuk1). The trend lines represent piecewise log-linear functional regressions fitted by least squares criteria and scaled to an initial cohort size of 1000. different (2=0.18; P>0.50), so juveniles of both sexes were pooled. For females, the rate of decline decreased beyond age 4 years, whereas for males the rate in- creased beyond age 4 years: XT _/V p-0.217S.t AT_ _IU p-0.1653.t *'/r«« iyftx)c AT -XT p-0.2878.1 for both sexes aged 1-4 year [17] for females aged >4 years [18] for males aged >4 years [ 19] The Ns,xl series were truncated at 29 years for females and 20 years for males. The truncation points, MA: and MAm, represent the oldest specimens collected by Bigg (1969) and also the ages by which the size of age- classes diminished to less than 0.5% the number of new recruits (see Tables 1 and 2). It should be noted that the rates of decline in the size of age-classes do not entirely reflect mortality because the population was non-stationary, such that the number of seals be- ing recruited (i.e., the initial size of cohorts) had in- creased over time. When corrected for an intrinsic rate of increase of 12.5% per annum (Olesiuk et al., 1990a), the exponential decays represent finite annual mortal- ity rates of 9.5% for juveniles aged 1-4 years, 4.6%> for adult females aged >4 years and 15.6% for adult males aged >4 years (Olesiuk3). The finite per capita birth rate, (3, was calculated to be 29.8%, which implied that mortality during the first year was 27.0%. The finite per capita mortality rate, d, was subsequently esti- mated to be 13.3%. The estimated number of seals in each sex- and age- class indicates that the population was markedly skewed toward younger age-classes (see Tables 1 and 2). The mean age was only 4.0 years (4.7 years for females and 3.2 years for males). A total of 74% of all individuals were aged <5 years and 91% were aged <10 years. As noted previously, this skewed age- structure was not due entirely to high mortality ( mean life expectancy was 8.2 years — 10.4 years for females and 6.0 years for males; Olesiuk4), but also because the population had been exponentially increasing since 1970 (Olesiuk et al, 1990a). Given a growth rate of 12.5%, the total population numbered only 8,770 in 1983 and 4,870 in 1978, such that the initial size of the 1983 cohort, represented by 5-year-olds in the popu- lation in 1988, was only 55% the intial size of the 1988 cohort, and the initial size of the 1978 cohort, repre- sented by 10-year-olds in the population in 1988, only 31% the initial size of the 1988 cohort. In other words, the population in 1988 was skewed toward younger animals because most of its constituents had been re- cruited in recent years. Bigg (1969) reported that mean body mass of pups increased from 10.2 kg at birth to 24.0 kg by the end of the 5-6 week nursing period, but there appeared to be little further increase in body mass during the 500 Fishery Bulletin 91(3), 1993 Table 1 Energetic life table for males. Parameters are: Nmtx, number in age-class; A/,,,,,, mean body mass (kg! MAT,,, , proportion of animals mature; JCF„„X multip ier to correct for elevated juvenile me tabolic rates; BMRm ,„ basal metabolic rate; MR1 „,» MR2„„„ and MRMX, gros s maintenance requirements based on ingestion rates in •aptivity, d< lily activity budget s, and the r average respectively; DGRmlx, mean daily growth requirements; DER,„ .,. mean daily total energy requirements; and FRl„„xn FR2m,x, and FR„„X daily food requirements (kg) based on energetic life tables volumes of stomach contents , and their average respectively. Parameters are given for the population of 15,810 at the end of the pupping season. Ml energet c parameters are in watts. Age Nm» M„„x, MAT, JCFm , BMR„„X, MR1 MR2 ~MR DGR DERmlxl FR1„„X, FR2m„, ~FR„,.„ 0 1814 24.0 0.00 1.80 67.1 130.3 170.5 150.4 2.9 153.4 1.99 1.07 1.53 1 1178 29.3 0.00 1.60 71.4 137.3 181.4 159.4 5.3 164.6 2.14 1.29 1.71 2 948 38.9 0.00 1.42 76.9 146.3 195.3 170.8 5.5 176.3 2.29 1.57 1.93 3 762 48.9 0.06 1.26 79.5 150.2 202.0 176.1 4.7 180.8 2.35 1.82 2.08 4 613 57.5 0.29 1.13 80.4 151.1 204.3 177.7 3.9 181.6 2.36 2.02 2.19 5 460 64.5 0.53 1.05 81.5 152.5 207.0 179.7 3.1 182.9 2.38 2.18 2.28 6 345 70.2 1.00 1.00 84.1 156.8 213.5 185.2 2.5 187.6 2.44 2.31 2.37 7 259 74.6 1.00 1.00 87.5 162.9 222.3 192.6 1.9 194.5 2.53 2.40 2.46 8 194 78.1 1.00 1.00 90.1 167.5 228.9 198.2 1.5 199.7 2.59 2.47 2.53 9 145 80.8 1.00 1.00 92.1 171.0 234.0 202.5 1.1 203.6 2.65 2.53 2.59 10 109 82.8 1.00 1.00 93.7 173.7 237.9 205.8 0.9 206.6 2.68 2.57 2.63 11 82 84.4 1.00 1.00 94.8 175.7 240.8 208.3 0.6 208.9 2.71 2.60 2.66 12 61 85.5 1.00 1.00 95.7 177.2 243.0 210.1 0.5 210.6 2.74 2.63 2.68 13 46 86.4 1.00 1.00 96.3 178.4 244.7 211.5 0.4 211.9 2.75 2.65 2.70 14 34 87.1 1.00 1.00 96.8 179.2 245.9 212.6 0.3 212.9 2.77 2.66 2.71 15 26 87.6 1.00 1.00 97.2 179.9 246.9 213.4 0.2 213.6 2.77 2.67 2.72 16 19 88.0 1.00 1.00 97.5 180.4 247.6 214.0 0.2 214.1 2.78 2.68 2.73 17 15 88.3 1.00 1.00 97.7 180.7 248.1 214.4 0.1 214.5 2.79 2.68 2.74 18 11 88.5 1.00 1.00 97.8 181.0 248.5 214.7 0.1 214.8 2.79 2.69 2.74 19 8 88.7 1.00 1.00 97.9 181.2 248.8 215.0 0.1 215.1 2.79 2.69 2.74 20 6 88.8 1.00 1.00 98.0 181.3 248.9 215.1 0.1 215.2 2.80 2.69 2.74 Table 2 Ener getic life table for females. Parameters are the same as described in Table 1 with the following addition! : FEC„ annual fecundity rate; and DRR , mean daily gross rep •oductive requirements. -"ammeters are jiven for the population of 15,810 at the end of the pupping season. All energetic parameters are in watts. Age Afo Mrw MAT,,, JCF& BMR,ix, MRlflxl MR2flx, ~MR7 years was 60.0 kg and the mean of the 8 male specimens aged >9 years was 86.6kg (Fig. 5). The mean body mass of all ages, calculated at the end of the pupping season and weighted according to the stable age-structure, was 44.2 kg for females, 45.3 kg for males, and 44.7 kg overall. Energetics Energetic life tables are shown in Tables 1 and 2. The main advantage of the energetics approach over stom- ach volume analyses was that total energy require- ments could be partitioned into energy required for basal metabolism, activity, growth and reproduction. This enabled an assessment of the relative magnitude and uncertainty associated with each component of the energy budget. The estimated gross maintenance requirements of adult harbor seals, based on the rate of energy inges- tion by captive phocids, MRl^, ranged from 1.85- 1.94X (X=1.90x) the predicted adult BMR (Eq. 11). IOO- A A 4— A— * 80- $f ° A ¥>i _ 60- rX^ P°. 0 40 - I f 20- o FEMALES Wt(x ,= 65.97 (l-|/3e-a363<*-°-347))3 A MALES- Wt (x) 89.16 Cl-l/3e"°-287(x-°-880')3 0 - AGE (years) Figure 5 Mean body mass (±SE) as a function of age. Data are from Bigg (1969) supplemented with data from Bishop (1967). Growth curves represent specialized von Bertalanffy equa- tions fitted to the data by least squares criteria. Multipliers to account for the elevated metabolic rates of juveniles, JCFslxl ranged from 1.8 at weaning to 1.0 at the onset of maturity. The corrected juvenile main- tenance requirements, MRls,,„ thus declined from 3.5 x the predicted adult BMR at weaning, to 3.1x the pre- dicted adult BMR at age 1.0 years, to 2.6 x the pre- dicted adult BMR at age 2.0 years, and ultimately to 1.9x the predicted adult BMR at the onset of matu- rity. These predictions appear to be consistent with the metabolic rates of juveniles reported in the litera- ture. Keiver et al. ( 1984) reported that the gross main- tenance requirements of harp seals aged 5-24 months were 2. 1-3.0 x predicted adult BMR levels and data in Innes (1984) indicate that the gross maintenance re- quirements (assuming 70% net efficiency) of grey and harp seals aged 1-28 months were 2.9 x predicted adult BMR levels. The unweighted mean correction for juve- niles of all ages was 1.34, which was similar to the 1.4 derived by Innes et al. (1987) for juveniles of various ages pooled. Similarly, the average MR1SIXI for harbor seals aged 0-3 years was 2.9 x the predicted adult BMR, which was similar to the 2.8 x adult BMR esti- mated for captive harbor seals aged 0-3 years by Markussen et al. (1990) who, interestingly, found no evidence of age-specific changes in maintenance re- quirements between 0 and 3 years of age. A second estimate of maintenance requirements, MR2, was obtained by weighting swimming and rest- ing metabolic rates, SMR and RMR, according to an activity budget for free-ranging harbor seals. Davis et al. ( 1985) reported that the SMR of a 63-kg adult har- bor seal swimming at 1.4 msec1, its preferred swim- ming speed, was 2.3x its expected BMR (i.e., 2.2X its resting metabolic rate which was reported to be l.lx BMR). Similarly, the net SMR of an 85-kg adult har- bor seal swimming at the same speed was 210 watts (Williams, 1987), or 2.4x its predicted BMR. In con- trast, a 33-kg yearling swimming at 1.4 msec _1 exhib- ited a relatively higher net swimming metabolic rate of 170 watts (Davis et al., 1985). Although this was 3.6X the predicted BMR for an adult of equivalent mass, it was only 2.3 x the corrected BMR for a year- ling (i.e., the juvenile SMR appeared to be elevated to the same extent as its total MR1 ). Age-specific SMfl(„'s were therefore assumed to be 2.3 x BMRSIX, (appropri- ately elevated for juveniles ) for all ages and both sexes. Assuming that seals spent 40% of their time resting on land and 60% swimming and that the RMR,x's were equivalent to BMRslxl (again appropriately elevated for juveniles), the net MR2SIX, was estimated to be 1.8x the corrected BMRilxl, and the gross MR2s!x, to be 2.5 X the corrected BMR,IXI. The second estimate of gross maintenance require- ments, MR2, was approximately 34% (range 31-37%) greater than the first estimate, MR1, for all age-classes. 502 Fishery Bulletin 91(3), 1993 As noted earlier, MR1 probably tended to be an under- estimate as it was derived from captive seals that were sometimes quiescent. On the other hand, MR2 may be an overestimate as it is unlikely that seals swim con- tinuously while in the water, and also because the RMR of sleeping seals may actually fall below BMR (Ashwell-Erickson and Eisner, 1981; Worthy, 1987a). The mean of MRlKlxl and MR2XIX„ denoted as MRS(X), was therefore adopted for subsequent calculations. Ambient sea and air temperatures in the study area were probably within the thermoneutral zone of free- ranging harbor seals. Oritsland and Ronald (1975) de- tected no change in the metabolic rate of an adult harp seal swimming in water temperatures ranging from 8.5-26°C and Gallivan (1977) no change in the metabolic rates of 3 adult harp seals in water tem- peratures ranging from 1.8 to 28.2°C. Hart and Irving (1959) reported that the lower critical temperature of harbor seals in air was 2°C and Matsuura and Whittow (1973) found that the metabolic rate of a resting har- bor seal was constant in air temperatures up to 35°C. Mean monthly sea surface temperatures in the study area typically range between 6.2 and 17.0°C/' and mini- mum and maximum monthly air temperatures between -0.4 (January minimum) and 23.5°C (July maximum) (Canadian Hydrographic Service, 1990). Thermo- regulatory costs were therefore assumed to be negli- gible (see also General Discussion). The costs associated with growth were calculated based on actual growth rates rather than applying Innes' et al. (1987) empirical equations for growing phocids because there may be substantial differences between growth rates of captive and free-ranging seals. For instance, growth rates reported for recently weaned grey and harp seals in captivity (Worthy, 1987a) were about an order of magnitude greater than those esti- mated for harbor seals in the wild. The apparent gross cost of growth, 201 WfkgdT1 (Innes et al., 1987), rep- resents, assuming net efficiency was 70%, a net cost of 141W(kg-d-')-' or 12.2 MJ-kg-1. Given the wet-weight energetic density of tissues (37.8 MJkg1 for blubber and 6.5MJkg~' for proteinaceous tissue; Olesiuk and Bigg, unpubl. data), this implies that post-weaning body growth was composed of about 20% fat and 80% pro- tein. Using these values to extrapolate the estimated body composition at weaning (see below), the adult body would be composed of approximately 30% blub- ber, which is consistent with the 27-30% reported for free-ranging harbor seals (Pitcher, 1986). Daily growth requirements, DGi?s,,„ were low for all age-classes, ranging from about 2.5-5.5 watts for juveniles to neg- ligible values for adults. l|H. Freeland, Institute of Ocean Sciences, Pat Bay. B.C.. pers. commun. 1989. In calculating the energy invested in foetal develop- ment, the energetic density of the fetus was assumed to be the same as the 7.2MJ-kg~' reported for harp seal neonates (Worthy and Lavigne, 1983). This im- plies that neonates are essentially devoid of fat, which is probably true as neonate and near-term harbor seals have very thin blubber layers (Pitcher, 1986). In addi- tion, neonates can tolerate little mass loss before dy- ing (Boulva and McLaren, 1979). Applying this value to the mean mass at birth of 10.2kg (Bigg, 1969), the total energy content of the term foetus is estimated to be about 73.4 MJ. If it is assumed that, as in harp seals (Worthy and Lavigne, 1983), the placenta con- tains an additional 5.9 MJ, the net foetal investment is estimated at 79.3 MJ, and the gross investment at 113 MJ. Additional energy would be required for the metabolism of the fetus. However, the foetal mass would represent only a negligible portion of a female's total mass through most of the pregnancy. Moreover, since the foetal masses were not subtracted from the total masses of pregnant females incorporated into the growth curves, it was assumed that the costs were absorbed into the maintenance requirements of adult females. In calculating the costs invested in lactation, the energetic density of the mass gained by pups was as- sumed to be the same as the 33.1 MJkg-1 reported for harp seal pups (Worthy and Lavigne, 1983), which im- plies that the mass gained was approximately 85% fat and 15% protein. In view of their rapid rate of growth, this value is probably also applicable to nursing har- bor seal pups. When applied to the 13.8 kg increase between the mean birth and weaning mass (10.2 and 24.0 respectively; Bigg, 1969), the energy assimilated by nursing pups is estimated to be 456.8 MJpup '. In addition, each nursing pup would require about 172.4 MJ for maintenance during the 5-1/2 week nurs- ing period (Bigg, 1969). Thus, the total net nursing investment was estimated to be 629.2 MJ, which rep- resents a gross investment of 946 MJ. The total an- nual cost of foetal development and lactation was thus estimated to be about 1060 MJ for each reproductive female. Estimates of the total daily energy requirements, DERX, were surprisingly constant with age, ranging from 150 W for yearlings of both sexes to 215 W for full-grown males (Tables 1 and 2). The range in DER,,,, was much narrower (1.4x) than the range in body mass (3.7x) because the major energy expenditures scaled to M"7\ and also because juvenile metabolic rates were elevated relative to adults of equivalent mass. The mean per capita DER was estimated to be 172 W. Most of the daily energy requirement, DER, was expended for maintenance and comparatively little for Olesiuk: Prey consumption of Phoca vitulina 503 production. Daily growth requirements accounted for, on average, only 1.2% of the total gross DER (1.7% of NE), and only 2.1% (3.0% of NE) of the gross DER for seals aged 1-2 years which exhibited the highest growth costs. However, the estimates reflect only the direct costs associated with growth. A portion of the elevated metabolic rates of juveniles, which account for 8.4% of the overall population energy budget (12.0% of NE) but were incorporated into maintenance costs, may be indirectly associated with growth. Although the DER of lactating females were 2.8 x those of non- lactating females of equivalent mass, and lactation ac- counted for 89% of the total costs of reproduction, lac- tation accounted for only 14.6% of the overall energy requirements of reproductive females when amortized over the entire year. Overall, net reproductive costs accounted for only 3.2% of the total population energy budget (4.5% of NE). Thus, growth and reproduction combined accounted for only 4.4% of the total popula- tion energy budget (6.2% of NE). With respect to maintenance, basal metabolism ac- counted for 33.9% of the total population energy bud- get (48.4% of NE) when the corrections for elevated juvenile levels are excluded, or 42.3% (60.4% of NE) when the corrections are included. If it is assumed that 13% of gross maintenance requirements is lost in faeces and urine and 17% expended in the heat incre- ment associated with feeding, and thermoregulatory costs were negligible, the remaining 23.4% of the total population energy budget (33.4% of NE) was expended on activity. However, there is considerable uncertainty in this estimate. Had total maintenance costs been directly extrapolated from captive animals using MR1, only 16.1% of the total budget (23.0% of NE) would have been available for activity. On the other hand, had total maintenance been estimated from the activ- ity budget using MR2, 29.0% of the total budget (41.4% of NE) would have been available for activity. Averag- ing MR 1 and MR2 thus introduced a potential error of about ±13% into the overall population energy budget. Based on the energetic densities of 10 of the 15 most important prey species, which accounted for 86.1% of the overall diet, the mean weighted energetic density of the diet was estimated to be 6.65 MJ kg-1 (see Table 3). The total daily energy requirements, DER^x:, there- fore translated into daily food requirements, FR1^,X„ ranging from 2.0 kg for yearlings of both sexes to 2.8 kg for full-grown males (Tables 1 and 2), which repre- sented 8.2% and 3.1% of their mean body masses re- spectively. The mean daily per capita food requirement, weighted according to the sex- and age-structure of the population, was estimated to be 2.2 kg, or 5.0% of mean body mass. Estimates of daily food requirements based on the volumes of stomach contents, FR2s!l), were consistently lower than the estimates based on energetic life tables, FR1S,XI (Fig. 6). The mean per capita food requirement estimate based on stomach volumes, FR2, was 1.6 kg, or 73% of the mean estimated of 2.2 kg based on ener- getic life tables, FR1. It is unlikely that much of the discrepancy between the two estimates can be attrib- uted to differences in the energetic density of diets on the east and west coasts. A crude estimate of the mean energetic density of the east coast diet, obtained by applying energetic densities of similar species on the west coast to 10 of the 15 most important prey on the east coast, which combined accounted for 81.2% of the east coast diet, was 6.31 MJ kg-1, or 95% of the diet in the study area. The magnitude of the discrepancy between FRl,,^, and FR2xlxl declined with age (Fig. 6). There was fairly good agreement between the two methods for adults: 2.1kg from stomachs versus 2.5 kg from energetic life tables. The former was almost within the ±13% poten- tial error introduced in the latter owing to uncertainty in the costs associated with activity (see above). In contrast, juvenile estimates based on stomach volumes were substantially lower than corresponding estimates based on the energetic life tables (Fig. 6). The average per capita requirement for immature seals based on stomachs, 1.3 kg, was only 65% of the estimate based on energetic life tables, 2.1kg. The estimate based on stomach volumes for yearlings, the age-class exhibit- ing the greatest discrepancy, was only 52% of the esti- mate based on energetic life tables. The reason for the larger discrepancy for immature age-classes was that the volumes of stomachs, when all ages were pooled, was scaled to M076 (Eqn. 15), which was close to the M075 expected for adults (Eqn. 11). In other words, there was no evidence that juveniles consumed greater quantities of prey than adults of equivalent mass, whereas significant corrections were applied in the en- ergetic life tables to account for the elevated metabolic rates of juveniles. One factor that might have contributed to this dis- crepancy may have been post-weaning changes in body composition. At weaning, about 50% of body mass of pups is composed of blubber which, as the animal ages, is used and displaced by proteinaceous tissue. Thus, although pups exhibited some post-weaning growth in body mass during the first year, they may actually have experienced negative growth in energetic terms. However, even if it is assumed that all of the post- weaning blubber reserves (37.8MJ-kg~') were replaced with proteinaceous tissue (6.5 MJ-kg~') during the first year, the daily energy requirement of yearlings would decrease by only 0.15 kg day-1, or 15% of the observed discrepancy. The discrepancy between FR1 ,. and FR2 , for juve- niles suggests that the metabolic rates of free-living 504 Fishery Bulletin 9 1 (3). 1993 Table 3 Contribution to the overall diet (upper and lower limits) estimated annual consumption in metric tons (range). and energetic density (MJkg ') of harbor seal prey in the Strait of Georgia. The range in annual consumpt on was calculated by assuming that daily food requirements were within ±35% of the point estimate and that realistic upper and lower limits of the importance of prey in the diet were half the width of the extreme limits (see Discussion). Energetic densities are from Perez and Bigg ( 1986) converted to joules assuming 1 calorie = 4.184 joules. Prey Percentage of Annual prey Energetic species overall diet consumption (t) density Pacific hake iMerluccius productus) 42.6 (26.3-57.2) 4,214 (2,215-( 5,664) 4.90 Pacific herring iClupea pallasi) 32.4 (19.8-54.7) 3,206 (1,679-, 5,818) 9.08 Pacific salmon iOncorhynchus spp.) 4.0 1.3- 8.6) 398 ( 171- 846) 8.41 Plainfin midshipman iPonchthys notatus) 3.4 0.8- 7.8) 335 ( 135- 745) — Lingcod iOphiodon elongatus) 3.0 1.3- 5.4) 294 ( 137- 556) — Surfperches (Family Embiotocidae) 2.3 0.5- 5.4) 230 ( 91- 520) — Cephalopods (Class Cephalopoda) 2.1 0.0- 5.9) 208 ( 68- 535) '5.06 Flatfishes (Order Pleuronectiformes) 1.2 0.5- 2.9) 123 ( 57- 284) 5.02 Sculpins (Family Cottidae) 1.2 0.1- 3.1) 114 ( 40- 276) — Rockfishes (Family Scorpaenidae) 1.1 0.4- 2.4) 112 ( 50- 241) 4.90 Pacific tomcod (Microgadus proximus) 1.0 0.6- 1.4) 101 ( 53- 164) — Walleye pollock iTheragra cha/cogramma) 1.0 0.6- 1.3) 97 ( 50- 151) 5.90 Pacific sand lance {Ammodytes hexapterus) 0.8 0.4- 2.1) 79 ( 39- 193) 5.10 Pacific cod (Gadus macrocephalus) 0.5 0.3- 0.7) 54 ( 28- 87) 4.18 Smelts; mainly eulachon (Family Osmeridae) 0.4 0.3- 1.8) 40 ( 23- 149) 25.90 Unidentified/other fishes 2.7 1.0- 3.0) 267 ( 119- 380) — Other invertebrates 0.2 0.0- 0.6) 20 ( 7- 54) — Total consumption / Weighted mean energetic density 9,892 (6,432-13,359) 6.65 'Mean of values for market and gonatid squid. 2Value for eulachon, the dominant smelt in the diet. 3.0 y REQUIREMENTS (kgs) 6 en * V / Q 1.5 - O O u. y 5 '-0- Q / 0.5 ; 10 40 60 80 IC O BODY MASS (kgs) Figure 6 Comparison of estimates of daily food requirements for each sex- and age-class based on energetic life tables, FRI (sex symbols), and those based on the volumes of undigested prey in harbor seal stomachs collected on the east coast of Canada, FR2 (trend line; from Boulva and McLaren 1979). juveniles may not be as elevated or may converge on adult rates more rapidly than those of captive juve- niles. For example, the magnitude by which juvenile metabolic rates are elevated may be a function of their growth rates, which were generally much higher in captivity than in free-ranging harbor seals. Alterna- tively, the elevated metabolic rates of juveniles may have been obscured in Boulva and McLaren's (1979) analysis by the indiscriminate pooling of juveniles and adults, or the average meal sizes of juveniles underes- timated due to seasonal biases. For instance, most of Boulva and McLaredn's ( 1979) specimens were collected in summer and fall, which coincides with a post-wean- ing reduction in food intake in several other species of phocids (Worthy, 1987b). In the absence of any compelling basis for favouring either FRlslv or FR2S,X„ their mean, denoted as FRvxl, was adopted in subsequent analyses. Since juvenile FR1M and FR2s(x) differed by about ±20% of their mean, and juveniles accounted for 62% of the total population energy budget, this introduced a potential error of ±12% into the overall population energy budget. Estimated daily food requirements, FRXI>I, ranged from 1.5 kg for yearlings to 2.7 kg for full grown males, which repre- sented 6.3% and 3.1% of their total body masses respec- tively. The mean per capita daily food requirement, FR, was estimated to be 1.9 kg, or 4.3% of mean body mass. Olesiuk. Prey consumption of Phoca vitulina 505 Diet composition Of the 2,841 scat samples collected in the Strait of Georgia, 2,765 (97.3%) contained identifiable prey. Samples typically contained one to three (X=1.91) dif- ferent prey species, but occasionally contained as many as seven. Marine and anadromous fishes, which ac- counted for 96.0% of all prey identified, were by far the most prevalent prey category. The diet included at least 48 species from 20 different families (Olesiuk et al, 1990b). The diet was dominated by gadoids and clupeids which were present in 62.0 and 59.2%, re- spectively of all samples containing identifiable prey. Other important families that occurred in at least 1% of samples were, in decreasing order of importance, salmonids, batrachoids, embiotocids, cottids, pleuro- nectids, hexagrammids, scorpaenids, ammodytids and osmerids. The second most prevalent prey category was cepha- lopods, which occurred in 168 (6.1%) of all samples containing identifiable prey and represented 3.5% of all prey items identified. A superficial examination of their beaks indicated that the vast majority were squid (mainly Loligo opalacens with lesser amounts of Gonatus spp. ), but at least one octopus was also con- sumed. The remaining prey categories, namely crusta- ceans, other molluscs, echinoderms, and birds, occurred in <1% of all samples containing identifiable prey and accounted for <0.5% of the total number of prey items identified. In assessing the relative importance of each prey (Eqn. 16), it was assumed that scats represented all prey consumed within a 24-hour period. Since pinniped gastrointestinal passage rates typically range from 5 to 30 hours (Pastukhov, 1975; Helm and Morejohn, 1979; Prime, 1979; Bigg and Fawcett, 1985; Prime and Hammond, 1987; Harvey, 1989), it is likely that most of the prey present in a scat sample had been con- sumed within the same 24-hour period. Although cepha- lopod beaks may be retained in stomachs over longer periods (Bigg and Fawcett, 1985), cephalopods consti- tuted only a small part of the diet and beaks were usually accompanied by cephalopod eye lenses, which probably pass rapidly. It is believed that essentially all species consumed in a meal were represented in scat samples, mainly because such a wide array of structures had been used to identify prey. Captive studies of otolith recovery rates have indicated that the fragile otoliths of small fishes, such as herring, may be completely digested and hence under-represented in scats (Hawes, 1983; Jobling and Breiby, 1986; Jobling, 1987; da Silva and Neilson, 1985; Dellinger and Trillmich, 1988; Harvey, 1989). Further- more, based on a comparison between harbor seal stom- ach and intestinal contents, Pitcher (1980) concluded that the otoliths of larger prey such as salmon would also tend to be under-represented in scats because their heads (i.e., otoliths) are sometimes discarded prior to being consumed. However, these studies merely dem- onstrate the inadequacy of relying exclusively on oto- liths (see also Fig. 7), and are therefore not pertinent in the present study. The improved resolution achieved by utilizing a wide array of structures can be illustrated by examining the prevalence of the above prey species in selected collections in which they constituted the dominant prey. For example, herring elements were identified in 86.4- 100% of samples (X=96.1%; 150 of 156 samples) in 5 selected collections in which they were the dominant prey; and salmonids in 73.8-90.0% of samples (X=77.7%; 73 of 94 samples) in 5 selected collections in which they were the dominant prey. The slightly lower prevalence of salmonids was probably due to the fact they were not consumed by all seals as most of the samples without salmonids contained other prey spe- cies. Nevertheless, even if it were assumed that all seals had consumed these prey, the prevalence of her- ring in the diet would only have been underestimated by a factor of 1.04, and the prevalence of salmonids by a factor of 1.29. In contrast, in the same collections herring otoliths occurred in only 62.7%- of the samples containing herring and salmonid otoliths in only 9.6% of the samples containing salmonids. Thus, the preva- lence of herring would have been underestimated by a factor of 1.59 and salmonids by a factor of 10.42 had only otoliths been used to identify prey. The assertion that scat samples provided an accu- rate representation of diets is further substantiated O 50 < 10- 8 30- MID HEX SFP SPECIES Figure 7 Percentage of scat samples in which various fish prey were represented by otoliths. Prey codes: GAD=gadoids; HER= herring; SAL=salmonids; MID=plainfin midshipman; HEX=hexagrammids; SFP=surfperches; FLF=flatfish; SCP=sculpins; and ROK=rockfish. 506 Fishery Bulletin 91(3). 1993 NON- ESTUARIES by the similarity of the prey identified in scat samples and those identified in stomachs collected within and adjacent to the study area. All but one of the 19 prey types identified in 69 stomachs collected throughout British Columbia (Fisher, 1952; Spalding, 1964) were represented in the scats, and all but 2 of the 22 prey types identified in 81 stomachs collected in Washing- ton State (Scheffer and Sperry, 1931; Scheffer and Slipp, 1944). The exceptions were sablefish and rat- fish, each of which occurred in one stomach, and bur- rowing crayfishes, which occurred in three stomachs. Furthermore, the mean number of prey species identi- fied in the scat and stomach samples was also similar (X=1.91 in scats versus 2.06 in stomachs). Seasonal changes in diet composition outside of es- tuaries are shown in Figure 8A. Salmonids composed a relatively minor portion (0-5.2%; X=3.1% ) of the diet in all months. The diet was domi- nated by gadoids (X=45.7%; 94.5% of which were hake) and herring (X=33.0%). There was a pronounced seasonal shift be- tween these two prey, with ga- doids dominating during April- October (54.3-73.7%) and herring during December-March (58.0- 70.2% ). Other important prey, de- fined as those constituting >1% of the overall diet or >2% of the diet in any month, were hex- agrammids (X=3.4%; 95.8% of which were lingcod and 4.2% greenling), especially during December-April; plainfin mid- shipman (X=3.4%), especially during April-June and Novem- ber-December; surfperches (X= 2.2%; 81.3% of which were shiner perch and 18.7% pile perch); cephalopods (2.1%); and sand- lance (0.9%). Incidental prey in- cluded, in decreasing order, rock- fish, flatfishes, sculpins, smelts, skates, gunnels, lamprey, prickle- backs, crabs, sticklebacks, cling- fishes, eelpouts and sea urchins. Unidentified prey accounted for a mean of 2.09J of the total non- estuary diet. Seasonal changes in diet com- position within estuaries are shown in Figure 8B. Salmonids were consumed in all months (X=10.3$ i, but were most preva- lent during during September-January (14.5-21.2%). As was the case outside estuaries, the dominant prey in estuaries were gadoids (X=42.9%; 94.2% of which were hake) and, to a lesser extent, herring (X=27.3%). Gadoids dominate (39.2-53.9%) in all months except February-March, when herring were dominant (49.5- 50.6%). However, the seasonal shift in the importance of these prey was not nearly as pronounced as it was outside estuaries. Other important prey included plainfin midshipman (X=3.6%) especially during May- June (9.7-10.5% ); surfperches (X=3.6%; 91.9% of which were shiner perch), flatfishes (X=2.8%), sculpins (X=2.6%), all of which were most prevalent during Au- gust-September (5.4-6.1%, 5.0-6.3% and 7.8-7.9% re- spectively); and cephalopods (X=2.3%), especially dur- ing November-March (2.9-5.5%). Other incidental prey were, in decreasing order, rockfishes, sandlance, stick- LEGEND Salmonids Gadoids Pacific herring Sculpins Flatfishes □ Surfperches - Hexagrammids Plainfin midship: Sandlance Cephalopods Other I I Uni< Figure 8 Seasonal changes in diet composition 100 t) included surfperches (23 1), cephalopods (20 t), flatfish (123 1), sculpins (114t) and rockfishes (112 tj. Annual consumption of all other prey combined, none of which composed >1% of the overall diet, was estimated at 208 1, and total consump- tion of unidentified prey at 181 1. From the seal's perspective, approximately 44.2% (29.1 TJ) of the total annual energy requirements were obtained from herring and 33.3% (21.9 TJ) from hake. Thus, even though a larger biomass of hake than her- ring was consumed, hake were energetically less im- portant than herring owing to their lower energetic density. Salmonids accounted for 3.3 TJ, which repre- sented 5.1% of the total energy consumed compared to 4.0% of the biomass consumed. The precise energetic importance of other important prey could not be di- rectly ascertained because their energetic densities were not known. General discussion Although formal statistical analyses are not possible with the available data, the model provides some in- sight into the likely accuracy of the annual prey con- sumption estimates. One potential source of bias in the consumption estimates are errors in the estimated daily energy requirements. Since the gross energy re- quired for growth and reproduction constituted only a small portion of the overall population energy budget (1.7% and 4.5% respectively), uncertainties in these parameters, unless grossly underestimated, are un- likely to have an appreciable effect on the overall en- ergy budget. Moreover, a large body of evidence indi- cates that the basal metabolic rates of adult phocids conform with Kleiber's (1975) equation (Lavigne et al., 1986). Thus, the two major potential sources of bias are the extent to which juvenile metabolic rates are elevated relative to adults, and the costs associated with activity. With respect to the former, the estimates of juvenile maintenance requirements based on cap- tive seals and stomach contents of free-ranging seals differed by about ±20% of their mean. Because the study population was increasing at its intrinsic rate, it was markedly skewed toward juvenile age-classes which accounted for 62% of the total population en- ergy budget. Hence, the discrepancy between the cap- tive and stomach content estimates introduced a po- tential error of ±12% in the overall population energy budget. With respect to activity, the requirements of free-ranging seals probably fall somewhere between Olesiuk Prey consumption of Phoca vitulma 509 those of captive seals and seals that swim continu- ously for 60% of the time. Since the latter two esti- mates differed from their mean by ±13% for all sex- and age-classes, the overall uncertainty in the esti- mated mean daily per capita energy daily requirements was probably on the order of ±25% of the point esti- mate. Additional inaccuracies may be introduced in the conversion between units of energy and units of bio- mass due to seasonal, year-to-year, and age-related fluctuations in the energetic density of prey. For ex- ample, the energetic density of herring in the study area varies seasonally from 7.6 to 10.4 MJ-kg~' (Bigg and Olesiuk, unpubl. data), a range of ±20% of the mean. However, because the seasonal fluctuations in herring are probably more pronounced than other prey, and because median energetic values of prey were adopted in the analysis, the typical potential bias is more likely on the order of ±10%. Since the potential biases in the energetic densities of prey and energy requirements are independent and additive, the total annual prey consumption estimate can be considered accurate to within about ±35% of the point estimate. Further errors may be introduced in partitioning the total prey consumption among the different prey species. The main potential source of bias is the un- derlying assumption that all prey comprising a meal had been consumed in equal quantities. The lower and upper limits of the potential magnitude of this bias tended to be narrower for the two dominant prey (62- 135% and 61-169% of the point estimates for hake and herring, respectively ) than for other important prey, which averaged 35-211% of the point estimates. How- ever, since it is very unlikely that a particular prey species was always consumed in negligible quantities or always comprised the entire meal when it was con- sumed along with other prey, these extreme limits un- doubtedly overestimate the actual range of importance of prey. For example, when applied to the frequency of various fishes in 10,699 northern fur seal stomachs, the split-sample index actually gave results very simi- lar to volumetric analyses (r2=0.929 by region and r2=0.978 overall with slopes and intercepts not signifi- cantly different from one and zero respectively) (Olesiuk et al., 1990b). If it is assumed that realistic lower and upper limits were half the width of the extreme limits, the total potential error in the estimates of annual consumption of hake and herring would be on the or- der of 50-170% of the point estimates, and 45-220% of the point estimates for other important prey species (see Table 3). It should be noted, however, that be- cause the potential sources of bias in the consumption estimates are largely independent, the errors are just as likely to cancel as they are to compound. The sex and age-structure of the population, which varies with the status of populations, had a surpris- ingly small effect on the mean per capita prey require- ment. In the study population, which represented a population with a stable sex-and age-structure that was increasing at its intrinsic rate of 12.5% per annum, mean daily per capita food requirements were 1.9 kg, or 4.3% of mean body mass. If stationarity is induced by reducing fecundity rates, the parameter with the greatest influence on per capita requirements, the mean daily per capita requirement would increase to 2.1kg, but decline to 3.9% of mean body mass. This is be- cause the stationary population would be skewed more toward adults which not only have higher daily re- quirements, but also greater body masses. Thus, as also concluded by Lavigne et al. (1985) for harp seals in the northwest Atlantic, the population energy bud- get was relatively robust to direct changes in sex- and age-structure. In contrast, Hilby and Harwood (1985) found that energy requirements of grey seals were very sensitive to demographic changes. Their anomalous findings can be attributed to the fact that individual energetic requirements were scaled linearly to mass rather than mass075, and also because the metabolic rates of juveniles were only marginally elevated ( 13%) relative to those of adults of equivalent mass. It should also be noted that populations may also experience density-dependent effects not directly related to demo- graphic changes. For example, foraging costs may in- crease as prey become scarce, or seals may switch to alternate prey that would presumably have a lower energetic density or require greater energy expendi- ture to capture. These indirect effects were outside the scope of the basic model and could therefore not be evaluated. Contrary to the assumption that thermoregulatory costs were negligible, some investigators have found that the metabolic rates of captive seals increased when immersed in colder water. Brodie and Pasche (1982) reported that resting metabolic rates of fasting grey seal pups increased in cold water as they depleted their blubber reserves, and on this basis suggested that per capita food requirements would increase in a population as prey became less abundant and the con- dition of animals declined. Hart and Irving ( 1959) found that the critical lower temperature of harbor seals in water was 20° C in summer and 13° C in winter, well above the ambient surface sea temperatures in the study area. However, because these experiments were conducted under artificial conditions, it is doubtful that the results can be validly extrapolated to free-ranging seals. For example, Figure 1 in Hart and Irving (1959) indicates that at 0°C, the resting metabolic rates of harbor seals were 1.4-1. 8x greater than those within 510 Fishery Bulletin 91(3), 1993 the thermoneutral zone. However, the seals employed in their tests were restrained and probably post-ab- sorptive. Even if it were assumed that basal require- ments, which account for an average of 43% of the total energy requirements, were elevated to this ex- tent for the 60% of the time seals spent in the water, thermoregulatory costs would only be on the order of 18-36 W. However, for a free-ranging seal, 17% of the GE ingested would be liberated as the heat increment associated with feeding and, since musculature is only about 25% efficient (Luecke et al., 1975), the remain- ing 75% of the 23%' of GE expended for swimming would be liberated as heat. Since this "wasted" heat amounts to about 59 W for an average seal, it would appear that the thermoregulatory needs of free-rang- ing seals would be met indirectly through other en- ergy expenditures. One important assumption underlying the model was that daily ingestion rates were constant both with sea- son and between regions. In contrast to other phocids such as harp and ringed seals (McLaren, 1958; Ser- geant, 1973), Boulva and McLaren (1979) found no discernible seasonal pattern in the percentage of empty harbor seal stomachs, which implies that harbor seal feed throughout the year. As have previous models, the bioenergetics model assumed that feeding rates were constant in terms of biomass. Alternatively, it is possible that seals alter their foraging patterns in re- lation to the energetic density of prey. For example, greater quantities of poor-quality prey such as hake may be consumed when they are readily available com- pared to high-quality prey such as herring. Because any differences between the amount of energy ingested and required would be reflected by changes in energy reserves, the magnitude of potential biases introduced by these effects can be assessed from the seasonal changes in the condition of animals. Pitcher (1986) reported that the blubber composition of individual harbor seals ranged between extremes of 21% and 55% of body mass. Even if it assumed that the mean blub- ber content declined from 55% to 21% during a 6-month period of reduced food intake (or consumption of poorer quality prey) and increased from 21% to 55% during a second 6-month period of increased food intake (or con- sumption of higher quality prey), daily food require- ments during the first and second 6-month periods would be only 125% and 75% of the annual mean re- spectively. Thus, gradual seasonal changes in food in- gestion rates would not have a major effect on the prey consumption estimates. Boulva and McLaren (1979) noted that the condi- tion (i.e., girth:length ratio) of harbor seals of all ages and sexes combined was highest in winter and early spring, decreased during late spring, and was lowest in summer and late autumn. Similarly, Pitcher (1986) found that the blubber layer of adult males and fe- males harbor seals were thickest in winter, thinned during the summer, and that females had even thin- ner blubber layers by the autumn moult. These obser- vations are consistent, at least qualitatively, with my bioenergetics model which predicts that animals would accumulate blubber during the winter while feeding mainly on energy-rich herring and deplete blubber dur- ing the summer while feeding mainly on energy-poor hake. The model also predicts that nursing females would utilize an additional 16.8 kg of blubber, or 22% of their total body mass, during the late July to early September nursing period (Bigg, 1969), and would therefore be in poorer condition than males by the autumn moult. In an earlier assessment for harbor seals in the Bering Sea, Ashwell-Erickson and Eisner (1981) esti- mated the mean daily per capita gross energy require- ments to be 216-238 watts, which is 26-38% greater than my estimate of 172 watts. The difference appears to be almost entirely attributable to geographic differ- ences in body size. The mean body mass of harbor seals (both sexes were combined) in the Bering Sea was 67.7 kg, which is about 50% greater than the mean body mass of 44.7 kg in British Columbia. This differ- ence translates into about a 37% difference in meta- bolic mass (M075), to which energy expenditures were scaled in both our models. After my model had been completed, Harkonen and Heide-Jorgensen ( 1991) published a very similar model for harbor seals in the Skagerrak. There is good agree- ment between our models on how the energy budget is partitioned among various components. Harkonen and Heide-Jorgensen (1991) estimated that, for an increas- ing population, 73.1% of metabolizable energy (NE and the heat increment) is expended on maintenance, 19.0% on activity, 2.0% on growth, 4.5% on reproduction, and 1.5% on the annual moult. According to my model, 68.1%- is expended on maintenance (including the heat increment), 26.9% on activity, 1.4% on growth, 3.7% on reproduction, and I made no allowance for any costs associated with the annual moult. However, there is a considerable discrepancy between our estimates of the mean daily per capita gross energy requirements, even though the mean body mass of seals in the Skagerrak and British Columbia are very similar (41.8 and 44.7 kg respectively). Harkonen and Heide-Jorgensen's (1991) estimate of 227 watts is 32% greater than my estimate of 172 watts. This discrepancy is due primarily to dif- ferences in the way energetic requirements were scaled to body mass. Harkonen and Heide-Jorgensen (1991) first estimated the requirements of juveniles and then extrapolated these estimates to adults by assuming Olesiuk: Prey consumption of Phoca vitulma 51 1 that energy expenditures were linearly related to body mass (their correction of 1.4 to account for non-mass related differences between juveniles and adults was almost identical to mine). In contrast, energy expendi- tures in my model were scaled according to metabolic mass (M075), which increases less rapidly with age than body mass. According to my growth curves, the in- crease in body mass between weaning and adulthood would be 29% greater than the increase in metabolic mass over the same period for females, and 39% greater for males. Despite the aforementioned limitations, my bioener- getics model provides some insight into the foraging patterns of harbor seals in the Strait of Georgia. Hake and herring, the two most abundant fishes in the study area (Shaw et al., 1990; Hay et al., 1989), are clearly also the two most important prey of harbor seals. Com- bined, they account for an estimated 75% (at mini- mum 63%) of the total consumption both in terms of biomass and energy. One would therefore expect hake and herring stocks to play an important role in regu- lating the size of the harbor seal population, if in fact carrying capacity is food-limited. The annual consump- tion of these prey by harbor seal population in 1988, which is thought to be near or perhaps slightly above historic levels (Olesiuk4), represents approximately 3.2% of the total hake biomass in the study area (Shaw et al., 1990) and 3.5% of the total herring biomass (Haist et al. 1988; Hay10). Interestingly, both hake and herring appear to be only seasonally available in the Strait of Georgia, but in a reciprocal manner to one another. The Strait of Georgia of herring stock is largely migratory. Adult herring normally reside off the west coast of Vancouver Island, but emmigrate into the Strait of Georgia dur- ing December-March to spawn (Taylor, 1964; Hay et al., 1987; Hay et al, 1989). Although the Strait of Georgia hake stock is resident, hake are scarce through- out much of the Strait during December, and by March have congregated in deep ( 150-300 m) spawning ag- gregations in offshore waters. Following spawning, the spawning aggregations disperse and during April- November hake occur in shallower waters (50-100 m) throughout much of the Strait of Georgia (McFarlane and Beamish, 1985), which coincides with the period herring are unavailable. Thus, hake and herring pro- vide an abundant year-round source of food, as re- flected by the seasonal shift in the predominance of these prey in the diet (Fig. 8A). The year-round avail- ability of these two abundant prey may account for the much higher density of seals in the study area compared to other regions of British Columbia (Olesiuk etal., 1990a). Seasonal fluctuations in the overall importance of hake and herring were much less pronounced in estu- aries (Fig. 8B). In some estuaries the seasonal shift between these two prey were similar to those outside estuaries, whereas in other estuaries hake dominated the diet in all months (see Olesiuk et al., 1990b for details). Seals in the latter estuaries were probably feeding on small, localized stocks that are non-migra- tory and are known to reside nearby11, or on juvenile hake which move inshore and inhabit shallow waters (McFarlane and Beamish, 1985). Although hake and herring represented the major food items, a wide array of other species were con- sumed in small quantities. Predation on these prey appeared to be largely limited to the particular areas and periods each was most available or vulnerable. For example, adult salmon were consumed primarily as they concentrated en route to spawning rivers, and especially within estuaries. Numbers of seals in most estuaries also increased during September-December coinciding with the return of spawning salmon to ad- jacent rivers (Olesiuk et al, 1990b). Plainfin midship- man and lingcod were also preyed upon primarily dur- ing their spawning seasons when males defend nests, and trout in a few localities as they were released in large quantities from hatcheries or returning to spawn in natal rivers (Olesiuk et al., 1990b). Since these prey comprised a minor component of the overall diet, they probably play little role in regulating harbor seal abundance. This study indicates that the harbor seal is an op- portunistic predator in that it is capable of adjusting its foraging patterns to take advantage of seasonally and locally abundant or vulnerable prey. Although the low efficiency of the population may be construed as reinforcing the premise that seals are "inefficient con- verters of fish flesh" (Sergeant, 1973), the ecological efficiency of harbor seals is actually comparable to that of many terrestrial mammals (see review in Lavigne et al., 1982) and slightly above the theoretical upper limit of 2-3% expected for homeotherms (Turner, 1970). In order to acquire a more complete understanding of the role of harbor seals in the ecosystem, it will be necessary to extend the bioenergetics model to the community level (Lavigne et al., 1982). The population model raises several complex questions that can only be answered by community models. For example, to what extent is the sex- and age-structure and pro- ductivity of herring stocks affected by harbor seal '"D. Hay, Pacific Biological Station, Nanaimo, B.C., pers. commun. 1989. "G. McFarlane, Pacific Biological Station. Nanaimo, B. C, pers. commun. 1989. 512 Fishery Bulletin 91(3). 1993 predation; or to what degree are the overall predatory pressures on herring affected indirectly by seal preda- tion on other herring predators such as hake, salmon, and lingcod? Such a bioenergetics community model will need to incorporate not only data on the size of prey consumed, which can be obtained from a more detailed analysis of scat contents (e.g., Bailey and Ainley, 1982; Antonelis et al., 1983), but also the dy- namics and compensatory and depensatory mechanisms of population regulation of prey populations. Acknowledgments I would like to extend special gratitude to the late M. Bigg for his contributions to the investigations under- lying this synthesis. I am also grateful to L. Barrett- Lennard, G. Ellis, C.-E. Neville, T. Pawloski, H. Reisenleiter, and N. Thompson for their assistance in the field; to S. Crockford and R. Wigen for identifying prey remnants in scats; to S. Innes and D. Lavigne for helpful discussions on phocid energetics; to E. Warneboldt for drafting the illustrations; and to T. Beacham, M. Bigg, and anonymous reviewers for con- structive comments on earlier drafts of the manuscript. Literature cited American Society for Testing Materials (ASTM). 1982. ASTM Standard for metric practice. ASTM, Philadelphia, PA. E 380-82. 42 p. Antonelis, G. A., C. H. Fiscus, and R. L. Delong. 1983. Spring and summer prey of California sea lions, Zalophus californianus, at San Miguel Island, Cali- fornia, 1978-79. Fish. Bull. 82:67-76. Ashwell-Erickson, S. and R. Eisner. 1981. The energy cost of free existence for Bering Sea harbor and spotted seals. In D. W. Hood and J. A. Calder (eds.). The eastern Bering seas shelf: oceanog- raphy and resources Vol II, p. 869-898. Univ. Wash- ington Press. Bailey, K. M., and D. G. Ainley. 1982. The dynamics of California sea lion predation on Pacific hake. Fish. Res. 1:163-176. Beddington, J. R., R. J. H. Beverton, and D. M. Lavigne (eds.). 1985. Marine mammals and fisheries. George Allen & Unwin. London, 354 p. Beverton, R. J. H. 1982. Report of IUCN Workshop on marine mammal/ fishery interactions. La Jolla, California, 30 March- 2 April, 1981. Intl. Union Conservation of Nature and Natural Resources, Morges, Switzerland, 57 p. 1985. Analysis of marine mammal-fisheries inter- actions. In J. R. Beddington, R. J. H. Beverton, and D. M. Lavigne (eds.). Marine mammals and fisheries, p. 3-33. George Allen & Unwin, London, U.K., 354 p. Bigg, M. A. 1969. The harbor seal in British Columbia. Bull. Fish. Res. Board Canada. No. 172, 33 p. 1981. Harbor seal, Phoca vitulina Linnaeus, 1758 and Phoca largha Pallas, 1811. In S. H. Ridgway and R. J. Harrison (eds.),. Handbook of marine mammals, Vol. 2: Seals, p. 1-27. Academic Press, London, 359 p. Bigg, M. A., and I. Fawcett. 1985. Two biases in diet determination of northern fur seals (Callorhinus ursinus). In J. R. Beddington, R. J. H. Beverton, and D. M. Lavigne (eds.), Marine mam- mals and fisheries, p. 153-169. George Allen & Unwin, London, U.K., 354 p. Bigg, M. A, and P. F. Olesiuk. 1990. An elutriator for processing marine mammals scats. Marine Mammal Sci. 6(4):350-355. Bishop, R. H. 1967. Reproduction, age determination, and behavior of the harbor seal, Phoca vitulina L., in the Gulf of Alaska. MSc. thesis, Univ. Alaska, College, Alaska, 121 p. Boulva, J. 1973. The harbor seal, Phoca vitulina concolor, in east- ern Canada. Ph.D. diss., Dalhousie Univ., Halifax, Nova Scotia, Canada, 134 p. Boulva, J., and I. A. McLaren. 1979. Biology of the harbor seal, Phoca vitulina, in" eastern Canada. J. Fish. Res. Board Can., Bull. No. 200, 24 p. Brodie, P. F., and A. J. Pasche. 1982. Density-dependent condition and energetics of marine mammal populations in multispecies fisheries management. In M.C. Mercer (ed.), Multispecies ap- proaches to fisheries management advice, p 35- 38. Can. Spec. Publ. Fish. Aquat. Sci. 59. Brody, S. 1945. Bioenergetics and growth, with special reference to the efficiency complex in domestic animals. Reinhold, New York, 1023 p. Canadian Hydrographic Service 1990. Sailing directions: British Columbia Coast (south portion). Vol. 1, 15th ed. Dep. Fisheries and Oceans, Canada, 397 p. Clark, C. W. 1985. Economic aspects of marine mammal-fishery interactions. //; J. R. Beddington, R. J. H. Beverton, and D. M. Lavigne (eds. I. Marine mammals and fisheries, p. 34-38. George Allen & Unwin, London, U.K., 354 p. Da Silva, J., and J. D. Neilson. 1985. Limitations of using otoliths recovered in scats to estimate prey consumption in seals. Can. J. Fish. Aquat. Sci. 42:1439-42. Davis, R. W., T. M. Williams, and G. L. Kooyman. 1985. Swimming metabolism of yearling and adult har- bor seals Phoca vitulina. Physiol. Zool. 58:590-596. Olesiuk Prey consumption of Phoca vitulma 513 Dellinger, T., and F. Trillmich. 1988. Estimating diet composition from scat analysis of otariid seals (Otariidae): is it reliable? Can. J. Zool. 66:1865-1870. Demaster, D. P., D. J. Miller, D. Goodman, R. L Delong, and B. S. Stewart. 1982. Assessment of California sea lion fishery interactions. Trans. 47th N. Amer. Wildl. and Natu- ral Resources Conf., 1982, p. 253-264. Fisher, H. D. 1952. The status of the harbor seal in British Colum- bia, with particular reference to the Skeena River. Fish. Res. Board Can., Bull. No. 93, 58 p. Gallivan, G. J. 1977. Temperature regulation and respiration in the freely diving harp seal (Phoca groenlandica). M.Sc. thesis, Univ. Guelph, Guelph, Ontario, Canada, 60 p. Gallivan, G. J., and K. Ronald. 1981. Apparent specific dynamic action in the harp seal (Phoca groendlandica I. Comp. Biochem. Physiol. Vol. 69A:579-581. Haist, V., J. F. Schweigert, and D. Fournier. 1988. Stock assessments for British Columbia herring in 1987 and forecasts of the potential catch in 1988. Can. MS Rep. Fish. Aquat. Sci. 1990, 63 p. Harkonen, T., and M.-P. Heide-Jorgensen. 1991. The harbor seal Phoca vitulina as a predator in the Skagerrak. Ophelia 34:191-207. Hart, J. S., and L. Irving. 1959. The energetics of harbor seals in air and in wa- ter with special consideration of seasonal changes. Can. J. Zool. 37:447-457. Harvey, J. T. 1989. Assessment of errors associated with harbor seal (Phoca vitulina) faecal sampling. J. Zool. Lond. 219:101-111. Harwood, J., and J. P. Croxall. 1988. The assessment of competition between seals and commercial fisheries in the North Sea and Antarctic. Marine Mammal Sci. 4:13-33. Hawes, S. D. 1983. An evaluation of California sea lion scat samples as indicators of prey importance, M.A. thesis, San Francisco State Univ., San Francisco, CA, 50 p. Hay, D. E., D. N. Outram, B. A. McKeown, and M. Hurlbert. 1987. Ovarian development and oocyte diameter as maturation criteria in herring (Clupea harengus pallasi). Can. J. Fish. Aquat. Sci. 44:1496-1502. Hay, D. E., M. C. Healey, L. J. Richards, and J. B. Marliave. 1989. Distribution, abundance, and habitat of prey fishes in the Strait of Georgia. In K. Vermeer and R.W Butler (eds.), The ecology and status of marine and shoreline birds in the Strait of Georgia, British Columbia, p. 37-49. Spec. Publ. Can. Wildl. Serv. Ottawa. 186 p. Helm, R. C, and G. V. Morejohn. 1979. Initial defecation time and intestinal length of three species of pinnipeds: Phoca vitulina, Zalophus californianus and Mirounga angustirostris. Abstract. Third Biennial Conf. Biol. Marine Mam- mals, Seattle, Washington. Hilby, A. R., and J. Harwood. 1985. The effects of variation in population param- eters on the energy requirements of a hypothetical grey seal population. In J. R. Beddington, R. J. H. Beverton, and D. M. Lavigne (eds.), Marine mam- mals and fisheries, p. 337-343. George Allen & Unwin, London, U.K., 354 p. Hourston, A. S. 1956. Investigations of the juvenile herring stocks in the Strait of Georgia region, 1955. Pac. Biol. Stn. Ms. Rep. BQ114, Nanaimo, B.C., 59 p. Innes, S. 1984. Swimming energetics, metabolic rates and hind limb muscle anatomy of some phocid seals. Ph.D. diss., Univ. Guelph, Guelph, Ontario, 186 p. Innes, S., D. M. Lavigne, W. M. Earle, and K. M. Kovacs. 1987. Feeding rates of seals and whales. J. Anim. Ecol. 56:115-130. Jobling, M. 1987. Marine mammal faeces as indicators of prey im- portance— A source of error in bioenergetics stud- ies. Sarsia 72:255-260. Jobling, M., and A. Breiby. 1986. The use and abuse of fish otoliths in studies of feeding habits of marine piscivores. Sarsia 71:265- 274. Ketchen, K. S., N. Bourne, and T. H. Butler. 1983. History and present status of fisheries for ma- rine fishes and invertebrates in the Strait of Georgia, British Columbia. Can. J. Fish. Aquat. Sci. 40:1095- 1119. Reiver, K. M., K. Ronald, and F. W. H. Beamish. 1984. Metabolizable energy requirements for mainte- nance and faecal and urinary losses of juvenile harp seals (Phoca groendlandica ). Can. J. Zool. 62:769-776. Kleiber, M. 1975. The fire of life: An introduction to animal ener- getics. Robert E. Krieger, Huntington, NY, 454 p. Lavigne, D. M. 1982. Similarity in energy budgets of animal populations. J. Anim. Ecol. 51:195-206. Lavigne, D. M., W. Barchard, S. Innes, and N. A. Oritsland. 1982. Pinniped bioenergetics. In Mammals in the seas. Vol. IV. Small cetaceans, seals, sirenians and sea otters, p. 191-235. FAO Fisheries, Ser. 5, Rome. Lavigne, D. M., S. Innes, R. E. A. Stewart, and G. A. J. Worthy. 1985. An annual energy budget for North-west Atlan- tic harp seals. In J. R. Beddington, R. J. H. Beverton and D. M. Lavigne (eds.). Marine mammals and fisheries, p. 319-336. George Allen & Unwin, Lon- don, U.K., 354 p. Lavigne, D. M., S. Innes, G. A. J. Worthy, K. M. Kovacs, O. J. Schmitz, and J. P. Hickie. 1986. Metabolic rates of seals and whales. Can. J. Zool. 64:279-284. 514 Fishery Bulletin 91(3). 1993 Luecke, R. H., V. Natarajan, and F. E. South. 1975. A mathematical biothermal model of the Cali- fornia sea lion. J. Thermal Biology 1:35—45. Malouf, A. 1986. Seals and sealing in Canada. Report of the Royal Commission on Seals and Sealing Industry in Canada. Vol.3. (Government of Canada Report), 679 p. Markussen, N. H., M. Ryg, and N. A. Ortisland. 1990. Energy requirements for maintenance and growth of captive harbor seals, Phoca vitulina. Can. J. Zool. 68:423-426. Mate, B. R., and J. T. Harvey. 1987. Acoustical deterrents in marine mammals con- flicts with fisheries. Workshop, 17-18 February, 1986, Newport Oregon. Oregon State University, Publ. No. ORESU-W-86-001, 116 p. Matsuura, D. T., and G. C. Whittow. 1973. Oxygen uptake of the California sea lion and harbor seal during exposure to heat. Am. J. Physiol. 225:711-715. McConnell, B. J., J. H. Prime, A. R. Hilby, and J. Harwood. 1984. Grey seal diet. In Interactions between grey seals and UK fisheries, p. 148-183 . Natural Environ. Research Council, Sea Mammal Research Unit, Cam- bridge, 241 p. McFarlane, G. A., and R. J. Beamish. 1985. Biology and fishery of Pacific whiting, Merluccius productus, in the Strait of Georgia. Mar. Fish. Rev. 47:23-34. McLaren, I. A. 1958. Biology of the ringed seal (Phoca hispida Schreber) in the eastern Canadian arctic. Bull. Fish. Res. Board Can. 118, 97 p. Melteff, B. R., and D. H. Rosenberg. 1984. Proceedings of the workshop on biological inter- actions among marine mammals and commercial fisheries in the southeastern Bering Sea. October 18-21, 1983, Anchorage, Alaska, 300 p. Oftedal, O. T., and S. J. Iverson. 1987. Hydrogen isotope methodology for measurement of milk intake and energetics of growth in suckling young. In A. C. Huntley, D. P. Costa, G. A. J. Wor- thy, and M. A. Catellini (eds.i, Approaches to marine mammal energetics, p. 67-96. Soc. Mar. Mammal. Spec. Publ. 1. Olesiuk, P. F., M. A. Bigg, and G. M. Ellis. 1990a. Recent trends in the abundance of harbor seals, Phoca vitulina, in British Columbia. Can. J. Fish. Aquat. Sci. 47:992-1003. Olesiuk, P. F, M. A. Bigg, G. M. Ellis, S. J. Crockford, and R. J. Wigen. 1990b. An assessment of the feeding habits of harbor seals (Phoca vitulina) in the Strait of Georgia, British Columbia, based on scat analysis. Can. Tech. Rep. Fish. Aquat. Sci. 1730. 135 p. Oritsland, N. A., and K. Ronald. 1975. Energetics of the free diving harp seal (Pago- philus groenlandicus). Rapp. P.-v. Reun. Cons. Int. Explor. Mer, 1969:451-454. Parsons, J. L. 1977. Metabolic studies on ringed seals (Phoca his- pida}. M.Sc. thesis, Univ. Guelph, Guelph, Ontario, Canada, 82 p. Pastukov, V. D. 1975. Some observations on the Baikal seal under ex- perimental conditions. Transl. Ser. No. 3544, Can. Fish, and Mar. Serv., 6 p. Perez, M. A., and M. A. Bigg. 1986. Diet of northern fur seals, Callorhinus ursinus, off western North America. Fish. Bull. 84:957-97 1 . Pitcher, K. W. 1980. Stomach contents and faeces as indicators of har- bor seal, Phoca vitulina, foods in the Gulf of Alaska. Fish. Bull. 78:797-798. 1986. Variation in blubber thickness of harbor seals in southern Alaska. J. Wildl. Manage. 50:463-466. Pitcher, K. W., and D. C. McAllister. 1981. Movements and haulout behaviour of radio- tagged harbor seals, Phoca vitulina. Can. Field Nat. 95:292-97. Prime, J. H. 1979. Observations on the digestion of some gadoid fish otoliths by a young common seal. Int. Counc. Explor. Sea CM 1979/N No. 14, 6 p. Prime, J. H., and P. S. Hammond. 1987. Quantitative assessment of grey seal diet from faecal analysis. In A. C. Huntley, D. P. Costa, G. A. J. Worthy, and M. A. Catellini (eds.), Approaches to marine mammal energetics, p. 165-181. Soc. Mar. Mammal. Spec. Publ. 1. Scheffer, T. H., and C. C. Sperry. 1931. Food habits of the Pacific harbor seal, Phoca richardii. J. Mammal. 12:214-226. Scheffer, V. B., and J. W. Slipp. 1944. The harbor seal in Washington State. Am. Midi. Nat. 32:373-416. Sergeant, D. E. 1973. Feeding, growth and productivity of northwest Atlantic harp seals (Pagoplulus groenlandicus). J. Fish. Res. Board Can. 30:17-29. Shardlow, T. F., and L. D. Collicutt. 1989. Strait of Georgia sport fishery creel survey sta- tistics for salmon and groundfish, 1987. Can. MS Rep. Fish. Aquat. Sci. 2035, 62 p. Shaw, W., G. A. McFarlane, and R. Reiser. 1990. Distribution and abundance of Pacific hake (Merluccius productus) spawning stocks in the Strait of Georgia, British Columbia, based on trawl and acoustic surveys in 1981 and 1988. Symp. Application Stock Assessment Techniques to Gadids. Intl. North Pacific Fish Coram. Bull. 50:121- 134. Spalding, D. J. 1964. Comparative feeding habits of the fur seal, sea lion and harbor seal on the British Columbia coast. Bull. Fish. Res. Board Can. No. 146, 52 p. Sullivan, R. M. 1979. Behavior and ecology of harbor seals, Phoca vitulina, along the open coast of California. M.Sc. thesis, Humboldt State Univ., Areata, CA, 115 p. Olesiuk; Prey consumption of Phoca vituhna 515 Taylor, F. H. C. 1964. Life history and present status of British Co- lumbia herring stocks. Fish. Res. Board. Can. Bull. 143, 81 p. Turner, F. B. 1970. The ecological efficiency of consumer populations. Ecology 51:741-42. Williams, T. M. 1987. Approaches for the study of exercise physiology and hydrodynamics in marine mammals. In A. C. Huntley, D.R Costa, G.A.J. Worthy and M.A. Catellini (eds.l. Approaches to marine mammal energetics, p. 127-146. Soc. Mar. Mammal. Spec. Publ. 1. Worthy, G. A. J. 1987a. Metabolism and growth of young harp and grey seals. Can. J. Zool. 65:1377-1382. 1987b. Ecological energetics of harp and gray seals: Im- plications from a simulation model. In A. C. Huntley, D. P. Costa, G. A. J. Worthy, and M. A. Catellini (eds.l, Approaches to marine mammal energetics, p. 227- 246. Soc. Mar. Mammal. Spec. Publ. 1. Worthy, G. A. J., and D. M. Lavigne. 1983. Changes in energy stores during postnatal de- velopment of the harp seal, Phoca groenlandica. J. Mammal. 64:89-96. Yoehem, P. K., B. S. Stewart, R. L. Delong, and D. P. Demaster. 1987. Diel haulout patterns and site fidelity of harbor seals (Phoca vitulina richardsi) on San Miguel Is- land, California, in autumn. Marine Mammal Sci. 3:323-332. Zullinger, E. M., R. E. Ricklefs, K. H. Redford, and G. M. Mace. 1984. Fitting sigmoidal equations to mammalian growth curves. J. Mammal. 65:607-636. AbStfclCt.— Habitat association, growth, and burrowing behavior were examined for a population of young-of-the-year juvenile queen conch, Strombus gigas L., in the southern Exuma Cays, Bahamas, during the winter 1988-1989. These early juveniles (35-54 mm siphonal shell length) primarily inhabit shal- low unvegetated zones where they burrow in the sediment during the day and surface at night. Labora- tory experiments showed that the burrowing rhthym was endogenous. Highest densities of 1- and 2-year- old juveniles (80-140mm) conch were observed in adjacent, deeper seagrass beds, suggesting that queen conch make an ontogenetic shift in habitat. Results of an enclosure ex- periment revealed that growth rates of early juvenile conch were higher in seagrass (0.11 mm/day) and in rubble (0.09 mm/day) than in bare sand (0.01 mm/day I where they were initially found. These observations suggest that emergence of juvenile conch and movement to vegetated habitats at 35-54 mm shell length is associated with changes in nutri- tional requirements. Changing habi- tat association with age may also be related to predator avoidance and burrowing capabilities. Habitat re- quirements of early juvenile conch are different from those of 1-year- old juveniles and stock enhancement programs will need to consider on- togenetic habitat shifts. Ontogenetic shift in habitat by early juvenile queen conch, Strombus gigas: patterns and potential mechanisms Veronique J. Sandt Allan W. Stoner* Caribbean Marine Research Center 805 E 46th Place Vero Beach. Florida 32963 Manuscript accepted 6 April 1993 Fishery Bulletin 91:516-525 (1993) Declining populations of queen conch (Strombus gigas L. ), usually attrib- uted to overfishing, have been re- ported for numerous areas in the Caribbean region (Brownell and Stevely, 1981; Appeldoorn et al., 1987; Berg and Olsen, 1989). Hatchery pro- duction and field release of juvenile conch have been suggested as a means of restoring depleted stocks (Berg, 1976; Brownell, 1977; Davis and Hesse, 1983; Iversen et al, 1986) but release methods have not been perfected (Stoner, unpubl. data). Al- though survival in small juvenile conch is low in natural populations (Appeldoorn, 1984), mass rearing of large juveniles (1- and 2-years-old) is costly (Siddall, 1983) and releases will need to be made with 0+ year class conch. Unfortunately, pilot re- leases of small conch have resulted in very low survival (Appeldoorn and Ballantine, 1983; Siddall, 1983; Appeldoorn, 1985), probably because of the lack of ecological information on early stage juveniles and problems associated with the identification of suitable habitats (Iversen et al., 1986; Stoner and Sandt, 1991; Stoner, unpubl. data). Despite high densities (1— 2/m2) of 1- and 2-year old conch (80-140 mm shell length) in certain seagrass meadows (Alcolado, 1976; Weil and Laughlin, 1984; Stoner and Waite, 1990; Wicklund et al, 1991; Stoner et al., 1993), individuals less than 50-60 mm in shell length (called early juveniles in this study) have rarely been seen in the field. Iversen et al. (1986) suggested that these early juvenile conch probably spend a large part of their time buried in the sediment. Attempts have been made to find them using suction dredges (Iversen et al., 1987; Stoner, unpubl. data) but few early juveniles have been collected and, to date, no quantitative data exist. An opportunity to gather informa- tion on the distributional ecology and behavior of early juvenile queen conch was provided by the fortuitous discovery of a 0+ year class popula- tion near Neighbor Cay, in the cen- tral Bahamas, in January 1989. The subsequent investigation was designed: 1) to quantify habitat- specific distribution and abundance patterns, 2) to examine habitat- specific growth rates, and 3) to ex- amine buiTowing behavior in relation to time of day and light cues. The findings are discussed in terms of stock rehabilitation and management. Study site The distribution and behavior of early juvenile queen conch was in- vestigated near the north shore of *To whom correspondence should be ad- dressed. 516 Sandt and Stoner: Ecology of earlyjuvenile Strombus gigas 517 Neighbor Cay in the southern Exuma Cays, Bahamas (28°48.84'N, 76°11.9'W) (Fig. 1). The study site was located between a section of shoreline beachrock and subtidal carbonate rubble to the east and a sand bar to the west which extended outward from the shore- line (Fig. 2). Tidal currents ran parallel to shore and increased from near zero at the shoreline to 50 cm/s or greater 200 m from shore. The tidal range was 1.0 m. Distribution and abundance of conch was examined along a 110-m section of nearshore shallow habitat (0- 1.9 m deep at mean low water). Several zones extend- ing to 25 m offshore from the low water mark were identified (Fig. 2). Zone A was intertidal, between high and low water marks. Zone B consisted of the shallow- est subtidal area extending from shore to the top of a steep surf berm (zone C). Zone D was a sparsely veg- etated transition area between either zone C (stations 1-7) or zone B (stations 8-12), and zone E. Zones A through C were bare sand habitats. Zones E, F, and G were areas vegetated with turtlegrass, Thalassia testudinum Konig and small amounts of shoal grass, Halodule wrightii Konig (Table 1). Measurements and experiments in zone E were made 2.5 m from the in- shore edge of the seagrass (Fig. 2). Measurements in Zones F and G were made 10 and 20 m, respectively, from the seagrass edge. The rocky area to the east was a carbonate cobble covered with macroalgae including an abundant turf of the green algae Cladophoropsis ~*5 ', Study Site 7^ NEIGHBOR * k ^ CAY ^\Vl EXUMA SOUND LEE -. STOCKING ISLAND 82" _l ^•?6°**^| Figure 1 Map of the Lee Stocking Island area in the centra] Bahamas, showing the location of the Neighbor Cay study site where earlyjuvenile queen conch were found. membranacea and Batophoro oerstedi, along with clumps of the branched red algae Laurencia obtusa, L. poitei, and Graciolaria compressa. The adjacent seagrass area (zones E, F, G) is a known, long-term (>5 years) nursery habitat for queen conch (>l-yr-old) studied earlier bv Stoner and Sandt ( 1991). Methods and material Environmental measurements Sediment grain-size and organic content, two factors frequently important in the distribution of infaunal ani- mals, were measured in each zone (A to G) and along each transect perpendicular to the beach where pos- sible (Fig. 2). Sediments to 5.0cm depth were collected in a 40-mm diameter core and frozen (-15°C) until analy- sis. In the laboratory, one thawed sediment subsample of approximately 100 g wet weight was dried for at least 24 hours in an oven at 80°C to constant weight and incinerated at 550°C in a muffle furnace for 4 hours. Organic content was estimated as the percent differ- ence between dry weight and ash-free dry weight. A second sediment subsample of approximately 50 g was used to measure granulometric properties. Grain-size for the sand fraction was determined by using standard sieve procedures (Folk, 1966), after removing salts and extracting the silt-clay fraction by washing with fresh- water on a 62-um mesh screen. Silt- clay fractions were analysed with a standard pipet procedure (Galehouse, 1971). Product moment statistics were generated for mean grain-size (McBride, 1971). Relative cohesiveness of the sedi- ment was measured in the field at every non-rocky station by using a 50-cm-long steel rod (15.3 mm di- ameter) equipped with a sliding lead weight (1.3 kg) and a stop 17 cm above the end of the device. The rod was placed vertically on the surface of the sediment, the weight was raised 25 cm on the steel shaft and dropped. Penetration was mea- sured in mm, and the average of two measurements was recorded. Seagrasses, macroscopic detritus (identified as being mostly senes- cent seagrass blades), and macro- algae were collected on transects 3, 5, 7, and 9 in the 3 vegetated zones E, F, and G. Samples were collected from a 25 X 25 cm quadrat into 3-mm mesh bags. Detritus and Fishery Bulletin 91(3), 1993 t GREAT BAHAMA BANK v \ /////////////^NN \ ■■■:■■■■:■■:.",::& » \ \ ■■■:.■::.:: '/ i > \ Sand Bar Sur( Berm IZone CI &\.\ . F , • ■__ » , c . V.Wj(///iC//M/////WWUW/ttU/ji/iZ. ID — LOW WATER B '•'"•'' ... ..-/<'•------ q . '.- NEIGHBOR CAY Figure 2 Detailed map of the Neighbor Cay, Bahamas, study site showing the habitat zones where environ- mental measurements and conch density were made. Zones A, B, C, and D are bare sand. Zones E, F, and G are vegetated with turtlegrass (Thalassia testudinum). Numbers at the mean low water line (MLW) indicate positions of transects which ran perpendicular to the beach. above-ground parts of the macrophytes were separated by species and dried at 80°C for 24 hours. The indi- vidual components were weighed (O.lg) and biomass was expressed in g dry weight/m2. Water depth was measured at mean low water at every station in each zone. Conch distribution Conch density (no./m2) was determined in January 1989 at each station by searching two 0.75-m radius circles (to 10cm depth) to extract buried conch. Preliminary observations at the Neighbor Cay study site showed that burial depth was limited to 3-4 cm for 35-54 mm conch. All conch were counted and measured for shell length (SL, spire to siphonal groove). In early February 1989, heavy mortality due to pre- dation reduced the early juvenile population by at least 50%. Crushed shells of small conch were abundant in all of the primary habitats, and crabs were the sus- pected predators. Few free-ranging, tagged conch were found alive after that time. Table 1 Habitat characteristics in seven zones (A -G) surveyed fo r early juvenile conch at Neighbor Cay, Bahamas (see Fig 2).' Values are mean ±SD and number of measurements for each zone are in parentheses. A B c D E F G (12) (10) (8) (9) (10) (7) (7) Depth -0.2 + 0.1 0.3 ± 0.1 0.7 ± 0.2 1.3 ± 0.4 1.4 ± 0.5 1.6 ± 0.1 1.9 ± 0.1 Sediments Grain-size 984 ± 116 974 ± 187 796 ± 250 705 ± 204 521 ± 90 490 ± 64 496 ± 64 Organics 2.59 ± 0.36 2.60 ± 0.28 2.62 ± 0.32 2.74 ± 0.22 2.73 ± 0.5 2.68 ± 0.26 2.70 ± 0.22 Relative cohesiveness 48+7 48 ± 13 127 ± 35 47 ± 9 41 + 7 40 ± 3 43 ± 6 Macrophyte biomass Thalassia testudinum 0 0 0 0 11.4 ± 5.6 8.8 ± 3.7 12.5 ± 4.0 Halodule wrightii 0 0 0 0 0.1 ± 0.2 0.1 ± 0.1 0.2 ± 0.3 Macrodetritus 0 0 0 0 4.8 ± 0.8 1.1 ± 0.6 1.3 ± 1.4 'Units for each characteristic are: depth m at mean low water: MLW) (negative depth is height above MLWi, sediment grain size (urn), sediment organics (9f of dry weight), relative cohesiveness (mm) , Thalassia. Halodule. and macrodetritus biomass (g dry weigh/nrl. Sandt and Stoner Ecology of early juvenile Strombus gigas 519 Enclosure experiment Habitat-specific growth rates were examined experi- mentally in covered enclosures. Circular cages, 1.5 m in diameter (1.77 m-) were constructed with polyethyl- ene mesh (1 x 1cm) wired to a reinforcement bar driven into the sediment. Two cages were built in zone B, near transect 6 in bare sand, two were placed in the seagrass zone E, and two were built in the rocky zone between transect 11 and 12 (Fig. 2). The mesh was pushed into the sediment approximately 5.0 cm to pre- vent escape of conch. Tops of the cages were covered with the same plastic mesh to exclude predators. Conch between 37 and 49 mm SL were collected in zone B. After removing all visible invertebrates from the 6 cages, 8 individually tagged and measured conch were randomly assigned to each cage. This yielded a density of 4.5 conch/nr, near the highest natural den- sities of similar size class conch in the field. The experiment was initiated on 3 February 1989, examined after 19 and 41 days for cage damage and conch loss, and terminated after 63 days. Growth rates were expressed in mm shell length per day. At the end of the enclosure experiment, conch were collected and frozen for measurement of body condi- tion factor and stomach analysis. After thawing, the animals were drawn from their shells and rinsed to remove feces and mucus. All undamaged conch were blotted lightly and weighed (0.1 g). Condition factor was expressed as the ratio of soft tissue weight:shell length (g/mm). No significant sexual dimorphism oc- curs in queen conch until the gonads begin to develop at approximately 2 years of age; therefore, sex of the experimental animals was not considered. Stomach contents, removed from the conch and pre- served in a 70% solution of ethanol mixed with dilute rose bengal, were quantified by the gravimetric sieve- fractionation method of Carr and Adams (1972). Pooled contents from the stomachs of conch from each experi- mental treatment were washed through a series of four sieves of decreasing mesh size (425, 250, 150, and 75 urn, and each sieve fraction was examined under a dissecting microscope (20-40 x). Food items were heavily macerated but easily classified into general taxonomic categories; these included algae, detritus, and small invertebrates such as foraminiferans, gas- tropods, and polychaetes. The proportion of each food type (as well as sand) in each sieve fraction was esti- mated by identifying and counting the individual par- ticles. After examination and identification, each sieve fraction was dried for 12 hours at 80°C. Dry weights of the fractions were summed and the relative impor- tance of the different food categories were reported as the percentage of total dry weight. Analysis of variance was used to test the signifi- cance of differences in growth and condition factor among the habitat treatments. Because multiple mea- surements from individual enclosures are pseudo- replicates (Hurlbert, 1984), mean values from the rep- licate cages were used in the analyses. Growth rates were log10-transformed to produce homogeneity in vari- ance (Box test, F = 0.511, P = 0.608). Transformation was unnecessary for condition data. Tukey's test was used for multiple comparisons. Burrowing in the field Early juvenile conch were first observed at Neighbor Cay late in the day (17:00 hours) on 8 January 1989, but none were visible the following morning. Assum- ing daytime burial, a survey of the population was initiated to examine diel periodicity in burrowing be- havior. On 17 January, 64 early juveniles were tagged with vinyl orange spaghetti tags (Floy Tag & Manu- facturing, Inc. ) tied to the shell spire and released in zone B of the study site. A 4-cm free end was left on the tag so that buried conch could be easily seen and counted. During the survey, all early juvenile conch were tagged when observed (500 total). Surveys for buried conch were made on 22 dates between 17 January and 28 February 1989. Between 17 January and 7 February, surveys were made by two divers swimming parallel to the beach. All tagged conch were counted and their behavior was recorded as either on the surface, or buried. Because massive mortality due to predation on early juveniles occurred during the first week of February, subsequent obser- vations were made on the tagged conch held in experi- mental enclosures (see previous section). Over the course of the study the beach was surveyed at nearly all times of day and night and at different stages of the tide. Twice in January, surveys were made each half-hour during each of the transitions from dark to light and from light to dark. This permitted observa- tions on the precise time of emergence and burrowing relative to times of twilight, sunrise, and sunset. Tim- ing on 53 surveys was such that observations were made at least once or twice every half hour through day and night. Burrowing in the laboratory The role of light stimulus on burrowing behavior of early juvenile conch was examined by subjecting them to different light-dark cycles in laboratory aquaria. Be- cause most of the early juveniles found in the field had been manipulated during previous experiments, we 520 Fishery Bulletin 91(3). 1993 used hatchery-reared conch (39-53 mm SL) as subjects for this experiment. Nine 190-L aquaria containing 5 cm of coarse sand from Neighbor Cay (zone B) were used in the experi- ments. Water temperature was maintained at approxi- mately 28°C, and aeration was provided by under- gravel filters. Ten juvenile conch were placed in each of the aquaria and subjected to a schedule of 12 hours light: 12 hours darkness. After 4 days, when the bur- rowing cycles of the animals became apparent, the ani- mals were subjected to three different light regimes during the subsequent 4 days: three aquaria were kept under natural cycle (12:12), three were placed under continuous darkness and three under continuous light. Observations on burrowing behavior were made twice during the day and twice during the night (0400, 1100, 1500, 0000 hours). Food, in the form of seagrass detri- tus, was placed on the surface of the sediment. subtidal, in the surf berm, and at the bottom of the slope, respectively). All others were found in the veg- etated zone E (Fig. 3). No early juvenile conch were found in the most offshore zones F and G. Mean den- sity of early juveniles (1.25 conch/m2) was greatest in zone C, in association with the coarser grain sizes and low sediment cohesiveness (Table 1). In contrast to the narrow, high beach distribution of early juveniles, later stage juveniles were found in every zone, and highest densities in the seagrass zones E, F, and G (Fig. 3). Most of the early juvenile conch were distributed between transects 3 and 9 (Fig. 2) over an area of approximately 250 m2. Based on a mean density of 1.25 conch/m2 in zone C and 4 conch/m2 observed in some areas of zone C, the early juvenile population was be- tween 275 and 1000 conch. The densities observed prob- ably accounted for most of the total population since 500 conch were tagged during night observations. Results Habitat characteristics Transects from the intertidal beach to the shallow subtidal seagrass bed were characterized by increas- ing water depth ranging from -0.2m (at MLW) in the center of intertidal zone A to 1.9 m in zone G. Sediment grain-size decreased with increasing depth (Table 1). In zones A through E sediments were coarse sands (984-521 urn), while medium sands (496-490 urn) were found in zones F and G. Organic content of the sediments varied little across the different zones, rang- ing from 2.59 to 2.74% of dry weight (Table 1). Sedi- ment cohesiveness was relatively constant across the offshore zones (40-48 mm), except in the surf berm (zone C), where cohesiveness was low (i.e., penetrabil- ity and porosity were high) (Table 1). Seagrasses, primarily Thalassia testudinum, were present in relatively low biomass in zones E, F, and G (Table 1). Dry weights of macrodetritus were also rela- tively low, with the highest mean value (4.8 g dry wt/ m2) observed in zone E. Conch distribution In the January survey, all conch found were less than 60 mm or greater than 80 mm SL. Given the summer spawning season of queen conch in the Exuma Cays (Stoner et al., 1992) and estimated growth, it was as- sumed that the former represented the young-of-the- year class (0+), while those larger than 80 mm were 1- and 2-year old conch. Eighty-six percent of the early juvenile conch (35-54 mm SL) were found in unvegetated zones B, C, and D (i.e., in the immediate Enclosure experiment No mortality occurred during the first two periods of the transplant experiment; therefore, no replacements were necessary and growth rates were calculated for the original conch (Table 2). At the end of the experi- ment, two conch were found dead, one on sand and one on seagrass. Five conch were unaccounted for, two in the sand treatment, two in the rocky zone, and orte in the seagrass habitat. Growth rate in the sand habitat was much lower (0.012 mm/day) than the rates in seagrass (0.112mm/ day) and rock habitats (0.094 mm/day) (Table 2). The differences were significant (ANOVA, F = 769.35, C 1 :■ tm 0+ CLASS CD 1+. 2+ CLASS o 1-5- z (a 1.0- z LlJ Q 5 0.5- z o 0.0- J. A li , L A B C D E F G ZONES Figure 3 Density of queen conch juveniles at Neighbor Cay. Bahamas, shown as a function of habitat zone (see Fig. 2). The 0+ year class comprises conch <60mm shell length. The >0+ category includes all juvenile conch >80mm. Values are mean +SD. Sandt and Stoner: Ecology of earlyjuvenile Strombus gigas 521 Table 2 Daily growth rate in shell length and body condition factor1 of early juvenile conch (37^19 mm shell length) held in enclosures on sand, seagrass, and rock habitats. Values are means ±SD. Two enclosures with eight conch each were used in the analysis. Differences among treatments were different for both growth rate (ANOVA on log- transformed data, F = 769.35, P < 0.001 1 and condition factor luntrans- formed data, F = 12.82, P = 0.034). Letter codes indicate mean values which were not significantly different I Tukey's multiple comparison test, P> 0.05). Treatment Sand habitat Seagrass habitat Rock habitat Growth rate (mm/day) 0.012 ± 0.001" Condition factor (g/mm) 0.027 ± 0.001" 0.112 ± 0.0101' 0.094 ± 0.002f' 0.042 ± 0.002h 0.039 ± 0.005"b 'Condition factor is defined as the wet weight of soft tissue divided by shell length of an individual. P < 0.001), and Tukey's multiple range test indicated that conch in sand grew at a rate lower than those in seagrass and rock habitats (P < 0.001). Growth rates (P = 0.124) for conch in seagrass and rock were the same. Condition factor was also lower in the sand treat- ment (0.027), than in seagrass (0.042) and rock habi- tats (0.039) (Table 2). The habitat effects were signifi- cant (ANOVA, F = 12.82, P = 0.034). Differences in mean condition factor in seagrass and sand treatments were significant (Tukey's test, P = 0.035), but the dif- ferences were not significant in sand and rock (P = 0.063 ) or seagrass and rock (P = 0.648). Stomach contents of conch from the three different habitats were primarily algae and sand (Table 3). Conch enclosed in the rocky zone had a lower percentage of sand (26.5%) in their stomachs than conch in the other two habitats (48.6-58.2%); however, algae made up over 87% of the organic constitu- ents of the stomach contents in all cases. Detritus particles, identified as T. testudinum, represented a small percentage of stomach contents of conch held in seagrass and rock habitats (Table 3). Fora- minifera and Gastropoda found in the stom- achs were small forms (507f of vitellogenic and mature oocytes; beta atresia. No spermatogenesis; residual spermatozoa in lobules and efferent duct. Resting Primarily CN (most >70um) and perinucleolar oocytes; beta, gamma, and delta atresia. Some mitotic regeneration of spermatogonia and interstitial tissues; lobules and efferent duct empty 528 Fishery Bulletin 9I|3). 1993 nia if CN oocytes were not present. Perinucleolar oo- cytes were not included because this stage was not present in all immature females. To develop the criterion, histological sections of ova- ries from randomly-selected resting (n=21; 319- 449mm) and immature (n=32; 252-410 mm) females were examined and the area of one lobe was measured with an image analysis system. Five specimens were selected, if available, per 25-mm size interval for each of the two reproductive states. Atresia, which is evi- dence of previous oocyte development, was present in each resting female, but it was not present in each immature female. The oocytes and/or gonad area in resting females were noticeably larger than those in immature females; additionally, body length (70 urn in resting ovaries. It was not necessary to use the CN oocyte/oogonium criterion often because >90% of the resting ovaries had beta or delta stage, and with less frequency gamma stage, atretic follicles. Gonad area was also larger in resting ovaries compared to imma- ture ovaries ([7=43, P<0.001, df=l; immature= 3.78+ 3.24mm2 per lobe; resting=9.93+3.33 mm2 per lobe); however, oocyte characteristics should be used to as- sess reproductive state. Use of the CN oocyte/oogonium criterion to assess maturity was not free of error, as the CN oocyte/oogo- nium size distributions for resting and immature fe- males overlapped at 60-80 urn (Fig. 2). This error was negligible because the length-frequency distributions of resting females and females with evidence of cer- tain maturity (e.g., developing, ripe, or spent state) overlapped, and the smallest individuals of both groups were in the same size interval (Fig. 3). Female Spanish mackerel become sexually mature later and at a larger size than males. Mature gonads Schmidt et al.: Spanish mackerel age. growth, and reproduction 529 Table 2 Mean observed lengths at capture (mm FL) and mean back-calculated lengths at age for Spanish mackerel, by sex and for sexes combined. Mean observed No. of length at Mean back calculated lengths at successive annuli speci- Age mens capture 1 2 3 4 5 6 7 8 9 10 11 Males 1 58 372 332 2 45 409 344 396 3 37 430 345 393 423 4 34 454 346 397 426 449 5 22 504 364 416 448 475 499 6 3 469 329 383 411 430 449 466 Total number 199 141 96 59 25 3 Weighted mean 343 399 429 458 493 466 Females 1 129 402 343 2 124 465 362 444 3 93 520 371 453 508 4 49 560 366 446 501 546 5 32 595 364 441 495 537 581 6 7 636 363 448 500 548 594 630 7 4 685 390 468 528 570 610 646 679 8 9 3 0 690 328 419 482 529 569 609 642 675 10 11 0 2 745 364 446 503 541 573 606 632 665 686 718 740 Total number 443 314 190 97 48 16 9 5 2 2 2 Weighted mean 359 447 503 544 584 627 657 671 686 718 740 Sexes combined 1 190 393 337 2 171 450 353 430 3 134 494 358 433 482 4 83 517 349 421 468 506 5 54 558 354 424 472 510 548 6 10 586 345 424 470 510 550 581 7 4 685 392 469 529 571 610 646 679 8 9 4 0 687 337 426 514 538 577 615 646 677 10 11 0 2 745 366 441 504 542 574 606 632 665 686 718 740 Total number 652 462 291 157 74 20 10 6 2 2 2 Weighted mean 349 429 477 511 553 603 657 673 686 719 740 were present in 5% of the females at age 0, 95% at age 1, and 100% at age >2 (Table 3). Eighty-nine percent of the males were mature at age 0 and 100% at age >1 (Table 4). Estimates of L50 were 35.8 + 0.2cm (SE) for females and 23.9 + 0.3cm (SE) for males. The smallest mature female was 288 mm, and the largest immature female was 450mm (Table 3). The smallest mature male was 209 mm, and the largest immature male was 336 mm (Table 4). The combined percentage total of developing, ripe, and spent females by month indicated that May through August was the spawning period (Fig. 4). Ripe females were collected in inner continental shelf waters with a depth of ca. 9 m. They were collected in August in Onslow Bay (North Carolina) and off Cumberland Island (Geor- gia), and in May off Charleston (South Carolina). Males were present in ripe condition for a longer period (April- November) than were females (Fig. 4). 530 Fishery Bulletin 91(3). 1993 be as > 120 100 80 60 40 20 1 1 1 1 r O = Immature • = Resting 200 250 300 350 400 450 500 mm FL Figure 2 Individual average diameters (minimum + maxi- mum dimensions/2) of the five largest chromatin nucleolar oocytes or oogonia along a randomly se- lected axis of a histological section from immature (n=32) and resting (n=21) Spanish mackerel, Scomberomorus macula t us. a v o S-, 0> PL, 20 15 10 ^] = Developing, ripe, or spent (n=171) ■i = Resting (n = 272) I I r> , n 301- 401- 501- 601- 701- 325 425 525 625 725 mm FL Figure 3 Length-frequency distributions for female Spanish mackerel, Scomberomorus maculatus, with evidence of certain maturity (e.g., developing, ripe, or spent state) and those in the resting state. All specimens examined histologically. Table 3 Percentage of mature sped nens by size class ir 1136 femal e Span- ish mackerel, Scomberomorus maculatus Specimens in the develop- ing, ripe, spent, or resting reproductive states were considered mature. Al' specimens were examined h stologically. n = nu mber of specimens. mm FL Ag eO Age 1 Age >1 Nc age % n % n % n % n 151-175 0 ( 2) 176-200 0 l 10) 201-225 0 ( 72) 226-250 0 (117) 251-275 0 ( 1) 0 (125) 276-300 0 ( 16) 67 ( 3) 4 ( 84) 301-325 7 ( 55) 83 ( 6) 14 ( 56 1 326-350 2 ( 48) 87 ( 15) 29 ( 56) 351-375 5 ( 20) 85 ( 13) 100 ( 1) 70 ( 33) 376-400 0 i Hi 100 ( 19) 100 ( 161 88 ( 25) 401-425 100 ( 1) 100 ( 26) 100 1 21) 96 ( 24) 426-450 100 ( li 100 ( 20) 100 ( 28) 94 ( 18) 451-475 100 ( 15) 100 i 30) 100 ( 12) 476-500 100 ( 5) 100 i 231 100 ( 7) 501-750 100 i li 100 I 68) 100 ( 32) Totals 5 (153i 95 (123) 100 (187) 673 Table 4 Percentage of mature specimens by size class in 606 male Spanish mackerel, Scomberomorut maculatus. Specimens in the develop ng, ri 3e, spent, or resting reproductive states were considered mature. All speci- mens were examined histologically, n = number of specimens. mm FL Age 0 Ag ?>0 No age « n -; n ', n 151-175 0 ( 1) 176-200 0 ( 12) 201-225 18 ( 50) 226-250 62 ( 68) 251-275 100 1) 79 ( 93) 276-300 33 3) 100 ( li 84 ( 571 301-325 87 23) 100 ( 3) 100 ( 21) 326-350 96 231 100 i 13) 94 ( 18) 351-375 100 2) 100 ( 18) 100 l 231 376-400 100 ( 28) 100 ( 27) 401-425 100 ( 35) 100 ( 9) 426-600 100 ( 54) 100 ( 23) Totals 89 52) 100 (152) (402) Schmidt et al.: Spanish mackerel age, growth, and reproduction 53! Females 27 12 0 14 46 40 41 82 21 34 5 30 100 1 2 3 4 5 6 7 8 9 10 11 12 MM = Developing Mi = Ripe fTTTTTI = Spent Males 1=3 = Resting 5 8 0 60 61 31 20 48 30 39 20 18 100 Figure 4 Percentage composition of reproductive states by month for 340 female and 352 male Spanish mackerel, Seomberomorus maculatus, based on histological criteria. Number of speci- mens examined is above the bar. Discussion While we lack direct evidence that growth rings on otoliths are deposited annually, indirect evidence sup- ports that hypothesis. Otolith radius and FL are pro- portional as defined by the linear regression analysis, and the mean observed and back-calculated lengths at age agreed reasonably. In addition, the growth rings on otoliths apparently formed during a 2 to 3 month period. The relatively high (29%) percentage of oto- liths with opaque margins in November was likely an artifact of the very small sample size ( 7 ) for that month. The period of annulus formation that we observed agrees well with previous reports. Powell (1975) found that annuli formed during May^July in "Florida" (i.e. Gulf of Mexico and Atlantic) Spanish mackerel. Annu- lus formation may be slightly earlier (March-May) in fish from the northern Gulf of Mexico (Fable et al., 1987). We found two 11-year-old fish, extending the reported longevity of Spanish mackerel slightly. Klima's (1959) oldest fish were 6 years old, Powell's (1975) was 8 years old, and Fable et al. (1987) reported one 9-year- old fish. Our mean back-calculated lengths at age tended to be slightly smaller than those of Powell ( 1975) after his standard lengths were converted to fork lengths (using his equations), and our values for von Bertalanffy asymptotic lengths (Lj were similar to his (555mm for males, 694mm for females). In compari- son to the back-calculated lengths (Atlantic and Gulf of Mexico samples combined) of Fable et al. ( 1987), our back-calculated lengths were greater at age 1 and smaller at older ages. Their estimate of asymptotic length for females was similar to ours, but their esti- mate for males (794 mm) was much larger. Fable et al. (1987) reported greater asymptotic lengths for males than for females, while other studies found the re- verse to be true (Powell, 1975; Helser and Malvestuto, 1987; present study). Powell (1975) excluded the old- est fish from all analyses owing to small sample sizes (19 fish age 6-8). Fable et al. (1987) used all age classes regardless of sample size and concluded that the re- sulting growth parameters were more realistic than those reported by Powell (1975). Whether the differ- ences in growth parameter estimates between this and previous studies are due to methodological differences (e.g., sectioned vs. whole otoliths) or to differences in geographical coverage is not known. Examination of 547 female Spanish mackerel <325 mm FL in the present study confirmed the lower limit of size at maturity reported by Finucane and Collins ( 1986), who based their estimate on nine speci- mens <325mm FL from Georgia and the Carolinas (Table 5). Our estimate of size at maturity for females was higher than the estimate of Klima ( 1959). We found that males matured at smaller sizes than reported by Klima (1959) and Finucane and Collins (1986). These differences probably reflect the greater accuracy re- sulting from our histological versus their macroscopic methods. Our age-at-maturity data generally concurred with the qualitative data in previous studies showing that female and male Spanish mackerel mature at ages 0- 1 (Table 5). Using a histological method, Powell ( 1975) 532 Fishery Bulletin 91(3). 1993 Table 5 Summary of information available on size/age at maturity in Spanish mackerel, Scomberomorus maculatus. Size at maturity (mm FL) Age at maturity (yr) Study and location female male female male Present study; N. Carolina to SE Florida Finucane and Collins (1986); Georgia and Carolinas Powell (1975); South Florida, both coasts Klima(1959); SE Florida 288-450 209-336 0-1 275-424 275-399 250-320 280-340 1-? '1-2 '1-2 'Ages should be reduced by one year I see Powell [19751). unable to acquire a sufficient number of specimens at the appropriate time to use their methodology. Our conclusions concerning size and age at matu- rity were based on samples solely from the Atlantic migratory group, on ages from sectioned sagittae, and on histological examination of gonad tissue. Most fe- male Spanish mackerel mature at age 1 and at lengths greater than 36cm FL (Lm). The minumum legal length for Spanish mackerel is currently set at 12 inches (30.5 cm) FL by the South Atlantic Fishery Manage- ment Council and most southeastern Atlantic states have adopted the same regulation. Thus, the harvest of age-0 immature females is permitted. Federal and state agencies responsible for the management of Atlantic group Spanish mackerel may wish to re- examine minimum length regulations in light of our results. Acknowledgments found vitellogenic and/or mature oocytes in >50% of age-1 females sampled in Florida (Atlantic and Gulf coasts) during April through September. Klima (1959), using a macroscopic method, found that unreported percentages of ages 1-2 females and males were ma- ture; however, Powell (1975) concluded that the ages in Klima (1959) should be reduced by one year. Thus, according to Powell ( 1975), the data reported by Klima (1959) showed that some age-0 specimens were ma- ture. We found that more males than females (89% vs. 5%) were mature at age 0; most females matured at age 1. Our data on the annual reproductive cycle agreed well with previous conclusions based on gonad condi- tion (Beaumariage, 1970; Finucane and Collins, 1986) and on occurrence of larvae (Collins and Stender, 1987). Spawning occurs in Atlantic group females from mid- spring through summer. For males, we concurred with Finucane and Collins (1986) that developing or run- ning ripe males are present from mid-spring through early fall. Our method of assessing maturity was less efficient and perhaps less accurate than the method used by Hunter et al. (1992) for Dover sole, Microstomas pacificus. Hunter et al. (1992) collected specimens prior to the spawning period, when oocyte development was at the vitellogenic stage. Ovaries without yolked oo- cytes, postovulatory follicles, or atresia were considered immature. The primary advantages of their method are that specimens are collected over a shorter period of time and that it is not necessary to develop a criterion to distinguish resting and immature females. We were We thank Oleg Pashuk and Kathy Grimball for his- tological preparations; Scott Van Sant, Byron White, William Roumillat, Bryan Stone, and SEAMAP project personnel for assisting with specimen acqui- sition and workup; Churchill Grimes and Doug DeVries of the NMFS Panama City (Florida) Labo- ratory for making NMFS samples available; and Charles Schaefer and John Tucker for arranging the purchase and shipping of specimens from Florida. Charles Wenner, George Sedberry, and Churchill Grimes reviewed the manuscript and provided many valuable comments. This work was conducted under a MARMAP contract between the South Carolina Wildlife & Marine Resources Department and the National Marine Fisheries Service. Literature cited Beaumariage, D. 1970. Current status of biological investigations of Florida's mackerel fisheries. Proc. Gulf Caribb. Fish. Inst. 22:79-86. Carlander, K. D. 1982. Standard intercepts for calculating lengths from scale measurements for some centrarchid and percid fishes. Trans. Am. Fish. Soc. 111:332-336. Collins, M. R., and B. W. Stender. 1987. Larval king mackerel (Scomberomorus cavalla), Spanish mackerel (S. maculatus), and bluefish (Po- matomus saltatrix) off the southeast coast of the United States, 1973-1980. Bull. Mar. Sci. 41:822- 834. Schmidt et al.: Spanish mackerel age, growth, and reproduction 533 Collins, M. R., D. J Schmidt, C. W Waltz, and J. L. Pinckney. 1989. Age and growth of king mackerel, Scorn- beromorus cavalla, from the Atlantic coast of the United States. Fish. Bull. 87:49-61. Fable, W. A. Jr., A. G. Johnson, and L. E. Barger. 1987. Age and growth of Spanish mackerel, Scom- beromorus maculatus, from Florida and the Gulf of Mexico. Fish. Bull. 85:777-783. Finucane, J. H., and L. A. Collins. 1986. Reproduction of Spanish mackerel. Scorn- beromorus maculatus, from the southeastern United States. Northeast Gulf Sci. 8:97-106. Greeley, M. S., Jr., D. Ft. Calder, and R. A. Wallace. 1987. Oocyte growth and development in the striped mullet, Mugil cephalus, during seasonal ovarian re- crudescence: relationship to fecundity and size at maturity. Fish. Bull. 85:187-200. Helser, T. E., and S. P. Malvestuto. 1987. Age and growth of Spanish mackerel in the northern Gulf of Mexico and management impli- cations. Proc. Annu. Conf. Southeast Assoc. Fish and Wildl. Agencies 41:24-33. Hunter, J. R., and B. J. Macewicz. 1985. Rates of atresia in the ovary of captive and wild northern anchovy, Engraulis mordax. Fish. Bull. 83:119-136. Hunter, J. R., B. J. Macewicz, N. C. H. Lo, and C. A. Kimbrell. 1992. Fecundity, spawning, and maturity of female Do- ver sole Microstomus pacificus, with an evaluation of assumptions and precision. Fish. Bull. 90:101-128. Klima, E. F. 1959. Aspects of the biology and the fishery for Scomberomorus maculatus (Mitchilll, of southern Florida. Fla. Board Conserv., Mar. Res. Lab. Tech. Ser. No. 27, 39 p. MacGregor, J. S. 1957. Fecundity of the Pacific sardine {Sardinops caerulea). Fish. Bull. 57:427-449. Poole, J. C. 1961. Age and growth of the fluke in Great South Bay and their significance in the sport fishery. N.Y. Fish Game J. 8:1-11. Powell, D. 1975. Age, growth, and reproduction in Florida stocks of Spanish mackerel, Scomberomorus maculat- us. Fla. Dep. Nat. Resour, Fla. Mar. Res. Publ. No. 5. 21 p. SAFMC (South Atlantic Fishery Management Council). 1988. Summary of the Gulf of Mexico/South Atlantic coastal migratory pelagic fishery management plan, amendment 2. SAFMC, Charleston, SC. Trent, L., and E. A. Anthony. 1978. Commercial and recreational fisheries for Span- ish mackerel, Scomberomorus maculatus. In E.L. Nakamura and H.R. Bullis Jr. (eds.), Proceedings of the mackerel colloquium, p. 17-32. Gulf States Mar. Fish. Comm. No. 4. Wallace, R. A., and K. Selman. 1981. Cellular and dynamic aspects of oocyte growth inteleosts. Am. Zool. 21:325-343. Wenner, C. A., W. A. Roumillat, and C. W Waltz. 1986. Contributions to the life history of black sea bass, Centropristis striata, off the southeastern United States. Fish. Bull. 84:723-741. West, G. 1990. Methods of assessing ovarian development in fishes: a review. Aust. J. Mar. Freshwater Res. 41:199-222. Wilkinson, L. 1987. SYSTAT: The system for statistics. Evanston, IL: SYSTAT, Inc., 1987. Abstract. -Bluefish (Pomatomus saltatrix) are found along the US At- lantic coast from Maine to Florida, and are the object of a major recre- ational fishery as they migrate northward from the South Atlantic Bight in the spring and return south- ward in the late fall. Acceptance of analytic assessment results for blue- fish has been hindered by the lack of a consistent time series of geo- graphically comprehensive age- length keys for converting bluefish lengths to age. We used bluefish length-age data from North Carolina commercial fisheries during 1986 to 1989 to compare the utility of two simple methods for estimating age from length-frequency data with two more rigorous statistical methods. The simple methods were cohort slic- ing using von Bertalanffy growth pa- rameters for bluefish and the appli- cation of a pooled age-length key compiled with data from different times and fishing areas. The two statistical approaches were the iter- ated age-length key method and MULTIFAN. Combined 1986-1989 proportions at age estimated by co- hort slicing, pooled age-length key, and iterated age-length key meth- ods were significantly different from those in the test data, based on a quantitative comparison using the Kolmogorov-Smirnov cumulative dis- tribution test. Proportions at age estimated by MULTIFAN were not significantly different from the test data. Our work suggests that MULTIFAN is the best alternative to a time series of fishery-specific age-length keys for the estimation of Atlantic coast bluefish ages from length data. A comparison of alternative methods for the estimation of age from length data for Atlantic coast bluefish [Pomatomus sa/tatrix) Mark Terceiro National Marine Fisheries Service Northeast Fisheries Science Center Woods Hole. MA 02543 Jeffrey L. Ross North Carolina Department of Environment, Health, and Natural Resources Division of Marine Fisheries Morehead City, NC 28557 Manuscript accepted 16 April 1993. Fishery Bulletin 91:534-549 1 1993). Bluefish (Pomatomus saltatrix) are found along the U.S. Atlantic coast from Maine to Florida, migrating northward from the South Atlantic Bight in the spring and returning southward in the late fall. They are the target of a major recreational fishery along the Atlantic coast, with catches averaging 47,000 metric tons (t) per year during 1980 to 1990. For the same period, the commercial catch of bluefish, mainly by otter trawls, averaged 6,400 1 per year. Atlantic coast bluefish exhibit fast growth during the first two years of life, attaining fork lengths of over 40cm by age 2 (Hamer, 1959; Lassiter, 1962; Richards, 1976; Wilk, 1977). They may reach ages of up to 12 years and sizes in excess of 100 cm fork length and 14 kg in weight. About fifty percent of bluefish reach sexual maturity by the second year of life (Wilk, 1977). Spawning occurs during two major periods: March and April in the South Atlantic Bight near the inner edge of the Gulf Stream, with a peak about 1 April; and June through September in the Mid- Atlantic Bight, with a peak about 1 August (Wilk, 1977; Kendall and Walford, 1979; Nyman and Conover, 1988). There is evidence that spring and summer spawning fish mix ex- tensively during their lifespan, as summer spawning fish have been ob- served to originate from both cohorts (Chiarella and Conover, 1990). Year classes of bluefish therefore consist of varying proportions of spring- and summer-spawned cohorts. In anticipation of a fisheries management plan to regulate com- mercial and recreational catches of Atlantic coast bluefish, the Atlantic States Marine Fisheries Commission (ASMFC) and the Mid-Atlantic Fisheries Management Council (MAFMC), in cooperation with the National Marine Fisheries Service (NMFS), began work on stock assess- ment in 1986. Length-frequency data collected by the Northeast Fisheries Science Center (NEFSC) (bottom trawls, 1976-1986) and by NMFS Marine Recreational Fishery Statis- tics Survey (MRFSS) (1979-1985) were available for analytical assess- ment (i.e., virtual population analy- sis). However, no consistent time se- ries of geographically comprehensive age-length keys were available for es- timating age-frequency from length- frequency data. Therefore, an age- length key (hereafter referred to as the ASMFC pooled key) was devel- 534 Terceiro and Ross: Estimation of age from length data for Pomatomus saltatrix 535 oped by pooling data collected by the fisheries agen- cies of several Atlantic coast states during 1982 to 1986 (Crecco et al.1; Terceiro2). There were concerns about the potential for biased results in using the ASMFC pooled key, because of the influence of interannual variation in growth, recruit- ment, or mortality rate, given the broad temporal and geographic scale over which the age-length data were collected (Westrheim and Ricker, 1978). As a result, no consensus was reached on the utility of the ASMFC pooled key, catch-at-age data, or subsequent analyses presented in Crecco et al.1 and Terceiro2. Following implementation of the bluefish Fisheries Management Plan in 1990, stock assessment work for bluefish was renewed by ASMFC. A yield-per-recruit analysis was developed to provide biological reference points for the new assessment. Parameters of the von Bertalanffy growth function L, = L,nl{l-e~A'"~'o']), de- rived from weighted mean calculated lengths at annu- lus formation presented in NOAA ( 1989), were used as an expedient alternative to the ASMFC pooled key to age MRFSS-sample bluefish length-frequency data by cohort slicing (i.e., solving for t in the von Bertalanffy growth equation, given L,nf, Lt, K, and t0). Length at age, weight at age, and partial recruitment vectors for the yield-per-recruit analysis were developed from this version of the MRFSS length-age data. Published estimates of mean calculated bluefish length at annulus formation exhibit considerable varia- tion (Barger, 1990; Chiarella and Conover, 1990; Hamer, 1959; Lassiter, 1962; NOAA, 1989; Richards, 1976; Wilk, 1977) (Table 1). The variation is likely due to 1) variation in growth over time, 2 1 sampling and avail- ability bias, 3) varying influence of Lee's phenomenon (Jones, 1958), 4) use of different conventions in fixing the assumed birthdate of bluefish, and 5) differential proportions of spring and spawned fish in the length frequency samples collected by different investigators (Wilk, 1977; Kendall and Walford, 1979; Chiarella and Conover, 1990). Variation in mean lengths at age and subsequently derived growth parameters, likely inher- ent when data from several disparate sources are con- sidered, could lead to biased results from cohort slic- ing. The Eleventh NEFSC SAW1 recommended testing 'Crecco, V..M. Terceiro, and C. Moore. 1987. A stock assessment of Atlantic coast bluefish, Pomatomus saltatrix. Special report pre- pared for the Atlantic States Marine Fisheries Commission, Wash- ington, D.C. -Terceiro. M. 1987. Status of Atlantic coast bluefish— 1987. U.S. Dep. Commer. NOAA, Natl. Mar. Fish. Serv., Northeast Fish. Cent., Woods Hole lab. Ref. Doc. 87-10, 54 p. 'NEFSC. 1990. Report of the Eleventh NEFSC Stock Assessment Workshop, Fall 1990; Woods Hole, MA. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Northeast Fish. Sci. Cent. Ref. Doc. 90-09, 121 p. alternative methods, such as mixture of distributions methods, and data sources in lieu of a time series of age-length keys for the estimation of bluefish ages from length-frequency data. This work compares results from the application of 1) the cohort slicing method using growth parameters, 2) a pooled age-length key applied in the standard manner, 3) the iterated age-length key method (IALK) as presented by Kimura and Chikuni ( 1987) and Hoenig and Heisey (1987), and 4) MULTIFAN, a mixture of distributions method (Fournier et al., 1990), to blue- fish length-age data collected by the North Carolina Division of Marine Fisheries (NCDMF) from 1986 to 1989. Our objective was to determine how these meth- ods interpret the known-age test data, and suggest whether any are an acceptable alternative to the re- source-intensive method of developing fisheries spe- cific age-length keys. Methods Data sources North Carolina bluefish length-age test data Bluefish are landed in the commercial long haul seine (April- October), pound net (May-October), and winter otter trawl (October-May) fisheries in the estuarine and coastal waters of North Carolina. The NCDMF has collected biological sample data from these commer- cial fisheries for bluefish since 1982. This time series is the largest and most consistent body of bluefish length-age data available. NCDMF bluefish length-age data collected from spring 1986 to winter 1989 in = 1469, 13-84 cm, ages 0 to 9, ages determined from scale annuli), aggregated by calendar year, were used as a test data set for comparing the performance of alternative methods for estimating bluefish ages from length data (Table 2, Figure 1). Investigators have not agreed on a birthdate con- vention for Atlantic coast bluefish. NCDMF fisheries biologists employ an unweighted average date of 1 June, between peak spawning of the spring ( 1 April ) and summer ( 1 August) aggregations. This date roughly corresponds with the time of annulus formation dur- ing late May to mid-June. Bluefish recruit to North Carolina fisheries during the summer fishing season. Bluefish with fork lengths ranging from 15 to 32 cm that are captured before 1 June and possess scales that do not yet exhibit the first annulus are typically classified as age 0. Inspec- tion of modes in length-frequency data and the timing of spawning suggests that the largest of these bluefish originate from the spring cohort of the previous calen- dar year, while the smallest are fish from the summer 536 Fishery Bulletin 91(3). 1993 Weighted mean calculated fork lengths for fish from the U.S. Atlantic coast. (cm) at Table 1 annulus formation of bluefish [Pomatomus saltatnx) from various studies Source N Age (yr) 1 2 3 4 5 6 7 8 9 10 11 12 Hamer 1959 ? 31.8 38.1 41.9 50.6 55.0 63.2 70.0 72.6 Lassiter 1962 spring spawned 290 26.6 37.6 48.4 53.5 61.2 68.2 74.1 79.2 summer spawned 154 14.8 30.3 41.9 48.4 Richards 1976 64 23.0 40.0 49.0 58.0 64.0 69.0 71.0 Wilk 1977 7,425 20.8 33.6 44.4 55.1 63.1 67.3 71.9 NOAA 1989 429 22.6 38.1 50.2 59.8 67.3 73.1 77.8 81.4 84.2 86.5 Barger 1990 588 29.0 36.1 41.5 47.3 Chiarella and Conover 1990 147 26.5 43.5 53.2 60.7 65.9 69.3 72.0 75.4 79.3 83.8 85.8 87.0 Table 2 No rth Carol na Division of Marine Fisheries 1986-1989 bluefish (Pomatomus saltatrix) length-age test data, n=1469. Length (cm) Numbers at age 0 1 2 3 4 5 6 7 8 9 14 1 15 2 16 11 17 4 6 18 15 5 19 18 9 20 10 8 21 9 11 22 11 5 1 23 10 6 24 18 6 25 28 10 26 38 14 27 39 21 28 54 10 1 29 29 21 1 30 21 21 1 31 13 17 1 32 17 28 2 33 4 37 4 34 3 32 3 35 1 52 3 36 43 4 Terceiro and Ross Estimation of age from length data for Pomatomus saltatm 537 Table 2 (Continued) Length (cml Numbers at age 0 1 2 3 4 5 6 7 8 9 37 1 35 2 38 36 8 39 45 1 40 29 7 41 24 8 42 13 9 43 8 3 44 1 8 45 6 4 46 4 6 47 1 7 48 2 12 49 1 7 50 10 51 11 52 5 53 2 2 54 5 2 55 2 3 56 5 57 2 9 58 1 5 59 1 6 1 60 7 3 1 61 2 5 1 62 2 7 11 63 5 9 2 64 2 5 1 65 3 12 4 66 1 11 11 67 1 9 16 2 68 5 9 2 69 6 11 10 70 3 21 8 71 16 14 1 72 10 13 4 1 73 5 8 6 74 6 8 1 75 1 7 6 1 76 3 5 3 77 1 2 4 2 78 3 3 1 79 1 1 2 3 80 1 4 1 81 1 1 2 82 5 1 83 1 1 84 1 1 Age totals 357 567 144 60 80 110 80 41 21 9 Mean length 25.7 33.7 44.7 59.4 64.8 68.9 72.2 75.2 79.2 79.6 (cm) SD 3.6 6.5 3.7 1.1 1.0 1.2 1.1 0.7 0.5 0.4 CV(%) 14.0 19.3 8.3 1.9 1.5 1.7 1.5 0.9 0.6 0.5 538 Fishery Bulletin 91(3). 1993 1986 n = 418 1989 n = 302 JUuJk, 10 20 30 40 50 60 70 80 90 LENGTH (cm) Figure 1 North Carolina Division of Marine Fisheries (NCDMFl 1986-1989 bluefish iPomatomus saltatrix) length-age test data: percent at length distributions by year. cohort of the previous calendar year. Depending on the exact time of spawning and sampling, the largest age- 0 fish may be 12-15 months old, while the smaller age-0 fish are about 10 months old. Under the birthdate convention of 1 January used by NEFSC fisheries bi- ologists, all of these fish would be classified as age class 1. This difference in classification schemes has caused confusion in previous stock assessments and other research for bluefish. The NCDMF length-age test data in this paper are based on the 1 June birthdate convention. Connecticut, New Hampshire, North Carolina (CNN) length-age data Length-age sample data (/? = 3883, 12-87 cm, ages 0 to 11, ages determined from scale annuli) were available for bluefish collected by state agencies in Connecticut (1984-1985, n = 1452), New Hampshire ( 1986, n = 76), and North Carolina (1982- 1985, n = 2355) (Fig. 2). The 1 June birthdate conven- tion was employed in aging. Connecticut and New Hampshire fish were collected in research surveys, mainly during the summer and fall when most recre- ational fishing occurs (June-October). North Carolina fish were taken throughout the year in proportion to the number of fish landed by commercial fisheries. These data were combined to form a matrix of the distribution of age at length for bluefish. This matrix was applied as a standard age-length key (CNN ALK) to the NCMDF 1986-1989 annual length distributions 30 40 50 60 70 LENGTH (cm) 80 90 40 e- 30 - 20 - 10 0 1234567891011 Figure 2 Connecticut, New Hampshire, and North Carolina length- age data (n=3,883l for bluefish (Pomotomus saltatrix) used as the basis for the Connecticut, New Hampshire, and North Carolina age-length key (CNN ALKl and the interated age- length key (IALK) method. Terceiro and Ross: Estimation of age from length data for Pomatomus saltatrix 539 to estimate proportions at age for comparison with the NCDMF ages. It is important to note that the North Carolina part of the CNN length-age data and the NCDMF 1986-1989 length-age data were compiled by the same age-reader and therefore were not completely independent. NOAA (1989) length-age data Bluefish length-age data collected during NEFSC autumn bottom trawl research surveys, 1985-1987, formed the basis for weighted, mean-calculated lengths at annulus forma- tion presented in NOAA (1989). Samples were aged according to the 1 January birthdate convention. These data provided the von Bertalanffy growth parameters of Linf = 94.6 cm, K = 0.242, and t„ = -0.128, which were used in cohort slicing for estimation of bluefish ages from length data in the revised yield-per-recruit analysis for bluefish3. The NOAA (1989) value of/,, was adjusted to reflect the difference between the 1 January (used by NEFSC) and 1 June (used by NCDMF) birthdate conventions. The difference of 151 days, or 0.414 years, was added to the ages of fish in the raw data used to estimate the parameters. Refitting the data provided a new value of t0 - -0.542 (e.g., a 25-cm bluefish sampled on 1 April would typically be classified as age 0 by NCDMF staff; a 25-cm bluefish aged by cohort slicing with NOAA 1989 parameters and ttl = -0.128 would be 1.140 years old, and classified as age 1; while a 25-cm blue- fish aged by cohort slicing with NOAA 1989 param- eters and t„ = -0.542 would be 0.726 years old and classified as age 0). The NOAA 1989 growth param- eters with the adjusted value of t„ were used in cohort slicing of the NCDMF 1986-1989 annual length- frequency data for comparison with the NCDMF 1986-1989 annual and combined age distributions. Statistical methods Iterated age-length key The iterated age-length key (IALK) method was introduced by Kimura and Chikuni (1987). The method combines the standard age-length key (Kimura, 1977; Westrheim and Ricker, 1978) and mixture of distributions approaches (Hasselblad, 1966; Macdonald and Pitcher, 1979; Schnute and Fournier, 1980) for the resolution of lengths to ages. Application of the standard age-length key method without bias requires that the length-age data and length data have the same underlying age composition (Kimura, 1977). The IALK method has been presented as a potential solution to the problem of applying age-length keys to length-frequency distributions with different time and space characteristics, and therefore potentially differ- ent underlying age distributions (Kimura and Chikuni, 1987; Hoenig and Heisey, 1987). Assumptions for the IALK method are somewhat less stringent than for standard application of the age- length key. The method requires only that the length- age data be adequately sampled and that the associ- ated probabilities of length at age are applicable to the observed length distribution to be aged, although the potential for bias increases if the two data sets have very different temporal and geographic characteristics. The goal of the iterative procedure is to modify the length distribution of the length-age data so that it more closely approximates the length-frequency data to be aged, until the underlying age distributions of the length-age data and length-frequency data also become similar, thus satisfying the assumptions of the standard age-length key concept. Kimura and Chikuni (1987) presented the IALK method in the form of an algorithm, gave conditions for which the algorithm converged to a unique solution, and showed the algorithm was an application of the expectation-maximization (EM) algorithm (Dempster et al, 1977) to mixtures of distributions. Hoenig and Heisey (1987) subsequently used a similar approach for the IALK as an EM algorithm, while making different as- sumptions about the sampling errors of the data. The steps of the IALK algorithm as implemented by Kimura and Chikuni ( 1987) are the following: 1) estimate the proportions at age, with all ages initially having equal probability. Hoenig and Heisey (1987) suggested that the initial estimate of propor- tions at age might be set the same as those in the length-age data; 2) using observed probabilities of length at age cal- culated from the length-age sample data and the cur- rent estimate of proportions at age, calculate the cor- responding age-length key (probabilities of age at length). An intermediate result of this step is the IALK estimate of the observed length distribution; 3) apply the observed length distribution to be aged to the calculated age-length key of step 2; 4) calculate the maximum absolute deviation over ages between the newly estimated proportions at age of the observed length frequency and the proportions at age in the previous iteration. If the deviation is less than some small constant, then stop; otherwise con- tinue with the next iteration of the age-length key. We used the CNN length-age data (Fig. 2) as the source for the probabilities of length at age used in the calculation of the IALK applied to the observed NCDMF 1986-1989 length distributions (Fig. 1). Ini- tial proportions at age were set to the same values as the CNN length-age data, as suggested by Hoenig and Heisey (1987), for ages 0 to 11: [0.272 0.355 0.138 0.060 0.055 0.048 0.040 0.023 0.007 0.001 0.0001 0.0009]. 540 Fishery Bulletin 91(3), 1993 The IALK algorithm was judged to have converged when the maximum absolute deviation summed over iterated proportions at age was less than 0.001 (0.1%). The IALK method was applied to the annual NCDMF 1986-1989 length distributions separately. MULTIFAN MULTIFAN was introduced by Fournier et al. (1990) as an extension of the methods of Schnute and Fournier (1980) and Fournier and Breen (1983). MULTIFAN is a likelihood-based method using the mix- ture of distributions approach for the classification of lengths to age, with consideration of biological con- straints, to simultaneously analyze several length- frequency distributions sampled at different times. The major assumptions of the method include the follow- ing: 1) the lengths of the animals in each age class are normally distributed about the mean length at age; 2) growth follows the von Bertalanffy function; and 3) standard deviation of lengths about the mean length at age varies as a simple function of mean length at age. The log-likelihood function used in MULTIFAN com- pares the expected probability that a fish chosen ran- domly will lie in a given length interval with the ob- served number of fish in that interval for the set of growth parameters being tested. The formal statistical basis of MULTIFAN allows for a structured and rela- tively objective means of evaluating alternative inter- pretations of the processes producing the observed length- frequency distributions (e.g., the growth rate and resulting number of significant age classes, gear selectivity on younger age classes, and evidence of a length dependent trend in the standard deviation of length at age). The basic set of parameters of the MULTIFAN model include the following: 1) the pro- portions at age; 2) the mean length of the first age group; 3) the mean length of the last age group; 4) the von Bertalanffy growth parameter K; 5) two param- eters which determine the standard deviation of length at age; 6) a parameter determining the extent of selec- tivity bias on the first age group; and 7) a parameter determining the overall variance of the sampling er- rors in the length-frequency data sets (see Fournier et al. 1990 for detailed notation of the MULTIFAN model parameterization ). For example, in the case of two suspected age classes underlying a set of five length-frequency distributions, the model would have a total of nine parameters. The addition of one age class to the model (one age, five distributions) would increase the number of param- eters by five, to 14. Accounting for a length dependent trend in standard deviation of length at age (LDSD) in the three age-class case would increase the number of parameters by one, to 15; accounting for selectivity bias on the first age group would again increase the number of parameters estimated by one, to 16. A chi- square test is used to determine the best fitting growth structure. The statistical significance of increases in the log-likelihood values obtained by the addition of age classes are determined by the 0.90 point of the chi-square random variable, to reduce the probability of rejecting an additional age class when it is actually present in the data. The significance of the addition of other parameters is determined at the more conven- tional 0.95 level (Fournier et al, 1990). After an initial inspection of modes apparent in the length-frequency data, initial constraints for MULTIFAN included reasonable ranges for 1) the num- ber of expected age classes, 2) the corresponding val- ues for the von Bertalanffy parameter K, and 3) a range for the mean length at age of an obvious mode. Initial constraints for the NCDMF 1986-1989 bluefish length-frequency data were potential significant age classes from 5 to 12, K range from 0.20 to 0.40, and mean length range for presumed age-0 fish in the 1986, 1987, and 1989 samples set at 15-30 cm, and for age 1 in 1988 at 30-40 cm. Preliminary MULTIFAN runs determined the number of significant age classes un- derlying the NCDMF 1986-1989 sample length fre- quencies. Additional runs with MULTIFAN were made to test for evidence of selectivity bias (SEL) on the first age class and for a length dependent trend in standard deviation of length at age (LDSD) as means of improving the fits. Comparison of methods A qualitative assessment of the utility of each method was made by inspection of the estimated mean lengths at age determined by the different alternatives. The result of ultimate interest from the application of these methods to the NCDMF length frequency distributions, however, is the number of fish per age class. We as- sessed the performance of the alternatives by using the Kolmogorov-Smirnov (K-S) cumulative distribu- tion test (Sokal and Rohlf, 1981) to compare the calcu- lated annual and combined 1986-1989 proportions at age from each aging approach with the NCDMF an- nual and combined 1986-1989 proportions at age. This non-parametric method tests for differences in both the shape and location of two frequency distribu- tions under the null hypothesis that the samples are taken from populations with the same underlying dis- tribution. The proportions at age being compared are considered as two samples from the commercial land- ings. If the maximum unsigned difference between cu- mulative frequency distributions (proportions at age) exceeds some critical value (D) for a given confidence level and sample size, then the null hypothesis is re- jected (Sokal and Rohlf, 1981). The test provides an Terceiro and Ross: Estimation of age from length data for Pomatomus saltatrix 541 indication of how closely the alternative methods can match the interpretation of the age structure commer- cial landings of bluefish in North Carolina that would be provided by the NCDMF 1986-1989 length-age test data. By extension, the utility of these methods as alternatives to a time series of geographically appro- priate age-length keys can be evaluated on a quantita- tive scale. Results Application of the alternative methods to the NCDMF 1986-1989 length data provided estimates of mean length at age that were lower than the NCDMF value for age-0 fish, and generally higher for ages 1 and older. Mean lengths at age estimated by MULTIFAN for ages 3 to 5 were the exception, as they were lower than the NCDMF values (Table 3). For the annual and combined NCDMF 1986-1989 length distributions, cohort slicing using the NOAA (1989) growth parameters consistently provided the poorest match with the NCDMF proportions at age (Tables 4-8, pages 543-547); there were large differ- ences relative to the NCMDF data for ages 0, 4, and 5. Differences were significant (P > 0.05) in 1988 and for the combined distribution. Standard application of the CNN ALK provided im- proved results over simple cohort slicing, and estimated age proportions were not significantly different from the true NCDMF proportions, except when data were combined. Estimated proportions at age for fish of less than 40 cm were very close to the NCDMF propor- tions. The largest value of D always occurred at age 4 or 5, reflecting problems in resolving fish in the 60- 75 cm length interval to the correct age when using the standard age-length key application of the CNN length-age data (Tables 4-8). The IALK method provided initial results that were surprisingly poor. The convergence criteria proposed by Kimura and Chikuni (1987) for IALK estimated proportions at age were intended to stop iteration of the key when the underlying age distributions of the length-age data and the length frequency to be aged closely matched, thus better satisfying the conditions necessary for unbiased application of the key. Two characteristics of bluefish biology, multiple spawned cohorts for the same year class and differen- tial availability to fisheries by age class, appear to have combined to result in enough difference in the length distribution patterns of the CNN and NCDMF data to significantly hinder the effectiveness of the IALK model when the final iteration of the key was used. Iteration of the annual CNN length-age distri- butions reduced the number of bluefish in the 50-60 cm interval, and increased the number of bluefish in the 20-40 cm (ages 0 and 1) and 65-75 cm (ages 4 and 5) length intervals. The resulting length and age distributions of the input CNN length-age data matched the NCDMF lengths more closely than the initial distributions. However, when the final iterated key was applied to NCDMF lengths, the cumulative proportions at age generally provided large values of D for ages 0 and 1, or ages 4 and 5 (Table 9, page 548). Intermediate iterations of the CNN length-age data (generally the first or second iteration) provided a bet- ter match to the NCDMF proportions at age than the final iteration (generally about the 15th iteration). In effect, we limited the degree to which the CNN length- age distribution could be changed by evaluating the application of each iteration of the CNN ALK, and selecting as the best solution the estimated propor- tions at age providing the smallest cumulative differ- ence from the NCDMF cumulative proportions. Under this procedure of evaluation, the IALK provided the same or slightly improved results relative to standard application of the CNN length-age data. Only for the combined 1986-1989 length distributions were the IALK estimated proportions at age significantly differ- ent from the NCDMF proportions at the 5% level (Tables 4-8). MULTIFAN results were inconsistent when con- sidered on an annual basis. The growth parameters selected by MULTIFAN as best describing the growth pattern of bluefish during 1986-1989 (Table 10, page 549) performed relatively poorly in matching the NCDMF 1986 and 1987 proportions at age, but provided a close match to the NCDMF 1988 and 1989 proportions (Tables 4-8). Variation in growth and recruitment to the sampled fishery among cohorts contributed to varying success in matching the annual NCDMF proportions with the MULTIFAN method. The best fitting MULTIFAN run provided the only non-significant value of D (at age 6, the "plus group" of the MULTIFAN model interpretation) for the combined 1986-1989 length distributions obtained among the compared methods. The largest values of D provided by the other methods occurred at ages 4 or 5, indicat- ing problems in resolving the ages of fish 60 cm and larger in accordance with the NCDMF interpretation. Only MULTIFAN provided proportions at ages 4 and 5 that accurately reflected the NCMDF proportions in the combined distributions (Table 8). The improved performance of MULTIFAN when the NCDMF data are evaluated in combined fashion reflects better ad- herence to the underlying concept of the MULTIFAN model (Fournier and Breen, 1983). 542 Fishery Bulletin 9 1 (3), 1993 Table 3 Combined mean fork lengths (cm) at age frcm application of alternative methods for conversion Marine Fisheries (NCDMF) 1986-1989 bluefish iPomatomus saltatnx) length-frequency data to values. of North Carolina Division of ages, compared with NCDMF Method Mean fork length at age ( cm I 0 1 2 3 4 5 6 7 8 9 10 11 CNNALK 25.0 34.6 47.9 60.7 67.3 71.3 74.9 77.8 81.1 82.2 NOAA 1989 /„ = -0.542 24.6 35.9 48.6 59.4 66.5 72.1 77.4 81.1 83.6 — — — IALK 25.1 34.7 47.8 60.7 67.3 71.3 74.9 77.7 80.2 82.4 83.1 82.0 MULTIFAN 24.9 34.5 45.5 54.5 62.0 68.2 73.3 — — — — — NCDMF 1986-1989 25.7 33.7 44.7 59.4 64.8 68.9 72.2 75.2 79.2 79.6 Discussion Cohort slicing, standard application of the CNN ALK, the IALK method, and MULTIFAN are increasingly complex solutions to the problem of estimating ages from length data when appropriate age-length keys are not available. All four methods provided reason- able interpretations of the NCDMF bluefish data by accurately reflecting the dominance of age-0 and age-1 fish in the length distributions. Cohort slicing using the NOAA (1989) parameters generally provided the poorest match with the NCDMF proportions at age, because the method fails to account for the variability in length at age owing to multiple spawning events and the variation in growth among cohorts. The CNN ALK and IALK provided satisfactory reso- lution of bluefish lengths to age, nearly matching those estimated by the best MULTIFAN fit for the combined NCDMF 1986-1989 data. Performance of the CNNALK and IALK was poorest for those bluefish ages with a high degree of overlap in the NCDMF length distribu- tions (usually ages 4 and 5) because of the inability of these methods to accurately portray the variation in growth among cohorts. However, the CNN ALK and IALK provided the most accurate estimates of mean length at age for very large bluefish (age 7 and older) which often may not be sufficiently abundant in length- frequency sample distributions to form significant modes. We note that with the IALK method, the length-age data would usually be applied to a length distribution of unknown age composition, in which case the alter- native method we used for selecting the best IALK solution would not be an option. Kimura and Chikuni < 1987) suggested the IALK method should reduce bias in the resulting age distribution if an age-length key is applied to length-frequency data with potentially different underlying growth or selectivity patterns. Our work confirmed that the IALK algorithm achieves the goal of reducing the differences in underlying age dis- tribution of the length-age data used as a key and the length frequency to be aged. However, when these dif- ferences are so large that even the less restrictive as- sumptions of the IALK method are violated, the IALK method provides little improvement over standard ap- plication of the age-length key. MULTIFAN performed well in resolving the under- lying age structure of bluefish length frequency samples, given the reasonable number (e.g., through age 5 or 6) of age groups that were apparent in the NCDMF 1986-1989 length samples. MULTIFAN was the only method to correctly estimate the relative pro- portions at ages 4 and 5, by accurately reflecting the variation in length at age of fish from the length fre- quency mode at 65-75 cm. Key attributes of the MULTIFAN model which con- tributed to good performance in this exercise with blue- fish data include 1) the large number of parameters estimated by the model, and consideration of variation of mean length at age and selectivity bias on the first age class, 2) use of the entire NCDMF length data series to estimate the growth pattern, and 3) the hy- pothesis testing approach to model evaluation. Based on qualitative evaluation of the estimates of mean lengths at age and the quantitative comparison of es- timated proportions at age with the Kolmogorov- Smirnov cumulative distribution test, we believe that MULTIFAN was the best current alternative to a time series of fishery-specific age-length keys for the esti- mation of bluefish ages from lengths data. The MULTIFAN model is a good starting point for future method development. MULTIFAN assumes Terceiro and Ross: Estimation of age from length data for Pomatomus saltatrix 543 2 m *C .12 a e£ W CM o ^ iO CO eo en IO o o o . est CN bold Q X O — -r CM o 3 o o S - d o o c o O o o o O o O o' c d 3 d o o" o o d d -C aj 5 _c -a ^> * n s 5 S3 • o ti -a a. < c 1 00 en CO iO CM O o o o © © _ IO ^* CN CM CD C r o 3 © © s CM t^ CO CO CO CO — q ~ 3 © © 6 z = t™ © o o o iO tr^ CO = zz O © © a. to ~ CO c zz CO 3 c zz 3 © © &■* CN ■* ~ = z: O i—l c CO 3 © © 03 til -*j 5 O o o o - d d o cd d © © to c =o c C J2 en *^ ° ■', c "S SP-J o « "fi W iO ■ 03 0J *3 Q > .3 _« "-s « CO CO iO CM -r o ^H CO CO CO 00 © cum /Nor jmu o l>- CN iO ~ iO CO - - 3. 3- 3 t CO c- /. ^ cn en en en en en en © 2] © o o o zz d d d d d © i-i ■gS" S2.S t age (P) w Hamp ifference W CO lO CO CO TZ CO -* CM iO o © CM Oh iO iO CN IO IO n zz 3 © © >~j CO ^ — - ZZ 3 3 ~ o 3 © © - o o o o zz o d o" d d © © ra Ol T3 S§ £ .2 3 3 +J CJ c z « §3 o CN CM t> CO CM t- -) f] 3 © © t- ■-- c Q CN o CO CM c 3 3 © © 4 propo annecl Maxi o O = — - — ■ s r zz 3 © © o O o o d o' cd d d d © © 0) k o z 1- ~ «~ Eh z « Z -J CO CO Id CM tj- c— ^ CO CO 3 © © CJ c- CM IO r iO X - - 3 o © ce r- CO 00 - en en en en 3 © © Z c ^ o < © o o - d cd d d d S.2S c* -^ c a >, o u ^ d appl ), and z « z ^ CO en CO CO CO CO -r CM to 7 1 © © a. ia iO CN IO iO CM 3 3 3 3 CO o o q o O O d o cd O d o d 3 d 3 d © © © © ^ 3 ^ w T3 J * b3 2 a i-l 3 -^ — 3 ™ -a *£ 00 X CO 3 iO *«* CO ■^ £>■ to 3 © © C] CO CN - r CM re 3 3 © © Z-. >> o oc -^ i O o> X O o ~ O C o O 3 3 © © Z rt C i O o o = d o' d d o" © © Ol a" « _ oj c s (NCDMF 9) paramet length key **£ 00 ^ CO CO — IO "3* ao CO o O © © o CO c- CN — z? e'- CO ZT- 3 3 © © O <" CO c-- CO CO a en en zr. q 3 © © z rH o o o o c d d d *-i ^ rH rl 0) no J. heri .(19 age <, GO _ CO in r- »o en ^f o CN 3 © © £5S a. CD •*r . — iO i - 3 3 © © o &> « "tf o ~ — o o o 3 3 © © a) O « Z rH o o — o C: d d d 3 d © © C Z u «SJ- « '53 J= h e— 3 ^ £ CO CO m uo ■^ IO 3 © © on o ing :ing o CN m CN CO t- CO ir- 3C en 3 © © Q c: c~- CO CO Q0 en en en - 3 © © O o o — d - d d d 3 i — 1—5 r— \ ■h ri w 2 'P«TJ Q "E .2 2 x 8-^ CO CO CN CN CO UO 00 i> ■^ O © © c o £■ o> CN CN C- rr CO CO - 3 © © — 3 «* O Q Oh a: ^ - 3 C c c 3 3 3 o © S >. >>6 O o o d d r d d d d d © © 03 -O JJ II 2; _C OJ J3 « "E T ™ -' o £ C J Z -S .2 Q 0> bo < o - CN CO ■** IO CO ir- oo en © .-H 544 Fishery Bulletin 91(3), 1993 5 « ; .las & & OH CO CN IC CO ** CD CO Ol o o o h est eCN bold Q l> 71 U3 [> CC 3- S CN r o o o fa '"I : r c s O o o 3 o o o 5 Ol o o o o O d d d odd .- J= T3 *s s - £ 5 cd "3 -a a. < c | 00 CN 35 CD CO CN o c o o o o CJ 00 O m CO CO c 3 c o o o s cc CD CD CD t^ t> 3 p 3 pop E 2 3 H d O o O o o 8Z c & — o ;_* ^;9 2 aj ci bJlJI S> & 00 ■^ CO c^ CN lO CO 3 C o o o a, /- lO p IT2 ■* CO O 3 o o o fa CO CN o 5 s c CN 3 3 o o o ca "Ejd +j K o O d o o o d d d odd to c « & C .2 m ■^ O !. c orti age rtio « Ol in lO ■* ■* COI -I CD in CN CM O prop lina ropo -J < Q in CO - X CO m c O O O o o o o — o 3 S o o qj o ol d o d 3 d o o o odd a, £ a ^ rt oj "43 CJ > .2 « '-c M 00 • r-- 00 cn 01 Ol 3! Ol Ol o d O o o o d d d d d d i-i T3 £ " £ 2 <= age (Pi Hamp ference M 00 »o CN o CM ira CN ■^ t> CN O CN < a. CD Ol lO -r ~ r> Tf CM o O O o CM CO o O o o o O o o p o d o o o o d d d d odd ■s * - CO CD -O c S E °33 e ■§ a 2 W CN m m ■* Tf X CD o o o o in CO c CO X r*T . — i o o o o 5 propo jnnecl Maxi Z^M o o c o o e O o o pop O < d o o o o =' d d d odd Table : saltatnx) n of the G ULTIFAN. 2 « rj < -^ lO m I> CN lO CO w o o o o CO CO lO ■* CN CO 00 - o o o CN CO t-- t^ CO Ol Ol Ol 3) poo o o o o o d d d d ^ ^ ^ 3 .2 S £ « >, O O j2 appl and 2 fcd o < 1 — 1 CN CN o CN in CN ■* t~ moo a. CO 3 lO T Ol c~ •* CN — o o o CM "<* o o 3 o O O 3 o o o &-o~ O o o o o d d d d odd j= « * M 73 J "5 c < S a J 3 +j — S " -a 1? °> <. 00 9 o CN CN U3 ■^ o -* ~ o o o Q a -r s -T Ol ** 71 - o o o 00 "^ i3 O 01 2 ^ S e O O o - o o 3 d o d o d - d o o o odd oi (q 53 ■h k a — 0) c fe ■£ >> s (NCDM 9) param length ke 3°> "^H 00 00 00 00 CO 00 ,_, •>* m c: o o o CJ c CD c m CN lO t^ Ol C o o o O 01 2 w CO o CC o' o 00 o Ol d Ol d d O r-i pop <. 00 00 o o lO lO CN ■* in o o o E < S a. ~ X "> >> o CN ■* - ~ c 3 o o O o o o U O d ~ o o d d 3 d odd ca .0 J II 2 r; 0) J3 " | 'S f 1 2 -S -2 Ci 1J < c - CN CO ^f m CO I> 00 O) Oh Terceiro and Ross Estimation of age from length data for Pomatomus saltatnx 545 8 * -rf 1 10 0 IO CM CM CO in in X 3 0 0 h esti bold CJ X Tl* CO = X 3 = 0 0 H ~ 0 O C ~ 3 0 0 c c 0 0 5 s 0 O O 0 S 0 0 d d d d •-■="? *s s -J "2 £ x § CO 5 -a S t~ 00 ■^i 00 0 CO = r 0 3 0 0 O CO CO ?1 10 t- ~ r ~ r 0 0 H •— 10 CO c^ r~ CO ~ - c 3 0 0 6 2 3 fr 0 0 0 ci c> c 8 2 c tg - CJ r ^9 5 M-"8 Si s c~ . — 1 CO tn CM CO XT' ~ 3 3 0 0 ~ in ft EC CM ri CM ~ 3 — 0 0 H •— -^1 — O = 1 — 1 1 — 1 s 3 0 0 0 eg ac ■+-> 2 0 0 0 O O ~ O 0 c d d d » s ffl C c 3 »l *^ O j c orti age rtio < CM CO CO 03 CM X t> CO = CO CO 0 prop lina ropo: Cj in 3 '" X CO = 3 = 0 0 O O = 0 0 c 0 O s 0 0 0 C 0 d 0 0 d d cv u- > CO cy S o > O - +3 3 ^5 ^ w O CO 10 OS -^i 00 CM 00 ■* r- r- 0 cum /Nor jmu < O C~ X 00 10 CO CO t~ 00 Ol r. CT5 o> o> <3> °? P O CD c 0 0 0 O 3 d d d i-i ■a g 5 S 2 c age IP) Hamp ference m O tJ* CM ■^1 in •^i Tji in X CO 0 CO < ft. t^ 1 — 1 z: t> ~ t> CO = = 0 0 1 — 1 ■* ■ — 1 0 ^ = O s 3 c 0 0 0 O 0 — 0 c: O 0 d d d d -s s - ffl D -CJ W 2 r- c S E rtio icu mu 2 « CM CM CO 10 O) X I> CO CO c 0 0 CJ CM 1- = -r X CO = ~ = 0 0 e6 x) propo Connect N. Maxi z ^ O O — ~ ~ = c 0 ~ = 0 0 o < O O 0 0 — =' 0 3 d d d d j« -S * a 2 « O O (N CO 1—1 00 CM 00 ^ = 0 0 0 CO 00 X ID CO CO I> -'- - = 0 0 h- ~ <~ h 2 J 1— 1 m CO C- 00 Oi C32 05 3: 3 0 0 a o J o < 0 0 0 O c 0 O c ci ^ r-i ^ § .2 2 £ o >> ° .2 -^ h tPoma ard appl K), and 2 « X (M CM I> in CO in 1 — 1 CO 03 3 O O O O o < a. 0 O O rH c> c - 0 0 0 - 0 d d O O d d M TJ J * C< 2 co >-< _3 Ji — < 00 0 0 2 rt a> CO O CO c X CO CM X 3 0 0 ,n ^o CJ 0 -f = ~ — X in i—l 3 = 0 0 oo "^ "S = 0 c c c e 0 O c c; 0 0 0 0 0 ci c =' 0 O d d d d Ol (Q* ffl _ cu c &H "£ >> § E.S 1 s-t < 00 O 05 2 ~ CO -* C3> 00 CM 0 ^1 C- c c 0 0 r~ Er- t- CO in in ~ - c c 0 0 O in CO t> 00 Oi ~ ■3-. r — 0 0 0 O 0 0 0 0 - O "-< "-< rH ,-J »5Ji Cu OO J. heri (19 age 3°> <. 00 O Ol 2 rH co 10 ~ -T in •& EC CO c O O £ < S r~ - = m c CO = c c O O a. -r c i-l 1 — 1 q 3 0 0 O O CO O g c 0 0 — c 3 — O d d d d .sz » ;»■- «3 C a. S; -a .e fa on of ing ui ling t 2 CN 00 - CM O in m ■* = 0 0 / CM C^ -^1 t^ CO 00 - c 0 0 n 0 ID •x t- 00 00 05 o> 35 3 q q 0 w -a Q t 3 co 2 o. t- Ed CM CO ^ ?1 — O) in 05 3: X 0 0 OC Tji 1- CO t^ 1C X -t = = 0 0 ^3 o « -H Q a. TC 1—1 ~ 0 3 - = c 3 0 0 £ >, >>d O CO o o o i est CN bold C| H -1 CO CO ~ CO c c o o fc* « o c 3 c o o o o o o o Crj o! d d d d d d d d d d d ri- CU S j3 "° p? s - g T3 C! — U EG H 5 V- ^1 CD CO ^ "O p LO CO CN 3-. ^^ tr- c = c c o o o CM en CO o i— CN = = = 5 o o fcH CO LO CO e- 00 o> q = - q q q e z => H d d d d o d 8 Z c g - o T ^9 01 CD 2$ « i lo CN CO CO CO CD CO o = o o o a, CN r> CO t- o r- - c ~ o o H CO CN o r >-H »— < c - r - o o CO M 4J g d d d d d d d d d d d d -S _c '■§ 3 -5 J3 w 00 CO en 00 ■^ = CO CO c s o o cum /Nor umu < o CD t~ LC CD o: -x 00 CT> o = o o CN LO CO r- oc O) Ol a> q o o o d d d d d d d d f—I i-H ^h r-i ■5 8" CO Oh c CJ ~ 5 c dj CO CD co "^ .« m 00 to CO a> CO CO CO o r- o O CN < a, CD o X o OJ CO CN 3 o O O CM CO o i— i r— = q = O o o o d d d d d d d d d d d d ■a s - CO 0) T3 W Z r- = 33 S.2E z « o c~ o o CO CO CO CO CO 3 o o Q CN c — CO l> CO CN c c o o 7 propo onnecl . Maxi Z -J O o o o c o o o c r o o d d d d =' d d d - d d d Table ; saltatrix) n of the C ULTIFAN z « LO CO CN CN t- ■* t- t~ o = o o CO tr- CO C~ ~ CO 00 03 = = o o z d o CN io CO t-~ 00 Oi en 3! = c o o d d d d d d d d '-*' •-< r* i-i 3 .2 S C CO >, O _o j3 oma appl and Z M z d 10 . — i CO -- CO CO CO o CO o O O a. CD CO c CM X CN o = o o CN CO o f-H ^^ = O o — ~ o o - "5 -' d d d d d d d d d d d d X rj * co -cJ hJ "S C < 3^ <. QO CO CO o CO OS CO CO c- CO o o o Q CO c 1 — 1 OC Tf CO ~ r o o 00 •" S O en z ^ o d d d d e e d - d o d d — d o o d d OS tn 0) — CD C fe « S? 1 -f 3°> <* 00 ,_, CO 05 LO -cf ^t1 c- c c = o o o IT- LO LO = c- C35 = r r o o O en CO LO CO t> - o> Oi 3 o 3 o o z -' d d d d ~ o" d «oi CD OO ' iheri .(19 age 3°> <. 00 CN CO CO C31 o CO CO = 3 o o £ < 3 a. CO « - ^f t~- CN - = c o o O en CO CN o = *-* o C ~ o c o o CD O g Z "" d d d d d d d d d d d d c z o3 Mar: sing he it &< ty 3 ■*» § LO o en CN i — i o ■** CO C~ r o o 2 5Pc? o ■* [■- CO ^f CN CO CO 00 en ~ o o a CN LO CD t> r CJ> C35 C35 en - q o sioi icir usii o d d d d d d d d d rA ^ i — i z > to T3 Q £ -2 El, 2 LO LO en CO en a> CO C CO r? o o c o B- rt a. "* CN ~ r- t- - 03 CN C o o 2 O * H Q CN CO - = = = O o ~ o o £ >, >. o O d ~ d d c d d d d ~ d d CO XI CD II z Ufl^ 1 ^C 4j _C » ° a> c = Z-oiiO CD ta < o - CN CO Tj- LO CO t- 00 cn o — Terceiro and Ross; Estimation of age from length data for Pomatomus saltatnx 547 x: 00 p q X — q q x - cd o © d d © o -4-J 03 s CO C 0] i cs § tic cfi I 00 I> CO Tf CM CO CN o X X o o t- o a. o ft cd c tt. c^ lO CO CO — /: " X X X o o a ° .5 € B & CO © d d o o d X d X d X d o o d © CD > O D. -H X £ w 00 OS — 00 t © CO ,—t CO CN CN © CO < Q n = lO — CN X X © o 2 © © o r e o o o X X © o i 3 o o ©' © s © d d © d d d CJ CD c T3 2 3 CO S- c w 1 — 1 00 00 CO t> CO LO o c^ 00 00 © B 7 CCJ CD j ^ CJ CO 1- CN CO i> ■* i> 35 © © © o a. CO © © © © © © q CO i-H CJ fa c o' o" d d d o d © d d O r-i -* 00 o 00 •-H [^ i— i © © i — © CN c j < a. CO OS i> »C3 - X CO X X o o -2 CN CO o o 3 c — X o © © o S-. a s © o' © d d d © d d d d d o ft o u ft S '§ cS2 — Tf CN CN o o o j-i "5 $ Cl C 1 - IS - CN X X o o L. O - - = = = ~ q X X o o a C 1—4 ©' o d d = = d d d d d © 00 ^ a c5 5 3 2 « o <: Ui 00 00 © c- -rf LO © X © © r-2 >1 O IC U3 CN oo r- ^0- c- © © X © © ft _Q T3 CO CN CO I> t^ 00 © © © OS X o o C3 o' © d d d = © d d 1-i r-i i-i s o I* 0, ea — ' V, w J3 2 M lO ■* © 00 rt C- H © LO -T o o cG a. in o r~ lO © X CO X X o o CO 1 — 1 CN — z: = r - = X X X © © _3 3 si o o d d - - d © d © d d to a- 05 00 cd S 03 " £ ff CD 2 -* <- 00 rH o CO i> © X LO 00 © X © o a) 00 C| lO y: s c — X r^ o X © o O os o c c — o X O o X X © a © 2 "" o o d d d X d © d © d d £ OS c § 00 J" i— t CD a 5 « <^ 00 •* © ■* lO x CO © I> X o © © ,o — t)XJ ^ cu CJ OS »o CN C- © © X ~ X X o o 5 O os CN CO r- t^ j- © © - X X o o 2 ""■ ©' o d d d d d X ^H ^H rA r* co a> 2 2 CD m£ CO c — 3 J= <. 00 ■* IT3 i/S o CO LO CO ,_, CO X o o £ a, OS CO £ '- 0! T- CN X X a © O os CN CO = = 3 c © X X X o o CD .5 c 2 ~ o c> d d d - X d d d d d ra § To 3 -^ "O fa 'q l« CD o :rs -C a. o a- u a ■ Q CO © t> r. CN t> CN X ■tf o o o Cj ■* ?1 CN © IN * LO 00 © X © o c CN CO t> c- 00 00 © © © X q q o O o d © d = d © d d ,— 4 ^ 1-i en > e> >. lo X CD o 2 s tj ■* d CO C Jx II, ai C " tj eg = fa s Q O CO CO 00 i — i *tf LO -* 00 •^ © o o a, -tf X ~. -*t lO t> LO CN X © o "o CN ©' CO ©' - d o d © d d o d © d X d o d o o © © 01 Si; 2 U X 111 cd O a. CD o 2 OB < © CN CO -"3* LO © c~ 00 © O i-H 548 Fishery Bulletin 91(3). 1993 Table 9 North Carolina D vision }f Marine Fisheries NCDMF l 1986-1989 bluefish (Pomatomus sa Itatrix) annual proportions at age (P) compared with es timates derived by initial application of CNN age -length key using the iterated age-1 ength key method (IALK). Cumulat ve proportions are omitted for brevity. Maximum difference in cun ulative proportions at age (D) in u nderlined bold print. 1986A,05.„ =4,8=0.094, 1987D1I„S„,42S=0.093 1988 A>„5.„S32<=0.107, 1989 D0.06,„=302=0.111. 1986 1987 1988 1989 NCDMF IALK IALK NCDMF IALK IALK NCMDF IALK IALK NCMDF IALK IALK Age P P D P P D P P D P P D 0 0.323 0.289 0.033 0.209 0.245 0.035 0.182 0.082 0.099 0.245 0.291 0.046 1 0.428 0.507 0.045 0.419 0.447 0.064 0.346 0.524 0.080 0.325 0.295 0.017 2 0.072 0.029 0.002 0.082 0.014 0.005 0.151 0.061 0.009 0.099 0.040 0.043 3 0.012 0.012 0.002 0.031 0.021 0.014 0.062 0.077 0.006 0.073 0.132 0.017 4 0.036 0.067 0.033 0.042 0.134 0.078 0.071 0.136 0.071 0.079 0.172 0.109 5 0.065 0.074 0.043 0.073 0.073 0.078 0.059 0.086 0.099 0.109 0.046 0.046 6 0.038 0.007 0.012 0.078 0.031 0.031 0.065 0.021 0.056 0.033 0.003 0.017 7 0.007 0.002 0.007 0.038 0.016 0.009 0.049 0.006 0.012 0.020 0.000 0.003 8 0.014 0.010 0.002 0.019 0.009 0.000 0.009 0.002 0.003 0.013 0.020 0.003 9 0.000 0.001 0.002 0.009 0.005 0.005 0.006 0.001 0.003 0.003 0.000 0.000 10 0.000 0.000 0.002 0.000 0.000 0.005 0.000 0.004 0.000 0.000 0.000 0.000 11 0.000 0.002 0.000 0.000 0.005 0.000 0.000 0.002 0.000 0.000 0.000 0.000 Table 1 0 Matrix of MULTIFAN inverse log-likelihood function values for alter- native interpretations of the number of significant age classes under- lying the North Carolina Division of Marine Fisheries 1986-1989 bluefish iPomatomus saltatrix) length-frequency test data. LISD = length independent standard deviation in length at age, LDSD = length dependent standard deviation in length at age, NOSEL = no selectivity bias on the first age group. SEL = selectivity bias on the first age group. Values in parentheses are the number of parameters in model being tested (e.g., proportions at age, mean lengths at age, A", LDSD, or SEL). Bold underlined value is combination of character- istics selected for final evaluation as the best fitting model, based on the significance of increases in value with added parameters. Ages LISD+NOSEL LISD+SEL LDSD+NOSEL LDSD+SEL 10 11 12 13 4171.7 4171.9 4172.4 4174.2 (24) (25) (25) (26) 4169.9 4180.7 4175.6 4178.4 (28) (29) (29) (30) 4160.4 4170.0 4167.1 4168.9 (32) (33) (33) (34) 4158.9 4169.7 4172.7 4185.8 (36) (37) (37) (38) 4161.9 4175.7 4168.5 4187.0 140) (41) (41) (42) 4160.5 4171.7 4169.9 4188.1 (44) (451 (45) (46) 4160.7 4170.3 4169.6 4187.9 (48) (49) (49) (50) 4161.1 4170.0 4169.1 4187.8 (52) (53) (53) (54) growth follows the von Bertalanffy function. Other growth functions such as the logistic or Gomphertz may be better models of growth for the youngest age classes of fish, or for species that experience periods of rapid growth at intermediate ages fol- lowed by slow growth as they approach asymp- totic sizes (Moreau, 1987; Terceiro4). The normal distribution of length at age as- sumed in MULTIFAN is appropriate for many fish species (Fournier and Breen, 1983; Fournier et al., 1990; Terceiro et al., 1992). Other prob- ability distributions might better describe the variation in length at age for species with special life-history characteristics. Species characterized by multiple spawning events or extreme varia- tion in growth among cohorts due to density de- pendence often exhibit log-normal (long-tailed) distributions of length at age. Consideration of alternative probability distributions would be a useful addition to the MULTIFAN model. For spe- cies subject to a high degree of size selectivity within age, or those experiencing very slow growth at the oldest ages, the inclusion of pa- rameters that account for departures from nor- mality (skewness and kurtosis) could provide a MULTIFAN model that better estimates age from asymmetric distributions of length at age. 4Terceiro, M. 1991. Length composition analysis of Atlantic sea scallop using the MULTIFAN method. In Report of the Twelfth Northeast Regional Stock Assessment Workshop ( 12th SAW), Spring 1991, Appendix 1. p. 1-29. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Northeast Fish. Sci. Cent., Woods Hole Lab. Ref. Doc. 91-03, 187 p. Terceiro and Ross Estimation of age from length data for Pomatomus saltatnx 549 Acknowledgments We appreciate the efforts of an anonymous reviewer, whose critical comments greatly improved the manu- script. Literature cited Barger, L. E. 1990. Age and growth of bluefish Pomatomus saltatrix from the northern Gulf of Mexico and U.S. South Atlan- tic coast. Fish. Bull. 88:805-809. Chiarella, L. A., and D. O. Conover. 1990. A Spawning season and first-year growth of adult bluefish from the New York Bight. Trans. Am. Fish. Soc. 119:455-462. Dempster, A. P., N. M. Laird, and D. B. Rubin. 1977. Maximum likelihood via the EM algorithm. J. Royal Stat. Soc. Ser. B: Methodological 39:1-22. Fournier, D. A., and P. A. Breen. 1983. Estimation of abalone mortality rates with growth analysis. Trans. Am. Fish. Soc. 112:403-411. Fournier, D. A., J. R. Sibert, J. Majkowski, and J. Hampton. 1990. MULTIFAN a likelihood-based method for estimating growth parameters and age composition from multiple length frequency data sets illustrated using data for southern bluefin tuna \Thunnus mac- coyii). Can. J. Fish. Aquat. Sci. 47:301-317. Hamer, P. E. 1959. Age and growth studies of bluefish (Pomatomus saltatrix) of the New York Bight. M.S. thesis, Rutgers Univ., New Brunswick, NJ. Hasselblad, V. 1966. Estimation of parameters for a mixture of nor- mal distributions. Technometrics 8:431-444. Hoenig, J. M., and D. M. Heisey. 1987. Use of a Log-linear model with the EM algo- rithm to correct estimates of stock composition and to convert length to age. Trans. Am. Fish. Soc. 116:232- 243. Jones, R. 1958. Lee's phenomenon of "apparent changes in growth-rate" with particular reference to haddock and plaice. In International Commission for the North- west Atlantic Fisheries, Special Publ. 1. Halifax, Canada. Kendall, A. W. Jr., and L. A. Walford. 1979. Sources and distribution of bluefish, Pomatomus saltatrix, larvae and juveniles off the east coast of the United States. Fish. Bull. 77:213-277. Kimura, D. K. 1977. Statistical assessment of the age-length key. J. Fish. Res. Board. Can. 34:317-324. Kimura, D. K., and S. Chikuni. 1987. Mixtures of empirical distributions: an iterative application of the age-length key. Biometrics 43:23- 35. Lassiter, R. R. 1962. Life history aspects of the bluefish, Pomatomus saltatrix Linnaeus, from the coast of North Carolina. M.S. thesis, North Carolina State College, Raleigh, NC. Macdonald, P. D. M., and 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. Moreau, J. 1987. Mathematical and biological expression of growth in fishes: recent trends and further developments. In R. C. Summerfelt and G. E. Hall (eds. ), Age and growth offish, p. 81-113. Iowa State Univ. Press. Ames, IA. NOAA. 1989. Guidelines for estimating lengths at age for 18 northwest Atlantic finfish and shellfish spe- cies. NOAA Tech. Mem. NMFS-F/NEC-66, 39 p. Nyman, R. M., and D. O. Conover. 1988. The relation between spawning season and the recruitment of young-of-the-year bluefish, Pomatomus saltatrix, to New York. Fish. Bull. 86:237-250. Richards, S. W. 1976. Age, growth, and food of the bluefish (Pomatomus saltatrix) from east-central Long Island Sound from July through November 1975. Trans. Am. Fish. Soc. 105:523-525. Schnute, J., and D. A. Fournier. 1980. A new approach to length frequency analysis: growth structure. J. Fish. Res. Board. Can.37:1337- 1351. Sokal, R. R., and F. J. Rohlf. 1981. Biometry (2nd edition). W. H. Freeman and Company, NY. Terceiro, M., D. A. Fournier, and J. R. Sibert. 1992. Comparative performance of MULTIFAN and Shepherd's Length Composition Analysis (SRLCA) on simulated length-frequency distributions. Trans. Am. Fish. Soc. 121:667-677. Westrheim, S. J., and W. E. Ricker. 1978. Bias in using an age-length key to estimate age- frequency distributions. J. Fish. Res. Board. Can. 35:184-185. Wilk, S. J. 1977. Biological and fisheries data on bluefish, Pomato- mus saltatrix (Linnaeus). NMFS, NEFC, Sandy Hook Lab. Tech. Ser. Rep. 11, Highlands, NJ, 56 p. A statistical method for evaluating differences between age-length keys with application to Georges Bank haddock, Melanogrammus aeglefinus Daniel B. Hayes Woods Hole Laboratory, Northeast Fisheries Science Center National Marine Fisheries Service, NOAA 1 66 Water Street, Woods Hole, MA 02543-1 097 Age-length keys are a critical com- ponent of many methods used to es- timate catch at age (Southward, 1976; Kimura, 1977; Gavaris and Gavaris, 1983; Quinn et al., 1983; Lai, 1987; Martin and Cook, 1990). When formulating age-length keys for use in stock assessment, the ap- propriate time interval for data col- lection and aggregation is an im- portant consideration. Potentially, an age-length key derived from data taken at a single time during the year could be used to compute catch at age for the entire year. This prac- tice runs the risk of producing in- accurate estimates of catch at age, however. Finer time intervals for age-length keys are typically used (e.g., quarterly or monthly) but at a cost of increased sampling require- ments. Generally, intensive sam- pling over brief time intervals provides the most accurate repre- sentation of age at length in the population (Kimura, 1977; Westr- heim and Ricker, 1978). In practice, however, such a strategy is expen- sive to implement because of the labor-intensive nature of collecting and processing large numbers of fish for age data. For the efficient and cost-effective use of sampling resources, it is desirable to sample and age the fewest fish, giving ad- equate precision while avoiding bi- ased results. The data needed can be considerably reduced if appropri- ate time intervals for age-length keys can be defined. 550 A further problem facing fishery scientists is the need for evaluation of samples collected with different gears. Although fishing gears may select for fish size, the question here is whether the gears select for fish with different proportions of age at length. For example, the Northeast Fisheries Science Center (NEFSC) of the National Marine Fisheries Service samples most of the com- mercially important fish stocks off the northeastern United States with bottom trawls during the spring and autumn. The fish collected in these surveys have a substantially differ- ent size composition than the com- mercial catch because the standard survey trawl has a small mesh codend liner. However, it is un- known whether fish of a given length caught in the research sur- vey trawl differ in age composition from fish of the same length har- vested commercially. If the age at length of fish captured is the same for these gears, then a pooled key, including data from research sur- veys, will increase precision by in- creasing sample size but will not bias catch at age estimates. How- ever, if age-length keys differ, then pooling can introduce bias into es- timates of catch at age (Westrheim and Ricker, 1978). The primary objective of this study is to present a method for de- termining whether two age-length keys differ statistically. This method is then applied to Georges Bank haddock to determine what time in- tervals should be used for age- length keys for this stock and to determine if research trawl-survey age data can be combined with age data from commercially harvested haddock. The benefit (in terms of increased precision) of pooling age- length keys, where appropriate, is also determined. Methods Biological sampling Scales were collected from commer- cially landed haddock in the NEFSC port sampling program. Because sampling requests to this program are filled quarterly, age-length keys and the age structure of landings were computed on a quarterly ba- sis as in previous stock assessments (Clark et al, 1982; Gavaris and Van Eeckhaute, 1990). Length at age by month was averaged over several years (1980 to 1988) to determine growth on a monthly basis. Had- dock scales were also collected in the NEFSC bottom trawl research survey. The NEFSC conducts this survey during the spring and au- tumn each year, providing a fishery independent source of age struc- tures. These surveys are conducted from Cape Hatteras, North Caro- lina, to Nova Scotia, Canada, with the use of a stratified random sampling design. Further details concerning sampling procedures in the NEFSC survey are outlined in Grosslein ( 1969) and Azarovitz (1981). Scales were removed from the lat- eral line region below the second dorsal fin. After removal, scales were dried and then impressed on a laminated plastic slide. Scales were viewed at approximately 40 x magnification to determine age. Ages were determined following cri- Manuscript accepted 2 April 1993. Fishery Bulletin: 91:550-557 1 1993) NOTE Hayes: A statistical method for evaluating age-length keys for Melanogrammus aeglefinus 55 1 teria outlined by Penttila (1988) and Jensen and Wise (1962). By convention, a 1 January birthdate was used. Statistical analysis Age-length keys are commonly formed first by obtain- ing a matrix of numbers at age by length interval (Table 1), and then by converting this to a matrix of proportion at age for each length interval. For statisti- cal tests between age-length keys, however, I used the matrix of numbers at age by length. Age-length keys were compared by making tests of significance sepa- rately for each length interval present in both keys where the sample size was greater than six for each age-length key. Fisher's exact test (Siegel, 1956) was used in these comparisons. Because of the large num- ber of tests needed to compare age-length keys, experimentwise error was compensated for by adjust- ing the significance level for the individual tests. The significance level for n individual tests (a*) needed to maintain a desired experimentwise error (aexp) was de- termined by the following formula derived from Sokal and Rohlf( 1981): lnll- = l-e Table 1 Number sampled at age by 2-cm length groups from the commercial catch of Georges Bank haddock Melanogrammus aeglefinus , January to March 1983 Length cm) Age Total 2 3 4 5 6 7 8 9+ 42 1 2 0 0 0 0 0 0 3 44 0 6 0 0 0 0 0 0 6 46 0 4 2 0 0 0 0 0 6 48 0 4 1 1 0 0 0 0 6 50 0 3 5 4 0 0 0 0 12 52 0 2 5 6 1 0 0 0 14 54 0 1 8 14 0 0 0 0 23 56 0 0 9 11 0 0 0 0 20 58 0 0 5 12 1 0 0 0 18 60 0 0 3 8 2 0 6 0 19 62 0 0 3 8 4 0 6 0 21 64 0 0 0 12 3 1 2 1 19 66 0 0 0 9 3 0 7 0 19 68 0 0 0 8 0 0 6 1 15 70 0 0 0 4 1 4 8 0 17 72 0 0 0 0 1 3 9 0 13 74 0 0 0 2 0 2 10 1 15 76 0 0 0 0 1 1 9 0 11 78 0 0 0 0 1 1 6 1 9 80 0 0 0 0 0 1 7 1 9 82 0 0 0 0 0 0 7 1 8 84 0 0 0 0 0 0 1 1 2 86 0 0 0 0 0 0 1 0 1 This method can be quite conservative if the power of the individual tests is limited because of small sample sizes (Sokal and Rohlf, 1981). Owing to the conserva- tive nature of this method, two subjective criteria for evaluating test results were used. The first criterion was the number of tests exceed- ing the nominal significance level (i.e., a=0.05) in the set of comparisons of interest. A second criterion for evaluating the significance of individual tests was the pattern of nominally significant tests. If significant dif- ferences were observed between quarters for a given length interval, adjacent length intervals were also expected to show significant differences. Thus, tests suggesting nominally significant differences between quarters for a single length interval adjacent to length intervals where tests indicate no difference were sus- pected of occurring by random chance. Age-length keys from the first and second, second and third, and third and fourth quarters were compared to determine which quarters could be com- bined into a single age-length key. Tests were con- ducted by comparing keys within a year because com- bining age-length keys between years can introduce bias in the resulting key (Westrheim and Ricker, 1978). Additional tests were conducted comparing NEFSC spring survey age-length keys with combined first and second quarter commercial age-length keys. Estimation of catch at age and variance As a check on the potential for intro- ducing bias by combining age-length keys, estimates of the age composition of the commercial catch for each year from 1982 to 1988 were made by us- ing age-length keys combined in four different ways. First, age-length keys were constructed for each quarter by using only fish sampled commercially during that quarter. These catch-at- age estimates served as a basis for comparison since they correspond to the level of temporal aggregation (quarterly) that has commonly been used in Georges Bank haddock assess- ments (Clark et al, 1982; Gavaris and Van Eeckhaute, 1990). The second set of catch-at-age estimates was formed by combining commercial age data from the first half of the year into a single age-length key that was then applied to the first and second quar- 552 Fishery Bulletin 91(3), 1993 ter length-frequency distributions. The third set of catch-at-age estimates was based on quarterly age- length keys but included data from haddock sampled in the commercial catch and from the NEFSC bottom trawl surveys. The final set of age-length keys con- tained data from haddock caught both commercially and in trawl surveys but combined all data for the first half of the year. I treated length-frequency samples as simple ran- dom samples from the entire stock area in the compu- tation of the variance of catch at age. Following Gavaris and Gavaris (1983), the proportion of catch at age for unpooled samples was estimated as Ac - 2. L P.,, where, Aiq = Estimated proportion at age i in quarter q Liq - Proportion of total individuals at length j in quarter q PIJq = Proportion of age i individuals at length j in quarter q. For samples where age-length keys were pooled across quarters, the proportion of catch at age was estimated as A,, = 2-, Ljq Pljq j where P',iq is the proportion of age i individuals at length j for the pooled quarters, and L*q is the pooled propor- tion at length j in the pooled quarters. These propor- tions were calculated as p, _ nllt + nlj2 (L^N. + iL^N, Ni+N2 where, nlj: = # of age i fish at length j> in first time period nlj2= # of age i fish at length./ in second time period n,, - # offish at length/ in first time period nj2 = # offish at length,/ in second time period n\ N, total # offish landed in first time period total # offish landed in second time period. This method of pooling age-length keys treats each observation as a random sample from a single popula- tion. No weighting factors were applied when age- length keys were pooled. A pooled proportion of total individuals at length was computed as a stratified ran- dom sample where the weighting for each quarter was the total number of individuals landed during that period. This allows for the possibility that the length composition of landings differed between quarters. Estimates of the variance of proportion at age for unpooled data were computed following Gavaris and Gavaris (1983): Var(A„)= X LIPu lk^> w Where nq is the number offish aged in quarter q. Estimates of the variance of proportion at age when age-length keys were pooled were computed with the above formulae, except P]n was substituted for P,jq. Catch at age and variance for each quarter were com- puted following Gavaris and Gavaris (1983). Because these computations do not depend on whether pooled or unpooled age-length keys are used, the formulae are not repeated. Results and discussion Seasonal growth and comparison of age-length keys Most of the annual growth of Georges Bank haddock takes place during the third quarter, from June through September (Fig. 1). From this pattern of annual growth, age-length keys would not be significantly different between the first and second quarters, but would dif- fer significantly betweens the second and third quar- ters, and the third and fourth quarters. Accordingly, tests were conducted between these pairs of quarters to determine if age-length keys could be pooled across any adjacent quarters. Summary statistics of these tests are presented in Tables 2 through 4. Although little growth takes place between the fourth quarter and the first quarter in the following year, landings data are usually finalized on an annual basis; pooling between years is therefore of limited use. Comparisons between the first and second quarters yielded 9 of 94 tests exceeding the 0.05 level (Table 2). None of these tests exceeded the adjusted significance level (a*=0. 00054) needed to maintain an error rate of 0.05 for this set of comparisons. The number of results exceeding 0.05 is not substantially greater than the number of significant results that would be expected based on random chance; further, these differences oc- curred sporadically among the length classes (Table 2). Thus, based on these tests, first and second quarter age-length keys within each year can be treated as samples drawn from the same population and can be pooled. NOTE Hayes A statistical method for evaluating age-length keys for Melanogrammus aeglefinus 553 LLI 60 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC MONTH Figure 1 Monthly mean length at age of Georges Bank haddock Melanogrammus aeglefinus in the commercial catch. 1980- 1988. Significant differences were observed between sec ond and third quarter age-length keys; 45 of 95 com parisons exceeded the 0.05 level, and 19 tests of these tests had probabilities of less than 0.00054 (a) (Table 3). Tests be- tween the third and fourth were not as definitive; 21 of 85 comparisons exceeded the 0.05 level, but no tests exceeded 0.00060 (a). Several tests, however, were close to a' (Table 4). The large number of tests exceeding 0.05 and the occurrence of contiguous blocks of tests with low probability (Table 4) indicate that age- length keys differ between the third and fourth quarters, but the power to detect these differences may have been limited by small sample sizes. In general, the number of fish aged from the fourth quar- ter was lower than those from other quar- ters as indicated by the smaller number of length intervals with sample sizes greater than six. From these results, age- length keys should not be combined across the second and third quarters or third and fourth quarters for haddock. The consequence of pooling age-length keys across quarters when statistical tests indicate significant differences would be to bias catch-at-age estimates (Westrheim and Ricker, 1978). In the case of Georges Bank haddock, the effect of pooling third quarter age-length keys with second quarter keys (for example) would be to bias catch-at-age estimates for the second quarter towards younger fish. This occurs because growth during the third quarter tends to increase the mean length at age, and conversely, fish of a given length tend to be younger during the third quarter than during the second quarter. Comparison of pooled first and second quarter age- length keys with spring NEFSC bottom trawl surveys keys resulted in 7 of 71 tests having probabilities less than 0.05 (Table 5), but none exceeding the a level of 0.00072. Although more "significant" results were ob- tained than would be expected by random chance, no consistent pattern among these differences was appar- ent (Table 5). Further, several of these differences oc- curred in the 60 to 66 cm size classes where the sample size from the bottom trawl survey was generally small ( < 10 fish ), while sample sizes from the commercial catch were often large (>50 fish). Examination of the contin- gency tables for these size classes indicated that the difference in proportion at age between commercial and survey age-length keys was small and was often due to a broader representation of age classes in the commercial data. A broader representation in the com- mercial samples would be expected, however, given the Table 2 Results of Fisher's exact test comparing first and second quarter age-length keys from Georges Bank haddock Mela wgrammus aeglefinus, 1982-1988. En- tries are probability of difference observed occurr ng by random chance Dashes indicate comparisons where the sample size was less than seven with n one of the two quarters being compared. Year Length (cm) 1982 1983 1984 1985 1986 1987 1988 42 _ _ 44 0.307 0.676 — — — — — 46 0.782 0.117 0.559 1.000 — — — 48 1.000 0.804 0.892 0.661 — — 1.000 50 0.888 0.033* 0.653 0.289 — — — 52 0.786 0.819 0.022* 0.363 — — 0.724 54 0.176 0.434 0.143 0.008* — 0.018* 0.026* 56 0.051 0.284 0.207 0.451 — 1.000 0.067 58 0.562 0.749 0.002* 0.074 0.753 0.434 0.744 60 0.177 0.002* 0.950 0.889 0.432 0.303 0.870 62 0.153 0.123 0.163 — 1.000 0.799 0.406 64 0.136 0.372 0.223 — 0.644 0.389 0.464 66 0.300 0.694 0.849 — 0.230 0.950 0.811 68 0.352 0.497 0.373 — 0.418 0.461 0.704 70 0.182 0.032* 0.676 — 1.000 1.000 0.799 72 0.097 0.119 0.584 — 0.143 1.000 0.439 74 0.730 0.808 0.036* — 0.128 — — 76 0.488 0.589 0.096 — — — — 78 — 0.852 0.249 — — — — 80 — 1.000 0.161 — — — — 82 — 0.466 — — — — — *P<0.05 554 Fishery Bulletin 91(3), 1993 Table 3 Results of Fisher's exact test compari ng second ind third quarter age length keys from Georges Bank haddock Melanogrammus aeglefinus, 1982-1988. Entries are probability of difference observed occurring by random chance. Dashes indicate com- parisons where the sample size was less than seven within one of the two quarters being compared. Year Length (cm) 1982 1983 1984 1985 1986 1987 1988 42 0.219 0.170 — — — — 44 <0.001** 0.075 — 0.592 — — — 46 <0.001** 0.017* — 0.020* — — — 48 <0.001** 0.110 0.620 0.013* — <0.001** 0.832 50 <0.001** 0.026* 0.747 0.108 — <0.001*s ' 0.773 52 <0.001** <0.001** 1.000 0.115 1.000 0.003* <0.001** 54 <0.001** <0.001** 0.688 0.011* 0.582 0.012* <0.001** 56 0.042* <0.001** 1.000 0.321 0.429 0.488 <0.001** 58 0.046* <0.001** 0.146 0.053 0.115 0.335 <0.001** 60 0.038* <0.001** 0.863 — <0.001** 0.075 <0.001** 62 0.011* 0.011* 0.106 — 0.002* 0.055 <0.001** 64 0.523 0.015* 0.662 — 0.004* 0.010* — 66 0.246 0.032* 0.354 — 0.292 0.004* — 68 0.068 0.016* 0.192 — 0.046* 0.122 — 70 0.544 0.002* 0.700 — 0.011* 0.046* — 72 0.379 0.022* 0.738 — 0.590 0.058 — 74 0.185 0.544 — — 0.020* 0.755 — 76 0.686 0.407 — — 0.569 — — 78 — 0.427 — — — — — 80 — 0.206 — — — — — 82 — 0.537 — — — — — *P<0.05 **P<0.00054=a* Table 4 Results )f Fisher's exact test comparing third ind fourth qu arter age -length keys from Georges Bank haddock Melanogra mmus aeglefinus, 1982-1988. Entries are probability of difference observed occurring by r; ndom chance Dashes i ndicate corn- parisons where the sample size was less than seven within one of the two quarters being compared. Year Length (cml 1982 1983 1984 1985 1986 1987 1988 42 0.286 1.000 — 44 0.298 0.014* — — — — — 46 0.158 0.003* — — — 1.000 — 48 0.051 0.001* — — — 0.192 0.124 50 <0.001* 0.123 0.123 — — 0.053 1.000 52 0.159 0.315 0.212 — — 0.024* 0.832 54 0.146 0.049* 0.720 — — 0.754 0.211 56 0.205 0.765 1.000 — — 0.050* 0.791 58 0.143 0.009* 0.154 — 0.477 0.675 0.016* 60 0.292 0.174 0.939 — 0.794 0.540 0.014* 62 0.028* 0.032* 0.758 — 0.272 0.244 <0.001* 64 0.265 0.042* 0.132 — 0.085 0.022* — 66 0.014* 0.034* 0.079 — 0.785 <0.001* — 68 0.484 0.937 0.759 — 0.068 0.107 — 70 0011 0.038* 0.218 — 0.207 0.046* — 72 1.000 0.813 0.550 — 1.000 0.408 — 74 0.756 0.162 — — 0.054 0.922 — 76 0.282 0.612 — — 0.713 — — 78 0.355 0.499 — — 1.000 — — 80 0.765 — — — — — — 82 — — — — — — — *P<0.05 larger sample size available from the commercial fishery for fish in these length classes. Based on the lack of pattern among the signifi- cance tests and the lack of tests ex- ceeding a*, it appears that survey gear does not more consistently se- lect for fish of a different age at a given length than do commercial fishing gears. Accordingly data on age at length obtained from fish col- lected in the research surveys can be combined with data from fish sampled commercially. Catch at age and precision Catch-at-age estimates obtained with the various age-length keys showed no indication of systematic bias (Fig. 2). This was expected, as the results of statistical tests indi- cated no significant difference be- tween the age-length keys that were pooled. The precision of catch-at-age estimates, however, did show a trend among the different levels of aggregation (Fig. 3). Combining first and second quarter age-length keys derived solely from commercially caught haddock increased precision for all age groups, particularly for older age classes that typically had small sample sizes within a single quarter (e.g., the 82 and 84cm length classes; Table 1). The inclu- sion of survey age data had a rela- tively smaller effect on the precision of estimates for most age classes ex- cept for age-2 fish (Fig. 3). This oc- curred because the sample size of small haddock was generally less in the commercial samples than in the research survey samples. Statistical considerations Fisher's exact test tests the hypoth- esis that the proportion at age within each length class is no dif- ferent among keys than would be obtained by random chance. Formally, the hypothesis tested NOTE Hayes A statistical method for evaluating age-length keys for Melanogrammus aeglefmus 555 for length class j is Ho: p,lX = p:j2 for all age classes i from source 1 and source 2 Ha: p«i * p,j2 for all age classes i from source 1 and source 2. Consider as an example haddock (Melanogrammus aeglefinus) in the 54-cm size class sampled from the commercial catch from the first and second quarters of 1983. Numbers at age from each of these samples are Age 0 1 2 3 4 5 6 7 8 Total Quarter 1 0 Quarter 2 0 0 0 0 0 1 7 8 14 10 24 0 1 0 0 0 0 23 42 The question asked is whether the two samples are likely to be drawn from the same population or whether they differ sufficiently to indicate that the sampled populations are different. Assuming fixed marginal to- tals, an appropriate test of this hypothesis is Fisher's exact test (Siegel, 1956). Previously this test was im- practical for contingency tables greater than 2x2 because of the amount of computational power required by the algorithms available for its solution. Recent improvements (Pagano and Halvorsen, 1981; Mehta and Patel, 1983; implemented in Ver- sion 6 of SAS [SAS Institute, 1990]) allow problems of this size to be readily solved. Alternative tests exist in the chi-square test of homogeneity (Hennemuth, 1965) and the G1 test (Bishop et al., 1975). These tests have the advantage that age-length keys can be compared in their entirety. Some studies indicate that these tests may also have greater power than Fisher's exact test (DAgostino et al, 1988; Storer and Kim, 1990), but others have disputed the validity of this these assertions (Little, 19894. Also, the chi-square and G2 tests are often viewed as inappropriate when some of the expected values are less than 5 (e.g., Sokal and Rohlf, 1981; Haberman, 1988); a situation that commonly occurs in comparisons of age- length keys. Although grouping data across age or length classes is a way of increasing the expected values for each cell in the con- tingency table, such a procedure results in a loss of statistical power (Cochran, 1952). With Fisher's exact test, there are no re- strictions on the expected values for any cell within the contingency table (Siegel, 1956). In practice, however, each source (i.e., time period) should contain at least six ob- servations because smaller sample sizes do not have sufficient power to resolve even major discrepancies between sources (Bennett and Hsu, 1960). In summary, Fisher's exact test provides a means of testing differences between age-length keys derived from different sources or from different time periods. When age-length keys are pooled across time periods or sources that do not differ significantly, the resulting estimates of catch at age are more precise and impor- tantly do not appear to be biased by the pooling proce- dure. For Georges Bank haddock, age-length keys from commercially harvested haddock from the first and sec- ond quarters can be combined, as well as keys from haddock sampled in the NEFSC research survey. In the future, allocation of sampling effort should con- sider the benefits of pooling age-length keys. Acknowledgments I thank S. Clark, J. Forrester, S. Gavaris, and F. Serchuk for their constructive comments on earlier drafts of this manuscript. Thanks are also expressed to J. Brodziak for his helpful discussions on this subject. Table 5 Results of Fisher's exact te st comparing age- ength keys for Georg es Bank haddock Melanogrammus aeglefinus derived from fish sampled in NEFSC bottom trawl surveys to commercially sampled fish from the first and see- ond qua rters, 1982-1988. Entries are probability of difference observed occurring by random chance. Dashes indicate comparisons where the sample size was less than seven w I bin Mm of the twc sources be ng compared. Length Year (cm) 1982 1983 1984 1985 1986 1987 1988 38 0.209 — 40 0.214 — — — — — — 42 — — 0.054 — 0.350 — — 44 — — 0.151 — 1.000 — — 46 0.087 — — 0.699 0.232 — 1.000 48 0.038* 1.000 — — 0.243 — 1.000 50 0.388 1.000 0.344 — 0.095 — — 52 0.151 0.175 — — 0.220 0.241 — 54 0.107 0.479 0.547 0.123 — 0.360 — 56 0.002* 0.014* 0.146 0.093 0.033* 0.465 1.000 58 0.139 0.127 0.057 0.549 — — — 60 0.019* 0.839 1.000 0.102 0.273 — — 62 0.268 0.470 0.554 0.165 0.179 0.079 — 64 — 0.526 0.163 0.391 0.641 0.302 — 66 0.821 0.302 0.026* 0.768 0.003 — — 68 — 0.191 0.128 0.725 — — — 70 0.625 0.139 0.423 0.745 — — — 72 0.777 0.682 — — — — — 74 — — — 0.066 — — — 76 0.927 0.194 — — — — — *P<0.05 556 Fishery Bulletin 91(3). 1993 tn c « D O I O O 700 600 500 400 300 - 200 100 Figure 2 Mean catch at age of Georges Bank haddock Melanogrammus aegleftnus, 1982-1988, based on different levels of aggregation in the age-length key. Unpooled commercial refers to quarterly age-length keys derived solely from commercial samples. Pooled commercial refers to age-length keys derived solely from commercial samples, but with quarter- 1 and quarter-2 data pooled into a single age-length key Pooled and unpooled commercial and survey refer to the same levels of temporal aggregation as above but include age-at-length data obtained from fish collected in the NEFSC bottom trawl surveys. Vertical lines indicate 2 SE of the mean catch at age. I unpooled commercial | pooled commercial J unpooled commercial and survey J pooled commercial and survey Figure 3 Mean coefficient of variation of estimates of catch at age for Georges Bank haddock Melanogrammus aeglefinus, 1982-1988, based on different levels of aggregation in the age-length key. Symbol definitions as in Figure 2. Literature cited Azarovitz, T. R. 1981. A brief historical review of the Woods Hole Laboratory trawl survey time series. In W. G. Doubleday and D. Rivard (eds.), p. 62-67. Canadian Spec. Publ. Fish. Aquat. Sci. 58. Bennett, B. M., and P. Hsu. 1960. On the power function of the exact test for the 2x2 contingency table. Bio- metrika 47:393-397. Bishop, Y. M. M., S. E. Fienberg, and P. W. Holland. 1975. Discrete multivariate analysis: theory and practice. MIT Press, Cam- bridge, MA. Clark, S. H., W. J. Overholtz, and R. C. Hennemuth. 1982. Review and assessment of the Georges Bank and Gulf of Maine haddock fishery. J. Northwest Atl. Fish. Sci. 3:1- 27. Cochran, W. G. 1952. The x2 test of goodness of fit. Annals of Mathematical Statistics 23:315-345. D'Agostino, R. B., W. Chase, and A. Belanger. 1988. The appropriateness of some com- mon procedures for testing the equality of two independent binomial popula- tions. Am. Statistician 42:198-202. Gavaris, S., and C. A. Gavaris. 1983. Estimation of catch at age and its variance for groundfish stocks in the New- foundland region, p. 178-182. In W. G. Doubleday and D. Rivard (eds.). Sampling commercial catches of marine fish and invertebrates. Canadian Spec. Publ. Fish. Aquat. Sci. 66. Gavaris, S., and L. Van Eeckhaute. 1990. Assessment of haddock on Eastern Georges Bank. Canadian Atl. Fish. Sci. Advisory Comm. Res. Doc. 90/86, 37 p. Grosslein, M. D. 1969. Groundfish survey program of BCF Woods Hole. Commer. Fish. Rev. 31:22- 30. Haberman, S. J. 1988. A warning on the use of Chi-squared statistics with frequency tables with small expected cell counts. J. Am. Statistical Assoc. 83:555-560. Hennemuth, R. C. 1965. Homogeneity of age-length frequen- cies among months and quarters of the year for haddock caught on Georges Bank, 1962. Int. Comm. Northwest Atl. Fish. Res. Bull. 2:76-76. NOTE Hayes A statistical method for evaluating age-length keys for Melanogrammus aeglefinus 557 Jensen, A. C, and J. P. Wise. 1962. Determining age of young haddock from their scales. U.S. Fish Wild]. Serv. Fisheries Bull. 195(61>:439-450. Kimura, D. K. 1977. Statistical assessment of the age-length key. J. Fish. Res. Board Canada 34:317-324. Lai, H. 1987. Optimum allocation for estimating age composi- tion using age-length key. Fish. Bull. 85:179-183. Little, R., J. A. 1989. Testing the equality of two independent bino- mial proportions. Am. Statistician 43:283-288. Martin, I., and R. M. Cook. 1990. Combined analysis of length and age-at-length data. J. Cons. Cons. Int. Explor. Mer 46:178-186. Mehta, C. R., and N. R. Patel. 1983. A network algorithm for performing Fisher's ex- act test in r * c contingency tables. J. Am. Statisti- cal Assoc. 78:427-434. Pagano, M, and K. T. Halvorsen. 1981. An algorithm for finding the exact significance levels of r x c contingency tables. J. Am. Statistical Assoc. 76:931-934. Penttila, J. 1988. Haddock Melanogrammus aeglefinus. In J. Penttila and L. M. Dery (eds.l, Age determination methods for Northwest Atlantic species, p. 23- 29. U.S. Dep. Commer., NOAA Technical Report NMFS 72. Quinn, T. J. II, E. A. Best, L. Bijsterveld, and I. R. McGregor. 1983. Sampling Pacific halibut (Hippoglossus stenole- pis) landings for age composition: history, evaluation and estimation. Int. Pacific Halibut Coram. Sci. Rep. 68, Seattle, WA. SAS Institute Incorporated. 1990. SAS/STAT users guide, Version 6, fourth edi- tion. SAS Inst,. Carey, NC, 1686 p. Siegel, S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill, NY. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, second edition. W. H. Freeman, NY. Southward, G. M. 1976. Sampling landings of halibut for age compo- sition. Int. Pac. Halibut Comm. Sci. Rep. 8, Seattle, WA. Storer, B. E., and C. Kim. 1990. Exact properties of some exact test statistics for comparing two binomial proportions. J. Am. Statis- tical Assoc. 85:146-155. Westrheim, S. J., and W. E. Ricker. 1978. Bias in using an age-length key to estimate age- frequency distributions. J. Fish. Res. Board Canada 35:184-189. Direct validation of black drum (Pogonias crom/s) ages determined from scales Gary C. Matlock Robert L. Colura Lawrence W. McEachron Texas Parks and Wildlife Department 4200 Smith School Road Austin. Texas 78744 In the 1980's increased harvest of black drum (Pogonias cromis) in the Gulf of Mexico raised concerns about the possibility of overfishing (Anonymous1; Murphy and Taylor, 1989). Growth information neces- sary to examine the effects of fishing on population size was limited to rate estimates based primarily on temporal changes in length- frequency distributions (Pearson, 1929; Simmons and Breuer, 1962) and length changes in recaptured tagged fish (Osburn et al., 1980; Doerzbacher et al., 1988). Age in- formation was limited to scale analysis of fish in Texas and Vir- ginia (Pearson, 1929; Richards, 1973). Consequently, scales and otoliths became the focus of study for estimating age and growth rates (Cornelius, 1984; Music and Pafford, 1984; Murphy and Taylor, 1989; Beckman et al., 1990; Peters and McMichael, 1990), but, age data were not directly validated. For ex- ample, Beckman et al. (1990) used the intra-year progression of annu- lus formation on otoliths to conclude that one annulus forms per year. While their indirect evidence is com- pelling, age was not directly vali- dated, and no attempt has been 'Anonymous. 1989. Saltwater finfish research and management in Texas, a report to the Governor and the 71st Legislature. Tex. Parks Wildl. Dep.. PWD R-3400-061-12/88, Austin, TX, 65 p. 558 made to validate age by using the scale method. Thus, the reliability of conclusions concerning the effect of fishing on black drum may be suspect. More cost-efficient studies would result from using scales in- stead of otoliths for ageing if reli- able data could be obtained from scales. The objectives of this study were to validate directly the forma- tion of annuli (one growth check per year) on scales, the time of annulus formation, and the effect of tagging on annulus formation. Methods Black drum were caught in gill nets at randomly selected sites in nine Texas bay systems (Dailey et al, 1991) during spring (April-June) and fall (Sept. -Nov.) 1985-1991. Captured fish were measured for to- tal length (TL) to the nearest mm, tagged with internal abdominal tags with external plastic streamers (Os- burn et al, 1980), and released. Prior to release, at least two scales were removed from the area beneath the distal end of the left pectoral fin immediately ventral to the lateral line (Matlock et al, 1987). Scales were removed from a total of 9,088 released tagged black drum (195- 1,257mm TL) between April 1985 and December 1991. Anyone captur- ing a tagged fish was requested through posters and news-media advertisements to report date, lo- cation, and TL of each recaptured tagged fish, and to return at least one scale collected from the same area from which scales were re- moved at release. Scales from 22 re- captured tagged fish (354-635 mm TL) were returned to the Texas Parks and Wildlife Department (TPWD) by fishermen between May 1985 and December 1991. All re- ported data were assumed accurate. Scales were prepared by washing in soapy water and impressing on a cellulose acetate slide with a roller press (Smith, 1954). The impres- sions were examined at 32 diam- eters magnification with a micro- projector by using incandescent light. Annuli, characterized by breaks in circuli and new radii, were identified following Pearson's (1929) description and separately counted by two examiners without collabo- ration. Scales were read blind (i.e., without knowing TL or date each scale was obtained). Agreement be- tween readers was obtained for all scales on the first reading. Scale ra- dii and distances from the focus to successive annuli were measured along a diagonal line (Fig. 1) to the right antero-lateral scale corner (Klima and Tabb, 1959). Magnified (32x) scale measurements are re- ported unless otherwise noted. Six- teen of the 22 fish had both usable scales and TL measurements at both release and recapture; 17 had usable scales at both times but one fish lacked TL data at recapture (Table 1). The usable data from the remaining five fish were used to ac- complish some of the objectives. Annulus formation was validated by comparing the number of scale growth checks observed at release to those at recapture (Fig. 2). The difference in number of annuli was compared to the expected difference under the null hypothesis: number of scale growth checks per year * 1. Manuscript accepted 24 dune 1993. Fishery Bulletin: 91:558-563 (1993). NOTE Matlock et al Validation of Pogonias cromis ages 559 Figure 1 Black drum scale indicating morphological features used to measure radii and annuli: A : focus, B = annulus, and C = scale margin. The time of annulus formation was estimated by nar- rowing the possible timet s) using the fish with increased numbers of scale annuli. The longest period between release and recapture defined the possible period of formation. This estimate was refined by using the fish free the next longest period, and so on. Fish with no annuli increases were then used to confirm or refine the possible period of annulus formation. The effect of tagging on scale growth was examined by comparing the relationship between scale radius (Y) and total length (X) and the mean distances from scale focus to each annulus at release to those at re- capture. Least-squares linear regression for single Y at each X and analysis of covariance (Sokal and Rohlf, 1981) were used to compare relationships. One-way analysis of variance (Sokal and Rohlf, 1981) was used to compare distances from focus to each of the first three annuli. The lack of scales with more than three annuli at release precluded comparison of other an- nuli. The probability level for all statistical analyses was set at 0.05. Results and discussion The scale method for aging black drum <4 years old is valid. Only one opaque zone was formed each year, between April and May. Eight of the 17 fish with us- able scales at release and recapture had scales with more annuli (1) at recapture than at release (Table 1). All fish except one (B71764) were free during May; the one exception was free through 21 April. This suggests that annulus formation is completed between late April and late May. Data from the five fish not free during April or May support this conclusion; the number of annuli did not increase between release and recap- ture. Two fish free for 1 day in May provided little information; both showed no change in number of an- nuli. Two other fish released in late April and early May did not show an increase in number of annuli. One fish (F31261) released in late April may have formed an annulus just prior to initial capture, since the second annulus was located 154 mm from the scale focus, and the scale radius was 163 mm (Table 1 ). 560 Fishery Bulletin 9 1(3). 1993 Table 1 Release \nd recapture data for each black drum tagged in Texas bays during \pril 1985 through December 1991. Tag When Bay Total length Scale radius- Distance (mm) to annulus No. measured' Date system (mm) (mm) 1 2 3 4 5 B71575 Release 10-28-86 Aransas 388 155 60 Recapture 06-26-87 Aransas 471 192 65 160 B71764 Release 10-28-87 Aransas 533 181 63 92 133 Recapture 04-21-88 Aransas 522 227 69 108 183 213 F21080 Recapture 05-02-85 Galveston 468 186 75 153 170 F24845 Release 10-02-85 Lower Laguna Madre 548 210 65 130 160 Recapture 01-07-86 Lower Laguna Madre 555 200 55 100 165 F24355 Recapture 04-29-86 Lower Laguna Madre 635 219 62 91 156 172 201 F27341 Release 04-18-89 East Matagorda 463 174 71 169 Recapture 10-16-90 East Matagorda 632 245 78 172 213 F27765 Release 05-30-90 East Matagorda 507 200 78 174 Recapture 12-02-90 Gulf of Mexico 554 219 85 180 208 F28057 Recapture 09-10-85 Matagorda 558 250 145 172 210 F28641 Recapture 11-16-87 Matagorda 572 177 111 150 F28988 Release 10-22-87 Matagorda 485 200 85 130 Recapture 01-18-89 Gulf of Mexico 620 230 80 115 183 F29119 Release 05-03-88 Matagorda 418 172 63 127 156 Recapture 03-14-90 Matagorda 610 204 71 134 175 194 F29583 Release 10-24-89 Matagorda 343 136 54 Recapture 05-24-90 San Antonio 441 142 54 113 F29867 Release 05-09-90 Matagorda 374 151 61 128 Recapture 08-14-90 Matagorda 384 165 81 134 160 F31261 Release 04-27-89 Corpus Christi 444 163 74 154 Recapture 03-07-90 Corpus Christi 526 218 88 179 F31642 Release 10-25-89 Corpus Christi 511 194 73 150 Recapture 12-15-89 Corpus Christi Not provided 210 65 157 F31801 Release 05-02-90 Corpus Christi 484 181 82 180 Recapture 01-22-91 Corpus Christi 457 143 80 118 F34101 Release 10-08-87 Galveston 370 124 62 Recapture 02-12-88 Galveston 354 135 66 F37668 Release 05-22-90 Upper Laguna Madre 417 161 68 158 Recapture 05-23-90 LIpper Laguna Madre 417 161 63 158 F37671 Release 05-22-90 Upper Laguna Madre 417 158 65 153 Recapture 05-23-90 LTpper Laguna Madre 417 164 62 159 F42134 Release 06-01-88 San Antonio 468 164 65 103 140 Recapture 12-02-88 San Antonio 584 232 65 154 210 F42489 Release 09-28-89 San Antonio 398 155 79 116 Recapture 01-29-90 Aransas 438 173 81 138 F71439 Recapture 10-28-86 Aransas 560 253 90 185 215 1 Data from all fish with TL measured at release were used to estimate scale radius-TL relationship at release in Table 2; data from all fish measured at recap ure (excludes tag F31642I were used to estimate scale radius-TL relationship at recapture in Table 2. All scale radius and TL data in this table lex cept recapture data for tag F3 16421 were used to est mate the "combined' regression in Tabl e2. 2 32 X magnification Tagging did not appear to influence scale growth or annulus formation, based on scale radius-TL relation- ship, mean distance from focus to annuli, and similar research with a related sciaenid, red drum (Sciaenops ocellatus). The relationship between scale radius and TL (Table 2, Fig. 3) at release was not significantly different from that at recapture (slopes: F=0.14, df=l,34; P=0.80; y-intercepts: F=0.37; df=l,35; P=0.80). The re- gression for all data combined explained 74% of the variation (r=0.86) and indicated scale radius increased 0.36 unit for each unit of TL increase (Table 2). Distance of each of the first two annuli from scale focus at release was not significantly different from that at recapture (first annulus: F=2.437; df=l,37; P=0.14; second annulus: P=0.179; df=l,33, P=0.63). However, the mean distance from scale focus to the third annulus was significantly (P=11.078; df=l,14; P=0.005) greater on scales from fish at recapture NOTE Matlock et al : Validation of Pogonias cromis ages 561 220 % 200 + focus (mm) *. c* co o o o + D Distance from en Co O to o o o o D + 40 20 12 3 4 5 Number of annuli (opaque zones) Figure 2 Scale radius distance ( 32 X magnification 1 to each successive annulus on tagged black drum at release (squares! and at recapture (+). (187 ± 7 mm) than on those from fish at release (147 ± 4 mm). The difference was due largely to small sample size; a third annulus occurred at 133 mm on one fish (B71764) at release but at 183 mm at recap- ture (Table 1). The same tagging procedures we used for black drum were used for red drum without af- fecting scale growth or annulus formation (Matlock et al., 1987). Annuli on black drum scales were formed at mean distances from scale focus (± 1 SE) of 73 ± 3 mm, 143 ± 5 mm, 179 ± 8 mm, 192 ± 20 mm, and 201mm (Table 3). The assumption that fishermen reported TL accu- rately appears valid, but they apparently reported TL less precisely than did TPWD personnel. Evidence for this is that the 95%. confi- dence interval associated with the scale radius-TL relationship was wider for re- capture data than for release data (Table 2), but the regressions were not signifi- cantly different from each other. Ferguson et al. (1984) demonstrated that anglers in Texas generally report accurate length of recaptured tagged fish, although length measurement reported by anglers were less precise than TPWD measurements for the same fish. The life history of black drum and an- nulus formation in related sciaenids fur- ther support that black drum scale an- nulus formation is completed during April and May. Adult black drum spawn dur- ing January through April; peak spawn- ing is in March or April (Murphy and Taylor, 1989). Growth in TL is continu- ous until water temperature cools in win- ter (December through February) when growth slows substantially (Doerzbacher et al., 1988). Growth increases in spring. If scale growth follows a similar pattern, circuli are closer to each other as growth increases after winter (i.e., annulus for- mation is completed in spring). Red drum (>l-year-old) growth pattern is similar to black drum and circuli deposition on red drum scales (annulus formation) occurs as expected (Colura et al., 1984; Matlock et al., 1987; Doerzbacher et al., 1988; Green et al., 1990; Bumgaurdner, 1991). If annulus formation in all black drum bony structures (e.g., scales and otoliths) occurs at the same time, then indirect evidence of annulus formation in black drum otoliths also supports the forma- tion of annuli between April and June (Bechman et al., 1990). Additional research is needed to validate annulus formation on scales with four or more annuli. Acknowledgments We thank the staff of the Texas Parks and Wildlife Department, Coastal Fisheries Branch, who removed scales from black drum at tagging and David Pina for preparing the final drawing of a black drum scale. We also appreciate two anonymous reviewers and Ronald Hardy who provided constructive comments and sug- gestions that improved the clarity of the manuscript. Table 2 Regression statistics for scale radius lY in mm X 32x magnification) — total body length (X in mm) relationships (Y = a + bX) for scales taken from tagged black drum released and recaptured in Texas bays during April 1985 through December 1991. See Table 1 for which fish were used in each analysis. Time measurements recorded No. fish measured a b 957r CI around b r F Release Recapture Combined 17 15.85 0.34 21 17.17 0.35 38 11.44 0.36 0.26-0.42 0.24-0.50 0.29-0.45 0.905H 0.807' 0.860' * 58.38** * 35.20 * 105.33** **P<0.01. 562 Fishery Bulletin 91(3), 1993 + 250 _ + + "c s* o '- 230 a S= 220 + + yS \s^ + C J^ J>^ g" 21 J^s^ E S^ jS + X 200 DO S' yS^ + M /Q j*r " 190 — J^ ^r a + y^ \y^ 6 i8o ySj* ° E a y\s^ m l7° 3 + sg/V D ■o 1 60 — l> CO Qs^\S^ O 150 ny\^ a 5 '«« \ <^ + + 130 340 380 420 460 500 540 580 620 Total length (mm) Figure 3 Relationship between fish TL and scale radius (32X magnification) for tagged black drum at release (squares, and bottom line) and at recapture (+, and top line). Literature cited Beckman, D. W., A. L. Stanley, J. H. Render, and C. A. Wilson. 1990. Age and growth of black drum in Louisiana waters of the Gulf of Mexico. Trans. Am. Fish. Soc. 119:537-544. Bumgaurdner, B. W. 1991. Marking subadult red drums with oxytetracycline. Trans. Am. Fish. Soc. 120:537-540. Colura, R. L., C. W. Porter, and A. F. Maciorowski. 1984. Preliminary evaluation of the scale method for describing age and growth of spotted seatrout (Cynoscion nebulosus) in the Matagorda Bay sys- tem, Texas. Tex. Parks Wildl. Dep., Coast. Fish. Branch, Manage. Data Ser. 57, Austin, TX, 17 p. Cornelius, S. A. 1984. Contribution to the life history of black drum and analysis of the com- mercial fishery of Baffin Bay. Volume II. Tech. Bull. 6, Caesar Kleberg Wildl. Res. Inst., Kingsville, TX, 53 p. Table 3 Mean I±1SE) distance (mm at 32X magnification) from focus to annulus (V) for each annulus on scales taken from black drum tagged and recaptured in Texas bays during April 1985 through December 1991. Mean distance to annulus Time No. (numbei of annuli measured) measurements scales recorded measured 1 2 3 4 5 Release 17 '69±2 (171 J140±7 (14) 3147±4 (4) Recapture 22 '77±4 H44+6 6187±7 192±20 K201 (22) (211 (12) (2) (1) Combined 39 "73±3 "14315 16179±8 ;192±20 "201 (39) (35) (16) (2) (1) 1 Tagged fish used in analysis were B71575, B71764, F24845, F27341, F27765. F28988, F29119. F29583, F29867, F31261, F31642, F31801, F34101, F37668, F37671, F42134, and F42489. 2 Tagged fish used in analysis were the same as in footnote 1 except B71575, F29583. and F34101. I agged fish used in analysis were B71764. F24845, F29119, and F42134. 1 Tagged fish used in analysis were B71575, B71764, F21080. F24845, F24355, F27341, F27765. F28057, F28641. F28988, F29119, F29583. F29867, F31261, F31642, F31801. F34101. F37668, F37671, F42134, F42489, and F71439. 5 Tagged fish used in analysis were the same as in footnote 4 except F34101. 6 Tagged fish used in analysis were B71764, F21080, F24845, F24355, F27341, F27765, F28057, 188. F29119, F29867, F42134, and F71439. ed fish used in analysis were B71764 and F24355. " Tagged fish used in analysis was F24355. NOTE Matlock et al. : Validation of Pogonias cromis ages 563 Dailey, J. A., J. C. Kana, and L. W. McEachron. 1991. Trends in relative abundance and size of selected finfishes and shellfishes along the Texas coast: No- vember 1975-December 1989. Tex. Parks Wildl. Dep., Fish. Wildl. Div, Manage. Data Ser. 53, Austin, TX, 241 p. Doerzbacher, J. F., A. W. Green, and G. C. Matlock. 1988. A temperature compensated von Bertalanffy growth model for tagged red drum and black drum in Texas bays. Fish. Res. 6:135-152. Ferguson, M. O., A. W. Green, and G. C. Matlock. 1984. Evaluation of the accuracy and precision of vol- unteered size data from tagged red drum returns. N. Am. J. Fish. Manage. 4:181-186. Green, A. W., L. W. McEachron, G. C. Matlock, and H. E. Hegen. 1990. Use of abdominal streamer tags and maximum- likelihood techniques to estimate spotted seatrout sur- vival and growth. Am. Fish. Soc. Sym. 7:286-292. Klima, E., and D. C. Tabb. 1959. A contribution to the biology of the spotted weak- fish, Cynoscion nebulosus (Cuvier), from northwest Florida with a description of the fishery. Fla. State. Bd. Conser. Tech. Ser. 30, St. Petersburg, FL, 38 p. Matlock, G. C, R. L. Colura, A. F. Maciorowski, and L. W. McEachron. 1987. Use of on-going tagging programs to validate scale readings. In R.C. Summerfelt and G.E. Hall (eds), The age and growth offish, p. 279-285. Iowa State Univ. Press, Ames. Murphy, M. D., and R. G. Taylor. 1989. Reproduction and growth of black drum, Pogonias cromis, in northeast Florida. N.E. Gulf Sci. 10:127-137. Music, J. F., and J. M. Pafford. 1984. Population dynamics and life history aspects of major marine sportfishes in Georgia's coastal wa- ters. Ga. Dep. Nat. Resour. Coast. Div. Contrib. Ser. 38:1-382. Osburn, H. R., G. C. Matlock, and H. E. Hegen. 1980. Description of multiple census tagging program for marine fisheries management. Ann. Proc. Tex. Chapter Am. Fish. Soc. 2:9-25. Pearson, J. C. 1929. Natural history and conservation of redfish and other commercial sciaenids on the Texas coast. Bull. U.S. Bur. Fish. 44:129-214. Peters, K. M., and R. H. McMichael Jr. 1990. Early life history of the black drum Pogonias cromis (Pisces: Sciaenidae) in Tampa Bay, Flor- ida. N.E. Gulf Sci. 11:39-58. Richards, C. E. 1973. Age, growth, and distribution of the black drum (Pogonias cromis) in Virginia. Trans. Am. Fish. Soc. 102:584-590. Simmons, E. G., and J. P. Breuer. 1962. A study of redfish, Scianeops ocellatus, and black drum, Pogonias cromis. Publ. Inst. Mar. Sci. Univ. Tex. 8:184-211. Smith, S. H. 1954. Method for producing plastic impressions of fish scales without using heat. Prog. Fish-Cult. 16:75- 78. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. W.H. Freeman, NY, 859 p. Incorporation of between-haul variation using bootstrapping and nonparametric estimation of selection curves. Russell B. Millar Department of Mathematics and Statistics. University of Otago PO. Box 56, Dunedin, New Zealand The most frequently used paramet- ric description of trawl selectivity is the logistic curve. If r(l) denotes the retention probability of a length / individual, the logistic selection curve is specified as nh- exp(q + 6/) (1) 1 + exp(a + bl) where a and b are parameters to be estimated. Under this formulation it can be seen that 6>0, because this is a requirement for r(l) to increase with /. Also, c<0 because we require the retention probability of a length- 0 individual to be (effectively) zero. A similar curve is provided by the probit function (the cumulative distribution function of the Normal distribution) which has slightly shorter tails than the logistic curve (McCullagh and Nelder, 1989). Both of these curves are symmetric about the length at which retention is 50f/r, which will be denoted by /so. More generally, lx will denote the length at which retention is x%. Some selectivity data suggest an asymmetric selection curve. The log-log curve (also known as the Gompertz curve) and complimen- tary log-log curve are two param- eter asymmetric curves that can be used in the analysis of count data (McCullagh and Nelder, 1989). Al- though Pope et al. (1975) mention the log-log curve as a potential se- lection curve, neither of these asym- metric curves appears to have been used in published selectivity stud- ies prior to this current study. Richards curves (Richards, 1959) are three parameter curves that generalize the logistic in the form r(l) = ( exp(q + bl) \l + exp(a + bl) (2) Parameter 8 controls the amount of asymmetry with 8>1 or 0<8<1 giv- ing longer tail to the left or right of /50 respectively, and 8=1 giving the symmetric logistic curve. The au- thor (Millar, 1991) has found that the Richards curve will often pro- vide an adequate fit to data in cases where the logistic curve is clearly inappropriate. Selectivity data are count data, and in fitting selection curves to these data it is usual to assume that the counts are binomially distrib- uted (McCullagh and Nelder 1989). Within a single selectivity haul, the binomial assumption is appropriate if the fish encountering the gear be- have independently. It is common practice to fit a selection curve to the data combined over all success- ful hauls, and for the binomial as- sumption to remain valid it is then also necessary to assume that se- lectivity does not vary between hauls. This assumption is not valid in general, owing in part to vari- ables such as catch size and haul duration. Gear saturation may oc- cur for high catch sizes because of reduced selectivity in the latter part of the tow caused by meshes becom- ing clogged with fish (e.g., Suuronen and Millar, 1992) or distorted by the strain upon the gear. Hauls of longer duration may increase selec- tivity (e.g., Clark 1957) by allowing fish more time to escape, notwith- standing that the effect will be con- founded with catch size. If between-haul variation is not of primary interest, then fitting a selection curve to the combined hauls data remains a reasonable approach because the estimated se- lection curve parameters are quite insensitive to violation of the bino- mial assumption (McCullagh and Nelder, 1989). The combined hauls approach can be viewed as model- ling the "average" r(l), where the average is over the population of all hauls that could be made on that fishery. The selectivity hauls must therefore be a representative sample from this hypothetical popu- lation. Between-haul variation does, however, invalidate the estimates of variability for the parameters of the combined hauls fit. To correct for this it is common to apply a good- ness-of-fit based correction to the es- timated standard errors (McCullagh and Nelder, 1989), but Fryer (1991) has demonstrated that this can un- derestimate the effect of between- haul variation. Suuronen and Millar ( 1992) corrected the standard errors by using the replication estimator of dispersion (McCullagh and Nelder, 1989, p. 127). This is a non- parametric estimator that is analo- gous to the pure error sums of squares estimator of linear regres- sion analysis (Myers, 1990). The replication estimate of dispersion has an approximate chi-squared dis- tribution when there is no between- haul variation and within-haul Manuscript accepted 27 April 1993. Fishery Bulletin: 91:564-572 (1993). 564 NOTE Millar: Incorporation of between-haul variation 565 variation is binomial. When between-haul variation is present the approximate chi-square distribution no longer holds because the replicates across different size classes are then not independent. Nonetheless, the es- timator provides a correction to the standard errors that incorporates both between-haul variation and within-haul variation. If between-haul variation is of specific interest then fits to individual haul data are required. Fryer (1991) and Reeves et al. (1992) modelled between-haul vari- ability by permitting parameters a and b of the logis- tic curve (1) to vary between hauls according to a bi- variate normal distribution which can be estimated from the individual haul fits. Suuronen et al. (1991) and Suuronen and Millar (1992) regressed the esti- mated /,,,'s for individual hauls against their catch sizes, and in four of five separate selectivity trials a decrease in /50 with catch size was indicated, though only one of these was statistically significant at the 5% level. These regressions used weights given by the inverse of the estimated variance of the individual haul /50's. In the next section it is shown that neither the indi- vidual haul or combined hauls data from a scallop dredge selectivity study could be adequately modelled by any of the above mentioned approaches and that extreme between haul variation was present. A non- parametric analysis of the combined hauls data was implemented and between-haul variability was incor- porated into the estimates of reliability through bootstrapping. The approach assumed Al) the selection curve r(l) is a nondecreasing func- tion of/. In addition, being a combined hauls approach, it was also assumed that A2) the selectivity tows were representative of tows on the fishery. Material and methods Selectivity trials Selectivity trials were performed onboard the 82-m stern trawler Gadus Atlantica during the last week of August 1991 as part of an Iceland scallop iChlamys islandica) biomass survey for the St Pierre Bank (off the South coast of Newfoundland). The objectives of the study were 1) to summarize the retention proper- ties of the survey dredge by estimating the shell heights Z25, /50 and ln corresponding to 25%, 50% and 75% re- tention, and 2) to estimate the survey dredge's percent retention (by meat weight) of commercial sized (>60 mm shell height) scallops. The survey dredge was a 3.66-m (12-ft) wide off- shore scallop dredge with belly constructed from 3 inch (inside-diameter) metal rings joined together with metal links. One selectivity tow was performed at each of ten locations randomly chosen within the survey area. For these tows, shrimp netting covers (35 mm inside mesh opening) were attached behind the dredge, and chafing gear was used under the bottom cover. The covered dredge was towed over the distance (1.0 nauti- cal mile) and at the speed (3.0kn) used in regular biomass survey tows. The contents of the dredge and covers were separately dumped, carefully picked over, and all Iceland scallops were removed. The scallop catch was weighed and a representative sample of between 20 and 40kg (200-400 scallops), or the entire scallop catch if less than 20 kg, was taken for measurement. Each scallop in the sample was measured to the near- est millimetre in shell height. The catch weights were then used to estimate the size frequencies for the en- tire catch in the dredge and covers. Selectivity analysis We had planned to perform parametric analyses of the individual haul and combined hauls data using the standard maximum likelihood (McCullagh and Nelder, 1989) theory of the binomial model to choose the most appropriate form of the selection curve from those dis- cussed above. However, as seen in the Results section, neither the individual haul nor combined haul data were amenable to parametric analysis. Although it may not be possible to specify a parsi- monious parametric form for the selection curves, it is at least reasonable to insist that they be nondecreasing. That is, the larger a scallop, the greater its chances of being retained in the dredge. The nonparametric sta- tistical technique of isotonic regression fits non- decreasing curves to data. When the data are binomi- ally distributed then the isotonic regression curve is the maximum likelihood fit to the data (Barlow et al., 1972, p. 38). Isotonic regression curves are piecewise linear and can be fitted in an intuitive way using the PAV (pool adjacent violators) algorithm (Barlow et al. 1972, p. 13). In this application, the essence of the PAV algorithm is to pool adjacent size classes whenever their observed retention proportions violate the non- decreasing constraint. Isotonic regression views this violation as an artifact due to insufficient numbers in the "offending" size classes and so the pooling results in a block of size classes having a common observed retention proportion. Barlow et al. (1972) show that the isotonic regres- sion curve is unique and does not depend on the order 566 Fishery Bulletin 91(3), 1993 in which violators are pooled. The PAV algorithm can be implemented on computer as follows: Initially, treat all size classes as blocks of size 1. At each step of the algorithm there is an active block which is compared with the adjacent block in the active direction. The latter will be denoted L or R for active directions left and right, respectively. The initial active block and active direction are the smallest size class and R, re- spectively. The smallest size class is deemed to satisfy the nondecreasing constraint in active direction L. The PAV algorithm proceeds as follows: Bl) If comparison in the active direction results in violation of the nondecreasing constraint, then the two blocks are pooled to form a larger active block and the active direction becomes (or re- mains) L. B2) If comparison in the active direction does not violate the nondecreasing constraint then the ac- tive direction becomes (or remains) R. In addi- tion, • if the active direction was L then the active block remains the active block. • if the active direction was R then the active block becomes the next block on the right. After a finite number of steps, the algorithm termi- nates when the rightmost block is active and R is the active direction. If the observed retention proportions are non- decreasing for increasing /, then the isotonic regres- sion curve is simply given by connecting all the pro- portions together with straight line segments. If the isotonic regression curve fitted to the observed reten- tion proportions is flat (corresponding to a pooled block) at 0.25, 0.50 or 0.75, then the estimated /2S, l-M, or lls is given by the shell size that is the midpoint of that pooled block. For this study, published FORTRAN code (Cran, 1980) for implementation of the PAV algorithm (Barlow et al., 1972) was interfaced to the Splus statistical package. Isotonic regression does not provide an estimate of the standard errors of the estimated /2S, l50, and l7S. These were obtained by bootstrapping the data (Efron, 1982). To this end, the individual selectivity hauls were used to define a "population" of hauls. To include be- tween-haul variability, the bootstrap resamples (with replacement) from this population. Within each resampled haul the retention proportions were also bootstrapped to include within-haul variability. That is, for each size class, the bootstrapped retention pro- portion was the proportion of dredge caught scallops in a sample taken with replacement from the captured (in dredge and covers) scallops of that size. The boot- strap samples were the same size as those represented in the data. For example, in haul 1 there were 48 scallops of 70-mm shell height, of which 28 were caught in the dredge. For this size class, a bootstrap sample of 48 scallops was taken by sampling with replace- ment from the 48 scallops whenever haul 1 was se- lected for the bootstrapped combined hauls. The above resampling scheme was performed 200 times and on each occasion /25, l5 and l7S were esti- mated from isotonic regression fits to the combined hauls data, and percent retention (by meat weight) of commercial sized scallops was calculated by using the shell height to meat weight relationship given in Naidu (1991). Results The first four tows were taken over a relatively smooth bottom consisting mainly of small stones and pebbles. The remaining six tows were taken over a rougher bottom consisting of larger stones, rocks and boulders. The data from tow 5 were discarded owing to a torn cover. The proportion of commercial-sized scallops was lower in hauls 1-4 (58%) than in hauls 6-10 (81%). The weight of trash (rocks, sea cucumbers, starfish, etc.) exceeded the weight of scallops in every haul, particularly so in hauls 6-10. A complete summary of the hauls can be found in Millar and Naidu ( 1991 ). The replication estimate of dispersion, calculated over size classes with a total combined catch of at least 10 scallops, was 828 on 480 degrees of freedom. Under the null hypothesis, H0: (No between-haul variation and binomial within-haul variation! the estimator has an approximate chi-square distribution, hence Hn is rejected with P-value <10 6. Binomial variation within hauls should be a reasonable assumption for scallops, so rejection of H„ suggests significant between-haul variation. Results of parametric analysis Figure 1 shows, for each successful haul and combined- over hauls, the proportions of the covered dredge's catch of scallops that were in the dredge. These retention proportions can be extremely variable, especially for the smaller scallops, because of the low numbers en- countered. Logistic curves fitted to these retention pro- portions are shown as dashed lines. The residual plots show that the fitted logistic curves are inadequate. This is particularly true for the combined hauls data. (The residuals plotted in Fig. 1 are deviance residuals, as defined by McCullagh and Nelder [1989, p.39].) NOTE Millar Incorporation of between-haul variation 567 Logistic curve fits to scallop selectivity data Haul! Residual plot c o a. ° «BBCO» a •* ■ n mil •2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 Haul 2 -2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 - 111 111 ll I II I II III! "I ll| |" "|'| If 1 11 ' -2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 Haul 3 -2 3 7 12 18 24 30 36 42 48 54 60 66 72 73 84 90 96 1 03 ■g ,- mill in '||i|'i| mm i II ||] •2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 Haul 4 ■2 3 7 1! 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 m ■ CO CT - ill P) - I.I aI I . . i i ill oil ■ U c a > 1 'H| 1 1 'T I1 6 U5 . ' r- - 2 3 7 12 18 24 30 36 42 48 54 60 66 72 ,78 84 90 96 103 Haul 6 •2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 <0 <■> "D P 1 III. 1" II III 11" -2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 Shell height (mm) -2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 Shell height (mm) Figure 1 Logistic selection curves fitted to the individual haul and combined hauls data, and the deviance residuals from the fits. 568 Fishery Bulletin 91(3), 1993 g , Haul 7 — 4 < > o 1 •• o o rt^^RTo^7^ ill 1L •2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 Haul 8 iiiiiiimiuiFiiTiTiiTini'iininTrnniiirTnTTinnnTiiiTiiiTiiTTmiTimiirnfTrtTTrriTTirnn 2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 O cm 6r d 1?1U*^" ^ .- - TTF Jl .!..... ,, l.i I. .ill li •2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 Haul 9 •2 3 7 12 18 24 30 36 42 48 54 60 66 72 76 84 90 96 103 8 - C 1 I w Fira nun in -2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 Haul 10 2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 98 103 3 O CO CD in CO o o c CO o > a III 1, III 1 '"If 'I1 1 1'llpllll 'If! -2 3 7 12 18 24 30 36 42 48 54 60 66 72 {8 84 90 96 103 Combined hauls -2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 9 rg en 2 o a e\j t id > CO & 1 1 1,1,1, li i ll 1 li 111 ill ,11m, H' ' 'l|'l ||i||! 18 ?4 30 36 42 48 54 60 66 72 78 84 90 96 103 Shell height (mm) ■2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 103 Shell height (mm) Figure 1 (Continued) Logistic selection curves fitted to the individual haul and combined hauls data, and the deviance residuals from the fits. NOTE Millar: Incorporation of between-haul variation 569 Fits of the complimentary log-log, log-log, and Richards curve were also used on the individual haul data. The complimentary log-log curve fits were mar- ginally better than the logistic fits but were still clearly inadequate. The log-log curve fits displayed worse re- sidual structure than the logistic fits. This is because the log-log curve has a longer tail to the right of Z50, whereas the data suggest a longer tail to the left. The three parameter Richards curve fits provided a big improvement and, though very hard to judge, fits to about half of the individual hauls appeared to be adequate. Figure 2 shows the combined hauls data fits of the logistic, complimentary log-log, log-log, and Richards curves. The Richards curve fit is the only one that could possibly be considered adequate, though there is an obvious clustering of negative residuals for shell heights between 61mm and 71mm. Since this group of residuals contains the estimated value of Z50 (66.8 mm) it is of some concern, and the fit was deemed to be inadequate. (One might consider performing a run's test (say) for independence of the residuals. How- ever, the run's test would be very approximate since it assumes that residuals will be positive or negative with equal probability 0.5. This is not the case for these data, especially for the very small and very large size classes, even when the model is correct.) Nonparametric analysis The nonparametric selection curve fits to the individual haul and combined hauls data are overlaid on propor- tion-retained plots in Figure 3 and the corresponding estimated sizes of 25%, 50%, and 75% retention are given in Table 1. Note that the flat portions of the curves (Fig. 3) correspond to size classes that were pooled. Considerable variability in the estimated Z50's is evident, the smallest being 45.3 mm (haul 7) and the largest 80.1mm (haul 3). Figure 3 suggests that the estimated Z50 for haul 7 may be very unreliable — there were relatively few scallops less than 70 mm in this haul and the observed retention proportions of the smaller scallops are extremely variable because of the low numbers caught. The estimates of Z25, Z50, and Z76 from the combined hauls fit were 50.5, 69.4, and 77.3 mm, respectively. The percentile method (Efron 1982, p. 78), was used to determine approximate confidence intervals from the bootstrap. This gave 95% confidence intervals for Z25, lm, and Z75 of 21.0-53.8 mm, 66.1-72.5 mm and 73.6- 80.5 mm, respectively. The extremely large confidence interval on Z2S reflects the paucity of data for the smaller scallops. Retention by meat weight of commercial sized scallops was estimated to be 73%- with a 95% confi- dence interval of 63%-82%.. Discussion Bootstrapping (resampling) the experimental units (se- lectivity tows) is a natural way to emulate the effect of between-haul variability. In doing so, one requires an automated procedure for fitting a selectivity curve to the bootstrapped combined hauls data. Isotonic regres- sion is well suited to this task because the selection curve for the bootstrapped combined hauls will always satisfy assumption Al. In contrast, parametric selec- tion curves may not be sufficiently flexible to adequately fit all the possible bootstrapped combined hauls data sets. One Referee of this paper made the interesting sug- gestion that it may not matter that parametric fits to the combined hauls data or bootstrapped combined hauls could be inadequate, because the bootstrap pro- cedure should nonetheless be applicable and any prob- lems with the fits would be indicated by wide confi- dence intervals or indications of bias (e.g., Efron, 1982, p. 33). To investigate this, Richards curves were fitted to the same bootstrapped combined hauls used in the nonparametric analysis. The combined hauls fit had Z25, Z50, and Z75 of 48.5 mm, 66.8 mm, and 77.6 mm, re- spectively, and the 95f# confidence intervals obtained from the bootstrap were 33.3-56.6 mm, 59.4-71.5 mm and 74. 1-80.3 mm, respectively. These confidence in- tervals have widths of 23.3, 12.1, and 6.2 mm, respec- tively, compared with 32.8, 6.4, and 6.9 mm from the nonparametric fits. The percent retention value and its confidence interval were the same as for the non- parametric analysis. However, the more subtle and pos- sibly more relevant consequence of bootstrapping with a parametric curve is that the bootstrap indicates the ability of the combined hauls parametric fit as an esti- mator of the parametric fit to the entire hypothetical population of tows on the fishery. The latter may not be adequately modelled by a parametric curve — but the bootstrap will not consider this. The isotonic curve can not suffer this deficiency since the selection curve for the entire hypothetical fishery will be non- decreasing. It was assumed that the selectivity tows were repre- sentative of survey tows on the scallop fishery (as- sumption A2). The selectivity gear used here was a covered survey dredge and it was deployed under sur- vey conditions on a random subsample of survey loca- tions. Assumption A2 will therefore by reasonable pro- vided that the covers on the dredge did not have significant impact on its selectivity. To address this question Millar and Naidu (1991) compared the catch in the covered dredge with that in an uncovered dredge that was towed simultaneously. There was evidence to suggest a possible cover effect for scallops below about 58-mm shell height. This is unlikely to affect the 570 Fishery Bulletin 9 1 [3). 1993 Logistic curve ol &--- 3g= -2 3 7 12 18 24 30 36 42 48 54 60 66 72 78 84 90 06 103 Complimentary log-log curve •nrminrrmn 1 1 1 1 1 1 utt r 2 3 7 12 18 24 30 38 42 48 54 60 imiiinuu imin 72 78 84 90 96 103 Log-log (Gompertz) curve n retained 0 6 0 8 10 jrT---" < < Proporlio 0 0 0 2 0 4 ::\ » ? In 1 54.0 69.8 77.1 2 52.0 70.0 75.1 3 53.5 80.1 84.3 4 29.8 56.5 71.2 6 67.9 76.7 83.9 7 20.5 45.3 74.4 8 36.9 47.8 63.1 9 58.3 72.4 79.1 10 62.4 69.7 82.1 Combined 50.5 69.4 77.3 isotonic fit estimates of /50, lls or percent meat weight retention because bad data for small sizes will not affect the fit to large sizes. The same may not be true of parametric fits. Acknowledgements This work would not have been possible without the logistical support provided by K.S. Naidu and the programing assistance of Noel Cadigan, both of the Department of Fisheries and Oceans, St. John's, Newfound- land, Canada. Literature cited Barlow, R. E., D, J. Bartholomew, J. M. Bremner, and H. D. Brunk. 1972. Statistical inference under order restrictions. Wiley & Sons. NY, 388 p. Clark, J. R. 1957. Effect of length of haul on cod end escapement. ICNAF/ ICES/FAO workshop on selectivity, Lisbon. Paper S25. Cran, G. W. 1980. Amalgamation of means in the case of simple order- ing. Appl. Statist. 29:209-211. Efron, B. 1982. The jackknife, the bootstrap and other resampling plans. SIAM Monograph No. 38, CBSM-NSF. Fryer, R. J. 1991. A model of between-haul variation in selectivity. ICES J. Mar. Sci. 48: 281-290. McCullagh, P., and J. A. Nelder. 1989. Generalized linear models, 2nd edition. Chapman and Hall, London, 511 p. Millar, R. B. 1991. Estimation of asymmetric selection curves for trawls. ICES CM. 1991/B:56. Millar, R. B., and K. S. Naidu. 1991. The size-selectivity of Iceland scallops (Chalmys islandica) in offshore dredges. CAFSAC Res. Doc. 91/81, 17 p. Myers, R. H. 1990. Classical and modern regression with applications, 2nd edition. PWS-Kent. Boston. Naidu, K. S. 1991. An estimate of exploitable Iceland scallop iChalmys islandica ) biomass on St. Pierre Bank, 1990. CAFSAC Res. Doc. 91/46 32 p. Pope, J. A., A. R. Margetts, J. M. Hamley and E. F. Akyuz. 1975. Manual of methods for fish stock assessment, Part III. Se- lectivity of fishing gear. FAO Fish. Tech. Pap. 41, 65 p. Reeves, S. A., D. W. Armstrong, R. J. Fryer, and K. A. ('mill. 1992. The effects of mesh size, cod-end extension length and cod-end diameter on the selectivity of Scottish trawls and seines. ICES J. Mar. Sci., 49:279-288. Richards, F. J. 1959. A flexible growth function for empirical use. J. Exp. Biol. 10, p. 290-300. Suuronen, P., and R. B. Millar. 1992. Size selectivity of diamond and square mesh codends in pe- lagic herring trawls: only small herring will notice the dif- ference. Can. J. Fish. Aquat. Sci. 49:2104-2117. Suuronen, P., R. B. Millar, and A. Jarvik. 1991. Selectivity of diamond and hexagonal mesh codends in pe- lagic herring trawls: evidence of a catch size effect. Finnish Fish. Res. 12:143-156. Biological observations from the Cobb Seamount rockfish fishery Donald E. Pearson National Marine Fisheries Service, Southwest Fisheries Science Center Tiburon Laboratory, 3 1 50 Paradise Drive Tiburon, CA 94920 David A. Douglas Oregon Department of Fish and Wildlife, Marine Field Station 53 Portway Street, Astoria, Oregon 97 1 03 Bill Barss Oregon Department of Fish and Wildlife. Marine Region Marine Science Drive, Bldg. 3 Newport, Oregon 97365 The Cobb Seamount was discovered in 1950 by the crew of the fishery research vessel John N. Cobb (Anonymous, 1950). This submerged volcanic mountain (Budinger, 1967) is located approximately 280 nauti- cal miles (nmi) off the southern Washington coast and comes to within 50 meters of the sea surface. At the 200-m isobath, its area is approximately 32 nmi2 (Budinger, 1967). Various efforts have been made to study the biological, oceanographic, and geologic charac- teristics of the seamount since its discovery (Budinger, 1967; Birkland, 1971; Dower et al., 1992), but these efforts have been somewhat limited in scope and duration. The studies have shown that the seamount has concentrations of commercially valuable species including rock- fishes (genus Sebastes). Most of the information about commercial fishing operations on the seamount is anecdotal and speculative, pri- marily because the seamount is lo- cated outside the Exclusive Eco- nomic Zone (EEZ) and therefore fishing vessels operating there are not subject to fishing regulations or routine sampling. What is known, however, is that bottom trawl, mid- water trawl, gill net, and longline fishing has occurred at various times. Foreign and domestic fleets have fished on the Cobb Seamount from at least the mid 1960's (Sasaki, 1986). Most of the fishing operations appear to have been aimed at sev- eral species of rockfish. Sasaki (1986) suggests that prior to 1985, there was intense fishing pressure on the seamount by the Japanese fishing fleet. In 1985, a few fishermen from Oregon attempted to trawl on the Cobb Seamount but were unsuccessful. It is not known why these efforts failed. Little fishing activity is believed to have occurred on Cobb Seamount be- tween 1985 and September 1991. In 1991, fishermen from Oregon and Washington returned to Cobb Seamount; this time they were suc- cessful. This success drew the at- tention of fishery managers because fish caught outside the EEZ are not regulated by the west coast Fisher- ies Management Plan governing nearshore stocks. Because of the regulatory problems presented by this new fishery, port samplers were instructed to give a high priority to sampling of landings from Cobb Seamount in part as an effort to identify any unusual characteristics which might make landings identi- fiable. In this paper we present some of the unique characteristics of the fish populations found on Cobb Seamount, particularly widow rockfish (S. entomelas). This sea- mount is isolated enough from coastal fish populations so that re- cruitment dynamics, density depen- dent growth, and bioenergetic stud- ies could provide valuable new insights into fisheries biology. Methods Landings of commercially important groundfish are routinely sampled by biological technicians in Oregon ac- cording to standard protocols. Land- ings of fish from Cobb Seamount were sampled by port samplers in Oregon at two or three times the normal level and the majority of landings were sampled. All samples of widow rockfish were taken from vessels by using midwater trawl gear with the same mesh size as used in the nearshore widow rock- fish fishery. During a four-month period, 11 landings were sampled by taking multiple 50-pound subsamples from each landing. Fork lengths were obtained from 891 widow rockfish, otoliths were re- moved from 724 of these fish for age determination by the broken and burnt method (Chilton and Beamish, 1982). Mean length-at-age, age compo- sition, and physical characteristics of the otoliths were compared to samples collected from landings of widow rockfish caught in the near- shore areas of the northern Oregon- southern Washington coast. To eliminate yearly and seasonal bias, comparisons of nearshore widow rockfish to Cobb Seamount widow rockfish were limited to the same months and years for both groups. To compare mean length-at-age of Cobb widow rockfish to the near- shore widow rockfish, only six-year- old fish could be used because of the small number of fish in other Manuscript accepted 27 April 1993. Fishery Bulletin: 91:573-576 ( 1993). 573 574 Fishery Bulletin 91(3). 1993 age groups for both areas. Otoliths were visually in- spected without a microscope to look for gross morpho- logical differences. Results and discussion Several unique features were observed in the landings of widow rockfish from Cobb Seamount. Mean length- at-age was similar for six-year-old widow rockfish in both areas; however, size composition, age composi- tion, and physical structure of the otoliths were very different between the two areas. Mean length of six-year-old widow rockfish on the Cobb Seamount was 35.3 cm FL (ra=72, SE=0.11) and 37.3 cm FL (re=140, SE=0.09) for males and females, respectively. In the nearshore samples, the mean length of six-year-old fish was 35.3cm FL (n=53, SE=0.22) and 37.8 cm FL (ra=52, SE=0.26) for males and females, respectively. This suggests that mean length-at-age, at least for young fish, is similar between fish caught on the Cobb Seamount and fish caught in the nearshore area. Assuming that the fish on the seamount spent the majority of their lives there, then it is reasonable to assume that conditions affecting growth (tempera- ture, food availability, competition, etc.) are probably comparable to the conditions found in the nearshore environment. In the northern Oregon-southern Washington area of the Pacific coast, the mean length of widow rockfish in the commercial landings is 38.4cm (Pearson and High tower, 1991). In contrast, the mean length of widow rockfish from Cobb Seamount was 33.1cm. Age compo- sition of the widow rockfish in the landings from the Cobb Seamount is also very different from those in the landings from the nearshore area (Fig. 1). Only 7.5% of all fish collected from the Cobb Seamount were more than six years old, while 85.7% of fish collected from the nearshore samples were more than six years old (Fig. 1). This difference in age composition cannot be explained by differences in gear type because the same gear and mesh size was used in both areas. There are several potential explanations for these differences in- cluding differences in fishing pressure, differences in fishing behavior, differences in discard practices, and actual differences in the population age composition. The argument for differences in fishing pressure re- lies on the belief that widow rockfish on the Cobb Sea- mount were heavily exploited prior to 1985 and there- fore their population was reduced to very low levels. This argument is supported by Sasaki (1986) who re- ported that the seamount had, in fact, been heavily fished. Widow rockfish were never specifically men- tioned as having been caught by any fishery prior to 1991. While intense fishing pressure on the Cobb Sea- mount may be responsible for the absence of older widow rockfish, there are arguments against this ex- planation. First, a sample of 50 harlequin rockfish (S. variegatus) had a mean age of 15 years, much older than would be expected if they had experienced simi- lar fishing mortality to the widow rockfish. In addi- tion, even if the Cobb Seamount had experienced ex- tremely heavy fishing pressure, it seems unlikely that virtually all the older fish would have been caught. 50 o c CD 3 cr CD k_ LL +-* C CD 2 cd a. 30 Cobb Swunount n-724 N««rahorB n-824 F^ 12 13 Age (years) Figure 1 Comparison of age composition from widow rockfish caught on Cobb Seamount and the coastal area of northern Oregon. NOTE Pearson et al.: Cobb Seamount rockfish fishery 575 The argument for differences in fishing behavior has some merit. Fishermen report that the terrain on the Cobb Seamount is very rugged and it is possible that the fishermen tend to keep their nets higher off the bottom to avoid snags than they would in the nearshore areas. This could result in a different age composition if young fish tended to disperse higher into the water column than older fish. The argument for the difference in age composition being due to a difference in discard practices also has some merit. Fishermen operating under trip limits in the nearshore areas, and having to deal with mar- ket demands for larger fish, would tend to discard smaller fish. The problem with this argument is that it would tend to explain a higher proportion of small fish in the Cobb Seamount widow rockfish landings, but it would not explain the virtual absence of older, larger fish which certainly would be retained by the fishermen. The possibility that the age composition of the two areas is actually different also has merit. It is not known whether the population of widow rockfish on the Cobb Seamount is self sustaining or if the fish recruit to the seamount from the nearshore popula- tion. Juvenile rockfish have an extended pelagic phase. During this stage they can be advected offshore. Re- cent studies of juvenile rockfish off the central Califor- nia coast have found juvenile rockfish more than 200 miles offshore1. If this also occurs off Oregon, then it is possible that some fraction of the existing population on the Cobb Seamount originated as juveniles or lar- vae, or both, in the nearshore area. While it cannot be completely discounted that adults migrate to the sea- mount it seems unlikely because adult widow rockfish are not considered pelagic and the seamount is more than 200 miles from the nearest suitable habitat. It is possible that older fish either emigrate from the sea- mount or experience high mortality for some unknown reason. It is also possible that there was a massive kill of widow rockfish on the Cobb Seamount prior to 1985. Another possibility is that there never were widow rockfish on the Cobb Seamount prior to 1985. This latter possibility may have some merit since yellowtail rockfish appear to be totally absent from the Cobb Seamount; yet, they are quite abundant in the nearshore area, although the habitat at the Cobb Sea- mount would seem to be ideal for them. If getting to the Cobb Seamount is a fortuitous event, then it is possible that yellowtail rockfish just have not been lucky yet and that widow rockfish have only been lucky recently. Only further studies can determine the cause for the apparent differences in age composition between the Cobb Seamount and the nearshore populations. Otoliths from many species can undergo a process of vaterite replacement ( Gauldie, 1986 ) in which the arago- nite in the otoliths is replaced by vaterite. This process results in a quite distinctive otolith morphology (Fig. 2) and has been observed by the authors in many species of rockfish, flatfish, and other groundfish species. Vateritic otoliths are sometimes called "resorbed" or "crystallized" owing to their somewhat crystalline appearance. In nearshore widow rockfish sampled in Oregon from 1991, only 2.5% of the fish had vateritic otoliths. In contrast, 28 percent of otoliths from Cobb Seamount widow rock- fish are vateritic. Gauldie (1986), working with salmon, suggested that vaterite replacement is under single lo- cus genetic control but can be overridden by temperature extremes. It is possible that Cobb Seamount could be subject to temperature extremes during certain oceano- graphic events like El Ninos. Three other interesting observations have been made about rockfish from Cobb Seamount during this fishery. Rosy rockfish (S. rosaceous) was caught; thus a north- ern range extension for this species was created.- Har- lequin rockfish have been caught; thus, their southern range has been extended from its previously reported southern limit of Queen Charlotte Sound, British Co- lumbia (Eschmeyer et al., 1983). Five shortbelly rock- fish (S. jordani) were caught, three of them in excess of 34-cm fork length, which exceeds the largest fish previously known to the authors (33 cm). Studies on other seamounts in the Pacific have shown that rare species, exceptionally large specimens, and range ex- tensions occur on other seamounts (Hughes, 1981). The Cobb Seamount undoubtedly holds many other surprises for fisheries biologists. Studies of rockfish communities on the Cobb Seamount and how coloniza- tion occurs could lead to a better understanding of the population dynamics of this large, valuable group of groundfish. One possible way colonization could occur would be if it were found that strong cohorts on the Cobb Seamount were weak cohorts in the nearshore area and vice versa. This would suggest the possibility that advection of juveniles has an important role in determination of cohort strength which would have major ramifications in management of nearshore fisheries. Continued examination of fish with vateritic otoliths could yield new information, particularly if other populations of fish are found to have a high per- centage of vateritic otoliths. Remotely operated vehicle studies would be a useful approach to examine the benthic community to determine its similarity to nearshore areas. ■Stephen Ralston, Nat. Mar. Fish. Serv., Tiburon, CA 94920. Pers. Commun., March 1993. -R. Lea. Marine biologist. Calif. Dep. of Fish and Game, Monterey, CA 93940. Pers. commun. June 1993. 576 Fishery Bulletin 91(3). 1993 Literature cited Anonymous. 1950. North Pacific exploratory fishery program. Com- mercial Fish. Rev., 12:10:32-33. Birkland, C. 1971. Biological observations on Cobb Seamount. Northwest Sci. 43:3:193-199. Budinger, T. F. 1967. Cobb Seamount. Deep Sea Res. 14:191-201. Chilton, D. E., and R. J. Beamish. 1982. Age determination methods for fishes studied by the groundfish program at the Pacific Biological Station. Can. FTEC Spec. Publ. Fish. Aquat. Sci. 60, 102 p. Dower, J., H. Freeland, and K. Juniper. 1992. A strong biological response to oceanic flow past Cobb Seamount. Deep-Sea Res. 39:7/8:1139-1145. Eschmeyer, W. N., E. S. Herald, and H. Hammann. 1983. A field guide to the Pacific coast fishes of North America. Houghton Mifflin Co., Boston, MA, 336 p. Gauldie, R. W. 1986. Vaterite otoliths from Chi- nook salmon {Onchorhynchus tshawytscha). New Zea. J. Mar. Freshwater Res. 20:2:209-219. Hughes, S. E. 1981. Initial U. S. exploration of nine Gulf of Alaska seamounts and their associated fish and shellfish resources. Mar. Fish. Rev. 43(1 ):26-33. Pearson, D. E., and J. E. Hightower. 1991. Spatial and temporal var- iability in growth of widow rockfish (Sebastes entomelas). U.S. Dep. Commer., NOAA Tech. Memo., NOAA-TM-NMFS- SWFSC-167, 43 p. Sasaki, T. 1986. Development and present status of Japanese trawl fisheries in the vicinity of seamounts. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 43:21-33. Figure 2 Two sagittal otoliths from a widow rockfish. Otolith on the left is normal, while the otolith on the right exhibits the phenomenon known as vaterite replacement. Effects of body size on probability of predation for juvenile summer and winter flounder based on laboratory experiments David A. Witting Kenneth W. Able Rutgers University, Institute of Marine and Coastal Sciences, Marine Field Station, Great Bay Blvd, PO. Box 278, Tuckerton, Ml. 08087 Predation by brown shrimp [Crang- on crangon ) has been hypothesized to impose significant mortality on settling juvenile plaice (Pleuro- nectes platessa) in European wa- ters (Edwards and Steele, 1968; Lockwood, 1980; Veer and Berg- man, 1987; Pihl, 1990). Laboratory experiments found that predation rate was dependent upon prey den- sity and that predation, estimated by gut content analysis, could be the source of previously unex- plained density-dependent mortal- ity of early juvenile plaice in the Wadden Sea (Veer, 1986; Veer and Bergman, 1987). An important finding of this work is that preda- tor and prey body size may have profound effects upon rate of pre- dation and that recently settled plaice (under 30mm) are much more vulnerable to predation than are larger fish. We were prompted to investigate the role of the sevenspine bay shrimp (Crangon septemspinosa) as a predator on metamorphosing and small juvenile summer {Paralichthys dentatus) and winter (Pleuronectes ameri- canus) flounder because 1) all of these species are abundant compo- nents of New Jersey estuaries, 2) they all co-occur temporally, 3) the sevenspine bay shrimp is morpho- logically similar to the European brown shrimp, and 4) fish scales were present in the guts of this shrimp (Wilcox and Jeffries, 1974). We conducted laboratory experi- ments to determine the size range over which juveniles of these fish species may be susceptible to pre- dation by sevenspine bay shrimp. Methods We used small cylindrical glass bowls, 197 mm diameter, containing 2cm of washed sand (sufficient for both predators and prey to burrow in), filled with seawater to a depth of 50 mm. We conducted all experi- ments at prevailing ambient water temperatures (Table 1) by immers- ing the bowls in a flow-through sea- water bath. We collected transforming and juvenile winter flounder in May- July, 1990 and summer flounder in January-February 1990, with nightlights, plankton nets, and seines. Adult sevenspine bay shrimp were collected (January- July) by using a seine. All collec- tions were made in the Great Bay-Little Egg Harbor estuarine system (New Jersey). In the labo- ratory, we fed experimental ani- mals in excess (shrimp-chopped fish and shrimp; flounder-live brine shrimp nauplii), and then Contribution No. 93-33 of the Institute of Marine and Coastal Sciences, Rutgers University. starved them for 24 hours before the start of each experiment. After anesthetization (25mg-l_1 MS-222) we measured body sizes (Table 1) of flounder (in mm, stan- dard length [SL]) and shrimp (in mm, total length [TLJ, from the tip of the antennal scale to the end of the uropod, Price, 1962). At the start of each trial, we placed one flounder into each con- tainer and allowed it to acclimate for 24 hours. We then introduced one shrimp into each container at approximately 1500 hours. All con- tainers were covered with perfo- rated clear plastic sheets and were left undisturbed for 18 hours. We ran the experiments under a natu- ral light cycle (11 hours dark, 13 hours light, lights out at 1800) us- ing fluorescent lighting. Because sevenspine bay shrimp are unlikely to forage during the day (Haefner, 1979), we introduced the shrimp three hours before darkness in or- der to reduce the likelihood of pre- dation immediately following their introduction. We recorded water temperature at the start and end of each trial. For each experiment, we set up several control containers (prey present, predator absent) and observed no flounder mortality in these controls (rc-15 for summer flounder, n = 12 for winter flounder). At the termination of each trial we scored predation as plus or minus based on the presence or absence of a live flounder. We repeated trials for winter and summer flounder and varied shrimp and flounder size (Table 1). Because settlement in summer flounder oc- curs earlier than winter flounder, experiments involving summer flounder were conducted during late winter and winter flounder experi- ments were conducted in the spring, resulting in different ambient wa- ter temperatures (Table 1). In tri- als involving winter flounder, we Manuscript accepted 19 March 1993. Fishery Bulletin: 91:577-581 (19931. 577 578 Fishery Bulletin 91(3), 1993 Table 1 Sample sizes and body sizes of sevenspine bay shrimp and flounder prey for laboratory experi- ments. Sample size refers to the number of individual predator-prey trials. Pearson correlation coefficients (r) are given for correlations of predator size with prey size to indicate random allocation of predator and prey sizes. Prey Size range (C.V.) Temperature Sample (°C) size Predator (TL) Prey(SL) Pleuronectes americanus 18 60 51-59mm (4.2%) 9-46mm (37.4% ) 0.15 0.24 Paralichthys 9-12 135 27-65 mm (17.5%) ll-16mm (7.1%) 0.03 0.70 dentatus used large shrimp to minimize the predator size effect, thereby concentrating on prey size to determine if there was a size refuge for winter flounder. In trials involv- ing summer flounder, we used a large range of preda- tor sizes, but a narrow range of flounder sizes (those in the last stage of eye migration) because of the greater availability of the latter. This combination of predator and prey sizes was used to establish the relationship between shrimp size and their ability to successfully prey upon small summer flounder. In all trials, we randomly allocated predators and prey to the bowls to avoid deliberate or inadvertent size biases. We used logistic regression analysis (SAS/STAT Users guide, Release 6.03 Edition 1988), which uses maximum likelihood analysis of the natural logarithm of the ratio of these response frequencies (logits) to estimate parameters of a linear model. Because the response is a frequency response rather than a con- tinuous response, a chi-square value is calculated to test for the significance of the treatment variables. Once parameter estimates of the linear model have been made, expected logits can be generated. Expected probabilities of mortality can then be calculated using the following relationship: l+eL where m is the probability of mortality from preda- tion, e is the root of natural logarithm and L is the logit predicted by the linear model. This relationship is obtained by solving the following simultaneous equa- tions for m: m+s=l and " (f)-L where s is the probability of survival. Results Sevenspine bay shrimp consumed both flounder spe- cies under laboratory conditions, and size effects were important in all interactions. In summer flounder trials, both prey and predator size significantly af- fected the outcome of prey-predator interactions (predator size x2=24.07, P<0.0001, prey size x2=7.75, P<0.01). Summer flounder matched with larger shrimp experienced generally greater mortality than those matched with smaller shrimp (Fig. 1A). Logis- tic regression of these data produced a positive rela- tionship between probability of predation and shrimp size (Fig. IB) which was stronger when the effect of flounder size was controlled using linear regression and the residuals were plotted against predator size (Fig 1C). Smaller summer flounder experienced higher mortality (Fig. 2A); however, probabilities from logistic regression show no clear pattern (Fig. 2B) because of the confounding effect of predator size. When we controlled the predator-size effect, a clear negative relationship was revealed between summer flounder size and probability of predation (Fig. 2C). For winter flounder, the effect of flounder size on the probability of predation was significant (\2=4.03, P<0.05), but no significant predator size effect occurred (X2=0.04, P>0.8) presumably because we deliberately selected large shrimp in these trials in order to mini- mize this effect. Only the smallest individuals (<17 mm SL) were preyed upon (Fig. 3A). Logistic regression analysis demonstrated that the highest probability of predation was on the smallest individuals (>60% for 9mm SL). This probability declined to zero at flounder lengths of approximately >17mm SL (Fig. 3B). Param- eter estimates of the prey size effect for both summer and winter flounder were similar (-0.4 ± 0.2 for winter and -0.6 ± 0.2 for summer), implying a similar size relationship for both species. NOTE Witting and Able Probability of predation for juvenile Paralichthys dentatus and Pleuronectes amencanus 579 M to 1 £ 15 A D Mortality ■ No Mortality 12 - 9 6 3 fi- i n 1 HI i n 1.00- 1 0.80- B • • •• * * s • • 0.60 ■ • • • • * * .*.•-: 0.40- « • * •! • •* * 0.20 • • 1 0.00 1.00 i C •I * • 0.80- ••• 0.60 ••• 0.40- ...1 0.20- •o.oo — i 1 i 25 35 45 55 65 Total Length of Predator (mm) Figure 1 Relationship between size of sevenspine bay shrimp preda- tors and predation on summer flounder prey. Graph presents original data with sample sizes above each bar (A), raw prob- abilities as predicted by logistic regression (B), and probabili- ties after being standardized for the effect of prey size (C). These probabilities and those for Figure 2 are based on the equation: Logit = 1.66+10.15 * predator sizeM0.59 * prey size). See text for discussion of logits. 80 A □ Mortality NoMortality 60- j§ 40- E =3 Z 20 _ 0 _m ■ 1 1 _ 1.00 i i 1 0.80- B • <••••• 0.60- • » <• • 0.40- • .- 0.20 5 • * ■§ 1 "8 o.oo £ 1.00 1 1 o .^ C ■ % 0.80- i-. S 0.60- 0.40- • • • 0.20 • • '0.00 9 11 13 15 17 Standard Length of Prey (mm) Figure 2 Relationship between prey size and probability of predation on summer flounder by sevenspine bay shrimp. Graph pre- sents original data with sample sizes above each bar I A), raw probabilities as predicted by logistic regression (B), and prob- abilities after being standardized for the effect of predator size(C). Discussion Summer and winter flounder that survive the egg and larval stage and settle to estuarine substrates inhab- ited by adult sevenspine bay shrimp may be subject to significant predation. Summer flounder of 11-16 mm SL, the size at which they enter the estuary (Szedlmayer et al. 1992, Keefe and Able, 1993), are vulnerable to predation by a large size range of sevenspine bay shrimp. This interaction is likely be- cause these species co-occur under natural conditions. Previous studies indicate that abundance of adult sevenspine bay shrimp, in estuaries to the south and north of the study area (Modlin 1980, Price 1962), be- gin to increase in the fall and continue to do so until they reach a peak in the spring. This temporal pattern overlaps completely with the period (October-May) that summer flounder enter the estuaries of New Jersey (Able et al. 1990, Szedlmayer et al. 1992). These shrimp range from 10-50 mm TL (Price, 1962). Our data sug- 580 Fishery Bulletin 91(3), 1993 E E 0 1.00 S 0.80 CO I £ 0.60 Hi i IIHvi Iff III £ 0.40 0.20 0.00 □ Mortality ■ No Mortality Nlin ~i 1 r oo — oo - OO OOO 14 21 28 35 42 Standard Length of Prey (mm) 49 Figure 3 Relationship between size of winter flounder and predation by sevenspine bay shrimp. Figure presents original data with sample sizes above each bar (Al and predictions made by logistic regression (B). Bars on prob- abilities reflect one standard error of the estimate. These probabilities are based on the equation: Logit = 4.39-10.42 * prey size I. gest that small summer flounder are vulnerable to pre- dation over a large fraction of this range, with a 50% chance of mortality when encountering a shrimp of 45 mm TL (Fig. 1). Winter flounder also settle (April-May) in north- eastern U.S. estuaries (Pearcy, 1962) including those in New Jersey (pers. observ. ) when adult sevenspine bay shrimp are abundant. Winter flounder settle at a smaller size than summer flounder (7.8 mm SL for win- ter flounder, Chambers and Leggett, 1987, vs. 13 mm SL for summer flounder, Keefe and Able, In press); however, our data indicate that the relationship be- tween flounder size and vulnerability to predation is similar for the two species. This suggests that winter flounder must approximately double their size (i.e., reach approximately 17mm) before they achieve a size refuge from predation by large sevenspine bay shrimp (Fig. 3). Both winter and summer flounder ap- pear to exhibit a pronounced decrease in vul- nerability to predation between 9 and 20 mm SL. In summary, the vulnerabilities of both spe- cies of flounder were significantly affected by small differences in prey size. The duration of time spent within this size range (i.e., growth rate) can be quite variable depending upon habitat for winter flounder (Sogard, 1992) and temperature effects for summer flounder ( Keefe and Able, 1993). Thus slight variation in size at settlement, or growth after settlement, may have important effects upon survival for both species. This scenario suggests that variabil- ity in stage duration, not rate of mortality, may be a critical determining factor of year- class strength as has been suggested by a num- ber of authors (Sissenwine, 1984; Chambers and Leggett, 1987; Houde, 1987; Bailey and Houde, 1989). Acknowledgments We are grateful to Matthew Pearson, Lynn Wulff, and Roger Hoden for assistance in con- ducting experiments, and R. Christopher Chambers who provided a constructive review of an earlier draft. This project was funded through a fellowship grant from the Electric Power Research Institute (EPRI); support was also provided from a Leatham grant (Rutgers University, Biological Sciences), Manasquan Marlin and Tuna Club Scholarship Fund, The New Jersey Sea Grant College Minigrant Pro- gram, and the Institute of Marine and Coastal Sciences C4 M 49 1.1 216 Nov 9 16 34 0.6 0.03 3 1 V2B-2L tag. Results Approximately 120 hours of daylight and 170 hours of nighttime lingcod tracking were conducted between 7 October and 23 December 1991 (Table 1). Seven ex- perimental fish were released at Yeo Island (fish Yl- Y7) and six experimental fish at Amelia Island (fish A1-A6). There were an additional four control fish (fish C1-C4). Based on size, most tagged lingcod were esti- mated to be sexually immature and 2-3 years of age. Lingcod from this area mature over a size range of 50- 62 cm for males and 60-67 cm for females (Richards et al., 1990). The largest male and female lingcod used in the study had lengths of 58 and 62 cm, respectively. Seven of the 13 experimental lingcod left the study area at least 7 days prior to their transmitter expiry dates (Table 1). We termed these lingcod as "transients" and the remaining experimental lingcod as "residents." Resident and control lingcod could be detected in the study area until their transmitter expiry dates. With the exception of fish Y2, transient lingcod were last observed 2-7 days after release. By contrast, most resi- dent lingcod could be located in the study area after 28 days. Control fish were recorded after even longer periods. For example, fish CI (V3-5HI tag) was ob- served after 50 days. Similarly, fish C3 (V2B-2L tag) was observed after 40 days. All the control fish were located in shallow water (depth of approximately 20 m) and maintained relatively constant positions over the course of the study. The amount of time spent in the holding tank did not appear to affect the tendency of experimental ling- cod to leave or remain in the study area. Lingcod in transient and resident groups did not differ signifi- cantly in the number of days held before release ( Wilcoxon rank sum test, P>0.10). However, fish length (P<0.05) and weight (P<0.01) did differ significantly between transient and resident groups of lingcod (Wilcoxon rank sum test). Although sample size was too small to consider sex or transmitter size effects, both sexes and transmitter sizes were represented in each experimental group. In general, individual tracks were similar for ling- cod in the transient group and in the resident group. For example, Fish Y2 (transient) was released at the south side of Yeo Island, moved counter clockwise around Yeo Island, travelled southwest to Vancouver Island, and then returned to the release site in 5 days (Fig. 1). Fish Y2 remained within 0.5 km of the release site over the next 2 days where it was observed by SCUBA divers. For the next 10 days, fish Y2 could not be detected in the study area. It was then found south of Yeo Island, moved 7 km west, and left the study area 19 days after release. Fish A2 (resident) was re- NOTE Yamanaka and Richards Movements of transplanted Ophiodon elongatus 585 leased at the north side of Amelia Island, moved in a westerly direction for 5 days, turned east, and then remained within a 0.3 km2 area for 17 days, after which the tag battery expired. All four control fish remained within 0.4 km- of their capture-release site over the entire tracking period. Transient lingcod were apparently more active than resident lingcod, and control lingcod moved little. An analysis of variance on the ranked movement rate data for the experimental lingcod indicated that rates dif- fered significantly among fish within a group (P<0.001) and, furthermore, that rates differed significantly be- tween fish in transient and resident groups (P<0.001). Similarly, movement rates differed significantly be- tween experimental and control groups of lingcod (P<0.001). Mean (±SE) movement rates were 4.9±0.5m/ min, 1.8±0.2m/min, and 0.3±0.2 m/min, respectively, for transient, resident, and control groups of lingcod. Move- ment rate was also related to fish size. For the experi- mental lingcod, rank correlation coefficients between mean movement rate and fish length (0.69) and weight (0.60) were significant (P<0.05). Thus, fish size may determine movement patterns, which in turn deter- mine the tendency of experimental lingcod to leave or remain in the study area. We calculated the total (horizontal) distance trav- elled by each lingcod (Table 1) as the sum of the dis- tances between sequential observations of fish posi- tion. This measurement obviously depends on the number of positions recorded for a fish during each tracking period. Furthermore, the number of positions recorded was highly variable, because of continuous observations for some fish. We computed daily move- ment rates (km/d) to ensure meaningful com- parisons among fish. These rates were calcu- lated from distances between estimated positions at 24-hour intervals for fish with three or more sequential daily position obser- vations. As demonstrated for the raw move- ment data, daily movement rates were signifi- cantly greater for transient fish than for resident fish (Wilcoxon rank sum test, P<0.05), and for resident fish than for control fish (Wilcoxon rank sum test, P<0.001). Mean (±SE) daily movement rates were 1.42+0.24 km/d U = 19), 0.80±0.13km/d (rc=51), and 0.05±0.01km/d (n=21) for transient, resident, and control groups of lingcod, respectively. Continuous observations for four experimen- tal fish were sufficiently long to examine diel activity patterns. We determined the distance travelled between estimated positions at se- quential 2-hour periods (Fig. 2). By late-Octo- ber, the study area was experiencing only 8 hours of daylight. Lingcod travelled the greatest dis- tances at night (Fig. 2). Movement rates for these fish averaged 2.1±0.3m/min (n=58> during the night (1600- 0800 hours) and 0.7±0.1 m/min (ra=26) during the day (0800-1600 hours). Discussion Movements of lingcod transplanted 250 km from their normal habitat were related to fish size. The larger fish tended to disperse from the study site within a few days of release. These results corroborate the findings of Buckley et al. (1984); when lingcod are transplanted over large distances, the smaller fish are more likely to remain in the new area. The tendency for larger experimental lingcod to leave the study area may be related to the onset of sexual maturity. Lingcod spawn in the Strait of Georgia be- tween December and March (Low and Beamish, 1978). Lingcod selected for the transplant were all smaller than the size at 50% maturity. However, sexual matu- rity in lingcod can occur over a wide size range (Richards et al., 1990). For example, all experimental male lingcod and one transient female lingcod were larger than the size at 20% maturity. Thus, some of the recorded movements could be associated with se- lection of spawning (nest guarding) sites. Ultrasonic tags provide short-term information on fish behaviour. Experimental lingcod could be followed for at most 28 days (Table 1) and we can not discount the possibility that resident lingcod left the study area after the end of the experiment. Furthermore, tran- ~i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — 8 12 16 20 24 4 8 12 16 20 24 4 8 12 16 20 Hour of day Figure 2 Distances travelled over 2-hour intervals by four lingcod observed con- tinuously for 48-60 hours. Distances are shown corresponding to the interval midpoint. The shaded rectangles indicate periods of darkness. 586 Fishery Bulletin 91(3), 1993 sient lingcod may have become resident just outside the boundaries of the study area or returned at a later date. Fish Y2, for example, was apparently absent from the study area for 10 days before returning (Fig. 1). Conventional tagging studies are required to address these long-term issues. However, conventional tagging studies suffer from other problems. For example, the lack of fishing effort in the area due to seasonal clo- sures would reduce the probability of tag recapture, and a minimum size limit (65cm) would impose a size- selectivity on the recaptured fish. Our measurements of lingcod movement rate were comparable to those reported by Matthews ( 1992). She observed movement rates of 1.17 km/d for displaced male lingcod, intermediate to the rates of 1.42 km/d and 0.80 km/d that we computed for lingcod in tran- sient and resident groups, respectively. Furthermore, our estimate of nocturnal ( 1600-0800 h) movement rate was 2.1m/min, identical to the rate that Matthews (1992) measured between 2400-0600 hours. The con- trol lingcod in our study remained within the maxi- mum home range size of 0.4 km2 recorded by Matthews (1992). It was not possible for us to determine whether ling- cod moved toward their capture site after leaving the study area. Homing has been documented for adult lingcod (Hart, 1943; Cass et al., 1983; Matthews, 1992), but other factors undoubtedly affect fish movement. For example, the prevailing current in the Strait of Georgia flows to the northwest owing to land masses and prevailing winds from the southeast (Thomson, 1981). Hart (1943) observed from tag returns that most lingcod released in the Strait of Georgia moved in a northwesterly direction. The purpose of this study was to determine whether stocking with juvenile lingcod could promote natural stock rebuilding. Because 6 of the 13 displaced fish did remain within the 12 X 4 km study area while their tags were active (28 days), restocking with ju- venile lingcod is a potential enhancement tool. For transplanted fish to contribute to the spawning stock, they must remain in the new area and survive there until reproduction (spawning/nest guarding) is com- plete. Annual natural mortality rates for lingcod are 24<7r (Schnute et al., 1989). Mortality rates due to sport fishing and marine mammal predation may be higher (Collicutt and Shardlow, 1989; Olesiuk et al., 1990; Smith et al., 1990). To compensate for these mortalities, the number of age 2-3 year lingcod trans- planted must be at least two to three times the num- ber required to achieve a target spawning biomass one year later. A harvest of this magnitude could be detrimental to the donor stock. Until the depleated stock is re-established, the enhanced area should be closed to lingcod retention, and if possible, to all fishing activities. Hook and release mortality is not known but could decrease survivorship if fishing pres- sure is intense. Acknowledgments The long hours of tracking endured by Claudia Hand, Greg Workman, and Anita Gurak are gratefully ac- knowledged. We also thank Antan Phillips and Doug Miller for sharing their boat-handling expertise. K. Groot, J. Rice, and K. R. Matthews provided helpful comments on the manuscript. Literature cited Buckley, R., G. Hueckel, B. Benson, S. Quinnell, and M. Canfield. 1984. Enhancement research on lingcod Ophiodon elongatus in Puget Sound. Wash. Dept. Fish. Prog. Rep. 216, 93 p. Cass, A. J., M. S. Smith, I. Barber, and K. Rinhofer. 1983. A summary oflingcod tagging studies conducted in 1978 by the Pacific Biological Station. Can. Data Rep. Fish. Aquat. Sci. 417, 283 p. Clark, D. S., and J. M. Green. 1990. Activity and movement patterns of juvenile At- lantic cod, Gadus morhua, in Conception Bay, New- foundland, as determined by sonic telemetry. Can. J. Zool. 68:1434-1442. Collicutt L. D., and T. F. Shardlow. 1989. Strait of Georgia sport fishery creel survey sta- tistics for salmon and groundfish, 1989. Can. MS Rep. Fish. Aquat. Sci. 2087, 75 p. Hart, J. L. 1943. Migration oflingcod. Fish. Res. Board Can. Pac. Prog. Rep. 57:3-7. Holland, K. N., R. W. Brill, S. Ferguson.'R. Chang, and R. Yost. 1985. A small vessel technique for tracking pelagic fish. Mar. Fish. Rev. 47:26-32. Holland, K. N., R. W. Brill, and R. K. C. Chang. 1990. Horizontal and vertical movements of yellowfin and bigeye tuna associated with fish aggregating devices. Fish. Bull. 88:493-507. Ketchen, K. S., N. Bourne, and T. H. Butler. 1983. History and present status of fisheries for ma- rine fishes and invertebrates in the Strait of Georgia, British Columbia. Can. J. Fish. Aquat. Sci. 40:1095- 1119. Low, C. J., and R. J. Beamish. 1978. A study of the nesting behavior of lingcod (Ophiodon elongatus) in the Strait of Georgia, British Columbia. Fish. Mar. Serv. Tech. Rep. 843, 27 p. NOTE Yamanaka and Richards: Movements of transplanted Ophiodon elongatus 587 Matthews, K. R. 1992. A telemetric study of the home ranges and hom- ing routes of lingcod Ophiodon elongatus on shallow rocky reefs off Vancouver Island, British Columbia. Fish. Bull. 90:784-790. Matthews, K. R., T. P. Quinn, and B. S. Miller. 1990. Use of ultrasonic transmitters to track demersal rockfish movements on shallow rocky reefs. Am. Fish. Soc. Symp. 7:375-379. Olesiuk, P. F., M. A. Bigg, G. M. Ellis, S. J. Crockford and R. Wigen. 1990. An assessment of the feeding habits of harbour seals (Phoca vitulina) in the Strait of Georgia, Brit- ish Columbia, based on scat analysis. Can. Tech. Rep. Fish. Aquat. Sci. 1730, 135 p. Quinn, T. P., B. A. Terhart, and C. Groot. 1989. Migratory orientation and vertical movements of homing adult sockeye salmon, Oncorhynchus nerka, in coastal waters. Anim. Behav. 37:587-599. Richards, L. J., J. T. Schnute, and C. M. Hand. 1990. A multivariate maturity model with a compara- tive analysis of three lingcod {Ophiodon elongatus) stocks. Can. J. Fish. Aquat. Sci. 47:948-959. Richards, L. J., and K. L. Yamanaka. 1992. Lingcod. In B. M. Leaman (ed), Groundfish stock assessments for the west coast of Canada in 1991 and recommended yield options for 1992, p 23- 56. Can. Tech. Rep. Fish. Aquat. Sci. 1866, 304 p. Ruggerone, G. T., T. P. Quinn, I. A. McGregor, and T. D. Wilkinson. 1990. Horizontal and vertical movements of adult steel- head trout, Oncorhynchus mykiss, in the Dean and Fisher channels, British Columbia. Can. J. Fish. Aquat. Sci. 47:1963-1969. Schnute, J. T., L. J. Richards, and A. J. Cass. 1989. Fish survival and recruitment: investigations based on a size-structured model. Can. J. Fish. Aquat. Sci. 46:743-769. Smith, B. D., G. A. McFarlane, and A. J. Cass. 1990. Movements and mortality of tagged male and female lingcod in the Strait of Georgia, British Columbia. Trans. Am. Fish. Soc. 119:813-824. Thomson, R. E. 1981. Oceanography of the British Columbia coast. Can. Spec. Publ. Fish. Aquat. Sci. 56, 291 p. Superintendent of Documents Subscriptions Order Form Order Processing Code: I — I YES, enter my subscription(s) as follows Charge your order. It's Easy! S ■." 2i :'C'.-: 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 $. change. (Company or Personal Name) (Additional address/attention line) (Street address) (City. State, ZIP Code) Prices include regular domestic postage and handling and are subject to For privacy protection, check the box below: ] Do not make my name available to other mailers Please choose method of payment: I I Check Payable to the Superintendent of Documents ] GPO Deposit Account (Please type or print) (Daytime phone including area code) (Purchase Order No.) -□ | VISA or MasterCard Account 1 1 Thank you for (Credit card expiration dale) your order! (Authorizing Signature) Mail To: Superintendent of Documents P.O. Box 37 1954. 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 used 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 17 (for decapod crustaceans). Text footnotes should be numbered in Arabic numerals and typed on a sep- arate sheet from the text. Footnotes are not used for reference material or per- sonal communications, but rather to explain or define terms in the text and for contribution numbers on the title page. Informal sources Personal communi- cations, unpublished data, and untitled manuscripts in preparation are noted parenthetically in the text (full name, affiliation, brief address including zip code, and month and year when appro- priate). Acknowledgments Gather all ac- knowledgments into a brief statement at the end of the text. Give credit only for exceptional contributions and not to those whose contributions are part of their normal duties. Citations All titled sources should be listed in the Citations section, including unpublished and processed material. In text, cite as Smith and Jones (1977) or (Smith and Jones 1977); if more than one citation, list chronologically (Smith 1936, Jones 1975, Doe In press). All sources cited in the text should be listed alphabetically by the senior authors' surnames under the heading CITA- TIONS. Abbreviations of periodicals and serials should conform to Serial Sources for the BIOSIS Data Base™. Indicate whether sources are in a langu- age other than English. For informal literature, include address of author or publisher. Authors are responsible for the accuracy of all citations. Tables should supplement, not dupli- cate, the text. Each table should be numbered and cited consecutively, with headings short but amply descriptive so that the reader need not refer to the text. For values less than one, zeros should precede all decimal points. In- dicate units of measure in column head- ings; do not deviate from the unit of measure within a column. Table foot- notes should be noted consecutively in Roman letters across the page from left to right and then down. Since all tables are typeset, they need not be submitted camera-ready. Figures Photographs and line draw- ings should be of professional quality- clear and concise— and reducible to 42 picas for full-page width or to 20 picas for a single-column width, and to a max- imum 55 picas high. All graphic ele- ments in illustrations must be propor- tioned to insure legibility when reduced to fit the page format. Line weight and lettering should be sharp and even. Let- tering should be upper and lower case, and vertical lettering should be avoided whenever possible (except for vertical, y, axis). Zeros should precede all deci- mal points for values less than one. Re- productions of line artwork are accepted in the form of high-quality photographic prints from negatives or photomechan- ical transfer (PMT). Halftones should be sharply focused with good contrast. Micron rules should be inserted on elec- tron micrographs, even when magnifi- cation is included in the figure legend. There should be clear distinction be- tween identifying letters (press-on or overlay) and background of photograph. Label each figure in pencil on the back. Send only xerox copies of figures to the Scientific Editor; originals or photo- graphic prints will be requested later when the manuscript is accepted for publication. Copyright Government publications are in the public domain, i.e., they are not protected by copyright. Submission of manuscript Disks Authors are encouraged to re- tain manuscripts on word-processing storage media (diskettes, floppy disks) and submit a double-spaced hardcopy run from the storage media. Submit disks as MS-DOS "print" or "non- document" files (often called "ASCII files"). If a disk cannot be converted to an ASCII file, the author should indicate on the disk the source computer and software language along with the file name. Either 5V4-inch or 3V2-inch disks from IBM-compatible or Apple/Macin- tosh systems (non-graphics only) can be submitted, double-sided/double-density or high-density, limiting each file to 300 kilobytes. All 8-inch word-processing disks (e.g., Wang or NBI) must be con- verted onto 5V4- or 3V2-inch MS-DOS print disks. Send original hardcopy and two dupli- cated copies to: Dr. 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. stamvm^s^am^aB^JB&mvsBm? U.S. Department of Commerce Volume 91 Number 4 November 1993 Fishery Bulletin Marine Biological Laboratory/ Woods Hole Oceanographie Institution Library DEC 17 1993 Woods Hole, MA 02543 U.S. Department of Commerce Ronald H. Secretary Brown National Oceanic and Atmospheric Administration D. James Baker Under Secretary for Oceans and Atmosphere National Marine Fisheries Service Rolland A. Schmitten Assistant Administrator for Fisheries _l (_ Scientific Editor Dr. Ronald W. Hardy Northwest Fisheries Science Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle. Washington 981 12-2097 Editorial Committee Dr. Andrew E. Dizon National Marine Fisheries Service Dr. Linda L. Jones National Marine Fisheries Service Dr. Richard D. Methot National Marine Fisheries Service Dr. Theodore W. Pietsch University of Washington Dr. Joseph E. Powers National Marine Fisheries Service Dr. Tim D. Smith National Marine Fisheries Service The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115-0070. Second class postage is paid in Seattle, Wash., and additional offices. POSTMASTER send address changes for subscriptions to Fishery Bulletin, Super- intendent of Documents, Attn: Chief, Mail List Branch, Mail Stop SSOM, Washington, DC 20402-9373. Although the contents have not been copyrighted and may be reprinted entire- ly, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this period- ical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $24.00 domestic and $30.00 foreign. Cost per single issue: $12.00 domestic and $15.00 foreign. See back page for order form. Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C 1 5700 Seattle, Washington 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. 1 103. Begin- ning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle. Washington Volume 91 Number 4 November 1993 Fishery Bulletin M/n^T? ftn0l08ica' Moratory/ Wood* Holt Ocianooraphic Institution Library DEC 1 7 1993 Woods Hole. MA 0254a Contents 587 Alados, Consuela L., Juan Escos, and John M. Emlen Developmental instability as an indicator of environmental stress in the Pacific hake [Merluccius productus) 594 Chang, Sukwoo Analysis of fishery resources, potential risk from sewage sludge dumping at the deepwater dumpsite off New Jersey 61 1 Chivers, Susan J., and Al C. Myrick Jr. Comparison of age at sexual maturity and other reproductive parameters for two stocks of spotted dolphin, Stenella attenuate 619 Chow, Seinen, M. Elizabeth Clarke, and Patrick J. Walsh PCR-RFLP analysis on thirteen western Atlantic snappers (subfamily Lutjaninae) a simple method for species and stock identification 628 Edwards, Elizabeth F, and Christina Perrin Effects of dolphin group type, percent coverage, and fleet size on estimates of annual dolphin mortality derived from 1987 U.S. tuna- vessel observer data 641 Gilpatrick Jr, James W. Method and precision in estimation of dolphin school size with vertical aerial photography 649 Labelle, Marc, John Hampton, Kevin Bailey, Talbot Murray, David A. Fournier, and John R. Sibert Determination of age and growth of South Pacific albacore ( Thunnus alalunga) using three methodologies 664 Lovrich, Gustavo A., and Julio H. Vinuesa Reproductive biology of the false southern king crab [Paralomis granulosa, Lithodidae) in the Beagle Channel, Argentina 676 Sampson, David B. The assumption of constant selectivity and the stock assessment for widow rockfish, Sebastes entomelas Fishery Bulletin 91(4), 1993 690 Scoles, Daniel R., and John E. Graves Genetic analysis of the population structure ofyellowfin tuna, Thunnus albacares, from the Pacific Ocean 699 Sedberry, George R., and Nicole Cuellar Planktonic and benthic feeding by the reef-associated vermilion snapper, Rhomboplites aurorubens (Teleostei, Lutjanidae) 710 Temte, Jonathan L. Latitudinal variation in the birth timing of captive California sea lions and other captive North Pacific pinnipeds 718 Thompson, Grant G. Variations on a simple dynamic pool model 732 Toole, Christopher L., Douglas F. Markle, and Phillip M. Harris Relationships between otolith microstructure, microchemistry, and early life history events in Dover sole, Microstomus pacificus 754 Van Waerebeek, Koen. Geographic variation and sexual dimorphism in the skull of the dusky dolphin, Lagenorhynchus obscurus (Gray, 1 828) 775 Wade, Paul R. Estimation of historical population size of the eastern spinner dolphin (Stenella longirostns orientalist 788 Weisberg, Stephen B., and William H. Burton Spring distribution and abundance of ichthyoplankton in the tidal Delaware River Notes 798 Brill, Richard W, and David B. Holts Effects of entanglement and escape from high-seas driftnets on rates of natural mortality of North Pacific albacore, Thunnus alalunga 804 Hannah, Robert W, and Neil T Richmond Weight change of pink shrimp, Panda/us jordani, after commercial harvest and handling 808 Loza no-Alvarez, Enrique, Patricia Briones-Fourzan, and Fernando Negrete-Soto Occurrence and seasonal variations of spiny lobsters, Panulirus argus (Latreille), on the shelf outside Bahia de la Ascension, Mexico 816 Witherell, David B., and Jay Burnett Growth and maturation of winter flounder, Pleuronectes americanus. in Massachusetts AbStraCt.-Developmental Insta- bility (DI) has been proposed as an inexpensive, quickly applied, and sensitive indicator of stress that can be utilized in early warning and in monitoring anthropogenic impacts on fish and other animals and plants. A problem arises, however, to the extent that natural stressors confound the effects of human-in- duced disturbances. Our objective in this work was to investigate whether a natural stressor, in the form of El Nino conditions, contributed to DI in the Pacific hake. Right-left (fluctuating) asymmetry of otolith length, width, growth rate, and weight, as well as right-left otolith shape differences, were used as mea- sures of DI. Results show that in- deed El Nino disrupts development, indicating stress. This outcome sug- gests that DI, as an early warning and monitoring tool for stress, must be used with caution. Developmental instability as an indicator of environmental stress in the Pacific hake (Merluccius productus) Consuela L. Alados Juan Escos Estacion Experimental de Zonas Andas Consejo Superior de Investigaciones Cientificas Almena. Spain Present Address: National Fisheries Research Laboratory U S Fish and Wildlife Service Naval Station Puget Sound. Seattle, WA 98 1 I 5 John M. Emlen National Fisheries Research Laboratory U S Fish and Wildlife Service Naval Station Puget Sound, Seattle, WA 98 1 I 5 Manuscript accepted 25 June 1993. Fishery Bulletin 91:587-593 ( 1993). The viability of fish populations and the ecological communities of which they form a part is a matter of con- cern to fisheries managers. At pres- ent, awareness of ecological problems often occurs only when stocks have already begun to decline, when there is clear evidence of disease or mor- bidity, or when die-offs occur. Much work has been done on meth- ods to assess adverse environmental impacts on fish via physiological stress measures (Adams, 1990a). These methods can, in theory, be used to monitor problems and, possibly, act as early warning indicators. However, most of these indicators either mea- sure short-term acute stress response (e.g., corticosteroid levels), or are ex- pensive and laboratory-intensive. They are cost-effective only in evalu- ating already obvious signs that a problem exists. In economically de- pressed areas, including the devel- oping nations, these methods are im- practical. In addition, they generally lack ecological relevance (Adams, 1990b). A possible alternative to these methods lies in morphological mea- sures of developmental instability by which changes (induced by environ- mental or genetical stress) in the ba- sic developmental strategy of organ- isms can be assessed. For example, chronic stress might aggravate right- left asymmetry in normally bilater- ally symmetric structures (Soule, 1967; Valentine and Soule, 1973; Val- entine et al, 1973) or increase the number of aberrations in circulus patterns on scales (Shackell and Doyle, 1990). Morphological measures are inexpensive to obtain and cause minimal or no damage to animals in- volved. They can be quickly and in- expensively applied not only in situ- ations where concern has already arisen, but also as standard moni- toring tools. Screening of fry might be valuable in assessing effectiveness of various hatchery practices or in evaluating smolt quality in Pacific salmon. The efficacy of right-left, or "fluctu- ating" asymmetry (FA) as an indica- tor of both genetic and environmental stress has been well documented. Within populations, homozygosity is associated with increased FA 587 588 Fishery Bulletin 91(4). 1993 (Dobzhansky and Wallace, 1953; Lerner, 1954; Lewontin, 1956; Bader, 1965; Bruckner, 1976; Soule, 1979; Vrijenhoek and Lerman, 1982; Leary et al, 1983, 1984; Mitton and Grant, 1984). Increased homozygos- ity is one consequence of inbreeding (one form of ge- nomic disruption). Within populations of outcrossing species, inbreeding is associated with lowered fitness (Lewontin, 1956; Charlesworth and Charlesworth, 1987). Bader (1965), Bailit et al. (1970), and Clarke ( 1992 ) have all found positive correlations among popu- lations between inbreeding and FA. The converse of inbreeding depression is outbreed- ing depression. Hybridization and subsequent intro- gression of sufficiently different populations can lead to the disruption of coadapted gene complexes and thus, in theory, promote both decreased fitness and develop- mental instability (Vrijenhoek and Lerman, 1982). Zakharov and Bakulina (1984) found that when indi- viduals of similar populations ofDrosophila virilis were crossed, no increases in FA over the parental types were observed. When more divergent strains were crossbred, FA was pronounced in the offspring. Out- breeding depression should be most apparent in inter- specific hybrids. Leary et al. (1985) have found FA to be higher in laboratory hybrids of rainbow (On- corhynchus my kiss) and cutthroat (0. elarki) trout than in either parental type. Environmental disturbance, or deviation from the conditions to which organisms are adapted, leads to lowered fitness and developmental instability. Tempera- ture stress causes increased FA in laboratory mice and rats (Beardmore, 1960; Siegel and Doyle, 1975, a and b; Siegel et al., 1977), in lizards (genus Lacerta, Zakharov 1982), and in chum salmon (O. keta, Beacham 1990). Audiogenic stress has the same effect on labora- tory rats (Siegel and Smookler, 1973; Siegel and Doyle, 1975c). Valentine and Soule (1973) demonstrated that FA of laboratory grunion {Leuresthes tenuis) popula- tions rose with concentrations of DDT. Clarke (1992) showed increased FA in pesticide-treated bush flies iMusca vetustissima) even at concentrations too low to produce statistically detectable changes in mortality. Zakharov and Rubin (1985) demonstrated the effects of contamination on FA in animals in the Baltic Sea. Thus FA may be an extremely sensitive indicator of stress. Chronic stress indicators may be sensitive to natu- rally occurring events as well as anthropogenic distur- bances. Zakharov et al. ( 1991 ), for example, found that intensified population density feedback led to increased FA during the decline phase of shrew population cycles. Data exist (Emlen, unpubl.) suggesting increased asym- metry occurs in the canine teeth of population-stressed northern fur seals. Clearly, if natural stressors are important contributors to developmental instability. then their effects must be considered when using DI as a tool for identifying or monitoring anthropogenic disturbances. In this paper we provide an example of FA responses to another natural stressor, the El Nino event of 1982-1983. The Pacific hake (Merluceius produetus) ranges along the Pacific coast of North America from Mexico to Alaska. Several genetic stocks can be distinguished. These include two spawning in Puget Sound, one from the Georges Straits, several from the fjords of Van- couver Island, and a so-called coastal population that ranges from San Francisco Bay to southern Baja Cali- fornia. Reproduction occurs from January to April, and young fish enter the fishery at about age three. In- dividual growth occurs largely between May and September. Environmental conditions associated with the 1982- 1983 El Nino may have caused significant dislocations of the coastal hake stock on the summer fishing grounds, since the entire population apparently moved northward (Francis and Hollowed, 1985). Length and weight information from the data set used by Francis and Hollowed and gathered by the National Marine Fisheries Service showed no significant differences be- tween the 1980 and 1984 year classes for either sex. However, the 1980 El Nino year-class fish were slightly longer and lower in weight, and a Mann-Whitney test showed them to have significantly lower condition fac- tor (weight/length1) than the 1984 control year-class (respective means=0. 06709, 0.07208; rc=85, 63; P=0.0001). Both the dislocation in position and the diminished condition factor of population members sug- gest a drop in population viability and, thereby, impli- cate stress. Might this stress also have provoked de- velopmental instability? Methods The development of otoliths, small calcareous struc- tures used in maintaining balance in some fishes, is known to reflect growth rate and transition points in life histories (Wilson and Larkin, 1982; Volk et al., 1984; Alhossaini and Pitcher, 1988; Sogard, 1991). As such, they may be useful indicators of developmental instability. We do not suggest that otoliths fit the cri- teria given above for cost effectiveness and harmless- ness to the organisms monitored. We use them here simply as a convenience (they were readily available and came with appropriate, attendant data on fish age, length, weight, and date of collection). The National Marine Fisheries Service, at its Alaska center in Se- attle, maintains a collection of otoliths taken by scien- tific observers aboard commercial hake fishery boats. From this collection we obtained paired right and left Alados et al : Developmental instability in Merlucaus productus 589 otoliths for 4-year-old hake caught in the first half of September for the years 1984 and 1988 (from the 1980 and 1984 year classes). These fish were the same from which the length and weight data, mentioned above, were taken. Individuals from the 1980 year class were in their second and third year of growth during El Nino; those from the 1984 year class were not present during the El Nino event. Time-sequential data is use- ful in the present context since the population was used as its own control, reducing or minimizing con- founding factors of geography, genetic stock, etc. Five otolith characters were used to assess instabil- ity: fluctuating asymmetry in weight, length, maxi- mum width, growth rate, and right-left differences in shape (Fig. 1). For comparison of shape, one otolith was aligned with the reverse image of the other, and right-left shape differences examined. At least two means of alignment exist. Since otoliths show annual growth rings, it was possible to superimpose long axes drawn through the first growth ring. In the 1980 year class this ring was laid down just prior to the El Nino event and indicates the beginning of the period of stress. Measured differences along the outlines of the third ring, marking the end of El Nino, provide an appropriate measure of right-left differences in growth during the period. This approach was impractical. Growth Shape = I,| d, |/length Figure 1 Hake (Merluccius productusi otoliths: (A) An otolith showing growth lines and the measure used to assess growth between ages 1 and 4; (B) Diagram of superimposed right and left otoliths showing how shape difference was characterized. While distances between the rings could be reasonably approximated with the aid of an optical micrometer, our equipment was inadequate to permit video or photo reproduction of clear ring patterns. Accurate axes could not be drawn. Minute angular changes in axis defini- tion generated large differences in the shape measure. Because of these difficulties, we used an alternative approach. Equally magnified images of the two oto- liths were projected onto paper and their outlines traced. One tracing was then turned over and placed atop the other. Each otolith possesses clearly marked "top" and "bottom" extreme points (Fig. 1), so align- ment could be accomplished by superimposing the axes drawn through these two points. This procedure led to an almost perfect fit of the two otoliths along their straighter side and to clear pattern differences along their opposite scalloped side (Fig. 1). Nine equally spaced lines were drawn perpendicular to this axis, and, along each such line, the absolute distance be- tween the scalloped margins of the two overlain out- lines was measured. Right-left differences among the otoliths were taken as the sum of these absolute dis- tances, normalized by dividing by the axis length. In- formation lost by neglecting similar differences along the straight side was negligible owing to the almost perfect fits along that margin. Growth rate was mea- sured by the maximum distance, parallel to the straight side, between the first and the fourth otolith growth ring at the broader end of the otolith (Fig. 1). Four-year-old fish caught in September, near the end of the growth season, may have displayed some vari- ance in size owing to differences in date of hatching. This variance is not likely to be large because little growth occurs between earliest and latest hatching in January and April. Nevertheless, to correct for pos- sible associations between asymmetry and early growth, we used a normalized index, I L-R I AL+R), with L and R designating left and right measures. Because collections for each of the three years were made at very nearly the same time, biases arising from pos- sible influences of growth stage on asymmetry were minimized. The measure of right-left difference in shape is not a measure of FA in the strict sense (Palmer and Strobeck, 1986). However, an increase in the measure would indicate an increased deviation from normal de- velopmental homeostasis. Thus it is a valid indicator of developmental instability. Tests for fluctuating asymmetry depend on an ab- sence of directional asymmetry (skew), anitsymmetry (bimodality or platykurtosis) and, according to one school of thought (Palmer and Strobeck, 1986), a nor- mal distribution for (L-R)/(L+R). The Shapiro- Wilk sta- tistic (Zar, 1984, p. 95) was used to test for normality. Because of the controversy surrounding the validity of 590 Fishery Bulletin 9! [4), 1993 FA tests for traits not normally distributed, further analyses were canned out twice, once with and once without such traits. A second order ANOVA was applied to test for influ- ences of sex and year class on growth rate. Finally, fish were catergorized as fast or slow growers, according to whether growth rate fell above or below the median for the year in question, and a three-way MANOVA was performed to test for effects of growth rate, sex, and year on FA and right-left otolith shape differences. Results For both the 1980 and 1984 year classes, we found that the distributions of weight, width, and length did not differ significantly from normality (Table 1). These variables Table 1 Tests for normality of (L-R)/(L+R) for various traits used in developmental instability. (L in- dicates measure on the left, R the correspond- ing measure on the right otolith ). Width of otolith in = 147) Mean = 0.002 Skew = 0.184 Kurtosis = 1.333 Shapiro-Wilk W = 0.990 P = 0.9456 Weight of otolith in = 135) Mean = 0.003 Skew = 0.242 Kurtosis = 0.683 Shapiro-Wilk W = 0.994 P = 0.9965 Length of otolith (n = 135) Mean = 0.001 Skew = 0.049 Kurtosis = 2.030 Shapiro-Wilk W = 0.975 P = 0.1981 Growth rate of otolith ( 1980 year class; n = 79) Mean = -0.001 Skew = -1.567 Kurtosis = 6.124 Shapiro-Wilk W = 0.981 P = 0.0001 Growth rate of otolith ( 1984 year class; n = 69) Mean = 0.002 Skew = -0.281 Kurtosis = 0.248 Shapiro-Wilk W = 0.981 P = 0.7044 can, therefore, be used in analysis of FA. Growth rate was not normally distributed. The two-way ANOVA indicated no influ- ence of sex (7^=1.42; df=l, 141; P=0.234) and no significant inter- action effect of sex X year (F=0.36; df=l,141; P=0.550) on growth. The main effect of year, however, was strong (F=14.20; df=l,141; P=0.0002). This last result, along with the possibility that growth rate might influence asymmetry of other characters, provided the rationale for incorporating growth rate into the three-way MANOVA, as noted above. Results were qualitatively the same — no differences with respect to which comparison deviated signifi- cantly— whether growth rate was or was not included in the MANOVA, and so only one result set, that including this vari- able, is presented in this paper (Table 2). There was a highly significant difference in developmental instability, but only be- tween the years. Univariate analysis showed that FA was greater for the El Nino fish with respect to otolith weight and shape, but not width or length (Tables 3 and 4). To examine the possibility that the between-year differences were not due to El Nino effects, we examined otoliths of 4-year- old fish from another year class (1977) unaffected by El Nino events. As with the above two groups, fish sampling took place during the first two weeks of September. These otoliths were analyzed for right-left asymmetry in weight and shape. Results of a Multiple Analysis of Variance for all three year classes, with fixed effects of year and sex, are given in Tables 5 and 6. The results, in conjunction with Tukey tests for differences among years (Table 7) show clear increases in right-left differences in both shape and weight for the 1980 El Nino year over 1984 and, in shape, over 1977 as well. Differences between the two years not affected by El Nino (1977 and 1984) were not statistically significant. Discussion The observation that only two of the five right-left asymmetry measures responded to stress at a statistically detectable level raises the question of consistency. If developmental instability is to be a useful indicator of stress, we need to know what measures will be relatively sensitive to stress and which will not. Highly Table 2 Multiple analysis of variance for various traits jsed in devel opmental m- stability analysis. Fixed effects are year, sex and growth rate. Variables are right-left differences in width, weight, length growth rate and shape. Effects Wilks Lambda F df P Year 0.8029 5.596 5. 114 0.0001 Sex 0.9509 1.176 5, 114 0.3253 Growth 0.9532 1.119 5, 114 0.3546 Year x sex 0.9407 1.438 5. 114 0.2159 Year < growth 0.9230 1.901 5, 114 0.0995 Sex ■ growth 0.9612 0.961 5. 114 0.4704 Year X sex X grow th 0.9936 0.146 5, 114 0.9807 Alados et al Developmental instability In Merluccius productus 591 Table 3 Univariate breakdown of multivariate ana ysis of variance for various traits used in developmental in- stability analysis. Fixed effect df SS F P Variable Weight Growth rate 0.00001 0.26 0.608 Sex 0.00009 2.28 0.133 Year 0.00036 9.07 0.003 Growth x sex 0.00010 2.63 0.107 Growth x year 0.00001 0.35 0.554 Sex x year 0.00008 2.09 0.151 3-Wav 0.00002 0.49 0.487 Error 118 0.00469 Width Growth rate 0.00000 0.00 0.951 Sex 0.00003 0.88 0.350 Year 0.00008 2.08 0.152 Growth x sex 0.00001 0.32 0.574 Growth x year 0.00030 2.54 0.114 Sex x year 0.00004 1.15 0.286 3-Way 0.00000 0.01 0.930 Error 118 0.00456 Growth rate Growth rate 0.00134 4.11 0.045 Sex 0.00058 1.77 0.186 Year 0.00073 2.24 0.137 Growth X sex 0.00000 0.00 0.998 Growth X year 0.00068 2.10 0.150 Sex x year 0.00067 2.06 0.154 3-Wav 0.00007 0.22 0.639 Error 118 0.03837 Shape Growth rate 0.00065 0.76 0.385 Sex 0.00015 0.17 0.679 Year 0.01856 21.79 0.000 Growth X sex 0 00220 2.58 0.111 Growth x year 0.00076 0.89 0.347 Sex x year 0.00292 3.43 0.066 3-Way 0.00006 0.07 0.753 Error 118 0.10049 Length Growth rate 0.00002 0.60 0.440 Sex 0.00001 0.31 0.581 Year 0.00000 0.00 0.986 Growth x sex 0.00002 0.62 0.433 Growth x year 0.00009 2.66 0.106 Sex x year 0.00001 0.39 0.534 3-Way 0.00000 0.01 0.933 Error 118 0.00409 Table 4 Means of developmental instability indices of Pacific hake (Merluccius productus) otoliths for 1980 and 1984 year classes. 1980 (El Nino Groupl 1984 (Non El Nino Group) Variable Males in) Females in) Males (ra) Females (n) Weight 0.0083 (361 0.0117 (36) 0.0065 (24) 0.0065 (30) Shape 0.0831 (36) 0.0950 (36) 0.0679 (24) 0.0603 (30) Table 5 Multiple analysis of variance over one El Nino year 1 1980) and two Non-El Nino years (1977, 1984). Fixed effects are year and sex. Variables are weight and shape. Effects Wilks Lambda F df P Year Sex Year X sex 0.7606 11.876 4,324 0.000 0.9930 0.574 2, 162 0.564 0.9577 1.770 4,324 0.135 Table 6 Univariate breakdown of multiple anal vsis of variance for two traits used in developmental instability analysis over three years ( 1977, 1980. 19841. Sum of squares df F P Variable = weight Sex 0.00002 1 0.63 0.430 Year 0.00038 2 5.01 0.008 Sex x year 0.00017 2 2.18 0.116 Error 0.00621 163 Variable = shape Sex 0.00062 1 0.82 0.366 Year 0.03576 2 23.65 0.000 Sex x year 0.00285 2 1.89 0.155 Error 0.12322 163 Table 7 Tukey test for differences among years (1977, 1980, 1984) in two measures of developmental instability. El Nino year class is 1980; Non El Nino year classes are 1977, 1984. NS = not significantly different. Difference Year class between means Sample sizes) P Shape 1980-1984 0.0250 (77.591 <0.05 1980-1977 0.0337 (77,25) <0.05 1977-1984 -0.0087 (25. 59) NS Weight 1980-1984 0.0033 (79,55) <0.05 1980-1977 0.0024 (79,38) NS 1977-1980 0.0009 (38,551 NS 592 Fishery Bulletin 91(4). 1993 Means of asymmetry (Merluccius productusl year. Table 8 in growth rate for Pacific hake in an El Nino and a non-El Nino Year El Nino Non El Nino (1980) (1984) Growth rate Slow Fast 0.0142 0.0132 0.0251 0.0157 canalized traits are unlikely to be sensitive indica- tors of stress. High phenotypic variance can gen- erally be considered an indication of low canaliza- tion, and Zakharov ( 1989) has shown that, at least for Lacerta lizards, FA rises with phenotypic vari- ability (see also Soule, 1967). A related observation is that traits directly and strongly affecting fitness, i.e., under intense stabi- lizing selection and, therefore, with low heritabil- ity, should be poor candidates for detecting stress (Soule and Cuzin-Roudy, 1982). Similarly, struc- tures whose development is affected by use may not be appropriate for FA analysis. Locomotion, for example, might discourage asymmetric growth of limb size. Long-bone length in laboratory rats shows no FA response to heat, cold, behavioral, or audiogenic stress, while long-bone density, not a size characteristic, does (Doyle et al., 1977). In choosing measures of developmental instabil- ity, it may be useful to consider structural detail. With respect to otoliths, weight, length, width, and growth rate represent growth along a single di- mension. These metrics, then, are a simple sum- mation of one or more growth processes. Shape, on the other hand, reflects growth processes whose parts are separably measurable. Thus, shape does not compound and thereby obscure information. We might, accordingly, expect to find differences in the shapes of corresponding right and left struc- tures to be more sensitive indicators of stress than simple metric differences, a prediction consistent with the hake otolith results presented here. Do natural environmental stressors, as well as man-caused disturbances, result in developmental instability? The results reported here suggest an affirmative answer, namely that while DI might be usefully applied in management, it must be used with caution. Use of fin asymmetries or scale circulus aberrations in comparing hatchery stocks for smolt quality, for example, should involve only a single year class and should be given decreasing consideration as stocks diverge genetically or geo- graphically. Suspected problems arising from water con- tamination in Puget Sound might be investigated or moni- tored by examining FA, but only so long as the control population comes also from Puget Sound. Literature cited Adams, S. M. (ed.). 1990a. American Fisheries Symposium 8. Am. Fish. Soc, Bethesda, MA, 191 p. 1990b. Status and use of biological indicators for evaluat- ing the effects of stress on fish. In S. M. Adams (ed.) American Fisheries Symposium 8, p. 1-8. Am. Fish. Soc, Bethesda, MA, 191 p. Alhossaini, M., and T. J. Pitcher. 1988. The relation between daily rings, body growth and environmental factors in plaice, Pleuronectes platessa L., juvenile otoliths. J. Fish Biol. 33:409-418. Bader, R. S. 1965. Fluctuating asymmetry in the dentition of the house mouse. Growth 29:291-300. Bailit, H. L., P. L. Workman, J. D. Niswander, and C. J. MacLean. 1970. Dental asymmetry as an indicator of genetic and environmental conditions in human populations. Hum. Biol. 42:626-638. Beacham, T. D. 1990. A genetic analysis of meristic and morphometric varia- tion in chum salmon iOncorhynchua keta) at three differ- ent temperatures. Can. J. Fish. Aquat. Sci. 68:225-229. Beardmore, J. A. 1960. Developmental stability in constant and fluctuating temperatures. Heredity 14:411-422. Bruckner, D. 1976. The influence of genetic variability on wing symme- try in honeybees (Apis mellifera ). Evolution 30:100-108. Charlesworth, D., and B. Charlesworth. 1987. Inbreeding depression and its evolutionary conse- quences. Ann. Rev. Ecol. System. 18:237-268. Clarke, G. M. 1992. Fluctuating asymmetry: a technique for measuring developmental stress of genetic and environmental ori- gin. Acta Zool. Fenn. 191:31-36. Dobzhansky, T., and B. Wallace. 1953. The genetics of homeostasis in Drosophila. Proc. Nat. Acad. Sci. U.S.A. 39:162-171. Doyle, W. J., C. Kelley, M. I. Siegel. 1977. The effects of audiogenic stress on the growth of long bones in the laboratory rat {Rattus norvegicus). Growth 41:183-189. Francis, R. C, and A. Hollowed. 1984. Status of the Pacific hake resource and recommen- dations for management in 1985. Appendix 2 in Status of Pacific coast groundfish fishery and recommendations for management in 1985. Pacific Fishery Management Council, Portland, Oregon. Leary, R. F., F. W. Allendorph, and K. L. Knudsen. 1983. Developmental stability and enzyme heterozygosity in rainbow trout. Nature 301:71-72. Alados et al.: Developmental instability in Merluccius produaus 593 1984. Superior developmental stability of heteroygotes at enzyme loci in salmonid fishes. Am. Nat. 124:540- 551. 1985. Developmental instability and high meristic counts in interspecific hybrids of salmonid fishes. Evolution 39:1318-1326. Lerner, I. M. 1954. Genetic homeostasis. Oliver and Boyd, London, 134 p. Lewontin, R. C. 1956. Studies on homeostasis and heterozygosity, I. General considerations. Abdominal bristle number in second chromosome homozygotes of Drosophila melanogaster. Am. Nat. 90:237-255. Mitton, J. B., and M. C. Grant. 1984. Associations among protein heterozygosity, growth rate, and developmental homeostasis. Ann. Rev. Ecol. Syst. 15:479^199. Palmer, A. R., and C. Strobeck. 1986. Fluctuating asymmetry: measurement, analysis, patterns. Ann. Rev. Ecol. Syst. 17:391^121. Siegel, M. I., and W. J. Doyle. 1975a. The effects of cold stress on fluctuating asym- metry in the dentition of the mouse. J. Exper. Zool. 193:385-389. 1975b. Stress and fluctuating limb asymmetry in vari- ous species of rodents. Growth 39:363-369. 1975c. The differential effects of prenatal and post- natal audiogenic stress on fluctuating asymmetry. J. Exper. Zool. 191:211-214. Siegel, M. I, and H. H. Smookler. 1973. Fluctuating dental asymmetry and audiogenic stress. Growth 37:35-39. Siegel, M. I., W. J. Doyle, and C. Kelley. 1977. Heat stress, fluctuating asymmetry, and prenatal selection in the laboratory rat. Am. J. Phys. Anthro. 46:121-126. Shackell, N. L., and R. W. Doyle. 1990. Scale morphology as an index of develop- mental stability and stress resistance of Tilapia (Oreochromis niloticus). Can. J. Fish. Aquat. Sci. 48:1662-1660. Sogard, S. M. 1991. Interpretation of otolith microstructure in juve- nile winter flounder (Psuedopleuronectes americanus): ontogenetic development, daily increment validation, and somatic growth relationships. Can. J. Fish. Aquat. Sci. 48:1862-1871. Soule, M. 1967. Phenetics of natural populations, II. Asymmetry and evolution in a lizard. Am. Nat. 101:141-160. 1979. Heterozygosity and developmental stability: an- other look. Evolution 33:396-401. Soule, M. E., and J. Cuzin-Roudy. 1982. Allomeric variation. 2: Developmental instabil- ity of extreme phenotypes. Am. Nat. 20:765-786. Valentine, D. W., and M. E. Soule. 1973. Effect of p,p'-DDT on developmental stability of pectoral fin rays in the grunion, Leuresthes tenuis. Fish. Bull. 71:921-926. Valentine, D. W., M. E. Soule, and P. Samollow. 1973. Asymmetry analysis in fishes: a possible statis- tical indicator of environmental stress. Fish. Bull. 71:357-370. Volk, E. C, R. C. Wissmar, C. A. Simenstad, and D. M. Eggers. 1984. Relationship between otolith microstructure and the growth rate of juvenile chum salmon (Oncorhyncus keta ) under different prey rations. Can. J. Fish. Aquat. Sci. 41:126-133. Vrijenhoek, R. C, and S. Lerman. 1982. Heterozygosity and developmental stability un- der sexual and asexual breeding systems. Evolution 36:768-776. Wilson, K. H., and P. A. Larkin. 1982. Relationship between thickness of daily growth increments in sagittae and change in body weight of sockeye salmon (Oncorhynchus nerka) fry. Can. J. Fish. Aquat. Sci. 39:1335-1339. Zakharov, V. M. 1982. Analysis of developmental stability as a method of determination of the optimal conditions of devel- opment. Doklady Akad. Nauk. SSSR 267:1016-1018. 1989. Future prospects for population phenogenetics. Soviet Scientific Rev., section F, Physiology and Gen- eral Biology Reviews (T. M. Turpaev and A. V. Yablokov, eds.)4(3), 79 p. Zakharov, V. M., and E. D. Bakulina. 1984. Disturbance of developmental stability at cross- ing different strains of Drosophila uirilis (variation in the number of ovarioles taken as an example). Genetica 20:1390-1391. Zakharov, V. M., and G. I. Rubin. 1985. The disturbance of developmental stability as an indicator of anthropogenic influence on animal populations in the Baltic Sea basin. In Symposium on ecological investigations of the Baltic Sea environ- ment: Riga, USSR, 1983, 526-536 p. Valtion Painatuskeskus, Helsinki. Zakhavor, V. M., E. Pankakowski, B. I. Shefter, A. Peltonen, and I. Hauski. 1991. Developmental instability in the common shrew. Am. Nat. 138:797-810. Zar, J. H. 1984. Biostatistical Analysis, 2nd ed. Prentice-Hall, Englewood Cliffs, NJ, 718 p. AbStfclCt.— Analytical and statis- tical procedures were applied to bot- tom trawl survey data in tests of hy- potheses about potential effects of sewage sludge dumping at a 106- mile dumpsite (106-MDS) off New Jersey on fishery resources assessed on the continental shelf and upper slope. Sludge dumping, even in deep ocean waters, was not discounted as one of several ecological and envi- ronmental perturbations influencing these resources measured as tempo- ral, spatial, and seasonal differences in abundance. Species abundances of silver and red hakes (Merluccius bilinearis and Urophycis chuss), summer flounder [Paralichthys dentatus), goosefish iLophius ameri- canus), and black sea bass (Centro- pristis striata) declined significantly over temporal and spatial scales dur- ing the disposal of contaminant- laden sewage sludge at the deep- water 106-MDS. There was also a decline in the array of all aggregated species, but to a lesser degree. Re- sults of these analyses of assessment data are considered in relation to ef- fects of ocean dumping in shallow waters at the southern California sewage outfalls and in the New York Bight apex, and in relation to in- creased contamination of the ecosys- tem around the 106-MDS. Further, large-scale coordination of environ- mental research surveys with fishery resource assessments would allow tests of more specific hypotheses and allow a more definitive interpreta- tion of offshore resource population data as presented here. Analysis of fishery resources: potential risk from sewage sludge clumping at the deepwater dumpsite off New Jersey Sukwoo Chang Sandy Hook Laboratory, Northeast Fisheries Science Center National Marine Fisheries Service, NOAA Highlands, New Jersey 07732 Sewage sludge and a variety of other wastes have been disposed of in coastal waters of the New York Bight for over 50 years (Squires, 1983). Offshore deepwater waste dis- posal of sewage sludge, however, is a relatively new activity of the last 20 years. With legally mandated closure of the coastal water 12- mile dump site ( 12-MDS) in the New York Bight apex (Ocean Dumping Ban Act of 1988), a deepwater dumpsite, 106 miles off New Jersey (106-MDS), was selected as a temporary alternative site for dis- posal of sewage sludge from the New York and New Jersey metro- politan area (O'Connor, 1983; O'Connor et al, 1983, 1985; Pearce et al., 1983; Norton, 1989; NOAA12; Battelle3). The 106-MDS had been used pre- viously for disposal of industrial wastes as well as sewage sludge Manuscript accepted 4 August 1993. Fishery Bulletin 91:594-610 ( 1993). 'NOAA. 1975. May 1974 baseline investiga- tion of deepwater dumpsite 106. NOAA Dumpsite Evaluation Report 75-1. U.S. Dep. Commer., Natl. Oceanic Atmospheric Admin., Natl. Ocean Survey, Rockville, MD, 388 p. -NOAA. 1977. Baseline report of environmen- tal conditions in deepwater dumpsite 106. NOAA Dumpsite Evaluation Report 77-1. U.S. Dep. Commer., NOAA, Natl. Ocean Survev. Rockville, MD. 798 p. 'Battelle. 1990. 106-miles deepwater munici- pal sludge site monitoring, research and sur- veillance plan. Battelle Memorial Institute, Duxbury, MA, 90 p. (O'Connor, 1983; Bisagni4; Ander- son5). Phased relocation of New York- New Jersey sewage disposal began in March 1986, and by December 1987 all sewage sludge disposed at sea was being barged to the 106-MDS site. At the 106-MDS, an average of 6.0 million metric tons (t) wet weight sewage sludge (range 1.2-9.9 million t) was dumped between 1986 and 1992. This mean is similar to an av- erage of 5.8 million t of sewage sludge (range 4.0-8.3 1) dumped at the 12- MDS from 1973 to 1987. New Jersey ceased disposal at 106-MDS in March 1991. New York City phased out 20% of its ocean dumping in December 1991, and the remainder by the end of June 1992. The fishing industry, citizens groups, state governments, and fed- eral agencies all voiced concerns about potential effects of heavy sew- age sludge dumping at the 106-MDS on the surrounding marine environ- ment and fisheries. Dumping of sew- age sludge and industrial waste in 4Bisagni, J. J. 1977. A summary of the input of industrial waste chemicals at deepwater dumpsite 106 during 1974 and 1975. NOAA Dumpsite Evaluation Report 77-1. U.S. Dep. Commer., NOAA, Natl. Ocean Survev, Rock- ville, MD, 487-497 p. Anderson, P. W. 1983. Current status — ocean dumping in the New York Bight. Paper pre- sented on 6 April 1983 in Atlantic City, NJ at the 15th National Conference and Exhibition on Municipal and Industrial Sludge Utiliza- tion and Disposal, 6 p. 594 Chang Effects of sewage sludge dumping on fishery resources 595 the comparatively pristine deep-water ocean environ- ment has been considered to have an adverse effect on fishery resource abundance and composition (Carlise, 1969; Dart and Jenkins, 1981; Mearns, 1981; Russo, 1982; Cross et al., 1985; Spies, 1984; Farrington et al., 1982; Capuzzo and Kester, 1987, a and b; O'Connor et al., 1985, 1987, a and b; Werme67; SCCWRP8; Zdanowicz et al.9; NMFS1" ■"; Studholme et al.12). A con- taminant distribution model presented by O'Connor et al. (1985) (see also Reed et al., 1985) indicated that New York and New Jersey sewage sludge has the im- mediate effect of increasing concentrations of polychlo- rinated biphenyls (PCB's), Zn, Pb, Cr and Cu. In addi- tion, the potential area of influence (PAI) of waste disposal at the 106-MDS extends over a wide region (Fig. 1) off the Middle Atlantic outer continental shelf and upper slope (Bisagni, 1983; O'Connor et al, 1985; Gentile et al, 1989; Warsh13, Bisagni14, Ingham15). A change in fish species population abundance and composition is known to have occurred near shallow- fiWerme, C, R. Shokes. W. Steinhauer. S. McDowell, P. Debrule, P. Hamilton, and P. Boehm. 1988a. Evaluation and recommenda- tions for bioaccumulation studies for the 106-mile deepwater mu- nicipal sludge site monitoring program. Battelle Memorial Institute. Duxbury, MA, 36 p. TWerme, C, K. M. Jop, S. Y. Freitas and P. Boehm. 1988b. Imple- mentation plan for the 106-mile deepwater municipal sludge site monitoring. Battelle Memorial Institute, Duxbury. MA, 51 p. "SCCWRP (Southern California Coastal Water Research Project). 1989. Recovery of Santa Monica Bay after termination of sludge discharge. Southern California Coastal Water Research Project. An- nual Report 1988-1989, 46-53 p. "Zdanowicz, V. S., M. C. Ingham, and S. Leftwich. 1990. Monitoring the effects of sewage sludge disposal at the 106-mile dumpsite using mid-water fish as sentinels of contaminant metal uptake: a feasibil- ity study. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Sen/., North- east Fish. Sci. Cent., Woods Hole, MA. Ref. Doc. 90-02, 6 p. "'NMFS. 1992a. Interim report on monitoring the biological effects of sludge dumping at the 106-mile dumpsite. U.S. Dep. Commer, NOAA, Natl. Mar. Fish. Serv., Northeast Fish. Sci. Cent.. Woods Hole, MA 02543, 128 p "NMFS. 1992b. Second annual report on~monitoring the biological effects of sludge dumping at the 106-mile dumpsite. U.S. Dep. Commer, NOAA, Natl. Mar. Fish. Serv., Northeast Fish. Sci. Cent., Woods Hole, MA 02543, 157 p ,2Studholme, A. L., J. O'Reilly, and M. C. Ingham (eds.l. 1993. Ef- fects of the cessation of dumping at the 12-mile site. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Northeast Fish. Sci. Cent., Sandy Hook Lab., Highlands, NJ 07732. (In review.! "Warsh, C. E. 1975. Physical oceanography hist, rical data for Deepwater Dumpsite 106. NOAA Dumpsite Evaluation Report 75-1. U.S. Dep. Commer., NOAA, Natl. Ocean Survey Rockville, MD, 105- 140 p. l4Bisagni, J. J. 1976. Passage of anticyclonic Gulf Stream eddies through Deepwater Dumpsite 106 during 1974 and 1975. NOAA Dumpsite Evaluation Report 76-1. U.S. Dep. Commer., NOAA, Natl. Ocean Survey, Rockville, MD, 39 p. '''Ingham, M. C. 1977. The general physical oceanography of deepwater dumpsite 106. NOAA Dumpsite Evaluation Report, 77-1. U.S. Dep. Commer., NOAA, Natl. Ocean Survey, Rockville, MD. 29- 54 p. water sewage sludge disposal outfalls in southern Cali- fornia (Mearns 1981; Sherwood1"). For example, a bothid flounder (Pacific sanddab, Cithaichthys sordidus) was replaced by a pleuronetcid (Dover sole, Micro- stomias pacificus). Increased growth rates and high prevalence of fin erosion of Dover sole, however, ap- pear to have resulted from exposure to sediments con- taminated with chemical wastes from the outfalls. Liver anomalies were noted as well and seem to have been associated with exposure to sludge-related con- taminants (Mearns, 1981). Spies (1984) reported that distribution of bottom-feeding fishes (e.g., Dover sole) is probably affected by degraded benthic habitats and abundance of benthic (e.g., polychaete, Capitella capitata) and pelagic prey around these outfalls in the southern California Bight. Some chemical contami- nants associated with the sewage accumulated at higher trophic levels in predatory fishes, e.g., Pacific sanddab and boccaccio, Sebastes paucispinis. Benthic- pelagic coupling was evident in the accumulation of chlorinated hydrocarbons in Dover sole and Pacific sanddab. Cessation of sewage sludge dumping on the biomass of the most frequently occurring fish species around the coastal waters near 12-MDS from 1986 to 1989 was examined by Studholme et al.12. There were no significant changes, although American lobster (Homarus americanus) biomass increased (Pikanowski, 1992; Wilk et al.,17). Possibly this 39-month-long study period was insufficient to detect a recovery of any fin- fish species alterations resulting from decades of sew- age dumping at the old 12-MDS. Interestingly, the high prevalence of fin-rot disease of winter flounder iPleuronectes americanus) around the 12-MDS reported earlier (Ziskowski and Murchelano, 1975; Murchelano and Ziskowski, 1976) declined significantly after ces- sation of the dumping (O'Connor et al.18; Pacheco and Rugg19). It was also shown earlier that Atlantic mack- '"Sherwood, M. J. 1978. The fin erosion syndrome. Southern Califor- nia Coastal Water Research Project. Annual Report 1978, 203-221 p. l7Wilk, S. J., R. A. Pikanowski, A. L. Pacheco, D. G. McMillan, and L. L. Stehlik. 1993. Response of fish and megainvertebrates of the New York Bight apex to the abatement of sewage sludge dumping: an overview. In A. L. Studholme, J. O'Reilly and M. C. Ingham (eds.). Effects of the cessation of dumping at the 12-mile site. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Northeast Fish. Sci. Cent., Woods Hole, MA, 14 p. (In review.) '"O'Connor, J. S., J. J. Ziskowski, and R. A. Murchelano. 1987. Index of pollutant-induced fish and shellfish disease. NOAA Spec. Rep. Ocean Assessment Div. U.S. Dep. Commer, NOAA, Natl. Ocean Survey, Rockville, MD, 29 p. ''Pacheco, A. L., and J. Rugg. 1993. Disease incidence of inner New York Bight winter flounder collected during the 12-mile dumpsite study, 1986-1989. In A. L. Studholme, J. O'Reilly, and M. C. Ingham (eds.). Effects of the cessation of dumping at the 12-mile site. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Northeast Fish. Sci. Cent., Sandy Hook Lab., Highlands, NJ, 14 p. (In review.) 596 Fishery Bulletin 91(4), 1993 □ Figure 1 Location of the 106-mile dumpsite and strata used in Northeast Fisheries Science Cen- ter bottom trawl surveys (1963 to present). The potential area of influence (PAIl by sewage sludge dumping is within the dashed line, which is adapted from Biasagni (1983). erel {Scomber scombrus) embryos developing in sur- face waters about the 12-MDS ( i974-1978) had greater mortality and gross malformation, more mitotic ab- normality and mitotic inhibition than those in less con- taminated bight areas, and this was linked to site con- tamination (Chang and Longwell, 1984; Longwell, 1988; Longwell et al., 1992). Unlike the 12-MDS in the New York Bight apex and the sewage outfalls in southern California, the 106-MDS is situated in a physically dynamic zone charac- terized by periodic shifts and overturns in water masses. The toxic water is rapidly diluted and dispersed. This may diminish the impact of dumped waste and as- sociated chemical contaminants, but it also makes determination of any adverse effects on the re- source more difficult to study than those associated with sew- age sludge dumping in shallow- waters. Still, the economic value of the nearby fisheries warrants some effort at directly measur- ing potential adverse impacts of the 1986-92 sludge dumping on the fishery resource in the vicin- ity of the 106-MDS. The study reported here is an attempt to determine if any change in fishery resource abun- dance on the adjacent continen- tal shelf and slope could be de- tected after sewage sludge disposal commenced at the 106- MDS in March 1986. Although ocean dumping is now banned, this period of sludge dumping in the deep ocean provides an in- teresting case study for consid- eration of effects of ocean pollu- tion in general on economically important fishery resources. This analysis is of potential interest in respect to any future recon- sideration of ocean dumping. There are no prior population level studies on abundance changes in marine fishes associ- ated with the sewage sludge dumping in deep ocean waters. The general null hypothesis that was tested is that no change in resource species abundance and composition was coincident with sew- age dumping at the deepwater 106-MDS during the period 1986-90. Assessments were made of temporal, spatial, and seasonal differences in 11 individual finfish DEPTH ZONES ( M ) 27-51 56-110 111-163 >183 Chang: Effects of sewage sludge clumping on fishery resources 597 species, and of all these species combined. The general hypothesis was subdivided into three more specific hy- potheses, each of which was likewise tested: 1 ) there are no temporal differences in species abundance be- tween the period prior to sewage sludge dumping at the 106-MDS and the period after dumping resumed; 2) there are no spatial differences in species abun- dance between the area north of the 106-MDS and the area south of the dump site, the latter more likely to be influenced by the sewage sludge that supposedly moved through the dump site toward the west-south- west shallow outer shelf (PAI; Fig. 1); and 3) there are no seasonal differences for individual and all species for the north and south areas within the pre- and post-dumping periods. Methods Data from NOAA's Northeast Fisheries Science Center (NEFSC) bottom trawl surveys were used for analysis. From 1963 to the present, NEFSC has conducted bot- tom trawl surveys on the Northwest Atlantic Shelf and upper slope, from the Scotian Shoals to Cape Hatteras including the areas adjacent to the 106-MDS (Fig. 1). The entire survey area has been stratified by depth down to 365 m, and surveys have been based on a stratified random design that provides statistically valid samples for estimating indices of species popula- tion abundance. All fishes and invertebrates are sorted, counted, and weighed by species to the nearest 0.1kg. Large catches are subsampled by weight and volume and expanded to estimate the entire catch. Detailed information pertinent to the rationale, descriptions of trawl gears and sampling schemes, as well as history of the NEFSC bottom trawl survey are provided by Grosslein (1974) and more recently Azarovitz (1981). Grosslein et al. (1979) summarized research, assess- ment, and management of the northwest Atlantic eco- system using data from NEFSC bottom trawl surveys. Data from the spring 1982 through spring 1986 NEFSC surveys were evaluated as the pre-dumping period, and data from autumn 1986 through spring of 1990 as the post-dumping period. Eleven economically important species were selected for analysis: silver hake (Merluecius bilinearis); red hake (Urophycis ehuss); summer flounder (Paralichthys dentatus); goosefish {Lophius americanus); black sea bass (Centropristes stria.tus); scup (Stenotomus chrysops); butterfish (Peprilus triacantus); longfin squid (Loligo pealei); American lobster (Homarus americanus); sea scallop iPlacopecten magellanicus); and spiny dogfish (Squalus acanthias). Species were selected on the basis of their commercial values and landings, and the known avail- ability of data for the entire 106-MDS studv area (Chang, 1990). In addition, all species, a category com- prising the total catch for each cruise, was also used. The multivariate rank sum test (Puri and Sen, 1971) was employed for temporal, spatial, and seasonal dif- ferences in the population abundance analyses by us- ing the catch per unit of effort (CPUE) as a measure of population abundance: 1) the average CPUE values for each bottom trawl survey stratum for all cruises occurring in the pre-dumping period cruises (spring 1982-spring 1986), and for all cruises occurring in the post-dumping period cruises (autumn 1986-spring 1990), and 2) the average CPUE data from the area defined as south of the 106-MDS, including strata 61- 76 and data from the area defined north of the site including strata 1-12 (Fig. 1). Test statistics for CPUE were calculated and compared with critical values from the chi-square table to determine the following: 1. there are no temporal differences in CPUE of pre- and post-dumping periods for spring and autumn cruises collectively in the regions north and south of the 106-MDS (cf. Table 1), 2. there are no spatial differences in CPUE north and south of the 106-MDS for pre- and post-dumping period cruises collectively in spring and autumn sea- sons (cf. Table 2), 3. there are no seasonal differences in CPUE of spring and autumn surveys for pre- and post-dumping pe- riod cruises collectively in the regions of north and south of the 106-MDS (cf. Table 3). Test statistics were also computed and compared with the critical values for testing CPUE for spatial differ- ences for individual species and for all species taken on each cruise in the area north and south of the 106- MDS. Test statistics and abundance indices of indi- vidual species in tabulated form are not included among tables given here. The values of test statistics for temporal, spatial, and seasonal differences are readily comparable (Tables 1-3 ). When examining values of the test statistics for temporal differences, secondary differences for species, regions, and seasons were also assessed. Test statis- tics were similarly examined for spatial and seasonal differences among species, regions, and seasons. Changes in significance or non-significance of the test statistics between pre- and post-dumping period surveys may be interpreted as indicative of temporal differences in CPUE. Significant changes of test sta- tistics between regions north and south of the 106- MDS indicate spatial differences. Changes between spring and autumn surveys indicate seasonal differ- ences. Significant differences in CPUE in time, space, and season may thus be interpreted as some shift in species abundance. Negative differences are taken here as a response of the population to some adverse 598 Fishery Bulletin 91(4). 1993 Table 1 Results of the rank sum test for temporal differences in species catch per unit of effort (CPUE) between pre- and pos -dumping periods using spring and autumn cruises collec- tively in north and south regior s. Species Pre-dumping period vs. post-dumping period Spring Cruises Autumn Cruises CPUE CPUE No. of CPUE CPUE No. of name (weight) (number) species (weight) (number) species North Silver hake 0.098 4.709* 2.064 0.621 Red hake 0.596 0.075 10.087* 3.320 Summer flounder 8.553* 2.349 1.409 1.817 Goosefish 2.290 0.001 15.520* 1.192 Black sea bass 0.423 0.140 0.482 2.964 Scup 0.127 1.035 7.656* 2.910 Butterfish 7.288* 11.810* 1.361 0.161 Longfin squid 0.225 0.039 4.443* 0.001 American lobster 3.170 0.168 1.184 0.639 Sea scallop 3.626 5.285* 1.199 1.736 Dogfish 17.166* 10.993* 0.003 0.094 All species' 10.077* 4.094* 2.503 0.671 0.051 3.278 South Silver hake 1.472 0.040 0.966 6.984* Red hake 6.391* 4.576* 0.707 0.700 Summer flounder 6.645* 10.785* 1.219 1.440 Goosefish 0.934 0.118 3.664 0.033 Black sea bass 5.207* 6.474* 0.386 0.240 Scup 1.184 0.077 0.943 0.087 Butterfish 0.147 0.050 8.443* 12.691* Longfin squid 1.801 1.395 2.548 0.327 American lobster 0.247 2.792 1.713 1.539 Sea scallop 16.057* 12.614* 0.973 1.600 Dogfish 0.242 0.043 0.362 1.297 All species' 14.641* 21.212* 1.834 9.099* 2.377 8.473* ' "All species" includes the eleven listed and all other species. * Significant difference with the critical values at 95% level in the chi-square table. biological or environmental perturbation, to fishing pressure or to some combination of these factors. Posi- tive differences are taken as increased species abun- dance in response to favorable conditions. Nonetheless, significant differences of this sort alone do not establish a direct causal association between change in popula- tion abundance and sewage sludge dumping, or alter- nately between fishing pressure, natural environmen- tal perturbations, or any of combination of these. Any such inference must come from further interpretation of the data in respect to what is known about natural ecology of the subject species, fishing pressure on these, and any known sensitivities to sewage sludge. Abundance indices for the 11 species individually and for all species combined for pre- and post-dumping period surveys (Tables 4-7, pgs. 601-604) were com- puted by using the groundfish survey analysis program -'"Kramer. W. P. 1985. Groundfish survey analysis program (SURVAN version 5.2). program report. Prepared for U.S. Dep. of Commer., NOAA,. Natl. Mar. Fish. Serv.. Northeast Fish. Cent., Input Output Computer Service Inc., Waltham, MA. 137 p. (Kramer2"). Basic assumptions and methodology for es- timation of abundance indices based on NEFSC bot- tom trawl survey data are detailed by Pennington and Brown (1981). Trends in indices and variabilities in time and space may also be interpreted as suggesting some general shift in species abundance attributable to sludge dumping, even though there may be no di- rect proof of an association between population abun- dance change and sludge disposal. Results Temporal differences Test statistics based on species CPUE, and species abundance indices, including estimates for both spring and autumn cruises, are summarized in Tables 1 and 4-7. Combining the information from these two sources reinforces the inferences of significant change in spe- cies abundance over time. Chang Effects of sewage sludge dumping on fishery resources 599 Table 2 Results of the rank sum test for spatial differences in species catch per unit of effort (CPUE) between north and south regions using pre- and post-dumping period cruises collec- tively in spring and autumn seasons. Species North region vs south region Pre-d umping cruises Post-dumping cruises CPUE CPUE No. of CPUE CPUE No. of name (weight) number) species (weight) (number) species Spring Silver hake 13.402* 28.189* 17.765* 35.719* Red hake 3.654 6.591* 15.014* 22.364* Summer flounder 5.833* 5.161* 2.356 0.624 Goosefish 0.228 0.469 0.467 0.023 Black sea bass 0.259 0.727 1.956 2.217 Scup 2.483 1.295 0.511 8.128* Butterfish 12.030* 10.791* 0.496 0.154 Longfin squid 1.201 4.823* 0.004 7.375* American lobster 0.160 1.747 2.235 6.298* Sea scallop 0.001 0.351 5.503* 4.505* Dogfish 16.084* 3.279 0.115 1.561 All species1 1.078 0.943 11.532* 0.113 6.313* 8.808* Autumn Silver hake 40.064* 8.555* 41.868* 24.898* Red hake 33.139* 17.622* 11.277* 7.116* Summer flounder 11.877* 1.987 12.920* 1.263 Goosefish 46.468* 4.564* 24.837* 0.819 Black sea bass 6.133* 2.938 14.058* 0.304 Scup 0.532 2.283 0.018 0.729 Butterfish 5.400* 0.062 16.728* 10.153* Longfin squid 10.717* 10.898* 5.531* 12.555* American lobster 0.178 1.747 0.140 4.371* Sea scallop 0.635 1.157 0.288 0.639 Dogfish 10.896* 3.279 29.296* 18.483* All species1 24.920* 9.559* 5.992* 50.827* 21.397* 43.765* 1 "All species" includes the eleven listed and all other species. *Significant difference with the critical values at 95% level in the chi -square table Analyses of red hake data are used here as an ex- ample of a detailed interpretation of temporal differ- ences in species CPUE. (Black sea bass could be equally well used because of its similar pattern of CPUE indi- ces.! Species test statistics for red hake revealed sig- nificant temporal differences in CPUE indices both by weight (biomass) and number of individuals in the south, but not in the north region (Table 1). Biomass indices declined from pre-dumping spring cruises to post-dumping spring cruises (Table 4). Numerical abun- dance indices also declined from pre-dumping spring cruises to post-dumping spring cruises (Table 5). Red hake autumn data were also analyzed, but the species test statistics were not significant (Table 1). Biomass indices (Table 6) and number indices (Table 7) declined from the pre-dumping to the post-dumping period. These negative temporal differences for red hake are interpreted as an indication of significant reduction in abundance over time. CPUE test statistics for several other species also suggest significant temporal differences between pre- and post-dumping periods. The number of significant differences was greater for spring cruises than for au- tumn cruises (Table 1). Biomass indices were relatively low for both spring and autumn cruises in both north and south regions generally, with the exception of high values for dogfish (Tables 4 and 6). Species number indices were mixed (Tables 5 and 7). Test statistics for summer flounder and sea scallop CPUE by weight and number for the spring cruises were significantly different in the southern region. In the northern region, only summer flounder biomass and scallop number indices were significantly differ- ent. Indices for summer flounder and sea scallop for autumn cruises, however, failed to be significant in either region (Table 1). Summer flounder abundance indices of both biomass and number indicate a decline from the pre-dumping period to the post-dumping pe- riod, while both indices for the sea scallop increased (Tables 5 and 7). This suggests a decline (negative differences) over time for summer flounder, and an enhancement (positive differences) for sea scallop. 600 Fishery Bulletin 91(4), 1993 Table 3 Results of the rank sum test for testing seasonal differences in species catch per unit of effort (CPUE) between spring and Autumn seasons using pre- and post-dumping period cruises collectively in north and south regions. Species Spring season vs autumn season Pre-dumping cruises Post dumping cruises CPUE CPUE No. of CPUE CPUE No. of name (weight) (number) species (weight) (number) species North Silver hake 1.444 1.473 8.382* 2.653 Red hake 8.745* 4.845* 0.057 0.291 Summer flounder 5.024* 0.471 8.835* 1.378 Goosefish 0.098 4.898* 3.337 0.681 Black sea bass 6.519* 2.850 8.597* 0.009 Scup 0.078 2.757 2.498 6.681* Butterfish 1.123 0.945 6.645* 14.920* Longfin squid 9.991* 21.686* 0.262 14.070* American lobster 0.988 0.033 4.137* 0.683 Sea scallop 4.774* 4.137* 1.490 0.806 Dogfish 11.294* 4.527* 20.414* 16.008* All species1 15.317* 41.133* 1.944 57.442* 22.629* 0.265 South Silver hake 35.795* 2.322 27.903* 2.594 Red hake 6.125* 0.291 0.707 0.108 Summer flounder 0.035 0.150 0.002 0.045 Goosefish 28.811* 0.509 16.162* 0.131 Black sea bass 3.476 0.506 10.751* 4.645* Scup 6.210* 3.409 0.019 0.085 Butterfish 4.236* 16.741* 0.292 0.742 Longfin squid 3.229 0.342 0.762 3.974* American lobster 0.095 5.804* 0.013 0.689 Sea scallop 6.425* 4.257* 0.289 0.206 Dogfish 16.927* 8.076* 53.442* 37.763* All species' 51.240* 12.546* 0.919 140.185* 7.022* 24.325* 1 "All species" includes the eleven listed and all other species. * Significant difference with the critical values at 95% level in the chi-square tabk Biomass and number test statistics for butterfish in spring cruises were significantly different between pre- and post-dumping periods in the northern region, but not significantly different in the southern region. Con- versely, those statistics from autumn cruises were not significant in the northern region, but significant in the southern region. Test statistics for spiny dogfish showed similar patterns (Table 1). Species abundance indices for butterfish and spiny dogfish fluctuated, but generated much higher values for the test statistics than did those of other species (Tables 4-7). Accord- ingly, interpretation of the butterfish and spiny dog- fish statistics suggests a positive difference in these two species over time. In spring cruises, test statistics in CPUE both by weight and number for goosefish, scup, longfin squid, American lobster, and silver hake were not significantly different for pre- and post-dumping periods for either the north or south region, with exception of silver hake number indices in the north region. For autumn cruises, however, there were significant differences between pre- and post-dumping periods for goosefish, scup, and longfin squid (Table 1). Abundance indices for goosefish declined from the pre- to the post-dumping period for both cruises and regions. Longfin squid indices declined from the pre to the post period in spring cruises, but increased in autumn cruises. All scup indices declined but increased in autumn cruises within the south re- gion. American lobster indices showed little change between pre- and post-dumping periods. Silver hake indices increased in spring cruises but declined in au- tumn cruises from the pre to post period (Tables 4-7). The lower abundance indices of these species may sig- nify that the populations responded negatively to ad- verse differences. In spring cruises, CPUE statistics for all species, both by biomass and number, were significantly differ- ent between pre- and post-dumping periods in both regions. This occurred though test statistics for some individual species were significantly different and Chang: Effects of sewage sludge clumping on fishery resources 601 Table 4 Species abundance i ndices (e.g.. mean and SSE [ standard error]) based on weight of stratified catch per tow and variabilities (CV in % 1 using pre- and post-dumping spring cruises collec- tively in north and south regions. Species Pre-dumping spring cruises Post-dumping spring cruises Abundance indices Abundance indices name Mean SSE CV(%) Mean SSE CV(%) North Silver hake 1.529 0.62509 41 1.217 0.31258 26 Red hake 1.549 0.78584 51 0.863 0.23815 28 Summer flounder 0.401 0.09030 23 0.211 0.05036 24 Goosefish 1.109 0.28732 26 0.656 0.14389 22 Black sea bass 0.104 0.05443 52 0.025 0.01060 42 Scup 0.510 0.23137 45 0.192 0.09896 52 Butterfish 1.820 1.03770 57 0.383 0.13748 36 Longfin squid 1.009 0.27199 27 2.038 0.55665 27 American lobster 0.440 0.19855 45 0.350 0.09429 27 Sea scallop 0.194 0.09369 48 0.877 0.38851 44 Dogfish 51.186 14.36400 28 115.671 36.00800 31 All spp.' 183.027 23.42700 13 259.398 38.04700 15 South Silver hake 0.426 0.14559 34 0.389 0.10256 26 Red hake 0.671 0.31264 47 0.147 0.06884 45 Summer flounder 0.413 0.10370 25 0.260 0.06533 25 Goosefish 0.938 0.33040 35 0.367 0.17910 49 Black sea bass 0.238 0.08226 20 0.429 0.22193 52 Scup 0.635 0.46357 73 1.293 0.61669 48 Butterfish 0.340 0.15347 45 0.682 0.36380 53 Longfin squid 0.811 0.20697 26 2.382 0.86509 36 American lobster 0.082 0.02381 29 0.151 0.07966 53 Sea scallop 0.092 0.03275 36 1.245 0.73640 59 Dogfish 92.764 36.88800 40 74.456 15.59700 21 All spp.1 226.566 53.61400 24 181.878 18.19100 10 '"All species" includes the eleven listed and all other species. others not. In the autumn cruises, biomass statistics were statistically significant between pre- and post- dumping period cruises, but for the southern region only (Table 1). These results substantiate the negative temporal differences. They suggest a reduction over time for all species abundance and biomass in both regions of the study area. Abundance indices for all species in spring cruises decreased from the pre- to the post-dumping periods but increased in autumn cruises. Variabilities for all species were consistently lower (9-24%) than those for individual species ( 18-88% ) (Tables 4-7 ). This suggests that fluctuations in total biomass in the study area from 1982 to 1990 for all species were lower as a whole and masked fluctuations in individual species. One may infer that the impacts of fishing, ecological perturba- tions, or natural factors acting on all species as a whole. therefore, were relatively low in the spring. Although impact on the abundance indices for all species was less than on some individual species, there were no significant differences in the number of species be- tween pre- and post-dumping periods in either region. Spatial differences To assess spatial differences for species CPUE between north and south regions, the same approach was used as for temporal differences. Test statistics for spatial differences are summarized in Table 2 and Tables 4-7. These are based on species CPUE, species abundance indices, and their variability estimates from collective cruises in pre- and post-dumping periods. More species test statistics for autumn cruises were significantly different between the north and south 602 Fishery Bulletin 91(4), 1993 Table 5 Species abundance indices (e.g., mean and SSE [standard error]) based on number of stratified catch per tow and variabilities (CV in %) using pre- and post-dumping spring cruises collec- tively in north and south regions. Pre-dumping spring cruises Post-dumping spring cruises Abundance indices Abundance indices Species name Mean SSE CV(%) Mean SSE CV(%) North Silver hake 7.143 1.66440 23 9.255 2.13410 23 Red hake 6.491 2.35950 36 5.298 1.52160 29 Summer flounder 0.591 0.12461 21 0.350 0.07453 21 Goosefish 0.410 0.08432 21 0.285 0.05213 18 Black sea bass 0.410 0.23043 56 0.101 0.05330 58 Scup 3.027 1.76110 58 0.574 0.31650 55 Butterfish 28.105 13.54800 48 9.375 4.93800 53 Longfin squid 17.936 5.27820 29 37.478 15.55800 42 American lobster 0.995 0.46534 47 0.897 0.27397 31 Sea scallop 3.214 1.77760 55 35.384 23.24100 66 Dogfish 49.555 16.10700 33 79.592 22.47900 28 All spp.1 511.574 48.78400 10 554.673 51.95500 9 South Silver hake 2.431 0.57454 24 4.874 1.80180 37 Red hake 2.311 0.79196 34 0.641 0.21119 33 Summer flounder 1.016 0.31625 31 0.760 0.18100 24 Goosefish 0.314 0.07206 23 0.147 0.04088 28 Black sea bass 0.816 0.26722 33 2.007 0.71847 37 Scup 21.419 18.81900 88 17.119 8.51020 50 Butterfish 11.695 6.16760 53 51.810 44.22500 85 Longfin squid 50.430 21.27200 42 119.842 57.41000 48 American lobster 0.133 0.03483 26 0.137 0.04451 32 Sea scallop 1.845 0.62564 34 25.703 17.49900 68 Dogfish 45.598 16.41000 36 39.621 7.73790 20 All spp.1 450.046 47.42500 11 675.431 70.50500 10 1 "All species" includes the eleven listed and all other species. regions than those for spring cruises (Table 2). Several patterns are evident in significant test statistics for the autumn cruises. Biomass CPUE for eight species showed highly significant differences for both pre- and post- dumping periods. These species were silver hake, red hake, summer flounder, goosefish, black sea bass, but- terfish, longfin squid, spiny dogfish, and all species. Fewer significant differences were observed in the num- ber of CPUE. No distinct pattern was obvious for spring cruises. Only silver and red hake CPUE showed signifi- cant test statistics for both periods. Scup, American lob- ster, and sea scallop CPUE showed generally lower val- ues that were not significantly different. There was a significant difference between north and south in num- ber of species for autumn cruises in both the pre- dumping and post-dumping periods (Table 2). Interpretation of these significant spatial differences requires special caution. In 1982, the first year for which data were used in this study, populations were not of comparable size in the north and south regions. One must consider how populations of unequal size could have accommodated differing environmental per- turbations and fishing pressure in the north and south regions. Silver hake, red hake, butterfish, and spiny dogfish indices were highest. Goosefish and American lobster indices were lower overall, but higher in the north than in the south, while scup were higher only in the south. Longfin squid indices were consistent with its migratory pattern (concentrated in the south in spring and in the north in autumn) (Tables 4-7). Test statistics of all eight species were significantly different between north and south (Table 2). Test statistics were tabulated for species CPUE be- tween north and south, and for species abundance indi- ces, including all species in the north and south regions for individual cruises, but are not presented here. This information supplemented the findings from collective cruises (Tables 2, 4-7). Indices from individual cruises Chang; Effects of sewage sludge dumping on fishery resources 603 Table 6 Species abundance in dices le.g.. mean and SSE [standard error]) based on weight o " stratified catch per tow and variabilities (CV in % ) using pre- and post ■dumping autumn cruises collec- tively in north and south regions. Species Pre-dumping spring cruises Post-dumping spring cruises Abundance indices Abundance indices name Mean SSE CVC*) Mean SSE CV(%) North Silver hake 1.573 0.31258 23 0.997 0.30246 31 Red hake 2.707 0.69497 26 0.380 0.09313 25 Summer flounder 0.328 0.08221 25 0.135 0.06060 45 Goosefish 1.539 0.34043 22 0.350 0.10007 29 Black sea bass 0.010 0.00598 60 0.005 0.00363 73 Scup 0.905 0.40331 45 0.203 0.07233 36 Butterfish 9.356 3.27910 35 7.924 2.74850 35 Longfin squid 7.771 2.77170 36 4.266 0.94622 22 American lobster 0.539 0.10552 20 0.630 0.11589 18 Sea scallop 0.284 0.10900 38 1.810 1.07450 59 Dogfish 53.835 32.34200 60 19.980 13.61500 68 All spp.1 129.958 21.28800 16 76.576 7.16770 9 South Silver hake 0.178 0.05286 30 0.060 0.01787 30 Red hake 0.225 0.11908 53 0.052 0.02049 39 Summer flounder 0.439 0.16250 37 0.072 0.02495 34 Goosefish 0.111 0.03029 27 0.032 0.01329 42 Black sea bass 0.082 0.04975 61 0.060 0.02440 41 Scup 1.779 1.05160 59 4.660 3.96680 85 Butterfish 2.605 1.04430 40 2.803 1.43660 51 Longfin squid 2.716 0.64098 24 1.433 0.35703 25 American lobster 0.125 0.02937 24 0.069 0.02283 33 Sea scallop 0.278 0.14392 52 1.126 0.50005 44 Dogfish 0.006 0.00351 59 0.022 0.00672 31 All spp.1 65.832 11.32200 17 44.661 9.18910 21 '"All species" includes the eleven listed and all other species. provided insights into how individual fish and shellfish populations responded in time and space. The variabili- ties were relatively consistent in both regions, despite confounding biotic and abiotic factors. Fluctuations of total biomass for eleven individual species, and for all species as a whole, were relatively stable in both north and south regions during the period 1982-90. Seasonal differences Seasonal differences between spring and autumn were determined by the same method used to evaluate tem- poral and spatial differences. Test statistics for seasonal differences based on species CPUE and species abun- dance indices with their variability estimates were drawn from collective cruises of the pre- and post-dump- ing periods. They are summarized in Tables 3 and 4-7. Species test statistics for the pre-dumping cruises showed more significant differences between spring and autumn seasons than did post-dumping cruises. Weight test statistics provided more significant differences than those based on number. Spiny dogfish CPUE revealed significant test statistics for all categories. Test statis- tics for black sea bass, based on both biomass and number, showed significant differences between sea- sons for the post-dumping cruises in the southern re- gion, but not for the pre-dumping cruises in the same region. Similarly, silver hake, butterfish, and Ameri- can lobster indices in the northern region were signifi- cantly different for the post-dumping cruises, but not significantly different for pre-dumping cruises (Table 3). Some species weight abundance indices declined from the pre-dumping to post-dumping periods for both spring and autumn seasons (Tables 4 and 7). These included silver and red hakes, summer flounder, goosefish, and black sea bass. Number of species for the post-dumping cruises in the southern region was also significantly different (Table 3). 604 Fishery Bulletin 91(4), 1993 Table 7 Species abundance in iices le.g., mean and SSE \> tandard error] 1 based on number of stratified catch per tow and variabilities (CV in %) using pre- and post-dumping autumn cruises collec- tively in north and south regions. Pre-dumping spring cruises Post-d jmping spring cruises Abundance indices Abundance indices Species name Mean SSE CV(%) Mean SSE CV(%) North Silver hake 32.869 9.57110 29 16.073 4.51730 28 Red hake 13.974 3.72120 27 2.968 0.72162 24 Summer flounder 0.268 0.07093 26 0.135 0.05503 41 Goosefish 0.658 0.13661 21 0.334 0.07429 22 Black sea bass 0.530 0.24114 46 0.137 0.05997 44 Scup 9.652 3.80610 39 7.770 5.12530 66 Butterfish 191.717 60.39400 32 241.977 87.52200 36 Longfin squid 307.679 141.09000 46 297.160 92.28000 31 American lobster 1.245 0.24816 20 1.573 0.30140 19 Sea scallop 4.165 1.58340 30 39.042 25.79400 66 Dogfish 32.848 18.35200 56 13.260 8.71090 66 All spp.1 1481.352 162.12000 11 1365.058 154.03000 11 South Silver hake 8.632 2.41370 28 3.340 1.03060 31 Red hake 1.436 0.65891 46 0.482 0.19790 41 Summer flounder 1.136 0.40306 22 0.200 0.06627 33 Goosefish 0.252 0.05769 23 0.128 0.03517 27 Black sea bass 1.003 0.62278 62 0.823 0.36718 45 Scup 40.626 24.08400 59 149.944 132.31000 88 Butterfish 85.958 32.22600 37 84.561 47.39100 56 Longfin squid 107.343 33.92100 32 74.365 38.98300 52 American lobster 0.152 0.03600 24 0.142 0.04323 30 Sea scallop 5.443 3.00540 55 22.849 9.84010 43 Dogfish 0.039 0.02346 60 0.131 0.03651 28 All spp.' 842.080 117.52000 14 815.743 133.00000 16 '"All species" includes the eleven listed and all other spec es. Discussion Analyses of fisheries abundance data, based on catch per unit of effort (CPUE) of NOAAs Northeast Fisher- ies Science Center (NEFSC) bottom trawl surveys, in- dicate that abundances of silver hake, red hake, sum- mer flounder, goosefish, and black sea bass declined about the 106-mile dumpsite (106-MDS) from spring of 1982 through spring of 1990. These declines are concurrent with the temporary dumping there of sew- age sludge from the New York and New Jersey metro- politan area. Changes in total abundance and biomass for all species combined occurred to a lesser degree because of increased abundance of spiny dogfish, skates {Raja spp.) and some pelagic species (e.g., Atlantic mackerel) (NMFS 1991, 1992). There were statistically significant, temporal differences in average abundance values for some individual species and for all species based on spring and autumn bottom trawl surveys within the north and south dumpsite areas. It is par- ticularly worth noting that much of the abundance decline occurred in the south area of the dumpsite. The south region is most likely to be influenced by sewage sludge introduced in waters passing toward the west-southwest shallow outer shelf through the dumpsite in the potential area of influence (PAIi of the sewage sludge (Fig. 1; Bisagni, 1983; O'Connor et al, 1985; Gentile et al., 1989; Inghamls; Warsh13; Bisagni'). There were also statistically significant spa- tial differences in average abundance of many indi- vidual species and all species sampled on the spring and autumn surveys within the pre- and post-dump- ing period at the 106-mile dumpsite. There were sea- sonal differences, as well, in average values for sev- Chang. Effects of sewage sludge dumping on fishery resources 605 eral individual species and all species for north and south areas at the site, and within the pre and post sludge dumping periods. The hypothesis that fishery resource abundance in the vicinity of the 106-MDS and adjacent outer conti- nental shelf waters was not affected by sewage sludge dumping there from 1986 to 1990 is thereby rejected. Also rejected are the three secondary hypotheses re- lated to the primary one. The likelihood remains that waste disposal, even in deep-ocean waters, has mea- surable adverse impacts on the resource abundance, at least in the area of direct influence of the toxic waste. Analysis of survey data alone, however, does not establish cause and effect. It is necessary to consider the likelihood that events neither directly nor indi- rectly related to sewage sludge dumping at the 106- MDS could have led to the measured decline in abun- dance and shift in species composition within the study period. Also, it is necessary to consider how an abun- dance shift could have occurred through sludge dump- ing within this period of time. Fishing is foremost among factors affecting fishery resource abundance and species composition unrelated to any effect of ocean disposal of toxic waste. Impacts of fishing pressure on the species sampled at the off- shore dumpsite and analyzed in this study are poorly known. There are also other poorly understood natu- ral factors affecting species population abundance around the 106-MDS in the PAI and elsewhere that could have influenced abundance and species composi- tion at this site over the period of the study reported here. These include shifts in spawning time and area, size of predator stocks, decline in size of the spawning biomass, and increase in early life mortality (Gross, 1976; Hempel, 1978; Mayer, 1982; Cross et al., 1985; Tiews, 1985). The dramatic reductions in stock size affected by powerful modern fishing techniques, how- ever, through sheer reduction in stock size, may influ- ence all these factors so that their fluctuation is no longer entirely natural. These same natural factors are likewise potentially affected by ocean dumping of toxic waste as evidenced in the environmental litera- ture (Carlise, 1969; Gross, 1976; Mearns, 1981; Mayer, 1982; Spies, 1984; Wolf and O'Connor, 1988; Champ and Park, 1989; Hood et al., 1989; Baumgartner and Duedall, 1990; Longwell et al. 1992; Longwell21; Young and Mearns22; NMFS1"11; Studholme et al.12). Changes -'Longwell, A. C. 1981. Cytological examination offish eggs collected at and near 106-Mile Site. In NOAA Special Report. Assessment Report on the Effects of Waste Dumping in 106-Mile Ocean Waste Disposal Site. Dumpsite Evaluation Report 81-1. U.S. Dep. Commer., NOAA. Office of Mar. Pollution Assessment. Rockville, MD. 257-276 p. --Young, D. R., and A. J. Mearns. 1978. Pollutant flow through food web. Southern California Coastal Water Research Project. Annual Report 1978, 185-202 p. in water mass patterns and global warming are possi- bly the only phenomena not potentially influenced by ocean dumping. Given the complexity of natural and anthropogenic forces acting on populations, existing data sets from fishery ecology are simply not suffi- ciently synoptic or complete enough to be useful in sorting out direct sludge dumping effects from natural or pseudo-natural processes driving population fluctuation. Oceanographic dynamics around the 106-MDS make it unlikely that any change in harvestable fishery re- sources or catch composition could be the result of direct mortality attributable to the sewage since relo- cated sludge dumping began there in 1986. There is some evidence from embryo studies of planktonic eggs collected in the wake of sludge and acid waste disposal at 106-MDS that direct kills of floating fish eggs can occur, but this dumpsite is not a significant spawning grounds for resource species (Longwell21). Changes in fishery resource abundance at 106-MDS could, however, have been affected by a change in fish behavior in response to dumped material as well as to natural environmental perturbations (Olla et al., 1980). Migration to other areas to escape sludge disposal and its aftermath or to search for other feeding grounds is likely to alter the time and place of spawning and subsequent early-life survival. Reduced reproductive success and recruitment re- sulting from increased baseline contaminant body bur- dens of spawners is another mechanism whereby the population changes measured in the study presented here could have resulted from sludge dumping (Cross and Hose, 1988; Longwell et al., 1992). Offshore con- tamination of fish has the same potential of affecting subsequent spawning and recruitment as does contami- nation in nearshore waters and on the spawning grounds. Shifts in species composition could have come about if predator or competitor, or both, species were more tolerant of the sludge than were the resource species. In addition, poor quality eggs could have re- sulted from inadequate maternal nutrition if there was much direct prey mortality or pollution-impaired re- production of prey organisms in the wake of the sludge. Contaminant burdens of the reproductive tissues or ripe eggs of the species analyzed in this study are unknown. Extrapolation and formal treatment of ex- isting data on muscle and liver tissues of these or on the reproductive tissues of other species are probably not worthwhile. Contaminant burdens of mature win- ter flounder eggs alone (Calabrese et al.23) suggest that -"Calabrese A.. A. C. Longwell, and F. P. Thurberg. 1989. Final re- port on early reproductive success of winter flounder from Boston Harbor with comparisons made to Long Island Sound. Prepared for U.S. Environmental Protection Agency. U.S. Dep. Commer., NOAA. Natl. Mar. Fish. Serv., Northeast Fish. Sci. Cent.. Milford Lab.. Milford. CT, 61 p. 606 Fishery Bulletin 91(4), 1993 safety margins for adverse reproductive effects of sev- eral contaminants (Susani, 1986; Sorenson, 1991) on fish species at the 106-MDS could easily have been exceeded. Only moderate concentrations of chlorinated hydrocarbons have been shown to be associated with detrimental reproductive effects (Susani, 1986; Westernhagen et al., 1989). An interdisciplinary effort was made to determine changes in contaminant burdens of fish at the 106- MDS once sludge dumping began (NMFS10'11). In 1991, concentrations of metals were relatively low in the epibenthic megafauna species of deepwater fishes, such as blue hake (Antimora rostrata), rattails iCor- yphaenoides carapinus and C. armatus), halosaur (Halosauropsis macrochir) and commercially valuable slope-dwelling tilefish (Lopholatilus chamaeleonticeps), shrimp iNematocarcinus ensifer and Glyphocrangon sculpta), and American lobster (Homarus americanus). This was generally true for American lobster, with the exception of elevated liver Cd, which could reflect its trans-shelf migration and possible exposure to coastal pollution. Finfishes, however, generally contained rela- tively high concentrations of chlorinated pesticides and total PCB's in their livers (NMFS1"11). Significantly elevated levels of several metals (Ag, Cu,and Cr) were found in midwater myctophids (e.g., Benthosema gla- ciate, Lobianchia dofleini, Ceratoscopelus maderensis, and Hygophum hygomii) at stations southwest of the 106-MDS in the principal area of influence of the sludge. Plankton (primarily copepods) provide a pathway for entry of potentially toxic chemical contaminants from the sludge into the ocean food-web, because it consti- tutes the main prey of smaller midwater fishes. Con- centrations of metals in plankton samples taken in 1991 were comparable to, or higher than, those ob- served in fishes collected in 1989. The geographic dis- tributions of metal concentrations in 1991 fishes and plankton suggest that the elevated levels of metals found in certain samples are probably attributable to dumping activity at the 106-MDS. Boehm (1983), how- ever, postulated that offshore, southerly transport of organic contaminants along the New Jersey shore and down-valley transport from the New York Bight apex caused increased metal contaminants. Organic contami- nants (PCB's, polynuclear aromatic hydrocarbons (PAH's) and pesticides) were present in lower con- centrations in zooplankton than in midwater fishes (NMFS1011). The highest concentration patterns of PAH's, PCB's, and Ag in the sediment surface layer ( 0-0.5 cm ) tended to follow the distribution patterns of spores of Clostridium perfringens (a bacterial indicator of sew- age) in sediments collected from depths of 100 to 2800 m in the vicinity of the 106-MDS. Clostridium perfringens counts declined gradually to the southwest. Surface sediments were not detectably contaminated with other trace metals. Organic contaminant concentrations in sub-surface strata (0.5-3.5 cm) were higher than in the sediment or deeper layers. Artifacts in sediment cores in the 106-MDS tend to confirm that chemical concentration is related to dumping. All data imply that material dumped at the 106-MDS reached the seabed in the area southwest of the site in the PAL Data imply that the chemicals associated with the sludge entered the food-web around the dumpsite in the area believed to have been under principal influ- ence of the sludge (NMFS1"11). This is the area ac- counting for the most of the decline in fish species measured in this study. Unless absorbed by particles or precipitated by other constituents in the water column, slowly sinking wastes must take a long time to reach the bottom because of the great water depth. This increases the likelihood of significant ingestion of waste and associated contami- nants by mid-water column prey species of resource fish. Deepwater food-web dynamics have been explored in the 106-MDS by Van Dover et al. (1992) by using the natural stable isotopes of organic carbon nitrogen and sulphur in sewage sludge to trace sewage-oriented organic matter. Organic matter from the sludge was found to reach the deep-sea floor and enter the benthic food-web through consumption by surface-deposit feed- ers, including a sea urchin (Echinus affinus) and a sea cucumber (Bentlwdytes sanquinolenta). Other surface- deposit feeders, including infaunal benthic species such as polychaetes and molluscs, probably also ingest this organic matter contributing to sludge impacts in the open ocean via food-chain dynamics. There are strong associations of benthic macrofauna with habitat types and sediment contaminants in the continental shelf of the New York Bight (Chang et al., 1992). Species most common in the contaminated area around 12-MDS were mainly polychaetes (e.g., Tharyx acutus, Nephtys incisa, Pherusa affinis, and Capitella spp.), as well as a nemertean (Cerebratulus lacteus), an anemone (Ceriantheopsis americanus), a phoronid (Phoronis architecta ) and the nut clam Nucula proximo. Another group of species was consistently associated with minimally contaminated sediments and appeared to represent a basic natural benthic macrofaunal as- semblage for the typical sandy habitat of the Bight shelf. This group included the sand dollar (Echin- arachnius parma) and several species of amphipods (e.g., Byblis serrata, Corophium crassicorne, and Am- pelisca agassizi), as well as polychaetes (e.g., Goniadella gracilis and Exogone hebes). The roles of these contaminant-sensitive and insensitive species within Chang: Effects of sewage sludge clumping on fishery resources 607 certain food-web dynamics for commercially important fisheries resources in the New York Bight are becom- ing known (Steimle, 1985; Steimle and Terranova, 199H. Chemical analyses of sediment cores tend to confirm that sediment chemical concentration around the 12-MDS is related to some types of dumping, and that sediment concentrations decrease appreciably in some areas 20 months after cessation of sewage sludge dumping (Zdanowicz et al., 199324). Distribution of the contaminant-sensitive and insensitive benthic inverte- brate species assemblages in the New York Bight (Chang et al., 1992) then imply that material dumped at the 12-MDS reached the bottom in the area where it was dumped, entered the food-web at the lower trophic levels of benthic organisms, and ascended to the higher trophic levels of predatory fish species (Steimle et al.25,26). Similar food-web dynamics could have influenced the temporal and spatial differences in abundance of commercially important species mea- sured in the principal area of sludge influence at the 106-MDS. Conclusions Effects of sewage sludge dumping even in the deep ocean at the 106-MDS, like sludge dumping at the 12- MDS in the New York Bight and release of sewage from outfalls in southern California outfalls, cannot be excluded as a factor measurably affecting fishery re- source abundance and composition. Natural factors may have caused the population fluctuation assessed around the 106-MDS, but there are no adequate data sets for testing the likelihood that such phenomena, and not deepwater dumping, are responsible. Data sets are even more inadequate for measuring interactive effects of dumping, fishing, and natural phenomena, which must certainly occur. On the other hand, there are identifi- able mechanisms by which changes in abundance and 24Zdanowicz, V. S., S. Leftwich. and T. W. Finneran. 1993. Reduc- tions in sediment metal contamination in the New York Bight apex with cessation of sewage sludge dumping. In A. L. Studholme, J. O'Reilly and M. C. Ingham (eds.), Effects of the cessation of dump- ing at the 12-mile site. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Northeast Fish. Sci. Cent., Sandy Hook Lab., Highlands, NJ, 14 p. (In review.) -^Steimle, F. W. Jr., V. S. Zdanowicz, S. L. Cunneff. and R. Terranova. 1993a. Trace metal concentrations in common benthic macrofaunal prey from New York Bight apex. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv, Northeast Fish. Sci. Cent., Sandy Hook Lab., High- lands, NJ. 21 p. (In review.) 26Steimle, F. W. Jr., D. Jeffress, S. A. Fromm, R. N. Reid. J. J. Vitaliano, and A. Frame. 1993b. Prey selectivity by winter flounder Pleuronectes americanus in the New York Bight apex. LIS. Dep. Commer., NOAA, Natl. Mar. Fish. Serv, Northeast Fish. Sci. Cent., Sandy Hook Lab., Highlands, NJ, 28 p. (In review.) composition of fishery resources could have come about in relatively few years as a result of sludge dumping after prior years of industrial and chemical waste dis- posal at this deepwater site. Increased contamination of the food-web and environment could well have di- rectly and indirectly, possibly differentially, affected both behavior and reproduction of ecologically linked species leading to the unfavorable temporal, spatial, and seasonal differences of the fishery resources around the 106-MDS. Analyses of bottom trawl surveys and fishery land- ings data in relation to environmental pollution are of increasing importance as efforts are made to measure any impact of toxic contaminants on fishery resource. The influence of natural and man-induced environmen- tal factors, such as waste dumping, on fishery resources can be treated as concomitant variables (Thomas et al., 1976; Butler and Schutzman, 1979; Marking and Kimerle, 1979; Geyer, 1981, a and b; Reed et al., 1985; Miller et al.. 1988; Wallace et al., 1988; Stoddard and Walsh, 1988; Connor, 1989; Gift et al., 1989; Stanford and Young, 1988, Furness and Rainbow, 1990; Word et al, 1990; Manning, 1991). Interpretation of fishery as- sessment data in relation to the sewage sludge dump- ing at the 106-MDS, as developed in the discussion above, makes clear that future efforts in all these re- gards would benefit from better and updated baseline data on the fishery resources. If there were food-web dynamics data as well as more behavioral and repro- ductive data on responses of marine fish to contami- nant levels known to occur in the ecosystems of con- cern, we would then be better able to understand and determine causal factors influencing species abundance and fishery resources in the vicinity of the 106-MDS and adjacent continental shelf waters. Much is still to be learned about both the natural fluctuation of envi- ronmental factors that influence species abundance and population dynamics and about anthropogenic effects on the natural environment. A better understanding of natural and man-induced effects on fishery resources is necessary to minimize adverse controllable impacts on fisheries resources and requires further synthesis of environmental and fishery assessments. Acknowledgments I would like to acknowledge the contribution of scien- tists from Resource Surveys Investigation, Conser- vation and Utilization Division, Northeast Fisheries Science Center. Their efforts and perseverance in the collection of bottom trawl survey data and maintenance of the survey data system produced the extensive data base used for the analysis in the paper. 608 Fishery Bulletin 91(4), 1993 I would like to thank NMFS colleagues: Arlene Longwell of the Milford Laboratory; Anthony Pacheco and Frank Steimle of the Sandy Hook Laboratory for valuable suggestions regarding presentation of the re- sults and discussion; Robert Reid, Anne Studholme, and Vincent Zdanowicz of the Sandy Hook Laboratory; and Robert Murchelano of the Woods Hole Laboratory, who also provided critical reviews. Francis Csulak of NOAA's Damage Assessment Center provided valuable information on ocean dumping in the New York Bight and Brenda Figeurido and Malcolm Silverman of the Woods Hole Laboratory, the bottom trawl survey figure. Critical comments of two anonymous reviewers and the constructive editorial suggestions of the scientific editor of this bulletin are greatly appreciated. Literature cited Azarovitz, T. R. 1981. A brief historical review of the Woods Hole Labo- ratory trawl survey time series. In W. G. Doubleday and D. Rivard (eds.). Bottom trawl survey. Can. Spec. Publ. Fish. Aquat. Sci. 58:62-67. Baumgartner, D. J., and I. W. Duedall (eds.). 1990. Oceanic processes in marine pollution. Vol. 6: Physical and chemical processes: transport and transformation. R. E. Krieger Publ. Co., Malabar, FL, 248 p. Bisagni, J. J. 1983. Physical oceanography. Chapter 2. In J. B. Pearce, D.C. Miller, and C. Berman (eds.), 106-mile site characterization update. U.S. Dep. Commer., NOAATech. Memo. NMFS-F/NEC-26, 44 p. Boehm, P. D. 1983. Coupling of organic pollutants between estuary and the continental shelf and the sediments and wa- ter column of the New York Bight region. Can. J. Fish. Aquat. Sci. 40 (Suppl. 21:262-276. Butler, P. A., and R. L. Schutzman. 1979. Bioaccumulation of DDT and PCB in tissues of marine fishes. In L. L. Marking and R. A. Kimerle (eds.), Aquatic toxicology. Am. Soc. Testing and Ma- terials, Philadelphia, PA, 212-220 p. Capuzzo, J. M., and D. R. Kester (eds.). 1987a. Oceanic processes in marine pollution. Volume 1: Biological processes and wastes in the ocean. R. E. Krieger Publ. Co., Malabar, FL, 265 p. 1987b. Biological effects of waste disposal: experimen- tal results and predictive assessments. In J. M. Capuzzo and D. R. Kester (eds.), Oceanic processes in marine pollution. Vol. 1: Biological processes and wastes in the ocean. R. E. Krieger Publ. Co., Malabar, FL, 3-15 p. Carlise, J. G. Jr. 1969. Results of a six-year trawl study in an area of heavy waste discharge: Santa Monica Bay, California. Calif. Dep. Fish and Game. 55( 1 1:26-46. Champ, M. A., and P. K. Park (eds.). 1989. Oceanic processes in marine pollution. Vol. 3: Marine waste management: science and policy. R. E. Krieger Publ. Co., Malabar, FL, 341 p. Chang, S. 1990. Seasonal distribution patterns of commercial landings of 45 finfish and shellfish species off the northeastern United States during 1977-88. U.S. Dep. Commer, NOAA Tech. Memo. NMFS-F/NEC- 78, 132 p. Chang, S., and A. C. Longwell. 1984. Examining statistical associations of malforma- tion, cyto-pathology and cytogenetic abnormality of 1977 and 1978 Atlantic mackerel embryos with indi- cator levels of environmental contaminants. Int. Counc. Explor. Sea (ICES). CM. 1984/E:11, 9 p. Chang, S., F. W. Steimle, R. N. Reid, S. A. Fromm, V. S. Zdanowicz, and R. A. Pikanowski. 1992. Association of benthic macrofauna with habitat types and quality in the New York Bight. Mar. Ecol. Prog. Ser. 89(2-3):237-251. Connor, M. S. 1989. Estimating the public health risk of organic carcinogens. In M. A. Champ and P. K. Park (eds.), Ocean processes in marine pollution. Vol. 3: Marine waste management: science and policy. R. E. Krieger Publ. Co., Malabar, FL, 221-229 p. Cross, J. N., and J. E. Hose. 1988. Evidence for impaired reproduction in white croaker iGenyonemus lineatus) from contaminated areas off southern California. Mar. Environ. Res. 24:185-188. Cross, F. A., D. S. Peters, and W. E. Schaaf. 1985. Implications of waste disposal in coastal waters on fish populations. In R. D. Cardwell, R. Purdy, and R. C. Bahner (eds.). Aquatic toxicology and haz- ard assessment, seventh symposium. Am. Soc. Test- ing and Materials, Philadelphia, PA, 383-399 p. Dart, M. C, and S. H. Jenkins (eds.). 1981. Disposal of sludge to sea. Pergamon Press, 230 p. Farrington, J. W., J. M. Capuzzo, T. M. Leschine, and M. A. Champ. 1982. Ocean dumping. Oceanus 25(41:39-50. Furness, R. W., and P. S. Rainbow. 1990. Heavy metals in the marine environment. CRC Press, Boca Raton, FL. Gentile, J. H., V. J. Bierman Jr., J. F. Paul, H. A. Walker, and D. C. Miller. 1989. A hazard assessment research strategy for ocean disposal. In M. A. Champ and P. K. Park (eds.). Ocean processes in marine pollution. Vol. 3: Marine waste management: science and policy. R. E. Krieger Publ. Co., Malabar, FL, 199-212 p. Geyer, R. A. (ed.). 1981a. Marine environmental pollution. 1: Hydrocar- bons. Elsevier Oceanography Series 27B, Elsevier Scientific Publ. Co.. Amsterdam, Netherlands, 591 p. 1981b. Marine environmental pollution. 2: Dumping and mining. Elsevier Oceanography Series 27B, Chang Effects of sewage sludge clumping on fishery resources 609 Elsevier Scientific Publ. Co., Amsterdam, Netherlands, 574 p. Gift, J. J., J. A. Feva, A. F. Maciorowski, and W. L. McCulloch. 1989. Evaluation of sewage sludge bioaccumulation and toxicity tests. In M. A. Champ and P. K. Park (eds.), Ocean processes in marine pollution. Vol. 3: Marine waste management: science and policy. R. E. Krieger Publ. Co., Malabar, FL, 213-220 p. Gross, M. G. (ed). 1976. Middle Atlantic Continental Shelf and New York Bight. Special symposium volume. Am. Soc. Lim- nology and Oceanography, 441 p. Grosslein, M. D. 1974. Bottom trawl survey methods. Int. Comm. Northwest Alantic Fisheries (ICNAF) Res. Doc. 74/ 96, 27 p. Grosslein, M. D., B. E. Brown, and R. C. Hennemuth. 1979. Research, assessment, and management of a marine ecosystem in the northwest Atlantic: a case study. In J. Cairns Jr., G. S. Patil, and W. E. Waters (eds.), Environmental biomonitoring, assessment, pre- diction and management: certain case studies and related quantitative issues. Statistical Ecology Se- ries. Vol. 11. Int. Cooperative Publ. House, Bur- tonsville, MD, 289-357 p. Hempel, G. 1978. North sea fisheries and fish stocks — a review of recent changes. Rapp. P.-V Reun. Cons. int. Explor. Mer. (ICES) 173:145-167. Hood, D. W., A. Schoener, and P. K. Park. 1989. Oceanic processes in marine pollution. Volume 4: Scientific monitoring strategies for ocean waste disposal. R. E. Krieger Publ. Co., Malabar, FL, 286 p. Longwell, A. C. 1988. Fish embryos: practical indicators of environmen- tal quality of significance to fisheries. Water Resour. Bull. 24:999-1005. Longwell, A. C, S. Chang, A. Hebert, J. B. Hughes, and D. Perry. 1992. Pollution and developmental abnormalities of the Atlantic fishes. Environ. Biol, of Fishes. 35:1-21. Manning, J. 1991. Circulation: current meter observation. //; A. L. Studholme, M. C. Ingham, and A. L. Pacheco (eds.), Response of the habitat and biota of the inner New York Bight to abatement of sewage sludge dumping: third annual progress report. Dep. Commer., NOAA Tech. Memo. NMFS F/NEC-82, 4-7 p. Marking, L. L., and R. A. Kimerle (eds.). 1979. Aquatic toxicity. Am. Soc. Testing and Materi- als, Philadelphia, PA, 392 p. Mayer, G. F. (ed.). 1982. Ecological stress and the New York Bight: sci- ence and management. Estuarine Res. Federation, Columbia, SC. 715 p. Mearns, A. J. 1981. Effects of municipal discharges on open coastal ecosystem. In R. Geyer (ed.), Marine environmental pollution. Vol. 2: Dumping and mining. Elsevier Oceanography Series 27B, Elsevier Scientific Publ. Co., Amsterdam, Netherlands, 25-66 p. Miller, D. C, M. Marcy, W. Berry, C. Deacutis, S. Lussier, A. Q. Kuhn, M. Heber, S. C. Schimmel, and E. -lack i in. 1988. The acute toxicity of sewage sludge to marine fish, mysids and copepods. In D. A. Wolf and T P. O'Connor (eds.), Oceanic processes in marine pollu- tion. Vol. 5: Urban wastes in coastal marine environments. R. E. Krieger Publ. Co., Malabar, FL, 103-113 p. Murchelano, R. A., and J. J. Ziskowski. 1976. Fin rot disease studies in the New York Bight. In M. G. Gross (ed.). Middle Atlantic Continental Shelf and the New York Bight. Am. Soc. Limnology and Oceanography Special Symposium 2:329-336. NMFS (National Marine Fisheries Service). 1991. Status of fishery resources off the Northeastern United States for 1991. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-F/NEC-86, 132 p. 1992. Status of fishery resources off the Northeastern United States for 1992. Dep. Commer., NOAA Tech. Memo. NMFS-F/NEC-95, 133 p. Norton, M, G. 1989. A rationale for monitoring ocean dumpsites. //; M. A. Champ and P. K. Park (eds.), Ocean processes in marine pollution. Vol. 3: Marine waste manage- ment: science and policy. R. E. Krieger Publ. Co., Malabar FL, 241-251 p. O'Connor, T .P. 1983. Contaminant inputs, fate and effects. Chapter 12. In J. B. Pearce, D. C. Miller, and C. Berman (eds.), 106-mile site characterization update. Dep. Commer., NOAA Tech. Memo. NMFS-FNEC-26, 61 p. O'Connor, T. P., A. Okubo, M. A. Champ, and P. K. Park. 1983. Projected consequences of dumping sewage sludge at a deep ocean site near New York Bight. Can. J. Fish. Aquat. Sci. 40(2):228-241. O'Connor, T. P., H. A. Walker, J. F. Paul, and V. J. Bierman Jr. 1985. A strategy for monitoring of contaminant distri- butions resulting from proposed sewage sludge dis- posal at the 106-mile ocean disposal site. Mar. Environ. Res. 16:127-150. O'Connor, T. P., W. V. Burt, and I. W. Duedall (eds.). 1987a. Oceanic processes in marine pollution. Vol. 2: Physicochemical processes and wastes in the ocean. R. E. Krieger Publ. Co., Malabar, FL, 236 p. O'Connor, T. P., M. G. Norton, A. J. Mearns, D. A. Wolf, and I. W. Duedall. 1987b. Scale of biological effects. In T P. O'Connor, W. V Burt, and I. W. Duedall (eds.), Oceanic pro- cesses in marine pollution. Vol. 2: Physicochemical processes and wastes in the ocean. R.E. Krieger Publ. Co., Malabar FL, p. 1-7. Olla, B. L., W. H. Pearson, and A. L. Studholme. 1980. Applicability of behavioral measures in environ- mental stress assessment. Papp. P.-v. Reun. Cons, int. Explor. Mer (ICES) 179:162-173. 610 Fishery Bulletin 91(4). 1993 Pearce, J. B., D. C. Miller, and C. Berman (eds.). 1983. 106-mile site characterization update. U.S. Dep. Commer, NOAATech. Memo. NMFS-F/NEC-26, 483 p. Pennington, M., and B. E. Brown. 1981. Abundance estimators based on stratified ran- dom trawl surveys. In W. G. Doubleday and D. Rivard (eds.), Bottom trawl survey. Can. Spec. Publ. Fish. Aquat. Sci. Publ. 58, 149-153 p. Pikanowski, R. A. 1992. The effects of ocean disposal of sewage sludge on the relative abundance of benthic megafauna. Chemistry and Ecology 6:199-212. Puri, M. L., and P. K. Sen. 1971. Nonparametric methods in multivariate anal- ysis. Wiley & Sons, New York, NY, 438 p. Reed, M., T. Isaji, J. Rosen, and S. Hurlbut. 1985. Modelling the long term fate and effects of an- thropogenic pollutants on the north American conti- nental shelf. Can. Tech. Rep. of Fish, and Aquat. Sci. 1368, 91-124 p. Russo, A. R. 1982. Temporal changes in fish community structure near a sewage ocean outfall, Mokapu, Oahu, Hawaii. Mar. Environ. Res. 6:83-98. Sorenson, E. B. 1991. Metal poisoning in fish. CRC Press, Boca Raton, Ann Arbor, 374 p. Spies, R. 1984. Benthic-pelagic coupling in sewage-affected ma- rine ecosystem. Mar. Environ. Res. 13:195-230. Squires, D. F. 1983. The ocean dumping quandary: waste disposal in the New York Bight. State Univ. New York Press, Albany, NY, 226 p. Stanford, H. M., and D. R. Young. 1988. Pollution loadings to the New York Bight apex. In Oceans 88 Conference Record. Mar. Technol. Soc. Washington D. C, 745-751 p. Steimle, F. W. Jr. 1985. Biomass and estimated productivity of the benthic macrofauna in the New York Bight: a stressed coastal area. Estuarine Coastal Shelf Sci. 21:539- 554. Steimle, F. W. Jr., and R. Terranova. 1991. Trophodynamics of select demersal fishes in the New York Bight. U.S. Dep. Commer, NOAA Tech. Memo. NMFS-F/NEC-84, 11 p. Stoddard, A., and J. J. Walsh. 1988. Monitoring oxygen depletion in the New York Bight: the water quality impact of a potential increase of waste inputs. In D. A. Wolf and T P. O'Connor (eds.), Oceanic processes in marine pollution. Vol. 5: Urban wastes in coastal marine environments. R. E. Krieger Publ. Co., Malabar, Fl, 91-102 p. Susani, L. 1986. Effects of contaminants on teleost reproduction: past and ongoing studies. Dep. Commer., NOAA Tech. Memo. NOS OMA 29, 18 p. Tiews, V. D. 1985. Review on the effects of pollution on marine fish life and fisheries. Z. angew. Ichthyol. 3:97-118. Thomas, J. P., W. C. Phoel, F. W. Steimle, J. E. O'Reilly, and C. E. Evans. 1976. Seabed oxygen consumption-New York Bight apex. In M. G. Gross (ed.l, Middle Atlantic Conti- nental Shelf and the New York Bight. Am. Soc. Lim- nology and Oceanography Special Symposium 2:354-369. Van Dover, C. L., J. F. Grassle, B. Fry, R. H. Garrit, and V. R. Starczak. 1992. Stable isotope evidence for entry of sewage- derived organic material into a deep-sea food web. Nature 360:153-156. Wallace, G. T. Jr., J. H. Waugh, and K. A. Garner. 1988. Metal distribution in a major urban estuary (Bos- ton Harbor) impacted by ocean disposal. In D. A. Wolf and T P. O'Connor (eds.). Oceanic processes in marine pollution. Vol. 5: Urban wastes in coastal marine environments. R. E. Krieger Publ. Co., Malabar, FL. Westernhagen, H. von, P. Cameron, V. Dethlefson, and D. Janssen. 1989. Chlorinated hydrocarbons in North Sea whiting (Merlanguis merlangus) and effects on reproduction. I: Tissue burden and hatching success. Halgolander Meeresuntersuchungen 43:45-60. Wolf, D. A., and T. P. O'Connor (eds.). 1988. Oceanic processes in marine pollution. Vol. 5: Urban waste in coastal marine environments. R. E. Krieger Publ. Co., Malabar, FL, 273 p. Word, J. Q., C. D. Boatman, C. C. Ebbesmeyer, R. E. Finger, S. Fischnaller, and Q. J. Stober. 1990. Vertical transport of effluent material on the sur- face of marine waters. In D. J. Baumgartner and 1. W. Duedall (eds.), Oceanic processes in marine pollu- tion. Vol. 6: Physical chemical processes: transport and transformation. R. E. Krieger Publ. Co., Malabar, FL, 33-149 p. Ziskowski, J. J., and R. A. Murchelano. 1975. Fin erosion in winter flounder. Mar. Pollut. Bull. 6:26-29. Abstract. -We examined the av- erage age at attainment of sexual maturity (ASM) and several other reproductive parameters for evidence of density compensatory responses in two stocks of the spotted dolphin, Stenella attenuata. The northern off- shore and southern offshore stocks were compared because both have been exploited by the yellowfin tuna purse-seine fishery, but at different rates. We predicted decreasing trends in the ASM and increasing trends in the pregnancy rate for each stock because both have declined in abundance. A lower ASM and a higher pregnancy rate were pre- dicted for the sample from the north- ern offshore stock because it has been exploited to a greater extent than the southern offshore stock. No statistically significant trends were found in the ASM, but the increase in the proportion of sexually mature females simultaneously pregnant and lactating for the northern off- shore stock and the decrease in the proportion of mature females in the population for the southern offshore stock over time were statistically significant. The mean estimate of ASM was significantly higher for the northern offshore stock, 11.1 years (SE=0.236), than for the southern offshore stock, 9.8 years (SE=0.264). No significant differences between stocks were found in the mean esti- mates of reproductive parameters. Our analyses of temporal trends in several biological parameters did not provide conclusive evidence for compensatory responses having oc- curred, and therefore some possible explanations are considered. Comparison of age at sexual maturity and other reproductive parameters for two stocks of spotted dolphin, Stenella attenuata Susan J. Chivers Al C. Myrick Jr. Southwest Fisheries Science Center, National Marine Fisheries Service, NOAA PO Box 27 1 , La Jolla, CA 92038-027 I Manuscript accepted 17 June 1993. Fishery Bulletin 91:611-618 (1993). As population abundance declines, compensatory responses, such as in- creases in survival and pregnancy rates and a decrease in the average age at attainment of sexual matu- rity (ASM), are thought to occur (Eberhardt, 1977; Eberhardt and Siniff, 1977). Each situation must be carefully evaluated to determine whether compensatory responses will be detected because the magnitude of the parameter response and the range of population densities over which measurements are made are likely to be species specific, environ- mentally variable, and subject to pa- rameter measurement error (Good- man, 1981; Fowler, 1981, 1988; DeMaster, 1984a; York, 1987). For species that are difficult to observe, unbiased estimates of survival rates are difficult to obtain, precluding their use as potential biological indi- ces (Barlow and Boveng, 1991). Re- productive rates, however, are gen- erally more easily measured and, if their relationship to population den- sity is determined, may provide indi- ces of compensatory responses for populations (Perrin and Donovan, 1984). For instance, Barlow (1985) reported that in the spotted dolphin (Stenella attenuata), the fraction of mature females pregnant, and the fraction of mature females simulta- neously pregnant and lactating, may correlate well with per capita re- source availability because these pa- rameters were found to be relatively free of collection biases. The identifi- cation of biological indices (i.e., bio- logical parameters that correlate with population abundance) would provide a useful management tool to distin- guish between trends in population abundance resulting from exploita- tion or changes in environmental con- ditions (Hanks, 1981; Gerrodette and DeMaster, 1990). The spotted dolphin is an ideal choice for this analysis because a lengthy time series of biological and population abundance data are avail- able. This species is used as a cue by purse-seine vessel operators to find yellowfin tuna in the eastern tropi- cal Pacific (ETP) (Perrin, 1969). Dol- phins are incidentally killed during fishing operations, and the spotted dolphin population has had the larg- est number of animals killed in al- most every year since 1959 (Smith, 1983; DeMaster et al., 1992). Two stocks of spotted dolphin, the north- ern offshore and southern offshore stocks, are recognized as separate management units in the ETP (Perrin et al., 1985). During the first decade of the purse-seine fishery, the 1960s, the large numbers of dolphins killed caused the abundance of these two stocks to decline. The northern offshore spotted dolphin was esti- mated to be between 35% and 50% of their pre-exploitation abundance (circa 1959) by 1979 (Smith, 1983), while the southern offshore stock was estimated to be at 92-98% of its 61 612 Fishery Bulletin 91|4). 1993 pre-exploitation size (circa 1973) in 1979 (Smith1). Since 1976, both stocks of spotted dolphin are believed to have declined significantly in abundance (Anganuzzi etal., 1992). As potential biological indices, we estimated ASM and associated reproductive parameters for two stocks of the spotted dolphin. Although for long-lived mam- mals, like the spotted dolphin, population growth rates are relatively insensitive to changes in ASM and re- productive rates (Eberhardt and Siniff, 1977; Fowler, 1981; Reilly and Barlow, 1986), these parameters have been correlated with changes in population abundance for other marine mammal species (Fowler, 1987). For instance, a decrease in ASM was correlated with a decline in population abundance for baleen whale stocks (Lockyer, 1984). Similarly, a decrease in ASM for crabeater seals was correlated with an increase in the per capita availability of food resources resulting from the exploitation and subsequent reduction in size of baleen whale stocks (Bengtson and Laws, 1985). Like- wise, ASM for the striped dolphin of the western Pa- cific declined in response to exploitation and reduced population abundance (Kasuya, 1985). Based on the predictions of population responses to changes in den- sity (i.e., increases in ASM and decreases in pregnancy rates as population abundance increases) presented by Eberhardt (1977) and Eberhardt and Siniff (1977), we predicted a decline in the ASM for both the north- ern offshore and southern offshore stocks of spotted dolphin after 1974 as well as a lower ASM for the more heavily exploited northern offshore stock. Simi- larly, a higher fraction of sexually mature females preg- nant and those simultaneously pregnant and lactating should be observed for the northern offshore stock. Methods Since 1968, when the National Marine Fisheries Ser- vice (NMFS) first placed observers aboard U.S. vessels to observe fishery activities, life history data has been collected from all cetacean species incidentally killed in the ETP yellowfin tuna purse-seine fishery. Begin- ning in 1974, life history data collection procedures were standardized, and the original sampling scheme that selectively collected large, female specimens was replaced by a random sampling scheme that selected the first available dead dolphins brought aboard (Perrin and Oliver, 1982). In 1979, the Inter-American Tropi- cal Tuna Commission (IATTC) joined the NMFS in 'Smith, T. D. (ed.). 1979. Report of the status of porpoise stocks workshop (August 27-31, 1979, La Jolla, California). U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. LJ-79-41, 120 p. placing observers aboard U.S. vessels and collecting life history data. The offshore stocks of spotted dolphin are distin- guished geographically as being north or south of 1 south latitude (Perrin et al., 1985). The specimens used in our study were collected from these areas, but the sample of northern offshore stock animals was addi- tionally restricted to those animals collected west of 120° west latitude (Fig. 1). The northern offshore stock was sub-sampled because exploitation has not been spatially uniform. The yellowfin tuna purse-seine fishery began primarily as a coastal fishery in 1959 and gradually expanded farther offshore. The Commission's Yellowfin tuna Regulatory Area (CYRA) was established in 1968 by the IATTC and provides a useful boundary that separates the inshore fishery area from the westernmost region of the ETP yellowfin tuna fishery (Peterson and Bayliff, 1985). The area west of the CYRA was not fished until the late 1960s, and then only for a few months a year. Exploitation in this region has been significantly less than inside the CYRA but much more than in the southern area which has been fished only sporadically since the early 1970s (Punsly, 1983). Although spotted dolphin move along the 10°N latitude as well as from the south to the west (Perrin et al., 1979a; Au and Perryman, 1985; Reilly, 1990), owing to the large size of the area, inter- change between areas is probably limited. A standard- ized data collection scheme was established at about the same time as the fishery expanded into the west- ern and southern regions of the ETP, and therefore we anticipated that these two distinct geographic regions would provide the best opportunity for testing poten- tial biological indices. Female spotted dolphins with complete life history data (i.e., geographic position, total body length, teeth collected, and both ovaries examined) collected between 1974 and 1988 were selected for this study. Few speci- mens have been collected from the southern offshore stock, and therefore all available specimens were pre- pared. Specimens from the northern offshore stock were selected randomly by year and the sample sizes dis- tributed as evenly as possible between years so that annual estimates of ASM could be calculated. A maxi- mum of 50 specimens per year was the target for se- lecting specimens from the northern offshore stock sample because Hohn (1989) showed that a sample size of 50 accurately estimated ASM and reduced the variance from an estimate based on a sample size of 25. Thin sections of the teeth were prepared and aged by one of us (ACM) with techniques described in Myrick et al. (1983). Prior to estimating age, all specimens were numerically coded in random order to disguise the specimen number so that no inference could be made about the stock or collection year for any speci- Chivers and Mynck: Comparison of age at sexual maturity for Stenella attenuata 613 Figure 1 The collection location and sample size of female spotted dolphin specimens selected for this study summarized by 2°-degree square block. The sample of northern offshore spotted dolphins was selected from north of 1° south and west of the Commission's Yellowfin tuna Regulatory Area I the dashed line), and the sample of southern offshore spotted dolphins was selected from south 1° south. men. The associated life history data were also with- held until each specimen was aged. ASM and its variance were calculated with the fol- lowing formula (DeMaster, 1984b; Hohn, 1989). ASM=x + Xa s2 = X {[(p; * (1-pMnr-l)}, Other reproductive parameters examined include the fraction of the females mature, and the fraction of ma- ture females that were either pregnant or lactating, or simultaneously pregnant and lactating. Trends in these parameters between 1974 and 1988 were tested by using a chi-square test for linear trends (Snedecor and Cochran, 1973), and the mean estimates for each pa- rameter, all years combined, were tested for statisti- cally significant differences with a chi-square test. where, x = age of the youngest mature animal, j = age class of the oldest immature animal, p, = proportion immature in age class i, n, - sample size for age class i . Because population abundance estimates are not avail- able for the geographic areas from which our samples were collected, we substituted year as a correlate for evaluating potential biological indices. We tested for temporal trends in ASM with a linear regression to determine whether the slope was statistically signifi- cantly different from zero. In addition, data for each stock were pooled for all years, and the null hypoth- esis of no difference between estimates of ASM for the two stocks was tested with a Student's ^-test. Results Average age at attainment of sexual maturity ASM was estimated for each year with sufficient data available (Table 1): 1976-1982 and 1984-1988 for the northern offshore stock sample, and 1976, 1979, and 1982 for the southern offshore stock sample. The re- gression line fit to the data for the northern offshore stock was not statistically significantly different from zero. We did not fit a regression line to the data from the southern offshore stock because only three annual estimates were available. The pooled estimate of ASM for each stock was 11.1 years (SE=0.236; n=520) for 614 Fishery Bulletin 91(4). 1993 Table 1 Number of samples by year and state of sexual maturity l immature or mature) for the northern offshore and southern offshore stocks of spotted dolphin l sed to estimate the aver- age age at attainment of sexual ma urity. Year Offshore stock Northern Southerr Immature Mature Immature Mature 74 0 0 0 0 75 0 0 0 5 76 13 33 41 101 77 16 28 2 6 78 17 27 1 23 79 18 25 16 49 80 15 28 2 9 81 13 24 6 5 82 12 20 42 68 83 9 9 11 11 84 11 27 1 4 85 17 29 0 0 86 18 28 0 0 87 22 26 0 0 88 14 21 0 0 Total 195 325 122 281 the northern offshore spotted dolphin and 9.8 years (SE=0.264; n=403) for the southern offshore spotted dolphin (Fig. 2). This difference in the ASM was sta- tistically significant (t= 14.32, df=13, P< 0.001). The ™ 15 b ,3 DC I- 12- < 11 —I < 3 10 X LLI 9 Li. O 8-| LU (J < ' Ml ' + YEAR Figure 2 Estimates of the average age at attainment of sexual matu- rity i ASM i by stock for the offshore spotted dolphin with 1 SE. If no SE bar was plotted, then there were no overlapping indeterminate age classes. Circles represent estimates of ASM for the northern offshore stock and triangles are the esti- mates for the southern offshore stock. The pooled estimate of ASM for the northern offshore stock IN) is 11.1 years (solid line) and 9.8 years (dashed line I for the southern offshore stock IS). range in age of animals attaining sexual maturity was largest for the northern offshore stock. The youngest mature animal was 7 years, and the oldest immature was 16 years. The same range for the southern off- shore stock was 6 to 11 years (Fig. 3). Reproductive rates The statistically significant trends in the reproductive parameters examined were an increase in the propor- tion of mature females simultaneously pregnant and lactating in the northern offshore stock, the predicted compensatory response, and a decrease in the propor- tion of mature specimens in the sample from the south- ern offshore stock. When the data for all years were pooled, no statistically significant differences between the mean parameter estimates for the northern offshore and southern offshore stocks were detected (Table 2). ( ) 5 10 15 2C 25 30 35 40 NORTHERN SPOTTED DOLPHIN 40 - n ■ O 30- z; UJ O S 20- in 10 - r 0 - J 1 . 1 llllll,^p^1 AGE (years) Figure 3 Age-frequency distribution of the northern offshore and south- ern offshore spotted dolphin specimens aged. The open bars represent the sexually immature females in the sample, and the solid bars represent the sexually mature females in the sample. Chivers and Myrick Comparison of age at sexual maturity for Stenella attenuata 615 Table 2 The mean proportion of females by repro- ductive condition for the northern offshore and southern offshore stocks of spotted dol- phin: the proportion of females in the sample that were sexually mature, and the propor- tion of sexually mature females that were pregnant or lactating, or simultaneously pregnant and lactating. Stock Reproductive condition northern southern Mature Pregnant Lactating Pregnant and lactating 0.62 0.33 0.55 0.06 0.69 0.40 0.47 0.06 Discussion Published evidence of density dependence in dolphin species is rare, and results are inconsistent. For example, in Japan, the striped dolphin, Stenella coeruleoalba, is heavily exploited, and the length of the lactation period is reported to have de- clined (Kasuya and Miyazaki2). For ex- ploited spotted dolphins in the western Pacific, a decline in the ASM from 10.1 years in the 1961-63 cohorts to 8.6 years in the 1964-66 cohorts was reported, al- though the regression coefficient was not statistically significant (Kasuya, 1985). The decline in ASM for female striped dolphin from 9.7 years for the 1956-58 cohorts to 7.4 years in the 1968-70 co- horts was statistically significant (Ka- suya, 1985). On the otherhand, for the spinner dolphin (S. longirostris) in the ETP, the less exploited southern white- belly stock has a shorter lactation period than the more heavily exploited eastern spinner dolphin, the opposite of what would be expected for a compensatory re- sponse. The annual pregnancy rate, how- ever, is lowest for the southern whitebelly stock and highest for the more heavily 2Kasuya, T., and N. Miyazaki. 1975. The stock of Stenella coeruleoalba off the Pacific Coast of Ja- pan. Pap. ACMRR/MM/SC-25 pres. at FAO Scien- tific Consultation of Marine Mammals, Bergen, Feb- ruary, 1976, 36 p. exploited eastern stock, as expected for density compensation (Perrin and Henderson, 1984). Barlow (1985) also reported differences in the proportion of mature female spotted dolphins (S. attenuata) that were pregnant between the eastern and western Pacific. The lowest proportion of mature females pregnant was observed in the more heavily exploited population, again, contrary to predictions based on the reported exploitation rates and estimates of population abun- dance for the ETP spotted dolphin. Also, for ETP spotted dolphins, the proportion of lactating mature females significantly increased from 46% in 1973 to 69% in 1981 (P<0.05) presumably in response to exploitation; no other statistically significant trends in reproduc- tive parameters were found (Myrick et al., 1986). Comparisons with other studies are difficult because in most cases the degree of exploitation relative to K has not been quantified, and techniques to estimate age and ASM vary. In our study, statistically significant temporal trends were not detected for ASM but were detected for some of the reproductive parameters. The increase in the proportion of mature females simultaneously pregnant and lac- tating for the northern offshore stock and the decrease in the pro- portion of mature females in the population for the southern stock both suggest that the populations are growing. The fact that we did not observe the predicted compensatory responses in ASM may be due to one or more factors, including 1) too few data, 2) parameter estimation or measurement error, 3) a time lag in the response, or 4) biological differences between the stocks. Too few data and envi- ronmental periodicity (e.g., El Nino) may increase the variability in annual estimates of parameters, and both factors would reduce the ability to detect differences in potential compensatory responses (Goodman, 1984). The order in which parameters respond to changes in population density is not known, but time lags in responses may be expected owing to the late age at which these animals reach sexually maturity and the multi-year breeding cycle characteristic of the spotted dolphin (Goodman, 1981). Data for the southern offshore stock of spotted dolphin, in par- ticular, were limited and precluded testing for trends in ASM or the selected reproductive parameters. Although we selected our sample size, a maximum of 50 specimens per year for the northern offshore stock, on the basis of earlier work describing the effect of sample size on the variance for estimates of ASM (Hohn, 1989), our results indicate that larger sample sizes would be required to reduce the variance of the estimates in order to detect small, but biologically significant, changes (e.g., a 1.5-year difference) in ASM. Sample sizes of fewer than 50 per estimate of ASM resulted in too few specimens in the indeterminate age classes (i.e., those age classes with both sexually immature and mature animals) to estimate ASM accurately. Another concern when analyzing data collected from a kill is whether the age distribution is representative of the population. All immature age classes appear to be underrepresented in both stocks, especially age classes between 10 and 15 years of the northern offshore stock (Fig. 3). This phenomenon was noted in a previous aging study of ETP spotted dolphin but occurred about five age classes earlier than the one we observed (Barlow and Hohn, 1984). This difference may be a result of different readers estimating age. 616 Fishery Bulletin 91(4), 1993 although a Kolmogorov-Smirnov test to compare the age-frequency distributions of sexually immature and sexually mature animals from the two studies con- cluded that the samples were likely drawn from the same population. The underrepresentation of the immature age classes may be a result of segregation in the population by age and sex, fairly typical in large mammal populations, and may affect the estimation of ASM. We investigated the effect of this un- derrepresentation by assuming a stable age distribu- tion and by apportioning the animals in the indeter- minate age classes by sexual maturity under different assumptions of sexual maturity for animals in those age classes, and then calculated ASM. On the basis of segregation by age and sex observed in other species of large mammals, we predicted that the most reason- able scenario for reconstructing the age distribution is that the "missing" animals would be sexually imma- ture. If this indeed were the case, we have underesti- mated ASM. Under this and other scenarios we inves- tigated, estimates of ASM were consistently over- or under-estimated for both stocks. Therefore, possible sampling biases were unlikely to be responsible for the observed differences in ASM between the northern offshore and southern offshore stocks of spotted dolphin. The hypotheses we tested required that several im- plicit assumptions be made, including: 1) constant data collection biases, 2) minimal interchange of animals between areas, 3) consistent estimation of specimen age, 4) linear compensatory response with change in population abundance, 5) constant carrying capacity (K), and 6) equivalent initial life history parameters for the two stocks. The first three of the assumptions are believed to hold reasonably well. Barlow (1985) tested several of the same life history parameters we examined and found them to be relatively insensitive to a number of potential data collection biases; no ma- jor changes in the collection of life history data were made after 1974, thus supporting constant data collec- tion bias as a reasonable assumption. Furthermore, the fishery operates in the same areas from year-to- year at approximately the same time of year. In fact, 60% of our sample from the northern offshore stock was collected between July and September, and a fur- ther 38% was collected between April and June; in the south, 57% of the sample was collected between Janu- ary and March, and 27% was collected between Octo- ber and December. This pattern of sample collection was consistent from year-to-year, and therefore would not affect the pooled or annual estimates of ASM or reproductive parameters. Geographic variability has been noted in earlier studies of spotted dolphin life history data (Hohn and Hammond, 1985; Barlow, 1985); thus selecting specimens from discrete areas as we did reduces the potential for geographic variability to ob- scure population compensatory responses. Similarly, estimating the age of all specimens at one time and in the blind ensured that age estimates were made as consistently as possible. Information about the valid- ity of assumptions 4) and 5) is not currently available, but violations of the assumptions may provide expla- nations for our results. There is evidence for differ- ences in the morphological and life history character- istics between the northern offshore and southern offshore stocks of spotted dolphin (see Perrin et al., 1976, 1979b, 199F; Barlow, 1985; Hohn and Hammond, 1985; Myrick et al, 1986; Bright and Chivers4) that are likely correlated with the different oceanographic environments of the areas inhabited by these stocks (Au and Perryman, 1985; Reilly, 1990). We did not find conclusive evidence for compensa- tory responses in these stocks of spotted dolphin as only one reproductive parameter for each stock showed a statistically significant trend. However, observed trends in both parameters suggest the populations are below K and declining. Our comparison of ASM's sug- gest further biological differences between the north- ern offshore and southern offshore stocks or popula- tions. Currently, the order of compensatory responses and their dynamics are not known. In order to be use- ful as a biological index, a measure of the status of the population must be known and the dynamics of the parameter over a wide range of population densities quantified (Fowler and Siniff, 1992). Acknowledgments We thank the NMFS and IATTC biological technicians who collected the life history data while aboard the tuna purse-seiners, often under difficult working con- ditions, and the Southwest Fisheries Science Center personnel that coded and edited the data for use. Priscilla Akin and Andrea Bright prepared the teeth for estimating the age of the specimens. We thank Jay Barlow, Douglas DeMaster, Andrew Dizon, Christina 'Perrin. W. F„ G. D. Schnell. D. J. Hough, J. W. Gilpatrick, and J. V. Kashiwada. 1991. Re-examination of geographical cranial variation in the pantropical spotted dolphin (Stenella attenuata) in the east- ern Pacific. Dep. Cummer., NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. LJ- 91-39, 46 p. "Bright, A. M., and S. J. Chivers. 1991. Post-natal growth rates: a comparison of northern and southern stocks of the offshore spotted dolphin. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. LJ- 91-30, 24 p. Chivers and Myrick: Comparison of age at sexual maturity for Stenella attenuate 617 Lockyer, Steve Reilly, and three anonymous reviewers whose comments on earlier drafts greatly improved the manuscript. Literature cited Anganuzzi, A. A., K. L. Cattanach, and S. T. Buckland. 1992. Relative abundance of dolphins associated with tuna in the eastern tropical Pacific, estimated from preliminary tuna vessel sightings data for 1990. Rep. Int. Whaling Comm. 42:541-546. Au, D. W. K., and W. L. Perryman. 1985. Dolphin habitats in the eastern tropical Pacific. Fish. Bull. 83:623-643. Barlow, J. 1985. Variability, trends, and biases in reproductive rates of spotted dolphins (Stenella attenuate). Fish. Bull. 83:657-669. Barlow, J., and A. A. Hohn. 1984. Interpreting spotted dolphin age distribu- tions. NOAA-Technical Memorandum-NMFS-48, 22 p. Barlow, J., and P. L. Boveng. 1991. Modeling mortality for marine mammal popula- tions. Mar. Mamm. Sci. 7:50-65. Bengtson, J. L., and R. M. Laws. 1985. Trends in crabeater seal age at maturity: an in- sight into Antarctic marine interactions. In W. R. Siegfried, P. R. Condy, and R. M. Laws (eds.), Ant- arctic nutrient cycles and food webs, p. 669- 675. Springer- Verlag, Berlin. DeMaster, D. P. 1984a. A review of density dependence in marine mammals. /;; B. R. Melteff and D. H. Rosenberg (eds. I, Proceedings of the Workshop on Biological In- teractions among Marine Mammals and Commercial Fisheries in the Southeastern Bering Sea. Alaska Sea Grant Rep. 84-1, 300 p. 1984b. Review of techniques used to estimate the av- erage age at attainment of sexual maturity in marine mammals. Rep. Int. Whaling Comm. Special Issue 6:175-179. DeMaster, D. P., E. F. Edwards, P. Wade, and J. E. Sisson. 1992. Status of dolphin stocks in the eastern tropical Pacific. In D. R. McCullough and R. H. Barrett (eds.). Wildlife 2001: populations, p. 1038-1050. Elsevier Applied Science. NY. Eberhardt, L. L. 1977. Optimal policies for conservation of large mam- mals. Environ. Conserv. 4:205-212. Eberhardt, L. L., and D. Siniff. 1977. Population dynamics and marine mammal man- agement policies. J. Fish. Res. Board Can. 34:183- 189. Fowler, C. W. 1981. Density dependence as related to life history strategy. Ecology 62:602-610. 1987. A review of density dependence in populations of large mammals. Current Mammalogy 1:401-441. 1988. Population dynamics as related to rate of in- crease per generation. Evol. Ecol. 2:197-204. Fowler, C. W., and D. B. Siniff. 1992. Determining population status and the use of biological indices in the management of marine mammals. In D. R. McCullough and R. H. Barrett (eds.), Wildlife 2001: populations, p. 1025-1037. Elsevier Applied Science, NY. Gerrodette, T., and D. P. DeMaster. 1990. Quantitative determination of optimum sustain- able population level. Mar. Mamm. Sci. 6:1-16. Goodman, D. 1981. Life history analysis of large mammals. In C. W Fowler and T. D. Smith (eds.). Dynamics of large mammal populations, p. 415^136. John Wiley & Sons, NY. 1984. Statistics of reproductive rate estimates and their implications for population projection. Rep. Int. Whaling Comm. Special Issue 6:161-173. Hanks, J. 1981. Characterization of population condition. In C. W. Fowler and T. D. Smith (eds.), Dynamics of large mammal populations, p. 47-73. John Wiley & Sons, NY. Hohn, A. A. 1989. Variation in life history traits: the influence of introduced variation. Ph.D. diss., Univ. California, Los Angeles, 123 p. Hohn, A. A., and P. K. Hammond. 1985. Early postnatal growth of the spotted dolphin, Stenella attenuata, in the offshore eastern tropical Pacific. Fish. Bull. 83:553-566. Kasuya, T. 1985. Effect of exploitation on reproductive param- eters of the spotted and striped dolphins off the Pacific coast of Japan. Sci. Rep. Whales Res. Inst. 36:107- 138. Lockyer, C. 1984. Review of baleen whale (Mysticeti) reproduction and implications for management. Rep. Int. Whal- ing Comm. Special Issue 6:27-50. Myrick, A. C, Jr., A. A. Hohn, P. A. Sloan, M. Kimura, and D. Stanley. 1983. Estimating age of spotted and spinner dolphins (Stenella attenuata and S. longirostris) from teeth. NOAA-Technical Memorandum-NMFS-30, 17 p. Myrick, A. C. Jr., A. A. Hohn, J. Barlow, and P. A. Sloan. 1986. Reproductive biology of female spotted dolphins, Stenella attenuata, from the eastern tropical Pacific. Fish. Bull. 84:247-259. Perrin, W. F. 1969. Using porpoise to catch tuna. World Fishing 18:42^5. Perrin, W. F., and C. W. Oliver. 1982. Time/area distribution and composition of the incidental kill of dolphins and small whales in the U.S. purse-seine fishery for tuna in the eastern tropi- 61! Fishery Bulletin 91(4). 1993 cal Pacific, 1979-80. Rep. Int. Whaling Comm. 32:429-444. Perrin, W. F., and G. P. Donovan. 1984. Report of the workshop. Rep. Int. Whaling Comm. Special Issue 6:1-24. Perrin, W. F., and J. R. Henderson. 1984. Growth and reproductive rates in two popula- tions of spinner dolphins, Stenella longirostris, with different histories of exploitation. Rep. Int. Whaling Comm. Special Issue 6:417-430. Perrin, W. F, J. M. Coe, and J. R. Zweifel. 1976. Growth and reproduction of the spotted porpoise, Stenella attenuata, in the offshore eastern tropical Pacific. Fish. Bull. 74:229-269. Perrin, W. F, W. E. Evans, and D. B. Holts. 1979a. Movements of pelagic dolphins (Stenella spp.) in the eastern tropical Pacific as indicated by results of tagging, with summary of tagging operations, 1969- 76. NOAA-Technical Memorandum-NMFS-SSRF- 737, 14 p. Perrin, W. F., P. A. Sloan, and J. R. Henderson. 1979b. Taxonomic status of the 'southwestern stocks' of spinner dolphin Stenella longirostris and spotted dolphin S. attenuata. Rep. Int. Whaling Comm. 29:175-184. Perrin, W. F, M. D. Scott, G. J. Walker, and V. L. Cass. 1985. Review of geographical stocks of tropical dol- phins (Stenella spp. and Delphinus delphis) in the eastern tropical Pacific. NOAA-Technical Memoran- dum-NMFS-28, 28 p. Peterson, C. L., and W. H. Bayliff. 1985. Organization, functions, and achievements of the Inter-American Tropical Tuna Commission. Inter- Am. Trop. Tuna Comm., La Jolla, CA. Special Report No. 5, 56 p. Punsly, R. G. 1983. Estimation of the number of purse-seiner sets on tuna associated with dolphins in the eastern Pa- cific Ocean during 1959-1980. Inter-Am. Trop. Tuna Comm. Bull. 18:229-299. Reilly, S. B. 1990. Seasonal changes in distribution and habitat differences among eastern tropical pacific dol- phins. Mar. Ecol. Prog. Ser. 66:1-11. Reilly, S. B., and J. Barlow. 1986. Rates of increase in dolphin population size. Fish. Bull. 84:527-533. Smith, T. D. 1983. Changes in size of three dolphin (Stenella spp.) populations in the eastern tropical Pacific. Fish. Bull. 81:1-13. Snedecor, G. W., and W. G. Cochran. 1973. Statistical methods, sixth ed. Iowa State Univ. Press, Iowa. York, A. E. 1987. On comparing the population dynamics of fur seals. In J. P. Croxall and R. L. Gentry (eds.). Sta- tus, biology, and ecology of fur seals, p. 133-140. Pro- ceedings of an international symposium and workshop, Cambridge, England, 23-27 April 1984. Abstract. -Restriction endonu- clease analysis of mitochondrial cy- tochrome b and 12S ribosomal RNA (rRNA) gene fragments amplified by polymerase chain reaction (PCR) was applied to thirteen western At- lantic snapper species (the genera Lutjanus, Ocyurus and Rhom- boplites), in order to assess reliabil- ity of this method for genetic spe- cies and stock identification studies. Eight species (L. apodus, L. buccanella, L. campechanus, L. cyanopterus, L. griseus, L. synagris, L vivanus and R. aurorubens) could be identified by haplotype analysis on either single or both fragments, while the others (L. analis, L. jocu, L. mahogoni, L. purpureus and O. chrysurus) were not separated from one another because of overlapping or identical haplotypes observed be- tween species. High nucleon diver- sity estimates (/; = 40-80f7r ) observed within four species (L. campechanus, L. cyanopterus, L. griseus and L. jocu) indicated that sufficient in- traspecific polymorphism could be detected by the PCR-[RFLP] restric- tion fragment length polymorphism analysis making it a useful method for investigating genetic stock structure. PCR-RFLP analysis on thirteen western Atlantic snappers (subfamily Lutjaninae): a simple method for species and stock identification Seinen Chow National Research Institute of Far Seas Fisheries Ondo 5-7-1. Shimizu, Shizuoka 424, Japan M. Elizabeth Clarke Patrick J. Walsh Division of Marine Biology and Fisheries Rosenstiel School of Marine and Atmospheric Science, University of Miami. 4600 Rickenbacker Causeway. Miami. FL 33149-1098 Manuscript accepted 14 May 1993. Fishery Bulletin 91:619-627 ( 1993). In order to manage fisheries re- sources, it is essential to know the recruitment mechanisms of a given species. Lutjanidae is one of the larg- est teleostean families, comprising 4 subfamilies, 17 genera, and 103 spe- cies (Allen, 1985). More than a half of the lutjanid species belong to the subfamily Lutjaninae in which 14 species under three genera {Lutjanus, Ocyurus and Rhomboplites) have been described in the western Atlan- tic Ocean. All are fine food fishes and very important for commercial and recreational fisheries, and consider- able concern has been focused on the need to understand recruitment in order to manage these reef species (Huntsman, 1981; Doherty, 1983; Roberts and Polunin, 1991). However, the close morphological similarity of larvae among lutjanid species, espe- cially within the subfamily Lutjan- inae, has made specific identification of larvae a difficult task (Richards, 1985; Leis, 1987; Richards and Lin- deman, 1987). Allozyme electro- phoresis has proven useful for identifying some fish species even at their embryonic or larval stages (Morgan, 1975; Smith and Crossland, 1977; Mork et al„ 1983; Graves et al., 1989). Since DNA is virtually the same in any cell type of an individual and can be extracted from ethanol- preserved specimens, restriction frag- ment analyses of DNA is becoming a preferred method to investigate variation and relationships between fish species (Billington and Hebert, 1991). Distinct restriction fragment patterns between three bass species of the genus Paralabra.x could be de- tected in ethanol-preserved indi- vidual eggs and larvae via Southern blot analysis (Graves et al., 1990). Although conventional mitochondrial DNA (mtDNA) methods are power- ful for detecting variation in restric- tion fragment length, intensive DNA analysis would be difficult especially on very small samples, such as fish embryos or larvae, which offer a very small amount of DNA. To overcome this problem, the use of the poly- merase chain reaction (PCR) method, which can amplify DNA sequences more than 10-million fold (Saiki et al., 1988), is neccessary In this paper, we report amplification of two mitochondrial genes (cytochrome b and 12S rRNA) using fresh or frozen samples, ethanol-preserved embryos and larvae, and alcohol- preserved museum samples of thir- teen western Atlantic snapper spe- cies, and results of restriction frag- ment length polymorphism (RFLP) 619 620 Fishery Bulletin 91|4). 1993 analysis on these two DNA fragments within and be- tween species. Materials and methods Sample collection The species used and sources of collection are listed in Table 1. All of the fresh or frozen specimens were col- lected in the Miami and Key West, FL, area during 1990 and 1991. Fertilized eggs and hatched larvae of Lutjanus synagris were raised at 28°C in 380-L tanks and fed a combination of cultured rotifers and wild- caught zooplankton. The eggs (4 to 6 hours after in- semination) and larvae (3 and 9 days old) were fixed with 95% ethanol and kept for two months at room temperature. Two eggs and two larvae were randomly chosen. All of the museum specimens were obtained from the Ichthyology Museum at the University of Mi- ami. The time since preservation ranged from 5 to 42 years and it is likely that most were fixed with formal- dehyde first and then transferred to ethanol or iso- propyl alcohol (C. R. Robins, Professor, Rosenstiel School of Marine and Atmospheric Science, Division of Marine Biology and Fisheries, Univ. Miami, 4600 Rickenbacker Causeway, Miami, FL 33149-1098, pers. commun.). How- ever, exact records on preservation methods were not available. Samples are labelled as to source, for example, Lcyl and LcyMl designate Lutjanus cyanopterus fresh or frozen specimen No. 1 and L. cyanopterus museum specimen No. 1, respectively. DNA extraction Fresh or frozen specimens Mitochondria-enriched or total genomic DNA samples were prepared from fresh or frozen specimens by the method of Chapman and Powers ( 1984) with slight modification. Museum specimens A piece of white muscle (0.1- 0.2 g) was dissected from each individual, shredded and rinsed in TEK buffer (50mM Tris, 10 mM EDTA, Table 1 Sample species of snappers (Lutjaniae: Lutjanus, Ocyurus and Rhomboplites) and source of collection. Fresh or frozen samples Museum samples Species Abbreviation locality and date label, locality, and date L. analis La Miami, FL, '90/91 < 6)* L. apodus Lap Miami, FL, '90/91(7) L. buccanella Lb Miami, FL, '91(4l L. campechanus Lc NA UMML 27092, 28°N, 80°W, '66 (M2)** UMML 27093, 28°N, 80°W, '66 (Ml ) UMML 27096, U°N, 60°W, '66 (M3) UM/M2422, Gulf of Mexico, '58 (M4-5) L. cyanopterus Ley Key West, FL, '91(1) UMML15923, Virgin Id., '59 (Mil UM/M4868, Bahama, '56 (M2) L. griseus Lg Miami, FL, '90/91 (9) L.jocu Lj NA Cat.No. 1944, Puerto Rico, '57 (Ml I Cat.No. 620, Cutler, FL, '49 (M2-4) UMML34424, Miami, FL, '86 (M6) UM/M5660. Miami, FL, '64 (M7) L. mahngoni Lm Key West, FL, '91(2) UMML7523, Knight Key, FL, '57 (Ml) UM/M4865, Miami.FL '50 (M2) L. purpureas Lp NA UM/M6093, Venezuela, '65 (Ml-2) St.5708, Venezuela, '65 (M3i St.5658, Venezuela, '65 (M4) UM/M6086, Venezuela, '65 (M5-6) L. synagris Ls Miami, Key West, FL, '90 (7) L. vivanus Lv Miami, FL, '90/91(5) STA5739, Panama, '65 (Mil < >. chrysurus Oc Miami. Key West, FL, '90 (8) R. aurorubens Ra Miami, FL, '90(6) NA=not available. =number of individuals examinee **=museum specimens were labelled, for example, as LcMl designating Lutjanus campechanus museum speci- men No. 1. All museum samples were obtained from the Ichthyological Museum of the University of Miami. UMML is the catalogue number for each individual; values in parentheses indicate lots in which several individuals were contained. Chow et al PCR-RFLP analysis on thirteen western Atlantic snappers 621 1.5% KC1, pH 7.5) for 10 minutes at room tempera- ture. The muscle fibrils were finely minced and ho- mogenized in 5-10 mL of ice-cold TEK buffer by means of a motor-driven, glass-teflon homogenizer. Five or more strokes of the homogenizer were neccessary to grind the toughest muscle tissue. The homogenate was transferred to a plastic centrifuge tube (50-mL capac- ity), and SDS and proteinase K were added as de- scribed in Chapman and Powers (1984). The sample was then incubated at 65°C for 2 hours or more and vigorously shaken at intervals. This crude sample was phenol-chloroform extracted and ethanol precipitated as described in Chapman and Powers (1984). The amount of DNA was usually too small to be seen as a pellet, but was rehydrated in 500 pi of TE buffer (ImM EDTA, 10 mM Tris-HCl, pH 8.0) and transferred to a 1.5 mL microcentrifuge tube. To concentrate this DNA sample, ethanol precipitation was repeated and the pellet was rehydrated in 10 ul of TE buffer. Embryos and larvae Each specimen was transferred to a 1.5-mL microcentrifuge tube. TEK buffer was added and decanted several times to rinse specimens and to remove ethanol. The specimen was then ho- mogenized in 500 pi of TEK buffer and total DNA was isolated as described above. The pellet was rehydrated in 10 pi of TE buffer. Amplification of mitochondrial cytochrome b and 12S rRNA genes The two pairs of primers used that targeted cytochrome b and 12S rRNA genes were abbreviated forms of those described by Kocher et al. (1989). The nucleotide se- quences of each set of primers were as follows: cyto- chrome 6, 5-GCTTCCATCCAACATCTCAGCATGATG- 3' and 5-GCAGCCCCTCAGAATGATATTTGTCCTC-3'; 12S rRNA, 5-TCAAACTGGGATTAGATACCCCACTAT- 3' and 5-TGACTGCAGAGGGTGACGGGCGGTGTGT- 3'. These primers were synthesized t>y R. K. Werner in the Department of Biochemistry and Molecular Biol- ogy at the University of Miami. Polymerase chain reaction was carried out in a final volume of 50 pi in a reaction mixture described by Kocher et al. (1989). This reaction mixture was pre- heated at 94° C for 3 minutes followed by 30-35 cycles of amplification (93°C for 1 min, 44°C for 1 min, and 72°C for lmin). The same cycle was applied to am- plify both cytochrome b and 12S rRNA genes. After amplification, the reaction mixture was separated from covering oil and transferred to a 1.5-mL microcentrifuge tube. TE buffer (0.5 mL) was added to the reaction mixture, followed by chloroform: isoamylalcohol (24:1) extraction and ethanol precipitation. After centrifuga- tion at 12,000 x g for 10 minutes, the pellet was dried under reduced vacuum, rehydrated with 10 to 50 pi of TE, and stored at 4°C. Endonuclease digestion for amplified DNA fragments Nine restriction endonucleases used were Alu I, Cfo I, Dde I, Hae III, Hin fl, Mbo I, Msp I, Rsa I, and Taq I, all recognizing symmetric 4-base pair sequences. One unit of each enzyme was applied to 1 to 2 pi of ampli- fied PCR product in a final reaction volume of 5 pi. The digested samples were electrophoresed through 27c NuSieve (3:1) (FMC BioProducts, Rockland, ME) agarose gels in TBE buffer (90 mM Tris-boric acid, and 2mM EDTA). DNA bands were visualized and photo- graphed following electrophoresis and staining with ethidium bromide. Data analysis The size of fragment amplified and re- stricted was estimated in comparison with a size stan- dard ( 1 kb ladder, BRL). Nucleotide sequence divegence (p) (Upholt, 1977) was calculated by using the propor- tion of shared restriction fragments between specimens. Restriction patterns by each endonuclease were desig- nated A, B or C for composite haplotype analysis, and the number of haplotype was used to estimate nucleon diversity (/; ) (Nei and Tajima, 1981) within species. Results Gene amplification Each pair of primers successfully amplified 355±5 bp and 450±5bp fragments of the cytochrome b and 12S rRNA genes, respectively. There were no apparent dif- ferences in the fragment size among species. Both DNA fragments in all of the fresh and frozen specimens and in ethanol-preserved embryos and larvae were well amplified. In contrast, the amount of DNA was usu- ally much less in museum specimens, to the extent that PCR did not always amplify one or both of the gene fragments. Restriction fragment analysis Figure 1, A and B, shows representative restriction patterns with diagnostic enzymes for amplified frag- ments of the cytochrome 6 and 12S rRNA genes. The restricted fragment distributions for nine enzymes of 72 individuals in cytochrome b and 59 individuals in 12S rRNA genes are shown in Tables 2 and 3. 622 Fishery Bulletin 91(4). 1993 154 134 Lane Figure 1 Restriction profiles of cytochrome b lAl and 12S rRNA iBi gene fragments amplified by PCR. IA) lanes 1 and 14 (size standard), 2 (blank), 3 and 4 (La and Ra digested by Alu Ii, 5 and 6 (Lap and Lv digested by Cfo I), 7 and 8 (La and Ra digested by Dde I), 9 and 10 (Ra and Ls digested by Hae III), 11 and 12 (Lap and Ra digested by Hm fh, and 13 (La no digestion); (B) lanes 1 and 14 (size standard), 2 (La no diges- tion), 3 (La digested by Alu I), 4 (Lb digested by Cfo I), 5 (Ls digested by Dde I), 6 (Lg digested by Hae III), 7 (Oc digested by Hin fl), 8 iLs digested by Mbo I), 9 and 10 (Lg and La2 digested by Msp I), 11 and 12 (Lg and La digested by Tag I), and 13 (blank). For the 355 bp cytochrome b fragment, four enzymes (Mbo I, Msp I, Rsa I and Taq I) had no restriction sites in any of the species. Dde I had restriction site(s) in all species examined, without apparent size difference in restricted fragments within and between species. Varia- tion on the restricted fragment length within and be- tween species was observed for four of the nine en- zymes examined. Alu I appeared to have no restriction sites in this fragment of all Lutjanus spp. and 0. chrysurus (Oc), but two restriction sites producing three fragments (170, 110 and 74 bp) were evident in R. aurorubens (Ra). Cfo I, Hae III and Hin ft digestions revealed polymorphism within species; L. griseus (Lg) for all three enzymes, L. analis (La) for Cfo I and Hin fl, L. eampechanus (Lc) for Hae III and Hin fl, and 0. chrysurus (Oc) for Cfo I only. All nine enzymes had recognition site(s) in the 450 bp 12S rRNA fragment of at least one species. No length variation of the restricted fragments between species was observed for Cfo I and Hae III digestions. Taq I digestion produced 230 and 170 bp fragments in L. griseus (Lg), while 280 and 170 bp fragments were observed in the other species. Intraspecific polymor- phism was observed for the other six enzymes; Alu I (for Lc, Ley, Lg, and Lj), Dde I (for Lc, Ley, and Lg), Hin fl (for La, Lap, Ley, and Lj), Mbo I (for La, Lc, and Lj), Msp I (for La, Ley, and Lj), and Rsa I (for Lc, Ley, and Lj). Data of restriction fragment patterns were summa- rized for haplotype analysis for species discrimination (Table 4). Straightforward diagnostic restriction pro- files were observed in the 12S rRNA fragment of L. griseus (Lg) digested by Taq I and in the cytochrome b fragment of/?, aurorubens (Ra) digested by Alu I, with- out RFLP within species. Haplotype analysis on cyto- chrome b fragment separated L. cyanopterus (Ley) and L. vivanus (Lv), and that on 12S rRNA gene fragment separated L. eampechanus (Lc). L. apodus (Lap), L. buccanella (Lb) and L. synagris (Ls) were deseriminated by haplotype analysis on both fragments. In contrast, identical haplotypes were found in five individuals of L. analis (Lai, 5, 52-54), all of L. mahogony (Lml, 2, M2) and in seven of O. chrysurus (Ocl-7), and in two individuals of L. jocu (LjM2, M3) and in one of L. purpureus (LpM2). One individual of L. analis (La2), two of L. jocu (LjMl, M7), and one of O. chrysurus (Oc50) had their own distinct haplotype. Mean percent nucleotide sequence divergence (p) within species ranged from 0 to 2.43 with an average of 0.65. The nucleotide sequence divergence for those between species within the genus Lutjanus ranged from 0.11 to 3.84 with an average of 1.69, for those between Chow et al PCR-RFLP analysis on thirteen western Atlantic snappers 623 cC CD cd O i-< rH rH —.00 rH rH O O i-H "-I O in rH O O O O ^ fH ~ O rH rH —IOO rH rH rH rH o o t> 1 ^-< O O O — ' O O O H H -h O O rt jj c OJ > lO iH rH S —■OOO — 1 O O ~ O i-l iH O — — 1 - ~ -1 - n o C CD en GO 2 « rH O O O O i-l i-H -^ O — — 1 — 1 O O ~ ^ rf • 1 0* j- a. n CO Q, § QQ CD J ] —■OOO rH O O ^ rH rH O — 1 O O 1—1 1-1 ^ ^ CJ 2 cd a CN a. s- of -H O O O -H O O - O .-1 — 1 ■-H O O - - - ^ T3 cd JD C "3 C 3 CO s 1 -H O O O — O O — < rH O rH O O s CD £ a. cd a. en CO ■-1 O O O rH O O rH —■-HO O — 1 — ^H ,_, ^ ,_, c T3 «j G CO CO en cm ►J 1 ° rH O O O O 1-H rH ^ O H — 1 —too OJ C G cn T -Q c a. a. ol rz & GQ "I 03 CN >> _s (_i cj ^ O O O iG ►J CD C s CD be OJ g o s- CO rH O O O H O O r- O t-l rH — < 0 0 rH rH rH ^ o S IN rH O O O rH O O rH rH O O rH rH CL _D CO 7 —■OOO rH O O - rH i-l O O H H - - - -^ i— o CO OJ a C^ tC CC 1 »H O O O O J CL c "■ o CJ CN rH O O O 'C to a aj j Pi *tf »"? -H O O O i-l O O ,_, O rH rH — 1 0 O T— ( ,_, ,_ ,_, ^ CN HO HO O O "*J* iq m 0 O OOO W W O m IO ia IO 1C N H ^ UO 00 [> CO CU CN O 10 r~ ao 10 IO iO IO CO ^ ^H CO ^H ^H CO HHH CO CN CO CO CO CO -O C « CO co -ti (U C C CO 53 5 cu -a Cu 0 -0 i i>: 05 e2 624 Fishery Bulletin 91(4). 1993 ra CO 03 1 o o rn rr rr rH O rH O O O rH H O O rH rH O O rH O -H o in 1 O rH rH rH ^ rH O rH O rH rH O -H -H O O rH rH O O -H O rH > s J U0 i O rt ■-! i-l rH O O rH rH O rH -H O O rH rH O O rH O -H ^J CO -J OS c 0) CD | O rH — ( -H l"H rH O rH rH O O O rH rH O O rH rH O O rH O rH W -D ea ii o p. CN i-T J S © rH rH rH '" ' -H O H O rH rH O -H -H O O -h rH O O rH O —■ C CU X CU CM a. n CM* OHrtH - rH O - O -H -H O -H rH O O i-h rH O O -H O — en" > cC 'o o 01 c 01 he 0) J rH 72 ^"S O rH rH -H .— 1 rH O f—i O -H rH O rH rH O O rH rH O O -H O -H < z lO K s -H o o o 1-1 O rH ■-1 rH O O rH O O -H O O O rH rH rH O -H GG en (M J rH r-J ■* § s O i-l rH -1 r— 1 rH O "^ rH O O O rH rH -H O O rH O O rH O rH o m O ►J I O i-l rH rt - rH O ^ O rH rH O rH rH O O rH rH O O rH O rH DQ CU <-G o J- [> O rH rH rH rH H O i— » O rH rH O -H rH © O -H rH O O rH O rr a. o. c ca o -J CO u 1 © rH rH rH rH O rH O O O rH rH O O rH rH O O ^ o ^ 'E T~* -*j to OS CN O rH rH rH ~ rH O -1 rH O O rH O O O rH O rH O O rH O rH d _1 ■<* "ST O -J — 1 rH rH -H O _, O rH rH O rH rH O O rH ■h O O rH O rH ■-< CN uo o o o o o O O o o o o O O O O O O O O O O O O lO CO CN O CM in cm U0 CO CN lO (N CO in o cm m m cm 00 CO t^ T3 T* CM rH rH n- ^* tji rji rji O] CM ^ CO H rj< ^ CO -H/ CM rH CM CM rH C rt X co *i -_ c 1 1 n' be C CO H tfcl | K_, B ts HH H | ! Oi. UJ £■* tn 0C 20 O Li. O ,6 DC HI JAN FEB MAR APRIL MAY JUNE JULY AUG SEPT OCT NOV DEC I I Cruises at sen MONTH ! ] Month departed Figure 1 Number of observed U.S. -registered tuna purse-seine vessels at sea, and number of observed vessels leaving port each month during 1987. 630 Fishery Bulletin 91 (4|. 1993 I >■ o z HI 3 o 12 3 4 5 MONTHS AT SEA Figure 2 Months spent at sea per trip by U.S. -registered tuna purse-seine vessels carrying observers during 1987. trips killed no spotted dolphins, and trips with higher kills occurred with decreasing frequency. Data for whitebelly spinner dolphins were intermediate in qual- ity. They were fewer but still relatively smoothly dis- tributed. About half the trips killed no whitebelly spin- ner dolphins, but of those incurring mortality, most trips were responsible for about 0.1 dolphins per day. Higher kills again occurred with decreasing frequency, with little evidence of extreme outliers. Data for com- mon dolphins exhibited the worst quality, being both sparse and very unevenly distributed. Of the 12 trips incurring mortality of common dolphins, 9 killed about 0.5 dolphins per day, 2 trips killed 1-1.5 dolphins per day, while the remaining trip was responsible for about 9 deaths per day. That one trip was responsible for 67% (594/882) of U.S.-caused common dolphin deaths during 1987. Simulations We first selected randomly and without replacement 5, 10, or 20 vessels from the total of 33 observed ves- sels (Fig. 4). We then sampled randomly and without replacement, 25% , 50%, or 75% of the trips made by each of these selected vessels, without regard to actual timing of those trips throughout the year. We did not force temporal stratification on these small fleet simu- lations because fleet sizes of 5 or 10 vessels generated so few data. Vessel selections were replicated 10 times at each combination of fleet size and percent coverage. Trip selections were replicated 50 times from each of the 10 sets of selected vessels. Thus, selected vessels were the same within each set of 50 replicate samples of trips but differed between sets of 50 replicates. Trips conducted by selected vessels in each sample fleet represent the "population" of trips for that fleet replicate. Each set of 50 replicate selections of trips from a given set of selected vessels generated 50 esti- mates of cumulative annual mortality (cumula- tive mortality through December 31). For each set of 50 estimates, we calculated the mean esti- mate, the coefficient of variation for the mean estimate, and the relative bias of the mean esti- mate. Thus for each combination of fleet size and percent coverage we generated 10 averaged esti- mates of annual mortality, 10 coefficients of varia- tion, and 10 estimates of relative bias. Because the "true" kill varied between fleet rep- licates depending on the particular vessels and trips selected, the average of these 10 averaged estimates (from the 50 replicates I of annual mor- tality is not a particularly useful measure in this study. This average represents only the average of the 10 sets of vessels that happened to be chosen for the 10 replicates of fleet size. However, the coefficients of variation and estimates of relative bias for each of the 10 averaged estimates are relevant indicators of the ability of the estimation process to precisely or accurately, or both, reflect the true kill, whatever it happens to be for a particular sample of vessels. Ac- cordingly, we present results only for the averages of the 10 estimates of CV and RB generated for each combination of percent coverage, fleet size, and dol- phin mortality type. We do not discuss directly the individual mortality estimates. Estimates The estimates in this simulation study were derived by using the ratio estimator kill per day, rather than the more precise estimator kill per set (Lo et al., 1982; Hall and Boyer, 1986) because at the time this study was designed and executed, kill per day was still the estimator of choice for NMFS1. 'Kill/day was used by NMFS up until 1987 when IOC* observer coverage began, because the quota system required that dolphin kills be monitored continuously throughout the year rather than simply summed at the end of the year. Estimates based on kill/day can be made on a reasonably real-time basis because the day that a vessel leaves port is a readily available datum, and observers report kills by radio on a biweekly basis throughout each observed trip. Number of days fished by the entire fleet (observed plus unobservedl can then be determined relatively simply by summing the number of days at sea since leaving port, and total kill estimated as the product of total fleet days times the estimator kill/day. Biweekly estimates based on kill/set are not possible because observers are not permitted to report set data over the radio. Set data do not become available until after vessels have returned to port. Edwards and Perrin Annual dolphin mortality 631 40 - Northern Offshore Spotted 35 - K = 8,437 >- o z LU o III DC 40 - 30 - — i T = 62 20 - 10 - 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 110 - 100 - 1 Common Dolphin 90 - K = 882 > o z LU o LU DC LL 80 - 70 - 60 - 50 - 40 - 30 - 20 - 10 - T = 12 n 0 1 23456789 10 11 KILL/DAY per TRIP Figure 3 Kill per day during observed trips ( 124 trips total) by 33 U.S. -registered tuna purse-seiners fish ng in the ETP during 1987, for three species of dolphins representing the range of mortality data types. Scales of ordinates and abscissas differ between panels. K is total observed kill during 1987; T is number of trips incurring kill. Factors and levels included dolphin group, U= 1,2,3; offshore spotted, whitebelly spinner, and common dol- phin), fleet size, {j = 1,2,3,4; for 5, 10, 20, or 33 boats), fleet replicate* (As=l,..,10), percent coverage, (/=1,2,3; for 25%, 50%, and 75% ) and percent coverage replicate„, (m=l,...,50). 632 Fishery Bulletin 91(4), 1993 I TOTAL FLEET = 33 BOATS J Fleet Rep (1) % Coverage % Coverage FtePO) Rep (50) I K. = 50) I • A K K _. 50 ? K1 50 A eg Fleet Size = \\ 5 Boats ( _i» ) /N [ REPS \ = 10 Fleet Rep (2) . . 1 TRIP "Pop" (2) . . . . Fleet Rep (10) \ j. II :: TRIP "pop" (10) 10 £ K. 50 Figure 4 Sampling procedure for simulation study. Within each replicate, cumulative annual mortality for a given group of dolphins (YHAT) is estimated as YHAT,jklm = (K,lklJDAYS,kJ TFDAYS > where K,]Um is the observed cumulative annual kill on all selected trips, DAYSjklm is the observed number of days, and TFDAYS is the total number of days spent at sea by all vessels selected for that replicate. For each set of 50 replicate estimates of annual mor- tality for a given fleet size, percent coverage, and dol- phin group type, we calculated the average estimate (YAVEha), the relative bias of that estimate (RBYS0), and the coefficient of variation of that estimate (CVYS0) as follows YAVErMljtl = (1/50) *lYHATllkln, RBYM„lk, = 100 * (YAVE^u-known kUl„k)lknown kill,jk and CVYmkl = l^MSE(YAVES0,lkl)]/known killllk\*100 where MSE(YAVEh0ljkl) = H(YHATllklm - known killijkf m = l and Edwards and Perrin: Annual dolphin mortality 633 BYAVEm, = YAVEMltl - known kill Jmjki SOijkl ' Lijk > where BYAVE-Mlkl is the bias of YAVEm, and V(YAVEMjkl) is the variance of YAVEMlkh calculated as V(YAVE50:iH) = (1/49) * l(YHATljklm - YAVEm)* (Cochran, 1977). The preceding equations produce 10 YAVE50's, 10 RBYM's, and 10 CVY50's for each combination of fleet size, percent coverage, and dolphin group type. The average of these 10 RBY^s is RBYm,k = (VIO) *liRBY50l]kl) and the average of the 10 CVY50's is CVYi0tjh = (1/10)* I CVY; bOijkl ■ The analytic confidence intervals for the 50 estimates illustrate the influence of various combinations of fleet size and observer coverage level, on the estimated precision that may be associated with any individual estimate. The analytic confidence intervals for individual rep- licate estimates of mortality were calculated by the International Mathematical and Statistical Library (IMSL) routine SMPRR for ratio estimates (IMSL, 1987). This routine calculates confidence intervals for ratio estimates using the analytic formula for approxi- mate variance of a ratio. The procedure is based on the assumption that a normal approximation to the ratio variance is appropriate (Cochran, 1977; IMSL, 1987). Where data are sparse (fewer than 30 data points in the data set; i.e., in most of the cases in these simulations) this assumption is generally inap- propriate (e.g., Cochran, 1977), but for single repli- cates we had no computationally simple alternative. Bootstrapping confidence intervals for these individual replicates would have eliminated any need for a nor- mal approximation but would have required signifi- cantly more computer time to convey essentially the same gross patterns and general message. Sampling distributions and confidence intervals To facilitate interpretation of patterns seen in relative bias and coefficients of variations, we plotted both fre- quency distributions and analytic 95% confidence in- tervals for an arbitrarily selected single set of 50 indi- vidual replicate estimates of mortality derived under various combinations of conditions. The frequency dis- tributions and confidence intervals illustrate, in par- ticular, variations due to differences between replicates in the fleet selected, in contrast to the relative bias and coefficient of variation, which pertain to sampling properties of the estimator. Only one set of 50 repli- cates, of the 10 sets generated under each combination of fleet size and coverage level, is illustrated for each combination because the general messages conveyed by the figures were the same for all sets. Frequency distributions and confidence intervals are plotted only for the cases of 5 and 20 boats at 25% and 75% coverage. Combinations of these values spanned the range of fleet sizes investigated and enabled us to examine whether problems might occur even with cov- erage as high as 75% when fleet size is as small as 20 (or worse, 5) boats. The frequency distributions of the 50 estimates il- lustrate graphically the influence of various combina- tions of fleet size and observer coverage level on the behavior (dispersion) of the estimator itself (kill/day). Results Relative bias (RB) RB was generally small overall, but exceeded the man- agement objective of 5% for common dolphins and whitebelly spinner dolphins when coverage was low (25%; Fig. 5). Coefficients of variation (CV) CV decreased with increasing percent coverage and with increasing fleet size in both the "best" data group (offshore spotted dolphin; Fig. 6) and in the "interme- diate" data group (whitebelly spinner dolphin). CV de- creased with increasing percent coverage but showed no consistent effect of fleet size in the "worst" data group (common dolphin). Sampling distributions Frequency distributions were affected somewhat by fleet size (primarily by shifting the central tendency), noticeably by percent coverage (primarily by decreas- ing the spread of the distribution), and very strongly by dolphin group type (primarily in terms of the num- ber of modes in the distributions; Figs. 7, 8, and 9). Bias and variability increased with small sample sizes and non-smooth data distributions. 634 Fishery Bulletin 91|4). 1993 10 > P < JJI -10 -■ i D[ DOLPHIN GROUPS ■■ Northern Offshore Spotted EH Whilebelly Spinner I I Common "D 0 "id "d[ 10 20 10 20 10 20 25% Coverage 50% Coverage FLEET SIZE (5,10,20) 75% Coverage Figure 5 Average relative bias (%) in estimates of annual dolphin mortality as a function of fleet size and percent observer coverage. Dashed line at +5"7r and -5f/r indicates management target. strongly but still noticeably by fleet size (Figs. 10, 11, and 12). Mortality estimates were most stable and confidence intervals dramatically reduced for offshore spotted dolphins at high (75%) observer coverage (Fig. 10). Mortality esti- mates were more variable and confidence intervals wide and ragged, even with a rela- tively large fleet (20 boats), for offshore spot- ted dolphins at low observer coverage (5%). Intervals were most unstable at the combi- nation of lowest observer coverage and smallest fleet size4. Confidence interval pat- terns for whitebelly spinner dolphins were intermediate, being more variable than pat- terns for northern offshore spotted dophins but being less variable than patterns for common dolphins (Fig. 11). At low coverage and small fleet size, confidence intervals for whitebelly spinner dophins showed the same bimodality characteristic of confidence in- tervals for common dolphins under all con- ditions. At high coverage, confidence inter- vals showed the same relatively stable and Distributions of mortality estimates for offshore spot- ted dolphins tended to be relatively narrow and unimodal for fleets of both 5 and 20 vessels, at high coverage (75%; Fig. 7). Distributions remained unimodel but were more dispersed, at low coverage (25%). Distributions for whitebelly spinner dolphins were also unimodal in general but tended to be more dispersed than was the case for offshore spotted dolphins (Fig. 8). Distributions for common dolphins were markedly bimodal and dispersed under all sampling conditions, reflecting the selection (or not) of the one trip with unusually high kill (Fig. 9). Modal values of mortality estimates for all three dolphin types increased with in- creasing fleet size (not surprisingly, because more boats generally kill more dolphins) re- gardless of coverage level. Dispersion also increased with fleet size, more obviously when observer coverage was low than when coverage was high, as more data became available for analysis. Confidence limits Confidence limits were affected similarly to frequency distributions. Limits were affected most strongly by dolphin group type, very noticeably by percent coverage, and less 4The irregular pattern of the confidence intervals observed for mortal- ity estimates from small fleets (e.g., 5 boatsl with low coverage (e.g., 5 boats, 259f coverage! are not surprising because this combination of fleet size and percent coverage means that the mortality estimates are being derived from about 4 observed trips during an entire year (assuming 5 boats make 3 trips/year, 15 trips X 0.25 = 4 trips). 25% Coverage 50% Coverage FLEET SIZE (5,10,20) 5 10 20 75% Coverage Figure 6 Coefficient of variation (percenti in estimates of annual dolphin mortality as a function of fleet size and percent observer coverage. Dark horizontal bar at 20rr indicates management target. Edwards and Pernn Annual dolphin mortality 635 u. O >- o z UJ u- O >- o z UJ o ui IT 40 35 30 25 20 15 10 5 0 40 35 30 25 20 15 10 5 0 Boats: % Cov: 5 25 D nn ET_ ■ i i i Boats: % Cov: 20 25 J3 on 2000 4000 6000 8000 10000 12000 14000 35 - 30 - Boats: 20 % Cov: 75 25 ■ 20 ■ 15 ■ 10 ■ 5 - n - n -. — i — i — i — i — i JWi. 2000 4000 6000 8000 10000 12000 14000 Figure 7 Effect of percent coverage and fleet size on frequency distributions of annual mortality estimates for offshore spotted dolphins, from one set of 50 replicates. 35 - <* 30- u. Boats: % Cov: 5 25 35 ■ 30 - Boats: % Cov: 5 75 O 25- | 20- 111 UJ DC 10 - u. 5- " pi |—j nn„ 25- 20- 15 - 10 - 5- 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 35- <* 30 1 u. O 25- O Z 20 • Boats: % Cov: 20 25 35 30 - 25 - 20 - Boats: % Cov: 20 75 O 15' UJ oc 10- 5 • mfl n n fin J Lru 15 - 10 - 5 - rp _ ~ _ 200 400 600 800 1000 1200 1400 0 200 400 600 BOO 1000 1200 1400 A A K K Figure 8 Effect of percent coverage and fleet size on frequency distributions of annual mortality estimates for whitebelly spinner dolphins, from one set of 50 replicates. 636 Fishery Bulletin 91(4). 1993 u. O >- u z IU 3 O' ill cc o z LU o LU CC 10- 0 .10 Boats; 5 % Gov: 25 oH ^n. , ni.i , M Boats; 5 % Cov: 75 35 ■ 30 • 25 20 • 15 10 ■ 5 0 20 30 40 50 60 70 80 90 100 0 40 10 20 30 40 50 60 70 80 90 100 Boats; 20 % Cov: 25 ~nl. Pl~lm 400 800 1200 1600 2000 2400 2800 3200 35 30 25 ■ 20 • 15 10 ■ 5 0 Boats. 20 % Cov: 75 nn 200 400 600 800 1000 1200 1400 Figure 9 Effect of percent coverage and fleet size on frequency distributions of annual mortality estimates for common dolphins, from one set of 50 replicates. cm 5! LU VT «S ? °* in r § 20 30 REPLICATE # 12 ■ 10 - Boats: % Cov: 8 - 5 75 6 ■ 4 , Lt|t^H/««^»*<*,*"llt*< 2 ,lMl^,w— ^■1» ■ H -■? 10 20 30 40 50 Boats: 20 % Cov: 75 ♦♦♦♦♦♦♦*" rta^JwH-Mit* M" 10 20 30 REPLICATE # 40 50 Figure 10 Effect of percent coverage and fleet size on analytic confidence intervals of replicate annual mortality estimates for offshore spotted dolphins. Replicate estimates sorted by estimate level, from one set of 50 replicates. Note differences between panels in Y scale. Edwards and Pernn. Annual dolphin mortality 637 LU M « I CO CM i/l ?! l » I en + l g <* b 10 20 30 40 50 0 10 20 30 REPLICATE # REPLICATE # Figure 1 1 Effect of percent coverage and fleet size on analytic confidence intervals of replicate annual mortality estimates for whitebelly spinner dolphins. Replicate estimates sorted by estimate level, from one set of 50 replicates. Mortality estimates expressed in thousands. Note differences between panels in Y scale. 20 30 REPLICATE # „_— ~AA/uy*V~ Boat: % Cov: 5 75 10 20 30 40 6- Boat: % Cov: _/W^*VS"^ 4- 20 75 J 2- ,Jr--*~*~*~~~~* v" ZJ 2- "m"'IK-». «*— *n 20 30 REPLICATE # 50 Figure 12 Effect of percent coverage and fleet size on analytic confidence intervals of replicate annual mortality estimates for common dolphins. Replicate estimates sorted by estimate level, from one set of 50 replicates. Mortality estimates expressed in thousands. Note differences between panels in Y scale. 638 Fishery Bulletin 91(4). 1993 narrow limits seen for offshore northern spotted dol- phins at high coverage. Confidence intervals for com- mon dolphins illustrate the problems inherent in rela- tively heterogeneous data, where mortality is sometimes quite high, usually relatively low, and data overall are relatively few. For this species, confidence intervals were extremely variable even with relatively high coverage (Fig. 12) and were very narrow (or non- existent) for replicates which fortuitously included only low-kill data, but very wide for replicates including a few very high kills. Although these analytic confidence intervals are clearly inappropriate measures for precise estimation of variance characteristics of the estimates (note nega- tive intervals in some cases; Figs. 10, 11, and 12), they are presented here because they provide an effective illustration of the effects of varying conditions on the variability of the mortality estimates. Despite the inter-group differences, the general response to increas- ing coverage and increasing fleet size is similar in all three dolphin group types. Confidence intervals be- come narrower and more stable as more data become available. Discussion Dolphin group type Of the three factors investigated here (dolphin group type, observer coverage level, and fleet size), dolphin group type had the greatest effect on dolphin mortal- ity estimates, followed by percent coverage and fleet size. This hierarchy of effects is controlled by two char- acteristics of the kill data for each dolphin group type — frequency (the number of times that mortality occurs) and variability (differences between times in the num- ber of dolphins killed). The total number killed can have relatively little influence on the quality of the estimate. This is illustrated by comparing results for whitebelly spinner dolphin and common dolphin. Al- though total kill was comparable for both dolphin group types (981 deaths of whitebelly spinner dolphin, 882 deaths of common dolphin; Fig. 3) the data sets dif- fered markedly both in number of trips incurring kill (62 for whitebelly spinner dolphin, 12 for common dol- phin) and in the distribution of kill per day among those trips (Fig. 3). The data set for common dolphins exhibits the worst of both characteristics; frequency of kill was low (few trips killed common dolphins) and variability between trips in kill per day was high. These problems exem- plify an unfortunate interaction between data collec- tion problems and the ecology of the dolphins them- selves. Mortality of common dolphins due to the U.S. fleet during 1987 was infrequent because the geographic range of this species is relatively limited and occurs primarily within Mexico's Exclusive Economic Zone. U.S. vessels rarely fish in this area, therefore common dolphins rarely die in U.S. tuna nets in this area. Mor- tality was variable at least in part because common dolphins have an unfortunate habit (in this context) of forming very large schools, pre-disposing them to the possibility of very large-kill "disaster" sets. The data set for whitebelly spinner dolphin exhibits a problem with only one of the characteristics; data are relatively infrequent. Unlike the case for common dolphins, kill per day was not extremely variable. This similarity in kill per day generates statistics for whitebelly spinner dolphin that are much less biased and variable than for common dolphin. Observer coverage level The effect of observer coverage level is influenced both by fleet size and dolphin group type. To achieve a de- sired level of precision and accuracy in mortality esti- mates, observer coverage levels will have to be higher in smaller fleets because the available data will be fewer, and higher in dolphin groups with "messy" data, because observer coverage levels affect the probability of encountering an unusually large kill5. Observer coverage will need to be relatively high even for large fleets, when estimating mortality of dol- phins with sparse and heterogeneous data. For ex- ample, with kill data as sparse and variable as was the case for common dolphins in 1987, nearly 100% coverage would be required to generate CVs lower than 20% regardless of fleet size (Fig. 6). With more fre- quent and less variable kill data, such as for northern spotted dolphin in 1987, CVs lower than 20% can be achieved with 50% coverage of 10-boat fleets (Fig. 6). When fleet size drops to 5 boats, even this relatively well-behaved data set requires coverage at about 75% to achieve CVs less than about 20f£ . With coverage as low as 25%, even a fleet size of 20 boats was insuffi- cient to meet the management objective of 20% CV. Fleet size In smaller fleets, each data point comprises a larger fraction of the available mortality data. In particular, the influence of unusually large mortalities (e.g., the 'The possibility of high kills is more important than the possibility of low kills; this is because all dolphin group types experience zero kill frequently (therefore it is not unusual or unexpected I. Also, zero kill is a definitive lower bound, while the upper bound on kill i^ limited only by the potential school size of the dolphin group. Edwards and Pemn Annual dolphin mortality 639 one trip responsible for over 500 deaths of common dolphins) will be much greater in smaller fleets. The accuracy and precision of mortality estimates are affected less by fleet size per se than by observer coverage level for a given fleet size, and the fleet's variability in kill per day. Kill per day is affected not only by the fishermen's choices of fishing methods and areas, but also by the type of dolphin found associated with a given school of tuna. Because these are not factors that can be controlled, small fleets will gener- ally require higher coverage level than larger fleets to achieve a given level of accuracy and precision in mor- tality estimates. Simulation procedures and estimates The simulation procedure used in this study was de- signed to reflect the sampling process as it would oc- cur in the real world. More precise and less biased estimates of mortality rates for the population of trips contained in the 1987 data set would have resulted from simple random sampling of the 124 trips in the data set as a whole. But simple random sampling im- plicitly assumes that all vessels are equal in fishing ability. This is not the case, and it is likely that some fleets as a whole may have greater (e.g., those newer to purse-seining and therefore less experienced) or lesser (e.g„ the more experienced fleets) mortality rates than the average for the ETP purse-seine fleet overall. In addition, in the real world, not all trips made by all purse-seine vessels fishing in the ETP would be avail- able for sampling. Only trips made by the vessels in a particular fleet would be available for sampling, and only those vessels actually observed would contribute data. If vessels (or more properly, the crew) differ in their ability to release dolphins unharmed (or not), then fleets with more (or fewer) "low kill" vessels will have lower (or higher) mortality rates than other fleets of comparable size. Although it would have been pos- sible to estimate the number of trips that would have been made, on average, by a fleet of a given size, and to have then randomly sampled that many trips from the 1987 data base, the results would have been unre- alistically precise. The cluster sampling resulting from the selection of trips only after selecting vessels adds variability in the estimates but is more realistic than simple random sampling. The sampling scheme used here is a single stage cluster sampling, for which a ratio estimator is the most appropriate choice of esti- mation procedure (Cochran, 1977). Discussion Although the kill-per-day estimator used in this simu- lation study is no longer used by NMFS because 100% observer coverage has made estimation unnecessary, the results of this study have general implications for current estimation procedures based on kill per set (e.g.. Hall and Boyer, 1986) and for mortality estima- tion procedures in general where data quality may vary between stocks. The uneven structure of the data set for common dolphins has unfortunate implications for deriving es- timates of mortality for dolphin groups that are char- acterized by having such infrequent and widely vari- able kill per day. Specifically, estimates of mortality can vary widely depending on which trips happen to be chosen. In our simulation, we could resample the total population of vessels repeatedly, thus generating relatively unbiased, though individually variable, esti- mates of mortality. In the real world, only one sample (one set of mortality data per dolphin group type) will be collected per year. If this sample is collected under low percent coverage, it appears very likely that the data may be affected by undetectable sampling biases. This bias is more likely to underestimate than to over- estimate mortality because sets with large kill are rare and likely to be underestimated, even though the mor- tality during such sets may be responsible for a dis- proportionately large percentage of the total kill. The problem with missing the rare large-kill sets is that the kill in these sets can apparently be one or two orders of magnitude greater than the "usual" kill. For very abundant groups, missing a few large kills will miss only a small percentage of the total number of dolphins in the group; underestimating mortality could be relatively harmless. For less abundant groups, the large kills might represent a significant proportion of the existing stock. Underestimating this mortality could lead to seriously underestimating the impact of mor- tality due to fishing operations on these stocks. In the case of the 1987 data set for dolphin kill by the U.S. fleet, coverage greater than 96% (the highest observed) would be required for all boats in order to generate mortality estimates for common dolphin with CYs less than 20% . Alternatively, if only the most abun- dant groups are considered (e.g., offshore spotted dol- phin), CV's less than 20% could be achieved with only 50% coverage of fleets as small as 5 boats. There ap- pears to be no unique solution that is optimal for all groups. In addition, the poor quality of the data presented here for common dolphin in fact underestimates the true extent of the problem for this species. In actual practice, the species is managed as three separate stocks rather than as one combined stock as presented here. The data are thus extremely sparse for the indi- vidual stocks, and the problems with estimating mor- tality, given anything less than full observer coverage, are greatly exacerbated. 640 Fishery Bulletin 91(4). 1993 The results of these simulations, in particular the results for five-boat fleets, and common dolphins, strongly influenced subsequent regulations related to comparability criteria for import of tuna caught by non- U.S. fleets. These regulations now require that mor- tality data for common dolphins be stratified separately from all other stocks, and that non-U.S. fleets meet the U.S. requirement for 100% observer coverage6. Of the three factors discussed here (dolphin group type, percent coverage, and fleet size), only percent coverage can be controlled by the sampling program. Our results imply that providing maximum protection for all dolphin groups would require mandating cover- age to achieve a desired level of statistical precision for the dolphin group type with the least statistically stable data. Acknowledgments The U.S. fleet simulation study was designed by Bruce Wahlen. Nancy Lo and Peter Perkins provided insight- ful advice about sampling procedures and statistical methods. We thank also the participants in the Work- 6U.S. regulations for acceptable levels of observer coverage for non- U.S. fleets have progressed from a requirement for 33% coverage of trips made by fleets with 10 or more vessels, 50% of trips on fleets of 5-9 vessels, and 100% of fleets with 1-4 vessels during 1989 and 1990 (Federal Register 12/19/89;Vol. 54 #242, p. 51918-51923), to 75^ of all trips regardless of fleet size during 1991 and 1992 (10/18/90; 55 FR 42235 and 10/8/91; 56 FR 50672), to the current requirement for 100% coverage regardless of fleet size (Federal Reg- ister 1/8/92; Vol. 57, #8, p. 668). shop on Estimating Dolphin Mortality, and two anony- mous reviewers, who provided many useful comments during the review process. Literature cited Cochran, W. G. 1977. Sampling techniques, 3rd ed. John Wiley & Sons, NY, 428 p. DeMaster, D. P., E. F. Edwards, P. Wade, and J. E. Sisson. 1992. Status of dolphin stocks in the eastern tropical Pacific Ocean. In D. R. McCullough and R. H Barrett (eds.). Wildlife 2001: populations. Elseiver Press, London, England. Hall, M. and S. Boyer. 1986. Incidental mortality of dolphins in the eastern tropical Pacific tuna fishery: description of a new method and estimation of 1984 mortality. Rep. Int. Whal. Comm. 36:375-381. IATTC Annual Reports. 1980-1991. InterAmerican Tropical Tuna Commission, 7o Scripps Institute of Oceanography, La Jolla, CA 92037, var. pagination. IMSL. 1987. SMPRR/DSMPRR. In IMSL, Inc. Stat. Library, International Mathematical and Statistical Library of Computer Programs, p. 751-758. Lo, N. C. H., J. E. Powers, and B. E. Wahlen. 1982. Estimating and monitoring incidental dolphin mortality in the eastern tropical Pacific tuna purse seine fishery. Fish. Bull. 80(21:396-400. Sakagawa, G. T. 1991. Are U.S. regulations on tuna-dolphin fishing driv- ing U.S. seiners to foreign-flag registry? North Am. J. Fish. Mngmnt. 11(31:241-252. Abstract.- Field and laboratory procedures were used to acquire dol- phin school size estimates from ver- tical aerial photographs. Multiple photographs were taken of 48 sepa- rate schools during a 1989 eastern tropical Pacific (ETP) dolphin abun- dance survey. During a 12-week "counting period," three readers did independent counts of dolphins in the photographs. For each school, the best photograph imagery was se- lected and the mean of the three in- dependent counts was used to esti- mate its "true" size. The coefficient of variation (CV) for school size esti- mates (between-reader precision) av- eraged 5.4% and ranged between 1.2% and 14.6%. Most (92%) of the schools were estimated with preci- sion, resulting in a CV of less than 9.0%. Within-reader CV averaged 3.5% and ranged 1.4%-7.1%, indi- cating that readers were quite pre- cise. To test if reader methods were constant during the counting period, temporal trends in estimates were tested by linear regression analyses and a repeated-counts experiment with repeated measures analysis of variance (RM-ANOVA). Regression analyses indicated no significant temporal trends or bias in the de- viation of counts from the means. The RM-ANOVA showed a signifi- cant "reader with time" interaction which was attributed to the rela- tively high variability between read- ers in counts made at the start of the experiment. Results suggested that methods were constant and counts were precise after an initial "warm-up" counting session. Method and precision \n estimation of dolphin school size with vertical aerial photography James W. Gilpatrick, Jr. Southwest Fisheries Science Center National Marine Fisheries Service, NOAA 8604 La Jolla Shores Drive, La Jolla, CA 92038 Manuscript accepted 11 May 1993. Fishery Bulletin 91:641-648 ( 1993). Visual estimation of the number of dolphins in a school is difficult be- cause dolphins dive and the entire school is rarely visible at the sea sur- face at one time. Shipboard observer estimates can be highly variable and they may be biased (Scott et al., 1985; Anganuzzi and Buckland, 1989). Con- sequently, estimates of dolphin popu- lation abundance derived from visual survey data may be biased, and re- sulting management decisions aimed at conserving these populations may be inappropriate. The Southwest Fisheries Science Center (SWFSC) has been conduct- ing surveys to monitor temporal trends in the abundance of eastern tropical Pacific (ETP) populations of the pantropical spotted dolphin iStenella attenuata), spinner dolphin (S. longirostris), striped dolphin (S. coeruleoalba) and common dolphin (Delphinus delphis). These surveys are one part of a multifaceted effort to conserve ETP dolphins subjected to incidental mortality in the inter- national purse-seine fishery for yel- lowfin tuna, Thunnus albacares (Perrin, 1975; Wade and Gerrodette, 1992). Starting in 1987, in response to the problem of potential bias in visual school-size estimates, annual ETP surveys were complemented with aerial photography of dolphin schools. This allowed shipboard- observer estimates to be compared with estimates (of the same schools) taken from counts of dolphins in the photographs. School size estimates derived from large-format (126-mm) aerial photo- graphs were validated during a 1979 study when five separate schools (size range; 161-396 dolphins) were pho- tographed and then captured in a tuna purse-seine net (Scott et al., 1985). Results showed that dolphin counts from the photographs were not statistically different from counts tallied by hand as dolphins were be- ing released from the net. This sug- gested that counts from aerial pho- tographs approximated "true" school size. This report details photographic and counting methods used to derive dolphin1 school size estimates from large-format vertical aerial photo- graphs. Many of the techniques origi- nated with Scott et al. (1985). The techniques were modified to support the photography and laboratory ef- forts associated with large-scale ETP dolphin population surveys (i.e., an- nual surveys of 120 days; covering a 19-million-km2 study area). After a survey in 1989, three readers did in- dependent dolphin counts from ap- proximately 200 dolphin school photographs (multiple photographs were taken for 48 separate schools) over 12 consecutive weeks. For each school, the best photograph imagery was selected (according to criteria de- scribed below) and the mean of the three independent counts was used to estimate its true size. The coeffi- cient of variation (CV) was used to characterize the precision (between- 'For purposes of this paper, "dolphin" refers to dolphins as well as small toothed-whales that are included in the photograph sample. 641 642 Fishery Bulletin 91(4), 1993 and within-reader) associated with the school size estimates. One concern was that during the 12-week "counting period," readers might be inconsistent in the application of criteria (cognitive or physical methods, or both) used for counting. If readers were inconsistent, then this might affect the accuracy and precision of the esti- mates. A two-step approach was taken to address this concern. First, temporal trends in the deviation of counts from the means (see Sokal and Rohlf, 1981, p. 50) were evaluated by linear regression. This ap- proach assumed that the mean of three reader counts was close to the true school size and that the sample error was normally distributed. If individual reader counts were not normally distributed, then this might indicate reader bias or a change in criteria. Second, readers did repeated-counts (at four time points dur- ing the counting period) of a known sample of photo- graphed schools. This was done to evaluate the consis- tency of individual readers (within-reader precision) and to test the hypothesis that dolphin counts (from the same photographs) done at the beginning, middle, and end of the counting period were independent of temporal effects. If the hypothesis proved true, then this would support the idea that counting criteria were constant throughout the counting period. Materials and methods At sea Aerial photographs analyzed in this report were taken with Chicago Aerial Industries KA-62 aerial cameras mounted vertically on a Hughes 500-D helicopter. The helicopter was stationed aboard the National Oceanic and Atmospheric Administration (NOAA) survey ship David Starr Jordan. All cameras had forward-motion compensation (to minimize photo image blur from air- craft movement) and a 76.2-mm lens. Large-format (126-mm) Kodak Aerochrome MS 2448 color film was used. In order to minimize the behavioral response (scattering and deep diving) of the dolphin schools to the helicopter, photographs were taken at 244 m ( 800 ft) altitude. The scale of the photographs at this altitude was 1:3200, the sea surface area in a photograph frame was 366 m2, and a 2-m dolphin measured 0.63 mm on the film. The camera cycle rate was programmed to expose for approximately 80% film image overlap, i.e., 807f of the area photographed in one frame was photo- graphed again in the next successive frame. Succes- sive exposed photograph frames for a school were re- corded as a complete "photo-pass." To enhance the probability of photographing an entire school, multiple photo-passes (avg.=5) were made over a school. In the laboratory Counts Light tables equipped with dissection micro- scopes (0.7x to 7x variable objective and 10 x wide- field oculars) were used to view and count the dolphin images during counting. Dolphins were counted by hand-tally while being plotted with a permanent marker on a clear acetate overlay. The marked over- lay, when moved to the image overlap area of adjacent frames and aligned over dolphins that were previously plotted, made it easier to identify those dolphins not yet counted in the photo-pass. The photo-pass was the unit on which school size estimates were based (Scott et al, 1985); each pass was counted independently by three readers. Criteria for selecting the "true" school size For each of the 48 photographed schools, the readers chose (by group consensus) the one photo-pass where the mean of the reader counts was the best estimate of true school size. This decision was based on the precision of the three replicate counts and the reader assigned "quality ratings" for the photo-pass. Ratings reflected how confident the reader was in the accuracy of the count. For each photo-pass, readers independently as- signed quality ratings ranging from 1 to 4. A rating of "1" indicated that a photo-pass had "excellent" quality. A rating of "2" indicated that the count was "good" despite the presence of some questionable images (i.e., it was difficult to discern and count dolphins accu- rately when images were partially obscured by light- glare or when photographic resolution was reduced for deep swimming dolphins because of loss of light with sea depth). A photo-pass was rated "3" or "fair" when more questionable images were encountered, but read- ers still believed the count was a close approximation of true school size. A photo-pass rated "4" was deemed unusable for size estimation because the reader felt there were too many questionable images and the count was not a reliable estimate of true school size. One source of between-reader variation in dolphin photograph counts was reader error, where dolphins were missed or counted twice. Variation also occurred because readers differed in their interpretations of whether questionable images were dolphins or merely background water turbulence in the photographs. Based on the premise that "precision leads to accu- racy" (Sokal and Rohlf, 1981), for each photo-pass, the CV of the three independent counts was used to moni- tor the reliability of the school size estimate. For photo- passes where the CV of the counts exceeded 10%, the dolphin school was re-counted (independently) to see if the precision of the estimate could be improved. If the CV was above 15% after a second count, the photo- pass was excluded from the study because the counts Gilpatnck: Estimation of dolphin school size with aerial photography 643 were considered too variable and little confidence was placed in the accuracy of the estimate. Analytical methods The CV was plotted against the variables "school size" and "quality rating" to evaluate how different school sizes and photograph image quali- ties affected precision. The CV was also used to char- acterize the precision of individual readers in repeated- counts as described below. The percent deviation (PD) is a measure of how dis- tant (+ or -) a count is from the mean. Because this difference is expressed as a percentage of the mean, the PD is a consistent index of deviation relative to changes in the mean. For the PD, let .r„ be the ith reader's determination of the size of school j. The av- erage determination (or mean) over three readers of the size of thejth school was given by 1 3 3 U (1) The PD of x is expressed as PD„ = -A 100 (2) PD values were plotted to evaluate whether indi- vidual reader counts were normally distributed or read- ers were biased (i.e., tended to count high or low rela- tive to the mean). To test for temporal trends in individual reader counts, PD values were regressed against the variable "time." Time represented the chronological sequence, unique for each reader, in which the 48 schools were counted. Logistically, it was im- practical for readers to follow the same sequence in working with the photo-passes. For the repeated-counts experiment, a known sample of six photographed dolphin schools (henceforth referred to as the "experiment schools"), which varied in school size and image quality, were counted four times. Counts were done once at the start of the counting period, then again every 25 days during the period. Changes with time (temporal trends) due to the variables of "school size," "image quality," and "reader" were tested by using a repeated measures analysis of variance (RM- ANOVA) model from Winer (1971; p. 337).2 Outlier counts were identified by using Shapiro-Wilk's test for normality at a = 0.05 (Shapiro and Wilk, 1965). After -Model detailed in: Gilpatrick, J. W. Jr. (19921. Using vertical aerial photographs to estimate dolphin school sizes: precision and consis- tency. U.S. Dep. Commer, NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. L.J-92-35, 20 p. log-transformation, data met F-test requirements of homoscedasticity according to Levene's test at a = 0.05 (computer program BMDP7D used, Dixon et al. 1988). The RM-ANOVA was computed using the software pro- gram SuperANOVA (Abacus Concepts, 1990). Results and discussion Scott et al. (1985) reported that dolphin school size estimates derived from aerial photographs were accu- rate and more precise than visual estimates. They found the standard deviation of estimates (log- trans- formed) averaged 6% of school size for photographic estimates and 10%-30% of school size for visual esti- mates. The CV for estimates (untransformed) of the 11 schools used in their precision analysis averaged 8.4% (range: 3.7%-15.1%) indicating slightly less precision when compared with estimates presented here (avg. CV: 5.4%; range: 1.2%-14.6%; Table 1). The difference is explained, in part, by their statistical model, which accounted for variance due not only to independent repetitive counts (2 to 4 per photo-pass), but also due to camera types (126- and 229-mm formats) and mul- tiple photo-passes (2 to 7) for a given school. In the present study, variability was minimized by use of one type of camera ( 126-mm format) and by including only counts of the single best photo-pass for a school. School size estimates averaged 146 and ranged be- tween 4 and 633 (Table 1). Most schools (92%) were estimated with precision that resulted in a CV of less than 9.0%, and precision varied little with school size (Fig. 1 and Table 2). Estimate precision tended to de- crease with decreased quality of the dolphin school photographs (Fig. 2). PD values (listed in Table 1) plot- ted for individual readers appeared normally distrib- uted, indicating no between-reader bias in counts of the dolphin images. Repeated-count data are presented in Table 3. Out- lier values for experiment school number III (Table 3) resulted from reader error when dolphins were missed as the marked acetate was moved from the dolphin low density area of the photo-pass (in this case, the beginning of the photo-pass) to the high density area. Alternatively, when dolphins in the high density area were counted first, the precision of the estimate was improved because the majority of dolphins in the school were plotted and counted at the onset; this made it easier to track individual dolphins on adjoining frames. Within-reader CV for repeated counts averaged 3.5% . Reader 2, the most experienced reader, was most pre- cise in repeated counts (avg. CV: 2.6%; range: 1.4— 3.8%) followed by reader 3 (avg. CV: 3.4%, range: 1.5- 5.1%), and Reader 1 (avg. CV: 4.7%, range: 2.5-7.1%). The RM-ANOVA showed significant differences between 644 Fishery Bulletin 91(4). 1993 Table 1 Counts, mean school size, CV, average quality ratings and PD values for schools photographed in the 1989 ETP dolphin population survey. Reader Reader Species counts- Mean school CV Avg. PD4 School quality no. codes' Rl R2 R3 size (%) rating3 Rl R2 R3 1 S.COE 37 35 39 37.0 5.4 1.3 0.0 -5.4 +5.4 2 S.ATT 250 227 221 232.7 6.6 2.7 +7.5 -2.4 -5.0 3 S.LON 153 139 147 146.3 4.8 1.0 +4.6 -5.0 +0.5 4 S.COE 57 54 56 55.7 2.7 2.3 +2.4 -2.9 +0.6 5 S.ATT/S.LON 124 127 130 127.0 2.4 2.3 -2.4 0.0 +2.4 6 S.ATT/S.LON 192 196 216 201.3 6.4 1.7 -4.6 -2.7 +7.3 7 S.ATT/S.LON 274 236 242 250.7 8.2 1.0 +9.3 -5.9 -3.5 8 S.COE 48 47 49 48.0 2.1 1.0 0.0 -2.1 +2.1 9 F.ATT 17 18 18 17.7 3.3 1.0 -3.8 + 1.9 + 1.9 10 S.ATT/S.LON 81 78 77 78.7 2.7 1.0 +2.9 -0.9 -2.1 11 S.ATT/S.LON 412 416 398 408.7 2.3 1.3 +0.8 + 1.8 -2.6 12 S.ATT/S.LON 101 106 118 108.3 8.1 2.3 -6.8 -2.2 +8.9 13 S.LON 20 20 21 20.3 2.8 2.7 -1.6 -1.7 +3.3 14 S.ATT/S.LON 107 113 113 111.0 3.1 1.3 -3.6 + 1.8 + 1.8 15 S.ATT 118 124 109 117.0 6.5 1.3 +0.9 +5.9 -6.8 16 S.ATT/S.LON 618 607 675 633.3 5.8 2.0 -2.4 -4.2 +6.6 17 P.ELE 400 391 399 396.7 1.2 1.0 +0.8 -1.4 +0.6 18 S.COE 58 53 60 57.0 6.3 1.3 + 1.8 -7.0 +5.3 19 S.COE 52 55 61 56.0 8.2 1.7 -7.1 -1.8 +8.9 20 D.DEL 317 323 326 322.0 1.4 1.0 -1.6 +0.3 + 1.2 21 D.DEL 326 312 367 335.0 8.5 1.3 -2.7 -6.9 +9.6 22 S.COE 24 23 25 24.0 4.2 1.3 0.0 -4.2 +4.2 23 G.MAC 19 18 19 18.7 3.1 1.0 + 1.8 -3.6 + 1.8 24 S.COE 75 77 68 73.3 6.4 2.0 +2.3 +5.0 -7.3 25 S.COE 34 30 32 32.0 6.3 1.3 +6.3 -6.3 0.0 26 S.COE 166 171 181 172.7 4.4 1.7 -3.9 -0.9 +4.8 27 D.DEL 66 64 60 63.3 4.8 1.7 +4.2 + 1.1 -5.3 28 S.ATT/S.LON 216 216 233 221.7 4.4 1.3 -2.6 -2.6 +5.1 29 S.ATT 25 24 23 24.0 4.2 1.0 +4.2 0.0 -4.2 30 G.GRI 56 64 55 58.3 8.5 1.3 -4.0 +9.7 -5.7 31 S.BRE 4 4 5 4.3 13.3 2.0 -7.7 -7.7 +15.4 32 S.LON 576 678 548 600.7 11.4 2.3 -4.1 +12.9 -8.8 33 S.ATT 38 50 49 45.7 14.6 2.7 -16.8 +9.5 +7.3 34 S.COE 40 39 40 39.7 1.5 1.0 +0.8 -1.7 +0.8 35 S.ATT 88 90 80 86.0 6.2 1.7 +2.3 +4.7 -6.7 36 S.LON 315 324 350 329.7 5.5 1.3 -4.5 -1.7 +6.2 37 S.COE 26 25 24 25.0 4.0 2.0 +4.0 0.0 -4.0 38 S.COE 150 147 156 151.0 3.0 2.0 -0.7 -2.7 +3.3 39 S.BRE 36 34 35 35.0 2.9 1.0 +2.9 -2.9 0.0 40 S.ATT/T.TRU 284 293 337 304.7 9.3 2.3 -6.8 -3.8 +10.6 41 G.GRI 20 19 18 19.0 5.3 1.0 +5.3 0.0 -5.3 42 S.COE 35 34 35 34.7 1.7 1.0 + 1.0 -1.9 + 1.0 43 S.COE 73 74 88 78.3 10.7 2.0 -6.8 -5.5 +12.3 44 S.COE 154 154 144 150.7 3.8 1.3 +2.2 +2.2 -4.4 45 D.DEL 450 526 474 483.3 8.0 1.3 -6.9 +8.8 -1.9 46 S.COE 80 87 92 86.3 7.0 1.7 -7.3 +0.8 +6.6 47 S.COE 23 24 23 23.3 2.5 1.3 -1.4 +2.9 -1.3 48 S.COE 39 39 37 38.3 3.0 2.0 + 1.7 + 1.7 -3.7 Average: 145.5 145.9 147.4 145.5 5.4 1.6 'Species codes: S.COE = Stenella coen leoalba; S.ATT = S. atlenuata; S.LON = S. longirostris; F.ATT = Feresa attenuata; P.ELE = Peponocephala electra; G.MAC = Globi ■cphala macrorhynchus; D.DEL = Delph nus delphis; G.GRI = = Grampus gr seus; S.BRE = Steno bredanensis; T.TRU = Tursiops truncatus. Multiple species codes indicate mixed -speci es schools. -Reader codes: Rl = Reader 1; RL = Reader 2; R3 = Reader 3. 'Obtained by averaging the thres independent reader assigned quality ratings. 'PD = percent deviation of reader count from the mean. Gilpatrick Estimation of dolphin school size with aerial photography 645 E s o u 13 " ■ 14 - 13 - ■ 12 - 11 - ■ ■ 10 - 9 - 8 - ■ ■ ■ ■ ■ ■ 7 - 6 - 5 - 4 ■ ■ ■ ■ I ■ ■ ■ ■ ■ ■ ■ ■ ■ 3 2 - 1 - ■ ■ ■ ■ ! ■ ■ ■ I I 200 400 School size 600 Figure 1 Comparison of mean school size and precision (using CVl of count estimates for schools (/?=48) photographed during the 1989 ETP dolphin population survey. ■ 14 - 13 12 - £ 11 - ■ ■ e .2 10 - SB ■ C q _ > ■ "3 8 - ■ ■ ■ ■ g 7 - ■ ffici ■ I" o ■ U 5 - ■ ■ 4 - ! - 3 i : ■ i ■ 2 1 C 0.4 0.8 12 16 2 2.4 2. 8 Average quality rating Figure 2 Relationship between photograph quality ratings (averaged for three independent readers) and precision (using CV) of count estimates for schools (;i=48) photo- graphed during the 1989 ETP dolphin population survey. Table 2 Comparison of precision of estimates for small (<125 dolphins), medium (125-350 dolphins), and large (>350 dolphins) schools. School size CV(%) Mean Range Small 5.3 (1.5-14.6) Medium 5.2 (1.4-9.3) Large 5.7 (1.2-11.4) overall mean: overall range 5.4 (1.2-14.6) levels of the factors "school size" and "quality rating" (see between-factors; Table 4). This was expected be- cause photographed schools of different sizes were in- tentionally selected for the analysis. Often, with RM- ANOVA, between-factor effects are confounded with different levels of the factors (as is true for the two factors above) and it is the temporal trends of the within-factor data that are of primary interest in the analysis (Winer, 1971, p. 299). There were no differences between readers in the overall means of their repeated counts (temporal trends not considered; P=0.7898; Table 4). The F value for the within factor "Time" indicated no significant linear trend for repeated-counts (with the mean of the three reader counts). How- ever, a significant interaction for "reader with time" 95c7c of the cross-checks; samples were rejected if the discrepancies in counts could not be resolved. A linear relationship (r-=0.94) was detected between ring counts for the 35th and 36th vertebrae of the first 200 alba- core processed, and no significant difference in age composition was found between counts from the two sets of vertebrae (\2 test, P=0.9). Even though the 35th and 36th vertebrae were both suitable for study, we chose the 35th vertebrae because it was larger and easier to read in most cases. Tagging operations For tagging, albacore were caught mainly with com- mercial troll gear (Dotson, 1980). The tagging proce- dure was similar to that described by Laurs et al. ( 1976). Immediately after a strike, the fishing line was retrieved manually. Once the specimen was on board, the hook was removed and the albacore was quickly inspected for injuries to the gills, eyes, mouth, and palate. Albacore that were vigorous and not visibly injured were placed on a measuring board or tagging cradle (Kearney and Gillett, 1982) and were tagged as rapidly as possible with a stainless steel tube applica- tor that contained a 13-cm-long serially numbered spa- ghetti-type plastic tag with a single barbed nylon head. The tag was inserted at an oblique angle so that the barb was anchored among the pterygiophores of the second dorsal fin. Fork length was measured and re- corded on audio tape, along with the date, time, alba- core condition, and tag number. Albacore were returned to the water head first immediately after tagging. Labelle (1993b) described the albacore tagging op- erations conducted in the south Pacific, and summa- rized the corresponding release-recapture statistics. Approximately 17,000 albacore were tagged and re- leased from troll fishing vessels during 1986-92; the season totals ranged from 815 to 6,524. As of Novem- ber 1992, recaptures of 42 tagged albacore had been reported; we obtained complete information on sizes and dates of release and recapture for 27 of 42 alba- core. Most of these 27 tags were returned by fishermen, although some were detected during catch sampling at canneries and landing sites. Recreational fishermen also tagged and released an additional 3,646 albacore along the south east coast of Australia during 1973-92 (Matthews and Deguara, 1992). As of November 1992, 14 of these had been recovered but only one tag recov- ery record was sufficiently complete to allow inclusion of the data. Data analysis Length-frequency analysis The MULTIFAN computer program was used to estimate VB growth parameters from the length-frequency data under the assumption that the modes in the data represent year classes. A detailed description of the model is given by Fournier et al. (1990) and the program is described in the MULTIFAN 3 User's Guide and Reference Manual (Ot- ter Research, 1991). MULTIFAN can incorporate spe- cific structural hypotheses into models being fitted to the length-frequency data. The simplest structural hy- pothesis assumes that the mean lengths-at-age lie on a VB growth curve and that the standard deviations of length-at-age are identical for all cohorts. More com- plex structural hypotheses can be tested to determine if they provide a statistically significant improvement in fit to the data. The more complex hypotheses tested assume that the following processes can occur in the population sampled: 1 Sampling bias for the first cohort. This could result from selectivity during the sampling process in- duced by the fishing gear or the sampling method. Size selectivity was assumed to apply only to the first cohort and to decrease linearly with age until fish reach the second cohort. 2 Age-dependent standard deviation in length-at-age. For some fish populations, variation in length-at- age is not constant across cohorts. This hypothesis allows the standard deviation of length-at-age to increase or decrease linearly with age. 3 Seasonally oscillating growth. Seasonal growth pat- terns are known to occur in some fish populations (Pauly and Gaschiitz, 1979). This process was in- corporated into the growth model by adding two parameters, one representing the magnitude of the seasonal effect and the other determining the time of the year at which growth is slowest because of the seasonal effect. We systematically fitted models incorporating all pos- sible combinations of the above structural hypotheses and used likelihood ratio tests to identify the most parsimonious model structure. Fitting and testing pro- cedures were done automatically by MULTIFAN, al- though some user-specified input was needed to en- sure that the model exhibited stable behavior (see Fournier et al, 1990, for explanations). Estimates of the VB growth parameters K and L„ were obtained along with parameters for sampling bias, age-specific standard deviations, and seasonal growth. In the absence of information on the age of the first age-class, MULTIFAN assumes that the VB curve passes through the origin (i.e., t0=0). Estimates of age- at-length are based on this assumption. Analysis of vertebrae data The VB growth param- eters were estimated from vertebral-ring-count data and length data, under the assumption that ring counts indicate total age (in years). We did not attempt to 656 Fishery Bulletin 91(4), 1993 estimate fractional age based on an assumed birth date; therefore, all ages are integers. Maximum-likelihood estimates (MLE) were obtained according to the proce- dure described by Kimura (1980). The objective func- tion ( 0) minimized to obtain the parameter estimates was the negative log-likelihood of the VB model: „, ro ,, ik-[i.(i-^-v)]}! 0= Nlo&.(2na-) + isi (1) 2 2a2 where: /, = fork length of individual i, t, = presumed age of individual i, TV = number of individual measurements, L,= estimated asymptotic fork length, K = estimated growth coefficient, tu = estimated hypothetical age at zero length, a2 = estimated variance associated with the length-at-age. The objective function was minimized by using the Quasi-Newton algorithm in the NONLIN statistical module of the SYSTAT microcomputer program (Wilkinson, 1989). Asymptotic standard errors and pa- rameter correlations were obtained from the Hessian matrix once the iteration process was complete (see Wilkinson, 1989). As suggested by Kimura (1980), the initial values supplied for the parameters were ob- tained from the Walford plot (Ricker, 1975). Comparison of growth curves with tag-return data We used the growth curves fitted to the two data sets and the observed length increments from the available tag-return data to test our assumption that the length- frequency modes represent year classes and vertebral- rings represent annual features. It must be acknowl- edged that there is no statistically correct means of comparing the models and the length increment data, because the growth curves derived from length fre- quencies and vertebral-ring-counts are based on age- length data. The only statistically correct predictions possible from these growth models are predictions of length (the dependent variable) from age (the inde- pendent variable). However, the length increment data can be related to the growth curves by first generating predictions from the fitted growth models which are then compared to the length increment data. This was accomplished by predicting lengths-at-recapture by us- ing the standard Fabens (1965) length increment model: lr =1, + (L -li){l-eKM) (2) where: /, = length at release of individual i, A„= time at liberty of individual i, lr, = estimated length-at-recapture of individual and L. and K are the VB parameters estimated from the length-frequency or vertebral-ring-count data. As noted above, this procedure is not correct in the strict statistical sense because the growth models are not used to predict length from age (see Francis, 1988). However, for comparative purposes, the second method should reveal any gross departures from the "annual features" assumption applied in fitting the growth mod- els. Some justification for this is presented later. Results Growth analysis based on length-frequencies The most parsimonious model structure for the alba- core length-frequency data set included seasonal growth and age-dependent standard deviation in length-at-age. The incorporation of first length bias did not signifi- cantly improve the fit. Parameter estimates and esti- mates of the means and standard deviations of length- at-age are given in Table 1. Note that we use the term "relative age class" to denote that estimates of abso- lute age are based on the assumptions that the length modes represent annual cohorts and t0=0. The seasonal growth phase estimate of 0.216 indi- cates that fastest growth occurs in February, and slow- est growth occurs six months later in August at the end of the austral winter. The seasonal growth ampli- tude estimate (0.949) is near the upper limit of 1.0, indicating that growth is almost non-existent at that time of the year. Standard deviation in length-at-age increased progressively with age for the nine signifi- cant age classes detected. The predicted aggregate length-frequency distribu- tions fitted the observed distributions very well over the entire range of sizes (Fig. 2), and the predicted modes closely matched the actual modes in most months. The predicted modal distribution pattern in- dicates that there were usually four prominent age classes in troll catch samples. Growth analysis based on vertebral-ring counts The fork lengths of the albacore sampled ranged from 44 to 110cm (Table 2), which corresponds to the size range of albacore caught in the surface and longline Labelle et al.: Determination of age and growth of South Pacific albacore 657 Table 1 Estimated parameters from the MULTIFAN analysis, with predicted fork lengths (FL) and standard deviations (SD) of lengths-at- ige. Estimated Relative Predicted FL Length-at-age Parameter value age class (x, cm) SD von Bertalanffy K (cm-yr1) 0.239 3.04 48.86 1.74 von Bertalanffy L_ 97.100 4.04 59.12 2.03 Brody's rho (p) 0.788 5.04 67.19 2.29 Age of first age class (yr) 3.040 6.04 73.55 2.53 Mean length sampling bias 0.000 7.04 78.56 2.72 Seasonal growth phase 0.216 8.04 82.50 2.89 Seasonal growth amplitude 0.949 9.04 85.60 3.03 Mean SD 2.373 10.04 88.05 3.14 Ratio of first to last SD 1.861 11.04 89.98 3.24 Table 2 Summary statistics on sizes of albacore aged from vertebral-ring counts. Ring Sample Range Mean Length-at Coefficient Mean size count size (FL, cm) FL -age SD of variation increment 1 7 45-51 47.2 2.21 4.68 2 79 44-65 50.1 4.23 8.38 2.9 3 71 46-72 57.1 6.86 12.03 7.0 4 68 48-78 65.9 6.63 10.01 8.8 5 61 57-95 73.6 8.01 10.88 7.7 6 53 58-99 79.6 8.74 10.98 6.0 7 54 70-105 85.1 8.42 9.89 5.5 8 26 73-102 88.9 7.34 8.25 3.8 9 20 80-101 95.0 4.58 4.81 6.1 10 32 88-110 96.6 4.61 4.77 1.6 11 20 92-107 97.6 3.69 3.78 1.0 12 2 103-108 105.5 3.53 3.35 7.9 13 1 107-107 107.0 — — 1.5 fisheries (Labelle, 1993a). Up to 13 rings, assumed to be annuli, were visible in the vertebrae sampled. Sample sizes for all relative age classes were greater than 20, except for classes 1, 12,, and 13. Significant differences in the standard deviation of length-at-age were detected among the age classes (Bartlett's test, XL>=64.683, P<0.001), and the standard deviation was greatest for relative age classes 5-7. This trend differs from the results obtained in the length-frequency analy- sis, which indicated increasing standard deviation with age (although only linear trends were possible in this analysis). Differences in mean size between the successive age classes were calculated to reveal other anomalies in growth patterns. In theory, absolute growth should be rapid at first and decrease progressively in later life (see Ricker, 1975). The pattern observed for albacore agrees with these theoretical expectations, except for the high growth rates observed be- tween relative ages 8->9 and 11->12 (Table 2). No statistical or biological criteria justified the omission of the data associated with ages 8— >9 from the analysis. However, the large growth incre- ment obtained for age 11-^12 could simply be biased because of the small sample size (2) for relative age class 12. The VB model was fitted to the aggregate vertebral-ring-count data for relative age classes 2-11. The data for relative age classes 1, 12, and 13 were excluded be- cause of their small sample sizes. The VB model provided a good fit to the remaining data (Fig. 3). The residuals were generally well cen- tered around zero and the slope of a linear regression of residuals against relative age was not sig- nificantly different from zero (F- test=0.001;P=0.973). Of the albacore sexed, 59 were females and 70 were males. As these were distributed in a simi- lar fashion with respect to rela- tive age class as the pooled data set (females, males, and unsexed), the VB parameters were estimated for each sex (Table 3). Likelihood ratio tests (Kimura, 1980) were used to test for differences in growth parameters between sexes. No significant dif- ferences were detected for any of the tests carried out (Table 4). Therefore, the growth parameters estimated from the pooled data set were considered to be repre- sentative of growth patterns of male and female alba- core within the size range sampled. Comparison of growth curves A direct comparison of the VB growth curves estimated from length-frequencies and vertebral-ring-counts is slightly complicated by the assumption of r„ = 0 for the length-frequency analysis and the estimation of t0 = -1.922 year for the vertebral-ring model. This incon- sistency would result in the growth curves being out of phase by almost two years, apart from any other dif- ferences that might be present. One way of comparing the two curves is to compare the expected growth of 658 Fishery Bulletin 91(4). 1993 120 no . . 100 1 ■ i ■ j-~ — ■ - £ 90 u £ 80 oi g 70 ■ ■ ■ i ^^ « i i i^i i ■ t /^ , , i 1 /\ '• X 1 - it S 60 b. 1 /\ 11/11 - 50 i i ■ ■ - 40 30 0 2 4 6 8 10 1 2 14 Vertebra] rings Figure 3 Observed albacore fork lengths (cm) for each ring-count category (relative age class). The von Bertalanffy curve fitted to data for relative age classes 2-11 is represented by the solid line. an albacore of a specific initial size over time. This is equivalent to using the L . and K estimates in a Fabens- type length increment model to make an approximate comparison of predicted growth trends independent of assumed or estimated t„. Despite substantial differ- ences in the L„ and K estimates from the two analy- ses, there was little difference in the predicted growth of a 40-cm albacore over time within the range of the data (Fig. 4). Growth rates of tagged albacore The 28 tag release-recapture records with com- plete growth data included a wide variety of sizes (61.0-96.5cm), times-at-large (54-1,790 days) and release-recapture locations well distributed throughout the principle fishing areas. Estimated average growth rates during the periods at large were 0.17-1.13cmmonth '. For albacore at lib- erty <500 days, growth rate showed the expected decline with increasing mid-size (half way be- tween release and recapture sizes); but this de- cline was negligible for albacore at liberty >500 days (Fig. 5). Given the relatively small number of records available, no attempt was made to es- timate the VB growth parameters based on tag release-recapture data. Comparison of tag return growth increments with the fitted models To compare the consistency of the VB models de- rived earlier with the tag-return data, we first calculated sets of predicted recapture lengths for the 28 tag return records by using the two sets of L„ and K estimates, derived from length-frequency and verte- bral-ring-count data, in Equation 2. We also made a third set of predictions using the L . estimate and twice the K estimate obtained from the length-frequency analysis (L„-97.1cm, K=0A78 yr~M. These parameters are consistent with an assumption that the observed Table 3 MULTIFAN parameter esti mates of the VB growth model based on the vertebral-ring-counts. Parameter esti- mates are based on all albacore of relative a ges 2-11 ( op); females only of age s 2-9 (midd e); and males only of ages 2-10 (bottom) Lower and upper bounds are the upper and lower limits of the 959f confidence intervals. MLE=maximum -likelihood estimates. Lower Upper Correlation matrix Sample content Parameters MLE SD bound bound L. K t„ All albacore L, 121.024 5.193 110.846 131.202 ages 2-11 K 0.134 0.016 0.104 0.165 -0.981 n = 484 to -1.922 0.291 -2.492 -1.350 -0.880 0.951 _' SS X" df P 1 none 122.037 169.320 0.168 0.077 -0.907 -2.573 6863.3 2 L.,=L„2 129.504 129.504 0.145 0.134 -1.180 -1.561 6911.0 0.104 1 >0.70 3 A',=A', 131.088 127.932 0.140 0.140 -1.265 -1.455 6922.4 0.129 1 >0.60 4 tvetv 131.197 128.411 0.138 0.141 -1.350 -1.350 6933.6 0.153 1 >0.50 5 Same L ... K, f0 131.157 131.157 0.136 0.136 -1.385 -1.385 6948.1 0.184 3 >0.50 120 4 5 6 7 Time interval (years) Figure 4 Growth of south Pacific albacore from an initial length of 40 cm as pre dieted by models fitted to length-frequency and vertebral-ring-count data. — . 1. : 500 days length-frequency modes are semestral, rather than an- nual features. We then compared each set of predic- tions to the observed recapture lengths. The growth models derived from length-frequency and vertebral-ring-count data, assuming that the modes and rings are annual features, appear to be consis- tent with the tag return length increment data (Fig. 6). Deviations of observed from predicted recapture lengths are both positive and negative and are generally within the range of expected individual variation. In con- trast, the growth model based on the assumption of semes- tral length-frequency modes overestimates the recapture lengths. On this basis, the tag return length increment data lend support to our initial assumptions concerning the length-frequency modes and vertebral-rings. Discussion Estimates of the VB growth parameters for south Pacific albacore based on the analy- sis of length-frequencies and vertebral-ring- counts were generally within the ranges of values reported for albacore from other re- gions (Table 5). The only exception was the L. estimate obtained with MULTIFAN, which was lower than all other estimates reported. This could be partly due to the fact that the data set used for the MULTIFAN analysis did not include length- frequency samples from longline catches, so the largest size classes were not well repre- sented. Nevertheless, we do not attach any biological significance to this or to the dif- ferences in VB parameters estimated from the two data sets in this study. Despite differences in the pa- rameters, the predicted growth patterns based on the two analyses are remarkably similar. This often noted feature of the VB growth model (e.g.. Knight, 1968; Francis, 1988) highlights the potential pitfalls of com- paring growth only on the basis of VB growth param- 1 5 © C 1.1 1 .91 .8 .7 .6 .5 .4 .3 .2 .1 >500 days 55 60 65 70 75 Mid-size (cm) 85 90 55 60 65 70 75 Mid-size (cm) 80 85 90 Figure 5 Individual growth rates of tagged south Pacific albacore for two time-at-liberty categories. Mid-size is the fork length halfway between the release and recapture lengths. 660 Fishery Bulletin 91(4), 1993 •& a. S 60 65 70 75 80 85 90 95 100 Observed recapture lengths (cm) Figure 6 Predicted versus observed recapture lengths of tagged south Pacific albacore. The three panels depict recapture lengths predicted on the basis of models fitted to length-frequency data assuming annual cohorts, vertebral-ring-count, data as- suming an annual formation rate, and length-frequency data assuming semestral cohorts. The lines cover the points corre- sponding to a perfect match between predicted and observed lengths. eters. In this study, the VB model was preferred to alternatives because it was used in available software such as MULTIFAN. There were no indications from the data that the VB model was not appropriate for modelling albacore growth, although a Gompertz model (Ricker, 1979) appeared to provide an equally good fit to the vertebral-ring-count data. In fishes, rings are thought to form in bony tissues as a series of growth checks separated by zones, usu- ally more opaque, that are associated with growth (Casselman, 1983). The usual inference that rings are formed yearly is based on the observation that growth rate, particularly for temperate species, is frequently seasonal as a result of its dependence on temperature. Growth checks are thought to occur during the period of the year when the water is coldest; high growth rates occur during periods of warmer water tempera- tures. Length-frequency analysis provided evidence that juvenile albacore growth is strongly seasonal; there- fore this seasonality could be the basis of ring forma- tion in vertebrae. However, the relationship of water temperature with this seasonality is unknown. Alba- core, while being predominantly a temperate species, are also found in tropical waters, particularly as adults. In common with other Thunnus species, they occupy a three-dimensional habitat within wide temperature lim- its (14-20°C) and are capable of physiological ther- moregulation, maintaining body temperatures of up to 15°C warmer than that of the surrounding waters (Morrison et al., 1978). It is possible that ring forma- tion in the bony tissues of albacore is influenced by these biological characteristics and by more complex biological cycles, such as feeding and reproduction, as well as the temperature of the water that they occupy. As a working hypothesis, the length-frequency modes and vertebral-rings were assumed to be formed each Table 5 Reported estimates of the von Bertal inffy growth parameter albacore from various regions around the world. Data Sampling Aging Size range source region method L, K Co (FL, cml Clemens (1961) N. Pacific Tagging 135.6 0.170 -1.870 54-77 Shomura ( 1966 ) N. Pacific Tagging 118.8 0.250 1.999 60-91 Shomura 1 1966) N. Pacific Scales 114.4 0.308 0.818 65-120 Shomura (1966) N. Pacific Scales 145.3 0.159 -0.056 50-120 Laurs and Wetherall (1981) N. Pacific Tagging 125.0 0.199 — 47-92 Shomura (1966) N. Pacific Vertebrae 104.8 0.431 1.504 69-90 Shomura (1966) N. Pacific Scales 145.3 0.150 -0.396 40-95 Bell (1962) N. Pacific Scales 108.8 0.225 -2.273 51-94 Huang et al. (1991) Indian Scales 128.1 0.162 -0.897 64-106 Beardsley(1971) Atlantic Length data 140.0 0.141 -1.830 68-114 Gonzalez-Garces and Farina-Perez (1983) N. Atlantic Dorsal spine 140.1 0.129 -1.570 38-110 Santiago (1992) N. Atlantic Length-freq. 117.9 0.200 — 48-106 Labelle et al. : Determination of age and growth of South Pacific albacore 661 year in the growth models presented in this paper. Laurs et al. (1985) validated the daily increment for- mation rates in sagittae of tagged north Pacific alba- core that had been injected with oxytetracycline (OTC) prior to release and concluded that the increment counts, adjusted upwards by 5%, provided an accurate estimate of absolute age. Preliminary estimates of ab- solute age of south Pacific albacore were obtained from counts of otolith increments, assuming similar incre- ment deposition rates as for the north Pacific popula- tion (Wetherall et al., 1989). These results indicated that south Pacific albacore grew about twice as fast as our estimates suggest. This would suggest that the length-frequency modes and vertebral-rings analysed here are more likely to be semestral than annual features. The length increment data from 28 tag returns pre- sented in this study clearly favor the hypothesis of annual length-frequency modes and vertebral-rings over a semestral formation rate. Given the assump- tion that length-frequency modes represent cohorts of albacore spawned during the same season, we would expect the temporal spacing of length-frequency modes to correspond to the frequency of spawning seasons. Both gonadosomatic indices and microscopic examina- tion of gonad tissue of south Pacific albacore clearly indicate a single annual peak in spawning activity dur- ing the austral summer (Ramon and Bailey, 1993). Most of the evidence therefore points towards an- nual length-frequency modes and vertebral-rings, al- though the conflicting results obtained from "daily" growth increment counts on otoliths are deserving of further study. It would be particularly useful to esti- mate increment formation rates in vertebrae and otoliths of tagged south Pacific albacore injected with OTC before release, as was done for the north Pacific population. During 1986-1989, 3,341 tagged south Pa- cific albacore were injected with OTC, and to date, three have been recaptured and the relevant hard parts obtained. It is hoped that examination of this material will, in due course, clarify increment deposition rates in south Pacific albacore. A previous review of the available information on stock structure lead Lewis ( 1990) to conclude that the south Pacific population probably had very limited ex- change with those of the North Pacific and the Indian Ocean. Parasite studies (Jones, 1991), tagging pro- grams, and ongoing electrophoretic surveys have not yet revealed the existence of separate sub-stocks within the south Pacific population. Thus, the growth rates reported here could be considered as representative of the growth patterns in the entire south Pacific alba- core population. However, we cannot reject the hypoth- esis that populations from other oceans have signifi- cantly different growth patterns. It is also plausible that local phenomena, such as land mass effects and upwelling zones that can affect the growth rates of albacore within certain regions of the south Pacific. Acknowledgments Several persons and agencies have contributed signifi- cantly to this research. Staff of the South Pacific Com- mission (SPC), New Zealand Ministry of Agriculture and Fisheries (MAF), and the United States National Marine Fisheries Service have contributed to the tag- ging, observer, and port sampling programs. Special thanks go to Patrick Swanson of MAF for his assis- tance in aging the vertebrae. We also thank all troll fishermen who assisted with the observer program and the release and recovery of tags. Financial assistance for the observer and tagging programs and for the data processing and analysis was provided by the Eu- ropean Community Fifth European Development Fund and the Canadian International Center for Ocean Development. Literature cited Anonymous. 1986. First South Pacific Albacore Research Workshop; Auckland, New Zealand, 9-12 June 1986. South Pac. Comm. Rep., Noumea, New Caledonia, 33 p. 1991. Fourth South Pacific Albacore Research Work- shop; Taipei, Republic of China-Taiwan. November 4-8, 1991. South Pac. Comm., Noumea, New Cale- donia, 40 p. Beardsley, G. L. 1971. Contribution to the population dynamics of Atlantic albacore with comments on potential yields. Fish. Bull. 69(4):845-857. Bell, R. R. 1962. Age determination of the Pacific albacore of the California coast. Calif. Fish Game. 48( 1 ):39-48. Berry, F. H., D. W. Lee, and A. R. Bertolino. 1977. Progress in Atlantic bluefin tuna ageing at- tempts. Collective Volume of Scientific Papers. Int. Comm. Conserv. Atlantic Tunas. 6(2):305-317. Berry, F. H. 1978. Vertebrae related studies. /;; J. J. Hunt (ed.). Proceedings of Atlantic bluefin tuna ageing workshop. Collective Volume of Scientific Papers. Int. Comm. Conserv. Atlantic Tunas. 7(2):337-339. Casselman, J. M. 1983. Age and growth assessment of fish from their calcified structures-techniques and tools. In E. D. Prince and L. M. Pulos (eds.), Proceedings of the international workshop on age determination of oceanic pelagic fishes: tunas, billfishes, and sharks. NOAA Tech. Rep. NMFS 8:1-17. 662 Fishery Bulletin 91(4), 1993 Clemens, H. B. 1961. The migration, age and growth of Pacific alba- core in the north Pacific (Thunnus germo), 1951- 1958. Calif. Dep. Fish Game, Fish. Bull. 115, 128 p. Dotson, R. C. 1980. Fishing methods and equipment of the US west coast albacore fleet. U.S. Dep. Commer., NOAA Tech. Memo. NMFS 8, 126 p. Fabens, A. J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. Fournier, D. A., J. R. Sibert, J. Majkowski, and J. Hampton. 1990. MULTIFAN a likelihood-based method for estimating growth parameters and age composition from multiple length frequency data sets illustrated using data for southern bluefin tuna (Thunnus maccoyii). Can. J. Fish. Aquat. Sci. 47:301-317. Francis, R. I. C. C. 1988. Are growth parameters estimated from tagging and age-length data comparable? Can. J. Fish. Aquat. Sci. 45:936-942. Gonzalez-Garces, A., and A. C. Farina-Perez. 1983. Determining age of young albacore, Thunnus alalunga, using dorsal spines. In E. D. Prince and L. M. Pulos (eds.), Proceedings of the international workshop on age determination of oceanic pelagic fishes: tunas, billfishes, and sharks. NOAA Tech. Rep. NMFS 8:117-122. Hampton, J., T. Murray, and P. Sharpies. 1989. South Pacific albacore observer programme 1988-1989. South Pac. Comm. Tuna and Bill. Assess. Prog. Tech. Rep. 22, 22 p. Hampton, J., T. Murray, and K. Bailey. 1991. South Pacific albacore observer programme 1989- 1990. South Pac. Comm. Tuna and Bill. Assess. Prog. Tech. Rep. 25, 25 p. Huang, C, C. Wu, C. Kuo, and W. Su. 1991. Age and growth of the Indian Ocean albacore, Thunnus alalunga, by scales. Proceedings of the ex- pert consultation on stock assessment of tunas in the Indian Ocean; Bangkok, Thailand, July 2-6, 1990. Collective Volume of Working Documents 4: 111— 122. FAO/UNDP Indo-Pacific Development and Man- agement Programme. Jones, J. B. 1991. Movements of albacore tuna {Thunnus alalunga) in the South Pacific: evidence from parasites. Mar. Biol. 111:1-9. Johnson, A. G. 1983. Comparison of dorsal spines and vertebrae as ageing structures for little tunny, Euthynnus alletteratus, from the Northeast Gulf of Mexico. In E. D. Prince and L. M. Pulos (eds.), Proceedings of the international workshop on age determination of oceanic pelagic fishes: tunas, billfishes, and sharks. NOAA Tech. Rep. NMFS 8:111-115. Kearney, R. E., and R. D. Gillett. 1982. Methods used by the Skipjack Survey and As- sessment Program for tagging skipjack and other tuna. South Pac. Comm. Tech. Rep. Skip. Surv. Asses. Progr. 7:21-43. Kimura, D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull. 77(4):765-776. Knight, W. 1968. Asymptotic growth: An example of nonsense dis- guised as mathematics. J. Fish. Res. Board. Can. 25:1303-1307. Labelle, M. 1993a. A review of the south Pacific albacore troll fishery 1985-92. South Pac. Comm. Tuna and Bill. Assess. Prog. Tech. Rep. 32, 25 p. 1993b. A review of albacore tagging in the South Pacific. South Pac. Comm. Tuna and Bill. Assess. Prog. Tech. Rep. 33. 17 p. Laurs, R. M., and J. A. Wetherall. 1981. Growth rates of north Pacific albacore, Thunnus alalunga, based on tag returns. Fish. Bull. 79(2):293- 302. Laurs, R. M., W. Lenarz, and R. Nishimoto. 1976. Estimates of rates of tag shedding by north Pacific albacore, Thunnus alalunga. Fish. Bull. 74(3):675-678. Laurs, R. M., R. Nishimoto, and J. A. Wetherall. 1985. Frequency of increment formation on sagittae of north Pacific albacore (Thunnus alalunga). Can. J. Fish. Aquat. Sci. 42:1552-1555. Lee, D., E. D. Prince, and M. C. Crow. 1983. Interpretation of growth bands on vertebrae and otoliths of Atlantic bluefin tunas, Thunnus thynnus. In E. D. Prince and L. M. Pulos (eds.), Proceedings of the international workshop on age determination of oceanic pelagic fishes: tunas, billfishes, and sharks. NOAA Tech. Rep. NMFS 8:61- 70. Lewis, A. D. 1990. South Pacific albacore stock structure: a review of available information. South Pac. Comm., Noumea, New Caledonia. Third South Pacific Albacore Research Workshop; Noumea, New Caledonia, 9-12 October 1990. WP 5, 13 p. Matthews, J., and J. Deguara. 1992. Gamefish tagging 1991-92. NSW Fish. Res. Inst. Annual Report, 13 p. Morrison, D. C, R. M. Laurs, and R. C. Ulevitch. 1978. Activity of albacore serum complement reflects its thermoregulatory capacity. In G D. Sharp and A. E. Dizon (eds.), The physiological ecology of tunas, p. 141-150. Academic Press. New York. Otter Research. 1991. MULTIFAN 3 user's guide and reference manual. Otter Research Ltd., Nanaimo, Canada. Pauly, D., and G. Gaschiitz. 1979. A simple method for fitting oscillating growth data, with a program for pocket calculators. ICES CM 1979/G: 24, 26 p. Pianet, R. J. Chabane, J. Cremoux, R. Etaix-Bonnin, R. Grandperrin, R. Leborgne, G. Moarii, G. Mou- Tham, J. Panche, M. Radenac. Labelle et al Determination of age and growth of South Pacific albacore 663 1990. Rapport de la campagne PROSGERMON 87, N.O. Coriolis, 21 Fevrier au 20 Mars 1987. ORSTOM, Noumea, 1990. Sci. Mer : Biol, mar.; 7: 69 p. Ramon, D., and K. Bailey. 1993. Reproductive patterns of south Pacific albacore, Thunnus alalunga, as indicated by gonosomatic in- dex and meiotic activity. South Pac. Comm. Fifth South Pacific Albacore Research Workshop; Papeete, Tahiti, French Polynesia. March 1993. WP 13, 27 p. Ricker, W. E. 1975. Computation and interpretation of biological sta- tistics of fish populations. Fish. Res. Board. Can. Bull. 191, 382 p. 1979. Growth rates and models. In W. S. Hoar, D. J. Randall, and J. R. Brett (eds.), Fish physiology. Vol- ume VIII: Bioenergetics and growth, 786 p. Academic Press. London. Santiago, J. 1992. Application of "MULTIFAN" to estimate the age composition of the North Atlantic albacore catches. Int. Comm. Conserv. Atlantic Tunas (ICCAT). Collective Volume of Scientific Papers XXXIXl 11:188-192. Sharpies, P., K. Bailey, P. Williams, and A. Allan. 1991. Report of observer activity on board Jamarc driftnet vessel RV. Shinhoyo Maru fishing for alba- core in the south Pacific ocean. South Pac. Comm. Tuna and Bill. Assess. Prog. Tech. Rep. 24, 24 p. Shomura, R. S. 1966. Age and growth studies of four species of tunas in the pacific Ocean. In T A. Manar (ed. ), Hawaii Governor's Conf. Center. Pac. Fish. Res. Proc, p. 203- 219. Wang, C. H. 1988. Seasonal changes of the distribution of south Pacific albacore based on Taiwan's tuna longline fisheries, 1971-1985. Acta Oceanographica Tai- wanica. 20:13-39. Wetherall, J. A., R. N. Nishimoto, and M. Y. Y. Yong. 1989. Age and growth of south Pacific albacore deter- mined from daily otolith increments. South Pac. Comm. Second South Pacific Albacore Research Work- shop; Suva, Fiji, Nov. 1989. WP 18, 7 p. Wilkinson, L. 1989. SYSTAT. The system for statistics. SYSTAT, Inc. Evanston, Illinoi. Abstract. -Reproduction of the false southern king crab (Paralomis granulosa) in two localities of the Beagle Channel, Argentina, was studied by monthly trap sampling during 1989 and 1990. Size at go- nadal maturity in males (50.2-mm cephalothoracic length, CL) and fe- males (60.6-mm CL) was signifi- cantly less than size at morphomet- ric maturity (57.0-mm CL in males; 66.5-mm CL in females). Embryonic development lasted 18-22 months. During this period, there was appar- ently a 10-12 month diapause. In one of the two localities, develop- ment of eggs in a given clutch was very heterogeneous, suggesting si- multaneous occurrence of eggs with 12- and 22-month development pe- riods. Larval hatching took place mainly during winter. Female P. granulosa molted during November and mated immediately after. Bien- nial reproduction was detected on the basis of ovarian and embryonic development, and on the basis of shell condition. Thus, two different female groups occur in the popula- tion of the Beagle Channel. Fecun- dity increases with size (1,441 to 8,110 eggs per female) and is sig- nificantly less at the end than at the start of embryogenesis. Ovaries and brood each represented at most 6- lr7i of body weight. Paralomis granu- losa is the only representative of its genus that inhabits shallow water and apparently retains some re- productive features of its deep- water relatives. Reproductive biology of the false southern king crab [Paralomis granulosa, Lithodidae) in the Beagle Channel, Argentina Gustavo A. Lovrich Consejo Nacional de Investigationes Cientificas y Tecnicas (CONICET) Centra Austral de Investigaciones Cientificas (CADIC) CC 92, 94 1 0 Ushuaia, Tierra del Fuego, Argentina Present Address Maurice Lamontagne Institute, Invertebrates and Biostatistics Division, 850 Route de la Mer, PO Box 1 000, Mont Joli, Quebec G5H 3Z4, Canada Julio H. Vinuesa Consejo Nacional de Investigationes Cientificas y Tecnicas (CONICET) Centra Austral de Investigaciones Cientificas (CADIC) CC 92, 94 1 0 Ushuaia, Tierra del Fuego, Argentina Manuscript accepted: 11 May 1993. Fishery Bulletin: 91:664-675 1 1993 1. Crabs of the genus Paralomis are lithodids that inhabit the Atlantic, Indian, and Pacific Oceans, ranging in depth from 5 to 4152 m. The false southern king crab, or centollon in Spanish, Paralomis granulosa (Jacquinot, 1847), inhabits the Pa- cific Ocean from Paso Tenaun (Chile; 40° S) to Cape Horn, and the Atlantic Ocean from 56° S to the Golfo San Jorge (Argentina; 47°S) including the Islas Malvinas (Falkland Islands) at depths of up to 50 m (Macpherson, 1988). Paralomis granulosa and the southern king crab (Lithodes santolla) constitute the main crusta- cean fisheries off the southern tip of South America. Commercial exploi- tation of P. granulosa started in the early 1970's. The largest catches were recorded in 1986 and amounted to 1,300 metric tons for Argentina and Chile. In Argentina, during the last three years, the yield of P. granulosa was at least twice as great as that of L. santolla. Although commercial fishing for P. granulosa began more than 15 years ago, studies on the life history of this species are few. Man- agement studies were conducted in Chile which resulted in fishing regu- lations, such as a minimum legal size and gear restrictions (Campodonico et al., 1983). There is virtually no information on the life cycle of P. granulosa in South America, except for reports on larval development (Campodonico and Guzman, 1981), larval ecophysiology (Vinuesa et al., 1989), and diets of juveniles and adults (Comoglio et al., 1990). In this study we document aspects of the reproductive biology of P. granulosa, from data collected dur- ing monthly sampling in the Beagle Channel during 1989 and 1990. Data on the reproductive cycle, fe- cundity, embryogenesis, and other life history traits were examined to ac- quire basic information on the biol- ogy of this commercially valuable species. Materials and methods The fishing area was located along the Beagle Channel between Lapa- taia Bay (west) and the Moat Chan- nel (east) (Fig. 1): fishing depths ranged from 10 to 60 m. The Beagle 664 Lovrich and Vlnuesa Reproduction of Paralomis granulosa 665 ov ^Kl 1 4 ! 1 z < 111 u o % c isla A ISLAS MALVINAS (FALKLAND) CJ li. lAV DE TIERWK ATLANTIC OCEAN < ■70.2mm CL; slope=1.473) (Fig. 3). The slope of the regression for known juvenile females (<59.7mm CL; slope=0.96) was significantly greater (F=223.0; P<0.001) than that of adult females (>75.1mm CL; slope=0.84). Estimated size at morphometric maturity 50% were 57.0 mm CL (959? confidence limits: 53.9-60.1) for males and 66.5 mm CL (95% confidence limits: 63.4-69.5) for fe- males (Fig. 2B). These results indicated that the method can be applied to both sexes. Embryogenesis Embryogenic development was divided into five stages as follows: Embryonic Stage I (ES I): Spherical or slightly ellipsoidal egg. Yolk bright yellow or orange. A peri- vitelline space present between the chorionic mem- brane and the yolk mass. No evidence of division. 55 CO. (3 o 1.8 MALES, N=758 E E 1.6 A 1— T 1.4 cu LU IE 1.2 i° UK 40% in frequency during the two sampled years. ES III never exceeded 5% in frequency and was observed in February and August 1989, and in April, July, and October 1990. These findings suggest that dia- pause ends between early winter and late summer, as observed in the laboratory. ES IV and ES V oc- curred mainly from late autumn to early spring. Post- ovigerous females appeared from late autumn to early spring. Samples of P. granulosa taken from the study area at any given time comprised females in at least two different molt stages: either EIM or AIM occurred with MIM (Table 1, Fig. 4B). PRM females occurred in Oc- tober and POM females were found in October and November 1989. Asynchronous embryonic development A portion of the females (21%; n=100) from the Becasses Islands, and some from the area near Ushuaia (2%; ft =395) carried clutches with eggs at different stages of development: ES II and ES IV or V were found in the same egg mass. This was observed around the Becasses Islands in August 1990, where 7 females car- ried eggs in ES II (at least 50% frequency in each clutch) and in ES IV In October 1990, 12 females had eggs in ES II (at least 50% frequency) and ES V. Two other clutches had eggs in ES II and ES IV. The more advanced eggs would probably hatch earlier than those that were less developed. We could not determine when hatching occurred because commercial fishing around the Becasses Islands was suspended in Octo- ber 1990 and the females were not held for further observation. Table 1 Embryonic development and shell condition of female Paralalias granulosa on 20 May 1990 in two localities, 10 km apart. ES=embryonic stages. Postovigerous females have empty egg cases and funiculi attached to the pleopodal setae. MIM=median intermolt; AIM=advanced intermolt. Frequencies for both embryonic development and shell-condi- tion are significantly different between the two localities (G-test for homogeneity on embryonic development = 12.66; P<0.005; G-test for homogeneity on shell condition = 23.45; P<0.005). Embryonic development Shell condition I .ocation ES I, II, and III ESIV andV Postovig. females MIM AIM Golondrina Bridges Is. 60 44 31 51 64 31 35 69 587 84 Lovrich and Vinuesa: Reproduction of Paralomis granulosa 669 LU CD ■< O DC B LU < O rr LU Q_ 100 90 80 70 60 50 40 30 20 10 0 100 90 80 70 60 50 40 30 20 10 0 18 18 12 20 33 27 30 17 10 19 16 24 21*8 25 15 44 71 JMMJSNJMMJSN FAJAODFAJAOD 1989 1990 55 19 25 13 2140 44816435 11 31 36 33961771937 45 40 61 J I I. JMMJSNJMMJSN FAJAODFAJAOD 1989 1990 Figure 4 (Ai Frequency of occurrence of egg stages through two years of sampling. Sample sizes are indicated above bars. Months with no samples are bare. (B) Frequency of occurrence of molt stages through two years of sampling. Sample size are indicated above bars. Months with no samples are bare. Embryonic growth Eggs from a single clutch increased significantly in size from the ES II ( 1.71 mm) to ES V ( 1.96 mm) (paired /-test: t=4.Q04; 3 df; P=0.02) in approximately one year of incubation in the laboratory. Within each asynchro- nously developing clutch, eggs in ES II ( 1.87 mm) were significantly smaller than those in ES IV or ES V (1.94mm) (completely randomized ANOVA; F=22.5; 1,12 df; P<0.001). Egg size at a given stage of develop- ment varied significantly among females (F= 13.28; 12,12 df; P<0.001). This demonstrates that embryonic growth cannot be generalized from eggs of individual female P. granulosa. 670 Fishery Bulletin 91(4), 1993 Development and maturity of ovaries Ovarian development of P. granulosa was described in detail by Lovrich (1991) and is similar to that of L. santolla (Vinuesa, 1984). The only difference with the latter was that oocytes in the ovaries of female P. granu- losa that had recently molted and spawned were al- ready in an early vitellogenic phase. These oocytes were surrounded by fibrous connective tissue and ra- diated outwards from the germ strand. Two groups of females, separated by differences in oocyte diameter (OD) and gonadosomatic index (GSI), were present in virtually all of the sampled months (Figs. 5 and 6). Discriminant analysis on a matrix con- sisting of CL, CH, OD and GSI (n=220 females sampled during 1990) gave 2 groups. The first included POM, EIM, or MIM females with eggs in ES I and ES II, the second comprised AIM or PRM females with eggs in ES III, ES rV, or ES V. The resulting discriminant function was y = 0.083 CL + 0.08 CH - 0.727 OD - 0.323 GSI, where coefficients were standardized by the standard deviations within groups. The canonical correlation was 0.798. The reproductive variables OD and GSI domi- nated the function, whereas morphological measure- ments contributed negligibly (canonical loadings: CL: 0.17; CH: 0.155; OD: -0.955; GSI: -0.866). An a posteriori classification was generated from the discriminant function. After applying the discriminant function to data for individual females we found that mean scores (i.e., y in function above) were signifi- cantly different for each group (Games and Howell test, t'=2A9; P<0.01). We thus conclude that there are two distinct groups differing in shell condition and development of ovaries and eggs. These two groups are a key feature of a biennial reproductive cycle. Females with asynchronous embryonic development were grouped by the discriminant analysis with fe- males in their first year of the reproductive cycle. In fact, for these females, mean GSI was 2.78% in August and 3.38% in October, representing 45 to 55% of maxi- mum GSI (fully developed ovaries in October, Fig. 5), respectively. Their average OD was 1.24 mm in August and 1.36 mm in October, whereas the maximum ex- pected at the end of the reproductive cycle was 1.8 mm (Fig. 6). Fecundity and reproductive effort Fecundity was studied in relation to carapace length in the Ushuaia (Fig. 7) and Becasses Islands. Clutches in late stages (ES IV or ES V) were analyzed sepa- rately to avoid underestimations due to possible loss of eggs (Kuris, 1991). Regressions of fecundity on cara- pace length were significant (Table 2; Fig. 7) and the slopes did not differ between females carrying clutches with eggs in ES I-II, in ES V, or with asynchronously developing eggs (from Becasses Islands) (F=1.12; P=0.33). ANCOVA indicated that females with ES V eggs or with asynchronously developing eggs had fewer eggs per clutch than those with ES-I and ES-II eggs (F=7.29; P<0.001) (Table 2). Adjusted fecundity was JASONMAMJJASOD (20) (21) (14) (21) (18) (21) (29) (29) (30) (30) (38) (31) (53) (9) 1989 1990 Figure 5 Relative frequency of oocyte diameter for female Paralomis granulosa caught in the Beagle Channel. Sample size is indicated in parentheses below each month. Lovnch and Vinuesa Reproduction of Paralomis granulosa 671 11 10 I- 7 - £ 6 to 5 o 3 - 20% mm MAMJJASOD (11) (30) (39) (42) (31) (44) (38) (53) (9) Figure 6 Relative frequency of gonadosomatic index (ratio ovaries weight : body weight * 100) of female Paralomis granulosa during 1990. Sample size is indicated in parentheses below each month. 3,688 eggs for females with ES-V eggs and 3,783 eggs for females with asynchronously developing eggs, rep- resenting about 88-90% of the estimated 4,201 eggs of females with ES I — II clutches for a constant carapace length (71.0mm). The ovaries and clutch of female P. granulosa were limited to less than 6-7% of body size. Brood weight varied from 3.67 to 30.39 g and scaled isometrically with weight of body, excluding clutch. Brood weight was at most 7% of female body weight: log clutch weight = -0.94 + 0.89 log weight (body excluding clutch) (n = 196; r2= 0.303 F„ 81.99, P<0.001; ^-statistic 40 3.9 3.8 3.7 >: 3.6 i- o 3.5 g 3.4 LU it 3.3 8 3.2 1 3.1 3.0 2.9 2.8 ^^"^ $ * ♦<>• ° - . ■•••.*..• . • O ° 0 Q O * o « ES 1 1 II • es iv a v . ECLOSION OUTLIERS 1. Scatter^ log cara area. E' n=4; ou tion for 76 1.78 1.80 1.82 1.84 1.86 1.88 190 1.92 1.94 1. LOG (CARAPACE LENGTH, mm) Figure 7 rams of log fecundity (eggs carried by a female) c pace length for Paralomis granulosa from Ushua 3 I and II, n=\92; ES TV and V, «=32; hatching egg tliers, n=9. The line represents the calculated equ clutches in ES I and II. 36 n a s, a- for#0 slope=l = -1.122; P=0.26). In addition, towards the end of the reproductive cycle in October (23rd month), GSI was about 6-7% (Fig. 6). Discussion Our data indicate that Paralomis granulosa has a bi- ennial reproductive cycle. First, female P. granulosa molt from late October to November, as evidenced by occurrence of PRM and POM stages (Fig. 4B). We con- sider that this event marks the beginning of the molt cycle and thus of the reproductive cycle. The frequen- cies of the different molt stages were consistent with those of the embryonic developmental stages, suggest- ing that embryogenesis and molting are phased. Em- bryogenesis lasts 18-22 months and is protracted by a 10-month diapause. Females with ES-I and ES-II eggs were always present throughout the year in the field samples and laboratory. Eggs are extruded from Octo- ber through November and embryonic development stops or slows at the beginning of cell division (ES II probably representing the diapause stage). After ap- proximately one year of incubation, embryonic devel- opment resumes between October and January, and 672 Fishery Bulletin 91(4). 1993 Table 2 Predictive regression of fecundity if) on crab size iCL, in mm) in two different locations of the Beagle Channel, during 1989 and 1990. Regressions for different stages of embryo lie development (ESi are presented for Ushuaia . r?-=coefficient of determination; F =F-statistic; ***=P<0.001. Embryo development CL range and location Equation of regression n r- F (mm) (1) ESI and II log F = -1.21 + 2.61 log CL 192 0.459 161.3*** 60.2-88.0 (Ushuaia) (2)ESV log F = -1.03 + 2.48 log CL 32 0.501 33.2*** 59.7-83.2 (Ushuaia) (3) Asynchronous log F = -2.72 + 3.41 log CL 43 0.528 45.9*** 61.1-78.4 (Becasses I.) Average log CL = 1.851 (±0.034) Adjusted log F ( 1 ) = 3.633(10.007) Adjusted log F (2) = 3.567 (±0.017) Adjusted log F(3 1 = 3.578 (±0.015) (1)*(2)&(3), P<0.01;(2l#(3l, P = 0.63 continues through the next winter, when eclosion oc- curs. Second, females separated into two groups on the basis of brood development, shell condition, and maturity of ovaries (Table 1; Figs. 5 and 6). In general, females that carried eggs in ES IV and ES V were in advanced molt stages (AIM and PRM) and had fully developed ovaries; by contrast, females with eggs in ES I, ES II, and ES III were in EIM or MIM stages and had small ovaries. Additionally, the presence of oocytes in early vitellogenesis in the recently spawned ovary denotes that oogenesis started before spawning and thus lasts more than two years. In Lithodes santolla, oogenesis lasts 24 months (Vinuesa, unpubl. data) while embryogenesis lasts 11 months (Vinuesa, 1984). The asynchronous embryonic development of P. granulosa is a novel feature among lithodid crabs. Our data suggest that some of the eggs within a single clutch develop in 12-14 months and hatch in early summer, while the remainder of the eggs com- plete their development during the next 10 months. We discount the possibility of a second mating and egg extrusion without molting be- cause 1) there is no seminal receptacle, and chitin- ous plates cover the gonopores during intermolt, and 2) the ovaries of females with asynchronously developing eggs were in the first year of their cycle. There is clearly a regional effect involved in asynchronous development since females from Ushuaia had more uniformly developed broods. However, with data presently available, we can- not speculate on the causes of this phenomenon. Since gonadal maturity (presence of gametes) does not necessarily imply morphometric matu- rity (crabs with differentiated secondary sexual characters), these two terms should be used to define different events in life history that may or may not occur simultaneously. We consider that these concepts are applicable to P. granulosa since different sizes at maturity were calculated by using different features (Figs. 2 and 3). Size at maturity for P. granulosa tends to increase with increasing latitude (Table 3). This re- lationship has been noted for L. santolla as well (Vinuesa, 1985). By contrast, in the Northern Pacific, size at maturity of lithodid crabs decreases with in- creasing latitude (Jewett et al., 1985; Somerton and Otto, 1986; Blau, 1990; Otto et al., 1990). Causes of geographical variation in size at maturity are still un- known, but environmental conditions such as bottom temperature may be a factor. Traps may give biased samples because 1) they se- lect for larger animals (Miller, 1990, but see Blackburn et al., 1990), 2) small crabs and females are excluded Table 3 Size at maturity (in mm) of Paralomis granulosa in different locali- ties. Size at morphometric maturity (SMM) is presented for males and size at gonadal maturity (SGM) for females. Where available, ±95'* confidence intervals are presented in parentheses. Males Females Area (SMM) (SGM) Malvinas I. (Falkland)1 (51°30'S) 52(2.15) 46 SenoOtway2(53°S) 64 52 Magellan Strait2 (52°30'S-53°S) 71 62 Beagle Channel ( 198 1-2 )2 ( 54°55'S I 75 66 Beagle Channel (this study 1 57(3.1) 60.6(2.3) 'Hoggarth, D.D., National Rivers Authority, Welsh Region, Gwyneedd, UK, pers. commun. 1992. 2Campodbnico et al. 1983. Lovrich and Vinuesa: Reproduction of Paralomis granulosa 673 by larger males, and 3) crabs near ecdysis and berried females are less vulnerable to trapping since they do not feed (Miller, 1990). Size at maturity of P. granu- losa may be overestimated because the 57.5-mm-CL size class was poorly sampled; this would particularly affect size at gonadal maturity. Also, the frequency of occurrence of females with late embryonic stages, es- pecially those in pre- and postmolt conditions, may be underestimated by trap sampling. Unfortunately, we have no way to assess these possible biases because we have no trawl surveys with which to compare data. Compared with other shallow-dwelling lithodids, P. granulosa has low fecundity and large eggs, resem- bling in this respect the deep-water species: Lithodes ferox (8,000 eggs maximum, 1.97 mm egg diameter; Abello and Macpherson, 1992); L. couesi (5,000, 2.3 mm; Somerton, 1981); L. murrayi (4,200, 2.45mm; Miquel and Arnaud, 1987; Miquel et al., 1985). By contrast, the shallow water Paralithodes camtschaticus and P. platypus carry up to 350,000 and 280,000 eggs, respec- tively, with an average egg diameter of 1.2 mm (Matsuura et al., 1971, 1972; Somerton and Macintosh. 1985). Fecundity off! granulosa was less in females with ES V or with asynchronously developing clutches for a constant carapace length (Table 2), because of egg loss. At the end of embryogenesis about 10-12% of the initial brood was lost. Diseases and egg predators are frequent causes of egg loss in crab species (Kuris, 1991). However we did not find evidence of epibiosis in broods of P. granulosa. Exceptionally small broods, out- liers in Fig. 7, may have resulted from delayed mat- ing, lack of mates, or to small size of mating males, as occurs in other lithodids (McMullen, 1969; Powell et al, 1973; Paul and Paul, 1990). This question requires further investigation. As the energy expended on each offspring increases, the number of offspring that parents produce decreases (Smith and Fretwell, 1974). Thus, in two related spe- cies similar energetic investment may result in many small or few large offspring. However, considering the two lithodids of the Beagle Channel, one finds that P. granulosa biennially produces eggs which are fewer but not larger than those of L. santolla, which annu- ally produces up to 59,000 eggs (Guzman and Cam- podonico, 1972) of 2. 1-2.2 mm in diameter (Vinuesa, 1987). There is no evidence that the biennial reproductive cycle of P. granulosa is more advantageous (i.e., an adaptative strategy) than the annual cycle of L. santolla. Paralomis granulosa larvae pass through fewer molting events (Campodonico, 1971; Campodonico and Guzman, 1981); thus mortality due to ecdysis is reduced. Shorter larval development would also re- duce predation risks in the plankton. Inhabiting the layer of water closest to bottom (Lovrich, unpubl. data) allows larvae to find refuge and thus reduces losses to predation. We speculate that the lesser fecundity of P. granulosa may be partially compensated for by a high survival during their development. In Paralithodes platypus, Jensen and Armstrong ( 1989) interpreted bi- ennial reproduction as a consequence of physiological and energetic constraints incurred by the species in a harsh environment. King crab species can be categorized into three groups on the basis of their reproductive cycles: Paralithodes camtschaticus and Lithodes santolla spawn annually, Paralomis granulosa and Paralithodes platypus (except primiparous females) spawn bienni- ally, and finally L. aequispina, L. couesi, and L. ferox spawn asynchronously. All Lithodes species, with the exception of L. santolla, inhabit deep waters whereas Paralithodes species inhabit shallow waters. Otto and Cummiskey (1985) hypothesized that king crabs in- habiting shallow waters (L. santolla, P. camtschaticus, and P. platypus) spawn synchronously during spring while deep-sea king crabs (L. aequispina and L. couesi) have protracted spawning periods. These authors sug- gest that this pattern could be related to a dependence on food sources by shallow water species. Paralomis granulosa inhabits shallow waters and spawns syn- chronously every two years, but we suppose that in this species synchronicity is not related to food depen- dence because larval hatching occurs mainly during winter when neither food nor potential competitors are abundant (Lovrich, unpubl. data). Paralomis is a deep-water genus (Takeda et al. 1984; Macpherson, 1988) and P. granulosa is the only spe- cies that inhabits shallow waters. This species prob- ably colonized the Beagle Channel relatively recently, i.e., 8,500 years ago after the last deglaciation occurred (Rabassa et al., 1986). This species still retains certain features of its deep-water relatives: low fecundity, large eggs, protracted reproductive cycle, and independence of larval hatching from food availability. Acknowledgments This study was funded by CONICET (PID # 3154700- 88). We thank the fishing companies Mar Frio S.A. and Pesquera del Beagle S.A. for allowing us to sample on board. Special thanks to Jose T Garcia, Hector Monsalve, and the crew of the fishing boats. Beatriz Gonzalez (Biological Dept, F.C.E.y N., University of Buenos Aires) supervised statistical analysis; Mercedes Cattaneo assisted in correction of the early English version; and Bernard Sainte-Marie (Dept. Fisheries and Oceans, Canada) as well as two anonymous reviewers 674 Fishery Bulletin 91(4). 1993 provided helpful comments that improved the manu- script. Literature cited Abello, P., and E. Macpherson. 1992. Epibiosis and rhizocephalan infestation patterns in relation to the reproductive biology of Lithodes ferox (Filhol, 1885) (Anomura: Lithodidae). J. Crustacean Biol. 12:561-570. Anderson, D. T. 1982. Embryology. In D. E. Bliss (ed.), The biology of Crustacea, Vol 2: Embrology, morphology and genet- ics (LG Abele, ed.), p. 1-41. Academic Press, New York. Blackburn, J., B. A. Johnson, and D. Smith. 1990. A comparison of trawl and pot surveys of red king crab populations [Paralithodes camtschatica) near Kodiak, Alaska. In Proc. Int. Symp. on King and Tanner Crabs, p. 517-531. Alaska Sea Grant College Program Rep. AK-SG-90-04. Blau, S. 1990. Size at maturity of female red king crabs (Paralithodes camtschatica) in the Adak management area, Alaska. In Proc. Int. Symp. on King and Tan- ner Crabs, p. 105-115. Alaska Sea Grant College Pro- gram Rep. AK-SG-90-04. Campodonico, I. 1971. Desarrollo larval de la centolla Lithodes autarch- cus Jacquinot, en condiciones de laboratorio. (Crusta- cea. Decapoda, Anomura, Lithodidae). An. Inst. Patagonia, 2:181-190. Campodonico, I., and L. Guzman. 1981. Larval development of Paralomis granulosa (Jacquinot) under laboratory conditions. (Decapoda, Anomura, Lithodidae.) Crustaceana 40:272-285. Campodonico, I., M. B. Hernandez, and E. Riveros. 1983. Investigation, manejo y control de las pesquerias de centolla y centollon de la XIP Region. Informe consolidado: Recurso Centollon. Inf. Inst. Patagonia 25, 97 p. Comoglio, L. I., J. H. Vinuesa, and G. A. Lovrich. 1990. Feeding habits of southern king crab Lithodes santolla and false southern king crab Paralomis granulosa in the Beagle Channel. In Proc. Int. Symp. on King and Tanner Crabs, p. 315-325. Alaska Sea Grant College Program Rep. AK-SG-90-04. Gabe, M. 1958. Techniques histologiques. Masson et Cie., Paris, 1113 p. Guzman, M. L., and I.Campodonieo. 1972. Fecundidad de la centolla Lithodes antarctica Jacquinot (Crustacea, Decapoda, Anomura, Lithodi- dae). An. Inst. Patagonia 3: 249-258. Hartnoll, R. G. 1978. The determination of relative growth in Crus- tacea. Crustaceana 34:281-293. Jensen, G. C, and D. A. Armstrong. 1989. Biennial reproductive cycle of blue king crab, Paralithodes platypus, at the Pribilof Islands, Alaska and comparison to a cogener, P. camtschatica. Can. J. Fish. Aquat. Sci. 46:932-940. Jewett S. C, N. A. Sloan, and D. A. Somerton. 1985. Egg number and adult size in Lithodes aequi- spina from northern British Columbia. //; Proc. Int. King Crab Symp., p. 137-143. Alaska Sea Grant Col- lege Program Rep. AK-SG 85-12. Kuris, A. M. 1991. A review of patterns and causes of crustacean brood mortality. In A Wenner and A. Kuris, (eds.), Crustacean egg production, p. 117-141. A. A. Balkema, Rotterdam, The Netherlands. Lovrich, G. A. 1991. Reproduction y crecimiento del centollon, Para- lomis granulosa (Crustacea, Anomura, Lithodidae) en el Canal Beagle. Tesis doctoral inedita, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, 160 p. Mc Mullen, J. 1969. Effects of delayed mating on the reproduction of king crab Paralithodes camtschatica. J. Fish. Res. Board Can. 26:2337-2740. Macpherson, E. 1988. Revision of the family Lithodidae Samouelle, 1819 (Crustacea, Decapoda, Anomura) in the Atlantic Ocean. Monografias de Zoologia Marina 2:9-153. Matsuura, S., K. Takeshita, H. Fujita, and S. Kawasaki. 1971. Reproduction and fecundity of the female king crab, Paralithodes camtschatica (Tilesius) in the wa- ters off western Kamchatka. I. Observations of the ovarian eggs and the spawned eggs attached to pleopods. Bull. Far Seas Fish. Res. Lab. 5:147-159. 1972. Reproduction and fecundity of the female king crab, Paralithodes camtschatica (Tilesius) in the wa- ters off western Kamchatka. II. Determination of the fecundity based on the counts of the ovarian eggs and of the spawned eggs attached to pleopods. Bull. Far Seas Fish. Res. Lab. 8:169-190. Miquel, J. C, and P. M. Arnaud. 1987. Aspects de la biologie de Lithodes murrayi (Crus- tacea: Decapoda) aux lies Crozet, Ocean Indien Subantarctique. (SCAR/SCOR Group of Specialists on Southern Ocean Ecosystems and their living resources: colloque sur l'ecologie marine des iles subantarctiques et antarctiques. Museum National d'Histoire Natur- elle, Paris, 25 juin 1985). CNFRA 57:81-89. Miquel, J. C, P. M. Arnaud, and T. Do-Chi. 1985. Population structure and migration of the stone crab Lithodes murrayi in the Crozet Islands, Subant- arctic Indian Ocean. Mar. Biol. (Berl.) 89:263-269. Miller, F. J. 1990. Effectiveness of crab and lobster traps. Can. J. Fish. Aquat. Sci. 47:1228-1251. Morrison, D. F. 1976. Multivariate statistical methods. McGraw-Hill- Kogakusha, Ltd., 415 p. Lovnch and Vinuesa Reproduction of Paralomis granulosa 675 Otto, R. S., and P. A. Cummiskey. 1985. Observations on the reproductive biology of golden king crab (Lithodes aeqinspina) in the Bering Sea and Aleutian Islands. In Proc. Int. King Crab Symp.. p. 123-135. Alaska Sea Grant College Pro- gram Rep. AK-SG-90-04. Otto, R. S., R. A. Macintosh, and P. A. Cummiskey. 1990. Fecundity and other Reproductive parameters of female red king crab (Paralithodes camtschatica) in Bristol Bay and Norton Sound, Alaska. In Proc. Int. Symp. on King and Tanner Crabs, 65-90. Alaska Sea Grant College Program Rep. AK-SG-90-04. Paul, J. M., and A. J. Paul. 1990. Reproductive success of sublegal size male red king crab with access to multiple mates. In Proc. Int. Symp. on King and Tanner Crabs, p. 37-50. Alaska Sea Grant College Program Rep. AK-SG-90-04. Powell, G. C, K. E. James, and C. L. Hurd. 1973. Ability of male king crab Paralithodes cam- tschatica to mate repeteadly. Fish. Bull. 72:171-179. Rabassa, J. O., C. Heusser, and R. Stuckenrath. 1986. New data on Holocene sea transgression in the Beagle Channel: Tierra del Fuego, Argentina. Quat. South America and Antartic Peninsula 4:291-309. Smith, C. C, and S. D. Fretwell. 1974. The optimal balance between size and number of offspring. Am. Nat. 108:499-506. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, the principles and practice of statis- tics in biological research. W.H. Freeman and Co., N.Y., New York, 859 p. Somerton, D. A. 1980. A computer technique for estimating the size of sexual maturity in crabs. Can. J. Fish. Aquat. Sci. 37:1488-1494. 1981. Contribution to the life history of the deep- sea king crab, Lithodes couesi, in the Gulf of Alaska. Fish. Bull. 79:259-269. Somerton, D. A., and R. A. Macintosh. 1985. Reproductive biology of the female blue king crab Paralithodes platypus near the Pribilof Islands, Alaska. J. Crustacean Biol. 5:365-376. Somerton, D. A., and R. S. Otto. 1986. Distribution and reproductive biology of the golden King Crab, Lithodes aequispina, in the East- ern Bering Sea. Fish. Bull. 84:571-584. Takeda, M., K. Hiramoto, and Y. Suzuki. 1984. Additional material of Paralomis cristata Takeda et Ohta (Crustacea, Decapoda) from Suruga Bay. Japan. Bull. Biogeogr. Soc. Jpn. 39:27-31. Vinuesa, J. H. 1984. Sistema reproductor, ciclo y madurez gonadal de la centolla (Lithodes antarcticus) del Canal Beagle. In Estudio biologico pesquero de la centolla (Lithodes antarcticus) del Canal Beagle, Tierra del Fuego. Argentina; 1984. Contrib. Inst. Nac. Invest. Desarr. Pesq. (Argent.) 441:75-95. 1985. Differential aspects of the southern king crab (Lithodes antarcticus) in two latitudinally separated locations. In Proc. Int. King Crab Symp, p. 267- 279. Alaska Sea Grant College Program Rep. AK- SG85. 1987. Embryonary development of Lithodes antarcticus Jacquinot (Crustacea, Decapoda, Lithodidae) devel- opmental stages, growth and mortality. Physis (B. Aires) 45:21-29. Vinuesa, J. H., G. A. Lovrich, and L. I. Comoglio. 1989. Temperature-Salinity effects on the development of false southern King Crab (Paralomis granulosa). Thalassas 7:51-56. Wenner, A. M., C. Fusaro, and A. Oaten. 1974. Size at onset of sexual maturity and growth rate in crustacean populations. Can. J. Zool. 52:1095- 1106. Abstract.— Modern methods for fish stock assessment are often based on age-structured models that sepa- rate each coefficient of fishing mor- tality at age into a time-specific fac- tor (the rate of fishing mortality on the fully exploited age-classes) and an age-specific factor (a selectivity coefficient that measures the rela- tive vulnerability of the particular age-class). The assumption that the selectivity coefficients are constant through time greatly simplifies the assessment process because it allows for a reduction in the number of un- known parameters. However, if the assumption is incorrect, it can lead to incorrect estimates of stock status. The most recent stock assessment for Pacific widow rockfish (Sebastes entomelas) was based on the un- tested assumption that the selectiv- ity coefficients have not changed over the years. This assessment was de- rived from an analysis of catch-at- age data by using an assessment method known as the Stock Synthe- sis program. The work described here examined the sensitivity of the assessment results to the assump- tion of constant selectivity. Simula- tion experiments with the Stock Syn- thesis program showed that the stock size estimates for widow rock- fish can be highly sensitive to mod- est changes in selectivity. Experi- ments with two other assessment techniques, which also assume con- stant selectivity (the CAGE AN pro- gram of Deriso, Quinn, and Neal and the multiplicative catch-at-age model of Shepherd and Nicholson), showed that these methods are similarly sensitive to changes in selectivity. The assumption of constant selectivity and the stock assessment for widow rockfish, Sebastes entomelas David B. Sampson Coastal Oregon Marine Experiment Station Hatfield Marine Science Center, Oregon State University Newport, OR 97365 Manuscript accepted 21 May 1993. Fishery Bulletin 91:676-689 1993) In general, individual fish in a stock are not equally likely to be caught and different age-classes of fish do not ex- perience identical rates of fishing mor- tality. In the fisheries literature this phenomenon is usually described as "selectivity" or "availability" or "par- tial recruitment" (Megrey, 1989). In some instances selectivity results from the physical properties of the fishing gear. For example, younger and smaller fish may pass unharmed through the meshes of a trawl, whereas older and larger individuals may sense and avoid an approaching net. Alternatively, selectivity can re- sult when different age-classes of fish occupy geographic regions that are not fished with the same intensity. If younger fish are offshore and older ones are inshore, for example, then the age distribution of fish in the catch will depend not just on the stock's age distribution but also on the fishing locations. Selectivity coefficients, which measure the relative influence of fishing on the age structure of the stock, are fundamental parameters in the analysis of catch-at-age data. Many stock assessment procedures attempt to reconcile observations of catch-at-age with an underlying age- structured population model and thereby reconstruct the demographic history of the stock (Megrey, 1989). If different age-classes experience the same relative susceptibility to fishing each year, then one can model each annual age-specific rate of fishing mortality as the simple prod- uct (Sa-Fy) of an age effect (the selec- tivity coefficient, Sa) and a year ef- fect (the fishing mortality coefficient, Fy). Because there are an infinite number of (Sa, Fy) pairs that corre- spond to a given age-specific rate of fishing mortality, the selectivity co- efficient for at least one age class must be assumed constant. If the largest S„ is set equal to one, then the FY values correspond to the rate of fishing mortality on the fully ex- ploited age classes. Formulating the fishing mortality coefficient as the product of a year- effect and an age-effect greatly sim- plifies an analysis of catch-at-age data because it reduces the number of essential parameters. For example, if the catch-at-age matrix contains data for A ages and Y years, and if the selectivity coefficients are con- stant for all years, then there are only iA+Y) unknown parameters. How- ever, if the selectivity coefficients change every year, then there are (A-F) unknown parameters. Constant selectivity, and the con- sequent separability of fishing mor- tality into age and year effects, is a fundamental assumption for numer- ous stock assessment procedures, in- cluding separable Virtual Population Analysis (Pope and Shepherd, 1982), the CAGEAN program (Deriso et al., 1985, 1989), the multiplicative model of Shepherd and Nicholson (1986, 1991), and the Stock Synthesis pro- gram (Methot 1989, 1990). Fish stocks that have recently been as- 676 Sampson Constant selectivity and stock assessment for Sebastes entomelas 677 sessed by using one or more of these procedures in- clude Pacific halibut (Hippoglossus stenolepis) in the northeast Pacific (IPHC, 1991), walleye pollock (Theragra chalcogramma) in the Gulf of Alaska (Megrey, 1991), Dover sole (Microstomus pacificus) along the U.S. west coast (Turnock and Methot, 1991), and scad {Trachurus trachurus) from Atlantic waters off Spain and Portugal (Borges, 1990). Despite the widespread application of assessment methods that are based on the notion that selectivity is time-invariant, I know of no published studies that examine the sensitivity of these assessment procedures to violations of the constant selectivity assumption. It appears that often these assessment methods are ap- plied without first verifying that selectivity was con- stant for the stock being assessed. Gudmundsson ( 1986) recommended extensive analysis to avoid mis-specify- ing the catch-at-age model (for example, incorrectly assuming that selectivity was constant), and he devel- oped a least-squares technique for testing the separa- bility assumption. However, his methodology does not seem to be used widely. Several published papers document variations in se- lectivity through time. Houghton and Flatman (1981) examined selectivity coefficients for cod (Gadus morhua) in the west-central North Sea and found sig- nificant changes in the "exploitation pattern," which they attributed to shifts in the fishing pressure ex- erted by different segments of the fleet. Gudmundsson (1986) speculated that changes in fish size-at-age coupled with variations in the composition of the fishing gear caused changes in selectivity for the Icelandic stock of cod (Gadus morhua). Gordoa and Hightower (1991) analyzed data from the fishery for Cape hake (Merluccius capensis) off southwestern Africa and as- cribed significant shifts in selectivity to the fishermen's targeting on strong year classes. The work described here has a different focus. In this paper I do not examine how selectivity in a fishery has varied. Instead, I investigate whether the stock assessment program used to evaluate a particular stock's status is robust to changes in selectivity. For this exercise I analyzed the stock assessment for widow rockfish (Sebastes entomelas), an economically impor- tant component of the complex of Sebastes species found along the Pacific coast of North America. Gunderson ( 1984) described the history and character- istics of the U.S. fishery for widow rockfish off the coasts of Washington, Oregon, and California. In this paper, I demonstrate that the stock size esti- mates for widow rockfish, which are based on the un- tested assumption that selectivity has been constant from year to year, can be seriously biased if the as- sumption is violated. Furthermore, I test two other assessment methods that also use the constant selec- tivity assumption and show that they produce simi- larly biased results. Finally, I establish that compa- rable problems with bias can arise in the assessment results for other fish stocks, whose biological charac- teristics differ significantly from widow rockfish. Methods One technique for testing the reliability of an estima- tion procedure is to produce artificial data sets with known characteristics and then to estimate the pa- rameter values from which the data were derived. I used this approach to determine whether certain stock assessment methods were sensitive to violations of the constant selectivity assumption. First, I simulated catch-at-age data for a fish stock in which the selectiv- ity coefficients were changing slowly from year to year. Next, I used the assessment programs to analyze the catch-at-age data and to estimate stock biomass and abundance-at-age. Finally, I measured the bias of the estimates by calculating the relative errors of the esti- mates. The relative error of an estimate is the differ- ence between the estimate and its true value; all di- vided by the true value. Sensitivity of the Stock Synthesis program when applied to data for widow rockfish To investigate the sensitivity of the stock assessment results for widow rockfish to the assumption of con- stant selectivity, I developed a spreadsheet model to generate artificial catch-at-age data, which I then ana- lyzed using the Stock Synthesis program. The Stock Synthesis program has been used by the Pacific Fish- eries Management Council (PFMC) since 1990 to ap- praise the status of many of the Pacific groundfish stocks (PFMC, 1990). Methot (1989, 1990) documented the principles and equations underlying the Stock Syn- thesis program. The spreadsheet model simulates the characteris- tics of an age-structured population and employs the same equations as the Stock Synthesis program for describing the temporal progressions in abundance- at-age, biomass, catch-at-age, and total catch. The model uses parameters for mortality and growth that are similar to those observed in the U.S. stock of widow rockfish (Table 1). I generated values for abundance- at-age and catch-at-age for 10 years and 20 age classes, ages 4 through 22, as well as age 23 and older (Table 2). In 1989, almost 95% of the U.S. coast-wide land- ings of widow rockfish were fish between the ages of 5-10 years; about 3% of the landings were fish older than 15 years (Hightower and Lenarz, 1990). 678 Fishery Bulletin 9I|4). 1993 Table 1 Parameters for simulating a stock of widow rockfish: annual recruitment was 10" fish per year; natural mortality1 was 0.15 per year. Selectivity parameters- (Curve B (Fig. 1), 100% selection at age 8): lower inflection age was 6.0 years; lower slope was 2.5 per year; upper inflection age was 12.0 years; and the upper slope was 0.3 per year. Age Selectivity Weight3 (years) coefficients (%) (kg) 4 0.8 0.549 5 8.9 0.662 6 56.2 0.780 7 99.0 0.936 8 100.0 1.025 9 93.1 1.150 10 84.6 1.255 11 75.2 1.400 12 65.5 1.427 13 55.7 1.596 14 46.4 1.798 15 37.9 1.899 16 30.3 1.965 17 23.9 1.965 18 18.6 2.008 19 14.3 2.099 20 10.9 2.031 21 8.2 2.704 22 6.2 2.299 23+ 4.7 2.388 'From Hightower and Lenarz ( 1990). -The equation for the double-logistic selec- tion curve and the meaning of these param- eters are described in Methot ( 1990). 'From Barss and Echeverria (1987). In the first set of experiments, I exam- ined combinations of three factors to de- termine how they affect bias in the esti- mates from the Stock Synthesis program. They were 1) the age of full selection was either gradually increasing or decreasing from year to year, or it varied randomly; 2) the annual fishing mortality coefficients were either increasing, decreasing, or con- stant; and 3) the program was either given the true values of annual recruit- ment or it was required to estimate these values. For simplicity, I limited my ex- periments to these three factors, although undoubtedly there are others that can also have significant effects. Examples are trends in annual recruitment, the level of natural mortality, or the shape of the selectivity curve. True values for the selectivity coeffi- cients were generated from a double lo- gistic function (Methot, 1990) and were similar to those reported for widow rockfish in Hightower and Lenarz (1990). The strongly domed shape of the selectivity curve (Fig. 1), which may be due to the movement of older individuals into deeper, less heavily fished waters, seems to be a common feature for many of the groundfish stocks in the U.S. Pacific Northwest. To simulate temporal changes in selectivity, I shifted the selectiv- ity coefficients forward or backward by one age class (Fig. 1). When selectivity increased, 100% selection occurred at age 7 for the first three years (Curve A), at age 8 for the next four years (Curve B), and at age 9 for the last three years (Curve C). When selectivity decreased, full selection occurred at age 9 for the first three years, at age 8 for the next four years, and at age 7 for the last three years. To measure the effects of "random" changes in selectivity, I generated data sets for 10 trials. Selectivity in the first year of each trial always followed selection curve B, but the sequence of curves that applied in the subsequent years came from a random shuffling of the sequence AAABBBCCC (Fig. 2). Curve A applied in three randomly chosen years, curve B applied in three other randomly selected years, and curve C applied in the remaining three years. I did not examine other forms of change in selectivity, such as varia- tion in the basic shape of the curve. When simulating an increasing trend in fishing mortality, the fishing mortality coefficients changed linearly from 0.10 to 0.28 per year, at increments of 0.02 per year. When the trend was decreas- ing, the fishing mortality coefficients varied from 0.28 to 0.10 per year, at increments of -0.02 per year. When there was no trend in fishing mortality, the fishing mortality was 0.20 per year, which is approximately the rate of fishing that reduces the reproductive out- put from this simulated stock to 35% of its unexploited level when full selection is at age 8 years (Fig. 1, Curve B). All methods for analyzing catch-at-age data require additional information with which to tune the analysis and thereby resolve a basic indeterminacy in the model for catch1 (Shepherd and Nicholson, 1986). In the most recent assessment for widow rockfish, Hightower and Lenarz (1990) tuned the Stock Synthesis analysis to a single fishing mortality coefficient, but in many contemporary assessments of other Pacific groundfish stocks the Stock Synthesis runs have been tuned to estimates of abundance-at-age or biomass from re- search vessel surveys. In the sensitivity analysis, I gave the Stock Synthesis program auxiliary data for tuning either in the form of the true annual fishing mortality coefficients or the true propor- tions-at-age. Sensitivity of other stock assessment programs that assume constant selectivity To confirm that the assumption of constant selectivity, rather than some unique feature of the Stock Synthesis program, was respon- sible for any bias in the results, I experimented with two other 'Catch-at-age is approximately equal to the product of stock abundance-at-age and fishing mortality-at-age. If only catch data are available, one cannot distinguish between a case of large abundance and low fishing mortality versus one of small abundance and high fishing mortality. Sampson Constant selectivity and stock assessment for Sebastes entomelas 679 Table 2 Simu ated widov i rockfish abundance and catch data. (A) Selectivity increasing. fishing mortality constant (0.20 year) iB) Selectivity decreasing, fish ng mortality constant (0.20/year). Selection curves are ndicted in aarentheses. A Year 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Age (A) (A) (A) (B) (B) (B) (B) (C) (C) (C) Initial population size 1000's offish) 4 1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 5 845.6 845.6 845.6 845.6 859.3 859.3 859.3 859.3 860.6 860.6 6 650.4 650.4 650.4 650.4 715.0 726.6 726.6 726.6 738.4 739.5 7 459.3 459.3 459.3 459.3 500.3 550.0 558.9 558.9 614.5 624.4 8 323.7 323.7 323.7 323.7 324.3 353.3 388.4 394.7 429.9 472.6 9 231.3 231.3 231.3 231.3 228.1 228.5 249.0 273.7 278.7 303.6 10 168.1 168.1 168.1 168.1 165.2 163.0 163.3 177.9 192.9 196.4 11 124.4 124.4 124.4 124.4 122.1 120.1 118.4 118.7 127.1 137.8 12 94.0 94.0 94.0 94.0 92.1 90.4 88.9 87.7 86.3 92.4 13 72.3 72.3 72.3 72.3 70.9 69.6 68.3 67.1 64.9 63.9 14 56.7 56.7 56.7 56.7 55.7 54.6 53.6 52.6 50.7 49.0 15 45.3 45.3 45.3 45.3 44.5 43.7 42.8 42.0 40.5 39.0 16 36.7 36.7 36.7 36.7 36.1 35.5 34.9 34.2 33.0 31.8 17 30.1 30.1 30.1 30.1 29.7 29.3 28.8 28.2 27.3 26.3 18 25.0 25.0 25.0 25.0 24.7 24.4 24.0 23.6 22.9 22.1 19 20.9 20.9 20.9 20.9 20.7 20.5 20.2 19.9 19.4 18.8 20 17.6 17.6 17.6 17.6 17.5 17.3 17.1 16.9 16.5 16.1 21 14.9 14.9 14.9 14.9 14.8 14.7 14.6 14.4 14.1 13.8 22 12.7 12.7 12.7 12.7 12.6 12.5 12.5 12.3 12.1 11.9 23+ 74.3 74.3 74.3 74.3 74.1 73.9 73.7 73.4 72.8 72.2 Catch UOOOVs offish) 4 16.30 16.30 16.30 1.49 1.49 1.49 1.49 0.13 0.13 0.13 5 83.61 83.61 83.61 13.78 14.01 14.01 14.01 1.28 1.28 1.28 6 108.74 108.74 108.74 64.31 70.70 71.85 71.85 11.85 12.04 12.06 7 77.51 77.51 77.51 76.78 83.64 91.95 93.44 55.26 60.75 61.74 8 51.17 51.17 51.17 54.62 54.73 59.62 65.54 65.98 71.88 79.02 9 33.49 33.49 33.49 36.56 36.06 36.13 39.36 46.18 47.03 51.23 10 21.85 21.85 21.85 24.34 23.93 23.60 23.65 28.12 30.49 31.05 11 14.21 14.21 14.21 16.18 15.88 15.61 15.40 17.19 18.41 19.96 12 9.22 9.22 9.22 10.73 10.52 10.33 10.15 11.40 11.21 12.01 13 5.96 5.96 5.96 7.10 6.96 6.83 6.70 7.67 7.42 7.29 14 3.85 3.85 3.85 4.68 4.59 4.50 4.41 5.16 4.97 4.81 15 2.48 2.48 , 2.48 3.07 3.02 2.96 2.90 3.46 3.34 3.22 16 1.59 1.59 1.59 2.01 1.98 1.94 1.91 2.32 2.23 2.15 17 1.02 1.02 1.02 1.31 1.29 1.27 1.25 1.54 1.49 1.44 18 0.65 0.65 0.65 0.85 0.84 0.83 0.81 1.02 0.99 0.96 19 0.42 0.42 0.42 0.55 0.54 0.54 0.53 0.67 0.66 0.64 20 0.27 0.27 0.27 0.35 0.35 0.35 0.34 0.44 0.43 0.42 21 0.17 0.17 0.17 0.23 0.23 0.22 0.22 0.29 0.28 0.28 22 0.11 0.11 0.11 0.15 0.14 0.14 0.14 0.19 0.18 0.18 23+ 0.48 0.48 0.48 0.64 0.64 0.64 0.63 0.84 0.84 0.83 680 Fishery Bulletin 91(4), 1993 Table 2 (Continued) B Year 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Age (C) (C) iC) iBi (Bl (Bl (Bl (Al (A) (Al Initial population size (1000^ offish) 4 1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 5 860.6 860.6 860.6 860.6 859.3 859.3 859.3 859.3 845.6 845.6 6 739.5 739.5 739.5 739.5 727.7 726.6 726.6 726.6 661.0 650.4 7 625.3 625.3 625.3 625.3 568.8 559.8 558.9 558.9 513.1 466.7 8 481.0 481.0 481.0 481.0 441.6 401.7 395.3 394.7 393.9 361.6 9 339.7 339.7 339.7 339.7 339.0 311.2 283.1 278.5 282.0 281.4 10 239.4 239.4 239.4 239.4 242.7 242.2 222.3 202.2 202.4 205.0 11 171.0 171.0 171.0 171.0 174.0 176.4 176.0 161.6 149.8 149.9 12 124.3 124.3 124.3 124.3 126.6 128.8 130.6 130.3 122.0 113.1 13 92.0 92.0 92.0 92.0 93.8 95.6 97.3 98.6 100.3 93.9 14 69.5 69.5 69.5 69.5 70.9 72.3 73.6 74.9 77.3 78.7 15 53.5 53.5 53.5 53.5 54.5 55.6 56.7 57.7 59.7 61.7 16 42.0 42.0 42.0 42.0 42.7 43.5 44.3 45.2 46.8 48.4 17 33.5 33.5 33.5 33.5 34.0 34.6 35.2 35.9 37.1 38.4 18 27.1 27.1 27.1 27.1 27.5 27.9 28.4 28.9 29.8 30.8 19 22.3 22.3 22.3 22.3 22.5 22.8 23.1 23.5 24.2 24.9 20 18.5 18.5 18.5 18.5 18.6 18.8 19.1 19.3 19.8 20.4 21 15.4 15.4 15.4 15.4 15.5 15.7 15.8 16.1 16.4 16.8 22 13.0 13.0 13.0 13.0 13.1 13.2 13.3 13.4 13.6 13.9 23+ 73.4 73.4 73.4 73.4 73.7 73.9 74.2 74.6 75.2 75.9 Catch (1000's »f fish) 4 0.13 0.13 0.13 1.49 1.49 1.49 1.49 16.30 16.30 16.30 5 1.28 1.28 1.28 14.03 14.01 14.01 14.01 84.97 83.61 83.61 6 12.06 12.06 12.06 73.12 71.95 71.85 71.85 121.48 110.50 108.74 7 61.83 61.83 61.83 104.55 95.10 93.58 93.44 94.32 86.59 78.76 8 80.42 80.42 80.42 81.17 74.52 67.78 66.70 62.40 62.27 57.17 9 57.32 57.32 57.32 53.70 53.59 49.20 44.75 40.34 40.84 40.75 10 37.84 37.84 37.84 34.66 35.14 35.07 32.20 26.29 26.31 26.64 11 24.77 24.77 24.77 22.23 22.61 22.93 22.88 18.45 17.10 17.12 12 16.16 16.16 16.16 14.19 14.46 14.71 14.91 12.79 11.97 11.09 13 10.51 10.51 10.51 9.03 9.21 9.38 9.54 8.13 8.27 7.74 14 6.82 6.82 6.82 5.73 5.84 5.96 6.07 5.08 5.24 5.34 15 4.41 4.41 4.41 3.63 3.69 3.77 3.84 3.16 3.27 3.37 16 2.84 2.84 2.84 2.29 2.33 2.38 2.43 1.96 2.03 2.10 17 1.83 1.83 1.83 1.45 1.47 1.50 1.53 1.22 1.26 1.30 18 1.18 1.18 1.18 0.92 0.93 0.95 0.96 0.76 0.78 0.81 19 0.75 0.75 0.75 0.58 0.59 0.60 0.61 0.47 0.48 0.50 20 0.48 0.48 0.48 0.37 0.37 0.38 0.38 0.29 0.30 0.31 21 0.31 0.31 0.31 0.23 0.24 0.24 0.24 0.18 0.19 0.19 22 0.20 0.20 0.20 0.15 0.15 0.15 0.15 0.12 0.12 0.12 23+ 0.84 0.84 0.84 0.63 0.63 0.64 0.64 0.48 0.48 0.49 assessment methods that also use the assumption of constant selectivity. I analyzed the simulated catch-at- age data with the CAGEAN program (CAGEAN-PC, version 4, release 2) of Deriso et al. (1985, 1989) and with the multiplicative catch-at-age model of Shep- herd and Nicholson ( 1986. 1991 ). The Stock Synthesis program and the CAGEAN pro- gram use similar approaches for modeling catch-at- age, but they differ in their assumptions about vari- ability in the observed data. Stock Synthesis assumes a multinomial error structure for the catch-at-age data. If pa is the predicted proportion of age class (a) cap- tured, then the variance associated with a random ob- servation ofp„ is proportional to p„- (1-pJ. CAGEAN, however, assumes a lognormal error structure. If ca is the predicted number of fish caught of age class (a ), then log, ,(ca/ca) is normally distributed with a zero mean and a constant variance. The multiplicative catch-at-age model is essentially an approximation to the catch model that underlies Sampson Constant selectivity and stock assessment for Sebastes entomelas 681 100 UJ 40 o CURVE B CURVE C SELECTION YEAR TREND CURVE 123456789 10 INCREASING B C A DECREASING B C Figure 1 Selectivity curves used in the simulations. To simulate the effects of changing selectivity, the selectivity curve was either shifted towards older fish through time (increasing selectiv- ity) or towards younger fish (decreasing selectivity). both the Stock Synthesis and CAGEAN programs. Un- like the other two assessment procedures, the multi- plicative model estimates relative, rather than abso- lute, abundance. To conduct the multiplicative catch-at-age analyses, I used the GLIM statistical pro- gram (Baker and Nelder, 1985) and assumed a log- normal error structure. In the experiments with CAGEAN and the multipli- cative catch-at-age model, I used a subset of the data from the earlier experiments with the Stock Synthesis program. Two sets of catch-at-age data were analyzed, one from a population with selectivity shifting to older ages (selectivity increasing) and fishing mortality con- stant (Table 2A), the other from a population with selectivity shifting to younger ages (selectivity decreas- ing) and fishing mortality constant (Table 2B). I tuned the CAGEAN program to the true fishing mortality coefficients, and constrained the multiplicative catch- at-age analysis to have a trend of zero in the annual fishing mortality coefficients. The experiments here cor- respond to the cases examined earlier in which the Stock Synthesis program was tuned to fishing mortal- ity and recruitment was estimated. Sensitivity of the Stock Synthesis program when applied to data from a heavily exploited stock To determine whether the results from the experiments with a simulated stock of widow rockfish would apply to fish stocks with different biological characteristics, I generated two additional data sets, one from a popu- lation with selectivity shifting to older ages (selectiv- ity increasing), the other from a population with selec- tivity shifting to younger ages (selectivity decreasing). Both simulated populations, which suffered an instan- taneous natural mortality rate of 0.30 per year and an instantaneous fishing mortality rate of 0.60 per year, had significantly fewer old animals compared to the populations in the previous simulations. I analyzed the two data sets with the Stock Synthesis program, with tuning to the true fishing mortality coefficients,' and with recruitment estimated. Results The assessment programs that I investigated all pro- duce a wide variety of estimates, including selectivity coefficients and matrices of abundance and catch by age and year. Rather than evaluating bias for all esti- mates, my analysis focussed on estimates of annual stock biomass, numerical abundance, and recruitment. Of special importance to a stock assessment scientist or fishery manager is the bias in the estimate of aver- age biomass for the final year of a data series. This estimate is approximately the biomass estimate on which the catch quota for the next year is based2. If the estimate of average biomass in the final year is, say, 20% too high, then the quota will be roughly 20% too high; if the estimate is 10% too low, the quota will also be about 10% too low. Selectivity of the Stock Synthesis program when applied to data for widow rockfish The results of the experiments with the Stock Synthe- sis program and the data for the simulated stock of widow rockfish suggest that some of the assessment results can be highly sensitive to slight trends in se- lectivity. For example, when selectivity shifted towards younger ages, the biomass estimate for the final year of the series was 74% too high (Table 3 A; selectivity decreasing, tuned to fishing mortality, fishing mortal- •The annual catch quota is derived from an estimate of the biomass at the end of the previous year, plus an appropriate amount for the new recruitment. In practice this differs little from the estimate of average biomass in the final year. 682 Fishery Bulletin 91(4), 1993 Figure 2 Simulations with random variation in selectivity. The effects of random varia- tion in selectivity were simulated by shuffling the sequence of selectivity curves (AAABBBCCC, Fig. 1) that applied in each of the nine years after the first. The ten sequences shown here were used. ity constant, and recruitment estimated), and the esti- mated age distribution in the final year was grossly incorrect (Fig. 3A). The estimated numbers of five-year- old to seven-year-old fish were much too high and the numbers of fish 15 years and older were all slightly too high. When selectivity shifted towards older ages, the biomass estimate for the final year was as much as 597c too low (Table 3A; selectivity increasing, tuned to fishing mortality, fishing mortality increasing, and recruitment estimated) and the program underesti- mated the numbers of very young and very old fish (Fig. 3B). In these experiments, bias in a particular estimate was not a simple linear function of the factors exam- ined, but instead involved complicated interactions be- tween factors. Nevertheless, some general effects seemed to apply. When the Stock Synthesis program was tuned to fishing mortality and used to estimate recruitment, shifts in selectivity towards older ages (selectivity increasing) always induced negative bias in the estimates of average biomass, and shifts in selectiv- ity towards younger ages (selectivity de- creasing) always induced positive bias (Table 3, A and B). When the recruit- ment values were known or tuning to the proportion-at-age data was used, however, trends in selectivity had no con- sistent effect on the direction of bias. Estimation of recruitment values, often, but not always, increased the magnitude of the bias in the estimates of bio- mass (Table 3, A and B) and abundance (Table 3C). Tuning to proportion-at- age data, instead of to annual fishing mortality coefficients, often decreased the amount of bias in the estimates of biomass, abundance, and recruitment. Bias in these estimates usually was smallest when the trend in fishing mortality was increasing and largest when the trend in fishing mortality was decreasing. When the Stock Synthesis program es- timated recruitment, improvements in the fit were observed relative to those obtained when recruitment values were known (Table 3D). To the assessment sci- entist interpreting these results, the im- proved fit would suggest that the pro- gram had provided better estimates, when, in fact, the estimates were more biased and less reliable. When selectiv- ity shifted toward older fish, there was a systematic change from year to year in the catch-at-age data, which the program attempted to match by imposing a decreasing trend in recruitment (Fig. 4). When selec- tivity shifted towards younger fish, the program im- posed an increasing trend in recruitment. Similar dis- tortions occurred when the program was tuned to the true proportion-at-age data. This last result suggests that using age-frequency data from research vessel sur- veys will not eliminate the bias induced by changes in selectivity, even though survey data may not be sub- ject to the changes in selectivity that the fishery might experience. When selectivity varied randomly (Table 4), the mag- nitude of the bias in the estimate of the final year's average biomass was usually less than what occurred when selectivity had a trend (Table 3A), but the gen- eral patterns seen in the earlier experiments remained. Sampson. Constant selectivity and stock assessment for Sebastes entomelas 683 Table 3 Sensitivity analysis of the Stock Synthesis program. (A) Bias' in the estimated average biomass in the final year. (B) Average and standard deviation2 of the bias in the estimates of annual average biomass. lO Average and standard deviation of the bias in the estimates of annual numerical abundance. (Dl Average and star dard deviation of the bias in the estimates of annual recruitment1 and the improvement in fit1 when recruitment was estimated. A Increasing selectivity Decreas ng selectivity Known recruit % ment Estimated recruitment Kn <7r own recruitment % Estimated recruitment % Tuned to fishing mortality F increasing -1.0 -58.8 -1.2 49.8 F decreasing 16.9 -54.4 13.6 72.0 F constant 12.1 -54.0 10.4 74.3 Tuned to proportion-at-age F increasing 3.4 -26.1 -2.9 1.0 F decreasing 23.3 -24.6 11.7 5.7 F constant 17.3 -14.0 7.4 5.2 B Increasing selectivity Decreasing selectivity Known recruitment Estimated recruitment Known recruitment Estimated recruitment Mean(%) SD(%) Mean(%) SD(%) Mean ( % ) SD(%) Mean(%) SD(%) Tuned to fishing mortality F increasing 3.9 2.1 -32.5 12.7 -4.5 1.3 24.6 16.2 F decreasing 46.0 21.6 -20.8 18.2 21.4 8.0 63.3 18.8 F constant 29.5 11.5 -20.5 17.0 11.8 4.0 53.2 18.2 Tuned to proportion-at-age F increasing 4.5 0.5 -12.1 8.4 -3.8 0.4 -6.3 6.9 F decreasing 51.7 20.2 7.6 20.5 24.9 9.6 9.1 11.0 F constant 32.0 10.0 10.8 15.2 13.7 4.9 4.3 7.9 C Increasing selectivity Decreasing selectivity Known recruitment Estimated recruitment Known recruitment Estimated recruitment Mean ( 9c ) SD(%) Mean (%) SD(%) Mean ( % ) SD(%) Mean 1 '"< i SD(%) Tuned to fishing mortality F increasing 2.9 3.0 -27.3 20.4 -2.1 3.4 23.6 27.7 F decreasing 23.1 14.0 -24 1 24.2 8.4 3.5 40.9 27.7 F constant 15.8 8.9 -21.5 24.3 4.7 2.3 38.5 29.9 Tuned to proportion-at-age F increasing 3.5 1.6 -10.8 14.8 -2.5 1.6 -3.8 14.0 F decreasing 26.8 13.3 -4.4 22.8 10.2 4.5 -1.3 13.6 F constant 17.7 7.9 1.6 19.7 5.6 2.1 -1.2 13.2 684 Fishery Bulletin 91(4). 1993 Table 3 (Continued) Increasing selectivity Decreasing selectivity Mean (%) SD(%) Improvement infit(%) Meaner) Improvement 3D(%) infit(%) 89.8 29.5 95.7 32.2 96.0 31.6 60.0 18.5 61.7 23.8 59.6 21.4 Tuned to fishing mortality F increasing -30.0 45.3 F decreasing -38.0 42.1 F constant -32.8 45.2 Tuned to proportion-at-age F increasing -13.5 40.0 F decreasing -21.9 39.2 ^constant -13.3 41.1 45.7 56.7 51.9 25.6 37.7 30.9 24.3 29.8 30.7 0.5 -4.6 -2.5 'Bias is measured here as the relative error of the estimated value, (estimate - true I / true. 2The Stock Synthesis program estimates average biomass for each year of the input data series. The values here are the averages and standard deviations of the ten annual estimates. 'The Stock Synthesis program can estimate recruitment for each year of the input data series. The values here are the averages and standard deviations of the ten annual estimates. 'The "improvement in fit" is measured here by the relative increase in the value of the log-likelihood when the Stock Synthesis program estimates the annual recruitment rather than being given the true values. This is defined as {V-L )l L where L is the value of the log-likelihood when the program was given the true annual recruitment values, and L' is the value of the log- likelihood when the program estimated the annual recruitment values. The log-likelihood is given by £» J, ■ X„ p».„ log,(pv„ I; where J is the number of fish in the (y )th catch-at-age sample, pya is the observed proportion of fish in the (_v)th sample that are from the (a)th age class, andplo is the predicted proportion offish in the (ylth sample from the (nlth age class (Methot, 19901. 3 - '', A ■ SELECTIVITY DECREASING J ', ■ FISHING MORTALITY CONSTANT ,' '• • TUNED TO FISHING MORTALITY I ■ ', ■ RECRUITMENT ESTIMATED u. 2 - Li. O • ■ MILLIONS ■ ', ESTIMATED 0 1.0 - ;' ME N. 5 10 15 20 y ■ SELECTIVITY INCREASING \ B • FISHING MORTALITY INCREASING \ • TUNED TO FISHING MORTALITY FISH \ ■ RECRUITMENT ESTIMATED u. O \ TRUE §0.5- _i .j \ 5 0.0 ESTIMATED \ 10 15 20 AGE (YEARS) Figure 3 True ve rsus estimated age distributions. For the age distribu- tion shi iwn in the upper panel (A) the Stock Synthesis program product d an estimate of average biomass in the final year that was 74 7r too large. For the distribution in the lower panel (B) the pro jram produced an estimate that was 59% too small. Sensitivity of other stock assessment programs that assume constant selectivity The results of the experiments with CAGEAN and the multiplicative catch-at-age model indicate that these assessment methods, like the Stock Synthesis program, are also sensitive to violations of the constant selectiv- ity assumption. When selectivity increased, both the Stock Synthesis and CAGEAN programs incorrectly produced a large decrease in the biomass of the simu- lated population during the last few years, and large declines in recruitment (Fig. 5). The multiplicative catch-at-age model does not provide estimates of bio- mass, but its estimates of relative recruitment were almost identical to those from the CAGEAN program. When selectivity decreased, the Stock Synthesis pro- gram and the CAGEAN program both overestimated the biomass in the last few years, but the estimates from the CAGEAN program were grossly incorrect ( Fig. 6). Both the CAGEAN program and the multiplicative catch-at-age model estimated very large increases in recruitment. Sensitivity of the Stock Synthesis program when applied to data from a heavily exploited stock When the Stock Synthesis program was applied to data for a simulated population suffering heavy exploita- tion, the program produced estimates of biomass and recruitment that were even more biased than in the Sampson Constant selectivity and stock assessment for Sebastes entomelas 685 TRUE CATCH-AT-AGE 5 10 15 20 AGE (YEARS) 1.0 0.5 - 3 cc u [11 0.0 TUNED TO PROPORTION AT-AGE / TRUE VALUES TUNED TO FISHING MORTALITY '81 '83 '85 YEAR '87 89 Figure 4 Trends in selectivity induce trends in recruitment. True catches-at-age (upper panel I shift through time towards older fish because of the changes in selectivity. Fishing mortality was constant for the entire period. The Stock Synthesis pro- gram attempted to mimic the changes in catch-at-age by gen- erating estimates of recruitment that had a decreasing trend (lower panel). earlier experiments with the simulated stock of widow rockfish. With increasing selectivity, the bias in the estimated average biomass in the final year was -86% (Fig. 7) as opposed to the bias of -54% found earlier (Table 3A; selectivity increasing, tuned to fishing mor- tality, fishing mortality constant, and recruitment es- timated). With decreasing selectivity, the estimated av- erage biomass in the final year was 213% too high (Fig. 8); in the earlier experiment the corresponding estimate was only 74% too high (Table 3 A; selectivity decreasing, tuned to fishing mortality, fishing mortal- ity constant, and recruitment estimated). Discussion In the experiments described in this paper, the assess- ment programs were unable to fit exactly the simu- lated catch-at-age data because the catch model was mis-specified. Selectivity was falsely assumed constant, Table 4 Sensitivity analysis of the Stock Synthesis program. Bias in the estimated average biomass in the final year when selec- tivity varied randomly. Known recruitment Estimated recruitment MeanC7r) SD(%) Mean(7r) SD(%) Tuned to fishing mortality F increasing -1.4 0.8 -19.7 21.9 F decreasing 11.3 4.4 -17.0 24.0 F constant 7.8 3.5 -18.6 23.2 Tuned to proportion-at-age F increasing -0.1 0.8 -4.2 3.5 F decreasing 13.3 4.6 -8.5 5.7 F constant 8.9 3.8 -6.0 4.9 and the biomass and abundance estimates were bi- ased as a consequence. In any real application, not only would the assessment program be ignorant of the true model structure, but the program would also have to contend with "noise" in the data due to measure- POPULATION BIOMASS (1000s mt) I I i CAGEAN STOCK SYNTHESIS "">• '81 '83 '85 ,87 '89 i o - LU 5 O 5 - Q_ 3 - 5 fe 4- 81 '83 '85 '87 '89 LU 2 ...•.'^"^ §3- cr CAGEAN / \ ° ? - LU ' ^ \ S MULTIPLICATIVE cr .*•>■ z 1 - .«****»*'*' •'''' \ Q ^•"-"*' „--' » 111 (1 - Q TRUE < -1 - t Q > Z <-2- CO STOCK SYNTHESIS \ '81 '83 '85 '87 '89 YEAR Figure 6 Estimated biomass and recruitment from different assess- ment programs with selectivity decreasing. The three assess- ment programs were also applied to the simulated widow rockfish data in which selectivity shifted to younger ages and fishing mortality was constant (Table 2B). Again, all three programs were sensitive to the shifts in selectivity. E tn O | 1.5 - en CO < o 10 - CO z O P 5 05 - z> Q_ O Q_ TRUE ESTIMATED \ '81 '83 '85 '87 '89 sz en U_ O £ 1,0 - o 2 1- z LU r? 05 - ID cr o LU cr '"-., TRUE ESTIMATED \ '81 '83 '85 '87 '89 YEAR Figure 7 Stock synthesis estimates of biomass and recruitment for a heavily fished stock with selectivity increasing. The Stock Syn- thesis program was applied to simulated widow rockfish data in which the fishing mortality rate was constant at 0.60 per year, the natural mortality rate was 0.30 per year, and selectivity shifted to older ages. The estimates were even more biased than the ones from the corresponding earlier experiment. merit errors and to randomness in the catch process. A complete analysis of the problem would measure how the assessment program transforms variability in the catch-at-age data into variability in the resulting esti- mates (e.g., Kimura, 1989). Because the catch model with constant selectivity has fewer unknown parameters, when applied to noisy catch-at-age data, the assessment program's estimates could obtain greater precision (but not accuracy) by assuming constant selectivity, even though the assump- tion was incorrect3. However, it seems unlikely that noise in the data could ever reduce the bias resulting from a structural deficiency in the underlying catch- at-age model. '.John Shepherd, Ministry of Agriculture, Fisheries, and Food, Fish- eries Laboratory, Lowestoft, Suffolk, NR.S3 0HT, U.K., pers. commun. April 1992. In the experiments with the simulated catch-at-age data, the year-to-year changes in selectivity were not particularly drastic, but I know of no studies to sup- port my conjecture that they are realistic for the stock of widow rockfish. The simulated changes in selectiv- ity were comparable to those observed by Houghton and Flatman (1981) for North Sea cod and by Gordoa and Hightower (1991) for Cape hake. The fact that experiments with "random" changes in selectivity produced results similar to those from experiments with trends in selectivity confirm that the biased esti- mates were not just artifacts of having a simple trend in selectivity, rather than a more complex type of variation. One surprising result of the experiments with dif- ferent assessment methods was the large discrepancy between the estimates from Stock Synthesis and CAGEAN when selectivity decreased (Fig. 6). The two programs differ primarily in how they account for vari- Sampson Constant selectivity and stock assessment for Sebastes entomelas 687 4.0 8 30 w to < O 2.0 m 3 1.0 ESTIMATED '83 '85 '87 40 - ESTIMATED ,»'\ 3.0 - t 2.0 - / 10 - _.-'' TRUE _. * '83 '85 '87 YEAR '89 Figure 8 Stock synthesis's estimates of biomass and recruitment for a heavily fished stock with selectivity decreasing. The Stock Synthesis program was applied to simulated widow rockfish data in which the fishing mortality rate was constant at 0.60 per year, the natural mortality rate was 0.30 per year, and selectivity shifted to younger ages. Again, the estimates were more biased than the ones from the corresponding earlier experiment. ability in catch-at-age. Stock Synthesis assumes mul- tinomial error, but CAGEAN assumes lognormal er- ror. Using simulation techniques, Kimura (1990) di- rectly compared estimates derived by using these alternative assumptions and found little difference be- tween the estimates obtained. Another difference between Stock Synthesis and CAGEAN is in their method for modeling selectivity. The Stock Synthesis program uses a double-logistic curve to model selectivity as a smooth function of age, but the CAGEAN-PC program estimates the selectiv- ity coefficients independently for each age. Because the true selectivity coefficients were based on double- logistic curves, this difference between the two pro- grams should be only a minor factor. Kimura (1990) found that the assumption of a functional form for selectivity had little effect on his analyses of simu- lated catch-at-age data, provided the true selectivity coefficients conformed to the general shape of the se- lectivity function. The dilemma for the assessment scientist is to de- velop a framework for analyzing fisheries data that is simple to use and yet is adequate to describe the com- plex dynamics of a living and constantly changing fish stock. The assessment scientist has the difficult task of interpreting diverse and possibly conflicting infor- mation. He needs tools with which to weigh these data objectively and to draw from them reliable conclusions about the status of a stock. Stock Synthesis, CAGEAN, and the multiplicative catch-at-age analysis were de- signed to be such tools. In principle, one can use the Stock Synthesis and CAGEAN programs to test for shifts in selectivity. Both programs support a limited form of variable selectivity in which abrupt changes can occur at pre-specified times with constant selectivity during the intervening periods. With either program it is a relatively simple, but tedious, matter to re-analyze the data by using different times for the selectivity changes. For example, I applied the CAGEAN program to the simulated widow rockfish data set given in Table 2A, with the data partitioned into two periods of constant selectivity, and I allowed the timing of the selectivity change to occur between all possible adjacent years. The resulting pat- tern in the residual sums of squares4 (Fig. 9, upper panel) clearly indicates the true change in selectivity that occurred between the third and fourth years. I repeated the process with the data series partitioned into three periods of constant selectivity, one for the first three years, and the other periods for the remain- ing years. The CAGEAN program was able to fit the data exactly when selectivity changed between the sev- enth and eighth years (Fig. 9, lower panel). Although the current versions of the Stock Synthe- sis and CAGEAN programs can be applied in the above fashion to explore systematically for changes in selec- tion, such a brute force approach to model building is extremely repetitious and time-consuming. I hope that the next generation of stock assessment programs will automate this process in a manner similar to existing stepwise regression programs, and thereby allow the user to test rigorously for variations in selectivity. The model used for stock assessment should not force the data to fit a particular structure unless there is evidence that the structure is real or that it does not appreciably distort the results of the assessment. The 'CAGEAN defines the residual sum of squares as II log,k-l-log,(c) I2 where c and c are the observed and predicted catch-at-age. 688 Fishery Bulletin 91(4), 1993 20 3 o 15 3 o 9 2 89/90 '85/86 '87/88 '89/90 SELECTIVITY TRANSITION Figure 9 Residual sum of squares resulting from applications of CAGEAN with changes in selectivity. The CAGEAN program was applied to the simulated widow rockfish data in which selectivity shifted to older ages and fishing mortality was constant (Table 2A). In the upper panel the selectivity coeffi- cients were allowed to vary abruptly between adjacent years, thereby dividing the data into two periods of constant selec- tivity. The minimum in the residual sum of squares corre- sponds to the true change in selectivity that occurred between 1983 and 1984. In the lower panel, selectivity was constant for the first three years but was allowed to vary between adjacent years in the remaining period. The zero in the re- sidual sum of squares corresponds to the true change in se- lectivity that occurred between 1983 and 1984. In these analyses the age at 100% selection was fixed at age 7 for the first selectivity period, at age 8 for the second period, and at age 9 for the third. situation is analogous to an application of two-way analysis of variance (ANOVA). In fitting a two-way ANOVA model, one should test for a significant inter- action term before drawing inferences about the main effects. By the same logic, in fitting a catch-at-age model, one should test for changes in selectivity before concluding that stock size has been increasing or decreasing. I have no real evidence of changes from year to year in the selectivity for widow rockfish off the coasts of Washington, Oregon, and California. However, the work described in this paper demonstrates that an incorrect assumption of constant selectivity can seriously dis- tort an assessment of widow rockfish stock size. Fur- thermore, there is at least one reason to suspect that selectivity for widow rockfish has varied through time. During the early years of the directed fishery for widow rockfish, vessels targeted schools of fish using midwater trawls. With the rapid expansion of the fishery, the Pacific Fishery Management Council began imposing increasingly restrictive limits on the amounts of widow rockfish that could legally be landed from any single fishing trip (Gunderson, 1984). One result of these "trip limits" was a reduction in the landings by midwater trawlers relative to the landings by bottom trawlers. Midwater trawlers accounted for roughly 75% of the widow rockfish landings in Oregon during 1984 through 1988, but they accounted for only 60% in 1990, and for less than 50% in 1991. It seems quite probable that the midwater trawls have different selection charac- teristics than do bottom trawls, and that the shift from a midwater fishery to a bottom fishery would cause changes in selectivity. Any stock assessment model will have to make sim- plifying assumptions to summarize succinctly the ma- jor features of the data. However, in my view the as- sumption of constant selectivity is an unnecessary and misleading oversimplification, use of which can result in catch quotas that are either needlessly con- servative, resulting in immediate losses to the fishing industry, or that are excessively liberal, producing losses in recruitment and catches at a more distant time. Acknowledgments I am grateful to staff at the Oregon Department of Fish and Wildlife facility at Newport, Oregon, for an- swering numerous questions about the fishery for widow rockfish and the Stock Synthesis program. Also, this paper benefited greatly from helpful suggestions by Ronald Hardy, Linda Jones, John Shepherd, and two anonymous referees. Funds for this research were provided by the Oregon Trawl Commission, the Fishermen's Marketing Association, the Oregon De- partment of Fish and Wildlife, and the Agricultural Experiment Station of Oregon State University. I ap- preciate the support and encouragement of these institutions. Literature cited Baker, R. J., and J. A. Nelder. 1985. The GLIM System Release 3.77. Numerical Algorithms Group Ltd, Oxford, 305 p. Sampson. Constant selectivity and stock assessment for Sebastes entomelas 689 Barss, W. H., and T. W. Echeverria. 1987. Maturity of widow rockfish Sebastes entomelas from the Northeastern Pacific, 1977-82. In W. H. Lenarz and D. R. Gunderson leds.i. Widow rockfish, p. 13-18. NOAA Tech. Rep. NMFS 48. Borges, M. F. 1990. Multiplicative catch-at-age analysis of scad (Trachurus trachurus L.) from western Iberian waters. Fish. Res. 9:333-353. Deriso, R. B., T. J. Quinn II, and P. R. Neal. 1985. Catch-age analysis with auxiliary information. Can. J. Fish. Aquat. Sci. 42:815-824. Deriso, R. B., P. R. Neal, and T. J. Quinn II. 1989. Further aspects of catch-age analysis with aux- iliary information. Can. Spec. Publ. Fish. Aquat. Sci. 108:127-135. Gordoa, A., and J. E. Hightower. 1991. Changes in catchability in a bottom-trawl fishery for Cape hake iMerluccius capensisl. Can. J. Fish. Aquat. Sci. 48:1887-1895. Gudmundsson, G. 1986. Statistical considerations in the analysis of catch-at-age observations. J. Cons. int. Explor. Mer 43:83-90. Gunderson, D. R. 1984. The great widow rockfish hunt of 1980-1982. N. Am. J. Fish. Manage. 4:465-468. Hightower, J. E., and W. H. Lenarz. 1990. Status of the widow rockfish fishery in 1990. In Pacific Fishery Management Council, Status of the Pacific coast groundfish fishery through 1990 and recommended acceptable biological catches for 1991: stock assessment and fishery evaluation. Appendix F. Pacific Fishery Management Council, Metro Cen- ter, Portland, OR 97201. Houghton, R. G., and S. Flatman. 1981. The exploitation pattern, density-dependent catchability, and growth of cod (Gadus morhual in the west-central North Sea. J. Cons. int. Explor. Mer 39:271-287. IPHC (International Pacific Halibut Commission). 1991. Annual Report 1990. Int. Pacific Halibut Comm., Seattle, WA, 52 p. Kimura, D. K. 1989. Variability, tuning, and simulation for the Doubleday-Deriso catch-at-age model. Can. J. Fish. Aquat. Sci. 46:941-949. 1990. Approaches to age-structured separable sequen- tial population analysis. Can. J. Fish. Aquat. Sci. 47:2364-2374. Megrey, B. A. 1989. Review and comparison of age-structured stock assessment models from theoretical and applied points of view. Am. Fish. Soc. Symp. 6:8-48. 1991. Population dynamics and management of wall- eye pollock (Theragra chalcogramma) in the Gulf of Alaska, 1976-1986. Fish. Res. 11:321-354. Methot, R. D. 1989. Synthetic estimates of historical abundance and mortality for northern anchovy. Am. Fish. Soc. Symp. 6:66-82. 1990. Synthesis model: an adaptable framework for analysis of diverse stock assessment data. Int. N. Pac. Fish. Comm. Bull. 50:259-277. PFMC (Pacific Fishery Management Council). 1990. Status of the Pacific coast groundfish fishery through 1990 and recommended acceptable biological catches for 1991: stock assessment and fishery evaluation. Pacific Fishery Management Council, Metro Center, Portland, OR 97201, 58 p. Pope, J. G., and J. G. Shepherd. 1982. A simple method for the consistent interpreta- tion of catch-at-age data. J. Cons. int. Explor. Mer 40:176-184. Shepherd, J. G., and M. D. Nicholson. 1986. Use and abuse of multiplicative models in the analysis of fish catch-at-age data. The Statistician 35:221-227. 1991. Multiplicative modelling of catch-at-age data, and its application to catch forecasts. J. Cons. int. Explor. Mer 47:284-294. Turnock, J., and R. Methot. 1991. Status of west coast Dover sole in 1991. In Pa- cific Fishery Management Council, 1991, Status of the Pacific coast groundfish fishery through 1991 and recommended acceptable biological catches for 1992: stock assessment and fishery evaluation. Appendix B. Pacific Fishery Management Council, Metro Cen- ter, Portland, OR 97201. Abstract. -The genetic basis of the population structure of yellow- fin tuna, Thunnus albacares, in the Pacific Ocean was investigated with restriction fragment length polymor- phism (RFLP) analysis of mitochon- drial DNA (mtDNA). Samples of 20 yellowfin tuna were examined from each of five Pacific locations and one Atlantic location. MtDNA analysis with 12 informative restriction en- donucleases demonstrated consider- able genetic variation, as evidenced by an overall nucleon diversity of 0.84 and a mean nucleotide sequence diversity of 0.91%. Estimates of within-sample variation were re- markably consistent across all six lo- cations. Despite high levels of varia- tion, there was no evidence of genetic differentiation among samples. Com- mon genotypes occurred with simi- lar frequencies in all samples, and, with one exception, all genotypes that were represented by more than one individual occurred at more than one location. We could not reject the null hypothesis that all yellowfin tuna share a common gene pool. Our results are consistent with the al- ternate hypothesis that there is suf- ficient gene flow within the Pacific, as well as between the Atlantic and Pacific oceans, to prevent the ac- cumulation of significant genetic differentiation. Genetic analysis of the population structure of yellowfin tuna, Thunnus albacares, from the Pacific Ocean* Daniel R. Scoles Virginia Institute of Marine Science School of Marine Science, College of William and Mary Gloucester Point. VA 23062 John E. Graves** Virginia Institute of Marine Science School of Marine Science, College of William and Mary Gloucester Point. VA 23062 Inter-American Tropical Tuna Commission 8604 La Jolla Shores Drive, La Jolla, CA 92037 Yellowfin tuna (Thunnus albacares) occur in the tropical and subtropical oceans and support major commer- cial fisheries throughout their range (Collette and Nauen, 1983). The eco- nomic importance of this species is indicated by high annual catches that have increased from 596,764 metric tons (t) in 1981 to 986,529 1 in 1990, of which 66 to 69% were from the Pacific Ocean (FAO, 1992). Recently, purse-seine and longline fisheries in the western Pacific (120°E to about 180°) provided a major share of yellowfin tuna landings, with a catch of 342,921 1 in 1990 (Lawson, 1991). In the eastern Pacific (east of 130CW) record landings near 270,000 1 oc- curred in each of the past 3 years1. A thorough understanding of yel- lowfin tuna population structure is necessary for the effective man- agement of this economically impor- tant, marine resource. A variety of studies, including tagging, morpho- metric, fishery statistic and genetic analyses, have been used to infer population structure. However, the pro- posed population structures differed. Tagging studies have indicated that movements of yellowfin tuna in the Pacific Ocean tend to be geo- graphically restricted. Fink and Bay- liff (1970) analyzed tag return data in the eastern Pacific and proposed a northern and southern group of fish with some exchange between groups. There was very limited westward movement of tagged fish reported in the study; however, as the eastern Pacific fishery expanded westward in subsequent years, several returns were obtained from farther offshore indicating possible mixing between eastern and central Pacific fish (Bay- liff, 1984). In the western Pacific tag- ging studies showed that most indi- viduals remained within the western Pacific region and did not make ex- tensive movements (Itano and Will- iams, 1992; Lewis, 1992). Although the majority of recaptured yellowfin tuna in all studies showed limited movement, some returns were ob- tained which demonstrated the po- tential for fish to move large dis- Manuscript accepted 12 August 1993. Fishery Bulletin 91:690-698 (1993). 690 'Anonymous. 1992. Inter-American Tropical Tuna commission. Annual Report, 1991. "Contribution No. 1811 of the Virginia Insti- tute of Marine Science, **To whom reprint requests and correspon- dence should be addressed. Scoles and Graves: Genetic analysis of the population structure of Thunnus albacares 691 tances and between regions (Fink and Bayliff, 1970; Bayliff, 1984; Itano and Williams, 1992). Population structure was indicated by investigations of both meristic and morphometric characters which revealed significant differentiation among yellowfin tuna from the eastern, central, and western Pacific regions (Schaefer, 1955; Kurogane and Hiyama, 1957), as well as clinal character variation across the equato- rial Pacific (Royce, 1964). Further investigation using discriminant function analysis of morphometric vari- ables suggested mixing occurs between morphologically differentiated northern and southern yellowfin tuna of the eastern Pacific (Schaefer, 1989), as well as across the Pacific (Schaefer, 1991, 1992). Analysis of fishery data also suggests population structuring of yellowfin tuna within the Pacific. Kami- mura and Honma ( 1963) provided evidence for two or more semi-independent subpopulations based on size composition and catch data of equatorial Pacific yel- lowfin tuna from longline landings. Suzuki et al. (1978) examined longline and purse-seine length composition data and suggested the existence of semi-independent eastern, central, and western Pacific subpopulations. Additionally, homogeneity within the western Pacific was indicated by the widespread distribution of fish contaminated by radioactivity resulting from the 1954 U.S. nuclear tests at Bikini Atoll (Suzuki et al., 1978). While the results of several analyses suggest yel- lowfin tuna exhibit population structure within the Pacific Ocean, genetic analyses have revealed no sig- nificant genetic differentiation. Suzuki ( 1962) reported that the blood agglutinogen Tg2 occurs in similar fre- quencies in samples from the Indian Ocean and the eastern Pacific. Additionally, allozyme analysis did not reveal genetic differentiation between samples of yel- lowfin tuna collected off Hawaii (n=529) and Baja Cali- fornia (a; =207) at the polymorphic serum esterase lo- cus, although overall variation was low (Fujino, 1970). Preliminary evidence for frequency differences occur- ring at two other loci (phosphoglucose isomerase and transferrin A) was reported for both within and be- tween samples of Atlantic and Pacific yellowfin tuna2-3. However these loci have not been used to examine population structure. Our understanding of the population structure of yellowfin tuna in the Pacific Ocean remains problem- atic. Much evidence is available which suggests that population structure exists, yet genetic analyses did not reveal differentiation among samples collected from -Anonymous. 1977. Inter-American Tropical Tuna Commission, An- nual Report, 1976. 'Anonymous. 1978. Inter-American Tropical Tuna Commission, An- nual Report, 1977. distant locations. To further examine the genetic basis of the population structure of Pacific yellowfin tuna we employed restriction fragment length polymorphism (RFLP) analysis of mitochondrial DNA (mtDNA). Be- cause mtDNA evolves rapidly (Moritz et al., 1987; Brown et al., 1979) and displays considerable polymor- phism within animal populations (Avise and Lansman, 1983), mtDNA analyses have been useful in revealing population structure within marine fishes (Ovenden, 1990). Using this technique, we demonstrated consid- erable genetic variability within yellowfin tuna, but we could not reject the null hypothesis that samples share a common gene pool. Materials and methods Hearts were taken from 50 yellowfin tuna at each of five Pacific locations and one Atlantic location (Fig. 1); however, only 20 specimens per location were analyzed. Samples from the Pacific were collected during 1990 at Manta, Ecuador (ECU); Revillagigedo Islands, Mexico (MEX); Oahu, Hawaii (HAW); Manus Island, Papua New Guinea (PNG); and New South Wales, Aus- tralia (AUS). The sample from the Atlantic was col- lected during 1991 at Hatteras, North Carolina (ATL). Hearts were dissected within 12 hours of capture and placed on crushed ice. Hearts from fish collected in the Pacific were frozen at -20°C and shipped to the Inter- American Tropical Tuna Commission, La Jolla, CA, where they were stored at -20°C for more than one year before shipment to our laboratory. Hearts from fish collected in the Atlantic were transported on wet ice and frozen at -70°C within four hours of dissection. MtDNA was purified from 3 g of heart tissue from Atlantic specimens following the CsCl-ethidium bro- mide gradient centrifugation protocol of Lansman et al. (1981). MtDNA yields averaged about 350 ng of supercoiled mtDNA per g of heart tissue. Aliquots of mtDNA were digested with the following 12 informa- tive restriction endonucleases (Stratagene and BRL) according to the manufacturers' instructions: Apal, Aval, Banl, Bell, Bgll, Dral, EcoRl, Hindlll, Neil, PstI, Pvull and Xhol. Restriction fragments were end-la- beled with a mixture of all four a-35S-dNTP's by using the Klenow fragment, electrophoresed at 2 volts/cm overnight in YJc agarose gels, and visualized by auto- radiography (Sambrook et al, 1989). Yields of supercoiled mtDNA from Pacific specimens were low, possibly due to sub-optimal storage condi- tions. For these specimens, mtDNA-enriched genomic DNA was isolated from 4 to 6 g of heart tissue follow- ing the protocols of Chapman and Powers ( 1984), modi- fied by the omission of sucrose step gradients and the 692 Fishery Bulletin 91(4). 1993 Figure 1 Thunnus albacares. Location of sampling sites used in this analysis: (11 Manta, Ecuador (ECU); (2) the Revillagigedo Islands, Mexico iMEXl; (3) Oahu, Hawaii (HAW); 1 4 1 New South Wales, Australia (AUS); (5) Manus Island, Papua New Guinea (PNG); and (6) Hatteras, North Carolina (ATL). use of 1.5% sodium dodecyl sulfate for mitochondria] lysis. Following restriction enzyme digestion and hori- zontal agarose gel electrophoresis, DNA fragments were transferred to nylon membranes by Southern transfer (Sambrook et al., 1989) and immobilized by long-wave UV irradiation. Prehybridization was conducted for two hours at 42°C in 50% formamide, 5x SSC, 5x Dendhardt's solution, 0.025 mM NaP04, pH 6.5, and 100 ug/mL heat denatured calf thymus DNA. Probe DNA (mtDNA purified from extra specimens from the Atlantic) was nick-translated to incorporate biotin-7- dATP (BRL), and separated from unincorporated nucle- otides by size exclusion chromatography. One (ig of probe was added to the prehybridization solution for each 200 cm2 blot and allowed to hybridize overnight at 42°C. Following post-hybridization washes (Sambrook et al., 1989), mtDNA fragments were visu- alized with the BRL BluGene Non-Radioactive Nucleic Acid Detection Kit. A 12-letter composite mtDNA genotype, indicating the fragment pattern for each restriction enzyme, was developed for each individual. Estimates of nucleon diversity (h ) for each sample and for the pooled samples were computed following Nei (1987). Nucleotide se- quence divergences among genotypes were estimated by using the site method of Nei and Li (1979). Esti- mates of nucleotide sequence diversity (p) within each sample, and mean nucleotide sequence divergences among samples (corrected for within-sample diversi- ties) were computed following Nei (1987). Nucleotide sequence divergences were clustered by the unweighted pair-group method with arithmetic means (UPGMA) by using the average linkage algorithm of the SPSS-X statistical package (Norusis, 1988). Values of G„, a measure of heterogeneity between samples, were esti- mated from sample genotype frequencies (Nei, 1987), and values of Njn. the absolute number of migrants between samples, were determined from the relation Scoles and Graves Genetic analysis of the population structure of Thunnus albacares 693 Njn = (VGsl - l)/2 (Birky et al., 1983; Nei, 1987). Chi- square analysis was conducted by using the Monte- Carlo method of Roff and Bentzen (1989) with 1000 randomizations of the data to evaluate heterogeneity of genotype frequencies among samples without com- bining rare genotypes. Results Analysis of mtDNA from 120 yellowfin tuna with 12 restriction enzymes revealed a total of 34 genotypes, comprising 83 unique fragments. The most common genotype consisted of 52 fragments, representing a sur- vey of 304 bp, or about 1.8% of the mtDNA genome. The mean size of the yellowfin tuna mtDNA genome. determined from the most common restriction frag- ment profiles for each of the 12 restriction enzymes, was 16,549 ± 309 bp(SD). RFLP analysis of yellowfin tuna mtDNA demon- strated considerable variation. While four restriction enzymes, Aval, Dral, Hindlll, and Xhol, showed no variation, the remaining eight revealed two to seven restriction morphs each. Of the 34 genotypes, two were represented by 20 or more individuals, five were rep- resented by four or more individuals, and 20 geno- types occurred only once (Table 1). Within-sample nucleon diversities were high and showed little varia- tion among samples, ranging from 0.82 to 0.86, with a value of 0.84 for the pooled samples (Table 2). Mean nucleotide sequence diversities were also high, ranging Table 1 Distributions of mtDNA genotypes among samples of yellowfin tuna. Thunnus a Ibacares . Letters represent fragment patterns produced by th ? following enzymes left to right): Eco RI, Hindlll, Pstl, Dral Aval, Pvull, Neil, Ban , Bell. Bgll. Xh >I, and Apa\ . A descr ption af fragment sizes is available upon request. mtDNA Sampl mg location genotype ECU MEX HAW AUS PNG ATL Total 1 AAAAAAAAAAAA 7 6 7 8 8 7 43 2 BAAAAAAAAAAA 2 5 3 4 3 3 20 3 BAAAAABAAAAA 3 — 1 — 1 1 6 4 AAAAAAAAECAA 1 1 — 2 — 2 6 5 AAAAAAAABAAA — — — 1 1 2 4 6 AAAAAABAEAAA 1 — 1 — — 1 3 7 AAAAAABABAAA 2 1 — — — — 3 8 AABAAAAAAAAA — 2 — 1 — — 3 9 AAAAAAABAAAA — — 1 — 1 — 2 10 AAAAAAAABBAA — — 1 1 — — 2 11 BAAAAAAAAAAB — — 1 — — 1 2 12 BAAAAAAACAAA — — 1 — 1 — 2 13 AAAAAAAAEAAF — 1 — — 1 — 2 14 AAAAAABAAAAA — — 2 — — — 2 15 AAAAAAAAAAAB — — — 1 — — 16 AAAAABAAAAAA — — — — — 1 17 AAAAAAAAAAAC — — — — 1 — 18 AAAAAAAAAAAD — — — — 1 — 19 AAAAAAAAFAAA — — — — 1 - — 20 AAAAAAAAAAAF — — 1 — — 21 AABAAAAABAAA — — — — — 1 22 BABAAAAAAAAA — — 1 — — — 23 AAAAAABAAAAE — — — 1 — — 24 AAAAAAAADAAB — — — — 1 — 25 CAAAAABAAAAA 1 — — — — — 26 AABAAAAAAAAG 1 — — — — — 27 AABAAAAAEAAA 1 — — — — — 28 AAAAAABAAAAB — 1 — — — — 29 BAAAAAAAAAAD — 1 — — — — 30 BAAAAAAABBAA — 1 — — — — 31 BAAAAABABAAA — 1 — — — — 32 BAAAAABAGAAA 1 — — — — — 33 AAAAAABAADAA — — — — — 1 34 AAAAAAACFEAA — — — 1 — — Total 20 20 20 20 20 20 120 694 Fishery Bulletin 91(4), 1993 Table 2 Genetic variation within samples of yellowfin tuna, Thunnus albacares, expressed as nucleon diversity (h) and percent nucleotide sequence diversity (p). Sampling Nucleon Nucleotide sequence location diversity (h) diversity Ip) % ECU 0.863 1.04 MEX 0.863 0.955 HAW 0.868 0.799 AUS 0.816 0.905 PNG 0.837 0.737 ATL 0.863 0.917 Pooled 0.840 0.907 from 0.74% to 1.04%, and 1.04% for the pooled samples (Table 2). Despite the high level of within-sample variation, there was little evidence for genetic differentiation among the six sampling sites. The two most common genotypes (1 and 2) were observed at all locations at similar frequencies (30-40%, 10-25%, respectively; Table 1). Of the 12 other genotypes that occurred in more than a single individual, 11 were found in two or more samples, and only one (genotype 14) occurred in a single sample, represented by two individuals. Pairwise estimates of corrected nucleotide sequence divergences between samples (including the Atlantic) were low, ranging from 0.012% to 0.104%, with a mean pairwise divergence of 0.04%. Similarly, estimates of G„ were low, ranging from 0.011 to 0.025, and values of Nem were correspondingly high, ranging from 19.5 to 44.5 females per generation. No clear pattern of phylogeographic structure was revealed in cluster analyses of nucleotide sequence divergences among mtDNA genotypes or sampling lo- cations. Chi-square analysis of het- erogeneity among samples was not significant, as randomizations of the data into six samples were more heterogeneous than the observed genotypic distributions 581 of 1000 times (P=0.581). To determine an appropriate number of individuals to examine from each location and the number of restriction enzymes to employ, we compared levels of genetic variation and differentiation revealed by a pi- lot study of 12 individuals per loca- tion with 5 enzymes with larger analyses of up to 20 individuals per location with 12 enzymes (Table 3). Levels of variation were influenced more by increasing the number of enzymes surveyed than the number of individuals. Increasing the num- ber of enzymes increased the number of genotypes, the nucleon diversities, and, to a lesser extent, nucle- otide sequence diversities. Analysis of a greater num- ber of individuals per location had little effect on di- versity estimates with either 5 or 12 enzymes, although the ranges of within-sample diversities among loca- tions decreased as more individuals were analyzed. Increasing the number of individuals or the number of enzymes had little effect on levels of genetic differ- entiation. No significant differences were found in the distributions of genotypes among locations in Roff and Bentzen ( 1989) chi-square tests (Table 3). Furthermore, increasing the number of individuals in the 12 enzyme analysis did not increase the frequencies of genotypes unique to a location. Instead, many unique genotypes in the analysis of 12 individuals occurred in other lo- cations as a greater number of individuals were used. Because increasing sample sizes from 12 to 20 indi- viduals did not reveal greater spatial partitioning of genetic variation, we decided that 20 individuals was an appropriate sample size. Discussion Population structure is typically manifested as the spa- tial or temporal partitioning of genetic variation. There- fore, to demonstrate if population structure exists, a technique must first reveal a reasonable level of ge- netic variation. RFLP analysis of mtDNA revealed con- siderable genetic variation in yellowfin tuna. The over- all nucleon diversity and mean nucleotide sequence diversity (0.84 and 0.91%, respectively) are in the up- Table 3 Pooled percent nucleotide sequence diversities (p and nucleon diversities l/il, and respective ranges in parentheses among samples obtained by varying the numbers of restriction enzymes and indiv duals of yellow in tuna, Thunnus albacares, per sampling location Probabilities of significance result from Roff and Bentzen's (1989) chi-square analysi s with 1000 ran domizations of the data. The five enzymes selected for the pilot study were EcoRI, Hit dill, PstI, Dral, and Aval. Number of Nucleon Nucleotide sequence Design genotypes diversity (h ) diversity (p) (%) Prob. 5 enzymes 6 0.496 0.705 0.960 12 individuals (0.41-0.64) (0.49-0.90) 5 enzymes 6 0.509 0.714 0.845 20 individuals (0.42-0.63) (0.62-0.9H 12 enzymes 27 0.829 0.907 0.975 12 individuals (0.76-0.94) (0.66-1.22) 12 enzvmes 34 0.840 0.907 0.581 20 individuals (0.82-0.87) (0.74-1.04) Scoles and Graves: Genetic analysis of the population structure of Thunnus albacares 695 per ranges reported for marine fishes (Avise et al., 1989; Ovenden, 1990; Gold and Richardson, 1991) in- cluding other large, pelagic fishes (Table 4). Genetic variability in tuna species has also been demonstrated by sequence analysis of a 307-base-pair region of the mitochondrial cytochrome 6 gene of blue- fin tuna, Thunnus thynnus; bigeye tuna, T. obesus; albacore, T. alalunga; and yellowfin tuna (Bartlett and Davidson, 1991). The sequence data presented for 33 yellowfin tuna result in a nucleoli diversity of h = 0.28, a value that is considerably lower than the nucleon diversity of 0.84 obtained in our study. Although nucleon diversity values are greatly influenced by the number of base pairs surveyed (Nei, 1987), the num- ber examined in this study and in Bartlett and Davidson (1991) were very close (304 and 307, respec- tively). This difference in nucleotide sequence diver- sity may reflect a slower evolutionary rate for the mi- tochondrial cytochrome b gene relative to the entire mtDNA genome. Significant genetic differentiation among yellowfin tuna from geographically distant locations was not found. Corrected nucleotide sequence divergences av- eraged only 0.049r, indicating that the mean differ- ence between two genotypes randomly chosen from any two samples was essentially the same as the differ- ence between two genotypes randomly drawn from the same sample. The frequencies of the two most com- mon genotypes were similar among all locations, and an overall chi-square test for heterogeneity was non- Table 4 Genetic variation within selected pelagic species determined by RFLP analysis of mtDNA employing 11-13 informative enzymes. Variation is expressed as nucleon diversity Oi) and percent nucleotide sequence diversity (p). Nucleon Nucleotide sequence Species diversity (h) diversity (p) (%) Thunnus albacares' - (Yellowfin Tuna) 0.84 0.91 Thunnus alalunga2 (Albacore Tuna) 0.60 0.31 Tetraplurus audax' (Striped Marlin) 0.74 0.54 Tetrapturus albidus3 (White Marlin I 0.70 0.35 Makaira nigricans' (Blue Marlin) 0.86 1.99 Istiophorus plalypterus* (Sailfish) 0.62 0.87 'This study. -Graves and Dizon (1989) and Graves, unpublished data. 'Graves and McDowell, in press. significant. Furthermore, nucleon diversities and nucle- otide sequence diversities were similar among loca- tions, although uniformity in the latter estimates was less pronounced. The apparent genetic homogeneity of yellowfin tuna is consistent with the hypothesis that there is genetic exchange among locations. Rare mtDNA genotypes had low frequencies of occurrence (Table 1), characteristic of high gene flow (Slatkin, 1985), a situation similar to that found in other species for which high gene flow is suggested: American eel, Anguilla rostrata (Avise et al., 1986); marine catfishes, Ariidae (Avise et al., 1987); weakfish, Cynoscion regalis (Graves et al., 1992a); and bluefish, Pomatomus saltatrix (Graves et al., 1992b). Low pairwise estimates of Gs, (0.011-0.025) also indi- cated homogeneity in yellowfin tuna and gave rise to high N,m values (>19) which are consistent with high rates of gene flow among regions (Birky et al., 1983). The absence of genetic differentiation between Pa- cific and Atlantic samples of yellowfin tuna is similar to that reported for other vagile pelagic species. RFLP analysis of mtDNA of skipjack tuna, Katsuwonus pelamis, and of albacore from the Atlantic and Pacific oceans revealed modest amounts of within-sample variation, but no significant differentiation was found between Atlantic and Pacific conspecifics (Graves et al, 1984; Graves and Dizon, 1989). Similarly, RFLP analysis of mtDNA of the pelagic dolphin, Coryphaena hippurus, revealed no significant differentiation be- tween samples from the Atlantic and Pacific oceans4. However, spatial partitioning of genetic variation oc- curs in at least some pelagic fishes. Significant genetic differentiation was shown between Atlantic and Pa- cific blue marlin, Makaira nigricans, by direct sequence analysis of the mitochondrial cytochrome b gene (Finnerty and Block, 1992) and RFLP analysis of mtDNA (Graves and McDowell, in press). Similarly, differentiation was found among striped marlin samples, Tetrapturus audax, and sailfish, Istiophorus plalypterus, from the Pacific Ocean by RFLP analysis of mtDNA (Graves and McDowell, in press). There are several characteristics of yellowfin tuna which could promote gene flow among locations. Yel- lowfin tuna are distributed circumtropically (Collette and Nauen, 1983) and occur around the Cape of Good Hope in the southern summer (Talbot and Penrith, 1962). Tagging studies demonstrated that adults are capable of traveling large distances between Pacific regions (Fink and Bayliff, 1970; Bayliff, 1984; Itano and Williams, 1992), and are capable of undergoing trans-Atlantic crossings (Bard and Scott, 1991). The existence of suitable spawning areas throughout the JCarol Reeb, University of Hawaii, pers. commun. April 1992. 696 Fishery Bulletin 91(4), 1993 tropical oceans, suggested by circumtropical occurrence oflarvae (Nishikawa et al., 1985), would permit unob- structed gene flow throughout the species' distribu- tion. Both genetic and morphological analyses revealed considerable variation in yellowfin tuna; however, ge- netic analyses indicated that differentiation does not occur. Evidence that morphological characters are en- vironmentally influenced was provided by comparison of our data to the morphological differentiation found in Schaefer (1991, 1992). In our analysis, 20 fish from four locations (Ecuador, Mexico, Hawaii, and Austra- lia) were the same as those used in Schaefer s studies. Although no genetic differences were found among these locations, morphometric characters and gill-raker counts differed significantly, which we conclude is the result of phenotypic plasticity. The finding of greater morphological variability among Pacific samples than occurred between the Atlantic and Pacific (Schaefer and Walford, 1950) also supports the hypothesis that phenotypic plasticity is the cause of morphological dif- ferentiation in the Pacific. The null hypothesis that yellowfin tuna in the Pa- cific Ocean share a common gene pool could not be rejected in this analysis. Our results were consistent with the alternate hypothesis that yellowfin tuna main- tain sufficient gene flow among areas to prevent the accumulation of significant genetic differentiation. As theoretical models indicate that very low levels of mi- gration (only a few individuals per generation) are needed to prevent genetic differentiation among large populations (Allendorf and Phelps, 1981; Hartl and Clark, 1989), the present data indicate only that some minimal amount of exchange is occurring. Other ap- proaches are necessary to quantify the amounts of mix- ing among regions and to judge more accurately whether or not separate stocks should be distinguished for management purposes. Acknowledgments Kurt Schaefer either directly collected or arranged col- lection of all specimens from the Pacific. Antony Lewis and Ziro Suzuki kindly made regional literature avail- able for our use. William Bayliff, Mark Chittenden, Bruce Collette, Edward J. Heist, and Kurt Schaefer critically reviewed the manuscript. Literature cited Allendorf, F. W., and S. Phelps. 1981. Use of allelic frequencies to describe population structure. Can. J. Fish. Aquat. Sci. 38:1507-1014. Avise, J. C, and R. A. Lansman. 1983. Polymorphism of mitochondrial DNA in popula- tions of higher animals. In M. Nei and R.K. Koehn (eds.), Evolution of genes and proteins, p. 147- 164. Sinauer, Sunderland, MA. Avise, J. C, G. S. Helfman, N. C. Saunders, and L. S. Hales. 1986. Mitochondrial DNA differentiation in North At- lantic eels: population genetic consequences of an un- usual life history pattern. Proc. Natl. Acad. Sci. USA 83:4350-4354. Avise, J. C, C. A. Reeb, and N. C. Saunders. 1987. Geographic population structure and species dif- ferences in mitochondrial DNA of mouthbrooding ma- rine catfishes (Ariidae) and demersal spawning toadfishes (Batrachoididae). Evolution 41:991-1002. Avise, J. C, B. W. Bowen, and T. Lamb. 1989. DNA fingerprints from hypervariable mitochon- drial genotypes. Mol. Biol. Evol. 6:258-269. Bard, F. X., and E. L. Scott. 1991. Sept traversers transatlantiques d'albacores marques. Thons migrateurs ou sedentaires? (SCRS/ 89/23. YYP/89/1). In Report of the ICCAT Yellowfin Year Program. ICCAT Coll. Vol. Sci. Papers XXXVI:205-222. Bartlett, S. E., and W. S. Davidson. 1991. Identification of Thunnus tuna species by the polymerase chain reaction and direct sequence analy- sis of their mitochondrial cytochrome b genes. Can. J. Fish. Aquat. Sci. 48:309-317. Bayliff, W. H. 1984. Migrations of yellowfin and skipjack tuna re- leased in the central portion of the eastern Pacific Ocean, as determined by tagging experiments. IATTC Internal Rep. 18, 107 p. Birky, C. W. Jr., T. Maruyama, and P. Fuerst. 1983. An approach to population and evolutionary ge- netic theory for genes in mitochondria and chloro- plasts, and some results. Genetics 103:513-527. Brown, W. M., M. George Jr., and A. C. Wilson. 1979. Rapid evolution of animal mitochondrial DNA. Proc. Natl. Acad. Sci. USA 76:1967-1971. Chapman, R. W., and D. A. Powers. 1984. A method for the rapid isolation of mitochon- drial DNA from fishes. Maryland Sea Grant College Program Tech. Rep. UM-SG-TS-84-01, 11 p. Collette, B. B., and C. E. Nauen. 1983. FAO species catalogue. Vol. 2: Scombrids of the world: an annotated and illustrated catalogue of tu- nas, mackerels, bonitos and related species known to date. FAO Fish. Synop. 125, 137 p. FAO (Food and Agriculture Organization of the United Nations). 1992. Yearbook of fisheries statistics. Catches and land- ings, 1990. Yearb. Fish. Stat., Vol. 70. Fink, B. D., and W. H. Bayliff. 1970. Migrations of yellowfin and skipjack tuna in the eastern Pacific Ocean as determined by tagging ex- periments, 1952-1964. Inter-Am. Trop. Tuna Comm. Bull. 15:1-227. Scoles and Graves Genetic analysis of the population structure of Thunnus albacares 697 Finnerty, J. R., and B. A. Block. 1992. Direct sequencing of mitochondria] DNA detects highly divergent haplotypes in blue marlin (Makaira nigricans). Mol. Mar. Biol, and Biotechnolog. 1:206- 214. Fujino, K. 1970. Immunological and biochemical genetics of tunas. Trans. Am. Fish. Soc. 99:152-178. Gold, J. R., and L. R. Richardson. 1991. Genetic studies in marine fishes. IV. An analysis of population structure in the red drum (Sciaenops ocellatus) using mitochondrial DNA. Fish. Res. 12:213-241. Graves, J. E., and A. E. Dizon. 1989. Mitochondrial DNA sequence similarity of Atlantic and Pacific albacore tuna (Thunnus ala- lunga ). Can. J. Fish. Aquat. Sci. 46:870-873. Graves, J. E., and J. R. McDowell. 1993. Genetic analysis of population structure. Inter- national Commission for the Conservation of Atlantic Tunas (ICCAT) special billfish workshop publication. SCRS/92/50. ICCAT, Madrid, in press. Graves, J. E., S. D. Ferris, and A. E. Dizon. 1984. Close genetic similarity of Atlantic and Pacific skipjack tuna (Katsuwonus pelamis) demonstrated with restriction endonuclease analysis of mitochon- drial DNA. Mar. Biol. 79:315-319. Graves, J. E., J. R. McDowell, and M. L. Jones. 1992a. A genetic analysis of weakfish Cynoscion regalis stock structure along the mid-Atlantic coast. Fish. Bull. 90:469-475. Graves, J. E., J. R. McDowell, A. M. Beardsley, and D. R. Scoles 1992b. Stock structure of the bluefish Pomatomus saltatrix along the mid-Atlantic coast. Fish. Bull. 90:469-475. Hartl, D. L., and A. G. Clark. 1989. Principles of population genetics. 2nd. rev. ed. Sinauer. Sunderland, MA. Itano, D. G., and P. G. Williams. 1992. Analysis of yellowfin tagging data and related information collected by the Skipjack Survey and As- sessment Programme. SPC Tuna and Billfish Assess- ment Programme (TBAP) Tech. Rep. 28, 39 p. Kamimura, T., and M. Honma. 1963. Distribution of the yellowfin tuna Neothunnus macropterus (Temminck and Schlegel) in the tuna longline fishing grounds of the Pacific Ocean. FAO, Fish. Rep. 6:1299-1328. Kurogane, K., and Y. Hiyama. 1957. Morphometric comparison of the yellowfin taken from the equatorial Pacific. Jap. Soc. Sci. Fish. Bull. 23:388-393. Lansman, R. A., R. O. Shade, D. F. Shapira, and J. C. Avise. 1981. The use of restriction endonucleases to measure mitochondrial DNA sequence relatedness in natural populations. III. Techniques and potential appli- cations. J. Mol. Evol. 17:214-226. Lawson T. 1991. Status of tuna fisheries in the SPC area during 1990, with annual catches since 1952. SPC Tuna and Billfish Assessment Programme (TBAP) Tech. Rep. 27, 73 p. Lewis, A. D. 1992. Stock structure of Pacific yellowfin — a review. W. Pacific Yellowfin Tuna Res. Group Work- shop; Honolulu, Hawaii, No. 7, 13 p. Moritz, C, T. E. Dowling, and W. M. Brown. 1987. Evolution of animal mitochondrial DNA: rel- evance for population biology and systematics. Ann. Rev. Ecol. Syst. 18:269-292. Nei, M. 1987. Molecular evolutionary genetics. Columbia Univ. Press, New York, 512 p. Nei, M., and W.-H. Li. 1979. Mathematical model for studying genetic varia- tion in terms of restriction endonucleases. Proc. Natl. Acad. Sci. 76:5269-5273. Nishikawa, Y., M. Honma, S. Ueyanagi, and S. Kikawa. 1985. Average distribution of larvae of oceanic species of scombroid fishes, 1956-1981. Far Seas Fish. Res. Lab.,SSer., 12, 99 p. Ovenden, J. R. 1990. Mitochondrial DNA and marine stock assess- ment: a review. Aust. J. Mar. Freshwater Res. 41:835-53. Norusis, M. L. 1988. SPSS-X advances statistics guide. 2nd ed. SPSS Inc., Chicago, IL. Roff, D. A., and P. Bentzen. 1989. The statistical analysis of mitochondrial DNA polymorphisms: chi-square and the problem of small samples. Mol. Biol. Evol. 6:539-545. Royce, W. F. 1964. A morphometric study of yellowfin tuna Thunnus albacares (Bonnaterre). U.S. Fish Wildl. Serv. Fish. Bull. 63:395-443. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning, a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Schaefer, K. M. 1989. Morphometric analysis of yellowfin tuna, Thunnus albacares, from the eastern Pacific Ocean. Inter-Am. Trop. Tuna Comm. Bull. 19:387-427. 1991. Geographic variation in morpFibmetric characters and gill-raker counts of yellowfin tuna Thunnus albacares from the Pacific Ocean. Fish. Bull. 89:289-297. 1992. An evaluation of geographic and annual varia- tion in morphometric characters and gill-raker counts of yellowfin tuna, Thunnus albacares, from the Pa- cific Ocean. Inter-Am. Trop. Tuna Comm. Bull. 20:133-163. Schaefer, M. B. 1955. Morphometric comparison of yellowfin tuna from southeast Polynesia, Central America, and Hawaii. Inter-Am. Trop. Tuna Comm. Bull. 1:89-136. hcJ8 Fishery Bulletin 91(4), 1993 Schaefer, M. B, and L. A. Walford. 1950. Biometric comparisons between yellowfin tunas (Neothunnus) of Angola and the Pacific coast of Cen- tral America. U.S. Fish and Wildl. Serv. Fish. Bull. 51:425-443. Slatkin, M. 1985. Rare alleles as indicators of gene flow. Evolution 39:53-65. Suzuki, A. 1962. On the blood types of yellowfin and bigeye tuna. Am. Nat. 96:239-246. Suzuki, Z., P. K. Tomlinson, and M. Honma. 1978. Population structure of Pacific yellowfin tuna. Inter-Am. Trop. Tuna Comm. Bull. 17:273-441. Talbot, F. H., and M. J. Penrith. 1962. Tunnies and marlins of South Africa. Nature 193:558-559. Abstract. -The feeding habits of the vermilion snapper, Rhomboplites aurorubens, were investigated to de- termine the role of this ecologically dominant and economically valuable species in the trophic ecology of southeastern U.S. reef habitats. Trophic ecology was studied by ex- amining stomach contents and by comparing them to samples of benthic invertebrates and near- bottom plankton. Vermilion snapper fed on a variety of infaunal, epifau- nal, and pelagic invertebrates, as well as on demersal and pelagic fishes and cephalopods. The diet was diverse but was dominated numeri- cally by planktonic species. The benthic prey eaten were species as- sociated with hard-bottom reef struc- ture or were infaunal species from sand bottom areas adjacent to the reef. Many species were members of the hyperbenthos or demersal zoo- plankton and were apparently con- sumed in the water column during their nocturnal emergence from the sand or reef. Small crustaceans, es- pecially copepods, sergestid deca- pods, and larvae of barnacles, sto- matopods and decapods, dominated the diet of small (<50mm SL) ver- milion snapper. Larger decapods, fishes, and cephalopods were more important in the diet of larger ver- milion snapper. Vermilion snapper, although reef associated, does not feed heavily on reef species, and may be important in transferring energy from the water column and adjacent sandy areas to the reef. Planktonic and benthic feeding by the reef-associated vermilion snapper, Rhomboplites aurorubens (Teleostei, Lutjanidae) * George R. Sedberry Marine Resources Research Institute PO Box 12559 Charleston SC 29422-2559 Nicole Cuellar Gnce Marine Biological Laboratory University of Charleston 205 Ft. Johnson Charleston SC 294 1 2 Manuscript accepted 15 June 1993. Fishery Bulletin 91:699-709 1 19931. The vermilion snapper, Rhomboplites aurorubens, is the most abundant lutjanid in the recreational and com- mercial fisheries of the southeastern U.S. (Grimes et al., 1982), and is a dominant component of the reef- associated ichthyofauna off the Caro- linas and Georgia (Sedberry and Van Dolah, 1984). Because of its abun- dance and habit of foraging on small pelagic crustaceans (Grimes, 1979), vermilion snapper may be an impor- tant trophic link between the water column and those reef habitats where it schools during resting periods. If vermilion snapper also forages on in- faunal benthos in adjacent sandy ar- eas, it may be very important in the trophic coupling between the reef and surrounding expansive sandy areas. Vermilion snapper is also fed on by other predatory fishes (Sedberry, 1988) and thus may provide trophic links among top-level carnivores and the infauna, holozooplankton, dem- ersal zooplankton isensu Alldredge and King, 1977), and reef benthos. Predation by fishes has been shown to be important in transferring en- ergy from the water column or adja- cent sand bottom areas to reef habi- tats (Bray et al, 1981; Meyer and Schultz, 1985; Rothans and Miller, 1991). Although information on feed- ing is available for some species of reef-associated fishes off the south- eastern United States (Manooch, 1977; Grimes, 1979; Sedberry, 1985, 1987, 1988), the importance of reef bottom versus adjacent water column and sand bottom habitats as feeding grounds is poorly understood. These reefs support not only a variety of large sessile invertebrates (e.g., sponges, corals, tunicates) and asso- ciated motile organisms ( Struhsaker, 1969; Wenner et al., 1983) but also a greater faunal abundance, diversity, and biomass than adjacent sand bot- tom areas of the open shelf (Struhsaker, 1969; Wenner. 1983; Wenner et al., 1983). Because of the rocky outcrops and the warming in- fluence of the Florida Current, these reefs support tropical and subtropi- cal families of fishes, and many eco- nomically valuable serranids, haemu- lids, sparids, and lutjanids, including vermilion snapper (Miller and Richards, 1980; Chester et al, 1984). The greater biomass and diversity of rocky reef habitats, compared with Contribution No. 327 of the South Carolina Marine Resources Center and Contribution No. 114 of the Grice Marine Biological Laboratory. 699 700 Fishery Bulletin 91(4), 1993 sandy areas, may be the result of trophic links through reef-asssociated fishes, such as vermilion snapper, with other ecotopes on the shelf. While an important component of regional reef eco- systems and fisheries, vermilion snapper and demer- sal-feeding fishes, such as red porgy (Pagrus pagrus) are being overfished (Low et al., 1985; Collins and Sedberry, 1991). As a result of this increased fishing pressure, vermilion snapper has increased in abun- dance relative to P. pagrus and other overexploited reef fishes off the southeastern United States, and the importance of its functional role in reef ecosystems may have changed. Increased fishing pressure since the late 1970 s has also apparently caused a concomi- tant decrease in mean length and a decrease in size at maturity of vermilion snapper (Collins and Pinckney, 1988; Collins and Sedberry, 1991). If small vermilion snapper consume different prey than do large vermil- ion snapper and if average fish size is decreasing in the region, the trophic structure of reefs may be affected. The objectives of this study were to describe the feeding habits of vermilion snapper and to evaluate its relative dependence on hard-bottom benthos, sand- bottom infauna, demersal zooplankton (Alldredge and King, 1977, 1985; Porter and Porter, 1977), holozooplankton (Cahoon and Tronzo, 1992) and nek- ton as food. An additional purpose was to describe differences in feeding habits with size. Methods Stomachs from vermilion snapper were collected dur- ing six cruises in 1980 and 1981 from 11 reef stations off South Carolina and Georgia. Stations were located in each of three depth zones representing the inner shelf ( 16-22 m depth, three stations), middle shelf (23-37 m depth, four stations) and outer shelf (46- 69 m depth, four stations). Delineation of depth zones was based on distribution of fish assemblages as noted in previous studies and on community analysis of catches in the present study (Struhsaker, 1969; Miller and Richards, 1980; Sedberry and Van Dolah, 1984). Fishes were collected primarily with a roller-rigged 40/54 high rise trawl (Hillier, 1974); a few were cap- tured with traps or hook-and-line. Sampling was con- ducted during the day and at night on hard-bottom reef habitat that was mapped for each station by means of underwater television. Detailed descriptions of sta- tion locations and fish sampling techniques have been described elsewhere (Sedberry and Van Dolah, 1984; Wenneretal., 1984). Standard lengths (SL in mm) were measured at sea. Stomachs were removed, individually labeled, and fixed in 10% seawater-formalin. Because of limited space and very large catches, only those stomachs that ap- peared to contain ingesta were preserved. No attempt was made in the field to determine the percentage of stomachs with food. Small (<50 mm SL) vermilion snap- per were preserved whole and dissected in the labora- tory. Fixed stomachs and small individuals were washed in tap water and transferred to 50% isopropyl alcohol. Contents of individual stomachs were sorted by taxa and counted. Volume displacement of food items was measured in a graduated cylinder or estimated by us- ing a lxl mm grid (Windell, 1971). The relative con- tribution of food items to the diet was determined by using three methods: percent frequency occurrence, F= number of stomachs with prey taxon number of stomachs with food percent numerical abundance, N-- percent volume displacement, V- number of individuals of prey taxon total number of prey items volume displacement of prey taxon total volume of all prey items 100 X 100 100 These values are presented only for those prey species that occurred with a frequency of at least one percent or that made up at least one percent of the total num- ber or volume of prey. Values of F, N, and V were also calculated for higher prey taxa, for stomachs pooled by 50-mm intervals of SL of vermilion snapper. The chi- square statistic (Tyler, 1979) was used to test for sig- nificance (0.05 level) of feeding heterogeneity between predator size classes. To determine selection of prey type and predator feeding habitat, stomach samples were compared to samples from the potential prey environment by using Ivlev's index of electivity (Ivlev, 1961), calculated as E -P, r, + P, where: E = electivity for the ;th potential prey species; r, = percent by number of species i in the diet; and p, = percent by number of species i in samples from the environment. Electivity values range from -1 to +1. Negative values imply that the species is avoided, not preferred, only incidentally ingested, or unavailable to the predator. Positive values imply that the predator prefers the prey species or that it is feeding on prey species that occur in a different habitat than that sampled by the prey sampler. A value near zero im- Sedberry and Cuellar Planktonic and benthic feeding by Rhomboplites aurorubens 701 plies no selectivity by the predator, that is to say the fish is feeding on the prey in proportion to the prey's relative abundance in samples taken in the habitat. The electivity index was calculated for species that were numerically dominant in prey environment samples or in fish stomach samples. Dominant species included those that ranked among the five most abun- dant species within stomach or environmental samples pooled by depth zone. Prey samples and stomach col- lections were pooled by depth zone (inner, middle, and outer shelf) for comparison. Samples of benthic prey were obtained at the 11 reef sites during 1980 and 1981 with a diver-operated suction sampler (inner and middle shelf) or a grab sampler (outer shelf). Details of benthic sampling are provided elsewhere (Wenner et al., 1983; Wenner et al., 1984) and are only summa- rized here. Briefly, divers obtained five replicate suc- tion samples during the day at each inner and middle shelf reef site by scraping the hard reef substrate en- closed by a 0.1-m2 quadrat box while simultaneously sucking with an airlift device similar to that described by Chess (1979). Suction samples were collected in 1.0-mm mesh bags. At the outer shelf stations, where water depth pre- cluded the use of the suction device operated by divers, quantitative 0.1 m'--samples were collected in reef habi- tat with a modified Smith-Mclntyre grab. After re- trieval, each sample was sieved through a 1.0-mm sieve and retained organisms were identified and counted. Most (75.0%) grab samples were taken during the day. Additional sampling was conducted to compare stom- ach contents with potential prey species occurring in the near-bottom water column above the reef. To de- termine extent of feeding on the near-bottom plank- ton, electivity values were calculated for samples ob- tained with an epibenthic plankton sled and compared with those from the suction and grab samples. The sled was similar to that described by Brattegard and Fossa (1991) but had a mouth opening of 0.5 m2 and runners that permitted it to sample 0.5 m off the bot- tom. The sled also had a mouth-opening mechanism designed to fish only when it was in contact with the bottom. A 0.947-mm mesh net was attached to the sled and two ten-minute tows were made per station at night to minimize net avoidance. Sled collections for invertebrates were made only in 1981. These samples were analyzed for decapods, stomatopods, cumaceans, and mysids, and comparisons with fish stomachs were limited to these taxa. Underwater tele- vision and diver observations were used to direct in- vertebrate sampling to reef habitat. Further details of all fish and invertebrate sampling can be found in Sedberry and Van Dolah (1984) and Wenner et. al. (1984). ' Daytime submersible observations of vermilion snap- per behavior were made during dives aimed at visu- ally censusing (transects and point counts) reef fishes of the outer shelf on 14-15 July 1985. The submers- ible used was the Sea Link /, which provided a pan- oramic view (Askew, 1985) of shelf edge reefs and their fishes. Results Vermilion snapper was very abundant in trawl catches. Although it was relatively infrequent at inner shelf stations (mean catch per tow of 12.4), vermilion snap- per was very abundant at middle (243.7 per tow) and outer shelf (140.1 per tow) stations. Approximately 115 species of prey were identified in 255 stomachs that contained food. Fish with stomachs containing food were found at all times of the day; 22% of stomachs with food were collected between 0001 and 0600 hours local time; 17%, between 0601 and 1200 hours; 18%, between 1201 and 1800 hours; and 43%, between 1801 and 2400 hours. Although no at- tempt was made in the field to quantify proportions of stomachs with food, there were more stomachs with obvious contents at night ( 1801-0600 hr). Most prey items found in vermilion snapper stom- achs were planktonic or nektonic organisms (Table 1). Amphipods, mainly planktonic hyperiids and caprellids (e.g., Lestrigonus bengalensis, Phtisica marina), cope- pods, and decapods (e.g., larval forms, Lucifer faxoni) were the most frequently consumed taxa. These small crustaceans were eaten in large numbers by smaller vermilion snapper, but, with the exception of deca- pods, contributed little to prey volume (Table 2). Mysids, cumaceans, and doliolids were also frequently con- sumed. The overall diet was dominated volumetrically by squids and fishes. Significant feeding differences between size classes were detected by the Tyler chi-square feeding hetero- geneity index (Table 2). Several groups of crustaceans and fishes demonstrated significant differences in fre- quency between size classes of vermilion snapper. De- capods were important prey for larger fish but domi- nated the diet of fish between 50 and 100 mm SL. Small crustaceans, such as copepods, stomatopods and amphipods decreased in relative number and volume in the diet with increasing size of predator, although amphipods remained a frequent food item in all size classes examined. Barnacles and cumaceans varied in relative volume of prey for different size classes, but cumaceans increased in frequency in larger fish. Al- though squids and fishes were eaten by most size classes, they were volumetrically most important in 702 Fishery Bulletin 91(4), 1993 Table 1 Food items that occurred with a freq uency IF) of at least one percent of stomachs with food, or that made up at least one percent of the total number (Ar) or volume 1 V) of food in vermi lion snapper (Rhambr plites aurorubens) stomachs (n=255). All higher taxonomic groupings of prey are listed, regardless of their requeney. number, or volume. P =planktonic, H=holoplanktonic, D=demersal zooplankton; B=benthic, E=epifaunal, I=infaunal; N=nektonic Taxon and Taxon and prey item F N V prey item F N V Cnidaria Hyperiidea nematocysts 1.2 0.1 0.1 Euprone sp. (H) 1.2 0.1 <0.1 Annelida Hyperiidae undet. (H) 5.1 0.5 0.1 Total Polychaeta 11.9 1.8 2.5 Hyperiidea undet. (H) 5.9 0.7 <0.1 Phyllodoce longipes (I) 3.2 0.4 0.3 Lestrigonus bengalensis (H) 17.8 4.1 0.2 Mollusca Lycaea sp. (H) 1.2 0.1 <0.1 Total Gastropoda 5.1 0.9 0.3 Phronima sp. (Hi 2.0 0.7 0.3 Natica pusilla (B) 1.6 0.4 0.1 Phronima sedenlaria (H) 1.6 0.3 <0.1 Total Pelecypoda 1.2 0.3 <0.1 Phronimella elongata (H) 2.0 0.8 <0.1 Ervilia concentrica (I) 1.2 0.3 <0.1 Simorhynchotus sp. (Hi 1.6 0.2 <0.1 Total Cephalopoda 5.9 0.4 43.8 Total Decapoda 56.5 20.8 11.7 Loliginidae undet. 4.4 0.3 22.4 Decapoda undet. zoea (P) 3.2 0.2 <0.1 Loligo plei (N) 0.8 <0.1 21.1 Decapoda undet. larvae 2.0 0.2 <0.1 Crustacea Natantia undet. zoea (Pi 2.4 0.3 <0.1 Total Ostracoda 12.6 11.1 0.2 Natantia undet. shrimp 12.2 1.3 1.4 OstracodaA(H) 11.1 11.0 0.2 Penaeidea Total Copepoda 43.5 13.4 0.5 Lucifer faxoni (Hi 14.2 8.5 0.8 Calanopia americana (D) 2.8 0.5 <0.1 Penaeidae zoea (Pi 1.2 0.1 <0.1 Candacia curta (Pi 2.8 0.4 <0.1 Sicyonia typica 1.2 0.1 2.0 Centropages furcatus (H) 4.0 0.3 <0.1 Solenocera atlantidis 1.2 0.1 0.2 Labidocera aestiva (P.Bi 6.3 0.7 0.1 Caridea Labidocera sp. (H) 1.6 0.2 <0.1 Caridea undet. shrimp 1.2 0.1 <0.1 Oncaea sp. (H) 1.6 0.3 <0.1 Leptochela sp. 3.2 0.3 0.2 Sapphirma sp. (H) 1.2 0.1 <0.1 Leptoehela papulata 16.6 2.2 3.0 Temora styltfera (H) 9.5 1.6 <0.1 Ogyrides sp. 2.0 0.2 0.2 Temora turbinata (Hi 21.7 5.3 0.2 Processa sp. 1.2 0.1 0.1 Undinula vulgaris (Hi 3.2 0.6 <0.1 Anomura Total Cirripedia 16.2 3.1 2.5 Albunea parettu 4.7 1.4 0.2 Barnacle larvae (D) 15.4 3.0 2.5 Rantlia muricata zoea (Pi 2.0 0.2 <0.1 Total Stomatopoda 9.9 2.2 1.0 Brachyura Stomatopod larvae (Pi 8.3 2.0 0.8 Brachyura undet. zoea (P) 4.0 0.4 <0.1 Stomatopoda adults (Di 1.2 0.1 0.2 Brachyura undet. Total Mysidacea 16.6 5.2 0.8 megalopae (P.Bi 7.9 1.1 0.2 Anchialina typica (D) 1.2 0.1 <0.1 Brachyura undet. crab 5.5 0.4 0.4 Bowmaniella portorieensis 9.9 1.0 0.3 Calappidae zoea (P) 2.0 0.2 0.1 Bowmaniella sp. 1.2 0.1 <0.1 Ovalipessp. (P,Bl 1.2 0.1 0.4 Mysidopsis bigelowi (B) 1.2 0.1 <0.1 Pinnotheridae zoea (Pi 2.4 0.3 <0.1 Promysis atlantwa (Pi 2.0 3.6 0.4 Portunidae megalopae (P.B) 4.4 0.6 0.1 Total Cumacea 14.6 7.4 1.1 Portunidae crab (P.Bi 2.8 0.2 0.2 Cyclaspis varians (D) 5.5 0.7 0.1 Portunus sp. (P.B) 2.0 0.2 0.2 Oxyurostylis smithi (Dl 11.1 5.9 0.8 Xanthidae crab (B) 1.6 0.1 0.3 Total Isopoda 2.0 0.2 <0.1 Sipunculida (Bl 1.2 0.1 <0.1 Total Amphipoda 47.4 13.7 1.8 Chaetognatha (Pi 13.8 1.9 0.3 Gammaridea Chordata Ampelisca abdita (D) 1.2 0.1 <0.1 Thaliacea Ampelisca vadorum (Dl 2.0 0.1 0.1 Doliolida undet. iHi 8.3 7.4 1.2 Corophiidae undet. 1.6 0.3 <0.1 Larvacea 0.8 <0.1 <0.1 Gammaropsis sp. 2.0 0.2 <0.1 Cephalochordata Lysianopi 1.6 0.1 <0.1 Branchiostoma caribbaeum (D 3.2 0.4 (1 1 I'll at IS Sp. 1.2 0.1 <0.1 Total Teleostei 27.7 6.9 32.2 Rudilemboides naglei 2.0 0.3 <0.1 Anguilhformes 0.4 <0.1 1.3 Synchelidium americanum 1.6 0.1 <0.1 Prionotus sp. larvae IPi 1.2 4.2 2.6 Tiron tropakis 1 1 2.0 0.1 <0.1 Sardmella aunta (Ni 1.6 0.2 12.4 Caprellidea Teleostei undet. eggs (P) 2.0 3.0 <0.1 Caprellidae undet. 1.6 0.1 <0.1 Teleostei undet. larvae (Pi 3.2 0.4 0.6 Phtisica marina (P.Bi 14.6 2.4 0.3 Teleostei undet. 14.2 1.3 14.7 Sedberry and Cuellar: Planktonic and benthic feeding by Rhomboplites aurorubens 703 Table 2 Percent frequency occurrence tF), percent number (N). and percent volume [V of higher taxonomic groups of food in the diet of vermilion snapper {Rhomboplites aurorubens ). by length interval An asterisk (* indicates significant differences in prey frequency between one length interval and the adjacent larger interval, based on Tyler's ( 1979) test. Prey Length intervals (mm SL) 1-50 51-100 101-150 >150 F N V F N V F N V F N V Cnidaria (nematocysts) — — — — — — 2.9 0.3 0.9 _ _ Annelida Polychaeta 7.1* 1.6 2.1 10.9 0.4 2.8 13.3 2.2 2.8 10.9 1.6 2.2 Mollusca Gastropoda 7.1 0.9 0.2 1.8 0.1 <0.1 5.7 1.6 1.0 6.0 1.4 0.1 Pelecypoda — — — — — — 2.9 1.2 0.2 — — — Cephalopoda — — — 1.8 0.1 1.0 4.8 0.6 24.2 13.4 1.0 51.4 Crustacea Ostracoda 28.6 3.7 1.5 24.4* 29.1 3.6 7.6 1.0 0.1 3.0 0.2 <0.1 Copepoda 82.1 33.9 12.0 60.0 15.3 4.0 41.0* 13.1 1.0 16.4 1.9 <0.1 Cirri pedia 21.4 6.1 4.2 20.0 1.3 0.5 11.4 2.6 2.5 18.0 4.8 2.6 Stomatopoda 25.0 2.8 9.6 20.0* 4.6 8.4 3.8 0.4 1.5 4.5 0.3 0.3 Mysidacea 21.4 1.9 4.6 12.3 0.7 0.5 20.0 16.0 4.0 11.9 1.5 0.1 Cumacea 3.6 0.9 0.8 7.3* 0.4 0.4 21.0 21.7 5.0 14.9 4.3 0.3 Isopoda 3.6 0.2 0.2 — — — — * — — 6.0 0.5 0.1 Amphipoda 35.7 20.3 12.7 49.1 12.9 6.2 49.5 14.2 4.0 46.3 11.5 0.9 Decapoda 35.7* 25.0 31.1 70.9* 29.4 42.0 52.4 14.1 19.0 58.2 14.3 7.9 Sipunculida — — — — — — 1.9 0.2 0.1 1.5 0.1 <0.1 Chaetognatha 21.4 1.6 3.4 10.9 1.3 0.9 11.4 2.3 0.8 16.4 2.3 0.1 Chordata Thaliacea 3.6 0.2 0.4 9.1 1.5 1.6 5.7 2.0 0.4 13.4 24.9 1.3 Larvacea — — — — — — — — — 3.0 0.2 <0.1 Cephalochordata — — — 1.8 0.1 0.2 2.9 1.0 0.2 6.0 0.4 0.1 Teleostei 7.1* 0.7 17.1 43.6* 2.9 27.8 21.9* 5.5 32.4 38.8 28.8 32.7 Examined stomachs with food: 28 55 105 67 Mean SL (mm) offish with food: 36.4 77.8 131.1 168.1 the diet of larger vermilion snapper (Table 2). Squids and fishes were 76% of the prey volume for all vermil- ion snapper, but 84.1% of the prey volume (and only 29.8% by number) offish greater than 150 mm SL. Vermilion snapper fed sparingly on invertebrates closely associated with the reef habitat and collected in suction and grab samples during the day (Table 3). Electivity values were negative (usually -1.00) for all dominant species in benthic samples. Polychaete spe- cies that dominated those samples consisted mainly of tube-reef building species iFilograna implexa) and spe- cies associated with sponges and corals (Exogone dispar and Syllis spongicola) (Gardiner, 1975; Wendt et al., 1985); none of these were consumed by vermilion snap- per. On the other hand, many species that dominated numerically in the diet (e.g., Oxyurostylis smithi and Lucifer faxoni) were collected in benthic samples but were not a major component of the daytime reef fauna. Because abundance of these species in benthic samples was so low, electivity values were positive. Those spe- cies that were higher in relative abundance in stom- achs than in benthic samples (e.g., O. smithi) may have been consumed in the water column during peri- odic emergence. As in the case of benthic samples, most species of cumaceans, mysids, stomatopods, and decapods that dominated samples from the sled were not as rela- tively abundant in stomach samples (Table 4). Most electivity values were negative; however, fewer ab- sences from stomach samples (£=-1.00) occurred with the dominant species from sled samples than with the benthic samples. The mysid Promysis atlantica at the inner shelf (£=0.77) and the decapod Lucifer faxoni at the middle shelf (£=0.63) were dominant species in sled samples that were positively elected as prey. Lu- cifer faxoni was by far the most abundant species in 704 Fishery Bulletin 9I|4), 1993 Table 3 Relative abundance (percent of total number of individ iials, N and electivity values {El for dominant benthic species in suction and grab samples and in vermil on snapper (Rhonibophtes aurorubens) stomachs. Dominant species for each type of sample (benthic or stomach) include those that ranked in the five most abundant species in any depth zone for that type of sample. Inner shelf Middle shelf Outer shell Relative abundance Relative abundance Relative abund ance Fish Benthic Fish Benthic Fish Benthic stomachs samples E stomach s samples E stomach s samples E Dominant species — benthic samples Chone americana — 0.33 -1.00 — 0.81 -1.00 — 0.59 -1.00 Erichthonius brasilien sis 0.09 2.89 -0.94 — 0.30 -1.00 — 0.13 -1.00 Erichthonius sp. A — 0.08 -1.00 — — — — 3.75 -1.00 Exogone dispar 0.44 3.71 -0.79 — 0.47 -1.00 — 0.01 -1.00 Filograna implexa — 20.42 -1.00 — 63.87 -1.00 — 21.90 -1.00 Luconacia incerta — 3.27 -1.00 0.05 1.03 -0.90 — 0.18 -1.00 Malacoceros glutaeus — 0.41 -1.00 — 0.81 -1.00 — 0.02 -1.00 Phyllochaetopterus socialis — 0.21 -1.00 — 0.12 -1.00 — 12.40 -1.00 Pista palmata — 0.09 -1.00 — 0.08 -1.00 — 8.60 -1.00 Podocerus sp. A — 2.87 -1.00 — 0.27 -1.00 — 0.14 -1.00 Spiophanes bombyx — 0.39 -1.00 — 0.46 -1.00 — 5.81 -1.00 Syllis spongicola — 2.14 -1.00 — 1.90 -1.00 — 1.38 -1.00 Total 0.53 36.81 0.05 70.12 0.00 54.92 Dominant species — stomachs Bowmaniella portoricensis 0.18 0.09 0.32 1.80 0.10 0.89 0.50 0.22 0.37 Leptochela papulata 0.35 0.04 0.81 3.45 0.09 0.95 1.86 0.12 0.88 Lucifer faxoni 12.05 0.18 0.97 9.79 0.01 0.99 0.62 0.04 0.87 Oxyurostylis smithi 15.74 0.67 0.92 2.46 0.14 0.89 — — — Phtisica marina 0.09 0.01 0.75 1.80 0.03 0.97 7.20 0.40 0.90 Phyllodoce longipes 1.14 0.06 0.89 0.05 0.04 0.10 0.12 0.20 -0.22 Promysis atlantica 11.79 — 1.00 0.05 <0.01 0.88 — — — Rhudilemboidea naglei 1.06 0.05 0.91 — <0.01 -1.00 — 0.04 -1.00 Total 42.40 1.10 19.40 0.42 10.30 1.02 Number of stomachs with food: 49 138 68 sled samples at the inner shelf and was also abundant in stomach samples but demonstrated negative elec- tivity at the inner (£=-0.53) and outer (£=-0.39) shelf sites. At the middle shelf, L. faxoni was the most abun- dant species in the analyzed taxa in fish stomachs and ranked third in abundance in sled samples. Many of the species that were high in relative abundance in fish stomachs also occurred in sled samples; however, electivity values were not always positive. The rela- tive abundance of cumaceans, mysids, stomatopods, and decapods in fish stomachs, compared with their abundance in sled samples, indicated that vermilion snapper often selected crustaceans in higher propor- tions than was their availability to the plankton sled at night (Table 4). Included were several orders of Crus- tacea, particularly the cumacean Oxyurostylis smithi and the mysid Promysis atlantica at inner shelf sta- tions, the decapods Lucifer faxoni and Leptochela papulata at middle shelf stations, and the decapods Leptochela papulata and Solenocera atlantidis at the outer shelf. Comparing dominant species in the diet of vermil- ion snapper with their relative abundance in benthic (lower half of Table 3) and sled (lower half of Table 4) samples indicated high positive selectivity for most prey species from both environments sampled. For suction and grab samples, this was due to the extremely low abundance (<19() of all the dominant prey in the benthos. For the sled samples, many dominant prey species were also dominant in the environment. Discussion The vermilion snapper is well adapted to foraging in the water column (Davis and Birdsong, 1973; Grimes, 1979). Grimes (1979) reported that the diet of vermil- ion snapper was dominated by planktonic organisms Sedberry and Cuellar: Planktonic and benthic feeding by Rhomboplites aurorubens 705 Table 4 Relative abundance (percent of total num Der of indivi duals. N) and electivity values (E) for dominar t species of decapods, cumaceans, mysids and stomatopods in s ed samples and vermilion snapper {Rhomboplites lurorubens) stomachs. Dominant speci is for each type of sample (sled or stomach I include those that ranked in the five most abundant s pecies of decapods, cumaceans, mysids and stomatopods in any depth zone for that type of sample. Inner shelf Middle shelf Outer shelf Relative abundance Relative abunc ance Relative abundance Fish Sled Fish Sled Fish Sled stomachs samples E stomachs samples E stomachs samples E Dominant species — sled samples Bowmaniella portoricensis 0.30 5.56 -0.90 5.52 10.45 -0.31 4.30 5.45 -0.12 Lucifer faxoni 20.57 66.77 -0.53 29.93 6.79 0.63 5.38 12.29 -0.39 Mysidopsis furca — 0.67 -1.00 0.17 4.49 -0.93 — 22.28 -1.00 Neopontonides beaufortensis — 4.10 -1.00 — 3.35 -1.00 — 0.77 -1.00 Periclimenes iridescens — 3.18 -1.00 — 3.75 -1.00 — 0.84 -1.00 Pontophilus gorei — — — — 0.03 -1.00 — 7.33 -1.00 Promysis atlantica 20.12 2.57 0.77 0.17 7.24 -0.95 — 0.63 -1.00 Thor manning! — 0.05 -1.00 — 0.43 -1.00 — 2.44 -1.00 Total 40.99 82.90 35.15 36.63 9.68 52.03 Dominant species — stomachs Bowmaniella portoricensis 0.30 5.56 -0.90 5.52 10.45 -0.31 4.30 5.45 -0.12 Cyclaspis varians 1.05 0.05 0.91 3.18 1.45 0.37 — 0.41 -1.00 Lucifer faxon i 20.57 66.77 -0.53 29.93 6.79 0.63 5.38 12.29 -0.39 Leptochela pap u lata 0.60 0.39 0.21 10.54 2.07 0.67 16.13 1.61 0.82 Oxyurostylis smithi 26.88 1.38 0.90 7.52 1.19 0.73 — 0.42 -1.00 Promysis atlantica 20.12 2.57 0.77 0.17 7.24 -0.95 — 0.63 -1.00 Solenocera atlantidis — — — — 0.03 -1.00 4.30 1.12 0.59 Total 69.52 80.82 56.86 32.56 30.11 22.70 Number of stomachs with food: 49 . 138 68 and he noted that the diet of juveniles (<100mm TL) was dominated volumetrically by copepods. In the present study, decapods dominated the diet volume of fish less than 101mm SL. The feeding habits of ver- milion snapper changed considerably with size in the present study, although pelagic prey dominated in all size classes. As vermilion snapper grows, it switches from a diet of many small crustaceans, to a diet domi- nated by a few large cephalopods, fishes, or decapod crustaceans. The switch to different prey taxa and to fewer, larger prey individuals was similar to ontogenetic diet changes noted by Sedberry ( 1983 ) for several de- mersal fishes from the outer continental shelf. In con- trast, Schmitt and Holbrook (1984) found that black surfperch iEmbiotoca jacksoni) switched to larger prey, but that gross taxonomic composition of the diet did not change and that surfperch continued to feed on macrocrustaceans. in spite of growth and changes in body size and foraging behavior. Vermilion snapper, like many other fishes (Sedberry, 1983), apparently becomes capable of taking larger prey as it grows and switches from picking plankton to pursuing and cap- turing active nektonic species, such as Spanish sar- dine (Sardinella aurita), and squids. Grimes (1979) suggested that vermilion snapper is a nocturnal forager and that selective feeding by ver- milion snapper on demersal zooplankton, such as cumaceans, indicates nocturnal foraging. These crus- taceans are infaunal in the sand or epifaunal in reef crevices during the day but emerge at night (Anger and Valentin, 1976; Alldredge and King, 1985), becom- ing prey for vermilion snapper ancl other fishes that forage on near-bottom zooplankton at night. Demersal zooplankton demonstrating nocturnal emergence and found in the diet of vermilion snapper included syllid polychaetes, some calanoid copepods. cumaceans, am- phipods, decapods, barnacle and stomatopod larvae, chaetognaths and cephalochordates (Williams and Bynum. 1972; Fincham. 1974; Anger and Valentin, 1976; Hobson and Chess, 1976; Alldredge and King, 1977; Hammer, 1981; Alldredge and King, 1985; Cahoon and Tronzo, 1988). These taxa are dominant members of the demersal zooplankton (Porter and Porter, 1977) and emerge from benthic habitats at night when they 706 Fishery Bulletin 91(4), 1993 feed, molt, reproduce or disperse (Alldredge and King, 1985). This behavior makes them subject to intense predation by vermilion snapper or other specialized nocturnal predators (Robertson and Howard, 1978). Demersal zooplankton are approximately as abun- dant as holozooplankton on the continental shelf off North Carolina (Cahoon and Tronzo, 1992), and the nocturnal emergence of demersal zooplankton prob- ably increases overall food availability at night, while allowing planktivorous fishes to feed under the cover of darkness. Small vermilion snapper are prey for di- urnal lutjanids and crepuscular serranids (South Caro- lina Wildlife and Marine Resources Department1; Parrish, 1987; Sedberry, 1988) and would be subject to predation during the day. Vermilion snapper also consumes holoplanktonic spe- cies such as copepods, hyperiids, Lucifer faxoni, and doliolids, as well as nektonic squids and fishes. As noted by Grimes (1979), squids were especially impor- tant in the diet of large juveniles and adults. These larger size classes of vermilion snapper also fed more on fishes, which included schooling pelagic species such as Spanish sardine, Sardinella aurita. Spanish sar- dine makes up 9.4% of fishes caught in trawls on the southeastern continental shelf2. Predation by vermil- ion snapper on nektonic foragers such as squids and Spanish sardine provides a trophic link between the pelagic nekton and the reef. In the role as an abun- dant consumer of nekton, holoplankton, and demersal zooplankton, vermilion snapper may be important in transferring energy from benthic sand habitats and the water column to the reef, in the form of feces. Feces that disintegrate just above the reef provide fine particles and nutrients for filter- and suspension- feeders, and fecal pellets that are less refractile can be used directly by small crustaceans and other organ- isms living in the reef (Rothans and Miller, 1991). Ver- milion snapper appear to be relatively inactive during the day, hovering or moving slowly in large schools along the reefs within a meter of the bottom (pers. observ. by GRS from Sea Link I). By feeding in the water column at night and swimming just above the reef during the day, vermilion snapper enhances the transfer of this organic matter to the benthos. During these diurnal resting periods, vermilion snapper prob- ably deposit feces, derived from water column noctur- 'South Carolina Wildlife and Marine Resources Department. 1984. Final Report. South Atlantic OCS area living marine resources study. Phase III. Volume 1. Prepared by Marine Resources Research Insti- tute, SCVVMRD, for Minerals Management Service, Washington. D.C.. under contract No 14-12-0001-29185. 223 p. -Sedberry. G. R.. C. A. Barans, C. A. Wenner, and V. G. Burrell Jr. The ichthyofauna of sandy bottom habitat on the continental shelf off the southeastern U.S. Manuscr. in prep. nal foraging, onto the reef. Bray et al. (1981) also found that a planktivorous reef fish functioned as a trophic link between the plankton and benthos by importing organic carbon to the reef in the form of feces. Meyer and Schultz ( 1985) found that grunts (Haemulon spp.) feeding on sand flats transferred significant amounts of nutrient and organic matter to reefs, thus enhanc- ing coral growth. Benthic polychaetes ranking high in relative abun- dance in the suction and grab samples were not im- portant in the diet of vermilion snapper, and electivity values for all dominant species in the benthic samples were negative. The dominant crustaceans in benthic samples were epibenthic species such as the corophoid amphipod, Eriehthonius brasiliensis, and the caprellid amphipod Luconacia incerta which attach to sessile invertebrates and are usually closely associated with the reef substratum (McCain, 1968; Bousfield, 1973). Because E. brasiliensis and caprellid amphipods have also been found in nocturnal zooplankton samples (Wil- liams and Bynum. 1972; Hobson and Chess, 1976), they may be consumed by predators at night in the water column. Some motile benthic crustaceans such as mysids, cumaceans, and decapods that were com- mon in vermilion snapper stomachs were much higher in relative abundance in stomachs than in benthic samples. Vermilion snapper apparently prey on these benthic crustaceans during their periodic migrations into the water column. The polychaetes that dominated benthic samples apparently do not undertake such migrations. Vermilion snapper feed on many of the same prey species as the benthic-feeding sparid Stenotomus chrysops, an abundant demersal fish of the continen- tal shelf (Sedberry and Van Dolah, 1984; Sedberry, 1988; Sedberry et al.2). Submersible observations indi- cate that S. chrysops feeds during the day on the benthos, living in sand adjacent to reef habitat, whereas vermilion snapper were not observed to forage in this manner. Vermilion snapper apparently consumes sand- dwelling benthos at night, when they emerge from the bottom, while S. chrysops feeds on benthos during the day when demersal zooplankton has burrowed in the bottom. This provides a temporal partitioning of prey resources between these two dominant species of reef- associated fish. For reef fishes of the southeastern shelf, there is large variation in the degree of dependence on hard substrate as a habitat for prey (Sedberry, 1985, 1987, 1988). Of the three most abundant species in our reef trawl catches (S. chrysops, Haemulon aurolineatum, and R. aurorubens), all occur over sand bottom, al- though they are much more abundant over reefs (Wenner, 1983; Sedberry and Van Dolah, 1984). The Sedberry and Cuellar Planktonic and benthic feeding by Rhomboplites aurorubens 707 former two species feed primarily on sand bottom benthos (Sedberry, 1985, 1988). Although H. auro- lineatum and S. chrysops feed heavily on sand infauna and are not completely dependent on reef habitat, ver- milion snapper is more restricted to reef habitat and does not range far from a home reef (Fable, 1980). Wenner (1983) collected only 2 specimens in 11 trawl- ing tows in sand bottom habitat on the southeastern continental shelf, whereas S. chrysops and H. auro- lineatum are dominant species in sand habitats (Wenner, 1983; Sedberry-). Vermilion snapper did not, however, feed directly on reef fauna, and its attraction to reefs may be behavioral rather than trophic. Although vermilion snapper feeds extensively on de- mersal zooplankton, many individuals also feed oppor- tunistically on concentrations of holoplankton that oc- cur on the southeastern shelf. Copepods, such as Undinula vulgaris and especially Temora turbinata, that were frequent food items for vermilion snapper are epipelagic oceanic species that are transported land- ward across the continental shelf (Hopkins et al., 1981 ). Off the southeastern United States, upwelling at the shelf edge and cross-shelf transport of deep, highly productive oceanic water provide mechanisms for shore- ward movement of shelf-edge biota, including copep- ods (U. vulgaris, T. turbinata, Oncaea spp.), doliolids, sergestid decapods (e.g., Lucifer faxoni), and other abundant oceanic zooplankters (Yoder et al., 1983; Paffenhbfer et al., 1984). Vermilion snapper, one of the most abundant fishes at shelf edge depths ( Sedberry and Van Dolah, 1984), apparently takes advantage of this abundant resource and performs the function of transferring some of this oceanic productivity to bot- tom habitats on the continental shelf. In conclusion, vermilion snapper feeds on a variety of prey above the substrate, much of which is demer- sal zooplankton. Although it is unknown how much of the biomass of the daytime benthos is composed of nocturnally emerging demersal zooplankton, it is ap- parent that these organisms are an important food source for vermilion snapper and other reef fishes. De- mersal zooplankton, such as mysids, cumaceans, ampeliscid amphipods, and certain decapods, composed at least eight percent of the total volume of food for vermilion snapper in the present study and are more important in the diet of smaller size classes. While a small contribution, these benthic species provide a feed- ing opportunity for vermilion snapper and may be more important in the diet during periods of low productiv- ity of holoplankton. Nektonic fishes and cephalopods provide the greatest volume of food for larger vermil- ion snapper. Because it is prey for other reef preda- tors, vermilion snapper is an important trophic link among several habitats on the southeastern continen- tal shelf. Acknowledgments For their help in identification and confirmation of prey items, we thank the following; D. M. Knott, amphi- pods and copepods; C. B. O'Rourke and E. R. Hens, polychaetes; and E. L. Wenner, decapods. D. M. Knott, R. F. Van Dolah, P. H. Wendt, and E. L. Wenner pro- vided the data on invertebrate distribution and abun- dance for the electivity analyses. D. M. Knott, M. R. Collins, W. D. Anderson Jr., and anonymous reviewers provided helpful comments that improved the manu- script. This work was supported with funds provided by the Bureau of Land Management (Contract No. AA551-CT9-27) and the Minerals Management Ser- vice (Contract Nos. AA551-CT1-18 and 14-12-0001- 29185), U.S. Dept. of the Interior; and the National Marine Fisheries Service (South Carolina MARMAP contract No 50WCNF006002). Submersible time was provided through the NOAA Undersea Research Cen- ter at the University of North Carolina at Wilmington. Literature cited Alldredge, A. L., and J. M. King. 1977. Distribution, abundance, and substrate prefer- ences of demersal zooplankton at Lizard Island La- goon, Great Barrier Reef. Mar. Biol. 41:317-333. 1985. The distance demersal zooplankton migrate above the benthos: implications for predation. Mar. Biol. 84:253-260. Anger, K., and C. Valentin. 1976. In situ studies on the diurnal activity pattern of Diastylis rathkei (Cumacea, Crustacea) and its im- portance for the "hyperbenthos." Helgol. Wiss. Meeresunters. 28:138-144. Askew, T. M. 1985. Johnson-Sea-Link user's manual. Harbor Branch Foundation, Inc., Fort Pierce, Florida. Misc. Publ. 17, 29 p. Bousfield, E. L. 1973. Shallow- water gammaridean Amphipoda of New England. Cornell Univ. Press, Ithaca, New York. Brattegard, T., and J. H. Fossa. 1991. Replicability of an epibenthic sampler. J. Mar. Biol. Assoc. U.K. 71:153-166. Bray, R. N., A. C. Miller, and G. G. Geesey. 1981. The fish connection: a trophic link between plank- tonic and rocky reef communities? Science 214:204- 205. Cahoon, L. B., and C. R. Tronzo. 1988. A comparison of demersal zooplankton collected at Alligator Reef. Florida, using emergence and reen- try traps. Fish. Bull. 86:838-845. 1992. Quantitative estimates of demersal zooplankton abundance in Onslow Bay, North Carolina. USA. Mar. Ecol. Prog. Ser 87:197-200. 708 Fishery Bulletin 91(4). 1993 Chess, J. R. 1979. An airlift sampling device for in situ collecting of biota from rocky substrata. Mar. Technol. Soc. J. 12:20-23. Chester, A. J., G. R. Huntsman, P. A. Tester, and C. S. Manooch III. 1984. South Atlantic Bight reef fish communities as represented in hook-and-line catches. Bull. Mar. Sci. 34:267-279. Collins, M. R., and J. L. Pinckney. 1988. Size and age at maturity for vermilion snapper {Rhomboplites aurorubens) (Lutjanidae) in the South Atlantic Bight. Northeast Gulf Sci. 10:51-53. Collins, M. R., and G. R. Sedberry. 1991. Status of vermilion snapper and red porgy stocks off South Carolina. Trans. Am. Fish. Soc. 120:116- 120. Davis, W. P., and R. S. Birdsong. 1973. Coral reef fishes which forage in the water col- umn. A review of their morphology, behavior, ecology and evolutionary implications. Helgol. Wiss. Meer- esunters. 24:292-306. Fable, W. A. 1980. Tagging studies of red snapper (Lutjanus campechanus) and vermilion snapper (Rhomboplites aurorubens) off the south Texas coast. Contrib. Mar. Sci. 23:115-121. Fincham, A. A. 1974. Periodic swimming behaviour of amphipods in Wellington Harbor. N.Z. J. Mar. Freshwater Res. 8:505-521. Gardiner, S. L. 1975. Errant polychaete annelids from North Caro- lina. J. Elisha Mitchell Sci. Soc. 91:77-220. Grimes, C. B. 1979. Diet and feeding ecology of the vermilion snap- per, Rhomboplites aurorubens (Cuvier) from North Carolina and South Carolina waters. Bull. Mar. Sci. 29:53-61. Grimes, C. B., C. S. Manooch, and G. R. Huntsman. 1982. Reef and rock outcropping fishes of the outer continental shelf of North Carolina and South Caro- lina, and ecological notes on the red porgy and ver- milion snapper. Bull. Mar. Sci. 32:277-289. Hammer, R. M. 1981. Day-night differences in the emergence of de- mersal zooplankton from a sand substrate in a kelp forest. Mar. Biol. 62:275-280. Hillier, A. J. 1974. URI high rise series bottom trawl manual. Univ. Rhode Island Mar. Bull. 20. Hobson, E. S., and J. R. Chess. 1976. Trophic interactions among fishes and zooplank- ters near shore at Santa Catalina Island, Cali- fornia. Fish. Bull. 74:567-598. Hopkins, T. L., D. M. Milliken, L. M. Bell, E. J. McMichael, J. J. Heffernan, and R. V. Cano 1981. The landward distribution of oceanic plankton and micronekton over the West Florida shelf as re- lated to their vertical distribution. J. Plankton Res. 3:645-658. Ivlev, V. S. 1961. Experimental ecology of the feeding of fishes. Yale Univ. Press, New Haven, Connecticut. Low, R. A. Jr., G. F. Ulrich, C. A. Barans, and D. A. Oakley. 1985. Analysis of catch per unit of effort and length composition in the South Carolina commercial hand- line fishery, 1976-1982. N. Am. J. Fish. Manage- ment 5:340-363. Manooch, C. S. HI. 1977. Food habits of the red porgy, Pagrus pagrus Linnaeus (Pisces: Sparidae), from North Carolina and South Carolina. Bull. Mar. Sci. 27:776-787. McCain, J. C. 1968. The Caprellidae (Crustacea: Amphipoda) of the western North Atlantic. U.S. Natl. Mus. Bull. 278:1- 141. Meyer, J. L., and E. T. Schultz. 1985. Tissue condition and growth rate of corals asso- ciated with schooling fish. Limnol. Oceanogr. 30:157- 166. Miller, G. C, and W. J. Richards. 1980. Reef fish habitat, faunal assemblages and fac- tors determining distributions in the South Atlantic Bight. Proc. Gulf Caribb. Fish. Inst. 32:114-130. Paffenhofer, G.A., B. T. Wester, and W. D. Nicholas. 1984. Zooplankton abundance in relation to state and type of intrusions onto the southeastern United States shelf during summer. J. Mar. Res. 42:995-1017. Parrish, J. D. 1987. The trophic biology of snappers and groupers. In J. Polovina and S. Ralston (eds.i. Tropical snap- pers and groupers: biology and fisheries management, p. 405-463. Westview Press. Porter, J. W., and K. G. Porter. 1977. Quantitative sampling of demersal plankton mi- grating from different coral reef substrates. Limnol. Oceanogr. 22:553-556. Robertson, A. I., and R. K. Howard. 1978. Diel trophic interactions between vertically mi- grating zooplankton and their fish predators in an eelgrass community. Mar. Biol. 48:207-213. Rothans, T. C, and A. C. Miller. 1991. A link between biologically imported particulate organic nutrients and the detritus food web in reef communities. Mar. Biol. 110:145-150. Schmitt, R. J., and S. J. Holbrook. 1984. Ontogeny of prey selection by black surfperch Embiotoca jacksoni (Pisces: Embiotocidae): the roles of fish morphology, foraging behavior, and patch selection. Mar. Ecol. Prog. Ser. 18:225-239. Sedberry, G. R. 1983. Food habits and trophic relationships of a com- munity of fishes on the outer continental shelf NOAA Tech. Rep. NMFS SSRF 773. 56 p. 1985. Food and feeding of the tomtate, Haemuhm aurolineatum (Pisces, Haemulidae), in the South At- lantic Bight. Fish. Bull. 83:461-466. 1987. Feeding habits of sheepshead, Archosargus pro- batocephalus, in offshore reef habitats of the south- eastern continental shelf. Northeast Gulf Sci. 9:29-37. Sedberry and Cuellar Planktonic and benthic feeding by Rhomboplites aurorubens 709 1988. Food and feeding of black sea bass, Centropristis striata, in live bottom habitats in the South Atlantic Bight. J. Elisha Mitchell Sci. Soc. 104:35-50. Sedberry, G. R., and R. F. Van Dolah. 1984. Demersal fish assemblages associated with hard bottom habitat in the South Atlantic Bight of the USA. Env. Biol. Fish. 11:241-258. Struhsaker, P. 1969. Demersal fish resources: composition, distribu- tion and commercial potential of the continental shelf stocks off the southeastern United States. Fish. Ind. Res. 4:261-300. Tyler, A. V. 1979. Statistical analysis of diet differences related to body size. In S. J. Lipovsky and C. A. Simenstad (eds.), Gutshop 78, fish food habits studies, proceed- ings of the second Pacific Northwest technical work- shop, p. 51-55. Univ. of Washington Sea Grant Publ. WSG-WO-79-1, Seattle. Wendt, P. H., R. F. Van Dolah, and C. B. O'Rourke. 1985. A comparative study of the invertebrate macrofauna associated with seven sponge and coral species collected from the South Atlantic Bight. J. Elisha Mitchell Sci. Soc. 101:187-203. Wenner, C. A. 1983. Species associations and day-night variability of trawl caught fishes from the inshore sponge-coral habi- tat. South Atlantic Bight. Fish. Bull. 81:537-552. Wenner, E. L., D. M. Knott, R. F. Van Dolah, and V. G. Burrell Jr. 1983. Invertebrate communities associated with hard bottom habitats in the South Atlantic Bight. Estuarine Coastal Shelf Sci. 17:143-158. Wenner, E. L., P. Hinde, D. M. Knott, and R. F. Van Dolah. 1984. A temporal and spatial study of invertebrate com- munities associated with hard-bottom habitats in the South Atlantic Bight. NOAA Tech. Rep. NMFS 18, 104 p. Williams, A. B., and K. H. Bynum. 1972. A ten-year study of meroplankton in North Caro- lina estuaries: amphipods. Ches. Sci. 13:175-192. Windell, J. T. 1971. Food analysis and rate of digestion. In W. E. Ricker (ed.), Methods for assessment offish produc- tion in fresh water. Blackwell Scientific Publ., Lon- don, p. 215-226. Yoder, J. A., L. P. Atkinson, S. S. Bishop, E. E. Hofmann, and T. N. Lee. 1983. Effect of upwelling on phytoplankton productiv- ity of the outer southeastern United States continen- tal shelf. Con. Shelf Res. 1:385-404. Abstract. -The relationship be- tween latitude and birth timing was assessed for captive-born California sea lions (Zalophus californianus), northern (Steller) sea lions (Eumetopias jubatus), northern fur seals (Callorhinus ursinus), and Pa- cific harbor seals iPhoca vitulina richardsi) from zoos and aquaria in the United States, Canada, South Korea, and New Zealand. The births of 466 viable California sea lion pups demonstrated highly significant, negative and curvilinear latitudinal variation in birth timing. Over the latitudinal range of captive births, this variation accounted for a change of approximately -0.6 days/°latitude. Furthermore, the variances of the mean dates of birth for the largest 18 captive populations were signifi- cantly dependent upon latitude; shorter birthing periods occurred at higher latitudes. Northern sea lions (n=9) had a similar, but non-signifi- cant latitudinal relationship in which birthing dates occurred ap- proximately 30 days later. No sig- nificant relationship between lati- tude and birthing date was found for northern fur seals (rc=13). The birth dates of 110 viable Pacific har- bor seal pups had highly significant, positive and curvilinear latitudinal variation, similar to that previously described for this subspecies in the wild between 30° and 47°N. Pupping dates for each species in captivity were comparable to those found for wild populations of North Pacific pin- nipeds. The described latitudinal variation and the temporal consis- tency between captive and wild populations of California sea lions and Pacific harbor seals support the hypothesis that photoperiod re- sponse maintains specific birth tim- ing in these species. Latitudinal variation in the birth timing of captive California sea lions and other captive North Pacific pinnipeds Jonathan L. Temte Department of Zoology, 42 1 Birge Hall. University of Wisconsin Madison, Wisconsin, 53706 Present address: Department of Family Medicine. University of Wisconsin 777 South Mills, Madison, Wl 537 I 5 Manuscript accepted 4 June 1993. Fishery Bulletin 9 1 : 7 10-7 1 7 ( 1993 1. 710 Captive pinnipeds are not only of educational and entertainment value to zoo and aquarium visitors, but also offer an exceptional opportunity to examine temporal aspects of their re- productive biology. Zoo records usu- ally document birth dates, outcomes, and parental histories which are dif- ficult, if not impossible, to obtain in field studies. Moreover, captive animals live in environments where availability of food is relatively con- stant and any movements (i.e., trans- fers between facilities) are well documented. A recent census (Asper et al., 1988) identified 924 pinnipeds in captivity in North America. Of these, four North Pacific species — the California sea lion (Zalophus californianus), the northern or Steller sea lion (Eumeto- pias jubatus), the northern fur seal (Callorhinus ursinus), and the har- bor seal (Phoca vitulina) — repre- sented 91 percent («=837). These spe- cies breed successfully in captivity over wide latitudinal ranges. Standard reviews of the pinnipedia (Mate, 1979; Mate and Gentry, 1979; Udell, 1981; Schusterman, 1981; King, 1983) identify fixed seasons of birth for each of the North Pacific sea lions. Moreover, Bigg (1973) found that captive California sea li- ons at the London Zoo (5 IN) breed at the same time as the parent popu- lation in California (33°N). In con- trast, Schusterman et al. (1982) noted that births of California sea lion pups at Sea Life Park, Hawaii (21°N), oc- curred significantly later than births in California at Marineland (34°N) and Marine World (38°N). Regional or latitudinal variation in birth timing, or both, has been re- ported for the harbor seal (Bigg, 1969a; Temte et al, 1991) and the grey seal (Halichoerus grypus: Coulson, 1981). Likewise, Temte (1985) demonstrated a 14-day shift in the mean date of pupping for northern fur seals between colonies on St. George Island, Alaska (57N), and San Miguel Island, California (34°N). These variations in birth tim- ing have been interpreted as 1 ) se- lection acting on discrete populations to match reproductive efforts with seasonal constraints (Bigg, 1973). 2) responses to latitudinally chang- ing temporal cues, such as tempera- ture or photoperiod (Coulson, 1981; Temte, 1985), or 3) a combination of the above (Temte et al., 1991; Boyd, 1991). This study identifies and defines the extent of latitudinal variation in birth timing of captive California sea lions, northern sea lions, northern fur seals and Pacific harbor seals from birthing records available from zoos and aquaria scattered across a wide latitudinal range. Comparisons of birthing periods were made, where applicable, between captive and wild populations. Temte: Birth timing of captive Zalophus californianus Methods The data Data for this analysis were initially compiled from the Marine Mammal Inventory Report (MMIR), a registry of captive marine mammals maintained by the Office of Protected Resources, National Marine Fisheries Ser- vice. To comply with the Marine Mammal Protection Act, U.S. facilities that exhibit marine mammals sub- mit yearly summaries of their captive populations, in- cluding animal identification, sex, age, date of acquisi- tion or birth, origin, and current status. The birth dates of 527 California sea lions born in captivity at 41 loca- tions, 9 northern sea lions born at two locations, 20 northern fur seals born at three locations, and 125 Pacific harbor seals born at 13 locations were extracted from this registry (Fig. 1). Seventy-one percent of birth date and outcome (viable pup, non-viable pup, stillborn pup) data were independently verified by contacting the appropriate facility. Of the data veri- fied, 98.6% were accurately reported on the MMIR (Temte, 1993). Eighty-three additional captive births of California sea lions, not reported on the MMIR, were provided by the facilities. Seven additional birth dates of captive north- ern fur seals were obtained from the report of Bigg ( 1984 ). Data on two births of harbor seals were obtained from the Vancouver Aquarium (Vancouver, British Columbia). To assess the effect of a major latitudinal translocation, birthing dates of 11 California sea lions born at the Auckland Zoo, New Zealand (37°S) were obtained. All cows at this facility could be traced, from zoo records, to the California Channel Islands population. Pup identification, location, latitude, date of birth, sex, and birth outcome were entered into the data- base. Dates of birth were converted into numerical equivalents by using a sequential astronomical calen- dar (day 0 = December 21; see Temte, 1985). Birth outcomes were rated as stillbirth (pup born dead), non- viable (pup lived less than one day), or viable (pup lived at least one day). Figure 1 Locations of zoos, aquaria and marine zoological parks where captive births of Cali- fornia sea lions (filled circles), northern sea lions lopen boxes), northern fur seals istars) and Pacific harbor seals (open circles) have been reported. Sea Life Park, Waimanalo, Hawaii ( 21°20'N: CSL), Auckland Zoo, Auckland, New Zealand <37°52'S: CSL) and Seoul Grand Park, Seoul. South Korea (37:31'N: PHS) are not shown. The shaded area indicates the natural breeding range of the California sea lion. Statistical analysis Data were treated separately for each species. Comparisons were made between the mean dates of birth for stillborn, non-viable, and viable pups by using the appropri- ate parametric or non-parametric statistic. The birth dates of viable pups were assessed for latitudinal var- iation. Least squares linear re- gression models were fitted to the latitude-date data, and where nec- essary, orthogonal polynomials were used to meet statistical assump- tions. Because the intra-colony vari- ance of the mean date of birth for California sea lions was related to latitude, regressions were per- formed with and without weighting to correct for unequal variances. Or- thogonal polynomial regression models using (latitude minus mean latitude )J as the quadratic term were chosen to reduce correlation between the linear and quadratic terms and to better estimate the linear coefficient (Snedecor and Cochran, 1980). To assess differ- ences between California and north- ern sea lions, analysis of covariance (ANCOVA: Kleinbaum and Kupper, 712 Fishery Bulletin 91(4). 1993 1978) was utilized to control for differences in latitudi- nal distributions. Results California sea lion The captive births of 610 California sea lions were evaluated (Table 1). Stillborn and non-viable pups ac- counted for 122 births (20.0%). Stillbirths occurred as early as 1 January and as late as 1 August. Whereas no significant difference was found between the tim- ing of stillbirths and non-viable births (Kruskal- Wallace: df=l; #=0.371; NS), stillborn and non-viable pups were born an average of 28.0 days earlier than viable pups, this difference being highly significant (Kruskal-Wallace: df=l; H=70.49; P<0.001>. Hence, only viable births (n=466; 22 pups with estimated birth dates were excluded) were examined for latitudinal variation. Viable births occurred over a 21°46' latitudinal range, between 2P20'N and 43°06'N. These were normally distributed over the birthing period (Fig. 2), occurring as early as 30 April and as late as 1 September. Mean dates of pupping for individual colonies occurred from 27 May to 25 June. The mean date of birth for all viable pups was 12 June ± 13.5 days (± SD). The variances of the mean birth dates were calcu- lated for the 18 largest captive colonies (each with at least 6 viable births). A significant latitudinal gradient existed for this variance (Fig. 3), with a decrease of 5.6 days2 for each degree of northward displacement. A plot of pupping date versus latitude indicated that latitude variation occurred within captive California sea lions (Fig. 4). A second order orthogonal polyno- mial model (Table 2) was highly significant (r-'=0.206; F ;, „ =60.23; P<0.0001 ) and estimated a negative lati- tudinal slope of approximately -0.6 days/°latitude over W 140- f20 CD XI E 3 California Sea Lion 466 100 110 120 130 140 150 160 170 160 190 200 210 220 230 240 250 Pupping Date Figure 2 Frequency of viable California sea lion births at all facilities during 10-day periods from 31 March (day 1001 to 6 Septem- ber (day 259). Dates are numbered sequentially from 21 De- cember. 0 California Sea L ion ■ 0° 0 0 <3D 0 r2= 0.405; P < 0.005 0 0 8 0 — . — . — . — . — 1 c o *i_ o '00 > .E 50 Q. Q_ 3 Q- o-l — , — , — , — i — i — i — i — i — , — i — i — i — , — i — i — .— . — —r 20 25 30 35 40 45 Latitude of Captive Colony (°N) Figure 3 Intra-colony variance of the mean pupping date from the 18 largest California sea lion colonies as a function of latitude. Each colony had at least six viable births. Linear regression model: Variance (days2) = 219 - 5. 58( latitude); r- = 0.405; ',!„„. = -3.30; P < 0.005. Table 1 Numbers, mean dates month/day; sequential day number), and standard deviations of viable births, non-viable births, and stillbirths in captive North Pacific pinnipeds. Species Type n Percent Mean date SD(days) California sea lion Viable 488 80.0 6/12(172.7) 13.5 Non-viable 50 8.2 5/16(145.6) 29.4 Stillborn 72 11.8 5/14(144.0) 46.1 Northern sea lion Viable 9 100.0 7/09(199.8) 16.5 Northern fur seal Viable 13 48 1 7/10(200.6) 14.8 Stillborn 14 51.9 6/23(184.2) 60.0 Pacific harbor seal Viable 110 88.0 5/20(150.3) 36.7 Non-viable 5 4.0 5/11(140.8) 47.5 Stillborn 10 8.0 5/11(140.8) 34.0 Temte: Birth timing of captive Zalophus cahfomianus 713 California Sea Lion - -o- ■ Latitude (°N) Figure 4 The birth timing of viable California sea lions as a function of latitude. See Table 2 for regression model. Northern sea lion The captive births of nine northern sea lions occurred at 32°45'N and 41°21'N (latitudinal range=8°36). All births were of viable pups. The mean date of birth was 9 July ± 16.5 days (± SD; Table 1). Although the plot of pupping date vs. latitude was similar to that for the California sea lion, a simple linear regression model (Table 2) failed to demonstrate significant latitudinal effects in this species (/-2=0.271; Fll7i=2.60; P=0.23); the lack of significance was possibly due to the small sample size. However, when northern sea lion data were combined with those of the California sea lion in a multiple linear regression model, no significant dif- ferences were found between the latitudinal slopes of the two species. Analysis of covariance identified sig- nificantly later birthing (P<0.001) for northern sea lions, occurring 30 days after California sea lions. the range of data. Weighting of the regression to cor- rect for the unequal variances did not appreciably change the parameter estimates. Eleven pups were born at the Auckland Zoo (Auckland, New Zealand: 37° S) from 28 November to 23 December and had a mean birth date of 11 Decem- ber. The polynomial model predicted that birthing at 37°N should occur on 6 June. Therefore, California sea lions translocated to the Southern Hemisphere ex- perienced an approximate 6-month shift in birth timing. Climatic conditions did not appear to significantly affect the timing of birth. For example, colonies at similar latitudes (Vallejo, California: 37°47'N and Kan- sas City, Missouri: 39°07'N) maintained similar pup- ping schedules despite having vastly different seasonal temperature cycles. Following an adjustment for lati- tude, no significant difference was found between the birth timing of male and female pups. Northern fur seal The captive births of 27 northern fur seals were re- ported from four locations between 32°45'N and 49'07'N (latitudinal range=16°22'l. Slightly more than half of the pups were stillborn (Table 1). The mean date of birth of viable pups was 10 July ± 14.8 days (± SD). No significant latitudinal variation was detected in birth timing (Table 2: r~=0.174; Fa „|=2.32; NS). Pacific harbor seal The birth dates of 127 captive-born Pacific harbor seals were identified from 14 locations between 32°45'N and 49°18'N (latitudinal range=16°33'). The birth outcomes of two pups were not known and they were excluded from further analysis. Of the remaining 125 births, 88.0^ were of viable pups (Table 1). The mean date of birth for viable pups was 20 May ± 36.7 days (± SD). No significant differences in the annual birth timing Table 2 Regression models for latitudinal variation in bii •th timing of North Pacific pinnipeds. Mean *Linear regression estimate of latitude PD = a + 6,(L) + b,lL-L)- Species n ("North a 6,(CI)±95% b, (CI) ±95', r2 Significance California sea lion 466 33.21 190 -0.61 (0.17i 0.057 (0.024) 0.206 0.0001 Northern sea lion 9 35.62 271 -2.00 (2.43) — — 0.271 NS Northern fur seal 13 45.03 145 1.23 (1.59) — — 0.174 NS Pacific harbor seal 110 41.60 -113 5.85 (1.10) 0.681 (0.065) 0.508 0.0001 *PD = Pupping date (sequential from 21 December); L = Latitude. 714 Fishery Bulletin 91(4), 1993 were found among viable, non-viable and stillborn pups. A simple linear regression model defined significant latitudinal variation in birth timing of viable pups with a gradient of 4.10 days/°latitude; this slope was not significantly different from that reported by Temte et al. (1991) for colonies of wild Pacific harbor seals be- tween 30° and 47°N. A second order polynomial model (Fig. 5; Table 2), however, was more statistically ap- propriate based on residual analysis. This model was highly significant (r2=0.508; Fl2107l=55.23; P<0.0001), defining a positive relationship between latitude and birth date. Data regarding sex of pup (n = 102), maternal age (n=88), and the previous annual cycle of the mother (pregnant vs. non-pregnant: rc=89) were available for subsets of the harbor seals. When entered, either sepa- rately or as a group, into a multiple regression model adjusting for latitude, none of these parameters was found to significantly alter birth timing. Discussion Use of captive birth data Events on rookery sites or pupping beaches hamper the estimation of true pupping seasons in pinnipeds. For example, premature pupping in the California sea lion occurs with increasing frequency over a 5-month period, melting into the normal pupping season be- tween mid-May and the end of June (Delong et al., 1973). In addition, seasonal movements of animals from breeding to feeding sites may occur (Braham, 1974; Mate, 1975), potentially obscuring latitudinal varia- tion of birthing if seasonal entrainment occurs at lati- Pacific Harbor Sea! Latitude (°N) Figure 5 The birth timing of viable Pacific harbor seals as a function of latitude. See Table 2 for regression model. tudes other than that of the rookery site. The use of captive populations permits accurate measurement of birth dates and birth outcomes, while controlling for potential latitudinal displacement during the repro- ductive year. California sea lion This study confirms the report of Schusterman et al. (1982) and demonstrates that marked latitudinal varia- tion occurs not only in the timing of birth, but also in the variance of the mean birth date. Furthermore, in- terpretations of results based on this data set repre- senting 466 viable births from 41 locations, as com- pared to the three locations used by Schusterman et al. (1982), are far less sensitive to possible confound- ing effects induced by captivity or differences in cli- matic conditions. The extreme example of sea lions translocated to the Southern Hemisphere provides fur- ther evidence of strong latitudinal effect. The birth timing of captives mirrors that of wild populations. Although a review by Mate (1979) con- cluded that pups, regardless of latitude, are born from mid-May to late-June over the range of California sea lion rookeries, slight latitudinal variation may exist. In Baja California, pupping occurs in late-June (Brownell et al., 1974), and Le Boeuf et al. (1983) re- ported a maximum pup count on 10 July at Los Islotes (28°N). Further to the north, at San Nicolas Island (33°15'N), pupping peaks during the first half of June (Peterson and Bartholomew, 1967; Odell, 1975; Heath and Francis1-). Data from more northerly colonies are anecdotal in nature. For example, Braham (1974) re- ported a 2-week-old pup on 11 June at San Luis Obispo County, California (35°30'N). As in captive sea lions, premature pupping in wild colonies occurred as early as January on San Nicolas and San Miguel Islands (Odell, 1970; Delong et al, 1973). Premature pups are probably represented in captivity by the stillborn and non-viable groups. The 209r rate of stillborn and non-viable pupping for cap- tive sea lions is noteworthy considering the 5-16f/< rate of premature pupping reported for San Nicolas Island during years with high incidences of prematurity (Odell, 1970). Whereas pesticide exposure and disease 'Heath. C, and J. Francis. 1983. California sea lion population dy- namics and feeding ecology with applications for management. Re- sults of 1981-1982 research on Santa Barbara and San Nicholas Islands. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., South- west Fish. Sci. Center. P.O. Box 271. La Jolla, CA 92038. Admin. Rep. LJ-83-04C. Heath, C, and J. Francis. 1984. Results of research on California sea lions, San Nicholas Island, 1983. U.S. Dep. Commer., NOAA. Natl. Mar. Fish. Serv., Southwest Fish. Sci. Center, P.O. Box 271, La Jolla, CA 92038. Admin. Rep. LJ-84-41C. Temte: Birth timing of captive Zalophus cslifornianus 715 have been suggested as agents of premature pupping in wild populations (Delong et al., 1973; Odell, 1970, 1981), the high rate of stillbirth and non-viability in captivity may be influenced by the level of surveil- lance inherent to captive populations. Male California sea lions were reported to be 3.37 cm longer and 1.41kg heavier than female pups at birth (Le Boeuf et al., 1983). Year-to-year differences in mean birth lengths and weights have been related to differences in the mean birth date of harbor seals (Boulva, 1975). No such differences were found be- tween the birth timing of male and female pups in the present study. Hence, discrepancy in birth size is likely due to greater growth velocity of fetal males. Northern sea lion Only limited data from two locations were available for captive northern sea lions. Although these animals appeared to have a temporal pattern of birth similar to, but 30 days later than, that of the closely related California sea lion, no significant latitudinal trend was found. The northern sea lion has a wide distribution of breeding (see Fig. 1 in Loughlin et al., 1984), but little evidence exists for latitudinal variation in birth tim- ing in the wild. For example, a review of median birth dates at rookeries from 37°N to 60°N by Merrick ( 1987) failed to demonstrate any latitudinal variation in this species. Northern fur seal A high rate of stillbirth was noted for the fur seal. This phenomenon was previously reported by Bigg ( 1984) during studies indicating that contact with sub- strate (e.g., arrival on shore) may stimulate parturi- tion. Captive environments, with shallow pools and access to platforms, cannot adequately reproduce the pelagic environment which female fur seals inhabit during most of active gestation. Hence, premature par- turition may be induced by enclosures. The mean date of pupping for viable pups in captiv- ity of 10 July was the same as that derived for St. George Island, Alaska (Temte, 1985). Unlike wild north- ern fur seals, however, the captives demonstrated no significant latitudinal variation. The sample set may have been too small to detect the estimated 0.6 days/iatitude trend reported by Temte ( 1985). Pacific harbor seal Captive harbor seals had a high rate of viable birthing with pups born over a pupping season comparable to that of wild counterparts (Temte et al., 1991). This species demonstrated stronger latitudinal effect than that of any other North Pacific pinniped covered in this report, with an average shift of 4.1 days/°latitude. Whereas this relationship is similar to that previously described for wild Pacific harbor seals on the North American west coast south of 47°N, the larger captive data set allows better definition of a curvilinear relationship. Temte et al. ( 1991) separated the Pacific harbor seal into three subgroups based on birth timing. The north- ernmost group, from northern British Columbia and Alaska has no latitudinal variation in birth timing. The late-birthing group inhabits Puget Sound, Wash- ington, and pups two months later than coastal seals at the same latitudes. The southern group, as noted above, has highly significant latitudinal variation. Mor- phometric analysis of skulls from each of these popu- lations supports the hypothesis of discrete populations (Temte, unpubl. data). No individuals from populations originating in Puget Sound, northern British Colum- bia or Alaska were included in this study The sex of the pup, maternal age, and the previous maternal cycle (pregnant vs. non-pregnant) had no ef- fect on birth timing. These findings are from captive animals with dependable food supplies. In contrast, Boyd ( 1984) has suggested that maternal condition af- fects implantation timing in grey seals. Latitudinal variation and photoperiod For captive California sea lions, a smooth and con- tinual temporal change of pupping dates across lati- tude occurred despite wide diversity of climatic condi- tions. Nevertheless, as almost all the captives could be traced to a single wild population on the California Channel Islands, a strong environmental component appears to influence birth timing. This pattern is highly suggestive of a response to a predictable, latitudinally dependent seasonal cue. The 6-month shift in birth timing between hemispheres, as demonstrated by Cali- fornia sea lions in New Zealand, strongly supports a photoperiod hypothesis. Furthermore, the decreasing variance to the north (in the Northern Hemisphere) is as expected if photoperiodism occurs (Bronson, 1985). At higher latitudes, organisms experience a greater rate of change in photoperiod. Consequently, responses to specific cues should occur over a compressed time period. Latitudinal variation in birth timing of the northern fur seal, a species with delayed implantation (Daniel, 1981), can be explained by a response to photoperiod occurring between ovulation and implantation (Temte, 1985). This sets the time of implantation and birth while allowing flexibility in estrus timing. For example, arrival at breeding grounds and mating occur as much 716 Fishery Bulletin 91(4). 1993 as two months earlier for parous cows than for virgin northern fur seals (Craig, 1964; Bigg, 1986); this dis- parity is not reflected in birth timing (Trites, 1992). Gentry (1981) noted, however, that individual parous females have fairly rigid year-to-year estrus timing, and Bigg (1984) has suggested that the breeding synchrony found in this species may be due to arrival on shore and on social factors. The harbor seal has been shown to have delayed implantation (Fisher, 1954; Bigg, 1969b) and the Cali- fornia and northern sea lions are thought to have this reproductive pattern as well (Odell, 1975; Boshier, 1981; Schusterman, 1981). Hence, a similar mechanism may exist to signal implantation in these species. Temte ( 1985) proposed that northern fur seals responded to a photoperiod of 12.5 hours per day (prior to the autum- nal equinox), thus explaining the positive slope in lati- tudinal variation. Likewise, the Pacific harbor seals in this study have a positive slope and may respond to long (>12. Oh/day) photoperiods. In contrast, California sea lions have a negative slope in latitudinal varia- tion, which suggests a response to a photoperiod slightly less than 12.0 hours per day (following the autumnal equinox), and occurring prior to implantation. Acknowledgments I thank the many dedicated personnel of the zoos, aquaria, and marine zoological parks for cooperation in data collection and verification, and am most grate- ful to Ann Terbush and Wanda Cain of the Office of Protected Resources, National Marine Fisheries Ser- vice, for providing Marine Mammal Inventory Reports on these species. M. D. Sibley of the Auckland Zoo made available birthing records from the Southern Hemisphere. I also thank Dan Odell of Sea World, Ted Garland of the University of Wisconsin, Department of Zoology, and two anonymous reviewers who pro- vided critical and highly valuable comments and sug- gestions. Partial support was provided by Lutheran Brotherhood through a Medical Research Fund M.D./ Ph.D. Scholarship to the author. Literature cited Asper, E. D., L. H. Cornell, D. A. Duffield and N. Dimeo- Ediger. 1988. Marine mammals in zoos, aquaria, and marine zoological parks in North America: 1983 census re- port. Int. Zoo Yearb. 27:287-294. Bigg, M. A. 1969a. Clines in the pupping season of the harbour seal, Phoca vitulina. J. Fish. Res. Board. Can. 26:449-455. 1969b. The harbour seal in British Columbia. Bull. Fish. Res. Board Can. No. 172:1-33. 1973. Adaptations in the breeding of the harbour seal, Phoca vitulina. J. Reprod. Fert. (Suppl.) 19:131-142. 1984. Stimuli for parturition in northern fur seals (Callorhinus ursinus). J. Mammal. 65:333-336. 1986. Arrival of northern fur seals, Callorhinus ur- sinus, on St. Paul Island, Alaska. Fish. Bull. 84:383- 394. Boshier, D. P. 1981. Structural changes in the corpus luteum and en- dometrium of seals before implantation. J. Reprod. Fert. (Suppl.) 29:143-149. Boulva, J. 1975. Temporal variation in birth period and charac- teristics of newborn harbour seals. Rapp. P.-v. Reun. Cons. int. Explor. Mer 169:405-408. Boyd, I. L. 1984. The relationship between body condition and the timing of implantation in pregnant grey seals (Halichoerus grypus). J. Zool., Lond. 203:112-123. 1991. Environmental and physiological factors control- ling the reproductive cycles of pinnipeds. Can. J. Zool. 69:1135-1148. Braham, H. W. 1974. The California sea lion on islands off the coast of San Luis Obispo County, California. Calif. Fish Game 60:78-83. Bronson, F. H. 1985. Mammalian reproduction: an ecological perspec- tive. Biol. Reprod. 32:1-26. Brownell, R. L., R. L. Delong, and R. W. Schreiber. 1974. Pinniped populations at Islas de Guadelupe, San Benito, Cedros, and Natividad, Baja California, in 1968. J. Mammal. 55:469-472. Coulson, J. C. 1981. A study of the factors influencing the timing of birth in the grey seal Halichoerus grypus. J. Zool., Lond. 194:553-571. Craig, A. M. 1964. Histology of reproduction and the oestrus cycle in the female fur seal, Callorhinus ursinus. J. Fish. Res. Board. Can. 21:773-818. Daniel, J. C. Jr. 1981. Delayed implantation in the northern fur seal {Callorhinus ursinus). J. Reprod. Fert. (Suppl.) 29:35-50. Delong, R. L., W. G. Gilmartin and J. G. Simpson. 1973. Premature births in California sea lions: asso- ciation with high organochlorine pollutant residue lev- els. Science 181:1168-1170. Fisher, H. D. 1954. Delayed implantation in the harbour seal, Phoca vitulina L. Nature 173:879-880. Gentry, R. L. 1981. Northern fur seal Callorhinus ursinus (Linnaeus, 1758). In S. H. Ridgway, and R. J. Harrison (eds.), Handbook of marine mammals. Volume I: The wal- rus, sea lions, fur seals and sea otter, p. 143-160. Academic Press, London. Temte: Birth timing of captive Zalophus cahfornianus 717 King, J. E. 1983. Seal of the world, 2nd ed. Comstock Publ. As- sociates, Cornell Univ. Press, Ithaca, NY, 240 p. Kleinbaum, D. G., and L. L. Kupper. 1978. Applied regression analysis and other multivari- able methods. Duxbury Press, North Scituate, MA, 556 p. Le Boeuf, B. J., D. Aurioles, R. Condit, C. Fox, R. Gisiner, R. Romero and F. Sinsel. 1983. Size and distribution of the California sea lion population in Mexico. Proc. Cal. Acad. Sci. 43:77-85. Loughlin, T. R., D. J. Rugh and C. H. Fiscus. 1984. Northern sea lion distribution and abundance: 1956-80. J. Wildl. Manage. 48:729-740. Mate, B. 1975. Annual migrations of the sea lions Eumetopias jubatus and Zalophus californianus along the Oregon coast. Rapp. P.-v. Reun. Cons. int. Explor. Mer 169:455-461. 1979. California sea lion. In Mammals in the seas. Vol. II: Pinniped species summaries and report on sirenians, p. 5-8. Food and Agriculture Organiza- tion of the United Nations, Rome. Mate, B., and R. L. Gentry. 1979. Northern (Steller) sea lion. In Mammals in the seas. Vol. II: Pinniped species summaries and report on sirenians, p. 1-4. Food and Agriculture Organi- zation of the United Nations, Rome. Merrick, R. L. 1987. Behavioral and demographic characteristics of northern sea lion rookeries. M.S. thesis, Oregon State Univ., Corvallis, OR, 124 p. Odell, D. K. 1970. Premature pupping in the California sea lion: proceedings of the 7th annual biosonar and diving mammal conference, p. 185-190. Stanford Research Institute, Menlo Park, CA. 1975. Breeding biology of the California sea lion, Zalophus califoi-nianus. Rapp. P.-v. Reun. Cons. int. Explor. Mer 169:374-378. 1981. California sea lion Zalophus californianus (Les- son, 1928). In S. H. Ridgway and R. J. Harrison (eds.). Handbook of marine mammals. Volume I: The walrus, sea lions, fur seals and sea otter, p. 67- 97. Academic Press, London. Peterson, R. S., and G. A. Bartholomew. 1967. The natural history and behavior of the Cali- fornia sea lion. Am. Soc. Mammal., Special Publ. no. 1. Schusterman, R. J. 1981. Steller sea lion Eumetopias jubatus (Schreber, 1776). In S. H. Ridgway and R. J. Harrison (eds.), Handbook of marine mammals. Volume I: The wal- rus, sea lions, fur seals and sea otter, p. 119— 141. Academic Press, London. Schusterman, R., I. Kang, B. Andrews, C. McDonald, and D. Ballou. 1982. Reproduction in captive California sea lions: photoperiod and the annual timing of their birth season. In B. R. Mate (ed.). Marine mammal infor- mation, p. 42 (Abstract). Oregon State Univ. Sea Grant College Program, Corvallis, OR. Snedecor, G. W., and W. G. Cochran. 1980. Statistical methods, 7th ed. Iowa State Uni- versity Press, Ames, IA. Temte, J. L. 1985. Photoperiod and delayed implantation in the northern fur seal iCallorhinus ursinus). J. Reprod. Fert. 73:127-131. 1993. The Marine Mammal Inventory Report: inde- pendent verification of a captive marine mammal da- tabase. Mar. Mammal Sci. 9:95-98. Temte, J. L., M. A. Bigg, and O. Wiig. 1991. Clines revisited: the timing of pupping in the harbour seal (Phoca vitulina). J. Zool., Lond. 224:617-632. Trites, A. W. 1992. Reproductive synchrony and the estimation of mean date of birth from daily counts of northern fur seal pups. Mar. Mammal Sci. 8:44-56. AbStr3Ct.— A simple dynamic pool model (the "base model"), de- fined by a linear weight-at-age rela- tionship and a Gushing (convex power) stock-recruitment relation- ship, results in an explicit solution for the fishing mortality rate corre- sponding to maximum sustainable yield FMSY. This solution's sensitivity can be examined by comparing it to solutions derived under alternative model specifications. Four such modifications are considered here: 1) replacing the Gushing stock-recruit- ment equation with an equation of the Beverton-Holt form; 2) general- izing from linear growth to a flexible form of von Bertalanffy growth; 3) allowing the ages of recruitment to the fishery a and the mature stock am to diverge; and 4) allowing for a finite maximum age in the stock. Ex- act polynomial solutions for FMSY are derived for each specification (except the fourth), and the potential bias introduced by use of the base model is examined for each. In all cases, the solution to the base model is within 10% of the solution to the al- ternative model under a range of pa- rameter values. Variations on a simple dynamic pool model Grant G. Thompson Resource Ecology and Fisheries Management Division Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA 7600 Sand Point Way NE., Seattle. WA 98 1 I 5-0070 Manuscript accepted 24 June 1993. Fishery Bulletin 91:718-731 ( 1993). Some fishery models can be solved analytically (i.e., by mathematical manipulation of the equations), while others can be solved only numerically (i.e., by the brute force of computer simulation). One advantage of ana- lytic models is that the generality of their solutions is more straightfor- wardly addressed. For example, it is easy to show that the stock size as- sociated with maximum sustainable yield (MSY) in a Schaefer (1954) model is always one-half the pristine stock size; this property follows di- rectly from the assumption of logis- tic growth upon which the model is based. Such generality is more diffi- cult to demonstrate for simulation models, however. For instance, if a particular simulation showed that MSY was obtained at a stock size equal to one-half the pristine level, there would be no way to tell whether this result was general, except by re- peated trial and error with different values for the input parameters. Another example of an analytic so- lution is the one obtained by Thomp- son (1992) for the fishing mortality rate at MSY (FMsy) in his simple dy- namic pool model. Because this solu- tion is an analytic one, it is com- pletely general in the sense that it will hold whenever the underlying as- sumptions of the model hold, regard- less of parameter values. Of course, the underlying assumptions may not hold in a particular instance, which raises the question: How sensitive is the solution to those assumptions? The purpose of this paper is thus to examine the sensitivity of Thomp- son's (1992) solution relative to the underlying assumptions of that model. This will be accomplished by developing four reasonable modifica- tions to the base model proposed by Thompson and by examining the range of errors that might likely be encountered if the base model were employed in situations where one of the modifications would have been more appropriate. Review of the base model Thompson ( 1992, see also Jensen, 1973) defined a simple dynamic pool model as one that reflects the follow- ing assumptions: 1) cohort dynamics are of continuous-time form; 2) vital rates are constant with respect to time and age; 3) fish mature and re- cruit to the fishery continuously and at the same invariant ("knife-edge") age; 4) mean body weight at age is determined by age alone; 5) the stock (or population) consists of the pool of recruited individuals; 6) maximum age is infinite; 7) the stock is in an equilibrium state determined by the fishing mortality rate; and 8) recruit- ment is determined by stock biomass alone. Within the framework pro- vided by these assumptions, particu- lar models are distinguished by the forms assigned to the weight-at-age and stock-recruitment functions. As an example, Thompson (1992) developed a particular simple dy- namic pool model than can be solved explicitly for FMSY. In terms of biom- ass per recruit, the model is basically 718 Thompson. Variations on a simple dynamic pool model 719 the same as that of Hulme et al. (1947), where body weight is assumed to be a linear function of age (e.g., Richards, 1969): w(a) \ar-aoJ (1) where a represents age, ar is the age of recruitment, au is the age intercept, w(a) represents individual weight at age, and wr is the weight at recruitment. For a given recruitment level, stock biomass in the model is given by *". me^f) (2) where M is the instantaneous rate of natural mortal- ity, F'=FIM, B(F') is the equilibrium stock biomass obtained under a relative fishing mortality rate off, b(F',ar) is the equilibrium biomass at a=ar obtained under a relative fishing mortality rate of F\ and K" M(a-an) (3) which can be interpreted in this model as the pristine ratio of growth to recruitment (Thompson, 1992). Thompson (1992) extended the model described in Equation 2 by incorporating a stock-recruitment rela- tionship of the convex power form suggested by Gushing (1971): biF'.ar)=pBlF'F, (4) where p and q are constants and 01, [«fa-«frJH] For the case where K"=l, (12) Q' = (F'USY+lf FiSy+l-(F'MSY-l)K" 115) Another difference between this model and the base model is that here F'MSY reaches zero at Q'=1/(K"+1), whereas F'MSY in the base model does not reach zero until q=l. Still another difference is that here the up- per limit to the ratio B{F'VSY)/B(0) is 0.5, contrasted with 1/e in the base model. In both models, this limit is reached as F'MSY approaches zero. Finally, the behavior of this modification differs from that of the base model in that extinction is possible here, owing to the Beverton-Holt curve's finite slope at the origin. Extinction occurs here at FL Q' + ^(4K" + Q')Q' -1. (16) The relative fishing mortality rate described by Equa- tion 16 need not be unrealistically high. For example, it will be less than F'max whenever the following rela- tionship holds: Q'< 4A'" K" (17) FW=(2Q')W-1. (13) For the case where K"<1 2 r MSY ^pf^cos ||) COS' [(gj~) \Q'J J _1 • (14) The parameter Q' functions inversely to q in the sense that Equations 11-14 reduce to Equation 7 as Q' approaches infinity (the F'„mx case), whereas Equa- tion 6 does so as q approaches zero. As Q' increases, F'MSY increases monotonically, whereas F'MSY decreases with increasing q in the base model. As with the base model, Equations 11-14 contain F'msy = 1 as a special case. Here this is obtained when Q' = 4 or when K" approaches infinity. This contrasts somewhat with the base model, where keeping F'MSY at a constant value of 1.0 required an inverse relation- ship between q and K". However, it should be pointed out that F'MSY= 1 is a very special case in the Beverton- Holt form of the model, since this turns out to be the only constant value of F'MSY that does not imply some sort of relationship between Q' and K". In fact, a di- rect relationship between Q' and K" is required for all constant values ofF'MSY > 1, as described below: Bias resulting from the assumption of Cushing recruitment Assuming that the stock-recruitment relationship follows the Cushing form when it actually follows a Beverton- Holt form can lead to a biased estimate of FMSY. To com- pare stock-recruitment curves, Kimura (1988) observed that a two-parameter function can be defined by any two points on the curve. In his example, Kimura used hypothetical stock-recruitment "observations" at the pris- tine biomass level and at one-half the pristine biomass level. Kimura conjectured that recruitment might be reduced to about 90% of the pristine level when biom- ass has been reduced by 50% relative to its own pris- tine level, a suggestion which has been endorsed by others (e.g., Clark, 1991). Given the other parameters used in his example, Kimura found that the FMSY value under a Beverton-Holt stock-recruitment relationship was much less than the value under a Cushing rela- tionship fit to the same two stock-recruitment points. However, there is no reason to believe a priori that a Cushing relationship is necessarily less conservative than a Beverton-Holt relationship in terms of its asso- ciated FMSY value. Note that Equation 6 can be solved explicitly for q as a function of K" and F'MSY as follows: K"+1-(K'-1)F'MSY (K" + 1+F'MSYWMSY+1) (18) Thompson Variations on a simple dynamic pool model 721 Substituting Equation 12, 13, or 14 into Equation 18 thus gives the q value that sets FMSY under a Gushing stock-recruitment relationship equal to FMSY under a Beverton-Holt relationship. Assuming that both stock-recruitment curves are parametrized to pass through the same pristine stock-recruitment point (B(0),6(0.ar)), this q value implies a second intersec- tion at some lower biomass level. It turns out that this lower level is always less than about 20% of B(0) (Fig. 1A) and greater than about 20% of b(0,ar) (Fig. IB). In other words, Kimura's (1988) placement of a lower intersection at 50% of B(0) would always cause the Cushing model to overestimate FMSyi Placing the lower intersection at a biomass level less than 20% of B(0), however, might result in either an over- or under- estimate. For example, one rule of thumb (Clark, 1991) holds that FUI (the fishing mortality rate that reduces the slope of the yield-per-recruit curve to one-tenth of the slope at the origin), Fas„ (the fishing mortality rate that reduces the level of spawning biomass per recruit to 35% of the pristine level) and M should be approxi- mately equal. In the base model, this rule of thumb holds exactly at if '=1.5 (Thompson, in press). In the base model with Beverton-Holt recruitment, then, FMSY=F01=F35r=M at Q'=4 and K"'=1.5. These param- eters imply a stock-recruitment curve in which recruit- ment is reduced from b(0,ar) by exactly 1/11 when bio- mass is reduced to 50% of B(0), and in which recruitment is reduced from b(0,ar) by exactly 50% when biomass is reduced to 1/11 of B(0). In the base model with Cushing recruitment, on the other hand, FSIS^F0 t=F35%=M at q=2/7 and K = 1.5, im- plying a stock-recruitment curve in which recruitment is reduced from b(0,ar) by about 18% when biomass is reduced to 50% of 5(0), and in which recruitment is reduced from 6(0,ar) by about 50% when biomass is reduced to 1/11 of S(0). Thus, in the "rule of thumb" case, the form of the stock-recruitment curve (Cushing or Beverton-Holt) has virtually no impact on the re- sulting estimate of FMSY so long as the curve passes through (S(0),6(0,or)) and (B(0)/ll,rj(0,a,.)/2). More generally, to cause Cushing and Beverton-Holt curves to intersect at (S(0),6(0,a,)) and at some frac- tion p of 5(0), set Alternatively, the lower intersection can also be de- fined in terms of relative recruitment (as opposed to relative biomass). To cause Cushing and Beverton-Holt curves to intersect at (B(0),6(0,ar)) and at some frac- tion Hot 6(0,ar), set 9 = 1- ln( 9) \n(0) -\n[(l - B)(K" + 1)Q'+ 0] (21) andp as in Equation 20. Biases resulting from placement of the lower inter- section at 50% and 10% of S(0) are compared in Fig- ure 2, A and B, respectively. Note that use of the 50% value (as in Kimura's [1988] example) causes large and uniformly positive biases in the base model's esti- mate ofFMSY. On the other hand, use of the 10% value constrains bias to the +/- 10% range over a large por- tion of parameter space. Biases resulting from placement of the lower inter- section at 90% and 50% of b(0,ar) are compared in Figure 2, C and D, respectively (the 50% reference point has been suggested by Mace1 and Myers et al.2). As with the relative biomass reference level, use of Kimura's (1988) relative recruitment reference level (90%) causes large and uniformly positive biases in the base model's estimate of FMSY over a large portion of parameter space. On the other hand, use of the 50% value constrains bias to the +/- 10%' range over a siz- able region. As Figure 2 illustrates, then, there is reason to be- lieve that the form of the stock-recruitment curve (Cushing or Beverton-Holt) may not be particularly important in terms of the resulting estimate of FMSY so long as the candidate curves intersect at a fairly low level. In other words, fishery managers need not al- ways view an estimate of FMSY as being critically de- pendent on the form of the stock-recruitment curve. A general growth function The linear growth function used by Thompson (1992) may be viewed as a special case of the following more ln(p[(^"+l)Q'-l] + D-lnlQ'l-lni/T+l) q = l - (19) and ln(p) M wQ'-h (^TlX i -i (20) 'Mace, P. M. 1993. Relationships between common biological refer- ence points used as thresholds and targets of fisheries management strategies. Dep. Commer., NOAA. Natl. Mar. Fish. Serv., 1335 East- West Highway. Silver Spring. MD 20910. -Myers, R. A., A. A. Rosenberg. P. Mace, N. Barrowman, and V. Restrepo. 1993. In search of thresholds for recruitment overfishing. Dep. Fisheries and Oceans. St. John's. New Foundland. Unpubl. manuscr., 14 p. 722 Fishery Bulletin 91(4), 1993 0.25 E o -O a o c o o a. o 0.10 2 3 4 5 6 7 Scaled recruitment parameter 0' 2 3 4 5 6 7 8 Scaled recruitment parameter Q' Figure 1 Location of the zero-bias intersection of the Gushing and Beverton- Holt stock-recruitment curves. Horizontal dashed lines represent limiting values, obtained in the limit as FMSY goes to zero. Proceed- ing from left to right along a given horizontal dashed line, the curves which intersect the line correspond to K" values of 2.0, 1.5, 1.0, and 0.5, respectively. Given values of the composite parameter K" and the scaled Beverton-Holt recruitment parameter Q'. fixing the lower intersection of the recruitment curves at the point identified by the appropriate locus in this figure causes the base model to give the same value for F[,SY as the Beverton-Holt modification. iA) Zero-bias intersection defined in terms of relative stock biomass. Points below and to the left of the curves result in negative bias, while points above and to the right of the curves result in positive bias. (Bl Zero- bias intersection defined in terms of relative recruitment. Points below and to the right of the curves result in negative bias, while points above and to the left of the curves result in positive bias. general form: w(a i (\ _g-«(o-o )-K K V< (22) where K is Brody's growth coefficient, K' = KIM, and n is a positive integer. In different parametrizations. Equation 22 corresponds to (or is a special case of) a number of general growth functions, including those of Richards (1959, see also Fletcher, 1975), Savageau (1980), and Schnute (1981). Schnute described his parametrization of Equation 22 as "generalized von Bertalanffy growth" (although he did not restrict n to integer values). When n=3. Equation 22 corresponds to the common ("specialized") von Bertalanffy (1938) curve, and, when n = l, the "monomolecular" curve of Putter ( 1 920 ) and Brody ( 1928 ) is obtained. In the limit as K approaches zero, Equation 22 gives an nth-degree polynomial in age: w(a) = wr(u-u"\ \ar-a0J (23) which has been used to describe growth (though not always in weight) by Mendelsohn (1963), Dethlefsen et al. (1968), Knight (1968), Rafail (1972), Roff (1980), Geoghegan and Chittenden (1982), Standard and Chittenden (1984), and Chen et al. (1992). Equation 1 thus represents the special case of Equation 22 where K ap- proaches zero and n=l. Polynomial solution The polynomial solution for this model is parti- tioned into two cases (K=0 and K>Q) and derived in the Appendix. When K=0, the solution for F,'„„, can be written as a polynomial of degree n, and the solution for F'MSY can be written as a polynomial of degree n+\. When K>0, the solution for F'max can be writ- ten as a polynomial of degree 2«, and the solu- tion for F'MSY can be written as a polynomial of degree 2n + l. As with the base model, the solu- tion in either case indicates that maintaining an F'MSY value of 1.0 requires an inverse relation- ship between q and K" (which, as in the base model, can be written explicitly). Likewise, the upper limit to the ratio B(F|/sv)/B(0) is the same as in the base model ( lie) in both cases. The polynomial solution can be manipulated easily to show how it varies across the range of possible K', q, and K" values. For example, sev- eral limiting values of F'„mx and F'MSY are shown in Table 1. Bias resulting from the assumption of linear growth Assuming that growth is linear when it actually fol- lows a generalized von Bertalanffy form can lead to a Thompson Variations on a simple dynamic pool model 723 Composite parameter K" A. b-0.5 ■-, b-0.4 ---.. b-0.3 Composite parameter K" b— 0.1 b-0\ b-*0.1 ( 112 3 4 6 6 7 Scaled recruitment parameter Q' Composite parameter K" J12345678 Scaled recruitment parameter Q' Composite parameter K" B b-0.5" b-0.4 b-0.3 b"0 1 \b-0 b— 0.1 012345678 012346678 Scaled recruitment parameter 0' Scaled recruitment parameter 0' Figure 2 Isobias loci obtained when the base model is used to approximate the Beverton-Holt modification under four calibration methods. (A) Isobias loci obtained when the lower intersection of the Gushing and Beverton-Holt stock-recruitment curves is fixed at 50% of pristine biomass. Loci corresponding to biases in F'usy of 30% (6=0.3), 40% (6=0.4), and 50% (6=0. 5i are shown. (Bi Isobias loci obtained when the lower intersection of the Gushing and Beverton-Holt stock-recruitment curves is fixed at 10% of pristine biomass. Loci corresponding to biases in F'MSt of -10% (6=-0.1), 0% (6=0), and +10% (6=+0.1) are shown. (Ci Isobias loci obtained when the lower intersection of the Gushing and Beverton-Holt stock-recruitment curves is fixed at 90% of pristine recruitment. Loci corresponding to biases in F'MSr of 30% (6=0.3), 40% (6=0.4), and 50% (6=0.5) are shown. (D) Isobias loci obtained when the lower intersection of the Cushing and Beverton-Holt stock-recruitment curves is fixed at 50% of pristine recruitment. Loci corresponding to biases in F'Msr of + 10% (6=+0.1i, 0% (6=01, and -10% (6=-0.1i are shown. biased estimate of FMSY- One way to compare the two types of curve is to require that they intersect at ivr and that they imply the same pristine biomass-per- recruit level. In the base model, stock biomass per recruit is obtained by multiplying Equation 2 through by wr/b(F',ar). When Equation 22 or 23 is used to rep- resent growth, stock biomass per recruit is given by Appendix Equation 5 or Appendix Equation 11. When growth curves are forced to intersect at wr and pris- tine biomass per recruit levels are equated, the follow- ing parametrization is defined: zr = u (-ini)ekK K ):K' + 1 1. (24i where (p is the binomial coefficient (Appendix) and K" is the estimated value of K" used to define the linear growth relationship (assuming that M and ar are the same under both weight-at-age relationships, K" is distinguished from K" by the fact that the age intercepts of the two curves differ — Equation 3). Substituting Equation 24 for K" in Equation 6 gives F'MSY in the base model when the linear growth func- tion is fit in the manner described above. This F'MSY value can be either higher or lower than the value given by the solution derived in the Appendix. For the case where n=3. Figure 3 shows the range of +/- l(Kr bias for four values of K", along with the loci of zero bias. Note that at low values of K" (e.g., 0.5), the two 724 Fishery Bulletin 91(4), i 993 Table 1 Limits on F'„ta, and F'its) under generalized von Bertalanffy growth. A" = 0 A' Limits on F'mal /T->~ F'=± FL Limits on F'MI A" = 0 FL K" = lim ( K \ _ I K" k -o\ln(/;A" + 1)/ n 1-g n + q 1-9 ^iisv = 1-g ( i+k; )+k;1+2K';+i] = o. The solution for F'ma:i in this model is the same as in the base model (Equation 7). Equation 31 contains F'MSY = 1 as a special case, obtained when the following relationship holds: 9 = 1 IK} + 2){K", + 1) Aa[K"f) + (K"f + 2f (32) Just as the base model required an inverse relation- ship between q and K" in order for F'MSY to equal 1.0 (Equation 8), this model requires an inverse relation- ship between q and K"f. Likewise, the upper limit to the ratio B(FUST)/B(0) is the same as in the base model (1/e). Age at maturity exceeds age at recruitment Another possible modification is to allow a,„ to exceed ar This requires rewriting Equation 27 as follows: biF',a,) = b(F '-(§) o + F1/1 fe-a (33) The previous expressions for recruitment (Equation 25) and equilibrium fishable biomass (Equation 26) can be applied without modification. However, because the entire mature stock is subject to both fishing and natural mortality, the previous expression for equilib- rium mature biomass (Equation 28) is simplified to BJF') _/b(F\a„,r\A+K"ni+ F\ ' \ M ) \ (1+FT ) ' (34) 726 Fishery Bulletin 91(4), 1993 Solving Equations 25, 26, 33, and 34 simultaneously gives i (35) The fact that F' appears in the exponent in Equation 35 complicates the solution for F'MSY somewhat, increas- ing the degree of the polynomial solution to four: \(K"fK"m + 2K';+2K';„ + 3) &-j£ta + (2K:, + 3)q+ K}-l]F„s,? + [(K; + UK"„, + l)(^r-^r) 9 + VS"f + 3Wm + 1 K"fK"m - K"m - 2]f;;sv - (K"f + 1 )(«-; + 1X1 - q) = (36) k/ + 0. The solution for F'nmx is the same as in the base model (Equation 7). Equation 36 contains F'„SY=1 as a special case, ob- tained when the following relationship holds: <7 = K" + 2 iK"t+2 )(K';„+2 ) tj^- ^-\+K"liK"„,+ 1 )+3K"m+4 (37) As with the base model (Equation 8), the above ex- pression implies an inverse relationship between q and K"f. Likewise, the upper limit to the ratio B(Fl1SY)IB(Q) remains the same ( lie). Bias resulting from the assumption of af=a When at>am, Equation 6 tends to underestimate F'MSY. Loci of -10% bias are shown in Figure 4A for four values of K"m. Parameter combinations above a par- ticular curve and below the horizontal line K"f = K"m result in an F'MSY estimate that is within 10% of the value given by Equation 31. Note that the base model's solution is fairly sensitive to K"f when K"m is low. For example, when K"„ =0.5, almost any value of K"f\), the base model's solution is less sensitive. The results for the case where all), the base model's solution is less sensitive. Finite maximum age As defined by Thompson (1992), all simple dynamic pool models exhibit mortality and growth rates which are independent of age (above the age of recruitment). This implies that there is no maximum age. However, in more complicated dynamic pool models, it is com- mon to specify a maximum age above which all re- maining fish die in knife-edge fashion. As noted by Fletcher (1987), misspecification of maximum age can introduce significant bias into some models. When the base model is modified so as to exhibit a finite maximum age (a,„,„), Equation 5 will tend to overesti- mate true equilibrium stock biomass, which can be written as i-i (38) where K"ma=ll[M(amax - a,,)], after Equation 3. The dif- ference inside the exterior parentheses in Equation 38 is proportional to the difference between two calcula- tions of stock biomass per recruit in a population with infinite maximum age, where the first calculation be- gins the integral (over age) at age ar and the second begins at age ama. Subtracting the second term from the first adjusts for the assumption of a maximum age at a=a„mx. Because of the presence of F' in the exponential term in Equation 38, it is not possible to solve for F'MSY explicitly in this modification. Bias resulting from the assumption of infinite maximum age Note that as K",mx becomes small (e.g., as c,„„, becomes large), the proportion surviving to the maximum age (the exponential term in Equation 38) goes to zero and Thompson Variations on a simple dynamic pool model 727 2.0 * A Composite parameter > bi b bi A:- 2.0 — ____^^ r^O^s . ___________ *'m" ' ° ■ K'mm°5 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 09 1.0 Cushing recruitment parameter q 4.0 ~£ ■ >site parameter to W b b // « /C'1.0 >C-1.5 Q 1.0- fi-1.0 O 0 Km'O.B 00 0.1 02 03 0.4 0.5 0.6 07 08 09 1.0 Cashing recruitment parameter q Figure 4 Isobias loci obtained when the base model is used to approximate the modification in which the ages of recruitment to the fishery and to the mature stock diverge. I A) Loci of -10% bias in F'USY obtained when the age of recruitment to the fishery is greater than the age of recruitment to the mature stock. (B) Loci of +10% bias in F'Msy obtained when the age of recruitment to the fishery is less than the age of recruitment to the mature stock. Equation 38 collapses to Equation 5. However, at any positive value of K\ ",„,, Equation 5 will tend to overesti- mate the true value of B(F') to some extent. Conversely, Equation 6 will tend to underestimate the true value of F't/gy. Figure 5 shows loci of -10% bias in Equation 6's estimate of F'MSY. Points above and to the left of the curves result in a bias of less than 10% (absolute value). For example, a stock with Af=0.2, o„=-l, and am„=24 would have a K",wx value of 0.2. For such a stock, Equa- tion 6's estimate of F'MSY would be within 10% of the correct value for any value of K ">0.5 so long as q was less than about 0.53. Conclusion Four modifications to the base model presented by Thompson (1992) have been considered (Beverton-Holt recruitment, generalized von Bertalanffy growth, divergent ages of recruitment and maturity, and finite maximum age). The first three modifications all increase the degree of the polynomial solution for FMSY (Table 2), while the fourth modification renders a polynomial solu- tion impossible. In order to make the Cushing stock-recruit- ment form of the model comparable to the Beverton-Holt form, an acceptable approximation can often be made by equating the pristine stock- recruitment points and placing the other (non- zero) intersection of the stock-recruitment curves at a fairly low level (e.g., at 10% of the pristine biomass level or at 50% of the pristine recruit- ment level). In order to make the linear growth form of the model comparable to the generalized von Bertalanffy growth form, an acceptable approxi- mation can often be made by equating the weights at recruitment and the pristine biomass-per- recruit ratios. When the ages of recruitment to the fishery and to the mature stock diverge sufficiently or when the maximum age in the stock is sufficiently low, the base model can produce a significantly biased estimate of FMSY. Except for the case in which the age of recruitment to the fishery pre- cedes the age of recruitment to the mature stock, though, it is helpful to note that the base model always errs on the conservative side. In conclusion, it appears that simple models (at least the base model considered here) may often perform adequately even when the true dy- namics of the system follow more complicated formulae. This tends to confirm the results of studies by Silliman (1971), Roff (1983), and Ludwig and Walters ( 1985), who also found that simple models could perform at least as well as more complex versions in a variety of situations. Acknowledgments I would like to thank James Balsiger, Nicholas Bax, Roderick Hobbs, Daniel Kimura, and Richard Methot of the Alaska Fisheries Science Center for reviewing all or portions of this paper in various stages of devel- opment. Comments provided by Ian Fletcher of the Great Salt Bay Experimental Station were especially 728 Fishery Bulletin 91(4), 1993 0.2 0.3 0 4 0 5 0 6 0 7 0 1 Cushlng recruitment exponent q Figure 5 Loci of -10% bias in F'usr obtained when the base model is used to approximate the modification in which the maximum age in the population is finite. Points above and to the left of the curves corre- spond to a bias of less than 10% (absolute value I, while points below and to the right of the curves correspond to a bias of greater than 10% (absolute value). helpful. Several anonymous reviewers also supplied constructive suggestions. Literature cited Alverson, D. L., and W. T. Pereyra. 1969. Demersal fish explorations in the northeastern Pacific Ocean — an evaluation of exploratory fishing methods and analytical approaches to stock size and yield forecasts. J. Fish. Res. Board Can. 26:1985- 2001 von Bertalanffy, L. 1938. A quantitative theory of organic growth (Inquir- ies on growth laws. II. Human Biol. 10:181-213. Beverton, R. J. H., and S. J. Holt. 1957. On the dynamics of exploited fish populations. U.K. Min. Agric. Fish., Fish. Invest. (Ser. 2) 19, 533 P- Brody, S. 1928. An analysis of the course of growth and sen- escence. In W. J. Robbins, S. Brody, A. G. Hogan, C. M. Jackson, and C. W. Greene (eds.), Growth, p. 31-64. Yale Univ. Press, New Haven, CT. Chen, Y., D. A. Jackson, and H. H. Harvey. 1992. A comparison of von Bertalanffy and polynomial functions in modelling fish growth data. Can. J. Fish. Aquat. Sci. 49:1228-1235. Clark, W. G. 1991. Groundfish exploitation rates based on life his- tory parameters. Can. J. Fish. Aquat. Sci. 48:734- 750. Cushing, D. H. 1971. The dependence of recruitment on parent stock in different groups of fishes. J. Cons. int. Explor. Mer 33:340-362. Table 2 Degree of polynomial solutions for F'Msy under various modifications. Modification Degree Beverton-Holt stock-recruitment Generalized von Bertalanffy growth A'=0 K>0 rc+1 2n+l Divergent ages of recruitment and maturity a„, 3 "i«i,„ 4 Dethlefsen, L. A., J. M. S. Prewitt, and M. L. Mendelsohn. 1968. Analysis of tumor growth curves. J.Natl. Cancer Inst. 40:389-405. Fletcher, R. I. 1975. A general solution for the complete Richards function. Math. Biosci. 27:349-360. 1987. Three optimization problems of year-class analysis. J. Cons. int. Explor. Mer 43:169-176. Geoghegan, P., and M. E. Chittenden Jr. 1982. Reproduction, movements, and population dy- namics of the congspine, Stenotomus caprinus. Fish. Bull. 80:523-540. Hulme, H. R., R. J. H. Beverton, and S. J. Holt. 1947. Population studies in fisheries biology. Nature 159:714-715. Jensen, A. L. 1973. Relation between simple dynamic pool and sur- plus production models for yield from a fishery. J. Fish. Res. Board Can. 30:998-1002. Kimura, D. K. 1988. Stock-recruitment curves as used in the stock- reduction analysis model. J. Cons. int. Explor. Mer 44:253-258. Knight, W. 1968. Asymptotic growth: an example of nonsense dis- guised as mathematics. J. Fish. Res. Board Can. 25:1303-1307. Ludwig, D., and C. J. Walters. 1985. Are age-structured models appropriate for catch- effort data? Can. J. Fish. Aquat. Sci. 42:1066-1072. Mendelsohn, M. L. 1963. Cell proliferation and tumor growth. In L. F Lamerton and R. J. M. Fry (eds.), Cell proliferation, p. 190-210. Davis, Philadelphia. Putter, A. 1920. Studien uber physiologische Ahnlichkeit. VI. Wachstumsahnlichkeiten. Pflugers Arch. Gesamte Physiol. Menschen Tiere 180:298-340. Rafail, S. Z. 1972. Fitting a parabola to growth data of fishes and some application to fisheries. Mar. Biol. 15:255- 264. Thompson Variations on a simple dynamic pool model 729 Richards, F. J. 1959. A flexible growth function for empirical use. J, Exp. Bot. 10:290-300. 1969. The quantitative analysis of growth. In F. C. Steward (ed.l. Plant physiology: a treatise. Vol. 5a: Analysis of growth: behavior of plants and their or- gans, p. 3-76. Academic Press, New York. Riordan. J. 1980. An introduction to combinatorial analysis. Princeton Univ. Press, Princeton, NJ, 244 p. Roff, D. A. 1980. A motion for the retirement of the von Berta- lanffy function. Can. J. Fish. Aquat. Sci. 37:127- 129. 1983. Analysis of catch/effort data: A comparison of three methods. Can. J. Fish. Aquat. Sci. 40:1496- 1506. Savageau, M. A. 1980. Growth equations: A general equation and a sur- vey of special cases. Math. Biosci. 48:267-278. Schaefer, M. B. 1954. Some aspects of the dynamics of populations im- portant to the management of the commercial ma- rine fisheries. Bull. Inter-Am. Trop. Tuna Comm. 1:27-56. Schnute, J. 1981. A versatile growth model with statistically stable parameters. Can. J. Fish. Aquat. Sci. 38:1128-1140. Silliman, R. P. 1971. Advantages and limitations of "simple" fishery models in light of laboratory experiments. J. Fish. Res. Board Can. 28:1211-1214. Standard, G. W., and M. E. Chittenden Jr. 1984. Reproduction, movements, and population dy- namics of the banded drum, Larimus fasciatus. Fish. Bull. 82:337-363. Thompson, G. G. 1992. Management advice from a simple dynamic pool model. Fish. Bull. 76:377-388. In press. A proposal for a threshold stock size and maximum fishing mortality rate. In S. J. Smith. J. J. Hunt, and D. Rivard (eds.). Risk evaluation and biological reference points for fisheries man- agement. Can. Spec. Publ. Fish. Aquat. Sci. 120. Appendix Some combinatoric terms In order to incorporate Equations 22 or 23 into the model, it is helpful to define a few concepts taken from combinatorial theory (the notation used here follows Riordan |1980]i. First, the number of permutations of n objects taken k at a time is given by with {n )k defined as zero for k<0 or k>n. The number of combinations (i.e., permutations with- out regard to order) of /; objects taken k at a time is given by the binomial coefficient 0 kUn-k)! (A2) with ( I) defined as zero for k<0 or k>n. The number of ways in which an n -element set can be partitioned into k subsets is given by Stirling num- bers of the second kind, written S(iuk) ■(s)^"(a)*- (A3) The coefficients of the polynomial expansion of Lr)„ are given by Stirling numbers of the first kind, written £* //i-l+Ax /2n-k \ sin, k)= L (-D1/ 11 1 S(n-k+\, A). iA4 \n-k+kJ\n-k-\J Polynomial solution for a generalized growth function K=0 Beginning with the simpler case where K=0 (Equation 23), stock biomass per recruit can be written BPR(F) = [ wr(a~arYe-M,i + F"°-°rda J° \ar-anJ (A5) (wr\ y (n)l!K"k In general, stock biomass in any simple dynamic pool model with Cushing recruitment can be written as the following function of biomass per recruit: m,)=(EBE^ (A6) Substituting Appendix Equation 5 (A5) into Appen- dix Equation 6 (A6), multiplying throttgh by MF, and differentiating gives the following expression: dY(F) dF \a-q)(l+F)"*2) ( X [(/! )kK"k a-kF')a+F' >" - *] - q X [(n)kK"k (1 + F')« + '-"]). (A7i (n)k [n-k)\ (All 730 Fishery Bulletin 91|4). 1993 The solution to Equation A7 can be written as a poly- nomial of degree n+\ as follows: n i n (A8) X(l[t%k(::\)](n)kK">- i=0 \ 4=0 [ ltt% )kK"^qyMS): - qFMS/ ♦* = 0. In the special case where <7=0, Equation A8 gives the following polynomial solution for F'max: X f I [ (":%k(::\)(n)kK"k])F;nJ=0. (A9) i=0 \ *=0 / Equation A7 can be solved explicitly for q. The locus at which F'MSY-\ is given by [2\ JMf)'] (A10) Note that Equations 6-8 constitute the special cases of Equations A8-A10 where n=l. K>0 When growth follows the form of Equation 22, stock biomass per recruit can be written , ~ /-l_p-Kla-a l-KIK"\n BPR(F')=l wr [ 1_e-KIK J e-Ma+FHa-"Jda (All) / wr \/y (-lHl)e-l"UK\ \M( l-*-K\>0\*=0 / / Finally, the solution to Equation A14 can be written as a polynomial of degree 2n + l as follows: i [ X f X (-1 WjX*y4 w. * + r, , *»e ^ /A' ) *'J - 1=0 L 7=0 V=0 / fXfi<-l>*(;>(/^,,,-,, + ?,.,*+ (A22) \J,=0 \*=0 Thompson Variations on a simple dynamic pool model 73 1 In the special case where <7=0, Equation A22 gives Equation A14 can be solved explicitly for q. The lo- the following polynomial solution for F'max: cus at which F'MSi =1 is given by £r£|i(-l «SX*% ,-, *+y,, ^kik)k']f;„j=0. ( A23 ) I [(-1 *®e*r'r(ll (miT+2)2W+l)l ,=oL>o\*=o /J q= * °L ^""* i J, (A24) Z|"(-l)*(5)er*r/W'll (mK'+2)2\kK+2)~\ *=o I Abstract. -Patterns of otolith microstructure and microchemistry (Sr/Ca ratios) are described in lar- val and juvenile Dover sole and re- lated to developmental events and habitat. The initiation of metamor- phosis is associated with a transi- tion from clear (protein rich) to opaque (protein poor) material and formation of the first accessory pri- mordium. Settlement is associated with the point at which growth from accessory primordia completely en- closes growth from the central pri- mordium and Sr/Ca ratio is mini- mal. No discrete otolith landmarks coincide with termination of meta- morphosis. Contrary to other flatfish that have a more abrupt metamorphosis, accessory primordia do not form un- til the left eye traverses the mid- dorsal ridge of the cranium in Do- ver sole, an event that may occur months after the eye first reaches that position. A period of 70 days or more can separate the first- and last- formed accessory primordia in Do- ver sole, suggesting that all acces- sory primordia do not form in response to a single event. A relationship between increments and days is validated for Stage 3-5 metamorphic and post-metamorphic stages. Duration of Stages 3 and 4 was in close agreement with predic- tions based on seasonal distribu- tions. Duration of Stages 1 and 2, as determined by unvalidated incre- ment counts, was about half as long as determined from seasonal distri- butions. Relationships between otolith microstructure, microchemistry, and early life history events in Dover sole, M/crostomus pac/ficus Christopher L. Toole Department of Fisheries and Wildlife 1 04 Nash Hall, Oregon State University Corvallis, OR 9733 1 -3803 Present address: Environmental and Technical Services Division National Marine Fisheries Service 91 I NE I 1th Ave., Suite 620, Portland, OR 97232 Douglas F. Markle Phillip M. Harris Department of Fisheries and Wildlife 1 04 Nash Hall, Oregon State University Corvallis, OR 9733 1 -3803 Manuscript accepted 18 May 199:!. Fishery Bulletin 91:732-753 1 1993 1. Changes in otolith microstructure often correlate, and are presumed syn- chronous, with morphological and be- havioral changes in larval and juve- nile fish (Brothers and McFarland, 1981; Victor. 1982; Fowler, 1989; Ozawa and Penaflor, 1990; Gartner, 1991). The formation of secondary growth centers (accessory primordia) on flatfish otoliths is associated with metamorphosis in starry flounder, Platichthys stellatus (Campana, 1984a); plaice, Pleuronectes platessa (Alhossaini et al., 1989, Karakiri et al., 1989); California halibut, Para- lichthys californicus (Kramer, 1991); and winter flounder, Pseudopleu- ronectes americanus (Sogard, 1991). Daily growth increments from acces- sory primordia are wider and are more likely composed of subdaily in- crements than those from the cen- tral primordium (Campana, 1984a; Karakiri et al., 1989; but also see Alhossaini et al., 1989). Metamorphosis in flatfish includes a number of morphological and be- havioral changes, most notably, eye migration and settlement from the water column to the sea floor. In most flatfish species, metamorphosis ap- pears to be rapid; eye migration and settlement occur nearly simulta- neously over a period of about 1 week (starry flounder, winter flounder, California halibut) to 3 weeks (pla- ice) (Ryland, 1966; Policansky, 1982; Campana, 1984a; Chambers and Leggett, 1987; Gadomski et al, 1992). Although timing of accessory primor- dium formation is not discussed in these studies, examination of pub- lished photographs suggests that the oldest and most recent accessory pri- mordia are separated by only a few daily growth increments. In plaice, the earliest-formed accessory primor- dium (Alhossaini et al., 1989) or an unspecified accessory primordium (Karakiri et al., 1989) have been in- terpreted as settlement marks. In winter flounder, the earliest-formed accessory primordium has been in- terpreted as the point from which post-metamorphic age of an indi- vidual is determined (Sogard and Able, 1992). Counts and measure- ments of growth increments from accessory primordia to the otolith edge have been used to infer settle- 732 Toole et al Otolith microstructure. microchemistry and early life history of Microstoma pxificus 733 ment dates and post-settlement growth and mor- tality rates (Alhossaini et al., 1989; Karakiri et al., 1989). In Dover sole (Microstomas pacificus), a commer- cially important northeast Pacific Ocean flatfish, meta- morphosis differs from the rapid process described above. Markle et al. (1992) define five stages in Dover sole development. Stage 1 includes planktonic pre- metamorphic larvae from 6.1 to 58.5 mm standard length (SL). Eye migration begins when Stage- 1 lar- vae are 10-15 mm SL but is arrested about midway through the process (Pearcy et al., 1977; Markle et al., 1992). Thus, Dover sole are optically asymmetrical dur- ing most of their planktonic life. Stage 2 includes plank- tonic metamorphic larvae from 42.3 to 60.4 mm SL, in which the eye has migrated past the mid-dorsal ridge and a pronounced shrinkage in body depth has begun. Stage 3 includes transitional metamorphic larvae from 40.7 to 74.9 mm SL, which are found in both the plank- ton and benthos. Stage-3 larvae are characterized by attainment of the adult configuration for a number of morphological characters. Stage 4 includes predomi- nantly benthic, metamorphic larvae from 41.7 to 79.3 mm SL, which have formed a characteristic intes- tinal loop in the secondary body cavity. Stage 5 in- cludes post-metamorphic juveniles from 48.9 mm SL to sexual maturity, which occurs at lengths greater than approximately 250 mm total length ( Yoklavich and Pikitch, 1989; Hunter et al., 1992). Based on seasonal collections of staged larvae, Markle et al. (1992) esti- mated that duration of the metamorphic period (Stages 2, 3, and 4) is approximately 9 months. The protracted metamorphic period in Dover sole provides an opportunity to examine changes in otolith structure and chemistry associated with each stage of metamorphosis. It may also be possible to distinguish between otolith landmarks associated with develop- mental processes and those associated with settlement from the water column to the bottom, because these events are not as coincidental in Dover sole as in other flatfish species. The objectives of this study were to 1) describe microstructure and microchemistry of Dover sole otoliths collected before, during, and after metamorphosis; 2) identify structural and chemical landmarks representing otolith growth during impor- tant morphological and behavioral transitions; 3) vali- date periodicity of increment formation in otoliths of metamorphic Stage-3 and Stage-4 larvae and post- metamorphic Stage-5 juveniles; and 4) using the otolith landmarks determined in objective 2 and increment counts between those landmarks and the otolith edge, re-examine the chronology of metamorphic events de- scribed in Markle et al. < 1992). Methods Dover sole collections Otoliths of six Stage-1, 103 Stage-3, 82 Stage-4, and 220 Stage-5 Dover sole were examined for microstruc- ture and microchemical analysis. Specimens were ob- tained on various dates between 1987 and 1990 from bottom trawl collections off Oregon (Markle et al., 1992), opportunistic bottom trawl collections of com- mercial fishermen from Oregon and Washington, and midwater trawl collections off Oregon and central Cali- fornia1 (Appendix 1). The following measurements were taken to the near- est 0.1mm on all specimens: standard length, body depth at anus (BD1A), and distance from snout to pos- terior extent of intestine (SINT). Weight of pat-dried specimens was determined to the nearest 0.1 g. Otolith preparation and analysis for microstructure Sagittae were removed after the fish were measured. Otoliths not immediately prepared for analysis were stored dry in vials. Terminology of otolith morphology follows Campana and Neilson (1985) and Secor et al. (1991), modified to account for features in Dover sole otoliths (Fig. 1 and Results section below). Left and right otoliths of Stage 3-5 Dover sole differ noticeably in shape (Fig. 1 and Results section below) and were analyzed separately. The longest and shortest axes of sagittae from Stage-1 larvae were measured to the nearest 0.01mm prior to mounting on slides in either a toluene-based medium (Histoclad) or an acrylic glue (Super Glue). Otoliths were then ground in the sagittal plane with 600-grit paper until growth increments became apparent. Sagittae from Stage 3-5 Dover sole were cleaned in ethanol, air-dried, weighed to the nearest 0.01 mg, and measured along the anterior-posterior axis (maximum length) and along the dorsal-ventral axis (maximum width) to the nearest 0.01 mm (Table 1). The number of translucent rings (annuli) outside the clear central growth area (Fig. 1 and Results section below) was determined upon examination of whole otoliths under reflected light against a black background. Most otoliths were mounted on slides in a toluene-based medium (Histoclad) and the lateral face was ground in the 'Whipple, J. 1991. Progress in rockfish recruitment studies. U.S. Dep Commer NOAA. Natl. Mar. Fish. Serv.. Southwest Fish. Sci. Center. P.O. Box 271, La Jolla. CA 92038. Admin. Rep. T-91-01. 57 p. 734 Fishery Bulletin 91(4), 1993 DORSAL ACCESSORY PRIMORDIUM CLEAR CENTRAL AREA INNER BOUNDARY OF ENCLOSED PERIPHERAL AREA ANTERIOR ■SCALLOPING" ANTERIOR ACCESSORY A PRIMORDIUM OPAQUE CENTRAL AREA UNENCLOSED PERIPHERAL AREA VENTRAL ACCESSORY PRIMORDIUM / POST-ROSTRUM ANTIROSTRUM _- ROSTRUM Figure 1 Schematic illustrations of Dover sole, Microstomus pacificus, otolith structure and orientation, showing terminology used in text. lAl Stage-4 left sagitta, sagittal section, lateral face towards the viewer, anterior to left. The anterior cavity is confluent with the anterior accessory primordium. (Bl Orientation of left and right Stage-4 sagittae, anterior to right. Table 1 Number of otoliths from Stage 3-5 benthic Dover sole. Microstomia pacificus. examined in study for each measurement or enumeration. AP = accessory primordium. Stage 3 Stage 4 Stage 5 Total Right otoliths Whole otoliths Otolith length 81 53 215 349 Otolith width 81 53 215 349 Otolith weight 87 63 218 368 Number of annuli 87 63 209 359 Sagittal sections Diameter of clear central area 55 33 0 88 Number of AP 70 41 6 117 Increments since last AP (enclosed 1 44 32 3 79 Increments since last AP (not enc.) 13 0 0 13 Increments since enclosure 51 46 3 100 Increments between checks 28 18 0 46 Left and right paired otolith comparisons Whole Otoliths Otolith length 67 50 198 315 Otolith width 67 50 196 313 Otolith weight 74 57 204 335 Sagittal sections Number of AP 41 22 0 63 Increments since last AP 23 12 0 35 First vs. last countable AP 14 11 0 25 Increments since enclosure 10 6 0 16 length anterior and posterior to AP 54 39 0 93 Toole et al Otolith microstructure, microchemistry. and early life history of Microstomus pacificus 735 sagittal plane to the level of the central primordium. If resolution was not sufficient, the medial face was also ground to the central primordium. Otoliths from Dover sole >75 mm were ground and read progressively because all increments could not be observed in a single plane in sagittal section. In order to examine otolith morphology in other planes, a subsample of otoliths from 17 Stage 3-5 Do- ver sole were embedded in low viscosity Spurr blocks (Haake et al., 1982) and ground to a level even with the central primordium in frontal or transverse sec- tions. Ground sections were covered with a mounting medium and cover slip. Additional otolith measurements were obtained from subsamples of Stage-3 and Stage-4 larvae (Table 1) to determine growth and development of features associ- ated with metamorphic stages. The diameter of the clear central growth area (Fig. 1 and Results section below) of 88 right otoliths was measured to the near- est 0.02 mm on a black background at 50x with a dissecting scope and reflected light. Otoliths were then examined at 100, 400. and 1000 x with a compound microscope and video monitor. At 100 x we measured length of the otolith anterior and posterior to the cen- tral primordium to the nearest 0.01 mm in paired left and right otoliths from 93 larvae. The number and position of accessory primordia, whether or not central primordium growth was com- pletely enclosed by growth from accessory primordia, number of increments from accessory primordia to otolith edge, and number of increments from enclosure of central primordium growth to otolith edge were de- termined for all Stage-3 and Stage-4 larval otoliths in which the features were legible (Table 1). Similar in- formation was obtained from 6 Stage-5 juveniles. Most counts and determinations were based on right otoliths; however, a subsample of left otoliths was also exam- ined for paired comparisons (Table 1). Counts of all growth increments from the first count- able increment outside the central primordium to the otolith edge were made for five of six Stage- 1 larval otoliths and for four Stage-3 larval otoliths exhibiting a variety of patterns of accessory primordium forma- tion. Counts near the central primordium were made with 1000 x light microscopy. In all otoliths for which total increment counts were made and in approximately 20% of otoliths for which counts from accessory primordia were made, there were areas where accurate increment counts were not pos- sible. In these areas, the number of increments was interpolated by linear approximation (Methot, 1983; Butler, 1989), based on the average widths of 5-20 increments on the distal or, preferably, both the proxi- mal and distal sides of the uncountable area. Stage 3- 5 larval otoliths with >5% interpolated increments were Table 2 Rearing conditions of juvenile Dover sole, Microstomus pacificus, during the first validation experiment. Group 1 Group 2 Group 3 Collection depth 110m 150 m and 163 m 108 m Number injected 4 10 21 Number controls 13 0 0 Average weight1 5.4g 3.4g 4.5 g Aquarium volume 152 L 76 L 114 L Average density 0.036 g/L 0.045 g/L 0.039 g/L Temperature 8°C 10°C 12°C 'Average for entire experiment, based on weights at death. excluded from further analysis. We present the per- centage of interpolated increments with total increment estimates for the few Stage- 1 specimens examined. Validation of increment deposition rate Two groups of juvenile Dover sole were given inter- peritoneal injections with a solution of 0.01 g oxy- tetracycline hydrochloride (OTC)/mL distilled water at a dose of 0.1 g OTC/kg fish weight (Campana and Neilson, 1982) to produce fluorescent marks on their otoliths. Dover sole in the first experimental group were held at three constant temperatures to evaluate temperature effects on increment formation. Those in the second experimental group were exposed to identi- cal temperatures and were injected once in March and again in September to evaluate seasonal effects on in- crement formation. First experiment An initial group of 48 juvenile Do- ver sole was collected at four stations off Cape Foul- weather, Oregon, on 17 March 1990 (Table 2). Thirty- five fish were injected on that day and 13 served as uninjected controls to check for naturally occurring fluorescence. The fish were held in aquaria contain- ing artificial seawater (Instant Ocean) and fed to sa- tiation with Tiibifex worms once per day in the morn- ing. Fluorescent lights in the room were set to natural cycles, and aquaria were partially covered to reduce incoming light. Rearing conditions are summarized in Table 2. Fish were sacrificed 21 («=6), 26 (n=4) or 29 (ra=9) days following initial injection. The remaining 29 fish were given a second injection of OTC 36 days after capture and were sacrificed 5 (n=6), 8 (ra=10), or 12 (« = 13) days after the second injection (i.e., 41, 44, or 48 days after initial injection). 736 Fishery Bulletin 91(4). 1993 Second experiment A second group of 10 juvenile Dover sole ranging from 55.3 to 118.3 mm was col- lected on 20 March 1991 at 77 m off Cape Foulweather, Oregon, and injected the following day. The fish were held in a flow-through filtered seawater tank. Ambient nearshore oceanic water entering the tank averaged 10.7°C (range: 8.7-14.6°C) during the experiment. Fish were exposed to diffused natural light and fed as in the first experiment. Seven days after injection, four fish, 55.3-63.9 mm, were sacrificed. The remaining six fish were individually marked with fin clips on 27 April. They were re-injected on 11 September, 174 days after first injection, and sacrificed 19 days later. Lengths ranged from 88.2 to 152.2 mm. Otoliths from both groups of fish were prepared as described above and examined at 100-1000 x on a Zeiss microscope equipped with a IV Fl epifluorescence con- denser. The preparation was examined under full spec- trum light to locate areas with distinct growth incre- ments and the OTC band was then observed under ultraviolet light. At least two documentary 35-mm slides of each otolith were taken: one showing the fluorescent band as well as the otolith edge and one showing incre- ments under full-spectrum illumination. Each slide was projected and traced onto the same sheet of pa- per, resulting in a diagram that showed the location of the fluorescent band in relation to increments and other landmarks. The diagram served as documen- tation for increment counts as well as a guide for further examination with light microscopy. Photo- graphs and counts were made at one to four loca- tions on each otolith, usually at the rostrum, anti- rostrum, or post-rostrum (Fig. 1). Increment counts were generally made on right otoliths; left otoliths were used only if right otoliths were unavailable or illegible. All otoliths were read at least twice on sepa- rate dates by the same reader and the average was used in subsequent analyses. The range of estimates for each otolith and standard deviation of mean counts were also determined. Results were analyzed by regressing number of in- crements between the proximal edge of the innermost fluorescent band and edge of the otolith against num- ber of days since OTC injection. When a second fluorescent band was obvious and the area distal to the second band was illegible, counts were made be- tween bands. To determine if all experimental groups could be combined into one regression, multiple re- gressions, including "dummy variables" corresponding to experimental groups, were analyzed with partial F-tests (Neter et al, 1989:364-370). The slope of the final regression model was compared to a slope of 1.0 (Neter et al., 1989). Otolith preparation and analysis for microchemistry Otoliths from two Stage-3 larvae (51.3 and 61.9mm), three Stage-4 larvae (54.6, 57.1, and 65.7mm), and two Stage-5 juvenile Dover sole (88.6 and 172.2mm) were examined with a wavelength dispersive electron microprobe to determine if microchemical changes were associated with microstructural changes. Preparation and analytical techniques followed Toole and Nielsen (1992). Beam diameter was 5|.im, counting time for each element was 20 seconds, accelerating voltage was 15 kV, and current was 20 nA. Results Validation of increment deposition rate First experiment Otoliths from uninjected control fish exhibited bright yellow-green fluorescence around the edge of the otolith and weaker fluorescence associated with strong checks and scratches. However, fluorescent marks observed on injected fish differed markedly in appearance. Twenty-seven injected fish had at least one otolith with one or two distinct fluorescent bands located inside the otolith edge (Fig. 2, A and B). Fluo- rescent bands extended over two to six increments. Increments formed after capture were often less dis- tinct than those formed in nature (Fig. 3), and four of 27 injected fish with fluorescent bands (14.8%) had otoliths with such poorly defined increments that they were considered unreadable and were eliminated from further analysis. Both daily and subdaily increments (Campana and Neilson, 1985) were observed. For those otoliths in which both subdaily and larger increments could be counted in the same area, the ratio of small to large increments averaged 2.78 (rc = 10 fish, range 1.36-3.48). In most cases, larger growth increments could easily be distinguished from subdaily increments by adjust- ing the focus. However, in otoliths of two fish (7.4% of those with fluorescent bands), only subdaily increments could be observed. These fish were also eliminated from further analysis. Second experiment Four Dover sole sacrificed 7 days after first injection had not fed. The remaining six fish began feeding approximately two weeks after the first OTC injection. Two of these fish grew rapidly through- out the experiment, whereas the other four fish ceased growing after about 20 weeks (Fig. 4). Two of four Dover sole sacrificed 7 days after first injection and five of six fish sacrificed 19 days after second injection had at least one otolith with a Toole etal.. Otolith microstructure. microchemistry, and early life history of Microstomas pacificus 737 Figure 2 Sagittal section of right otolith from 67.2 mm Stage-4 Dover sole, Microstomias pacificus, larva marked with OTC and held in a 12°C aquarium for 48 days. Magnification = 400X. (A) Photograph taken with ultraviolet light; arrows show fluorescent OTC marks. (B) Same preparation taken with full-spectrum light; arrows show position of fluorescent OTC marks. fluorescent mark and countable increments (70% of injected fish). Increment width was narrow, ranging from 0.63 to 1.88 um. Because all counts underesti- mated the true number of days since injection and because larger marks were not apparent by adjusting focus, the marks were not equiva- lent to subdaily marks observed in the first experiment. The OTC mark deposited after first injection was not visible in fish held until second injection, so increments deposited during the entire 183-day period could not be counted. Precision of counts Maximum and minimum increment counts for each otolith differed from mean counts, by an average of 5.7% (Range=0- 207c). Standard deviations associ- ated with mean increment counts averaged 1.84 (Range=0-7.8). Validation relationships Neither variances (Bartlett's test, S=1.48, P=0.11) nor mean observed/expected increment counts (ANOVA, df=27, F=2.21, P=0.10) differed between the five experimental groups offish (three aquarium temperatures in the first experiment, March and September marking groups in the second experiment). Similarly, regressions of observed versus expected counts did not improve when a model containing separate slopes and intercepts for each ex- perimental group was compared with a model contain- ing separate intercepts and one slope (P=0.313, df=5,22, Figure 3 Left otolith, anterior end of frontal section; from 68.0mm Stage-4 Dover sole, Microstomus pacificus, larva marked with OTC and held in a 10°C aquarium for 44 days. Arrow marks capture point, as identified by OTC fluorescence. Increments prior to capture are well-defined, whereas increments formed in aquarium are nearly indiscernible. Magnification = 400 ■ . 738 Fishery Bulletin 91(4), 1993 T: I O z UJ _J o a. < Q Z < 80 120 160 200 JUUAN DATE 240 280 Figure 4 Relationship between standard length and calendar date of Dover sole. Microstomas pacificus, held in an ambient flow- through seawater tank, 21 March-30 September 1991 (Julian dates 80-272). Smaller fish were not marked initially; dotted lines indicate uncertainty regarding growth trajectories. however, the power of this test was also too weak to conclude that increments were deposited daily (1-13=0.05). Error in the preceding regressions was due, in part, to poor resolution of laboratory-formed increments (Fig. 3; Campana and Neilson, 1985) and may there- fore be greater than the error associated with ageing wild-captured fish. However, several resolution prob- lems appeared common to otoliths of wild-captured and laboratory-reared fish: difficulty interpreting incre- ments at the otolith margin, difficulty resolving in- crements around stress checks (Campana and Neilson, 1985) in certain regions of the otolith because of com- pression (and sometimes fusion) of increments, and occasional difficulty distinguishing daily and subdaily increments. For these reasons, use of both slope and intercept estimates was considered reasonable when back-calculating days from increments in otoliths of wild-captured fish. The same relationships were ap- plied to increment counts outside the range of obser- vations, which may introduce an unavoidable source of additional error. P=0.899) or when the model with separate intercepts and one slope was compared with a model containing one slope and one intercept (P=1.14, df=5,22, P=0.369), so all experimental groups were initially combined into one regression. However, inspection of residuals from that regression suggested that a greater error was as- sociated with counts from Stage-5 juveniles than with counts from Stage-3 and Stage-4 larvae, regardless of experimental group. Mean ratios of observed:expected counts for fish <80mm (0.924) than for fish >80mm (0.718) U-test, df=26, P=0.012). Therefore, each size group was treated separately in the final regression models. The relationship for Dover sole <80 mm was Days since marking = 3.81 + 10.962 * observed increments), where n=21, r2 = 0.85, SEINTERCEPT = 2.90, and SESL0PE = 0.091 (Fig. 5A). The calculated slope was not signifi- cantly different from 1.0 (/ =0.419, df=26, P=0.34); how- ever, the power of this test (Peterman, 1990) was too weak to conclude that increments were deposited daily (1-13=0.06; Rice, 1987, Neter et al., 1989). The relationship for Dover sole >80 mm was Days since marking = 6.32 + (1.026 * observed increments i where n=l, r2 = 0.93, SEINTERCEPT = 2.81, and SESL0PE = 0.122 (Fig. 5B). The calculated slope was not sig- nificantly different from 1.0 (**=0.211, df=6, P=0.42); o UJ o z CO CO > < a 50 --"' .<&&- 40 .-■ tx-p^p- .--' 30 -*'"' -'Jk^^''0 -"""" 20 '• .' - - ' ' a p, - '\^\' ' , - ' 'a IU ''' A 0 i '.' i 10 20 30 40 50 SO ; .--' .-"' -/o -'* O'* ^s^ ,--' „ 40 r .-''-- ' ^^^ - - ' ' ' --' .'' ^s^ --* --" 30 r , -' .>\^^''' .**"* - , - * ' *' "x^T- ' „*■■*" 20 r -''' .-'b^af] e'' .-'' : ---^X-(2) - ' 10 [ << >" B 0 0 10 20 30 40 50 OBSERVED INCREMENTS Figure 5 Relationship between days since OTC injection and number of growth increments in Dover sole, Microstomus pacificus, otoliths. Inner dashed lines are 95^ confidence limits for mean response; outer dashed lines are 95% prediction limits. (Ai Relationship for Stages 3 and 4 <80mm SL. Formula: days since injection = 3.81 + (0.96 * observed increments); n=21; r2=0.85. 80mm SL. Formula: days since injection = 6.32 + (1.03 * observed increments I; n=l: r'=0.93. Toole et al Otolith microstructure, microchemistry, and early life history of Microstomus pacificus 739 Table 3 Characteristics of five Stage- 1 Dover sole, Microstomus pacificus, otoliths. Number of increments equals number observed plus number interpolated by linear approximation. The range of values presented for one larva reflects uncertainty in interpreting possible subdaily increments Mean Standard Collection Maximum Number of Percent increment length (mm) date diameter (mm) increments interpolated (%) width (um) 30.1 January, 1987 0.10 77 52 1.30 20.6 March, 1991 0.15 103 33 1.46 46.0 July, 1991 0.25 160-180 0 1.39-1.56 51.5 July, 1991 0.28 229 20 1.22 47.7 Oct., 1991 0.30 182 5 1.64 Structural patterns Growth from central primordium Two prominent areas were seen within the field of growth emanating from the central primordium. Clear central area Stage- 1 larval otoliths were translucent and hemispheric (appearing nearly circu- lar in sagittal section), with all growth emanating from the central primordium (Figs. 1 and 6). The left eye had previously migrated to the dorsal ridge of the cranium in all Stage-1 larvae examined. The diameter of Stage-1 larval otoliths ranged from 0.10 to 0.30 mm and the number of increments ranged from 77 to 229 (Table 3). When fully formed in Stage-3 larval otoliths, this clear central area was 0.24- 0.49mm (mean=0.39, n=60) and contained 187-230 increments (Table 4). Mean increment width in the clear central growth area ranged from 1.22 to 1.56 um, and increment widths nearer the cen- tral primordium were approxi- mately 0.5 um, suggesting that, when counting increments, under- estimation may have occurred ow- ing to resolution limitations of light microscopy (Campana et al., 1987). A structural feature of some Stage-1 larval otoliths was the presence of one or two conical cavi- ties, which radiated towards the anterior and posterior edges (Fig. 7). Each cavity was enclosed laterally and medially and was open at the anterior or posterior end. The anterior cavity may be continuous with the sulcus, which was visible on one Stage-1 larval otolith. Opaque central area An opaque area surrounded and was continuous with the clear central area in all Stage 3-5 larval otoliths (Figs. 1 and 8, A-D). Incre- ments in this region were wider than increments in the clear central area ( 1.6-1.8 um) and there was higher contrast between continuous and discontinuous zones within increments. Continuity of increments was in- terrupted by the conical cavities described previously and by accessory primordia. The number of increments in this area ranged from 46 to 95 (Table 4). Growth from accessory primordia We recognized two zones of growth from accessory primordia, distinguished by the point at which growth from the central primor- Figure 6 Sagitta from 20.6mm Stage-1 Dover sole, Microstomus pacificus. larva collected on 1 April 1990 at NMFS Station 9003-35. Estimated number of increments was 103 (Table 3). Magnification = 1000 x. 740 Fishery Bulletin 9 1 (4), 1993 Table 4 Increments counted from central primordium to structural landmarks on four Stage-3 Dover sole, Microstomia pacificus, otoliths shown in Figure 8, A-D. Assigned dates (in parentheses) are determined from the increment:day regression for growth distal to accessory primordia. Periodicity of increments proximal to accessory primordia has not been validated. Ranges indicate uncertainty interpreting possible subdaily increments or determining exact location of the clear to opaque transition. The first distinct accessory primordium is not the first-formed accessory primordium. NP = not present; AP = accessory primordium. Landmark Standard length of larvae (mm) 55.6 46.9 64.9 55.5 Central primordium 0 0 0 0 Clear to opaque transition 187-213 205-219 =230 =200 First distinct AP 196-222 (19 Nov. 88) NP 322 (3 Dec. 89) 206 (17 Dec. 89) Last AP 224-250 (16 Dec. 88) NP 338 (18 Dec. 89) 222 (1 Jan. 90) Central primordium enclosure NP 300-314 (20 Feb. 90) 376 (24 Jan. 90) 252 (30 Jan. 90) Otolith edge 255-281 322-336 425 296 Capture date 19 Jan. 89 17 Mar. 90 17 Mar. 90 17 Mar. 90 dium was completely enclosed. This determination is based on sagittal sections at the level of the central primordium. Unenclosed peripheral area Otoliths of all Stage 3-5 larvae had at least two accessory primordia ( AP), when viewed in sagittal section ( Fig. 8, A-D). A maximum of seven AP were observed, although additional AP on the lateral surface of the otolith (Fig. 9) were obscured when otoliths were prepared in sagittal section. Ac- cessory primordia always formed ad- jacent to growth from the central pri- mordium, rather than adjacent to growth from previously formed AP. On right otoliths, AP were easier to discern than on left otoliths, possibly because the plane of growth changes at about the time of AP formation in left, but not right, otoliths (Fig. 10, A and B). Increments emanating from AP exhibited higher contrast and were wider (about 3.0pm) than those emanating from the central primordium. The most anterior AP appeared to form first and was closely associ- ated with the transition from clear to opaque central areas. However, the exact origin of the anterior AP was usually impossible to discern because it merged / Figure 7 Sagitta from 51.5 mm Stage-1 Dover sole. Microstomas pacificus, larva collected on 12 July 1991 near Destruction Island. Washington. Anterior and posterior cavities are indicated by arrows. The lateral face of the otolith (surface toward viewer) has been ground to the level of the central primordium; the medial face has been left intact. Estimated number of increments was 229 (Table 3). Magnification = 400x. Toole et al Otolith microstructure, mlcrochemistry, and early life history of Microstomas psoficus 74 A B y D Figure 8 Stage-3 Dover sole. Microstomas pacificus, otoliths showing variations in accessory primordium formation; anterior to left, dorsal toward top in A-C, dorsal toward bottom in D. AP = accessory primordium, CCA = clear central area, CP = central primordium, OCA = opaque central area. (Al Left otolith of 55.6-mm larva collected 19 January 1989, showing unenclosed central primordium growth (indicated by triangle). Magnification = 100x. (B) Left otolith of 46.9-mm larva collected on 17 March 1990, showing two accessory primordia (anterior and posterior) completely enclosing growth from the central primordium. Magnification = 100*. (C) Left otolith of 64.9-mm larva collected 17 March 1990, showing growth from four accessory primordia (anterior, posterior, dorsal, and ventral) completely enclosing growth from the central primordium. Magnification = 100 • . (D) Right otolith of 55.5-mm larva collected on 17 March 1990. showing growth from four accessory primordia (one anterior, two posterior, and one dorsal) completely enclosing growth from the central primordium. Magnification = 100x. with the conical cavity described previously (Fig 8, A-D). This situation also applied to the origin of the most posterior AP in left otoliths (Fig. 8, A-C). The chronology of additional AP formation did not follow a consistent pattern. Dorsal and ventral AP were often associated with dark bands in the opaque central area (Fig. 11). Dorsal and ventral AP were formed 196-338 incre- ments distal to the central primordium (Table 4). When increments were counted back from the otolith edge. dates of last AP formation varied between fish, rang- ing over a six-month period from October to March (Fig. 12). The number of days between the first-formed and last-formed dorsal and ventral AP averaged 28.8 (range 1-75, n-25). This range would be greater if it were possible to make accurate counts to the anterior AP. Enclosed peripheral area Growth from accessory primordia completely enclosed growth from the cen- 742 Fishery Bulletin 91(4), 1993 . • -' ' Figure 9 Transverse section of right otolith of 47.5-mm Stage-3 Dover sole. Microstomas pacificus, larva collected on 22 March 1989, showing recently-formed accessory pri- mordium near medial surface (toward top) that would be undetectable in sagittal section. Magnification = 400 x. tral primordium in at least one otolith of 78% of benthic and 71.4% of pelagic Stage-3 larvae, 98.4% of benthic and 100% of pelagic Stage-4 larvae, and 100% of Stage- 5 larvae, suggesting that most otoliths were enclosed before or during Stage 3. Enclosure of growth from the central primordium occurred an average of 33.7 days after formation of the last accessory primordium (range 0-87, n=108). During Stages 3 and 4, SINT increases as the intes- tinal loop extends into the secondary body cavity and BD1A decreases as body depth shrinks (Markle et al., 1992). Consequently, fish with low SINT/BD1A ratios are presumably at an earlier stage in the metamor- phic process than fish with higher ratios. Benthic Stage- 3 larvae with at least one unenclosed otolith had a lower ratio of SINT/BD1A (mean=1.06) than benthic Stage-3 larvae with enclosed otoliths (mean=1.15) (t- test, df=46, P=0.0096), as did pelagic Stage-3 larvae (0.90 vs. 1.32, t-test, df=28, P=0.019), suggesting that Stage-3 larvae with unenclosed otoliths were at an earlier stage of development than Stage-3 larvae with enclosed otoliths. Stage-3 larvae with enclosed otoliths were collected an average of 24.4 days after enclosure (ranges 5.7-65.4, n=51) while Stage-4 larvae in the same col- lections averaged 51.9 days (range=9.6-125.9, n=46). The difference (27.5 days) may represent the average duration of Stage 3. Otolith elongation along the anterior-posterior axis became more pronounced following enclosure, as did asymmetry between left and right otoliths. Left otoliths were heavier (paired t-test, df=335, P<0.001), longer (paired /-test, df=314, P<0.001) and wider (paired /-test, df=312, P<0.001) than right otoliths. In addition, the proportion of otolith length anterior to the central pri- mordium was higher in right than left otoliths (i-test, df=92, P<0.001) by an average of 8.5% (Fig. 8, A- D ). The number of accessory primor- dia tended to be higher on right than on left otoliths (paired /-test, df=62, P=0.036), while more incre- ments formed after the last AP (paired t-test, df=34, P=0.006) and after the point of enclosure (paired t-test, df=15, P=0.011) on left than on right otoliths. Stress checks became prominent in sagittal section following enclo- sure (Fig. 8, B-D). Enumeration of stress checks and determination of periodicity was difficult because checks were often discontinuous around the otolith and, even when continuous, the number of observable increments between checks var- ied regionally. However, mean periodicity of check formation in 46 Stage-3 and Stage-4 otoliths collected over a three-day period in March, 1990, suggested synchrony of check formation in otoliths of different fish and regularity in periodicity of check formation (Fig. 13). Calculated dates of check formation did not correspond to particular lunar phases. Number of days between checks ranged from 5.7 to 33.6, with a mode of 12.5 days and an average of 15.3 days (Fig. 14). Other features of the enclosed peripheral zone in older fish were the development of a rostrum (Figs. 1 and 15) and formation of translucent growth zones (Fig. 16). All Stage-5 otoliths collected in January and March had at least one translucent growth zone. This first post-settlement translucent growth zone was ini- tiated in late fall of the settlement year (240-293 days after formation of the last AP) and was completed the following spring (125 to 141 days later) in otoliths of three Stage-5 juveniles. Based on daily increment counts and seasonal deposition patterns, the first and subsequent translucent growth zones are interpreted as post-settlement annuli formed during slow winter growth periods (Fig. 17, page 746). This interpretation is consistent with Hagerman (1952) and Chilton and Beamish (1982). Although there was little difference in weight be- tween the largest Stage-4 and smallest Stage-5 Do- Toole et al.: Otolith microstructure, microchemistry. and early life history of Microstomus paaficus 743 B Figure 10 Frontal sections of left and right Dover sole, Microstomus paaficus, otoliths showing differences in orientation of the clear central growth area; ante- rior to right, lateral surface to top. (A) Left otolith from 58.0-mm Stage-3 larva collected on 16 March 1990. Central area tilts towards the lateral surface posteriorly and towards the medial surface anteriorly. (Bl Right otolith from 64.9-mm Stage-3 larva collected on 17 March 1990. Central area is oriented in the same direction as surrounding peripheral growth. Stage-5 juveniles; and the third, of only Stage-5 juveniles. Microchemical patterns Ratios of Sr/Ca exhibited similar patterns along transects through the otoliths of all lar- vae and juveniles examined (Figs. 20, A-B, page 747, and 21, page 748). An initial Sr/Ca spike (0.007-0.010) occurred approximately 48 urn from the central primordium, inside the clear central area. Following the spike, Sr/Ca levels within the clear central area fluctuated at intermediate levels (0.004-0.006). Sr/Ca ratios began a second, more gradual, decline beginning near the clear-to-opaque transition in the central area of the otolith, which also corresponded to the area in which the ante- rior AP was forming. This decline began in late summer or fall of the calendar year prior to settlement. Sr/Ca ratios often reached minimum levels following formation of accessory primordia. In a 65.7-mm Stage-4 larval otolith (Fig. 20A), Sr/Ca ratios reached a minimum (0.002) at about the end of January. This minimum point was approximately 10 days beyond enclosure of growth from the central primordium and 31 days beyond the most recently formed acces- sory primordium. In a second Stage-4 larva, a minimum (0.003) was first reached in mid-De- ver sole collected in January and March, otolith weight in Stage-5 juveniles was nearly double that of Stage-4 larvae (Fig. 18, page 746). A discontinuity in the rela- tionship between Stage 3-5 right otolith length and fish length dur- ing the first winter following settlement was also apparent (Fig. 19, page 746). This relationship was best described by a three- stanza segmented linear model (Bacon and Watts, 1971; Laidig et al., 1991). Parameter estimates in Table 5 (page 746) resulted in a good fit (r-=0.955) and no discernable pattern in residuals. The first segment consisted of Stage-3 and Stage-4 larvae; the second, of Stage-4 larvae and ¥ Figure 1 1 Posterior accessory primordium on right otolith of 58.0-mm Stage-3 Dover sole, Microstomus paaficus, larva collected on 16 March. Narrow dark band in the opaque central area is associated with origin of the accessory primordium. Magnification =400> 744 Fishery Bulletin 91(4), 1993 >• o z LU 3 O LU Q DJ JANUARY 1989 N = 28 ~ MARCH 1989 - N = 27 : u\i[ i* "f!i*k i'i-l-ji H i i ■ i ■ ■ i ■ ■ i i OCT NOV DEC JAN FEB MAR APR Figure 12 Dates of last accessory primordium formation on right otoliths of Stage-3 and Stage-4 Dover sole, Microstomas pacificus, larvae collected during three sampling cruises in 1989 and 1990. Only enclosed otoliths with three or more accessory primordia were considered because origins of single anterior and posterior accessory primordia are usually indistinct. 40 30 >■ o LU O 20 O LU rr LL 10 0 : j • — — ^T^ Freq five r 4 Do betw 0 10 20 30 40 DAYS BETWEEN CHECKS Figure 14 jency distribution of number of days between the nost distal stress checks on 46 Stage-3 and Stage- rer sole, Microstomas pat ifii us, otoliths collected een 15-17 March 1990 (n=158). LU DC 3 < o Cl O ti n. to > < Q 80 60 -• 60 52 1 47.6, N= 56.5,N=29 MODE = 4 39 40 45 38 MODE = 48.0 38.6, N=45 20 • 30 23 + MODE = 36.5 25.4, N=46 MODE = 25.0 -o :3 16 8 + 13.7, N= MODE = 46 12.5 0 -• 0 CHECKS PRIOR TO CAPTURE Figure 13 Relationship between the five most distal stress checks and number of days from their formation to the otolith edge (cap- ture) in 46 Stage-3 and Stage-4 Dover sole, Microstomas pacificus, otoliths collected between 15-17 March 1990. Vertical bars represent 95"7r confidence intervals for means (horizontal bars I. Additional error is associated with convert- ing increments to days, with the regression described in the text. Corresponding lunar phases are displayed on the verti- cal axis. cember. The ratio then rose and fell to a second mini- mum (0.003) in mid-January. In each of these lar- vae, the ratio rose again (0.003-0.0035) prior to cap- ture in mid-March. Two Stage-3 larvae and a third Stage-4 larva did not exhibit distinct minima; rather, Sr/Ca ratios remained at about 0.0025 until cap- ture. Sr/Ca ratios first reached this level prior to enclosure. An 88.6-mm Stage-5 juvenile (Fig. 20B) with one peripheral annulus had narrow growth increments, making it difficult to determine exact ages associated with microprobe samples. However, the pattern proxi- mal to AP was similar to that of larvae, with one 0.007 spike and a 0.004-0.006 peak near the center of the otolith, followed by a decline occurring prior to and during formation of accessory primordia. Sr/Ca ratios remained low for >200 days following formation of the last AP until a second 0.004-0.005 peak formed in the translucent annulus near the otolith edge, which rep- resented capture in March. A 172.2-mm juvenile, collected in November 1989 and judged to have three complete post-settlement an- nuli and a fourth one forming, had four Sr/Ca peaks, which corresponded to translucent annuli (Fig. 21), distal to accessory primordia. Toole et al.: Otolith microstructure, microchemistry and early life history of Microstomas paaficus 745 Figure 15 Left otolith of 65.0-mm Stage-4 Dover sole Microstomus paaficus larva collected on 17 March 1989: anterior to left. Note stress checks, which are clearest in the ante- rior field behind the anti-rostrum. Magnification = 50x. Discussion Relation of otolith microstructure to metamorphic stages were developed in Stage-1 larval otoliths, yet both were present in all Stage-3 larval otoliths, it is prob- able that they formed during Stage 2. Thus, the transition from clear to opaque growth and initiation of AP formation occur after the left eye moves beyond the dorsal ridge and metamorphosis (as defined in Markle et al, 1992) begins, but be- fore larvae are first collected on the bottom. Formation of the first accessory primordium in plaice and winter flounder also does not occur until the left eye has migrated at least to the dorsal ridge of the cranium (Alhossaini et al., 1989; Sogard, 1991). However, unlike Dover sole, AP in otoliths of these spe- cies can form before the eye has moved past that point. Compari- sons with other flatfish species are more difficult. Accessory pri- mordia form "at or shortly after metamorphosis" in starry flounder (Campana, 1984a) and "after meta- morphosis" in California halibut (Kramer, 1991); however, the definition of metamorphosis relative Premetamorphic larvae Sagittal otoliths of Stage-1 larvae between 20.6 and 51.5 mm were uniformly translucent and lacked accessory primordia. Since the left eye had migrated to the middorsal ridge of the cranium in each specimen, for- mation of AP is not triggered by the initiation of eye migration or move- ment to this position in Dover sole. , Cavities noted in Stage-1 larval sagittae have not been described previously and their derivation and function are unknown. Metamorphic larvae Because Stage-2 larval otoliths were not available, the transition from clear to opaque central areas and devel- opment of the first two accessory primordia cannot be attributed to this stage with certainty. However, because neither of these features Figure 16 Right otolith of 100.4-mm Stage-5 juvenile Dover sole, Microstomus pacificus. caught on 18 March 1991 I anterior to right), photographed with reflected light against black background. Note the first annulus, which appears darker than the inner area. Magnification = 50 x. A = annulus. 746 Fishery Bulletin 91(4). 1993 200 1 N=9 -170 5 ' N = 11 LENGTH T N=51 TANDARD o N = 123 « 80 - + N = 146 50 - 0 12 3 4 NUMBER OF POST-SETTLEMENT ANNULI Figure 1 7 Relationship between standard length and number of post- settlement annuli in Dover sole. Microstomus pacificus, col- lected in January 1989, March 1989. and March 1990. Vertical bars represent 95% confidence intervals for means (horizon- tal barsl. to cranial morphology was not specified in these studies. Formation of accessory primordia has also been noted in several non-pleuronectiform taxa (Brothers, 1984). Gartner (1991) described a clear to opaque transition with accompanying formation of AP in otoliths of myctophid species and suggested these features were 2 i \~ a z LU — I Q I < □ z < 240 200 A o<&g* o 160 ^$0%°° 120 80 . nji fr^ 40 0 : % 2 3 OTOUTH LENGTH (MM) 2 240 T 200 I- W i«n Z LU _J 120 D rr 80 < a 40 s < 0 UJ 0 1 2 3 4 5 OTOLITH LENGTH (MM) Figure 19 (A) Relationship between standard length and right otolith length in Stages 3-5 Dover sole, Microstomus pacificus. For- mula and parameter estimates are included in Table 5. n = 351; r- = 0.955. Points representing standard length and right otolith diameter of five Stage- 1 larvae are also included. While Stage-2 otoliths were not available, it appears that otolith length increased between Stage 1 and Stage 3, whereas fish length did not. (Bl Polygons circumscribing areas bounded by fish in Stages 1 and 3-5. 3 • STAGE 3 WEIGHT (MG) ro n STAGE 4 ** * * *,* ** * ** * * *^ * STAGES * * * * * J ' ' • ' m % * * * „ * * * * * Vn * * X i- ° D OTOL .*&£&?****' *" 0 0 12 3 4 5 6 7 BODY WEIGHT (G) Figure 18 Relationsh p between weight of right otolith and body weight in Stages 3-5 Dover sole, Microstomus pacificus. Note change in the relationship during the transition from Stage 4 to Stage 5 i/i=368). Table 5 Estimated parameters for segmented linear regression model of the form Standard Length (mm) = e + rf,(.v-C|) + djix-c , Is, + f/,(.r-c,)s2, where x = right otolith length immi, s, = |2*(l/(l+exp ( — [ X— c, ]/0. 1)1)1—1 (i.e.. the logistic cumulative distribution function, used as a transition function between segments. with the constant 0.1 specifying a rapid transition be- tween segments), e, = fitted inflection points and e and , SW Africa (iV=40) and Chile (iV=22). Both a one-way analysis of variance (ANOVA) and an ANCOVA with skull CBL as covariate and with geographic region as classification variable were uti- lized to assess among-group variation in skull size and shape. Next, most morphologically similar group pairs. Van Waerebeek: Variation in skull morphology of Lagenorhynchus obscurus 757 namely Peru/Chile and SW Africa/New Zealand were compared by means of /-tests to identify discriminat- ing variables. Discriminant analyses (DISCRIM) were performed on samples limited to specimens with a complete set of measurements. This sample included 135 skulls from Peru, 24 from New Zealand, 22 from SW Africa and 7 from Chile. A first DISCRIM with 21 characters maxi- mized discrimination between geographic groups. Sub- sequently, a series of discriminant analyses was ex- ecuted by considering various combinations of a small number of highly variable cranial characters so that an optimum segregation between groups was obtained. These discriminant functions can be used tentatively to assign undocumented museum skulls to respective populations. It must be cautioned however that no specimens from Argentina have been examined and the discriminant locus of that population remains unknown. Non-metrical characters Initially, 35 non-metrical cranial characters (NMC), fol- lowing Perrin et al. (1982111), were determined, but se- rious problems with score repeatability over time re- duced the number of usable NMC to 27 (Table 1 and Fig. 1). Similar problems have been reported for the common dolphin by Perrin et al. ( 198811, 1989). Analysis of sexual dimorphism was limited to skulls of sexually mature specimens from central Peru (Ar=109), eliminating influences of growth and geogra- phy. Sexually monomorphic characters (the full set, see Results) were then tested for dependence on cra- nial maturity status in 173 mature and 48 immature skulls. Geographic variation in non-metrical traits among the four geographic areas was studied after eliminating variables influenced by sex and maturity. All significance testing of variation in NMC was done by means of chi-square contingency analyses. Tooth counts were compared by /-tests. Table 1 Non-metrical cranial characters utilized in present study (from Perrin et al., 1982"); see Figure 1. Character 1L/1R (left/right). Character 2L/2R. Character 3. Character 5L/5R. Character 6L/6R. Character 9L/9R. Character 10. Character 12L/12R. Character 13. Character 16. Character 17. Character 18L/18R. Character 19. Character 23. Character 24. Character 25. Character 30. Character 33L/33R. Character 34L/34R. Position of the foramen (A in Figure 1) in the premaxillary at the base of the prenarial triangle, relative to the level of the anteriormost projection of the antorbital processes. Codes: l=at; 2=ahead; 3=behind the level. Position of the anteriormost large foramina (B) in the maxillary relative to the level of the anteriormost projection of the antorbital processes. Codes: l=at; 2=ahead; 3=behind the level. Asymmetry of position of the two anteriormost large maxillary foramina 1B1. Codes: l=symmetrical; 2=left foramen more anteriorly placed; 3=right foramen more anterior. Number of small foramina (C) in the maxillary anterior to the anteriormost of the large foramina. Codes: l=one foramen; 2=two foramina;3 etc.; 6=no foramina. Number of foramina ID) in the maxillary behind a line at the level of the anterior edge of the external nares and perpendicular to the long axis of the skull. Codes: 1=1 foramen; 2=2 foramina; 3 etc.; 6= no foramina. Contact between maxillary and occipital, at point where occipital crest intersects margin of temporal fossa (E). Codes: l=contact lor space of lmml; 2=closed. Anterior contact between hamuli (Nl. Codes: l=open (gap>lmml; 2=closed. Spine of palatines visible in anterior gap between hamuli (Ol. Codes: l=visible; 2= not visible. Mesial fenestration in basisphenoid (Rl. Codes: l=foramen present; 2=absent. Number of mental foramina. Codes: 1=1; 2=2; etc. Arrangement of mental foramina. Codes: l=foramina lie along a simple curve; 2=foramina do not lie along a simple curve. 758 Fishery Bulletin 91(4), 1993 CHARACTER 13(H) ' FRONTAL 3 Figure 1 Non-metrical cranial characters utilized in this study following Perrin et al. (19821. Letters refer to Table 1. Results Cranial maturity Relationships between the degree of fusion in cranial sutures and sexual maturity for L. obscurus are shown in Table 2. Less than 25% of the adults exhibited ad- vanced rostral distal fusion (MXDIST=2) which, as in L. obliquidens (see Walker et al., 1986), eliminates MXDIST as a selection criterion for adult skulls in this species (too many useful skulls would be rejected ). However, three other sutures, frontale-supraoccipitale (FROC), zygomatico-parietale/exoccipitale (ZYG), and lacrimale-maxilla/frontale (LAC), had advanced fusion (state=2) in 95%, 89r/r, and 85%, respectively of the sexually mature specimens. Equally important, only two sexually immature specimens from New Zealand showed full fusion in any of these sutures: one was a pubescent female with large follicles (ZM39957) and the other, an unusually large female without ovarian corpora (ZM39561, 187 cm). Moreover, based on over- all size, those skulls fitted into the adult series. Skulls with at least some fusion (state 1 or 2) in either MXDIST, SYM, or PTPAL are very likely from sexu- ally mature dolphins (see Table 2). Metrical characters Individual variation Summary statistics for cranial characters in adult female and male dusky dolphins of central Peru are presented in Table 3. Of 74 charac- ters tested, only one (upper left tooth count, for fe- males) departed significantly from normality (Kolmogorov-Smirnov, P=0.017), which could be attrib- uted to chance fluctuations. Homogeneity of variance (95% confidence, TWOSAM, STSC Inc., 1989) was con- firmed for all sex-paired characters. Statistics summarizing individual variation in cra- nial characters are given also for mature dusky dol- phins of northern Chile, southwestern Africa and New Zealand (Table 3). No significant departures from nor- Van Waerebeek: Variation in skull morphology of Lagenorhynchus obscurus 759 Table 2 Percent frequencies rounded to nearest integer, for degrees of fusion seen in cranial sutures of Lagenorhynchus obscurus specimens of known sexual maturity status (sexes pooled i. Mature am mals are all from Peruvian waters; subadults include specimens from Peru, southwe stern Africa, and New Zealand States of fusion are: 0 = unfused; 1= some fusion; 2= advanced fusion. Cranial sutures are defined in text. Suture Subadults (0.5). Sexual dimorphism The overall size of the adult dusky dolphin skull, based on condylobasal length and zygomatic width, is equal in males and females it- tests, P>0.50; Tables 3 and 4). Highly significant sexual dimorphism (/-tests, P<0.005) was found in the width of the rostrum over most of its length (RW60, RW1/4, RW1/2, PMX1/2, RW3/4) and in the length of the tem- poral fossa; males had consistently higher mean widths. Analyses of covariance showed sexual dimorphism in exactly the same variables and with very similar sig- nificance levels as in the two-sided /-tests, indicating proportional differences (shape). However, it is uncer- tain whether the observed, small dissimilarities have biological meaning, particularly as there is wide over- lap in ranges. I was unable to produce a satisfactory model to predict the sex of an unknown animal through stepwise variable selection in a ^multiple regression (see Douglas et al, 1986) exactly because of those subtle deviations. Twenty-four of 36 characters showed significant cor- relations with the condylobased length (CBL) covariate, including five of the six dimorphic characters (Table 4), implying that much of the variability is explained by variation in CBL. Eliminating dimorphic charac- ters from the geographic analysis therefore entails little loss of information. Seasonal variation An analysis of seasonal variation was feasible only for Peruvian specimens. Skull size (CBL) was not seasonally dependent (MANCOVA, F=1.45, df=3, P=0.23) but six other skull characters (RW60, RW1/4, APL, RAL, TCUL, and TCUR) demon- strated seasonal variation at a low level of significance (MANCOVA, 0.05 < P < 0.02). However, at least in the case of RW60, RW1/4, and RAL, signifi- cant interactions due to sexual dimorphism in these measure- ments seemed to be the cause (P- values for 2-factor interactions were respectively 0.039, 0.047, and 0.037). In APL, TCUL, and TCUR, which are sexually mono- morphic characters (see above), there were no sex/season inter- actions (P>0.50). The significance in these few cases could be due to chance in view of the large number of /-tests. This is sup- ported by the results of a series of Runs tests (Runs above and below median, Runs up and down, STSC Inc., 1989) on chronologically ordered CBL, tooth counts and tooth widths. These results did not suggest any non-ran- domness (Z>0.05 for all tests). Geographic variation Exclusion of the few sexually dimorphic characters allowed combining samples of males and females and the use of cranially mature skulls of unknown sex. Highly significant differences between geographic regions were confirmed for 29 of 31 cranial characters (for all ANOVA, P<0.005 and mostly P<0.0001; Table 5). Tympanic bulla width (BW) proved divergent between groups but with less confi- dence (ANOVA, F=4.1, df=3,117, P=0.019) and no dif- ference could be detected for periotic length (ANOVA, F=0.24, df=3,201, P=0.87). Variance generally was ho- mogeneous (Cohran's C-test and Bartlett's test, P>0.01 and mostly P >0.05); exceptions include TCLL (Cohran's, P=0.006) and PEL (Cohran's, P=0.003; Bartlett's, P=0.009). ANCOVA analyses revealed discernible variation in 16 of 30 skull characters which can not be attributed to divergences in skull size alone but are due to pro- portional differences (i.e., shape); 11 characters were highly significantly different between geographic groups (P<0.01, Table 5). Means and 95^f confidence intervals are plotted for each variable against geographic region (Fig. 2). The most striking element is the overwhelm- ing difference in skull size, reflected in most measure- ments, grouping skulls from Peru and Chile and speci- mens from New Zealand and Southwest Africa. Skulls of dusky dolphins from the latter pair-group have, on average, a CBL about 3 cm shorter than skulls from Peru and Chile. 760 Fishery Bulletin 91(4), 1993 (— 1 s x m x t- ac c— cn --t CO CO CO CM O CD CO H(OH[- CO as t> co co m m •— " cm co o OJ OC i~> o t— r- CO uo CO m c- as as co oc co t- m m ** CO CO X os cm r- co p? CO CO J, A =j> 1 7 oi J: co 0) bo B cd CO CN <-* 1 1 [ 1 I i-l CO ^P 00 OS CD CD CO CN CO CN CM 4 in co t~ ao co as o co in co "* CO "* -^ CN ■«P -* -H 7 3 ■= °° CO ^P CD CN CM 3 CM 1 CM p3 ^t t> 00 as rf CO CD CO m ^ X 3 O CO -H CM r-> ^ ^ i-> 1-1 CN Q Oi O CD tJ- CO iO m „ ^ CM t- Tp M- m co CD O O -H CO !>; O CM in co co x p CM CM m CO CM t- X CM CO i-4 o Q CO o os -*t co co CO CN CO CN as m m m m co CM Tt CO CM CO CM Tp CO CO CD CN -h rH CM CN OHO^TfHCN 3 c O -3- CN t- CO CN X OS X ,_, OS CO t-- CM OS cm co m x m C- t~- CM t-; CO CO p O m ■^P m p P CO O CO Tp CM CO 0) C^ od t> cd 00 oi CN ■^; OS CO x in t> o iri co os os os in co t- d 0 co iri i-i CD CD ■H- m co i-t O O CO t~- 0 0 cm x m m m cm i-h Q- Tf Tj- Tj- ■**■ rj- ■* ■* CO ^t CO Tf "^ ^f "^F 'I* ■^■^M1 Tf Tj- ^P CO TP TP ■■* "H. CO co -p ^ COHH^ Tji H rf 0) CJ CO m CO X o CO o o OSCOHOifi >-t X O CD «-• t- CO CM t- 0 1-* -# CO CM ■*p cm co m as -h cm 0 u bD c cd CO CN f-t 1 1 1! 1 0)4 CM CO in m CO 4 i CO as CO CN 4 CM o> h 4 lO ci CO "t lOCO OS CO CO CM I CM CM 1 CM o o < (3 Tt t- r- m co co co m m x 00 i-t CN CM --1 CM X i-H s 1—1 CU CO Q © © -*cP CM X <7> — c O O t> CO CM -- CM CD TP- X Tp X lOOTf m i-h co co t^ p in p >-H p p CO CM i-h ca CO i- t^ irinN oi CM CO 00 CO ^" tj" iri iri co CN Tp CO CM CN CN f' CD hdoin 1-1 ^ 1-1 O f-i O ^ ^ i-H CM CD X c CO CO h;COH CN ,_, IC CO ,_, ^r x x o Tf" CO ft O X • 16 as in i> o co ■rf- as i-J i-i rp CN OS i-i oi 00 iri 0 CD OS OS t ' ^ oi i' w 0 id CO 05 O! CO t- CD CO "* CO CO m r- r- in c^ m cd m in co id 10 r~ t- O CO CO CO CM CM CO — ' O CM CO CO c CO — t CN CN rtHHH ^ t-t CO .2 Z Tp CD OS O CD CD CD' CO m X NCO -t C-O CD CO [^ ID f- I- 0 CM tJ- CM i-H OS O ,_ CM CD O O t- C- 0- O cd a, o CO CO CO ^3" CO CO CO CO CN CO CO CO CO "tf CO CO CO CO CO CO CO CO CO CO CO CM CO CO CO CO CO CO CO CO CN CO O m OS O CN rt m t* ■^T x m os co co os as- r- co o CO CO "* m CO -1 rp CO m m r^ cd I i-h co co I O, bo c CO cc ■-> CO O 00 00 t- 3 CO CO m t— t— t- CD OS CO CO m r~ m m "rp O CD 00 t> CN in co T ^p co CO CD O OS -P t— CO CO CO CO _J CO 1 CM CO CO 1 J4 47at ^ co or-o a cu ts- cm ^h i 1 1 OS o r- oo os cd r— oi 1 CN OS CN 4 in co ci t- 1 1 CM CO t— J m m 0 7 1 CO CD CN i-h CD CM 2 CM i CM l CM OJ J3, i - r x CN ■<* m c- t— Tf X r- 1-1 CO r-t i-H 1 O CO CM CN CN rt "" " 1-1 CO -5 Q CO O -H i-i t- CO CO tO m X x m CD CD WHOOSH r- ^p m OS CM CM t> t> CD x os m co rp m ■* > ■* CO OS UO ■* X OS O q CO r-t rH CM in O CN -h O l> N 0 in os 0 P p CM CM X CN as CD OS OS -^ CO CO CO CN ■* CN X cb in in co co CM CO CO CN CN CO CM t-" as 0 i-i cm 1-1 1-1 CN 0 0 I iri co i-h I £ c as oo m oo •— CO t> ■* OS CD CO O CM CO CM cm o m x co OS -h CM in p oq os CO CM -P co co os os m a cd cu Tf O O -CP CD CD CO as as as -t Tf i~; Tf cd Tp tP CM CM CO CD CN ■"-< CO f~ X 0 O O O tp Tp I os iri 0" 1 as cn o c- r- CD CO ■^ m in cd x x m t— ID t- ID ID CO m co os X CM CO CO CO CO CO CO 1 O CM CO 1 X CO CM ■-" CN CN i-H CO u C 2 ^COH H CO CD CD CD ■^i o as cd t— m cm CM CO CO X t- CO h CO -P -p r~ 2 m | s 2 . "3 o bo c cd o —• co tj- m IC CO t— ■^r CO X X CM CD CO os co m as co f Tf f. O O "*P CO CD CO CD ID CO ID Ol T CN ID cm Tf i— i x ao t> 7 CN m X X r^ os o cd x m x co m ^p o co 4 cc « CO c- 0 n co r- co CO CO CO *> 7 -H- Uh 2 7 CM ^2ii&4 k o a. T^toiS d> J, CN 1 CM 1 ^H -H 2 oi, JE1 --♦ oo --< t- t- ■<* -1 as m co x X CO O CO t— ■fl1 h/ f- CO OS CO rH c cd cu cp ej i-i co t> t> C~" o CM o ^Tfco'co'r' ^ rp CM CM CO t- ^ CD OHJCD-H ,— i _H _H tp iri os 0' r~ 0 t^ >i '-£» cn — < o t— c- CO CO in CD CD co x x m c- m c— m m co m co as as co co co CO CO CO CO 1— ■— CN CO 03 CJ _0; CO CN i-H CN CN HHHH '""' i-4 CO I— 1 ^H 3 -a 0) CO 2: 3! ® O O CJ) Oi OS OS CO OS o o c o o o o o o o 0 0 as t~ t^ r- os OS t> c- OS OS 0 0 O OS X 0J U. *3 f I1 iC KJ Tt -i- -* tT- «* -* io m in io io m m m m m m in t Tf -^ Tp ^p •r •HT -p Hi «t CM ID ID h/ T)1 3 CO bo c CO ^ CN CM CO CN T X X t~~ as as co m CM O CM x m m c- i-h O O O c- OS c-~ t^ -1 os m as x co m £ CO C N (NH COCO 3 OS in l CN CN CM 1 CN CN HHHHQO 1 1 1 1 CN co x co m -p 1 T J _L 1 X CM CO 00 CM cp r- cm T 1 CM 1-1 tr~ I CN CO X CO CM CO 1 I 1 co as CO CO 3 CO 2 1 co ^ 1 7 cc 7 c CD c5 00 X t^ CD CD ua CN ■T CN CN i-i x x r- c~ -tf CO Tp "* CO m m —• «-h CM CD CN CM CM CM 11 . co JL i-h co cm JL co 2®H - E "o CU Q. CO CO Ol CO CO CO in co r- co t~~ r*- «-h Cfl 1 0) CO i-H CN CN HHHH •"* 1-1 CO " r-f H CO C 3 Q O) CO m rH UO "^r tJ- as tp r- «-* x co r- — m os Tp X Tf MDCTl -P CO ^P co as as m co co c fa CO CD — t CD CO OJ OS tC Tf O CN oo in i-h ■-* os in oc ^p cq oq O CO CO in p p in V- O CO h CD hub HiH CJ CN OS Tt CO CO CN CN CM tj" iri iri CD CN CM CO CO CM CM CO ^J- 00 r-^ oi cm i-i CM cm' d h d ^ v h in -C OJ 3 ■"* 1-1 *""' t- CD o c 0) c Tf x >— i r- i/D CO lO CN i— j m CD f CO COO TP Tj- Tp CN IH NIOCD C- -P CM O O •-H X •

— cd cu X — — -* lO uo iri X ■^J' ,-H CO Tt" t- CN X iri f* CM CM 00 r-- iri cn CM iri OS CM CM ^U 0' -p iri OS OS CO CD t-^ CJ: W O M- CD CO Tf CO CD to ao co m t- m t— m m co m co os OS CO CO CO CO CO CO CO ^h O CM CO -o c 3 ep £ o CO C0 CN i-t CN CN -— CO DO 2 a^ as os ~ x: Os a-. OS X X t- t- t-~ as as as as as cc x as m as m m ^ x X m m m m 0 x x rp cm E iC m U0 lO lO 1(0 iO m in m m in in m t. m to if: in if? in m m m m m m m m m m m cm m m m m E cu 3 c ■i. c U E j3 u c CO c - CQ ~ So ~ _ ia (v* -C - „ s5ti boos * c c = - - - - 1 ill? - | E E--Z -& 2 c - - is O O Q - U K « « tf tab c JO. c CN -C c S. cu t- cd C D9 a* cd c ill i 5 j __ - to P T3" 1 Cd | bo cr: cu bo C £ "j ^5 egg 5 j= - cc cc C — ^ CU Cb CU r- III 2Si= ^ 6 6 ||J | ■e 3 3 -^ •a „ " - C C w ^ — w 8 J5 J5 <" § fc s OJ Ih 3 X E -a: CG c '3 a) — CO c DO CU IS u f z — 5 *c ^ > CM *CN -C ^ E cu a* — "3! CO s cc; 2i cu c a |1 o — C o Q. EZ o — cc: s s s ^S ■= - S = = ' It j;5 z 3 3 j: = u ■'. to *- cu - ; >, fl - Cu Cu rs: Cu O si * a: S « « co O y 0> tn tn — o — U W W r~ ^ ■S — -r -_; c w 9 | fg;5 3 5 c3 e^ f^ 6 •§ g c fc n -i -5 1 a z e t f jC 2 H 3- - U3D-D - CC CC q_ S i 1 0 t 0 cd co 0 J CC CC F cu 0- 0. 3 c _ 3 cc 0 UJ fg e2 CU c c G e2 CU h _c y f2 Van Waerebeek: Variation in skull morphology of Lagenorhynchus obscurus 761 Table 4 Results of two-sided r-tests and multifactor ana ysis of covarianee iMANCOVAl to determine sexual dimorphism and seasonal variation (see text) in skull measurements of mature female (At59l and male iA'=50l Lagenorhynchus obscu -us from Peru. Condylobasal length (CBL) is used as the covariate. Measurements are defined in the text. Variable Two-sided r-test MANCOVA Covariate (CBL) Factor '. (sex) t P /•"-ratio P F-ratio P CBL 0.63 0.53 — 0.37 0.55 RL 1.33 0.19 328 <0.0001 2.0 0.16 RWB 0.49 0.62 0.43 0.52 0.32 0.58 RW60 3.32 <0.002 12.3 <0.001 12.1 <0.001 RW1/4 3.62 <0.0005 4.87 <0.05 12.8 0.0005 RW1/2 3.58 <0.0001 12.3 <0.0001 13.1 0.0005 PMX1/2 3.59 <0.0001 0.29 0.59 10.4 0.002 RW3/4 4.2 <0.0001 8.72 <0.005 18.7 <0.0001 RTEN 0.93 0.36 201.5 <0.0001 0.001 0.98 RTIN 0.74 0.46 470.4 <0.0001 0.29 0.59 PROW 0.78 0.44 13.3 <0.0005 0.71 0.41 POOW 0.31 0.75 12.38 <0.001 0.04 0.84 ZYW 0.64 0.52 16.8 0.0001 0.62 0.44 PARW 1.17 0.24 2.41 .012 1.40 0.24 GWPMX 1.29 0.20 1.68 0.20 1.54 0.22 ENW 1.74 0.09 0.91 0.35 2.98 0.09 TFL 3.36 <0.005 7.35 <0.01 10.54 0.002 TFW 0.61 0.54 0.05 0.94 0.13 0.72 ORL 0.19 0.85 35.1 <0.0001 0.63 0.44 APL 1.26 0.21 2.70 0.10 1.54 0.22 INW 0.10 0.92 3.31 0.07 0.01 0.92 PTEL 1.39 0.17 30.5 <0.0001 1.1 0.30 UTRL 1.75 0.08 281.5 <0.0001 3.56 0.06 LTRL 1.10 0.27 135.1 <0.0001 0.26 0.61 RAL 0.42 0.67 386.2 <0.0001 0.54 0.47 RAH 0.06 0.95 13.4 <0.0005 0.07 0.79 TCUL 0.76 0.45 4.79 <0.05 1.41 0.24 TCUR 0.65 0.51 4.73 <0.05 0.95 0.34 TCLL 0.53 0.60 i.42 0.24 0.08 0.78 TCLR 0.54 0.60 1.32 0.25 0.09 0.77 TW 0.08 0.94 4.63 <0.05 0.36 0.55 BL 0.05 0.96 15.6 0.0002 0.003 0.96 BW 0.41 0.69 0.26 0.62 0.008 0.93 HBR 0.95 0.35 5.14 - <0.05 1.72 0.19 LBR 1.20 0.23 42.3 <0.0001 2.55 0.11 PEL 0.64 0.52 29.1 <0.0001 0.84 0.37 HCR 1.26 0.21 2.67 0.11 1.98 0.16 Southwest African versus New Zealand popu- lation Results of two-sided /-tests comparing adult skulls from southwest Africa and New Zealand are pre- sented in Table 6. Skulls from these regions do not differ in size, as indicated by condylobasal length (/=0.48, P=0.63) and zygomatic width i/=-0.074, P=0.94). Nev- ertheless, the two populations are strikingly divergent in dentition. Mean tooth width for New Zealand dusky dolphins (3.56 mm, SD=0.35, N=32) is highly signifi- cantly smaller (/=-7.14, P<0.0001) than in SW African dusky dolphins (4.18 mm, SD=0.37, N=36); the difference is visually obvious even from a quick inspection of skulls. In addition, New Zealand crania have on average 1.5-1.8 teeth more in each row than African skulls (two-sided /-tests, P<0.001). Maximum tooth counts in the four rows range 32-34 (SA) and 36-39 (NZ). Finally, SW African crania have a significantly lower supraoccipital crest (/=3.43, P=0.005), slightly wider external nares and temporal fossa and, possibly, a somewhat higher braincase (0.050.05. Variable ANOVA ANCOVA F-ratio df P F-ratio P CBL 120.2 3,275 <0.0001 — — RL 118.3 3,274 <0.0001 2.48 NS RWB 10.1 3,287 <0.0001 0.66 NS RTEN 63.5 3,261 <0.0001 0.83 NS RTIN 82.4 3,245 <0.0001 1.05 NS PROW 40.0 3,273 <0.0001 2.46 NS POOW 44.6 3,264 <0.0001 3.78 <0.02 ZYW 54.5 3,276 <0.0001 4.14 <0.01 PARW 6.88 3,282 <0.0005 0.87 NS GWPMX 4.39 3,294 <0.005 1.99 NS ENW 5.35 3,293 <0.0005 1.66 NS TFW 13.2 3,285 <0.0001 3.65 <0.02 ORL 18.0 3,280 <0.0001 3.31 <0.05 APL 22.9 3,280 <0.0001 3.38 <0.05 INW 34.1 3,282 <0.0001 4.82 <0.005 PTEL 33.5 3,242 <0.0001 2.89 <0.05 UTRL 122.0 3,270 <0.0001 4.33 <0.01 LTRL 109.3 3,242 <0.001 8.92 <0.001 RAL 118.3 3,239 <0.0001 6.18 <0.001 RAH 34.9 3,232 <0.0001 0.14 NS TCUL 7.4 3,256 <0.0005 6.68 <0.0005 TCUR 9.65 3,256 <0.0001 7.74 <0.0005 TCLL 11.7 3,236 <0.0001 8.85 <0.0001 TCLR 10.1 3,237 <0.0001 8.25 <0.0001 TW 50.3 3,260 <0.0001 25.2 <0.0001 BL 13.9 3,208 <0.0001 0.35 NS BW 4.1 3,117 <0.02 0.70 NS HBR 19.5 3,285 <0.0001 4.59 <0.005 LBR 28.4 3,285 <0.0001 1.08 NS PEL 0.24 3,201 NS 2.31 NS HCR 6.33 3,235 <0.0005 3.47 NS Peruvian and Chilean crania, explainable by the rela- tively greater morphological similarity and geographic proximity. Overall there were successfully identified to geographical population 164 out of 188 (87.29c) of the skulls, increasing to 179 out of 188 (95.2%) when skulls from Chile and Peru were pooled. Six characters with high discriminatory power were identified from the standardized discriminant function coefficients, /-tests and ANOVA's and an additional dis- criminant analysis was performed using those vari- ables (Table 8). Classification results and the discrimi- nant function plot (Fig. 4) demonstrate that specimen prediction remains fairly effective (143 of 188 or 76.1% and 171 out of 188 or 91.0% when the SE Pacific sample is pooled as one group). Non-metrical characters Tooth counts — As expected, tooth counts derived from skulls were consistently greater (P<0.0001) than counts made on heads in the flesh because small, apical teeth frequently do not erupt from the gums (see Van Waerebeek, 1992b). Differences of mean tooth counts between skull and fresh specimens varied from 2.29 to 2.67 per jaw ar- cade. Considering confidence intervals, the best estimate (closest integer) for each of the four ar- cades was two more teeth visible in skulls than in heads in the flesh. Tooth counts are bilaterally symmetrical in up- per and lower jaws and for all sample pairs it- tests, P>0.16). The upper jaw has significantly more teeth than the lower jaw, (one-sided /-tests; P<0.005 in all cases); confidence intervals in- cluded one (tooth) as the largest integer but not zero. Sexual dimorphism — The 28 non-metrical cranial characters (NMC) tested are, without exception, independent of sex (chi-square contingency tests, P>0.05). Sexual dimorphism neither exists in the number of teeth in Peruvian dusky dolphins, both as counted from skulls and from fresh animals (/-tests, P>0.25). It is considered safe to assume that this is generally true for the other popula- tions of L. obscurus and therefore the sexual fac- tor is disregarded in the subsequent analyses. Developmental variation — Twenty-two of the 28 NMC characters tested (78.6%) in Peruvian dusky dolphins are independent of maturity status (Table 9). Variables which may be correlated with age to some degree (chi-square, P<0.05) include position of the left premaxillary fora- men at the base of the prenarial triangle (1L), presence of fenestration* s) in occipitals (18L) or basisphenoid (30), visibility of palatine spine (25), and the number of mental foramina (33L). At least one of these is expected to be due to chance fluctuation (a=0.05). The more frequent occurrence of a notch in the upper margin of the foramen magnum (17) in im- mature animals as compared to adults proved highly significant (chi-square=8.83, P<0.005). Geographic variation — The six growth dependent char- acters were removed from the data set before varia- tion between populations was tested. By doing this, samples available for geographic variation analysis were greatly enhanced as juvenile specimens (but not neonates) could be included. Table 10 lists absolute Van Waerebeek: Variation in skull morphology of Lagenorhynchus obscurus 763 488 I 398 388 - 378 1 1 ' 368 1 . [ 338 228 I 218 ■ 288 I I • 198 - CH NZ PE 5m Condylobaeal length 183 . 181 I : 99 T 97 - 95 - CH N2 PE SA Rostrum length CH NZ PE SA Rostrum width at baa 278 268 I 258 - 248 1 I ; 238 1 278 " - 268 I 258 248 238 I i ; CH NZ PE SA CH NZ PE SA Roatrum tip to external nares Rostrum tip to internal nares 168 165 - I 162 - 159 I I ■ 156 - 153 158 1 CH NZ PE SA Praorbital width 188 - ' 185 . I 182 " 179 • 176 * I f ■ 173 178 - 198 186 I 182 178 l I ' 174 178 '. CH NZ PE SA Poatorbital width 168 - . 156 - 164 162 ' 168 ■ 148 CH NZ PE SA Zygomatic width CH NZ PE SA Pariatal width 89 : 79 - 78 . 77 - I T - 76 76 1 . 74 56 ' SB - f I ' J 54 - - 53 • 63 CH NZ PE SA Grsataet width premaxillary CH NZ PE SA External nares width I " I - 1 CH NZ PE SA Temporal fona width Figure 2 Comparison of means in cranial measurements (in mm) and meristics for adult Lagenorhynchus obscurus from Chile (CHl. New Zealand (NZl, Peru (PEi and southwestern Africa (SA). Error bars are 951* confidence intervals for means. frequencies of NMC character states in skulls from Peru (iV=221), Chile (iV=40), New Zealand (iV=69), and southwestern Africa (Af=65). Surprisingly, only three characters out of 22 ( 14% ) were found to demonstrate significant differences between geographic groups (chi- square contingency analysis; see Table 10). These are relative position of the foramen in the right premaxil- lary at the base of the prenarial triangle (variable 1R, 764 Fishery Bulletin 91(4). 1993 54 : 5.3 - - 52 L I SI ■ SI ■ .19 ■ 39 38 I 37 - 38 36 - 34 : 33 : 59 " 57 I 55 . 53 - - 51 - - . CH NZ PE SA Orbit length CH NZ PE SA Antorbitel procitt length CH NZ PE SA Internal mm width 66 - - 84 I as - - 88 56 " - 56 288 - 198 ■ 1 I : 188 - 178 I I : 168 . 288 198 I ■ 188 : 178 I I . 188 . CH NZ PE SA Pterygoid length CH NZ PE SA Upper tooth row length CH NZ PE SA Lower tooth row length 348 - T I 338 • 328 318 I I ■ 388 - CH NZ PE SA Remus length 72 " 7e I 68 66 j " 64 1 - 33 3a I 31 r 3t> ■ 29 - CH NZ PE SA Remus height CH NZ PE SA Tooth count upper lert 33 " 32 I 31 " 38 . ^ 29 ~ CH NZ PE SA Tooth count upper right 3 3 4 32 - 31 I 36 29 - 28 - 33 " 35 I 31 - 3 8 29 - CH NZ PE SA Tooth count low«r l«ft Figure 2 (Continued) CH NZ PE 3A Tooth count uppar left P<0.0001), presence of an accessory foramen above the foramen magnum (variable 16, P=0.02), and the de- gree of contact, posteriorly, between pterygoid hamuli (variable 23, P=0.03). Upon comparison of frequencies for variables 16 and 23, two similarity pairs became apparent: New Zealand/southwestern Africa and Peru/Chile. Within each subgroup the monomorphic status for both char- acters was confirmed (chi-square, df=l, P>0.41 in four tests). Southwest African L. obscurus differed very sig- nificantly in character 1R from all other stocks. By removing the African subsample, frequencies for 1R in the three remaining geographic units became homoge- neous (chi-square=0.58; df=4, P=0.96). Van Waerebeek Variation in skull morphology of Lagenorhynchus obscurus 765 4.6 4.4 4.2 4 1 ' 1 ■ 3.8 3.6 3.4 : I \ 36 36.5 i ; 35 34.5 34 : 33.5 33 19.9 " 19.7 - 19.5 - 19.3 . 19. 1 ■ 18.9 - CH NZ PE SA Tooth width CH NZ PE SA Bulla l«tngth NZ PE SA Bulla width 113 : 111 I 199 T 187 166 ■ I 183 128 126 124 CH NZ PE Sfl Height braincne 12 - 18 " 8 I I ■ 6 I 1 4 - CH NZ PE SO Length braincaae Figure 2 (Continued) CH NZ PE 3A Height aupraoecipital craat Discussion Advanced fusion in the frontal-supraoccipital suture (FROC) is the most reliable single cranial criterion of sexual maturity in the dusky dolphin (95% effi- ciency). In combination with fusion=2 in LAC and ZYG cranial sutures, it can select sexually adult specimens from an undocumented series of L. obscurus skulls with even a higher degree of cer- tainty. Distal fusion of premaxillary and maxillary bones has served as the indicator of choice for sexual maturity in Stenella spp. (Dailey and Perrin, 1973; Perrin, 1975a; Schnell et al., 1982s; Douglas et al., 1984, 1986; Perrin et al., 1981, 1987, 1989), common dolphin (Evans, 1975, 1982), and bottlenose dolphin (Walker, 19817; Mead and Potter, 1990; Van Waerebeek et al., 1990). Its reliability and useful- ness however depends on the species and the re- quirements of selection; Mead and Potter (1990) for instance reported that some immature Tursiops skulls were erroneously assigned to the adult group. Almost half of sexually mature dusky dolphins did not show any rostral distal fusion. The character reportedly was of no use in physically mature Pa- cific white-sided dolphins Lagenorhynchus obliquidens (Walker et al., 1986). Classification of specimens with heavily eroded sutures from weath- ering, boiling or chemical treatment is difficult, and if available sample size permits, these skulls should probably not be used. Moderate heating applied over short periods, to soften tissues for instance, inflicts little damage. Skulls of L. obscurus exhibit little sexual dimor- phism, therefore it was not possible to define usable classification functions for sexing specimens. Of 28 non-metrical cranial characters, none exhibited di- morphism. In addition, the number of dimorphic met- rical characters (6 of 37, 16.2%) was remarkably low in comparison with that in other small delphinids such as the pantropical spotted dolphin Stenella attenuata (23 of 36, 63.9%; Schnell et al., 1985) and the spinner dolphin Stenella longirostris ( 13 of 36, 36.1%; Douglas et al., 1986). L. obscurus and the named Stenella spp. share the characteristic of males having a broader rostrum and a longer temporal fossa than females, while the overall skull size is not dif- ferent between sexes (see Schnell et al., 1985; Dou- glas et al., 1986). Intersexual differences in these species may be governed, at least partly, by similar selection factors. A number of authors have invoked common (genetic) ancestry between L. obscurus and the Stenella species group to explain certain mor- phological similarities (True, 1889; Fraser, 1966; Fraser and Purves, 1960; Mitchell, 1970). However, I concur with the opinion of W. F. Perrin1* that con- 'W. F. Perrin, Natl. Mar. Fish. Serv. Pers. commun. March 1992. 766 Fishery Bulletin 91(4), 1993 Table 6 Geographic variation in skull measurements and meristics of Lagenorhvn chus obscurus from Peru (PE, N=IS9) Chile (CH, N=22), New Zealand (NZ, N=47) and southwestern Africa (SA, N=40). Variable PE versus CH NZ versus SA t P t P CBL -0.48 0.40 0.48 0.63 RL 0.03 0.98 -0.26 0.80 RWB -0.67 0.50 -0.45 0.65 RTEN -1.12 0.27 1.6 0.11 RTIN -0.34 0.73 1.32 0.19 PROW 2.50 <0.05 -0.094 0.93 POOW -0.41 0.68 0.29 0.77 ZYW -0.53 0.60 -0.074 0.94 PARW 0.84 0.40 -0.52 0.60 GWPMX 0.69 0.49 -0.14 0.89 ENW -1.33 0.19 -2.02 <0.05 TFW -0.11 0.91 -2.18 <0.05 ORL 1.03 0.30 -2.38 <0.05 APL -2.89 <0.005 1.11 0.27 INW -1.07 0.29 2.32 <0.05 PTEL -2.23 <0.05 -1.28 0.21 UTRL -0.17 0.87 -0.56 0.58 LTRL -3.29 <0.005 -0.86 0.39 RAL -2.10 <0.05 -0.97 0.36 RAH -0.50 0.62 0.41 0.68 TCUL -2.71 <0.01 3.45 <0.001 UR -2.70 <0.01 4.25 <0.0001 TCLL -3.01 <0.005 4.24 <0.0001 TCLR -2.76 <0.01 3.95 <0.0002 TW -1.05 0.30 -7.14 <0.0001 BL -1.32 0.19 0.33 0.74 BW — — -1.12 0.27 HBR 0.09 0.93 -2.09 <0.05 LBR -1.37 0.17 -1.02 0.31 PEL 0.52 0.61 1.0 0.33 HCR — — 3.43 <0.005 vergence or parallellism is a more likely cause of the cited resemblances. The near absence of sexual dimorphism in the skull and external features of the dusky dolphin, coupled to the exceptionally large relative testis size, led me to suggest that sperm competition should occur in this species (Van Waerebeek, 1992b). Both bivariate and multivariate analysis provide evi- dence of extensive geographic variation in skull mor- phology among the dusky dolphin stocks studied. Most strikingly, adult skulls from the Southeast Pacific stock are considerably larger than either New Zealand or SW African specimens (difference in weighted CBL means is 31mm or 8.5% of CBL); skulls from New Zealand do not differ significantly in size from those of SW Africa. This is corroborated by the greater mean body length (roughly 10cm) at sexual maturity and asymptotic length in Peruvian dolphins than in SW African and New Zealand specimens (Van Waerebeek, 1992b). Two non-metrical cranial characters also sup- port the existence of the Peru/Chile and New Zealand/ Table 7 Statistics for discriminant analysis utilizing 21 cranial charac- ters of adult Lagenorhynchus obscurus skulls from Chile (CH, N=7), New Zealand (NZ, JV"=24), southwestern Africa (SA, N=22) and Peru (PE, JV=135l. Group numbers (under centroidsl refer to these used in the discriminant function plot (Fig.3). SEP= Southeast Pacific (Chile and Peru combined). Discriminant Relative Canonical function Eigenvalue percentage correlation 1 2.202 76.02 0.829 2 0.523 18.06 0.586 3 0.171 5.91 0.382 Unstandardized discriminant function coefficients 1 2 3 CBL 0.0096 -0.0466 -0.0730 RL 0.0134 0.0025 0.0672 RWB 0.0038 -0.0618 -0.0057 RTEN -0.0074 -0.0209 0.0268 PROW -0.0285 -0.0416 -0.0647 POOW 0.0191 0.0528 -0.1291 ZYW 0.0152 -0.0285 0.1545 GWPMX -0.0204 0.0431 0.0021 ENW -0.0031 0.1414 -0.0817 TFW 0.0635 0.0256 -0.0028 ORL -0.0190 0.0361 0.1625 APL 0.0683 -0.0198 -0.0739 INW 0.0481 -0.1066 0.0586 UTRL 0.0209 -0.0352 0.1097 LTRL 0.0664 0.0670 -0.1218 RAL 0.0067 0.0118 0.0047 TCUL 0.0195 -0.0537 -0.0424 TCUR -0.1049 -0.1707 -0.1628 TW 0.6215 1.7912 0.6270 HBR -0.0044 0.0669 -0.0467 LBR 0.0276 0.0058 0.0239 CONSTANT -33.9754 5.8979 15.3204 Classification results Actual group Predicted group (percentage) CH NZ SA PE SEP CH 85.7 0 0 14.3 100 NZ 0 91.7 8.33 0 0 SA 0 4.6 90.9 4.6 4.6 PE 10.4 1.5 2.2 85.9 96.3 Group eentrou Is Discr minant functions 1 2 3 CH(=1) 0.398 -0.705 2.040 NZ(=2) -2.773 -1.254 -0.182 SA(=3) -2.345 1.582 0.133 PE(=4) 0.855 0.0016 -0.095 Van Waerebeek Variation in skull morphology of Lagenorhynchus obscurus 767 u c J -< a 4.2 3.8 3.4 3 2.6 2.2 1.8 F 1.4 ii 8.6 fr- e.2 -e.2 -8.6 -1 -1.4 -1.8 -2.2 -2.6 -3 -3.4 -3.8 -4.2 F 4 4 J_ -4.8 "3.4 -2 -8.6 6.8 Discriminant function 1 (score) 2.2 3.6 Figure 3 Scores of discriminant functions I and II based on 21 cranial characters for adult skulls of Lagenorhynchus obscurus from Chile (II, New Zealand (21, southwestern Africa (3), and Peru (4). Group centroids are indicated by a black dot. Corresponding statistics are presented in Table 7. □ U c 3 L U * H D 4.4 4 3.6 3.2 F 2.8 F- 2.4 E 2 1.6 1.2 8.8 8.4 8 -e.4 -8.8 -1.2 -1.6 -2 -2.4 4 * A 4* «*■ * 4* * 4* * " » . -4.7 -3.7 -2.7 -1.7 -9.7 8.3 1.3 2.3 3.3 Discriminant function 1 (icon) Figure 4 Scores of discriminant functions I and II based on the six most discriminating cranial characters (CBL, PROW, TFW, LTRL, TW, and TCUR) for adult skulls of Lagenorhynchus obscurus from Chile (1), New Zealand (2), southwestern Africa (3), Peru (4). Group centroids are indicated by a black dot. Corresponding statistics are presented in Table 8. 768 Fishery Bulletin 91(4), 1993 Table 8 Statistics of discriminant analysis utilizing a minimum num- ber (6) of sexually monomorphic cranial variables with maxi- mum discriminating power. Samples include adult skulls of Lagenorhynchus obscurus from Chile (CH, N=7), New Zealand (NZ, N =241. southwestern Africa ISA, iV=22) and Peru (PE, AT=135). Discriminant Relative Canonical function Eigenvalue percentage correlation 1 2.0188 84.35 0.818 2 0.3096 12.94 0.486 3 0.0649 2.71 0.247 Unstandardu ed discriminant function coefficients 1 2 3 CBL 0.0350 -0.0588 -0.1105 PROW 0.0353 -0.0363 0.1205 TFW 0.0650 0.0484 0.0601 LTRL 0.0663 0.0449 0.1227 TW 0.7749 2.0950 -0.5369 TCUR -0.0582 -0.3046 0.1422 CONSTANT -36.6466 18.7270 -5.0552 Classification results Predicted group (percentage) Actual group CH NZ SA PE SEP CH 71.4 0 0 28.6 100 NZ 0 87.5 12.5 0 0 SA 9.1 18.2 72.7 0 9.1 PE 19.3 2.2 3.7 74.8 94.6 Group centro ids Discriminant functions 1 2 3 CH(1) 0.301 0.0355 -1.281 NZ(2) -2.674 -0.986 0.0236 SAO) -2.218 1.235 0.0400 PE(4) 0.821 -0.0278 0.0557 SW Africa similarity pairs. The New Zealand form is set apart from all others by minimal tooth size and is further distinguished from the SW African form by the significantly higher number of teeth. A commonly used criterion of subspecific level dif- ference between populations is whether or not 90% of the specimens can be unequivocally assigned to the proper group (Mayr, 1970). Perrin (1975a, 1990 1. for instance, described three subspecies of spinner dol- phin, Stenella longirostris, based on this principle. Discriminant analysis allows correct identification of a group (96.5%, 91.7%, and 90.9%) of L. obscurus skulls from respectively the Southeast Pacific, New Zealand, and SW Africa, which warrants allocation of subspecific status. It seems premature, however, to formally describe subspecies and assign trinomial names, because no skulls from Argentina have been studied and the Chilean sample was too small to per- mit definitive conclusions on the degree of heteroge- neity in the Southeast Pacific group. In particular, no skulls have been examined from dusky dolphins oc- curring off southern Chile (south of 46°S), which are thought to belong to the Argentinian stock (Van Waerebeek, 1992a). Considering the persisting mor- tality levels in gillnet and harpoon fisheries off Peru and northern Chile and in purse-seine operations off Argentina (Van Waerebeek and Guerra, 1986s; Guerra et al. 1987; Crespo and Corcuera, 1990"; Van Waerebeek and Reyes, 1993; Van Waerebeek, unpubl. data), the question of separate subspecies or popula- tion status is of major importance with respect to conservation. The MANOVA results, suggesting some seasonal morphological variation in L. obscurus caught off cen- tral Peru, were not confirmed by other tests. However, the picture may be blurred by multi-year pooling or other sampling bias. High-resolution molecular tech- niques, such as DNA-fingerprinting, should preferably be used to explore this issue. Although results were complementary, the resolv- ing power of non-metrical cranial characters (not in- cluding tooth counts) was inferior to that of measure- ments (i.e., fewer characters differed); therefore non-metrical characters are not recommended as sole method for subspecific discrimination. Based on stud- ies of the common dolphin, Delphinus delphis, Perrin et al. (1989) independently concluded that non-metri- cal characters be used only in combination with met- rical characters and that special attention be paid to controlling conditions to maximize comparability of data sets. Although little information is available on parasite prevalences in dusky dolphins other than these from Peru, preliminary indications of differential infesta- tions add weight to the morphological findings. Crassicaudid nematodes were practically absent in the cranial sinuses of many hundreds of Peruvian dusky dolphins but were retrieved from the cranial sinuses of one out of two specimens from SW Africa. "Crespo, E. A., and J. F. Corcuera. 1990. Interactions between ma- rine mammals and fisheries in some fishing areas of the coast of Argentina and Uruguay. Doc. SC/090/G2 presented to the IWC, La Jolla, CA, 20-21 October 1990. Van Waerebeek Variation in skull morphology of Lagenorhynchus obscurus 769 Table 9 Frequency distributions for selected non-metrical character states (coded 1-6, see Perrin ?t al.. 1982) in 173 mature (MAT) and 48 immature (IMMl skulls of L. obscurus (sexes pooled) from central Peru. Chi-square contingency analyses indicate levels of developmental variation. For df=l a Yates' corrected chi-square was computed and the higher of the two values (lower Pi tabulated . Characters are explained in Table 1. Non-metrical character states X2 P df 1 2 3 4 5 6 1L MAT IMM 72 14 61 28 40 6 8.5 0.014 2 1R MAT IMM 56 22 17 7 99 19 4.89 0.09 2 2L MAT IMM 5 0 167 48 1 0 1.71 0.43 2 2R MAT IMM 9 0 162 48 2 0 3.21 0.20 2 3 MAT IMM 45 6 83 23 45 19 5.36 0.07 2 5L MAT IMM 5 1 166 46 0.09 0.77 1 5R MAT IMM 2 1 1 0 168 46 0.52 0.77 2 6L MAT IMM 147 42 22 6 3 0 1 0 1.14 0.77 3 6R MAT IMM 162 47 10 1 1 0 1.38 0.50 2 9L MAT IMM 2 0 171 48 0.56 0.45 1 9R MAT IMM 6 0 166 48 1.72 0.19 1 10 MAT IMM 172 48 — — — 12L MAT 173 12R MAT IMM 172 48 — — — 13 MAT IMM 1 0 165 48 1 0 0.58 0.75 2 16 MAT 3 170 0.84 0.36 1 17 MAT IMM 77 33 96 15 8.83 0.003 1 18L MAT IMM 35 8 2 4 133 36 7.3 0.026 2 18R MAT IMM 41 11 3 2 126 35 0.97 0.61 2 19 MAT IMM - 2 0 169 48 0.57 0.45 23 MAT IMM 149 38 5 3 1.36 0.24 24 MAT IMM 123 38 34 4 3.15 0.08 25 MAT IMM 118 39 38 3 5.97 0.01 30 MAT IMM 9 0 158 43 0 1 6.20 0.04 33L MAT 7 48 60 22 3 11.2 0.02 IMM 0 24 12 2 1 33R MAT 1 12 52 52 18 4 4.57 0.47 5 IMM 0 1 17 13 8 0 34L MAT IMM 115 35 23 4 1 0 1.25 0.53 2 34R MAT IMM 118 36 21 0 0 1.43 0.23 1 770 Fishery Bulletin 91(4). 1993 Table 10 Geographic variation in non -metrical cranial (NMC ) characters for 395 skulls of Lagenorhynchus obscurus from Peru (PE.JV =221), Chile (CH N = =40), New Zealand (NZ, N =69), and southwestern Africa (SA, N =65). Frequency distribu- tions of NMC states (coded '. -6, see Perrin et al. 1982) for each region and results of chi-square contingency analyses are presented. Samples include all sex/maturity groups except neonates characters known to be maturity dependent (see Table 91 have been exclu ded. NMC REG 1 2 3 4 5 6 x2 P df 1R PE CH NZ SA 78 10 21 24 24 4 6 22 118 16 37 15 31.8 <0.0001 6 2L PE CH NZ SA 5 0 1 0 215 30 62 61 1 0 0 0 2.8 0.83 6 2R PE CH NZ SA 9 0 1 0 210 30 62 61 2 0 0 0 5.9 0.43 6 3 PE CH NZ SA 51 7 17 11 106 17 33 27 64 6 13 23 5.9 0.43 6 5L PE CH NZ SA 6 1 0 3 0 0 0 1 212 26 62 55 8.33 0.21 6 5R PE CH NZ SA 3 1 0 3 1 1 0 0 214 25 62 57 10.8 0.10 6 6L PE 189 28 3 1 7.9 0.54 9 CH 23 5 0 1 NZ 51 12 0 0 SA 49 10 0 1 6R PE CH NZ SA 209 27 59 57 11 2 3 3 1 0 0 0 0.89 0.99 6 9L PE CH NZ SA 2 1 0 0 219 26 63 60 3.85 0.28 3 9R PE CH NZ SA 6 3 2 3 214 24 60 57 4.94 0.18 3 10 PE CH NZ SA 220 28 64 60 12L PE CH NZ SA 221 27 61 57 0 0 1 2 7.38 0.06 3 12R PE CH NZ SA 220 28 62 57 0 0 1 2 7.38 0.06 3 Van Waerebeek: Variation in skull morphology of Lagenorhynchus obscurus 771 Table 1 0 (Continued) NMC REG 1 2 3 4 5 6 x2 P df 13 PE CH NZ SA 1 0 1 0 213 25 62 59 1 0 0 0 2.41 0.88 6 16 PE CH NZ SA 3 0 3 5 218 21 59 53 9.59 0.02 3 18R PE CH NZ SA 52 2 12 7 5 1 0 0 0 0 0 1 161 15 47 48 7.06 0.32 6 19 PE CH NZ SA 0 0 0 1 2 0 0 0 217 16 60 55 6.48 0.37 6 23 PE CH NZ SA 187 12 41 41 8 1 6 7 8.83 0.03 3 24 PE CH NZ SA 161 12 34 34 38 2 13 14 3.8 0.28 3 33R PE 1 13 69 65 26 4 7.24 0.95 15 CH 0 0 7 4 1 0 NZ 0 5 16 16 5 1 SA 0 5 24 18 3 1 34L PE CH NZ SA 150 5 36 41 27 5 8 9 1 0 0 0 8.59 0.20 6 34R PE CH NZ SA 154 10 36 47 24 3 8 5 0 0 0 0 2.41 0.49 3 Cyamid amphipods have never been found on L. obscurus from Peru (Van Waerebeek, 1992b), whereas in New Zealand they seem to be fairly common (Cipriano, 1985). The actual distribution of L. obscurus populations is generally comparable to that of the four species of Cephalorhynchus, and the geographical dispersion may have similarly evolved under the influence of the eastflowing Westwind Drift, as proposed for Cephalorhynchus spp. by van Bree ( 1986) and Robineau (1989). The close affinity of the large-bodied Peruvian dusky dolphin with the Pacific white-sided dolphin (Webber, 1987; Van Waerebeek, 1992b) suggests that the former probably represents the most ancestral (plesiomorph) form of dusky dolphin from which the Argentinian, and small-bodied SW African and New Zealand forms were derived. Acknowledgments I am especially grateful to Julio Reyes, Peter B. Best. Carlos Guerra, and Walter Sielfeld for their kind per- mission to examine skulls they themselves collected. Special thanks are due also to Alan Baker, Peter van Bree, J. Darby, Ewan Fordyce, James G. Mead, Mar- tin Sheldrick, R. Thompson, and G. Tunnicliffe for free access to collections under their care, and to J.C. Reyes, Mark Chandler, Laura Chavez, Anne-Catherine 772 Fishery Bulletin 91(4). 1993 Lescrauwaet, Tony Luscombe, and Andrew Read for assistance with field work in Peru. Peter van Bree, Robert Clarke, Ronald Hardy, Linda Jones, William Perrin, and Jan Stock provided comments which greatly improved the manuscript. Research and travel funds were gratefully received from the King Leopold III Fund for Nature Research and Conservation, the Whale and Dolphin Conservation Society, Van Tienhoven Foundation, IUCN/UNEP, Greenpeace In- ternational, Cetacean Society International, Rotary Club Oudenaarde and Rotary International District 162. The Statgraphics program was donated by Steve Dennison. Literature cited Anonymous. 1975. Report of the meeting on smaller cetaceans; Montreal, 1-11 April, 1974. J. Fish. Res. Board Can. 32(7):889-983. Baker, A. N. 1983. Whales and dolphins of New Zealand and Australia. Victoria Univ. Press, 133 p. Borobia, M., and D. Sergeant. 1989. Variation in skull morphology of South Ameri- can dolphins of the Genus Sotalia. Abstract, Fifth Int. Theriological Congress, Rome, 22-29 August 1989: 480. Bree, P. J. H. van. 1986. Over soortvorming bij walvisachtigen (Cetacea). Natuurkundige Voordrachten 64:91-95. Brownell, R. L., Jr. 1965. A record of the dusky dolphin, Lagenorhynchus obscurus, from New Zealand. Norsk Hvalfangst- Tidende 8:169-71. 1974. Small odontocetes of the Antarctic. In VC. Bushnell (ed.). Antarctic mammals, p. 13-19. Ant- arctic Map Folio Series. American Geographical So- ciety, Folio 18, 199 p. Cipriano, F. 1985. Dusky dolphin research at Kaikoura, New Zealand: a progress report. Mauri Ora 12:151-158. Dailey, M. D., and W. F. Perrin. 1973. Helminth parasites of porpoises of the genus Stenella in the eastern tropical Pacific, with descrip- tions of two new species: Mastigonema stenellae gen. et sp.n. (Nematoda: Spiruroidea) and Zalophotrema pacificum sp.n. iTrematoda: Digenea). Fish. Bull. 71:455-71. Douglas, M. E., G. D. Schnell, and D. J. Hough. 1984. Differentiation between inshore and offshore spotted dolphins in the eastern tropical Pacific Ocean. J. Mammalogy 65(31:375-87. 1986. Variation in spinner dolphins (Stenella longirostris) from the eastern tropical Pacific Ocean: sexual dimorphism in cranial morphology. J. Mam- mal. 67(31:537-44. Evans, W. E. 1975. Distribution, differentiation of populations, and other aspects of the natural history of Delphinus delphis Linnaeus in the Northeastern Pacific. Ph.D. diss., Univ. Calif. Los Angeles, 145 p. 1982. Distribution and differentiation of stock of Del- phinus delphis Linnaeus in the northeastern Pa- cific. Mammals in the Seas. FAO Fish. Ser. 5(41:45-66. Fraser, F. C. 1966. Comments on the Delphinoidea. In K. Norris (ed.), Whales, dolphins and porpoises, p. 7-31. Univ. California Press, Berkeley. Fraser, F. C, P. E. Purves. 1960. Hearing in cetaceans: evolution of the accessory air sacs and the structure and function of the outer and middle ear in recent cetaceans. Bull. Brit. Mus. (Nat. Hist.), Zoology, 7(11:1-140. Gaskin, D. E. 1968. Distribution of Delphinidae (Cetacea) in rela- tion to sea surface temperatures off eastern and south- ern New Zealand. N.Z. J. Mar. Freshwater Res. 2:527-34. Gray, J. E. 1828. Spicilegia zoologica. Treutel: London. Hershkovitz, P. 1966. Catalog of living whales. Smithsonian Institu- tion, Washington D.C., 259 p. Guerra, C, K. Van Waerebeek, G. Portflitt, and G. Luna. 1987. Presencia de cetaceos frente a la segunda region de Chile. Estudios Oceanologicos 6:87-96. Kinze, C. C. 1985. Intraspecific variation in Baltic and North sea harbour porpoises (Phocoena phoeoena L.1758). Vidensk. Meddr. dansk naturh. Foren. 146:63-74. Leatherwood, S., and R. R. Reeves. 1983. The Sierra Club handbook of whales and dolphins. Sierra Club Books, San Francisco, 301 p. Leatherwood, S., J. S. Grove, and A. E. Zuckerman. 1991. Dolphins of the genus Lagenorhynchus in the tropical South Pacific. Mar. Mammal Sci. 7(21:194- 197. Mayr, E. 1970. Populations, species and evolution. The Belknap Press of Harvard Univ. Press, Cambridge, MA, 453 p. Mead, J. G., and C. W. Potter. 1990. Natural history of bottlenose dolphins along the central Atlantic coast of the United States. In S. Leatherwood and R. R. Reeves (eds.l, The bottlenose dolphin. Academic Press, San Diego, 653 p. Mikkelsen, A. M. H. 1991. Population studies of the two North Atlantic dolphin species, Lagenorhynchus albirostris and L.acutus through multivariate data analyses of cra- nial characters, and a comparative study of diagnos- Van Waerebeek Variation in skull morphology of Lagenorhynchus obscurus 773 tic skull character differences between these species and Peponocephala electra. M.S. thesis, Univ. Copenhagen. Mitchell, E. D. 1970. Pigmentation pattern evolution in delphinid ce- taceans: an essay in adaptive coloration. Can. J. Zoology, 48(41:717-40. Morzer Bruyns, W. F. J. 1971. Field guide of whales and dolphins. Uitgeverij Tor, Amsterdam, 258 p. Perrin, W. F. 1990. Subspecies of Stenella longirostris (Mammalia: Cetacea: Delphinidae). Proc. Biol. Soc. Wash. 103(2):453-463. 1975a. Variation of spotted and spinner porpoise (ge- nus Stenella) in the eastern tropical Pacific and Hawaii. Bull. Scripps Inst. Oceanogr. 21:206 p. 1975b. Distribution and differentiation of popula- tions of dolphins of the genus Stenella in the east- ern tropical Pacific. J. Fish. Res. Board Can. 32:1059-1067. Pen-in, W.F., P. A. Sloan, and J. R. Henderson. 1979. Taxonomic status of the 'Southwestern Stocks' of spinner dolphin Stenella longirostris and spotted dolphin S. attenuata. Rep. Int. Whal. Comm. 29: 175-184. Perrin, W. F., E. D. Mitchell, J. G. Mead, D. K. Caldwell, and P. J. H. van Bree. 1981. Stenella elymene, a rediscovered tropical dolphin of the Atlantic' J. Mamm. 62(31:583-98. Perrin, W. F., E. D. Mitchell, J. G. Mead, D. K. Caldwell, M. C. Caldwell, P. J. H. van Bree, and W. H. Dawbin. 1987. Revision of the spotted dolphins, Stenella spp. Mar. Mammal Sci. 3(2):99-170. Perrin, W. F., A. V. Yablokov, M. V. Mina, and J. Barlow. 1989. Comparison of the resolving power of metrical and non-metrical cranial characters in defining geo- graphic populations of the common dolphin, Delphi- nus delphis. Abstracts, Fifth Int. Theriological Congress, Rome, August 1989:493. Read, A. J., K. Van Waerebeek, J. C. Reyes, J. S. McKinnon, and L. C. Lehman. 1988. The exploitation of small cetaceans in coastal Peru. Biol. Conserv. 46:53-70. Robineau, D. 1989. Relationships among the species of Cephalorhynchus (Cetacea, Delphinidae). Abstract, Fifth Int. Theriological Congress, Rome, 1989:494. 1991. Les cetaces des lies Kerguelen. Mammalia 53(2):265-278. Rose, B., and A. I. L. Payne. 1991. Occurrence and behavior of the southern right whale dolphin Lissodelphis peronii off Namibia. Mar. Mammal Sci. 7(11:25-34. Ross, G. J. B. 1977. The taxonomy of bottlenose dolphins, Tursiops species, in South African waters. Ann. Cape Prov. Mus. (Nat. Hist.) 11:135-94. Schnell, G. D., M. E. Douglas, and D. J. Hough. 1985. Sexual dimorphism in spotted dolphins (Stenella attenuata) in the eastern tropical Pacific Ocean. Mar. Mammal Sci. 1:1-14. Sergeant, D. E. 1962. The biology of the pilot or pothead whale Globicephala melaena (Trail) in Newfoundland waters. Fish. Res. Board. Can., Bull. 132:1-84. Sjovold, T. 1977. Non-metric divergence between skeletal populations. Ossa 4, supplement 1:1-133. STSC, Inc. 1989. Statgraphics 4.2 Statistical Graphics System. Statistical Graphics Corp., Rockville, Maryland, USA. True, F. W. 1889. Contributions to the natural history of the ceta- ceans, a review of the family Delphinidae. U.S. Natl. Mus. Bull. 36:1-191. Van Waerebeek, K. 1992a. Records of dusky dolphins Lagenorhynchus obscurus (Gray, 1828) in the eastern South Pacific. Beaufortia43(4):45-61. 1992b. Population identity and general biology of the dusky dolphin Lagenorhynchus obscurus (Gray, 1828) in the Southeast Pacific. Ph.D. diss., Univ. Amsterdam, 160 p. 1993. Presumed Lagenorhynchus skull at Tasmanian Museum reidentified as Lissodelphis peronii. Australian Mammalogy. (In press.) Van Waerebeek, K., and J. C. Reyes. 1990. Catch of small cetaceans at Pucusana port, cen- tral Peru, during 1987. Biol. Conserv. 51:15-22. 1993. Interaction between small cetaceans and Peru- vian fisheries in 1988-1989 and analysis of trends. Rep. Int. Whaling Comm. (Special issue.) (In press.) Van Waerebeek, K., J. C. Reyes, A. J. Read, and J. S. McKinnon. 1990. Preliminary observations of bottlenose dolphins from the Pacific coast of South America. In S. Leatherwood and R. R. Reeves (eds.) The bottle- nose dolphin, p. 143-154. Academic Press, San Di- ego. 653 p. Walker, W. A., S. Leatherwood, K. R. Goodrich, W. F. Perrin, and R. K. Stroud. 1986. Geographical variation and biology of the Pa- cific white-sided dolphin, Lagenorhynchus obliquidens, in the north-eastern Pacific. //; M. M. Bryden and R. R. Reeves (eds.) Research on dolphins, p. 411- 465. Oxford Univ. Press, Oxford. Webber, M. A. 1987. A comparison of dusky and Pacific white-sided dolphins (Genus Lagenorhynchus): morphology and distribution. Masters thesis, San Francisco State Univ., 102 p. 774 Fishery Bulletin 91(4). 1993 Webber, M. A., and S. Leatherwood. Yablokov, A.V., W. F. Perrin, and M. V. Mina. 1990. Dusky dolphin Lagenorhynchus obscurus. In R. 1983. Evaluation of phenetic relations among groups Martin (ed.), Whales and dophins, p. 154-155. Sala- of dolphins using analysis of non-metric cranial mander Books, London. variation. Zoologicheskiy Zhurnal 62:1887-96. (En- glish summary.) AbStfelCt.— An assessment of the population status of the eastern spinner dolphin (Stenella longirostris orientalis) in the eastern tropical Pa- cific is required by the U.S. Marine Mammal Protection Act (MMPAl, be- cause dolphin are killed in the tuna purse-seine fishery. A pooled esti- mate of abundance from recent (1986-90) research vessel surveys, in combination with estimates of fisheries kills from tuna vessel ob- server data, was used to estimate the historical (pre-exploitation) population size with a population dynamics model. Estimates of rela- tive population size (current popu- lation size divided by historical popu- lation size) were calculated by using a range of values for the maximum net recruitment rate and the maxi- mum net productivity level (MNPL). The resulting estimates of relative population size ranged from 0.32 to 0.58, with a best estimate of 0.44 based on available life history data. Estimates of relative population size were all below the value of MNPL used to calculate each estimate. Cal- culation of confidence limits for rela- tive population size by Monte Carlo simulation showed that the precision of the estimates was sufficient to make a status determination. The results indicated that, as of 1988, the stock of eastern spinner dolphin was depleted as defined by the U.S. MMPA. Estimation of historical population size of the eastern spinner dolphin (Stenella longirostris orientalis) Paul R. Wade Scripps Institution of Oceanography University of California. San Diego La Jolla. CA 92093 Southwest Fisheries Sciences Center. National Marine Fisheries Service. NOAA PO Box 271, La Jolla, CA 92038 Manuscript accepted 15 June 1993. Fishery Bulletin 91:775-787 (1993). The range of the eastern spinner dol- phin, Stenella longirostris orientalis (Perrin, 1990), is entirely contained within the eastern tropical Pacific ( Fig. 1 ). An assessment of population condition or status of this stock is required under the U.S. Marine Mammal Protection Act (MMPA), be- cause eastern spinner dolphins are killed in the tuna purse-seine fishery, which includes some U.S. vessels, that occurs in this region. The MMPA requires that each marine mammal population be maintained at an "op- timum sustainable population" (OSP) level, which has been defined by the U.S. National Marine Fisheries Ser- vice as a population size between the maximum net productivity level (MNPL) and carrying capacity (Fed- eral Register, 21 December 1976, 41FR55536). Therefore, assessing the status of a marine mammal stock in- volves, if possible, determining if it is above its MNPL. Populations shown to be below MNPL are consid- ered depleted under the MMPA. One method for determining a population's status relative to MNPL is to estimate historical abundance, meaning abundance prior to significant fisheries mortality, which is assumed equivalent to the equilibrium popula- tion size (i.e., carrying capacity). The current population size is then com- pared with what is thought to be the MNPL for the population, given the estimate of equilibrium population size (Gerrodette and DeMaster, 1990). The historical abundance of several ceta- cean populations has been estimated by back-calculating from a current abundance estimate, with a population model and annual records of the num- ber of animals harvested (Reilly, 1981; Breiwick et al, 1980, 1984; Lankester and Beddington. 1986). Smith (1983) described a method for back-calculat- ing historical population size (Nh) for spinner and spotted dolphins (Stenella spp.) from estimates of the current population size (A7,), the historical kill in the tuna fishery, the maximum net recruitment rate (i?,„), and the maxi- mum net productivity level. He used this technique to estimate historical abundance for the eastern spinner dol- phin, resulting in estimates of relative population size ( NJNh ) for 1979 rang- ing from 0.17 to 0.25. An estimate of Nh for a population of spinner or spotted dolphins, which have a relatively low /?,„, can be very sensitive to the estimate of A7,, as long as the time period between Nc and Nh is not too great (Smith and Pola- check, 1979). Over a long time pe- riod (138 years), the estimate of A7,, has been shown to be insensitive to the estimate of A7 for a baleen whale iBalaena mysticetus) population with a similarly low R,„ (Breiwick and Braham, 1990). However, for a popu- lation that has experienced a rela- tively recent decline from known losses, the estimate of Nh should still be sensitive to the estimate of N, (Gerrodette and DeMaster, 1990). 775 776 Fishery Bulletin 91(4). 1993 160° 140" 130' 120" 110° WEST LONGITUDE 90' Figure 1 Distribution of the eastern spinner dolphin (Stenella longirostris orientalis) in the eastern tropical Pacific. Squares represent positions of all sightings from the 1986-90 Monitoring of Porpoise Stocks (MOPS) surveys used in the abundance estimate (a total of 236 sightings). The outer grey line represents the MOPS study area, and the inner solid line represents the area occupied by the eastern spinner dolphin. For JV,., Smith (1983) used an estimated abundance of 293,000 animals for the eastern spinner dolphin, which was based on combined data from aerial and research vessel surveys conducted in 1979 (Holt and Powers, 1982). Recently, the U.S. National Marine Fish- eries Service conducted large-scale research vessel sur- veys annually for five years (1986-90) as part of the Monitoring of Porpoise Stocks (MOPS) program, re- sulting in a revised estimate of abundance of 632,700 (Wade and Gerrodette, 1992b1). For a number of rea- sons, discussed below, this estimate should be more reliable (both more precise and less biased) than the 1979 estimate of abundance used by Smith (1983). This revised abundance estimate was sufficiently dif- ferent from the 1979 estimate to justify re-estimation of historical population size for the eastern spinner dolphin. Additionally, estimates of the historical kill have also been revised since Smith (1983), although they did not differ greatly from the previous estimates (Lo and Smith, 1986; Wahlen, 1986). Therefore, I esti- mated the historical population size for the eastern spinner dolphin using the same methods and the same ranges for the parameters R,„ and MNPL as Smith (1983), but with revised abundance and fishery mor- 'Wade. P. R., and Gerrodette, T. 1992b. Estimates of cetacean abun- dance in the eastern tropical Pacific. Paper SC/44/018 presented at the annual meeting of the Int. Whal. Comm.. June 1992. tality estimates. This resulted in new estimates of rela- tive population size for this stock. Confidence limits for the estimates of relative popu- lation size were calculated by using Monte Carlo simu- lation methods (Buckland, 1984). These confidence lim- its only incorporated uncertainty due to sampling error of the current population estimate and the mortality estimates. They did not incorporate uncertainty in the model parameters R,„ and MNPL. Therefore, confidence intervals were calculated for all parameter combina- tions. Population abundance estimate The MOPS cruises (1986-90) had approximately five times more kilometers of survey effort in the region occupied by eastern spinner dolphins than the 1979 survey. About 75% of the 1979 survey was concentrated within 1,000 km of the coast, whereas the range of the eastern spinner dolphin is up to 2,000 km from the coast (Fig. 1, Perrin et al., 1985). Therefore, the 1979 survey provided little coverage of the western half of the area occupied by eastern spinner dolphin (Holt and Powers, 1982, fig. 1). Raw sample sizes show the large difference in the quantity of data: a total of 285 schools containing eastern spinner dolphins were re- corded during the MOPS surveys; a total of only 41 schools, during the 1979 survey. The large increase in Wade Population size of Stenella longirostris onentalis 777 the quantity of data made the MOPS estimates of abun- dance more precise for this stock, whereas the increased coverage of the stock range reduced the potential bias of geographical variation in abundance. Both Holt and Powers (1982) and Wade and Gerrodette (1992b1) used line-transect analysis meth- ods (Burnham et al., 1980) to estimate abundance. However, the relatively low number of sightings that resulted from the 1979 survey required an analysis technique that pooled sightings of different stocks and species of dolphin to estimate the abundance of each stock (Holt and Powers, 1982). Although the same tech- nique was used initially to estimate annual abundance for the first four years of the MOPS data (Holt and Sexton, 1989, 1990, a and b; Gerrodette and Wade, 1991), the greater number of sightings in each year made this unnecessary. Therefore, to examine trends in abundance, a revised analysis of all five years of MOPS data was undertaken in which annual estimates of abundance for each stock were made only from sightings of that stock (Wade and Gerrodette, 1992a). These estimates were considered to be less biased esti- mates of abundance than earlier estimates available for eastern tropical Pacific dolphins (IWC, 1992). No significant trend in abundance for eastern spinner dol- phins was observed over this short period, but the power of detecting a trend was low (Gerrodette, 1987; Wade and Gerrodette, 1992a). The five annual esti- mates of abundance for the eastern spinner dolphin ranged from 391,200 to 754,200, with a mean of 588,500. Wade and Gerrodette ( 1992a) discussed in detail the differences between their analysis technique and the Holt and Powers (1982) technique, but I will briefly summarize the two major differences here. First, Holt and Powers (1982) calculated a single effective strip width (i.e., 2.0/flO), Burnham et al., 1980) for all dol- phin species, whereas Wade and Gerrodette (1992a) estimated a separate value for each stock. These effec- tive strip widths varied substantially between the dif- ferent dolphin stocks, ranging from a low of 2.5 km to a high of 11.9 km (Wade and Gerrodette, 1992a), indi- cating that the Holt and Powers (1982) technique may have introduced considerable bias by pooling across different stocks and species. Second, Holt and Powers ( 1982) estimated the abun- dance of each stock by making a pooled estimate for each species, and then divided the species estimate between the stocks of that species according to the relative size of the area occupied by each stock. For example, an estimate of spinner dolphin abundance was made by pooling sightings of eastern spinner dol- phins with sightings of whitebelly spinner dolphins, a different morphological form that is distributed far- ther offshore and partially overlaps the area occupied by the eastern spinner dolphin (Perrin et al., 1985, 1991). The abundance estimate for the eastern spin- ner dolphin was then made by multiplying this pooled estimate by the ratio of the area occupied by the east- ern spinner dolphin to the sum of that area plus the area also occupied by the whitebelly spinner dolphin. This approach would only be un-biased if the two stocks had exactly the same density (number of animals per unit area) within their respective stock areas. There is no reason to assume this is true; therefore, an analy- sis based solely on sightings of eastern spinner dol- phin, as in Wade and Gerrodette (1992a), is likely to be less biased. To obtain a best estimate of absolute abundance, the five years of MOPS data were pooled across years for a second analysis to estimate average abundance for the period for 25 stocks of cetaceans in the ETP, including the eastern spinner dolphin (Wade and Gerrodette, 1992b1). The analysis technique of Wade and Gerrodette (1992a) was used, supplemented by a technique for prorating sightings from unidentified categories. Abun- dance estimate from this analysis should represent the (least biased and most precise) abundance esti- mate currently available for eastern spinner dolphin, and was therefore used as the starting point for the back-calculations. A summary of the methods and re- sults from that paper for the eastern spinner dolphin has been presented here. Methods Pooled (1986-90) abundance estimate The methods of Wade and Gerrodette (1992a) were mostly repeated but were applied to all five years of data together rather than separately to each year by itself. Population abundance (N) of eastern spinner dolphins was computed by line-transect methods (Burnham et al., 1980) as: where and N = X N, nkfk(0) N„ = _ SkAk 2L, (1) (2) N,, = abundance estimate for eastern spinner dol- phins in stratum /?, nk = number of eastern spinner dolphin schools in stratum k, ft(0) = detection function in stratum k, evaluated at zero distance, Sk = mean school size for eastern spinner dolphin schools in stratum k, 778 Fishery Bulletin 91(4), 1993 Lh = total effort in stratum k in kilometers, A/. = total area in stratum k in square kilometers. This represents a stratified analysis, where only sightings from a stratum were used to calculate the density and, therefore, abundance within the stratum. Abundance estimates for each stratum were summed across the four strata to get a total estimate for the stock. The only change in methodology from Wade and Gerrodette (1992a) involved the calculation of f(0). In that analysis, f(0) was estimated by pooling across strata because of inadequate sample sizes in each stra- tum in each year. With the larger sample sizes avail- able from pooling the five years of data, there were enough sightings in the inshore and middle strata (Wade and Gerrodette, 1992a; fig. 1) to estimate f(0) independently in each stratum. A third stratum (west) on the edge of the stock area had only four sightings, so a single pooled estimate of f(0) was estimated for the middle and west strata. As expected, because it was outside of the range of eastern spinner dolphin (Perrin et al., 1985), there were no sightings in the fourth stratum (south). A hazard rate model (Buckland, 1985) was fit to the data to estimate f(0). The perpen- dicular distances were truncated at 5.5 km, because not all dolphin schools further than 5.5 km perpen- dicular distance were pursued for species identifica- tion and school size estimation. Eastern and whitebelly spinner dolphins partially overlap in range, but can be distinguished from each other by their color pattern and morphology (Perrin, 1990; Perrin et al., 1991). Out of 134 sightings of spin- ner dolphins in the area of overlap between the two stocks, 16 were, for various reasons, unidentified to stock. Those sightings were prorated to the eastern stock of spinner dolphin by using the estimated pro- portion of spinner dolphin in the overlap area from the eastern stock (Wade and Gerrodette, 1992b1). Simi- larly, sightings of unidentified dolphins were prorated to the eastern stock, based on the estimated propor- tion of dolphins from the eastern stock in each stra- tum (Wade and Gerrodette, 1992b1). The prorated por- tions of unidentified spinner dolphin and unidentified dolphin were added to the original estimate to give a final estimate of abundance. The standard error of the abundance estimate was calculated by bootstrap meth- ods (Efron, 1982), by using legs of effort as the re- sampling unit, with 1,000 iterations. Fisheries kill estimates Estimates of dolphin kill from the tuna fishery in the ETP have been revised since Smith (1983). Lo and Smith ( 1986) presented revised kill estimates for 1959- 1972, and Wahlen (1986) presented revised kill esti- mates for 1973-1978, in each case with associated stan- dard errors. Additionally, kill estimates for 1979-87, with associated standard errors, have been published (IATTC, 1989). However, Lo and Smith ( 1986) reported total dolphin kill and did not divide it into stock cat- egories, while Wahlen (1986) reported kill estimates by stock, but only for the U. S. tuna vessel fleet. There- fore, I divided the estimates of Lo and Smith ( 1986) to stock by the same stock proportions used in Smith (1983). I adjusted the estimates of Wahlen (1986) us- ing the estimated total number of sets, as reported in Punsly (1983). Wahlen (1986) reported the estimated number of sets by the U.S. fleet. I multiplied the kill estimate in each year from Wahlen ( 1986) by the ratio of the sets made by the entire fleet to the sets made by the U.S. fleet to produce an estimate of the total num- ber of eastern spinner dolphins killed in each year. This assumes that the kill rates of the unobserved international fleet were the same as the U.S. fleet. Population model The methods of Smith (1983) were duplicated, by us- ing the simple recursive relationship NM = Nt-Kt + Rt{Nt-\Kt\ (3) where N, = population abundance in year t K, = fisheries kill in year t R, = net recruitment rate in year t. Density-dependence is incorporated into the equa- tion through the net recruitment rate, which is de- fined as R, = R, im (4) where Rm = maximum net recruitment rate z = shape parameter that sets the maximum net productivity level (MNPL) Nt, = historical population size (assumed to be the equilibrium population size). For any value of R,„ and MNPL, z can be calculated as in Polachek (1982). Equation 1 can be solved for N, as a function of iV,+1, R„ and Kr Therefore, by specify- ing an initial population size, the number of animals killed in each year, the maximum net recruitment rate, and the maximum net productivity level, these two equations can be iteratively solved for Nb. Wade Population size of Stenella longirostris orients/is 779 Estimates of R and MNPL m Values used by Smith (1983) for R,„ were 0.0, 0.03, and 0.06, which he thought to encompass the range of pos- sible values of Rm for spinner dolphins. No direct esti- mate of net reproductive rate (R) exists for eastern spinner dolphins because of the difficulty in estimat- ing survival rates. The calving interval is approximately three years (Perrin and Reilly, 1984). The age of sexual maturity (ASM) has been reported as five years (Perrin and Henderson, 1984). However, a new study using a much larger data set estimated ASM for the eastern spinner dolphin to be approximately 10 years, by us- ing data collected from 1974 to 1990-. This is similar to the estimate of approximately 11 years for the con- gener northern spotted dolphin, Stenella attenuata (Chivers and Myrick, 199P; Myrick et al., 1986), which is found in the same region of the eastern tropical Pacific. There are no estimates of survival rates for eastern spinner dolphin. Therefore, estimating the net repro- ductive rate for eastern spinner dolphin required us- ing estimates of survival rates from another species. Among the best estimates of survival rates for a delphinid come from a long-term study of known indi- viduals of a coastal population of Tursiops truncatus, with estimates of adult and calf survival of 0.96 and 0.80, respectively (Wells and Scott, 1992). From Reilly and Barlow (1986), those survival rates in combina- tion with a calving interval of three years and an ASM of nine years resulted in an R of 0.03, which could be considered the best estimate of R for the eastern spin- ner dolphin. Those survival rates may be low, how- ever, because the Wells and Scott (1992) study was of a population that was thought to be at equilibrium, as it had been relatively constant in abundance for many years. Using the maximum survival rates considered by Reilly and Barlow (1986) with the same calving interval (3 yr) and ASM (9 yr) results in an R of 0.05. If the eastern spinner dolphin was well below half its equilibrium population size in 1979 (Smith, 1983), then its net reproductive rate should have been very close to its maximum, /?,„. For this paper I therefore consid- ered 0.04 as the best estimate of R,„ currently avail- able for the eastern spinner dolphin, with 0.06 the greatest value of Rm possible. Therefore, the same range of values as in Smith (1983) was used fori?,,,, ranging -'Susan Chivers, Southwest Fish. Sci. Cent., La Jolla, CA. Pers. commun. 'Chivers, S. J., and A. C. Jr. Myrick. 1991. Comparison of age at sexual maturity for two stocks of offshore spotted dolphins sub- jected to different rates of exploitation. Dep. Commer.. NOAA. Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271 La Jolla CA 92038. Admin. Rep. LJ-91-31. 19 p. from 0.00 to 0.06 by increments of 0.002, for a total of 31 values. Values used by Smith (1983) for MNPL were 0.50, 0.65, and 0.80 (MNPL is expressed as a fraction of equilibrium population size in this paper), correspond- ing to z values (see Eq. 4) of 1.0, 3.482, and 11.216, respectively. These encompassed the range of actual values of MNPL for long-lived marine mammals, such as dolphins, based on work by Fowler (1981). No di- rect estimate of MNPL exists for the eastern spinner dolphin. Fowler (1984) gave evidence that MNPL was greater than 0.50 for cetaceans. A value of 0.60 is cur- rently being used for management of cetaceans under the U.S. MMPA (Federal Register, 31 October, 1980, 45FR64548), and for this paper, will be considered the best working value of MNPL currently available for the eastern spinner dolphin. Values of z were used so that MNPL ranged from 0.50 to 0.80 (the same range as in Smith, 1983), by using increments of 0.01, for a total of 31 values. The exact value of z necessary to give the specified MNPL for any value of Rm was cal- culated as in Polachek (1982). The 31 values used for both R„, and MNPL produced a total of 961 parameter combinations for which rela- tive population size was estimated. This large number of parameter combinations allowed the calculation of contours for the estimate of relative population size as a function of the 2 parameters of the model. Confidence limits for N n For every combination of the parameters R„, and MNPL, confidence limits for relative population size were calculated by a Monte Carlo simulation (Buckland, 1984) which incorporated the sampling error of the current abundance and kill estimates. On each of 1,000 iterations, an artificial data set was randomly gener- ated by sampling values for the current abundance and for the fisheries kill in each year. These values were each drawn from Gaussian distributions with means and variances equal to the appropriate point estimates. Relative population size was then estimated for each of these artificial data sets, and 95% confi- dence limits for relative population size were calcu- lated using the percentile method (Efron, 1982). The kill estimates for 1959-1972 were not indepen- dent from each other, as Lo and Smith (1986) esti- mated the kill in each year by multiplying an average mortality-per-set for 1959-1972 by the number of fishing sets in each year. Therefore, on each simula- tion iteration the kill values for 1959-1972 were ran- domly generated with the same random deviate. This resulted in the kill values for those years being per- fectly correlated amongst themselves from simulation 780 Fishery Bulletin 91(4). 1993 trial to trial, which correctly reflected the lack of inde- pendence in the actual estimates. The kill values for all other years were sampled independently. Results Estimates of abundance and kill The five years of the MOPS surveys resulted in 236 sightings of eastern spinner dolphins used in the abun- dance estimate. The abundance estimate based solely on these sightings was 568,100. Adding prorated num- bers of unidentified spinner and unidentified dolphin sightings resulted in a final estimate of 632,700, with a CV of 0.167 (Table 1). The fisheries kill estimates ranged from a high in 1961 of 138,000 to a low in 1983 of700(Table2). Contours of relative population size (N,./Nh ) as a func- tion of R„, and MNPL ranged from 0.35 to 0.55 (Fig. 2). Relative population size increased with both R„, ( growth rate) and MNPL (the amount of non-linearity in the density-dependence response). The lowest relative population size was 0.32, for the case of #,,,=0.00, (i.e., no net growth in the population before fisheries kill was included). The highest relative population size was 0.58 for the case of the highest growth rate and MNPL (0.06 and 0.80, respectively). These low and high esti- mates of relative population size correspond to esti- mates of pre-exploitation abundance of 1,956,000 and 1,100,000, respectively. Relative population size in- creased by approximately 0.03 for every increase of 0.01 in Rm. The influence of MNPL was greater at higher growth rates, as relative population size in- creased by approximately 0.02 for every increase of 0.10 in MNPL at #,,=0.02, but increased by approxi- mately 0.05 for every increase of 0.10 in MNPL at #,,,=0.06. There were no combinations of parameter values such that relative population size was estimated to be above MNPL. The upper 957c confidence limit for relative popula- tion size as a function of R,„ and MNPL, based on the sampling error of the abundance and kill estimates, ranged from 0.45 to 0.91 (Fig. 3). The upper confidence limit was always above MNPL when R,„ was greater than 0.046 (Fig. 3, shaded region). The lower 95% con- fidence limit for relative population size as a function of R„, and MNPL, ranged from 0.22 to 0.36 (Fig. 4). All population trajectories declined until 1973 (Fig. 5), after which the estimated fisheries kill declined substantially (Table 2). For the highest growth rate, the population trajectory showed an increasing trend from 1976 to 1988 (Fig. 5, line C), whereas for the lowest growth rate the model resulted in a relatively stable population level between 1976 and 1988 (Fig. 5, line A). The confidence limits around relative population size broadened with increasing Rm. For example, for a MNPL of 0.60, the confidence limits ranged from 0.23 to 0.44 for #,,,=0.00, whereas they ranged from 0.33 to 0.72 for #,,=0.06 (Fig. 6). As in Smith (1983), relative population size was a linear function of Rm. Discussion For all parameter values of Rr„ and MNPL equal to those in Smith ( 1983), estimates of relative population size were higher in this analysis. For example, for R„ =0.03 and MNPL=0.65, Smith (1983) reported a rela- tive population size of 0.20 versus a result of 0.42 Table 1 Estimate of abundance (in thousands of animalsl of the eastern spinner dolphin {Stenella longirostris orientalis) from the Monitoring of Porpoise Stocks surveys (1986-90). Strata are identified in Figure 1. Abundance estimates Total Inshore Middle West From eastern spinner dolphin schools 568.0 364.8 160.2 43.1 Prorated from unidentified spinner dolphins 15.4 9.0 6.0 0.4 Prorated from unidentified dolphins 49.2 37.5 10.9 0.8 Final estimate 632.7 Standard error 105.7 Coefficient of variation 0.167 Upper 95% confidence limit 778.9 Lower 95% confidence limit 403.2 Wade Population size of Stenella longirostris orientate 781 Table 2 Estimates of fisheries kill in thousands by year for the eastern spinner dolphin ( Stenella longirostris orientalis). CV is the coefficient of variation of the kill estimate. Sources for the estimates are 1) 1959-72 from Lo and Smith (1986), using the stock proportions of Smith (1983 ; 2) 1973-78 from Wahlen (1986), adjusted for number of sets of total fleet in Puns ly (1983); 3) 1979- 87 from IATTC (1989). See text for explanation. Year Mortality CV 1959 14.3 0.32 1960 124.3 0.31 1961 138.8 0.28 1962 56.2 0.25 1963 62.4 0.22 1964 101.4 0.20 1965 119.6 0.20 1966 97.2 0.15 1967 66.8 0.16 1968 59.5 0.15 1969 106.0 0.15 1970 107.4 0.15 1971 58.4 0.17 1972 87.4 0.16 1973 18.4 0.16 1974 17.8 0.11 1975 17.1 0.11 1976 14.7 0.12 1977 1.8 0.12 1978 1.1 0.11 1979 1.5 0.24 1980 1.1 0.20 1981 2.3 0.28 1982 2.6 0.33 1983 0.7 0.38 1984 6.0 0.52 1985 8.9 0.16 here. The different results must be due to either the use of revised estimates of abundance and kill or the use of 1988 as a starting point rather than 1979; these were the only differences between the analyses. As will be shown, most of the difference resulted from the higher esti- mate of current population size, although the lower revised kill estimates also con- tributed to a higher estimate of relative population size. Repeating the back-cal- culation of Smith (1983) from 1979, but using revised population and kill esti- mates, resulted in nearly the same esti- mate of relative population size as re- ported here. For example, for R„=0.03 and MNPL=0.65, back-calculating from 1979 as opposed to 1988 resulted in an estimate of relative population size of 0.41 versus 0.42, whereas Smith (1983) reported a value of 0.20. An inspection of the popula- tion trajectories (Fig. 5) confirms that the difference was not due to the different starting year, as the model trajectories, except at the highest growth rates, indicated little change in the population size between 1979 and 1988. This also agrees with the independent re- sults of Buckland et al. (1992), which indicated little difference in relative population size between those two years. Therefore, the difference in the results reported here and those of Smith (1983) should not be interpreted as a recovery in the population between 1979 and 1988. These new, higher estimates of status should in- stead be interpreted as a revision of the estimate of relative popula- tion size, due mostly to the improved abundance estimate available from the MOPS surveys. The new estimates of relative population size, although higher than Smith (1983), are still below MNPL for all parameter combina- tions. Because the parameter values used encompassed those values possible for a spinner dolphin (Reilly and Barlow, 1986), this result indicated that, as of 1988, the eastern spinner dolphin was still well below its 1959 population size. With Rn =0.04 and MNPL=0.60, the population was estimated to be at 44% of its historical size. Even with the maximum value of Rm of 0.06, the population in 1988 was estimated to be 43% (MNPL=0.50) to 58% (MNPL=0.80) of its size in 1959. However, careful consideration must be given to several issues before accepting these results as valid. These issues include the precision (reflecting the precision of the abundance and kill esti- mates) and potential biases (reflecting either biased abundance and kill estimates or mis-specification of the model) of the result, and the quality of pre- 1972 fisheries kill data. Precision The precision of the estimates of relative population size was inves- tigated by simulation to explore the uncertainty of the results due to sampling error, under the assumptions that the population model and parameter values were true. This addresses the question of how likely the estimates of relative population size were below MNPL if the true relative population size was above MNPL, solely because of variability associated with sampling the current abundance and fisheries kill estimates. The upper 95% confidence limit of relative population size was below MNPL for the majority of the parameter combinations, moving above MNPL only for values of R,„ greater than 0.018 (Fig. 3). If MNPL was assumed to be 0.60, then the upper 95% confidence limit of relative population size was only above MNPL for values of Rm greater than 0.034 (Fig. 3). The upper confi- dence limit was always above MNPL if/?,,, was at least 0.046. Viewed in a hypothesis testing context, this result indicated that the null hypothesis that relative population size was greater than MNPL in 1988 could be rejected for most of the parameter combinations. Only at higher growth rates could this hypothesis not be rejected. From sampling error alone, it was equally possible that the population was actually worse off than estimated, as the lower 95% confidence limits go as low as 0.22, and were as low as 0.28 even at the highest growth rate of i?„=0.06. 782 Fishery Bulletin 91(4), 1993 1 \.55 0.75- \.45 V50 0.70- _l §? 0.65 ■ 2 L35 \.40 0.60- 0.55- 0.50- \ i > ' ■ ' i 0.00 0.01 0.02 0.03 R 0.04 0.05 0.06 Figure 2 Contours of relative population size (current abundance divided by historical abundance) for the eastern spin- ner dolphin iStenella longirostris onentalis), as a func- tion of maximum net recruitment rate (/J„,l and maximum net productivity level (MNPLi. Vas .80\ 0.75- (.75 \ \ .7d 0.70- \.60 1.65 _J ^ 0.65 - 2 \.50 \.55 0.60- .45 \ 0.55- \4i 0.50- X- . -i— -— 0.00 0.01 0.02 0.03 R 0.04 0.05 0.06 Figure 3 Contours of the upper 95% confidence limit for rela- tive population size for the eastern spinner dolphin iStenella longirostris orientalis), as a function of maxi- mum net recruitment rate (R,„) and maximum net pro- ductivity level (MNPLi. The shaded region represents the area where the confidence limit was above MNPL. The confidence limits around relative population size were not much greater proportionally than the confi- dence limits around A/,. (Fig. 6). For example, from the simulation the confidence limits for relative popula- tion size with values of 0.04 for Rm and 0.60 for MNPL were 0.29-0.62, representing a coefficient of variation (CV) of 19%, whereas the CV of Nc was 17% (Table 1). Although one might have expected the precision of the estimate of relative population size to be much less than the precision of AT , this was not the case because of the independence of most of the kill estimates. Sam- pling variance in the kill estimates would therefore 0 8Q 1 \ \'35 0.75- l \ \ 1 \.30 \ 0.70- 1 \ _i ^ 0.65 - \.25 \ 0.60- \ \ 0.55- \ \ 0 50 - \ \ 36 0. DO 0.01 0.02 0.03 0.04 0.05 0. R m Figure 4 Contours of the lower 95' I confidence limit for relative population size for the eastern spinner dolphin IStenella longirostris oriental is ), as a function of maximum net recruitment rate iR,„) and maximum net productivity level (MNPL). 2000 ' 1750 1500- 1250 1000 750 = 500 250 Figure 5 Population model trajectories for the eastern spinner dolphin (Stenella longirostris orientalis\ for three different parameter combinations of maximum net recruitment rate i/f,„i and maximum net productivity level (MNPL): (Al fl,„=0.00 and MNPL=0.50. iBi R,„=0.04 and MNPL=0.60, and (Ci ft,, =0.06 and MNPL=0.80. A and C represent tin- lowest and highest estimates of relative population size, respectively. B represents the combination of the best estimates for the parameters based on available life history data. Wade Population size of Stenella longirosths orientalis 783 a: 0.06 Figure 6 Point estimates and 95fr confidence limits for relative population size for the eastern spinner dolphin {Stenella longirostris orientalist, as a function of the maximum net recruitment rate (/?.„). for the estimate of the maximum net productivity level (MNPL=0.60i currently used for management under the U.S. Ma- rine Mammal Protection Act. tend to cancel itself, as over-estimates of kill in some years would be balanced by under-estimates in other years. However, systematic bias in the kill estimates would lead to a poor estimate of relative population size, creating a relatively precise yet inaccurate estimate. The same would be true for bias due to the use of an inappropriate model, or in the estimate of N,.. Potential biases in these three areas must therefore be considered. Bias Two major sources of bias may have existed in the fisheries kill estimates from 1972 to the present. One source of bias, that of observer effects, has been demonstrated and implies that less dolphin kill occurred on observed trips due to modifications in fishing behavior in response to the observer's presence (Wahlen and Smith, 1985). Unfortunately, there is no way to esti- mate the magnitude of the effect. The second potential source of bias was the lack of partici- pation in data collection by some countries dur- ing some years, especially if significant differ- ences in kill rates existed between countries. Most important may have been the lack of 1979-1986 data from Mexico (Edwards, 1989), a major component of the fishery whose kill rates may have been higher than average during that time period. These bi- ases would lead to under-estimates of kill and thus over- estimates of relative population size. Additional sources of bias existed in the pre-1972 kill estimates, because of both a lack of observations of mor- tality-per-set (MPS) in many years and because the MPS data prior to 1971 were not collected as part of a system- atic observer program. Data on the number and types of sets were collected in every year, starting in 1959 (Punsly, 1983), but an observer program for collecting MPS data was not started until 1971 and random placement of ob- servers until 1972 (Edwards, 1989). The moderate amount of MPS data collected in 1971 was potentially biased be- cause most of the boats with observers were smaller and older, and may have had a higher MPS (Edwards, 1989; Lo and Smith, 1986, table 1). Most of the pre-1971 data were from scientists who were on the tuna boats for the purpose of collecting dolphin specimens, but who also re- corded MPS data on their own initiative (Smith and Lo. 1983). There is no obvious reason why tuna vessels that agreed to allow scientists to collect specimens during their fishing operations would tend to have different mortality rates, but in a strict sense these were not random samples of fishing trips. Data from one fishing trip in 1964 were recorded and reported by a fisherman, who may have done so because of the magnitude of the kill, making those data potentially biased (Smith and Lo, 1983). However, the MPS data did not differ greatly in those years from the data collected in 1972 (Lo and Smith, 1986, table 1). Because of this and the greater quantity of MPS data available from 1972, estimates of 1959-1972 fisheries kill made by multiplying the average 1972 MPS rate by the number of sets in each year would not differ greatly from the estimates used here from Lo and Smith (1986), which were made by using the pooled 1964-1972 MPS rate. Therefore, the fisheries kill would only have been over- estimated if the MPS in the pre-1971 unobserved years was lower than in 1972. However, MPS has consistently declined over time, declining most rapidly following the passage of the Marine Mammal Protection Act in 1972 (Smith 1983). No evidence exists that MPS could have been lower from 1959 to 1970 than it was in 1971-72. MPS may have been higher, especially before use of the back-down procedure had become widespread and well practiced (Perrin, 1969; Edwards, 1989). If it is assumed that MPS has only declined since the beginning of the fishery, the 1959-1970 kill esti- mates of Lo and Smith ( 1986) were likely under-estimates of the true kill. Thus, the major sources of bias in fisheries kill estimates all suggest that kill estimates were negatively biased. Bias in the estimate of abundance could also bias the estimate of relative population size. Wade and Gerrodette 784 Fishery Bulletin 91(4), 1993 (1992a) discuss a number of sources of potential bias when applying line-transect theory to the MOPS sur- vey data. Several potential sources of bias do not ap- pear to have a major effect. Independent observer ex- periments indicate that few schools (and no large schools) were missed on the trackline (Wade and Gerrodette, 1992a). Aerial photographs have confirmed that little bias has been introduced by the observer's estimate of school size (Gerrodette and Perrin, 19914). One partially unresolved issue is that of vessel avoid- ance by dolphin schools, which would bias the esti- mate downwards, although this may not have been a major problem (Au and Perryman, 1982; Hewitt, 1985). Additionally, mean school size is likely over-estimated owing to the decreased probability of detection of small schools at larger perpendicular distances (Drummer and McDonald, 1987). Although some stocks in the MOPS surveys appeared to be biased by as much as 20% by this problem, the eastern spinner and other stocks were not (Wade and Gerrodette, 1992a). Finally, the distribution of the eastern spinner dolphin is well known (Perrin et al., 1985) and is well within the MOPS study area (Fig. 1), so it can be concluded that the abundance estimate applies to the entire population. Therefore, the estimate of abundance did not contrib- ute any major bias to the estimate of relative popula- tion size. Bias may also have been introduced by assuming that the simple model specified in Equation 3 correctly models eastern spinner population dynamics, although a simulation study has shown that a simple model can perform as well as a more complex model for this type of analysis (Lankester and Cooke, 1987). The most important feature of eastern spinner population dy- namics for this analysis is their inability to undergo large increases in population size from one year to the next. Their relatively low maximum population growth rate, which is due to the biological constraints of their life history discussed above, was incorporated into Equation 3 by using only biologically plausible values of Rm. The only way in which the actual population could have substantially differed from the model would be if the population had a much lower growth rate than expected in some years. For example, large inter- annual variations in oceanographic conditions related to El Nino events in the eastern tropical Pacific (Fiedler et al., 1992) may lead to large changes in the quantity of prey available for the dolphins. This could lead to lower growth rates in some years, which would cause the specified model to over-estimate relative popula- tion size. Of more concern is the lack of age-structure in the model (Goodman, 19845). The age-distribution of the northern spotted dolphin (S. attenuate/,) kill of 1974 to 1983 was heavily biased towards mature animals (Barlow and Hohn, 1984). If the kill of eastern spinner dolphin was similar for all years, then the simple model used would have over-estimated relative population size. Removing proportionally more mature animals, whose reproductive value was highest, would have tem- porarily reduced the growth rate of the population and caused the population to decline for a longer period than predicted by the simple model. In fact, an independent abundance index derived from data on sightings of dolphin schools from tuna vessels estimated that the population experienced a statistically significant decline from 1975, the first year the index was available, until 1982 (Buckland et al., 1992). This is different from the population trajectory I estimated here, which declined only until about 1977 (Fig. 5). Additionally, Buckland et al.'s (1992) trajec- tory indicated that the population level in 1988 was not substantially different from that of 1979, which conflicts with the model trajectories presented here with higher growth rates (Fig. 6), in which substantial growth occurs over 1979-1988. If Buckland et al.'s (1992) estimated population trajectory was an accu- rate assessment of the true population trend, then the results presented here suggest either 1 ) that the popu- lation growth rate was less than R„,=0.04; or 2) that kill was under-estimated during the 1980s for reasons discussed above; or 3) that a skewed age-structure led to a lagged response to the large decrease in kill dur- ing the 1970s; or 4) some combination of these possi- bilities. Current status Estimated kill from the fishery in recent years has been as high as 19,526, with an average kill of 13,900 from 1986 to 1990 (DeMaster et al., 1992), which rep- resented a kill rate of 2.1% of the population estimate of 632,700. As indicated by Equation 3, the estimates of historical population size presented here, which are back-calculated from 1988, were based only on kill data through 1987. Estimated kill was 18,793 in 1988 (IATTC, 1989) and 15,245 in 1989 (Hall and Boyer, 1991), representing 3.0% and 2.4% of the abundance 'Gerrodette, T., and C. Perrin. 1991. Calibration of shipboard esti- mates of dolphin school size from aerial photographs. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. U-91-36, 24 p. ''Goodman, D. 1984. Consideration of age structure in back projec- tion calculations for the northern offshore spotted dolphin popula- tion. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. LJ-84- 26C, 25 p. Wade: Population size of Stenella longirostns onentalis 785 estimate, respectively. These recent kill estimates were the highest since 1976 (Table 2), and may have been high enough to prevent recent recovery of the popula- tion. The most recent estimates of the abundance in- dex from tuna vessel sighting data indicated the popu- lation was declining from 1986 to 1991 (Anganuzzi et al., 1992fi). However, the most recent kill information indicated a substantial reduction in kill to less than m of the population in both 1990 (5.378, Hall and Boyer, 1992) and 1991 (5,879, Hall and Lennert, in press), which resulted in an average kill per year for 1988-91 of 11,324, or 1.8% of the population estimate of 632,700. Therefore, the current status of the popu- lation is unlikely to be substantially different from what it was in 1988. Managing kill levels so that they do not exceed some fraction of the expected maximum net recruitment rate may be the most reasonable man- agement strategy for promoting recovery of the popu- lation (DeMaster et al, 1992). The U.S. National Ma- rine Fisheries Service has recently proposed listing the eastern spinner dolphin as depleted under the U.S. MMPA (Federal Register, 17 June 1992, 57FR27010). A separate proposal to list the eastern spinner dolphin as "threatened" under the Endangered Species Act was not warranted at this time, as the population is in no immediate danger of extinction (Federal Register, 19 October 1992, 57FR47620). Proposed international quo- tas on fisheries kill for each dolphin stock in the east- ern tropical Pacific (IATTC, in press; MMC, 1993), if implemented, would ensure that mortality levels stayed low enough to allow recovery of the population to the OSP level. dolphin population was well below historical abundance levels in 1988. Most uncertainties appear to lead to over-estimates of relative population size, indicating the population may be at a lower level than indicated here. Calculation of confidence limits for relative popu- lation size showed that the precision of the estimates was sufficient to make a status determination except for higher values of Rm. However, higher growth rates (Rm > 0.04) were not supported by independent evi- dence available about the population trend since 1975. The results indicated that, as of 1988, the stock of eastern spinner dolphins was depleted as defined by the U.S. MMPA. The substantial fisheries kill that oc- curred after 1988 makes it unlikely that the popula- tion has experienced any significant recovery since then. Acknowledgments I would like to thank Douglas P. DeMaster, Tim Gerrodette, and James T Enright for the many edito- rial comments that improved this manuscript consid- erably. I would also like to thank an anonymous re- viewer who pointed out the method of exactly calculating the z values. Finally, I would like to thank the many people who contributed to the collection and analysis of the data used in this paper, particularly the observers on the research and tuna vessels. Literature cited Conclusions Based on the best data available on abundance, kill, and population dynamics, the population size in 1988 of the eastern spinner dolphin was estimated to be below MNPL, within the range of 32% to 58% of pre- exploitation population size. Based on available life history data, the population size was estimated at 44% of pre-exploitation population size. Relative population size was estimated to be higher than Smith's (1983) estimate for 1979, but this difference was due mostly to the use of a new, better estimate of abundance, rather than to a recovery of the population between 1979 and 1988. Although there are uncertainties asso- ciated with this analysis, especially with the early kill data, the results indicated that the eastern spinner "Anganuzzi. A. A., S. T. Buckland, and K. L. Cattanach. 1992. Rela- tive abundance of dolphins associated with tuna in the eastern Pa- cific Ocean: analysis of 1991 data. Paper SC/44/SM23 presented at the annual meeting of the Int. Whal. Comm. Au, D., and W. Perryman. 1982. Movement and speed of dolphin schools respond- ing to an approaching ship. Fish. Bull. 80( 2 ):37 1-379. Barlow, J., and A. Hohn. 1984. Interpreting spotted dolphin age distribu- tions. NOAA Tech. Memo. NMFS-SWFC-48, 22 p. Breiwick, J. M., and H. W. Braham. 1990. Historical population estimates of Bowhead whales: sensitivity to current population size. Rep. Int. Whal. Comm. 40:423^126. Breiwick, J. M., E. D. Mitchell, and D. G. Chapman. 1980. Estimated initial population size of the Bering Sea stock of bowhead whale. Balaena mysticetus: an iterative method. Fish. Bull. 78( 4 1:843-853. Breiwick, J. M., L. L. Eberhardt, and H. W. Braham. 1984. Population dynamics of western Arctic bow-head whales (Balaena mysticetus). Can. J. Fish. Aquat. Sci. 41:484-496. Buckland, S. T. 1984. Monte Carlo confidence intervals. Biometrics 40:811-817. 1985. Perpendicular distance models for line transect sampling. Biometrics 41:177-195. 786 Fishery Bulletin 91(4), 1993 Buckland, S. T., K. L. Cattanach, and A. A. Anganuzzi. 1992. Estimating trends in abundance of dolphins as- sociated with tuna in the eastern tropical Pacific, us- ing tuna vessel sightings data. Fish. Bull. 90(11:1-12. Burnham, K., D. Anderson, and J. Laake. 1980. Estimation of density from line transect sam- pling of biological populations. Wildl. Monogr. No. 72, 202 p. DeMaster, D. P., E. F. Edwards, P. R. Wade, J. E. Sisson. 1992. Status of dolphin stocks in the eastern tropical Pacific. In D. R. McCullough and R. H. Barret (eds.), Wildlife 2001:populations. Elsevier Press, London, England. Drummer, T. D., and L. L. McDonald. 1987. Size bias in line transect sampling. Biometrics 43:13-21. Edwards, E. F. 1989. Using tuna-vessel observer data to detect trends in abundance of dolphin populations: history and re- search to date ( 1988). NOAA Tech. Memo. NMFS- SWFC-122, 123 p. Efron, B. 1982. The jackknife, the bootstrap, and other re- sampling methods. SIAM, monograph #38, CBMS- NSF, 92 p. Fiedler, P. C, F. P. Chavez, D. W. Behringer, and S.B. Reilly. 1992. Physical and biological effects of Los Ninos in the eastern tropical Pacific, 1986-1989. Deep-Sea Res. 39(2):199-219. Fowler, C. 1981. Density dependence as related to life history strategy. Ecology 62(3 ):602-610. 1984. Density dependence in cetacean popula- tions. Rep. Int. Whaling Comm. Spec. Issue 6:373- 379. Gerrodette, T. 1987. A power analysis for detecting trends. Ecology 68(51:1364-72. Gerrodette, T., and D. P. DeMaster. 1990. Quantitative determination of optimum sus- tainable population level. Mar. Mammal Sci. 6(1 ):1- 16. Gerrodette, T. and P. R. Wade. 1991. Monitoring trends in dolphin abundance in the eastern tropical Pacific: analysis of 1989 data. Rep. Int. Whal. Comm. 41:511-515. Hall, M. A., and C. Boyer. 1991. Incidental mortality of dolphins in the tuna purse-seine fishery in the eastern Pacific Ocean dur- ing 1989. Rep. Int. Whal. Comm., Volume 41:507- 509. 1992. Estimates of incidental mortality of dolphins in the purse-seine fishery for tunas in the eastern Pa- cific Ocean in 1990. Rep. Int. Whal. Comm., Vol. 42:529-531. Hall, M. A., and C. Lennert. 1992. Estimates of incidental mortality in the eastern Pacific ocean tuna fishery in 1991. Rep. Int. Whal. Comm. Vol. 43. (In press.) Hewitt, R. P. 1985. Reaction of dolphins to a survey vessel: effects on census data. Fish. Bull. 83(2):187-193. Holt, R. S., and J. E. Powers. 1982. Abundance estimation of dolphin stocks in the eastern tropical Pacific yellowfin tuna fishery deter- mined from aerial and ship surveys to 1979. U.S. Dep. Commer., NOAA-TM-NMFS-SWFC-23, 95 p. Holt, R. S., and S. N. Sexton. 1989. Monitoring trends in dolphin abundance in the eastern tropical Pacific using research vessels over a long sampling period: Analyses of 1987 data. Rep. Int. Whal. Comm. 39:347-351. 1990a. Monitoring trends in dolphin abundance in the eastern tropical Pacific using research vessels over a long sampling period: Analyses of first year's data. Fish. Bull. 88(11:105-111. 1990b. Monitoring trends in dolphin abundance in the eastern tropical Pacific using research vessels over a long sampling period: Analyses of 1988 data. Rep. Int. Whal. Comm. 40:471-477. IATTC (Inter- American Tropical Tuna Commission). 1989. Annual Report. 1988. Inter-Am. Trop. Tuna Comm., Scripps Inst, of Ocean., La Jolla, 288 p. In press. Annual Report, 1992. Inter-Am. Trop. Tuna Comm., Scripps Inst, of Ocean., La Jolla. IWC (International Whaling Commission). 1992. Report of the Subcommittee on Small Cetaceans. Rep. Int. Whal. Comm. 42:178-234. Lankester, K., and J. R. Beddington. 1986. An age structured population model applied to the gray whale (Exchrichtius robustus). Rep. Int. Whal. Comm. 36:353-358. Lankester, K., and J. G. Cooke. 1987. On the adequacy of simplified population mod- els for the assessment of exploited whale stocks. Rep. Int. Whal. Comm. 37:383-386. Lo, N. C. H., and T. D. Smith. 1986. Incidental mortality of dolphins in the eastern tropical Pacific, 1959-72. Fish. Bull. 84:27-33. MMC (Marine Mammal Commission). 1993. Annual Report to Congress. 1992. Mar. Mamm. Comm., Washington, D.C. Myrick, A. C, A. A. Hohn, J. Barlow, and P. A. Sloan. 1986. Reproductive biology of female spotted dolphins, Stenella attenuate!, from the eastern tropical Pa- cific. Fish. Bull. 84:247:259. Perrin, W. F. 1969. Using porpoise to catch tuna. World Fishing 18:42-45. 1990. Subspecies of Stenella longirostris (Mammalia: Cetacea: Delphinidae). Proc. Biol. Soc. Wash. 103(2), p 453-463. Perrin, W. F., and J. R. Henderson. 1984. Growth and reproductive rates in two popula- tions of spinner dolphins, Stenella longirostris, with different histories of exploitation. Rep. Int. Whal. Comm. (Spec. Issue) 6:417-430. Perrin, W. F., and S. B. Reilly. 1984. Reproductive parameters of dolphins and small Wade Population size of Stenella longirostris orients/is 787 whales of the family Delphinidae. Rep. Int. Whal. Comm. (Spec. Issue 61:97-133. Perrin, W. F., M. D. Scott, G. J. Walker, and V. L. Cass. 1985. Review of geographical stocks of tropical dol- phins [Stenella spp. and Delphinus delphis) in the eastern Pacific. U.S. Dep. Commer.. NOAA-TM- NMFS-SWFC-28, 28 p. Perrin, W. F., P. A. Akin, and J. V. Kashiwada. 1991. Geographic variation in external morphology of the spinner dolphin, Stenella longirostris, in the east- ern Pacific and implications for conservation. Fish. Bull. 89:411-428. Polacheck, T. 1982. Local stability and maximum net productivity levels for a simple model of porpoise population sizes. U.S. Dep. Commer., NOAA-TM-NMFS-SWFC- 17, 14 p. Punsly, R. 1983. Estimation of the number of purse-seiner sets on tuna associated with dolphins in the eastern Pa- cific Ocean during 1959-1980. Inter-Am. Trop. Tuna Comm. Bull. 18:229-299. Reilly, S. B. 1981. Population assessment and population dynam- ics of the California gray whale (Eschrichtius robustus). Ph.D. diss., Univ. Washington. Reilly, S. B., and J. Barlow. 1986. Rates of increase in dolphin population size. Fish. Bull. 84( 3 ):527-533. Smith, T. D. 1983. Changes in size of three dolphin (Stenella spp.) populations in the eastern tropical Pacific. Fish. Bull. 81:1-13. Smith, T. D., and T. Polachek. 1979. Analysis of a simple model for estimating his- torical population sizes. Fish. Bull. 76:771-779. Smith, T. D., and N. C. H. Lo. 1983. Some data on dolphin mortality in eastern tropi- cal Pacific tuna purse seine fishery prior to 1970. U.S. Dep. Commer., NOAA-TM-NMFS-SWFC- 34, 26 p. Wade, P. R., and T. Gerrodette. 1992a. Estimates of Dolphin Abundance in the East- ern Tropical Pacific: preliminary analysis of five years of data. Rep. Int. Whal. Comm., Vol. 42:533-539. Wahlen, B. E. 1986. Incidental dolphin mortality in the eastern tropi- cal Pacific tuna fishery, 1973 through 1978. Fish. Bull. 84(31:559-569. Wahlen, B. E., and T. D. Smith. 1985. Observer effect on incidental dolphin mortality in the eastern tropical Pacific tuna fishery. Fish. Bull. 83:521-530. Wells, R., and M. Scott. 1992. Estimating bottlenose dolphin population param- eters from individual identification and capture-re- lease techniques. Rep. Int. Whal. Comm. (Spec. Issue 12):407-415. AbStraCt.-Water quality in the tidal freshwaters of the Delaware River has improved substantially over the last decade. Areas near Philadelphia that were once anoxic now rarely experience dissolved oxy- gen concentrations lower than 3 ppm. To assess how fish spawning and nursery activity in the river has changed following these water qual- ity improvements, ichthyoplankton were collected from April to June in 1987 and 1988 in the tidal freshwa- ters of the Delaware River and com- pared to data available from studies conducted during the mid-1970's. Eighteen taxa were collected in the present study; larval Morone americana (white perch), Alosa sapidissima (American shad), and Alosa spp. (river herring) were the most abundant. Total density of eggs and larvae was highest upstream of Philadelphia, where water quality has historically been highest. Ichthyoplankton density and taxo- nomic composition in the tributar- ies were similar to the most produc- tive portions of the mainstem, except in the Schuylkill River. There, wa- ter quality is still poor, and ichthyoplankton density was an or- der of magnitude less than in the mainstem. In comparison to data from the previous decade, the present assemblage of the Delaware River was more diverse. Nine taxa, most notably A. sapidissima, were significantly more abundant in at least one portion of the river in the present study, though total ichthyoplankton density was no higher than prior to the water qual- ity improvements. Five taxa, most notably Cyprinus carpio (carp), were significantly less abundant in the present assemblage. Spring distribution and abundance of ichthyoplankton in the tidal Delaware River Stephen B. Weisberg William H. Burton Versar, Inc. 9200 Rumsey Road Columbia, MD21045 Manuscript accepted 15 June 1993. Fishery Bulletin 91: 788-797 1 1993). The tidal freshwater portion of the Delaware River was once among the most polluted estuaries in the United States. Water quality was so poor dur- ing the 1940s that gases released from the water reportedly tarnished metal and corroded engine parts of naval vessels moored near Philadelphia, PA (Albert, 1988). Conditions were poor- est in the Philadelphia area, where much of the river became anoxic dur- ing warmer months. This so-called pollution block substantially reduced the upriver migration of anadromous fishes such as A. sapidissima and Morone saxatilis (striped bass) (Chittenden, 1971, 1974). Installation of secondary treatment at municipal sewage treatment plants along the river has reduced organic loading by more than 75% since the early 1970's (Albert, 1988). As a result, dissolved oxygen concen- tration in the river has increased, and the magnitude and duration of the pollution block near Philadelphia has been reduced. In areas that were once anoxic, dissolved oxygen concentra- tions during the spring are now rarely lower than 5 ppm, or 3 ppm during the summer (Albert, 1988). Fishery resources appear to have re- sponded to the improved water qual- ity conditions; the size of the A. sapidissima spawning population has increased substantially (Maurice et al., 1987) and average catch of young- of-year M. saxatilis has increased by more than an order of magnitude1. Despite the size and importance of the Delaware estuary, little is known about which fish use the tidal freshwater region of the river as a spawning and nursery ground. The most recent regional ichthyoplankton survey was con- ducted in 1981, but covered 100 km of river with only 20 samples per week and only data for M. saxatilis were reported2. Numerous ichthyo-plankton collections were conducted in the mid-1970s as part of power plant or industrial- related impact studies, but these were individually limited in spa- tial extent and have never been integrated to provide a regional description. Most important, how- ever, no regional study has been conducted since the dramatic im- provements in water quality oc- curred. In this study we provide new data to describe present ichthyoplankton distribution and abundance in the tidal Delaware 1 Weisberg, S. B., W. H. Burton, and H. A. Wilson. 1991. Delaware River striped bass studies: population estimate of the 1990 year class and an evaluation of the young- of-year index of abundance. Rep. to the Delaware Basin Fish and Wildl. Manage. Coop., Trenton, NJ. - Himchak, P. J., J. Carlson, and R. Tilton. 1981. Spawning and recruitment of the striped bass, Morone saxatilis, in the Dela- ware River. Rep. available from New Jer- sey Dep. Environ. Protection, Trenton, NJ, Bureau of Marine Fisheries. 788 Weisberg and Burton Ichthyoplankton abundance and distribution in the Delaware River 789 River and its tidal tributaries, integrate data from the many ichthyoplankton collections conducted during the 1970s, and compare these data to describe how spawn- ing and nursery activity in the area has changed. Methods Ichthyoplankton samples were collected weekly from April to June in 1987 and 1988, following a stratified random design. Sampling in 1987 was conducted be- tween Riverton, New Jersey (river kilometer [rkm| 174) and the fall line at Trenton, New Jersey (rkm 214) for nine weeks beginning on 22 April (Fig. 1). The sam- pling area was stratified into four 10-km regions. Each region was stratified further into three habitats: shoal (river depth of less than 8m), channel bottom (within 1 m of the bottom in areas with an overall depth >8 m ), and channel mid-water. Three randomly placed samples were collected from each habitat in each region during each week. Sampling in 1988 was conducted between Bristol, Pennsylvania (rkm 191), and Artificial Island, New Jersey (rkm 87), for eight weeks beginning on 18 April (Fig. 1). Allocation procedures in 1988 were the same as in 1987, except that the area was stratified PENNSYLVANIA TRENTON NEW JERSEY RANCOCAS CBEB< UPPER REGION CHRISTINA RIVER CSD CANAL MID-RIVER REGION LOWER REGION 1987 SAMPLING AREA fil 1988 SAMPLING AREA Figure 1 Map of the study area. into nine equally-sized regions, and in the channel bot- tom habitat, only two samples were collected. The sampling of the river in 1987 also included tribu- tary stations in Martin's Creek, Neshaminy Creek, Dredge Harbor, and Rancocas Creek. In 1988, tribu- tary stations were established in the Schuylkill River, Raccoon Creek, Oldman's Creek, the Christina River, and at two stations in the Chesapeake Bay and Dela- ware River (C&D) Canal. Tributary stations were lo- cated approximately 1 km upstream of the confluence with the Delaware River, except for the second C&D Canal station, which was located near the Rt. 13 bridge, about 4 km from the river. One mid-water sample was collected from each tributary station each week, ex- cept for the C&D Canal stations, where both mid-wa- ter and bottom samples were taken weekly. Samples from the bottom habitat were collected by using a 1-m diameter epibenthic ichthyoplankton sled; water column and shoal habitats were sampled using stepped oblique tows with a 0.5-m bongo net. Both types of gear were fitted with 505-^m mesh plankton nets. The volume of water that each net filtered was estimated with a General Oceanics digital flowmeter mounted at the mouth of the epibenthic sled and to one side of the bongo net. Tows were conducted against the current for five minutes at a speed of approximately l.Om/sec. The catch was preserved in 101 formalin, stained with rose bengal, and taken to the laboratory for analysis. Physico/chemical data (temperature, dissolved oxygen, pH and conductivity) were collected weekly with a calibrated Hydrolab Surveyor II in each of the river regions, the C&D Canal, and at each of the tributary stations. Measurements in the river and in the C&D Canal were taken at surface, mid-depth, and bottom. In the tributaries, measurements were taken at surface and bottom. Water trans- parency was measured with a 20-cm Secchi disk. In the laboratory, all eggs with diam- eters greater than 1.0 mm were removed from the collections and inspected to de- termine if they were M. saxatilis or A. spidissima based on the size of the perivitelline space relative to the diameter of the egg. Small eggs ( less than or equal to 1.0mm in diameter) consisted mostly of M. americana, Alosa aestivalis (blueback herring), or Alosa pseudoharengus (alewife), but were not differentiated quantitatively. M. americana yolk-sac larvae were dis- 790 Fishery Bulletin 91(4), 1993 tinguished from M. saxatilis larvae based on the smaller size of M. americana at this stage and on gut morphology, which in M. americana parallels the noto- chord for about six myomeres before turning down- ward. Post yolk-sac M. americana and M. saxatilis lar- vae less than 8 mm in length were separated by size and the formation of fin rays, (formation of fin rays occurs at about 5 mm and 8mm, respectively). Post yolk-sac larvae of 8 mm and longer were identified by clearing and staining the specimens and examining the pterygiophore/neural spine interdigitation patterns (Olney et al., 1983). A. sapidissima larvae were sepa- rated from Alosa spp. by counting post-anal myomeres (Lippson and Moran, 1974:i; Chambers et al., 1976). For the purpose of data presentation, the Delaware River sampling effort was post-classified into lower river, mid-river, and upper river regions (Fig. 1). These regions corresponded to Artificial Island (rkm 87) to the Christina River (rkm 114), Christina River to the Schuylkill River (rkm 148), and Schuylkill River to Trenton, NJ (rkm 214), respectively. Density estimates were calculated for each region during each week as the simple average of all samples collected in that re- gion. A Friedman's rank-sum test, blocked on week, was used to test for differences in density among re- gions (Conover, 1980). Data from over 2,000 ichthyoplankton samples col- lected from the Delaware River during the 1970s were obtained from reports submitted as part of regulatory requirements or environmental impact studies (ANSP, 19744; Anselmini, 19745; Anselmini 1976K; Potter et al., 1974, a and b78; Harmon and Smith 19759; PECO, 1977, a_eio,ii.i2.i3,i4. ^ 1979> a and buue. RMC, 197917). While none of these studies individually provided coverage of the entire tidal river, collectively they provided excel- lent coverage. To compare present ichthyoplankton com- position and abundance with that in the 1970s, these data were digitized and summarized. To maximize comparability of data between periods, we extracted for analysis from the historical reports only those samples which were collected between mid- April and mid-June with a 500-(im mesh gear that was towed for at least five minutes and filtered at least 40 m3. Where studies differed in taxonomic level of identi- fication, data from all studies were restated to the higher taxonomic group. From these data, weekly average den- sities were estimated for each region and year. Historic abundances for each region were taken as the simple averages of weekly estimates. We included data for only years in which collections were taken in at least eight weeks between April and June to ensure that our com- parison was not biased by data sets with limited tempo- ral coverage. Comparisons of average abundance with the present study were accomplished separately for each region by using a Friedman's test blocked on week. Results Water quality Water temperature during the 1987 and 1988 survey ranged from 11 to 26° C. Differences in temperature between regions were less than 1°C in both years. ' Lippson, A. J., and R. L. Moran. 1974. Manual for identification of early developmental stages of fishes of the Potomac River estuary. Available from Maryland Department of Natural Resources. Ref. No. PPSP-MP-13, 282 p. I ANSP (Academy of Natural Sciences of Philadelphia). 1974. Eco- logical studies in New Jersey, Oldman's Creek, Raccoon Creek, Birch Creek, and the Delaware River 1972-1973. Prep, by ANSP, Phila- delphia, Pennsylvania for the Shell Oil Company, Philadelphia, PA. 5 Anselmini, L. D. 1974. An ecological study of the Delaware River in the vicinity of the Mercer Generating Station, Trenton, NJ. Prep, by Ichthyological Associates, Inc., Ithaca, NY, for Public Service Elec- tric and Gas Co., Newark, NJ. h Anselmini, L. D. 1976. An ecological study of the Delaware River in the vicinity of Newbold Island. Prep, by Ichthyological Associates. Inc., Ithaca, NY, for Public Service Electric and Gas Co., Newark, NJ. ; Potter, W. A., D. C. Smith, and P. L. Harmon. 1974a. An ecological study of the Delaware River in the vicinity of Chester Generating Station. Chester Progress Rep. 1. Prep, by Ichthyological Associates, Inc., Ithaca, NY, for the Philadelphia Electric Co., Philadelphia, PA, 94 p. "Potter, W. A., D. C. Smith, and P. L. Harmon. 1974b. An ecological study of the Delaware River in the vicinity of Eddystone Generating Station. Eddystone Progress Rep. 3. Prep, by Ichthyological Associ- ates, Inc., Ithaca, NY, for the Philadelphia Electric Co., Philadel- phia, PA, 42 p. " Harmon, P. L., and D. C. Smith. 1975. An ecological study of the Delaware River in the vicinity of Eddystone generating station. Eddystone Progress Rep. 4 to Philadelphia Electric Co., Philadel- phia, PA. 1,1 PECO (Philadelphia Electric Company). 1977a. Chester Generat- ing Station, 316(b) Rep., Permit No. PA 0011614. Rep. to the EPA, Philadelphia, PA. " PECO. 1977b. Richmond Generating Station, 316(b) Rep., Permit No. PA 0011649. Rep. to the EPA, Philadelphia, PA. 12 PECO. 1977c. Southwark Generating Station, 316(b) Rep., Permit No. POA 00116-65. Rep. to the EPA, Philadelphia. PA. 1:1 PECO. 1077d. Delaware Generating Station, 316(b) Rep.. Permit No. PA 0011622. Rep. to the EPA, Philadelphia, PA. II PECO. 1977e. Eddvstone Generating Station, 316(b) Rep., Permit No. PA 00137-14. Rep. to the EPA, Philadelphia, PA. '' Ichthyological Associates (IA). 1979a. Effect of the cooling water intake structure, intrainment and impingement of fishes. Burlington Generating Station Demonstration for Section 316(b) of the Federal Water Pollution Control Act Amendments of 1972, PL 95-500. Prep, by IA, Ithaca, NY, for Public Service Electric and Gas Co., Newark, NJ. 16 Ichthyological Associates (IA). 1979b. Effect of the cooling water intake structure, intrainment and impingement of fishes. Mercer Gen- eration Station Demonstration for Section 316(b) of the Federal Wa- ter Pollution Control Act Amendments of 1972, PL 95-500. Prep, by IA, Ithaca, NY, for Public Service Electric and Gas Co., Newark, NJ. 17 Radiation Management Corporation (RMCl. 1979. An evaluation of the cooling water intake at the Edge Moor Power Station. Edge Moor Power Station section 316(b) Evaluation. Permit No. DE- 000058. Prep, by RMC, Pottstown, PA, for Delmarva Power and Light Co., Wilmington. DE. Weisberg and Burton: Ichthyoplankton abundance and distribution in the Delaware River 791 Temperature, dissolved oxygen, and salinity were not vertically stratified in the water column and did not differ substantially between tributaries and the near- est mainstem water quality sampling locations. Average dissolved oxygen values decreased over the sampling period during each year of the study, rang- ing from greater than 10 mg/L during the first week to less than 8 mg/L during the last week. Dissolved oxy- gen differed by as much as 3 mg/L among regions, with the lowest values occurring downstream of Philadel- phia. However, only a single measurement below 5 mg/ L was recorded in 1988, and no measurements below 6 mg/L were recorded in 1987. The river was consistently fresh upstream of Pea Patch Island (km 98); salinity averaged less than 2 ppt in the regions downstream of Pea Patch Island. The highest salinity (5 ppt) was observed in the C&D Ca- nal during the second week of sampling in 1988. A storm and subsequent freshwater runoff during the sixth week of the 1988 sampling turned the entire mainstem sampling area fresh but had little effect on salinity in the C&D Canal, which never fell below 2 ppt. Distribution of eggs and larvae Eighteen taxa, comprising 15 families, were collected from the Delaware River during 1987 and 1988 (Table 1). Seventeen of these taxa also were found in the tributaries (Table 2); no taxon occurred only in the tributaries. Most taxa were not abundant. Only three, Table 1 Average density (no./lOO m1 of ichthyoplankton in the lower, mid- and upper regions of the Delaware River during the present and previous (1972-1978) studies . Values in bold represent regional densities that were significantly (P<0.05) different between the present and previous studies. Present study Previous studies Lower Mid Upper Lower Mid Upper river river river river river river Percichthyidae Morone saxatilis (eggs) 0.2 0.2 <0.1 0.1 0.3 0 Morone saxatilis 0.1 0.4 0.1 0.3 0.1 0 Morone americana 0.5 8.4 48.3 1.1 0.4 27.2 Clupeidae Alosa sapidissima (eggs) 0.1 0.1 0.4 0 0 0.1 Alosa sapidissima 0.8 15.0 49.0 0 0 0 Other Alosa spp. 0.2 5.4 33.5 4.1 22.8 339.7 Breioortia tyrannus 0.1 0.1 0 0 0 0 Dorosoma cepedianum 0 0 0 0 <0.1 0 Ictaluridae 0 0 0.1 0 0 0 Anguillidae Anguilla rostrata <0.1 <0.1 <0.1 0.8 0.4 0.2 Percidae Perca flavescens <0.1 0 0.7 0 0 <0.1 Etheostoma sp. <0.1 <0.1 0.2 0 0 0.6 Cyprinidae 0.1 1.3 0.5 1.0 6.8 10.5 Soleidae Trinectes maculatus 0.1 0.1 <0.1 0 0 0 Centrarehidae Lepomis sp. <0.1 <0.1 0.1 0 0 0 Pomoxis sp. 0 0 0 0.1 0 <0.1 Catostomidae Carpiodes cypruuis <0.1 0.4 0.6 0 0 3.4 Castostomus sp. 0 0 0 0 0 1.0 Engraulidae Anchoa mitchilli 0.6 0.1 <0.1 0.2 0 0 Gasterosteidae <0.1 <0.1 <0.1 0 0 0 Cyprinodontidae Fundulus sp. 0 0 <0.1 00.1 <0.1 Sciaenidae Leiostomus xanthurus 0.5 0.1 <0.1 <0.1 0 0 Atherinidae Menidia sp. 0.2 <0.1 <0.1 0.4 0 0 Umbndae Umbra pygmaea 0 0 0 0 0 <0.1 792 Fishery Bulletin 91(4), 1993 Table 2 Average density (no./lOO m3) of ichthyoplankton in the Delaware River tributaries among samples collected in 1987 and 1988. C&D Christina Oldman's Raccoon Schuylkill Dredge Rancocas Neshaminy Martin's Canal River Creek Creek River Harbor Creek Creek Creek Percichthyidae Morone saxatilis (eggs) 5.7 0.2 0 0.2 0 0 0 0 0 Morone saxatilis 0 0 0.7 1.5 0 0 0 0 0 Morone americana 0.1 21.5 8.0 30.9 1.4 4.0 36.9 8.8 2.4 Clupeidae Alosa sapidissima (eggs) <0.1 0 0 0 0 0 1.6 0 0 Alosa sapidissima 0 41.4 168.4 247.3 1.3 80.7 199.9 46.6 75.5 Other Alosa spp. 0 28.6 106.4 160.6 1.0 52.0 194.5 52.7 57.8 Anguillidae Auguilla rostrata 0 0 0 0 0 0 0 0.1 0 Percidae Perca flavescens <0.1 1.0 0 0 0 0 1.3 0 0 Etheostoma sp. <0.1 2.2 0.2 0.1 0 0 0.1 0.1 0 Cyprinidae 0 0.6 0.8 7.4 0.6 0 <0.1 2.1 0 Soleidae Trinceetes maeulatus 0.1 0 0 0 0 0 0 0 0 Centrarchidae Lepomis sp. 0 0 0 0.4 0 0.3 0.6 0.4 0.5 Catostomidae Carpiodes cyprinus 0 2.8 0.2 0.8 0 0.1 0.7 0.2 0 Engraulidae Anchoa mitchilli 0.1 0 0 0 0 0 0 0 0 Gasterosteidae 0.1 0 0 0 0 0 0 0 0 Cyprinodontidae Fundulus sp. 0 0 0 0 0 0 <0.1 0 0 Sciaenidae Leiostomus xanthurus 1.8 0 0 0 0 0 0 0 0 Atherinidae Menidia sp. 0.1 0 0 0 0 0 0 0 0 I PRESENT STUDY (1987-1988) j PREVIOUS STUDIES ! (1972-1978) O ■■ S ■> - o z LU o w UJ I * A a* ,«P .A ,<* * A <(• * >A/osa spp. larvae Morone americana larvae £ o S. LU X Perca flavescens larvae a UJ E E > .* A.* J> J? -V-0 A* A* A* V 4* *,' Morone saxatilis larvae >- o f z a" LU E« LU | 3 * LU IT if A .» ,$> A * <> j.^ .J- * WEEK K* A^ J** / A* A* A* V J jf WEEK Figure 4 Temporal patterns of abundance for the dominant larval fish of the tidal Delaware River. Solid and dashed lines represent 1987 and 1988, respectively. stream (Glebe and Leggett, 1981; Maurice et al., 1987), could get past Philadelphia before hypoxic conditions developed. Second, only fish migrating out of the river late in the season, that is fish spawned in the most upstream reaches, arrived in the Philadelphia area after the lethal dissolved oxygen conditions dissipated. Our data suggest that the spatial extent of A. sapidissima spawning habitat in the Delaware Rivet- has expanded over the last decade, coincident with the water quality improvements near Philadelphia. While we found no evidence of spawning or nursery activity in tidal freshwater during the 1970s, prior to water quality improvements, by the late 1980s the density of larvae there was as high as in non-tidal upstream reaches. The suggestion of expanded spawning ground for shad is consistent with Maurice et al. (1987), who documented substantial increases in spawning activ- ity during the last decade in the lower reaches of the 796 Fishery Bulletin 91|4), 1993 non-tidal river. Presumably, expansion of the spawning and nursery grounds is a first step in recovery of the stock. There was little difference in M. americana abun- dance in the Delaware River between our study and the historical data (Tables 1 and 3), suggesting that the poor water quality may have had little impact on M. americana spawning and nursery activities in the Dela- ware River. We observed that most M. americana lar- vae were concentrated upstream of Philadelphia in an area where water quality was not as seriously impacted as below Philadelphia (Albert, 1988). Unlike A. sapidissima, which leave the estuary and must return through the Philadelphia region to spawn, M. americana remain in the estuary throughout their life cycle. Pre- sumably, these life history characteristics have allowed M. americana to avoid effects of the poor water quality of the Philadelphia area more effectively than shad. We found that larval M. saxatilis were primarily con- centrated in the area downstream of Philadelphia. Since this area historically had the worst water quality, the M. saxatilis population should have benefitted from wa- ter quality improvements. The Delaware River beach seine monitoring survey, conducted annually by the State of New Jersey, showed a consistent increase in young- of-year M. saxatilis abundance throughout the 1980s1. However, we did not find the density of eggs to be higher in our study than in the 1970s, and the density of larvae was only slightly higher. We did find that present densities of M. saxatilis eggs and larvae in the Dela- ware River are considerably less than in the Hudson River1" or the Chesapeake Bay ( Setzler-Hamilton et al., 1981; Grant and Olney, 1991). This may be due, in part, to a year effect, since juvenile abundance in the Dela- ware River in 1988, as measured by the beach seine survey, was lower than in any other year since 1985. However, it is possible that higher juvenile abundance in the late 1980s resulted from better survival of larvae as water quality improved (Burton et al., 1992), rather than from increased egg production. We found that the density of M. saxatilis eggs was considerably higher in the C&D Canal than in the mainstem Delaware River. This finding seems to sup- port studies suggesting that the mechanism for main- tenance of the Delaware River M. saxatilis population is transport of eggs spawned in the Chesapeake Bay through the C&D Canal (Chittenden, 1971; Johnson and Koo, 1975). Our data, however, are probably more consistent with that of Kernehan et al. (1981), who 18 Lawler, Matusky, and Skelly Engineers iLMS). 1989. 1986 and 1987 year class report for the Hudson River estuary monitoring program. Prep, bv LMS. Pearl River. NY, for Consolidated Edison Co., New York, NY. suggested that most of the eggs transported through the canal are not viable. In spite of finding consider- ably elevated densities of eggs in the C&D Canal, no larvae were found there, and we collected fewer than 10 larvae within 10 km of where the C&D Canal emp- ties into the Delaware River estuary. Most of the M. saxatilis larvae we found in the Delaware River were collected between the Delaware Memorial (rkmlll) and Commodore Barry (rkml32) bridges, more than 30 km upstream of the C&D Canal. This distance rep- resents about three tidal excursion distances, and while upstream transport of eggs is possible in some estua- rine systems (Norcross and Shaw, 1984), it is unlikely to occur in this portion of the Delaware because of the lack of a well-defined thermocline or pycnocline that would allow for two-layer circulation. Thus, it is more likely that most of the M. saxatilis larvae found in the Delaware River were actually spawned in the river. Acknowledgments We would like to thank P. Kazyak, J. Gurley, A. Brindley, J. McGroder, M. Young, and C. DeLisle for their consid- erable efforts in both the field and laboratory aspects of this project. We would also like to thank W Richkus, J. Frithsen, S. Beck and C. DeLisle for helpful comments on the manuscript, and J. Miller for his help in all aspects of the project. This work was funded as part of a continuing striped bass management effort by the Delaware Basin Fish and Wildlife Management Coop- erative, which includes the Delaware Division of Fish and Wildlife, Pennsylvania Fish Commission, New Jer- sey Division of Fish, Game and Wildlife, New York Di- vision of Fish and Wildlife, U.S. Fish and Wildlife Ser- vice, and the National Marine Fisheries Service. Literature cited Albert, R. C. 1988. The historical context of water quality manage- ment for the Delaware estuary. Estuaries 11:99-107. Burton, W. H., S. B. Weisberg, A. Brindley, and J. A. Gurley. 1992. Early life stage survival of striped bass in the Delaware River, USA. Archives for Environmental Contamination and Toxicology 23:333-338. Chambers, J. R., J. A. Musick, and J. Davis. 1976. Methods of distinguishing larval alewife from lar- val blueback herring. Chesapeake Science 17:93-100. Chittenden, M. E. 1971. Status of the striped bass, Morone saxatilis, in the Delaware River. Chesapeake Science 12:131-176. 1974. Trends in the abundance of American shad, Alosa sapidissima, in the Delaware River basin. Chesa- peake Science 15:96-103. Weisberg and Burton Ichthyoplankton abundance and distribution in the Delaware River 797 1976. Present and historical spawning grounds and nurseries of American shad, A/osa sapidissima, in the Delaware River. Fish. Bull. 74:343-352. Conover, W. J. 1980. Practical nonparametric statistics. Second ed., New York. John Wiley and Sons. Inc. Glebe, B. D., and W. C. Leggett. 1981. Temporal, intra-population differences in energy allocation and use by American shad (A/osa sapidissima) during the spawning migration. Can. J. Fish. Aquat. Sci. 38:795-805. Grant, G. C, and J. E. Olney. 1991. Distribution of striped bass Morone saxatilis (Walbaum) eggs and larvae in major Virginia rivers. Fish. Bull. 89:187-193. Johnson, R. K., and T.S.Y. Koo. 1975. Production and distribution of striped bass {Morone saxatilis I eggs in the Chesapeake and Dela- ware Canal. Chesapeake Science 16:39-55. Kernehan, R. J., M. R. Hendrick, and R. E. Smith. 1981. Early life history of striped bass in the Chesa- peake and Delaware Canal and vicinity. Trans. Am. Fish. Soc. 110:137-150. Maurice, K. R., R. W. Blye, and P. L. Harmon. 1987. Increased spawning by American shad coinci- dent with improved dissolved oxygen in the tidal Dela- ware River. In Proceedings of the international symposium on common strategies of anadromous and catadromous fish, p. 79-88. Am. Fish. Soc, Bethesda, MD. Norcross, B. L., and R. F. Shaw. 1984. Oceanic and estuarine transport offish eggs and larvae: a review. Trans. Am. Fish. Soc. 113:153-165. Olney, J. E., G. C. Grant, F. E. Shultz, C. L. Cooper, and J. Hageman. 1983. Ptergiophore interdigitation patterns in larvae of four Morone species. Trans. Am. Fish. Soc. 112:525-531. Setzler-Hamilton, E. M., W. R. Boynton, J. A. Mihursky, T. T. Polgar, and K. V. Wood. 1981. Spatial and temporal distribution of striped bass eggs, larvae and juveniles in the Potomac estuary. Trans. Am. Fish. Soc. 110:121-136. Effects of entanglement and escape from high-seas driftnets on rates of natural mortality of North Pacific albacore, Thunnus alalunga Richard W. Brill Honolulu Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service, NOAA 2570 Dole Street, Honolulu. Hawaii 96822-2396 David B. Holts La Jolla Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service, NOAA RO. Box 271, La Jolla, California 92038-0271 North Pacific albacore {Thunnus alalunga ) are heavily fished and may have experienced recent reductions in recruitment (NOAA, 1991). U.S. troll fishery landings and catch per unit of effort (CPUE) have been de- creasing since the 1960s and are cur- rently at only about 50% of their peak levels (Kleiber and Perrin, 1991; NOAA, 1991). Landings and CPUE by the Japanese pole-and-line fishery have also been falling con- tinuously since the late 1970V. North Pacific albacore are addition- ally targeted by longline fleets, and considerable numbers are taken by high-seas large- and small-mesh driftnet fleets. The small-mesh fleets of Japan, Taiwan, and Korea target neon flying squid (Ommastrephes bartramii) and incidentally capture albacore. The large-mesh fleets of Japan and Taiwan target albacore, skipjack tuna (Katsuwonus pelanus I, and various billfishes (Istiophoridae and Xiphiidae). Foreign driftnet fleets and U.S. trailers taking albacore overlap geo- graphically (Fig. 1). They also tar- get a common stock even when fishing in widely separated areas because of the trans-Pacific migra- tion of albacore (Otsu and Uchida, 1963). Albacore that encounter drift nets, and escape to survive long enough to be recaptured by another fishery can bear some external marks- '. These marks provide di- rect evidence of interactions among fisheries. U.S. trailers operating in the North Pacific have reported an increased frequency of net-marked fish with the expansion of the Tai- wanese, Korean, and Japanese high-seas driftnet fleets (NOAA, 1991;"). Albacore that become entangled in drift nets face a number of pos- sible fates (Fig. 2). The fraction that become entangled and subsequently drop out (alive or dead) is not known. If it is significant, the North Pacific albacore stock could be af- 'Tsuji. S., H. Nakano, and N. Bartoo. 1992. Report of the Twelfth North Pacific Albacore Workshop. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. LJ-92-04, 15 p. 798 -Bartoo, N., D. B. Holts, and C. Brown. 1991. Report of the 1990 cooperative albacore ob- server project. Dep. Commer, NOAA, Natl. Mar. Fish. Serv., NOAA, Southwest Fish. Sci. Cent.. B.O. Box 271. La Jolla. CA 92038. Admin. Rep. LJ-91-09, 16 p. 'Bartoo. N., D. B.. Holts, and L. Halko. 1992. Report of the 1991 cooperative North Pacific albacore observer project. Dep. Commer, NOAA, Natl. Mar. Fish. Serv., Southwest Fish. Sci. Cent., P.O. Box 271, La Jolla, CA 92038. Admin. Rep. LJ-92-07. fected more seriously by high-seas drift nets than landings alone indi- cate. Our study, part of a larger ef- fort to determine the impacts of drift nets on the North Pacific alba- core stock, was undertaken to de- termine whether albacore that en- counter drift nets, and subsequently escape alive, suffer higher than nor- mal rates of natural mortality. Our strategy was to look for dif- ferences in measures of fitness, and differences in short-term and long- term growth rates between net- marked (i.e., fish that have encoun- tered drift nets and escaped) and unmarked albacore. We also looked for indications of bacterial infection and long-term stress in net-marked albacore. These parameters could evince, although not quantify, dif- ferences in rates of natural mortal- ity We had to use indirect measures because rates of natural mortality (or physiological changes) could not be directly observed in albacore en- countering simulated drift nets in shoreside tanks. Unlike skipjack and yellowfin tuna {Thunnus albacares) (Brill, 1992), albacore have yet to be successfully main- tained in captivity. Selection of specific parameters was constrained by the necessity of collecting measurements and samples at sea aboard small com- mercial fishing vessels. We could ac- quire only measurements and samples that could be obtained un- der difficult conditions and that did not interfere with normal fishing operations. We were further limited to collecting samples that could be stored for later processing in the laboratory and that would not re- duce the market value of the catch. Given these constraints, we chose the following parameters. Weight to length and maximum girth to length ratios were used as indica- tors of general fitness (Bolger and Manuscript accepted 14 May 1993. Fishery Bulletin, U.S. 91:798-803 1 1993). NOTE Brill and Holts. Natural mortality of North Pacific Thunnus alalunga from drift nets 799 55 45 35° 25- 15' -^2 DRIFT NET FISHERY 180° 170° Hawaiian Islands 1 . I 140° "l30° 120 160 150° Figure 1 Area of operation of the U.S. troll fishery (light cross-hatching) catching albacore. Track lines of U.S. vessels carrying National Marine Fisheries Service observers in 1991 are shown by the solid line. Area of operation of foreign drift net fleets (heavy cross-hatching) extends west to approximately 140°W. infection and long term stress (Blaxhall, 1972; Anderson, 1990; Goede and Barton, 1990). Although muscle lipid content can be used as a measure of recent feeding history (Dotson, 1978), it was not used in this study. The lipid content of tuna muscle is dependent on sampling site (Dotson, 1978; Vileg et al, 1983; Boggs and Kitchell, 1991) and this factor could not be rigor- ously controlled. Materials and methods ENTANGLED /x LANDED ESCAPE DROP OUT DEAD ALIVE BADLY INJURED NOT BADLY INJURED, DO NOT BEGIN FEEDING BEGIN FEEDING Figure 2 The possible fates of albacore that become entangled in high- seas drift nets. Note that the condition of fish which escape drift nets alive probably form a continuum from those that are so badly injured that they succumb quickly, to those that are only slightly injured and resume feeding immediately. Connolly, 1989; Cone, 1989, 1990; Murphy et al, 1991). Relative otolith weight (i.e., the ratio of otolith weight to body weight) was used as an indicator of long-term growth rate (Boehlert, 1985; Pawson, 1990; Fletcher, 1991). (Because otolith weight is proportional to fish age rather than to body size, slower-growing fish have larger otoliths at a given body size because they have taken longer to reach that body size. ) RNA:DNA ratio of white muscle was measured because it is an estab- lished indicator of short-term (days to weeks) growth rates (Bulow, 1970; Bulow et al, 1981; Ferguson and Danzmann, 1990; Mustafa et al., 1991). Relative leu- kocyte abundance (i.e., the ratio of leukocytes to red blood cells) was assessed as an indicator of bacterial During 1990 and 1991, observers were placed aboard ten U.S. trollers by the Southwest Fisheries Science Center (NMFS, NOAA), in cooperation with the Western Fishboat Owners Association. Ob- servers documented the number of net- marked albacore landed and collected data and samples for our study. Boats operated in an area of the North Pacific shown in Figure 1. Albacore were classified according to the se- verity of net damage with a scheme (Table li devel- oped by Bartoo et al.2. Fork length, maximum girth (both to the nearest centimeter), and body weight (to the nearest kilogram) were measured. Dried blood smears, sagittae (otoliths), and white-muscle samples (frozen in sealed tubes in the boats' commercial fish freezers) were also obtained from a subsample of landed fish selected by the observers. Muscle samples (taken from the small amount of muscle remaining near the skull following decapitation) and otoliths could be ob- tained only on vessels that processed (i.e., "gilled and gutted") their catches. Complete sets of samples were not, therefore, obtained from every fish. In the laboratory, otoliths were dried, cleaned, and weighed to the nearest 0.1 mg on a Sartorius 1207 MP2 electronic balance. RNA and DNA were extracted from the frozen muscle samples by standard proce- dures (Hutchison and Munro, 1961; Munro and Fleck, 1966; Wilder and Stanley, 1983). The only modifica- tions were that the RNA was extracted with 1.0 N NaOH, and the DNA with 1.2 N perchloric acid, both at 37°C for 60 minutes. RNA. DNA, and protein con- centrations were measured with a Beckman model 35 UV spectrophotometer using the dual wave length method (Tsanev and Markov. 1960) and standard ex- tinction coefficients (Wilder and Stanley, 1983). Red cells and leukocytes were counted in Geimsa-stained blood smears with the aid of light microscopy (oil im- mersion lens) in a minimum of three microscope fields, or until at least 200 red cells were counted. Eosino- phils could be easily differentiated from the remainder 800 Fishery Bulletin 91(4). 1993 Table 1 Net damage classification scheme (Bartoo et al.2) used to assess severity of caused by entanglement and escape from drift nets, and the number fish in category examined for our study. injury to albacore each damage code Damage code Damage description Number offish 0 No damage or scarring. 217 1 Minor damage along side! si offish, pattern of stripes due to minor scale loss where fish forced its way through or along the net. 39 2 Minor damage to head, chiefly forward of pectoral fins, brush-like pattern of scale loss. 5 3 Severe damage from bruising or scraping away of parts of the skin, primarily in area of greatest girth and mostly on dorsal surface. 1(1 4 Damage of any degree that was partially or completely healed. 60 of the leukocyte population and were excluded from the total leukocyte count. Although Alexander et al. (1980) described the distinguishing features of alba- core blood cells, we were only able to differentiate eosi- nophils clearly from the remainder of the leukocytes. Leukocytes were, therefore, not further classified. By using data from unmarked fish only, regressions of body weight on fork length, maximum girth on fork length, and otolith weight on body weight were fitted by a least squares procedure (Statgraf Statistical Soft- ware) to the exponential equation: o-x" Regression parameters a and b were estimated for each case. Relative condition factor, relative girth, and relative otolith weight for individual fish were then calculated with the regression parameters: K=W/(a-X"). When K = relative condition factor, W = body weight and X = fork length; when K = relative girth, W = maximum girth and X = fork length; and when A' = relative otolith weight, W = otolith weight and X = body weight. Relative condition factor, relative girth, and relative otolith weight were calculated to permit direct com- parison of groups containing individuals of different body sizes (Pollard, 1972; Brill et al„ 1987). The standard condition factor (C) was calculated separately for unmarked and net-marked fish: C = W/FL3 where W = body weight and FL = fork length. Fish were grouped only as net-marked (damage codes 1-4, Table 1) and unmarked (damage code 0). Differ- ences in mean values for unmarked and net-marked fish were evaluated by unpaired Student's i-tests with P < 0.05 taken as the maximum level for statistical significance. Results and discussion The only significant differences between net-marked and unmarked fish were in measures of absolute size (mean fork length, body weight, and maximum girth), which were greater for unmarked albacore (Table 2). These data suggest that smaller fish escape drift nets, survive, and are recaptured by the troll fleet more frequently than large fish. This appears to be con- firmed by size-frequency distributions for the 217 net- marked and 114 unmarked fish included in this study (Fig. 3A). The two peaks in Figure 3A (at approxi- mately 65 and 80 cm) represent 3- and 4-year-old fish, respectively' and imply that 4-year-old individuals pre- dominate the unmarked group. In the net-marked group, 3- and 4-year-old fish are approximately equally represented. This result is, however, more likely due to either a sampling bias of the observers when choos- ing fish for inclusion in this study, or to time-area variations in fish size composition and the percentage of net-marked fish. The size-frequency distributions of NOTE Brill and Holts Natural mortality of North Pacific Thunnus alalunga from drift nets 801 Table 2 Parameters measured to evaluate the effects of escapement from drift nets on the natural mortality of albacore alalunga. N equals the number of observations. Thunnus Parameters Unmarked fish Net-marked fish Mean ± SD N Mean ±SD N Body weight* ikg) 10.6 ±2.4 182 8.6 ±2.6 83 Fork length* (cml 79.1 + 7.4 215 73.5 ±7.9 114 Maximum girth* (cm) 56.1+5.7 214 51.8 ±6.2 113 Condition factor (kg/cm3) 0.0208 + 0.0015 182 0.0210 ±0.0017 83 Relative girth 1.00 + 0.02 214 1.00 ±0.03 113 Relative condition factor 1.00 ±0.07 182 1.00 ±0.07 83 RNA to DNA ratio 6.99 ±0.97 34 6.90 ± 1.49 36 Relative otolith weight 1.00 ±0.09 175 0.99 ±0.11 47 Leukocytes/200 red cells 7.26 + 5.30 15 6.44 ±3.59 35 *P<0.05. a much larger number of albacore ( 12,085 unmarked and 381 net-marked individuals^; and from which the fish used in this study are a subsample, show that the frequency distributions of 3- and 4-year-old net-marked and unmarked fish are more nearly identical (Fig. 3B); 3-year-old fish were found to be only slightly dominant in both groups. Our objective in calculating condition factor, relative girth, and relative condition factor (also referred to as relative weight) was to determine if there was any indi- cation of differences in fitness between unmarked and net-marked fish. No differences were found (Table 2). The use and misuse of condition factor and relative condition factor have received much recent attention (Bolger and Connolly, 1989; Cone, 1989; Murphy et al., 1991). The use of such indices in groups of fishes with different mean lengths and size-frequency distributions may not be justified because disparities may be due to differences in body size, rather than to physiological fitness (Cone, 1989, 1990). Because indices of fitness were not different for unmarked and net-marked fish, we considered the question of suitability moot. No differences were found either in measures of short-term (RNA:DNA ratio) or long-term (relative otolith weight) growth rates (Table 2). The RNA:DNA ratio can be affected by the age of the fish (Haines. 1973); however, it was not corrected for body size (i.e.. 'N. Bartoo. D. Holts, C. Brown, and L. Halko. La Jolla Laboratory, Southwest Fish. Sci. Cent., unpubl. data. age) because we found no correlation between it and body weight. The relative leukocyte counts also did not differ be- tween unmarked and net-marked albacore (Table 2). We assumed that net-marked albacore could show ei- ther elevated or reduced leukocyte counts. Acute bac- terial infection has been shown to increase leukocyte counts in fish (Wedemeyer et al., 1990). Conversely, reduced leukocyte counts can be caused by elevated circulating corticosteroid levels (an acute stress re- sponse) or by chronic bacterial infection causing leukocytolysis (Shreck, 1981; Wedemeyer et al., 1983). The lack of difference between unmarked and net- marked albacore suggests that none of the processes that can affect relative leukocyte counts was occurring in the latter. We were hampered in interpreting our results be- cause we could not determine precisely how long the fish were at liberty following escapement from drift nets. Based on observations of skipjack and yellowfin tuna held in shoreside tanks at the Kewalo Research Facility (described in Brill, 1992), skin damage result- ing from capture generally heals completely within weeks in feeding fish (R. W Brill, unpubl. observa- tions). Fish with damage codes 1-3 (Table 1), there- fore, probably encountered drift nets within one month of recapture. It was impossible to estimate how long fish with damage code 4 had been at liberty or how long net damage remains visible, although Bartoo et al.2 speculated that fish with damage code 4 may have been at liberty more than a year since encountering a 802 Fishery Bulletin 91(4), 1993 16 14 -- 12 w 10 >- o 3 Or UJ o ■ - UNSCARRED SCARRED | I . . .~N ■ 45 50 55 60 65 70 75 80 FORK LENGTH (cm) 95 10 >- <-> 6 Z LJ Z> S 4 UNSCARRED SCARRED B i it'i i 45 50 55 60 65 70 75 80 85 90 95 FORK LENGTH (cm) Figure 3 Size-frequency distribution of the unmarked and net-marked albacore sampled for our study (A) (215 unmarked and 114 net-marked fish) and much larger data set (B) (12085 un- marked and 318 net-marked fish I from which the fish used in our study are a subset (N. Bartoo, D. Holts, C. Brown, and L. Halko, SWFSC, La Jolla Laboratory, unpubl. data). drift net. It is unknown if this is sufficient time for differences in relative otolith weight to become mea- surable. Boggs and Kitchell (1991) showed that sig- nificant weight loss occurs in tunas within one week of starvation. Therefore, it was likely that net-marked fish were at liberty long enough for measures of physi- ological fitness to become apparent, assuming net- marked albacore had not been feeding prior to recap- ture. The rate at which relative leukocyte abundance changes in albacore following stress or infection is not known, but in coho salmon (Oncorhynchus kisutch) and rainbow trout (O. mykiss) leukocyte abundance changes within 96 hours of imposition of stressful conditions (McLeay and Gordon, 1977). Leukocyte abundance should be a good measure of differences in the health of net-marked and unmarked albacore, but the large standard deviation (Table 2) may limit its usefulness in this species. Failure to detect differences (Table 2) between un- marked and net-marked albacore suggests that the fish escaping from drift nets, and living long enough to become vulnerable to recapture by U.S. trailers, do not suffer increased natural mortality. Most of the fish however, recaptured by trailers are found to be only minimally damaged or to have healed scars (Table 1), which may explain the lack of differences between net- marked and unmarked fish. As shown in Figure 2, there is most likely a spectrum of damage caused dur- ing encounters with drift nets. We do not know if the distribution of damage severity in the population of albacore vulnerable to the troll fishery is the same as that of all the albacore that escape drift nets alive. Also, the number of albacore falling out dead, or es- caping so badly injured that they succumb before be- coming vulnerable to recapture by troll vessels, may well be significant but remains to be determined. Literature cited Alexander, N., R. M. Laurs, A. Mcintosh, and S. W. Russell. 1980. Haematological characteristics of albacore, Thunnus alalunga (Bonnaterre), and skipjack tuna, Katsuwonus pelamis Linnaeus. J. Fish. Biol. 16:383- 395. Anderson, D. P. 1990. Immunological indicators: effects of environmen- tal stress on immune protection and disease outbreaks. Am. Fish. Soc. Symp. 8:38-50. Blaxhall, P. C. 1972. The haematological assessment of the health of freshwater fish. J. Fish. Biol. 4:593-604. Boehlert, G. W. 1985. Using objective criteria and multiple regression models for age determination in fishes. Fish. Bull. 83:103-117. Boggs, C. H., and J. F. Kitchell. 1991. Tuna metabolic rates estimated from energy losses during starvation. Physiol. Zool. 64:502-524. Bolger, T., and P. L. Connolly. 1989. The selection of suitable indices for the mea- surement and analysis offish condition. J. Fish. Biol. 34:171-182. Brill, R. W. 1992. The Kewalo Research Facility, 1958-1992: over 30 years of progress. Southwest Fish. Sci. Cent., NMFS, NOAA, Honlulu, HI, Tech. Mem. TM 171, 41 p. Brill, R. W., R. Bourke, J. A. Brock, and M. D. Daily. 1987. Prevalence and effects of infection of the dorsal aorta in yellowfin tuna, Thunnus albacares, by the larval cestode, Dasyrhynchus talismani. Fish. Bull. 83:767-776. Bulow, F. J. 1970. RNA-DNA ratios as indicators of recent growth rates of a fish. J. Fish. Res. Board Can. 27:2343- 2349. NOTE Brill and Holts. Natural mortality of North Pacific Thunnus alalunga from drift nets 803 Bulow, F. J., M. E. Zeman, J. R. Winningham, and W. F. Hudson. 1981. Seasonal variations in the RNA-DNA ratios and in indicators of feeding, reproduction, energy storage, and condition in a population of bluegill, Lepomis macrochirus Rafinesque. J. Fish. Biol. 18:237-244. Cone, R. S. 1989. The need to reconsider the use of condition indices in fishery science. Trans. Am. Fish. Soc. 118:510-514. 1990. Comments: Properties of the relative weight and other condition indices. Trans. Am. Fish. Soc. 119:1048-1058. Dotson, R. C. 1978. Fat deposition and utilization in albacore. /;; G. D. Sharp, and A. E. Dizon (eds.), The physiological ecology of tunas, p. 343-355. Academic Press, NY. Ferguson, M. M., and R. G. Danzmann. 1990. RNA/DNA ratios in white muscle as estimates of growth in rainbow trout held at different temperatures. Can. J. Zool. 68:1494-1498. Fletcher, W. J. 1991. A test of the relationship between otolith weight and age for the pilchard Sardinops neopilchardus. Can. J. Fish. Aquat. Sci. 48:35-38. Goede, R. W., and B. A. Barton. 1990. Organismic indices and an autopsy-based assess- ment as indicators of health and condition of fish. Am. Fish. Soc. Symp. 8:93-108. Haines, T. A. 1973. An evaluation of RNA-DNA ratio as a measure of longterm growth in fish populations. J. Fish. Res. Board Can. 30:195-199. Hutchison, W. C, and H. N. Munro. 1961. The determination of nucleic acids in biological materials. Analyst 86:768-813. Kleiber, P., and C. Perrin. 1991. Catch-per-effort and stock status in the U.S. North Pacific albacore fishery: reappraisal of both. Fish. Bull. 89:379-386. McLeay, D. J., and M. R. Gordon. 1977. Leucocrit: a simple hematological technique for measuring acute stress in salmonid fish, including stressful concentrations of pulpmill effluent. J. Fish. Res. Board Can. 34:2164-2175. Munro, H. N., and A. Fleck. 1966. Recent developments in the measurements of nu- cleic acids in biological materials. Analyst 91:78-88. Murphy, B. R., D. W. Willis, and T. A. Springer. 1991. The relative weight index in fisheries manage- ment: status and needs. Fisheries 16:30-38. Mustafa, S., J. -P. Langardiere, and A. Pastoureaud. 1991. Condition indices and RNA:DNA ratio in over- wintering European sea bass, Dicentrarchus labrax, in salt marshes along the Atlantic coast of France. Aquaculture 96:367-374. NOAA. 1991. Status of the Pacific ocean living marine re- sources of interest to the USA for 1991. U.S. Dep. Commer., NOAA Tech. Memo., NOAA-TM-NMFS- SWFSC-165, 78 p. Otsu, T., and R. Uchida. 1963. Model of the migration of albacore in the North Pacific Ocean. Fish. Bull. 63:33-44. Pawson, M. G. 1990. Using otolith weight to age fish. J. Fish. Biol. 36:521-531. Pollard, D. A. 1972. The biology of a landlocked form of the normally catadramous fish Galaxias maciilatus (Jenyns). IV. Nutritional cycle. Aust. J. Mar. Freshwater Res. 23:39-48. Shreck, C. B. 1981. Stress and compensation in teleostean fishes: re- sponse to social and physical factors. In A. D. Pickering (ed.), Stress and fish, p. 295-321. Academic Press, London. Tsanev, R., and G. C. Markov. 1960. Substances interfering with the spectrophotomet- ric estimation of nucleic acids and their elimination by the two wave-length method. Biochim. Biophys. Acta 42:442-452. Vileg, P., G. Habib, and G. I. T. Clement. 1983. Proximate composition of skipjack tuna Katsuwonus pelamis from New Zealand and New Caledonian waters. N. Z. J. Sci. 26:243-250. Wedemeyer, G. A., R. W. Gould, and W. T. Yasutake. 1983. Some potentials and limits of the leukocyte test as a fish health assessment method. J. Fish. Biol. 23:711-716. Wedemeyer, G. A., B. A. Barton, and D. J. McLeay. 1990. Stress and acclimation. In C. B. Shreck and P. B. Moyle (eds.), Methods for fish biology, p. 451- 489. Am. Fish. Soc, Bethesda, MD. Wilder, I. B., and J. G. Stanley. 1983. RNA-DNA ratio as an index to growth in salmo- nid fishes in the laboratory and in streams contami- nated by carbaryl. J. Fish. Biol. 22:165-172. Weight change of pink shrimp, Pandalus jordani after commercial harvest and handling Robert W. Hannah Oregon Department of Fish and Wildlife Marine Region, 2040 SE Marine Science Drive Newport, OR 97365 Neil T. Richmond Oregon Department of Fish and Wildlife Charleston Marine Laboratory 4475 Boat Basin Boulevard Charleston, OR 97420 The northeast Pacific trawl fishery for pink shrimp, Pandalus jordani, is unusual in that it is primarily managed by an aggregate size limit. Vessels are required to land catches with a minimum average size of 353 shrimp per kg (160 shrimp per lb). The regulation is intended to limit fishing mortality on age-one shrimp, and also to maintain the economic value of the catch (PFMC, 1981). The decision to retain shrimp from a par- ticular tow is made by fishermen at sea; however, the regulation is en- forced when the shrimp are landed in port, up to five days later. In gen- eral, a knowledge of how commer- cial fishing practices can influence the average weight of target species is important for effective manage- ment by an aggregate size limit (Kirkley and DuPaul, 1989). Accord- ingly, the objective of this study was to quantify changes in the average weight of pink shrimp from initial retention to shoreside delivery, us- ing normal commercial handling and icing procedures. Methods To assess the weight change of shrimp from commercial handling practices, we compared the weight of shrimp samples taken directly from the trawl nets of commercial 804 vessels at sea to the weight of the same samples during the subse- quent shoreside delivery. Four to eight replicate samples of approxi- mately 500 g were collected from each tow and weighed to the near- est 0.1 g with a magnetically damp- ened triple beam balance. The ac- curacy of the scale was checked by using a known weight prior to each trip and again prior to weighing the samples at the dock. The scale used had also been previously tested for accuracy at sea with blind weighings and had been shown to produce measurements with a coefficient of variation of less than 0.5% under a variety of sea conditions'. Each sample was placed in a 5-mm mesh bag, labeled and passed to a crew member. Shrimp were taken directly from the vessel's hopper and were not washed prior to weighing and bag- ging. The crew was instructed to distribute the sample bags throughout the other shrimp from the same tow. After placement of the samples, the crew shoveled and raked flake ice into the layers of shrimp, according to that vessel's normal procedures. 1 M. Saelens and R. W. Hannah. 1988. Or- egon. Dep. of Fish and Wildlife, Newport, OR 97365. Unpub. data. At the completion of the fishing trip, the bags were retrieved dur- ing unloading and inspected for punctures. Punctured bags were dis- carded. Each intact sample was opened, placed in a tub with a 7-mm mesh bottom to drain for one minute, and weighed again with the same scale. An estimate of the per- centage of ice in the gross load was obtained from processing plant records. Weight change was expressed as a percentage of the original sample weight. The average percent weight change for all samples combined, for each trip and for each day of fishing, was compared to a hypothetical weight loss of zero percent by using a one sample /-test. To facilitate comparisons of weight change be- tween trips or between individual days of fishing, 95% confidence in- tervals were constructed for each estimate of average weight loss. Scatter plots of percent weight change for individual samples ver- sus elapsed time were constructed for each trip and for all trips com- bined. Each graph was tested for a significant slope with simple linear regression. Results and discussion Seven sampling trips were com- pleted aboard commercial shrimp vessels between April and August 1988, lasting from 1 to 5 days. We recovered 70% of the mesh bags in- tact, resulting in weight change es- timates for 713 samples from 132 tows. Elapsed time from catch to unloading ranged from 4 to 121 hours. Percent weight change for the combined samples showed an ap- proximately normal distribution centered on a mean weight loss of 2.2%, with a 99% confidence inter- Manuscript accepted 4 June 1993. Fishery Bulletin, U.S. 91:804-807 ( 1993). NOTE Hannah and Richmond Weight change in Pandalus jordani after harvesting 805 Table 1 Mean percent weight change, standard error of the mean and number of days on ice, between capture and unloading, for 500-g samples of pink shrimp, Pandalus jordani (number of samples is in parentheses). Asterisks denote results significantly different from zero percent weight change. Trip Days on ice All days combined 1 9 3 4 5 1 -1.2 ±0.9 (15) -1.2 ±0.009 (15) 2 0.4 ±0.4 -0.3 ±0.5 -1.8 ±0.8 -4.1 ±1.3 -0.7 + 0.4 (11) (14) (9) (3) (37) 3 -5.9" ± 0.8*** -0.4 ±0.5 -1.8 ±0.5** 1.3 ±0.6* -3.1 ±0.9** -1.5 ±0.4*** (14) (25) (17) (35) (33) (124) 4 -2.1 ±0.3*** -1.3 ± 0.3*** 0.7 ±0.5 0.9 ±0.4* -0.2 + 0.2 (20) (28) 142) (30) (120) 5 -2.6 ±0.3 ' -5.4 ± 0.9*** -2.5 ±0.5*** -2.2 ± 0.8* -4.6 ±0.7*** -3.4 ±0.3*** (23) (23) (24) (19) (12) (101) 6 -4.7 ±0.3*** -4.1 ±0.4*** -5.2 ± 0.5*** -4.4 ±0.5*** -4.6 ±0.2*** (34) (43) (46) (32) (155) 7 -2.7 ±0.2*** -2.0 ±0.3*** -1.3 ±0.4" -2.5 ±0.3*** -0.4 ±0.5 -1.9 ±0.2*** (24) (36) (35) (45) (21) (161) All trips -2.2 + 0.1*** (713) ° No ice used on this day of fishing. -P<0.05; **P<0.01; ***P<0.001. Table 2 Summarj of sampling trips. ncluding trip ength (d), total net landed catch (kg) and the estimated percentage of ice in the gross oad ( from processing plant records Trip Total Estimated length catch percentage Trip (d) (kg) ice 1 1 600 35 2 4 7,900- 16 3 5 15,000 17 4 4 11,900 23 5 5 5,200 32 6 4 25,200 7 7 5 14.000 11 val of 1.9% to 2.5% (Table 1). Four of the seven trips showed a statistically significant average weight loss (P<0.05), whereas three did not. None of the trips showed a gain in average weight of the shrimp samples, despite ice deductions as high as 35% of the gross load (Table 2). The average weight change for samples from indi- vidual days and trips showed a large degree of vari- ability. Several of the regressions of percent weight change versus elapsed time were statistically signifi- cant (P<0.05); however, the sign of the slope coefficient was inconsistent (Table 3). No consistent patterns were observed between average weight loss and trip length, total catch, or the percentage of ice in the load (Tables 1 and 2). The average weight loss observed in samples which had been on ice for only a short period of time (Table 1 and Figure 1), in combination with the sig- nificant positive slope of several of the regressions, suggests that commercially handled, iced pink shrimp generally lose some weight very shortly after capture, and, in many instances, slowly gain weight for some time afterwards. However, the one regression with a positive intercept and significant negative slope (Table 3, trip 2), in combination with the variability of aver- age weight change (Table 1), demonstrate that weight change of shrimp under actual commercial handling and icing conditions is quite variable. The lack of a truly consistent pattern of weight change versus elapsed time suggests that Figure 1 may portray shrimp weight change under actual commercial fishing conditions as well as anything else presented. 806 Fishery Bulletin 91(4), 1993 Table 3 Summary of linear regression analysis of percent weight change on elapsed time for each trip and all trips combined. Model is Dercent Weight Change = a + b Elapsed Time). NS = not significant. Trip Intercept Coefficient P>¥ 1 -0.207 0.013 0.0002 2 0.023 -0.001 0.0001 3 — — NS 4 -0.024 0.00041 0.0001 5 — — NS 6 — — NS 7 -0.027 0.00014 0.0238 All trip ! -0.029 0.00013 0.00027 20 -20 20 40 60 80 100 120 140 Hours after capture Figure 1 Percent weight change versus elapsed time from capture to unloading for samples of trawl caught pink shrimp (N = 713) Billiard and Collins ( 1978), that iced Pandalus borea- lis gained 11% in weight in the first 1.5 days and maintained this gain for 8.5 days. However, the dif- ferent methods used readily explain the different findings of the two studies. Bullard and Collins ( 1978) worked with shrimp which had been held at -1.7° C for two hours and then rinsed and drained for 30 minutes to produce stable weights prior to icing. The shrimp employed in this study were very fresh, some- times jumping off the scales. If the average weight loss observed in this study occurs very rapidly, per- haps owing to loss of fluids just after the shrimp have died, or from the initial crush of the catch in the hold, such a loss could not have been detected in the Bullard and Collins (1978) experiment. It is not clear what percentage of ice was used by Bullard and Collins (1978), but they refer to it as "an excess of ice." Given the weight gain observed in some of the samples in this study, and the modest percentages of ice (Table 2), the results of the two studies are not inconsistent. Perhaps, however, the differences do sug- gest the importance of studying the weight change of shrimp under both laboratory and actual commercial fishing conditions. An average weight loss of 2.2% from capture to un- loading is sufficiently small as to be insignificant from a management standpoint. The minimum aggregate size limit of 353 shrimp per kg (160 shrimp per lb), which is intended to provide some protection from har- vest for age-one shrimp, is based on average size-at- age data. Interannual and geographic variability in shrimp growth easily exceeds 2.2c/r (Hannah and Jones, 1991). The findings of this study are significant how- ever, principally in that they demonstrate that the change in weight of pink shrimp that should be ex- pected, when using normal icing procedures, is fairly small. Accordingly, fishermen can readily comply with the minimum size regulation. Interestingly, there were daily average weight change values which exhibited a significant positive change (Table 1), probably owing to absorption of meltwater from the ice. Samples from two trips showed large deviations from the average weight change which were readily explainable. On trip 3, the vessel ran out of ice on the final day of fishing, and, on trip 6, some spoil- age was found throughout the load suggesting inad- equate icing. In both cases the samples showed high weight loss values, 5.9% and 4.6% respectively. These data suggest that there are small but fairly consistent losses in average weight of pink shrimp between commercial capture and unloading, at least when the shrimp are adequately iced. The results of this study appear to conflict with the findings of Acknowledgment This project was financed in part with Federal Interjurisdictional Fisheries Act funds through the U. S. National Marine Fisheries Service. The Oregon shrimp industry was crucial to the success of this project, particularly the owners and operators of the following fishing vessels: Betty A, Florence May. Maranatha, Olympic, Pacific Hooker and Pacific Hus- tler. Steve Jones, Mark Saelens, and Jean McCrae all participated in the at-sea sampling. Rick Starr con- ceived of this project and designed the sampling strat- egy. Terry Link assisted with sample workup and ves- sel arrangements. NOTE Hannah and Richmond. Weight change in Pandalus jordani after harvesting 807 Literature cited Bullard, F. A., and J. Collins. 1978. Physical and chemical changes of pink shrimp Pandalus borealis, held in carbon dioxide modified refrigerated sea water compared with pink shrimp held in ice. Fish. Bull. 76:73-78. Hannah, R. W., and S. A. Jones. 1991. Fishery induced changes in the population struc- ture of pink shrimp {Pandalus jordani). Fish. Bull. 89:41-51. Kirkley, J. E., and W. D. DuPaul. 1989. Commercial practices and fishery regulations: the U.S. northwest Atlantic sea scallop, Palcopten magellanicus (Gmelin, 1791), fishery. J. Shellfish Res. 8(11:139-149. PFMC (Pacific Fishery Management Council). 1981. Discussion draft fishery management plan for the pink shrimp fishery off Washington, Oregon, and California. PFMC, 169 p. Occurrence and seasonal variations of spiny lobsters, Panulirus argus (Latreille), on the shelf outside Bahia de la Ascension, Mexico Enrique Lozano-Alvarez Patricia Briones-Fourzan Fernando Negrete-Soto Universidad Nacional Autonoma de Mexico Instituto de Ciencias del Mar y Limnologia, Estacion "Puerto Morelos" Ap Postal I 152. Cancun, Q.R 77500 Mexico The spiny lobster, Panulirus argus, is one of the most valuable fishery resources on the Mexican Caribbean coast (Secretaria de Pesca, 1987). Lobsters are caught with traps and tangle nets, as well as by SCUBA diving, only on the shelf around Isla Mujeres and Isla Contoy (Fig. 1A), at significantly greater depths than elsewhere on the coast (Seijo et al. 1991). From Puerto Morelos to Xcalak, where the continental shelf is extremely narrow, lobsters are captured mainly by skin diving in the shallow coral reef and lagoon areas (Lozano-Alvarez et al. 1991a). In Bahia de la Ascension (Fig. IB), the fishery for P. argus is based on the use of artificial shelters, called "casitas." Casitas are de- ployed on the bottom and the lob- sters sheltering beneath them are caught by skin-diving fishermen. Both casitas and the fishing method are thoroughly described elsewhere (Lozano-Alvarez et al. 1989, 1991a; Briones et al. in press). Casitas are installed only within the bay, but some fishermen also skin dive for lobsters in the shallow coral reefs adjacent to the bay. No lobster fishing is conducted on the coastal shelf outside the bay deeper than 15 m. Only the tails are used. Tails are graded according to weight, packed in 10-pound (4.65 kg) boxes, and frozen. Fishing regulations in- clude a minimum size limit of 13.5 cm tail length (=74 mm carapace length, CD, a prohibition on the catching of egg-bearing females, and a closed season from 1 March to 30 June. In 1985-87, results from a tag- ging program showed that the spiny lobster population within the bay comprised mostly juveniles and su- badults (Lozano-Alvarez et al. 1991a). Growth of lobsters in the bay was fast, and they moved from the bay area towards the coral reefs as they grew. No evidence of repro- ductive activity in bay lobsters was found (Lozano-Alvarez et al. 1991a). Size of onset of sexual maturity is =80 mm CL, but females do not be- come fully mature until approxi- mately 90mm CL (Fonseca, 1990). This raised the hypothesis of the existence of the reproductive seg- ment of the population on the shelf outside the coral reef, at greater, currently unfished depths. Lozano- Alvarez et al. (1991a) stressed the need to test this hypothesis for fu- ture management plans. In this paper, we present evidence of the occurrence of adult reproduc- tive P. argus on the continental shelf outside Bahia de la Ascension, and discuss seasonal variations in abun- dance and some biological aspects of this segment of the population, as well as its potential importance as a refugium in space. The fisheries aspects of the study were published elsewhere (Lozano-Alvarez and Negrete-Soto, 1991). Methods Bahia de la Ascension (Fig. IB) is a large (=740 km2) and shallow (<6m deep) bay, bordered by man- groves and grass swamps. A sub- stantial part of the bottom of the bay is covered with seagrass (mainly Thalassia testudinum) and dense aggregations of red and green algae (Briones et al., in press). A coral reef tract exists along its mouth, protecting the inner waters of the bay from wave surge. The bay is a nursery area for P. argus (Lozano-Alvarez et al. 1991a). Fishing operations for spiny lob- sters were carried out on the conti- nental shelf outside the coral reef tract (Fig. IB) during two different seasons, summer ( 11 July-10 Sep- tember 1989, six fishings), and win- ter ( 14 November 1989-27 Febru- ary 1990, five fishings). Standard lobster traps of the kind used by fishermen at Isla Mujeres were set. The traps were rectangular, mea- suring 121 X 91 X 40cm. The frame, built with 12-mm diameter rod, was covered with plastic coated galvanized wire mesh (5 X 2.5cm), and an entrance was left on one of the small sides. Traps were baited with cow hide. A maximum of 39 traps was in- stalled in any fishing operation, in lines parallel to the coral reef. The continental shelf in front of the bay is narrow, not exceeding 4 km from the coast. Complex reefs exist in this area at depths of <10m and from 30 to 40 m (Jordan-Dahlgren et al., in press). After 40 to 60m the slope becomes very steep, rap- idly reaching depths in excess of 400 m. Echosounding was used to Manuscript accepted 17 May 1993. Fishery Bulletin, U.S. 91:808-815 ( 1993) NOTE Lozano-Alvarez et al.: Occurrence and seasonal variations of Panulirus argus 809 v. I '87.' 89W ~S USA GULF OF MEXICO \l»t« Mulam Punta Allen ishladela Ascension hla da I Eaplrltu Santo Punta Harraro 191 '•' .Y*VBanco Chlochorro V *"'" 8 6- lA: I Punta Pa|aroa Punta Tupac _19'20 J Figure 1 (A) The coast of Quintana Roo, showing the main fishing localities for Panulirus argus. (B) Bahia de la Ascension. Black areas denote coral reef tracts; shaded areas show the sector of the shelf where exploratory fishings with traps were carried out. Casitas for lobster fishing were used coastward from the coral reef. locate suitable bottoms for installing traps to minimize trap loss. In certain areas where the deep reefs are well developed (e.g., offshore from Punta Allen, Fig. IB ), no suitable bottoms were found. Thus, fishing operations were constrained to a range of depths between 15 and 30 m, from Nicchehabim Reef to Punta Pajaros (Fig. IB). Because soak time does not affect catch rate of the traps (Lozano-Alvarez et al., 1991b, Lozano-Alvarez and Negrete-Soto, 1991), we considered the number of lobsters per trap-lift as a gross index of abundance. Carapace length (CL) of lobsters was measured (±0.1 mm) from between the rostral horns to the poste- rior dorsal edge of the cephalotorax. Carapace state of lobsters was recorded following an arbitrary scale from 1 (newly molted and completely clean) to 4 (heavily fouled), as a relative indicator of the time elapsed since the last molt (Kanciruk and Herrnkind, 1976). Females were considered non-reproductive (NR) when eggs were absent, the carapace was clean, and no traces of sper- matophores were found on the sternum. Reproductive females comprised those that presented a new sper- matophore attached to the sternum (SP); external eggs attached to the pleopods (EE); or empty egg capsules remaining on the pleopods (EC). To compare the size distribution and general fea- tures of the lobsters found on the shelf outside the bay with those from the bay, two different groups of data were obtained from the bay area. Throughout the fishing season, July 1989-February 1990, monthly data from a number of boxes of lobster tails, categorized by tail weight (TW, g) class, were obtained from a local processing plant. This plant receives all the catch from the bay area. The data were transformed to CL by applying the following equation: Log TW (g)= 2.550 log CL (mm) -2.69298 (Lozano-Alvarez et al, 1989). However, because the size structure of the catch is constrained by the minimum size limit and includes lobsters caught on the fore-reef down to a depth of 15 m, and the catch data gives no information on sex or reproductive activity, we also used data from Lozano- Alvarez et al. (1991a) taken during a tagging opera- tion carried out in May-June 1986, which comprised lobsters found solely beneath casitas throughout the bay, as a less biased estimate of size structure, sex ratio, and carapace state of lobsters in the bay. Student's t-test (Zar, 1984) was used to compare log- transformed data on seasonal abundance of lobsters (number of lobsters per trap-lift), mean sizes between males and females, and mean sizes between shelf and bay lobsters. Sex-ratios were compared with a x2 test. Mean sizes of summer and winter lobster catches taken on the shelf outside the bay were compared with a 810 Fishery Bulletin 91(4). 1993 Table 1 Number of traps lifted. number of lobsters (Panulirus argi s) caught, and number of lobsters obtained per trap-lift during exploratory fishings with traps (summer 1989 and winter 1989-1990), on the continental shelf outside Bahia de la Ascension Mexico No. of traps No. of Lobsters per Date of trap-lifting lifted lobsters trap-lift Summer Jul 19. 1989 36 0 0 Jul 24 21 16 0.76 Aug 10 18 23 1.28 Aug 16 13 5 0.38 Aug 22 34 14 0.41 Sep 05 34 9 0.26 Total summer 156 67 0.43 Winter Nov 23, 1989 39 106 2.72 Nov 26 30 29 0.97 Dec 05 39 204 5.23 Dec 08 13 21 1.62 Feb 27, 1990 39 35 0.90 Total winter 160 395 2.47 Table 2 Catch composition of Panulirus argus caught on the shelf outside Bahia de la Ascension during the exploratory fishings with traps (summer 1989 and winter 1989-1990), and of those obtained from beneath casitas in the bay during May^June 1986 (M=males, F=females, CL=carapace length in mm). Offshore Summer Winter N Male: Female Mean CL, F. CL range, F. Mean CL, M. CL range. M. 67 1.31:1 89.3 55.0-156.9 111.0 76.5-162.1 395 1.04:1 78.9 52.4-117.2 88.7 56.0-134.0 Bay" 1402 1.04:1 64.4 22.0-100.3 65.9 29.8-113.1 Data from Lozano-Alvarez et al. ( 1991a). Table 3 Percentage of Panulirus argus in each of the four carapace states, caught on the shelf outside Bahia de la Ascension during the exploratory fishings with traps (summer 1989 and winter 1989-1990), and of those obtained from beneath casitas in May-June 1986. Carapace states 1-4 represent a range from neA'ly molted and completely clean to heavily fouled. Carapace state Offshore Summer (Ar=67) Winter (N=395) Bay (^=1402) non-parametric Mann-Whitney £/-test with normal ap- proximation (Zar, 1984). The level of significance con- sidered throughout this paper is P = 0.05. Results Size structure and catch composition of lobsters on the shelf outside the bay A total of 462 lobsters, 239 (52%) males and 223 (48%) females, were caught during fishing operations out- side the bay. The observed sex ratio was not signifi- cantly different from the expected value 115 0 1 0 0 1 0 0 1 Total 18 3 2 6 166 5 1 22 during the winter than during the summer (?=4.21, df=9, P< 0.01) (Table 1). No significant differences in sex ratio were found between the two seasons (X2= 1.332, df=l). The mean CL of the whole catch (Z=8.25, df=460, P<0.001), and of both males (Z=5.99, df=237, P< 0.001) and females (Z=4.15, df=222, P<0.001), was significantly greater in summer than in winter (Table 2). Newly molted lobsters (carapace state 1) were more abundant in winter (74.4%) than in summer (35.8%), whereas lobsters with old, heavily fouled exoskeletons were more evident in summer (8.95%) than in winter (0.5%) (Table 3). During the summer, 11 (38%) of the 29 females were reproductive, while only 28 (14%) of the 194 females caught in winter were in this condi- tion (Table 4). In both seasons, reproductive females measured over 80 mm CL, and most^were in stage EC (Table 4). Only 2 females in summer and 1 in winter were actually carrying eggs (EE). Size structure and catch composition of lobsters from casitas within the bay The 1,402 lobsters caught beneath the casitas during May-June 1986 comprised 713 (51%) males and 688 (49%) females. No significant difference from unity was found in sex ratio. Mean CL of males (65.9 mm) was significantly larger (r=2.50, df= 1000, P<0.01) than that of females (64.4mm) (Table 2, data from Lozano- Alvarez et al., 1991a). Most of the lobsters within the bay were in carapace states 1 and 2, i.e., recently molted (Table 3). Only five of the 688 females from within the bay were reproductive (stage EC ), but these were among the largest in size (87.0-100.3 mm CL), and were found in sites close to the reef (Lozano- Alvarez et al. 1991a). Comparison of size distribution between the shelf and bay segments of the population Figure 2 shows the size distribution of a) the sample taken from casitas within the bay (modified from Lozano-Alvarez et al. 1991. a and b) the total sample taken on the shelf outside the bay. Both males and females were significantly larger on the shelf than in the bay 90.0mm CL(Fig. 2B). Size structure of lobsters in the commercial catch from the bay and adjacent shallow reefs Table 5 shows the monthly size distribution of the catch taken by the fishermen. Peak catches occurred in the 69.3-76.3 mm CL size class (which includes the minimum legal size) throughout the fishing season, except in July, where the peak was located in the 76.3- 82.4 mm CL class. Lobsters >102.2mm CL probably represented the catch from the adjacent shallow reefs, and represented a small proportion of the catch dur- ing most of the fishing season. 812 Fishery Bulletin 91(4), 1993 15 n=H02 T'T' I ' I ' I ' I ' I r>=4B2 ''i m i i i i'T ^Tf'i iT1' I I I I I I 20 30 40 SO 60 70 80 80 100 110 120 110 140 150 ISO Carapace length (mm 1 IH Females I I Males Figure 2 Size-frequency distribution of male and female Panulirus argus caught (A) beneath casitas within Bahia de la Ascension (modified from Lozano et al. 1991a l, and (B) in traps on the continental shelf outside the bay. Discussion The occurrence of a reproductive segment of the population of Panulirus argus on the shelf outside Bahia de la Ascension has been documented in this study. The large sizes of lobsters and the evidence of mating and spawning activity on the continental shelf indicate that reproduction takes place in this deeper, unfished habitat, as opposed to the smaller sized, non-reproductive segment of the population within the bay. Munro (1974) suggested that in coralline areas it is unlikely that the amount of shel- ter would be a limiting factor for spiny lob- sters, but that this might be important in shelf areas with sparse coral cover. The deep and complex reef formations on the shelf outside Bahia de la Ascension (Jordan- Dahlgren et al., in press) likely offer abun- dant natural shelters for P. argus. Lobsters are attracted to traps for refuge rather than for food (Heatwole et al., 1988), and traps apparently fail to adequately sample resi- dent individuals (Herrnkind, 1980). Thus, the poor catches obtained during our sum- mer fishings probably reflected the occupa- tion of the available natural shelters by resi- dent adult lobsters. The resident nature of these lobsters seems to be further supported by the large proportion of reproductive fe- males and the high incidence of lobsters with fouled carapaces. The existence of resi- Table 5 Monthly size-frequency distribution of the commercial catch (July 1989-March 1990) of Panulirus argus in Bahia de la Ascension, obtained by converting tail weight (TW. g) to carapace ength (CL, mm)*. Commercia Size Number of lobsters category range (TW.g) (CL.mrni Jul Aug Sep Oct Nov Dec Jan Feb 60-100 56.7-69.3 1,484 848 636 1,274 636 424 0 371 100-128 69.3-76.3 17,720 6,680 5.240 4,640 3,200 2,680 2,401 1,600 128-156 76.3-82.4 18,464 5.408 3,680 3,456 2,144 1,760 2,240 1,184 156-185 82.4-88.1 12,717 3,672 2,592 2,241 1,431 1,134 540 621 185-213 88.1-93.2 6,785 2,461 1,656 1,472 897 690 690 391 213-241 93.2-97.8 3,480 1.460 960 880 560 420 402 240 241-270 97.8-102.2 1,692 900 612 468 306 342 180 162 270-298 102.2-106.3 944 432 352 288 160 240 0 128 298-340 106.3-111.9 810 480 285 255 150 180 0 105 340-397 111.9-118.9 572 338 195 195 117 156 130 91 >397 >118.9 486 348 191 110 81 171 190 91 Total 65,154 23.027 16,399 15,279 9,682 8,197 6,773 4,984 * Log TW = 2.55 Log CL -2.69298 (Lozano-A varez et ; 1., 19891. NOTE Lozano-Alvarez et al.: Occurrence and seasonal variations of Panulirus argus dent populations of P. argus has been suggested in the Bahamas (Kanciruk and Herrnkind, 1978; Herrnkind and Lipcius, 1989) and the Dry Tortugas (Davis, 1974). In contrast, the greater abundance of lobsters on the shelf during the winter, in addition to their smaller mean size and recently molted carapaces, indicates a recent recruitment of subadult lobsters to the deep habitat. The recruits likely originated in the bay, from where they exited in pulses, as sug- gested by the large number of lobsters caught in two fishings (Table 1). These two fishings were conducted immediately after the passage of two severe cold fronts. Offshore movements of P. argus associated with cold fronts have been documented in other parts of the Caribbean and Florida (Herrnkind, 1980). Moreover, catches increase notably in winter in the deep fishery off Isla Mujeres, whereas catches from more shallow areas, such as Bahia de la Ascension and Banco Chinchorro (Fig. 1A) decline at that time of year (Lozano-Alvarez, 1992). However, lobsters moving along the coast (Davis, 1979; Gregory and Labisky, 1986; Fonteles-Filho and Correa-Ivo, 1980; Herrnkind, 1980) could also account for the greater abundance found on the shelf. Some lobsters tagged in Bahia de la Ascension were recovered at locations several kilometers to the north and south of the bay (Lozano-Alvarez et al. 1991a). The extent of these movements along the offshore depth contours re- mains unknown, and warrants further investigation. In contrast to the sex-ratio in our shelf samples, females outnumbered males in offshore P. argus popu- lations in Florida (Davis, 1974; Lyons et al, 1981; Hunt et al., 1991M. Although in the Mexican Caribbean the reproductive season spans from March to November, with significant peaks in August-September (Fuentes- Castellanos, 1988), fully mature females, particularly those bearing eggs, were very scarce in our shelf samples. However, considerable evidence exists sug- gesting that traps fail to sample gravid females in several species of spiny lobsters (Herrnkind, 1980). We believe that gravid females were underestimated in our shelf samples owing to the availability of natural shelter. The distribution of lobsters in and outside Bahia de la Ascension seems to conform to the known size- related distribution of P. argus, that is to say juve- niles inhabite shallow habitats (lagoons, bays) and 1 Hunt. J. H., T. R. Matthews, D. Forcucci, B. S. Hedin, and R. D. Bertelsen. 1991. Management implications of trends in the popula- tion dynamics of the Caribbean spiny lobster. Panulirus argus, at Looe Key National Marine Sanctuary. Final Rep., NOAA Office of Ocean and Coastal Resource Management, Sanctuary Programs Div., Contract 50-DGNC-6-00093, 81 p. adults dwelling on deeper reef habitats. However, some large adult lobsters were found in the bay and shallow adjacent reefs (Fig. 2A, Table 5). This fact suggests that some adult lobsters return to inshore habitats after their ontogenetic migration to deeper habitats. Inshore movements of large P. argus have been documented in Antigua and Barbuda (Peacock, 1974), Brazil (Fonteles-Filho and Correa-Ivo, 1980) and Florida (Gregory and Labisky, 1986; Hunt et al., 1991). Hunt et al. (1991) suggested that these in- shore movements are controlled by changes in behav- ior associated with reproduction. The P. argus population in Bahia de la Ascension and adjacent offshore areas is highly dynamic. Fishing pressure on juveniles in the bay and near shallow reefs is heavy (Lozano-Alvarez, 1992), but the closed season is strictly observed, and no lobster fishing is conducted offshore. Developing an alternative trap fishery off Bahia de la Ascension would be impractical and costly, because of the narrowness of the shelf, the complex morphology of the bottom, the strong currents encoun- tered in the zone, and the small and variable CPUE of legal-sized lobsters obtained (Lozano-Alvarez and Negrete-Soto, 1991). Consequently, the fishery will con- tinue to be focused on the small-sized lobsters in the shallow areas. Despite heavy fishing in the bay, we found evidence of a winter offshore migration. Lozano-Alvarez (1992) estimated a high emigration rate from the bay through- out the year, so other offshore recruitment pulses may occur at other times of the year (e.g., the closed sea- son). Because of their large sizes, the resident females on the shelf outside the bay have high indices of egg productivity (Fonseca, 1990), and spawn more than once in the reproductive season (Lipcius, 1986). Thus, the unfished deep habitat outside the bay may bear importance as a refugium in space (Campbell, 1989; Caddy, 1990) for reproductive lobsters. Other deep, currently unfished lobster areas were found offshore from Puerto Morelos (Lozano-Alvarez et al., 1991b) and probably exist elsewhere along the coast of Quintana Roo, particularly on the narrow shelf from Puerto Morelos to Xcalak (see Fig. 1A), and in many other areas of the Caribbean as well (e.g. Cuba, Gonzalez et al., 1990). Their existence might serve to mitigate, through the production of larvae, the pres- sure exerted on more heavily fished stocks. We pro- pose that the deep habitats on the shelf outside Bahia de la Ascension should be left undisturbed as a nucleus of protected spawning stock, regardless of the unknown final destination of the larvae. We similarly encourage the protection of other deep, currently unfished areas throughout the Caribbean, as a means to preserve groups of reproductive lobsters that contribute larvae to the regional pool. 814 Fishery Bulletin 91(4). 1993 Acknowledgments We acknowledge the crew of the FV Fipesco-207 — Cap- tain Daniel Duran, Pedro Gaona, and Michel Moreno, as well as Maria Eugenia Ramos, for their assistance in the shelf fishing operations; the directives and mem- bers of the Sociedad Cooperativa de Produccion Pesquera "Pescadores de Vigia Chico" at Punta Allen for provid- ing logistic support, and the directors and staff of "Ocean Garden, Inc." at Canciin, for allowing us ac- cess to their production files. The shelf fishing opera- tions were partially funded by World Wildlife-Fund U.S. through Asociacion de Amigos de Sian ka'an, A.C. Literature cited Briones, P., E. Lozano, and D. B. Eggleston. The use of artificial shelters ("casitas") in research and harvesting of Caribbean spiny lobsters in Mexico. In B. F. Phillips, J. S. Cobb and J. Kittaka (eds.), Spiny lobster management: current issues and perspectives. Fishing News Books, Oxford. (In press). Caddy, J. F. 1990. Population dynamics, stock assessment and man- agement — opportunities for future research: a per- sonal overview. The Lobster Newsl. 3 (2):9— 11. Campbell, A. 1989. The lobster fishery of southwestern Nova Scotia and the Bay of Fundy. In J. F. Caddy, (ed.), Marine invertebrate fisheries: their assessment and manage- ment, p. 141-158. Wiley Interscience, NY. Davis, G. E. 1974. Notes on the status of spiny lobsters, Panulirus argus, at Dry Tortugas, Florida. In W. Seaman, and D. Y. Aska (eds.), Conference proceedings: research and information needs of the Florida spiny lobster fishery, p. 22-32. State Univ. Syst. Florida Sea Grant Rep. SUSF-SG-74-201, Gainsville, FL 1979. Management recommendations for juvenile spiny lobsters, Panulirus argus, in Biscayne National Monu- ment, Florida. South Florida Res. Cent. Rep. M-530, Homestead, FL 33030, 32 p. Fonseca, M. 1990. Fecundidad de la langosta Panulirus argus (Latreille, 1804) en el norte de Quintana Roo, Mexico. Tesis profesional, Univ. Simon Bolivar (Mexico), 49 p. (In Spanish.) Fonteles-Filho, A. A. and C. T. Correa-Ivo. 1980. Migratory behaviour of the spiny lobster Panulirus argus (Latreille) off Ceara State, Brazil. Arq. Cienc. Mar. 20 (l-2):25-32. Fuentes-Castellanos, D. 1988. Investigaciones pesqueras de la langosta en el Caribe mexicano. In Los recursos Pesqueros del Pais. Secretaria de Pesca Mexico, p. 441-462. Mexico, D.F (In Spanish.) Gonzalez, G., A. Herrera, E. Diaz, R. Brito, G. Gotera and C. Arrinda. 1990. Bioecologia y conducta de la langosta (Panulirus argus, Lat.) en las zonas profundas del borde de la plataforma de la region suroccidental de Cuba (Ab- stract, in Spanish). Int. Workshop on Lobster Ecol- ogy and Fisheries, Havana, Cuba, 12-16 June, 1990. Gregory, D. R. Jr., and R. F. Labisky. 1986. Movements of the spiny lobster Panulirus argus in South Florida. Can. J. Fish. Aquat. Sci. 43:2228- 2234. Heatwole, D. W., J. H. Hunt and F. S. Kennedy Jr. 1988. Catch efficiencies of live lobster decoys and other attractants in the Florida spiny lobster fishery. Fla. Mar. Res. Publ. 44, 15 p. Herrnkind, W. F. 1980. Spiny lobsters: patterns of movement. In J. S. Cobb, and B. F. Phillips (eds.). The biology and man- agement of lobsters, vol. I, p. 349-407. Academic Press, NY. Herrnkind, W. F. and R. N. Lipcius. 1989. Habitat use and population biology of Bahamian spiny lobster. Proc. Gulf Caribb. Fish. Inst. 39:265- 278. Jordan-Dahlgren, E., E. Martin-Chavez, M. Sanchez- Segura and A. Gonzalez de la Parra. The Sian ka'an Biosphere Reserve coral reef system. Atoll Res. Bull. (In press.) Kanciruk, P., and W. F. Herrnkind. 1976. Autumnal reproduction in Panulirus argus at Bimini, Bahamas. Bull. Mar. Sci. 26:417-432. 1978. Mass migration of spiny lobsters, Panulirus argus (Crustacea: Palinuridae): behavior and envi- ronmental correlates. Bull. Mar. Sci. 28:601-623. Lipcius, R. N. 1986. Size-dependent reproduction and molting in spiny lobsters and other long-lived decapods. In A. Wenner, led.). Crustacean issues, Vol. 3., p. 129-148. Balkema Press, Rotterdam. Lozano-Alvarez, E. 1992. Pesqueria, dinamica poblacional y manejo de la langosta Panulirus argus (Latreille, 1804) en la Bahia de la Ascension, Quintana Roo, Mexico. Tesis doc- toral, Fac. Ciencias, Univ. Nal. Auton. Mexico, 142 p. (In Spanish, Engl, abstr. ) Lozano-Alvarez, E. and F. Negrete-Soto. 1991. Pesca exploratoria de la langosta Panulirus argus con nasas frente a la Bahia de la Ascension en el Caribe Mexicano. Rev. Invest. Mar. (Cuba) 12(1-31:261-268. (In Spanish, Engl, abstr.) Lozano-Alvarez, E., P. Briones-Fourzan and B. F. Phillips. 1989. Spiny lobster fishery at Bahia de la Ascension, Q.R., In E. Chavez (ed.). Proceedings of the Work- shop Australia-Mexico on Marine Sciences, p. 379- 391. Centro de Investigaciones y Estudios Avanzados, Merida 97310, Mexico. NOTE Loza no-Alvarez et al. Occurrence and seasonal variations of Panulirus argus 815 Lozano-Alvarez, E., P. Briones-Fourzan, and B. F. Phillips. 1991a. Fishery characteristics, growth, and movements of the spiny lobster Panulirus argus in Bahia de la Ascension, Mexico. Fish. Bull. 89:79-89. Lozano-Alvarez, E., P. Briones-Fourzan, and J. Gonzalez-Cano. 1991b. Pesca exploratoria de langostas con nasas en la plataforma continental del area de Puerto Morelos, Q.R., Mexico. An. Inst. Cienc. del Mar y Limnol. Univ. Nal. Auton. Mexico 18 (l):49-58. (In Spanish, Engl, abstr. ) Lyons, W. G., D. G. Barber, S. M. Foster, F. S. Kennedy Jr., and G. R. Milano. 1981. The spiny lobster, Panulirus argus, in the middle and upper Florida Keys: population structure, sea- sonal dynamics and reproduction. Fla. Mar. Res. Publ. 38, 38 p. Munro, J. L. 1974. The biology, ecology, exploitation, and manage- ment of Caribbean reef fishes. Part V.I.: The biology, ecology and bionomics of Caribbean reef fishes: crus- taceans (spiny lobsters and crabs). Res. Rep. 3, Zool. Dep., Univ. West Indies, Kingston, Jamaica, 57 p. Peacock, N. A. 1974. A study of the spiny lobster fishery of Antigua and Barbuda. Proc. Gulf Caribb. Fish. Inst. 26:117- 130. Secretaria de Pesca. 1987. Pesquerias mexicanas: estrategias para su administracion. Dir. Gral. Admin. Pesq. Sria. Pesca, Mexico, 1061 p. (In Spanish.) Seijo, J. C, S. Salas, P. Arceo, and D. Fuentes. 1991. Analisis bioecondmico comparative de la pesqueria de langosta Panulirus argus en la plataforma continental de Yucatan. FAO Fish. Rep. No. 431, Suppl.:39-58. (in Spanish, Engl, abstr.) Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, NJ, 718 p. Growth and maturation of winter flounder, Pleuronectes americanus, in Massachusetts David B. Witherell Massachusetts Division of Marine Fisheries 18 Route 6A, Sandwich. MA 02563 Present address. North Pacific Fishery Management Council 605 West 4th Ave.. PO. Box 1 03 1 36. Anchorage. AK 995 1 0 Jay Burnett Northeast Fisheries Science Center National Marine Fisheries Service, NOAA Woods Hole. Massachusetts. 02540 Winter flounder, Pleuronectes americanus, is an important com- mercial and recreational species along the Atlantic coast. Previous studies have shown that the winter flounder stock is composed of sev- eral substocks that may consist of estuarine specific populations (Saila, 1961; Poole, 1966a; Pierce and Howe, 1977). Coastal stocks of winter flounder are managed by in- dividual states; area specific growth and maturity information is incor- porated into spawning stock bio- mass-per-recruit models (ASMFC, 1992). For management purposes, two stocks of winter flounder inhabit Massachusetts waters; one stock north of Cape Cod and the other stock south and east of Cape Cod (Lux et al., 1970; Howe and Coates, 1975; Pierce and Howe, 1977). For these stock units, growth parameters and sex ratios were reported from 1964 to 1968 coastal tag returns (Howe and Coates, 1975), but no studies us- ing aged scale or otolith samples had been undertaken. Growth rates estimated from tagging may not be equivalent to those based on age-length data (Francis, 1988). The two methods often provide dif- ferent results; faster growth rates are generally estimated from tag returns. 816 The primary objective of our study was to determine growth rates of juvenile and adult winter flounder based on aged scale samples. Our second objective was to determine average maturity schedules for the two winter flounder stock units, as maturation is an important life history para- meter for spawning stock biomass- per-recruit analyses. Materials and methods Winter flounder were sampled dur- ing the 1983-91 Massachusetts Di- vision of Marine Fisheries spring (May) bottom trawl surveys. State waters (0-3 mi) were surveyed with an otter trawl equipped with a fine mesh (13-mm stretch) codend liner that retained small fish. An aver- age of 95 stations per year were sur- veyed in a stratified random man- ner, at a sampling intensity of 1 station per 19 square nautical miles. Additional details of trawl survey methodology were provided by Howe (1989). Captured winter flounder were measured to total length (±0.5 cm), sampled for scales and otoliths, and assigned sex and maturity stage based on examina- tion of the gonads, by using estab- lished macroscopic criteria (Burnett et al., 1989). Because samples were collected at the end of the winter flounder spawning season, determi- nation of sex and maturity stage was relatively straightforward. Scale samples, taken from the cau- dal peduncle region, were used for age determination. Scales were im- pressed in acetate and aged as de- scribed for Georges Bank winter flounder (Fields, 1988). Von Bertalanffy growth curves were fitted to length-at-age data (1983-91 pooled) for males and fe- males of each stock unit, by using mean lengths observed for ages 2-8. Data from age-1 fish were not in- cluded, as preliminary analysis sug- gested that the smaller fish of this age group were either not available or not effectively captured by the sampling gear. Mean lengths at age for older (>8yrs) fish were not used in fitting growth models, because sample sizes were small. Statistical comparison of the results from this study and those of Howe and Coates ( 1975) was not attempted, as growth parameters derived from tagging and age-length information may not be equivalent (Francis, 1988). Stock-specific maturity schedules based on aged fish were calculated for both males and females. For each sex within each stock unit, a logistic function was fitted to the proportion of mature winter flounder, ages 2-5, by using non- linear regression. Lengths at 50% maturity (L60) were the inflection point on these curves. Similarly, ages at 50% maturity (A50) were de- termined from logistic curves fitted to observed maturity at age data. Results A total of 3,035 winter flounder (731 males and 986 females south and east of Cape Cod and 509 male and 809 female winter flounder north of Manuscript accepted 13 May 1993. Fishery Bulletin, U.S. 91:816-820 1 1993). NOTE Witherell and Burnett: Growth and maturation of Pleuronectes amencanus 817 North of Cape Cod South of Cape Cod 8 7 "to 6 ^ c . 0) 4 < 3 2 Males [ Females [ E L Males t Females E 1 r ■i ■1 ■■ ■i 1 wp ■ 300 200 100 0 100 200 300 400 300 200 100 0 100 200 300 400 Number examined Number examined Samj Figure 1 le sizes of male and female winter founder taken from each stock area. 1983-91. Cape Cod) were successfully aged and assigned matu- rity stages. Ages ranged from 1 to 16 years in the south- ern stock, and 1 to 15 years in the northern stock. For both stocks, females predominated at ages 3 and older (Fig. 1). Growth rates of winter flounder differed for each gen- der and stock unit. Females grew faster than males, and obtained larger mean asymptotic lengths (Fig. 2). For each gender, winter flounder from the south- em stock were generally larger at age than those from the northern stock, at ages 3-8. Mean lengths at age observed from our samples were similar to, but slightly smaller than, those determined from 1964-68 tag re- turn data (Table 1). For both stocks and sexes, maturation generally began at age 3 and was nearly complete at age 5. Maturation of age-3 and age-4 fish was quite variable and was some- what sensitive to size at age (Fig. 3). Owing to their somewhat faster growth rate, win- ter flounder matured at younger ages south of Cape Cod. For females, A50's (and corre- sponding L50's) were 3.0 years (28.3 cm) south of Cape Cod and 3.3 years (28.7cm) north of Cape Cod. For males, A50's (and corresponding L5l,'s) were 3.1 years (28cm) south of Cape Cod and 3.3 years (27.2 cm) north of Cape Cod. Discussion Adult winter flounder south and east of Cape Cod were larger at age than adults from other coastal popula- tions. For example, age-5 female winter flounder in North of Cape Cod South of Cape Cod 50- 50- 45- Males 45~ Males , j__=£3- 40- 40- 35- I -b^\~ 35- ijf^^^ E 30- ks^\ E30" ■H25- JL/| ' L»- 39.8 ~*° xy\ l°°- 45 9 £20- +T 5 20- O 7l ' K • 0 41 -3, Xr k ■ 0.31 £ 15~ /I ™15" ,/ I -I 10- T/ t0- 0.38 <° 10. y t0- o.ie 5- / 5" ( 112345678 01 2 3 4 5 6 7 8 Age (years) Age (years) 50- 50 -| 45- i- i X, 45' Females . ry- -e- Females X-~^? — 40 -jL; "^f 40- 35 J_^ 35- ■g-30- X-yS\ E 30" £25- 1 -PTT ° 91- 1 /| L-- 49.0 — Z0 rn/i !-"•■ 49 o ^20- -Qf £ 20- ?15- . >| K • 0.27 o) ,_ yK> K - 0.31 3 10- y >0- 0.07 «> 10. V t0- 0 25 5 / 5 i )12345678 01 2 3 4 5 6 7 8 Age (years) Age (years) Figure 2 Disti ibutions of length at age for male and female winter flounder sampled from both stock units . Von Bertalanffy growth curves are fitted through mean lengths, ages 2-8. Box plots of the medi an, quartiles, and range of lengths at age are shown. Fishery Bulletin 91 (4). 1993 Table 1 Calculated total length (cm) at age (yr) for female winter flounder in Massachusetts from 1983 -91 scale samples (present study) and from 1964-68 tag returns (Howe and Coates, 1975). North of Cape Cod South of Cape Cod Age 1964-68 1983-91 1964-68 1983-91 2 — 19.9 20.5 3 29.0 26.8 29.4 28.1 4 34.1 32.0 35.0 33.6 5 37.6 36.1 39.0 37.7 6 40.1 39.1 41.8 40.7 7 41.8 41.4 43.8 42.9 8 43.0 43.2 45.2 44.6 Massachusetts averaged 38 cm south and east of Cape Cod, versus 36 cm north of Cape Cod, 21 cm in Long Pond, Newfoundland (Kennedy and Steele, 1971), and 35cm in Narragansett Bay, Rhode Island (Berry et al., 1965). On Long Island, New York, Poole (1966a) found that growth rates of winter flounder increased from south to north; age-5 females ranged from an average of 31cm in Great South Bay to 36cm in Peconic Bay. Only the offshore population on Georges Bank is known to have faster growth rates than winter flounder in Massachusetts waters south and east of Cape Cod (Lux, 1973; Howe and Coates, 1975). The area south of Cape Cod is the center of the species geographical range (ASMFC, 1992) and may provide better environmental conditions for growth of winter flounder than other coastal areas. For both stocks of winter flounder in Massachu- setts, growth was apparently not affected by changes in exploitation or biomass. Biomass of these stocks declined over 50% from 1983 to 1990 owing to high fishing mortality (Howe et al., 1990). However, no trends in mean lengths at age were observed for the 1983-91 period, and growth rates were similar to those observed by Howe and Coates (1975) for the 1964-68 period (Table 1), when winter flounder were more abun- dant and exploitation rates were lower1. Under high levels of exploitation, faster growing fish are differen- ' Foster, K. L. 1987. Status of winter flounder iPseudopleuronectes americanus) stocks in the Gulf of Maine, southern New England, and middle Atlantic areas. Dep. Commer., NOAA. Natl. Mar. Fish. Serv.. Northeast Fish. Sci. Cent. Ref. Doc. 87-06. Age North of Cape Cod South of Cape Cod 88 848frmam -n r- M ales 75- 84 81 80 / /eesr ' 86 60- 85 / /B0 83 25- <#_ — /87 / 83 ' 88 26 30 Length (cm) Age Figure 3 Variability in maturity and mean length at age for winter flounder, ages 2-5, observed from 1983-91 bottom trawl surveys. Percent mature at length of age-3 and age-4 fish are plotted by year of collection (age-4 fish shown in italics I. Data for age-2 and age-5 fish are plotted by stars and squares, respectively; outliers are identified by year of collection in parenthesis. NOTE Withered and Burnett Growth and maturation of Pleuronectes amencanus 819 tially removed from populations (Ricker, 1975), and thus we were unable to separate the effects of crop- ping faster growing fish from potential density depen- dent effects. Winter flounder in Massachusetts waters mature at larger sizes than winter flounder from other coastal populations. For winter flounder in Massachusetts, only 50^ of the females were mature at 28 to 29 cm; L50's of females in Newfoundland were 25cm (Kennedy and Steele, 1971) and 27 cm in New Jersey (Danilla, 1978). Maturity ogives derived from our age-specific mean length at maturity data were similar to those reported from pooled length at maturity observations for Mas- sachusetts winter flounder during the period 1985-89 (O'Brien et al, 1993). They found lengths at 50% ma- turity of females to be 27.6 cm south of Cape Cod and 29.7 cm north of Cape Cod. Because populations of long lived, late maturing fish are more susceptible to overexploitation than are populations with early ma- turity and shorter life spans (Pitcher and Hart, 1982), each substock of winter flounder may require different management strategies (including size limits) for opti- mal harvest levels. Sex ratios of winter flounder favored females at older ages (Fig. 1), indicating a higher natural mor- tality rate for males. An increasing proportion of fe- males by size was observed for winter flounder in Rhode Island, but the differences in sex ratios were thought to be related to the faster growth of females (Berry et al., 1965). Female winter flounder on Georges Bank live longer than males (Lux, 1973). A number of mature male winter flounder in Concep- tion Bay, Newfoundland, failed to produce gametes, and there was a trend towards an increasing propor- tion of nonreproductive individuals with advancing age, leading to the possibility that males enter senes- cence at younger ages than do females, and thus have a higher natural mortality rate (Burton and Idler, 1984). A higher natural mortality rate for males has been found for other flounders, including European plaice, Pleuronectes platessa , (Bidder, 1925) and sum- mer flounder, Paralichthys dentatus, (Poole. 1966b), and it may be a general pattern that male flounders have shorter life spans than females. Acknowledgments This project was partially funded by the Commercial Fisheries Research and Development Act, Project 3- 375-R, the Interjurisdictional Fisheries Act, Project 3- IJ-3, and the Sportfish Restoration Act, Project F-56- R. We thank Arnold Howe, Tom Currier, and Steve Correia for their help with data collection and edi- torial assistance. Special thanks to personnel at the NMFS, Northeast Fisheries Science Center, for pre- paring scale samples for ageing. Literature cited ASMFC (Atlantic States Marine Fisheries Commission). 1992. Fishery management plan for inshore stocks of winter flounder Pleuronectes americanus. Fisheries Manage. Rep. No. 21, 138 p. Berry, R. J., S. B. Saila, and D. B. Horton. 1965. Growth studies of winter flounder (Pseudo- pleuronectes americanus) (Walbaum), in Rhode Island. Trans. Am. Fish. Soc. 94:259-264. Bidder, G. P. 1925. The mortality of plaice. Nature (London) 115:495-496. Burnett, J., L. O'Brien, R. K. Mayo, J. A. Darde, and M. Bohan. 1989. Finfish maturity sampling and classification schemes used during Northeast Fisheries Center bot- tom trawl surveys, 1963-89. NOAA Tech. Mem. NMFS/NEC-76, 14 p. Burton, M. P., and D. R. Idler. 1984. The reproductive cycle in winter flounder, Pseudopleuronectes americanus (Walbaum). Can. J. Zool. 62:2563-2567. Danilla, D. J. 1978. Age, growth, and other aspects of the life history of the winter flounder, {Pseudopleuronectes americanus) (Walbaum), in Southern New Jersey. M.S. thesis, Rutgers Univ., New Brunswick, NJ, 79 p. Fields, B. 1988. Winter flounder {Pseudopleuronectes americanus). In J. Penttila and L. M. Dery (eds.). Age determination methods for Northwest Atlantic species. NOAA Tech. Rep. NMFS 72, 135 p. Francis, R. I. C. C. 1988. Are growth parameters estimated from tagging and age-length data comparable? Can. J. Fish. Aquat. Sci. 45:936-942. Howe, A. B. 1989. State of Massachusetts inshore bottom trawl survey. In T. R. Azarovitz. J. McGurrin, and R. Seagraves (eds.), Proceedings of a workshop on bottom trawl surveys, p. 33-38. ASMFC Special Report 17. Howe, A. B., and P. G. Coates. 1975. Winter flounder movements, growth, and mor- tality off Massachusetts. Trans. Am. Fish. Soc. 104:13-29. Howe, A. B., T. P. Currier, S. J. Correia, and D. B. Witherell. 1990. Massachusetts fishery resource assessment. NOAA, NMFS Comm. Fish. Res. Dev. Act Annual Rep. 3-IJ-3. Kennedy, V. S., and D. H. Steele. 1971. The winter flounder, Pseudopleuronectes americanus in Long Pond, Conception Bay, Newfoundland. J. Fish. Res. Board Can. 28:1153-1165. 820 Fishery Bulletin 91(4), 1993 Lux, F. E. 1973. Age and growth of the winter flounder Pseudopleuronectes americanus, on Georges Bank. Fish. Bull. 71:505-512. Lux, F. E., A. E. Peterson, and R. J. Hutton. 1970. Geographical variation in fin ray number in win- ter flounder, Pseudopleuronectes americanus (Walbaum) off Massachusetts. Trans. Am. Fish. Soc. 99:483-488. O'Brien, L., J. Burnett, and R. K. Mayo. 1993. Maturation of nineteen species of finfish off the northeast coast of the United States, 1985- 1990. NOAA Tech. Rep. NMFS 113, 66 p. Pierce, D. E., and A. B. Howe. 1977. A further study on winter flounder group identi- fication off Massachusetts. Trans. Am. Fish. Soc. 106:131-39. Pitcher, T. J., and P. J. B. Hart. 1982. Fisheries ecology. Croom Helm Ltd., London, 414 p. Poole, J. C. 1966a. Growth and age of winter flounder in four bays of Long Island. N.Y. Fish Game J. 13:206-220. 1966b. A review of research concerning summer flounder and needs for further study. N.Y. Fish Game J. 13:226-231. Ricker, W. E. 1975. Computation and interpretation of biological sta- tistics offish populations. Fish. Res. Board Can. Bull. 191, 382 p. Saila, S. 1961. A study of winter flounder movements. Limnol. Oceanogr. 6:292-298. Fishery Bulletin Index Volume 91 (1-4), 1993 List of Titles 91(1) 1 Comparisons of spiny lobster Panulirus marginatus fecundity, egg size, and spawning frequency before and after exploitation, by Edward E. DeMartini, Denise M. Ellis, and Victor A. Honda 8 Spawning time, growth, and recruitment of larval spot Leiostomus xanthurus into a North Carolina estuary, by Cesar Flores-Coto and Stanley M. Warlen 23 Temperature effects on spontaneous behavior of larval and juvenile red drum Sciaenops ocellatus, and implica- tions for foraging, by Lee A. Fuiman and David R. Ottey 36 Abundance patterns of marine fish larvae during spring in a southeastern Alaskan Bay, by Lewis Haldorson, Marc Pritchett, David Sterritt, and John Watts 45 Age, growth, and reproduction of tautog Tautoga onitis (Labridae: Perciformes) from coastal waters of Virginia, by E. Brian Hostetter and Thomas A. Munroe 65 Early growth, behavior, and otolith development of the winter flounder Pleuronectes americanus, by Ambrose Jearld Jr., Sherry L. Sass, and Melinda F. Davis 76 Larval development of three roughy species complexes (Pisces: Trachichthyidae) from southern Australian waters, with comments on the occurrence of orange roughy Hoplostethus atlanticus, by Alan R. Jordan and Barry D. Bruce 87 Distribution and abundance of rockfish determined from a submersible and by bottom trawling, by Kenneth J. Krieger 97 Ontogenetic shift in the diet of young-of-year bluefish Pomatomus saltatrix during the oceanic phase of the early life history, by Rick E. Marks and David O. Conover 107 Trawl survey estimation using a comparative approach based on lognormal theory, by Robert A. McConnaughey and Loveday L. Conquest 119 A comparison of early-life-history traits in Atlantic menhaden Brevoortia tyrannus and gulf menhaden B. patronus, by Allyn B. Powell 129 Loss of shrimp by turtle excluder devices (TEDs) in coastal waters of the United States, North Carolina to Texas: March 1988-August 1990, by Maurice Renaud, Gregg Gitschlag, Edward Klima, Arvind Shah, Dennis Koi, and James Nance 138 Food habits of the sandbar shark Carcharhinus plumbeus off the U.S. northeast coast, with estimates of daily ration, by Charles E. Stillwell and Nancy E. Kohler 151 Spatial and temporal occurrence of Spanish mackerel Scomberomorus maeulatus in Chesapeake Bay, by Mark E. Chittenden Jr., Luiz R. Barbieri, and Cynthia M. Jones 159 165 171 Uncoupling of otolith and somatic growth in Pagrus auratus (Sparidae), by Malcolm P. Francis, Maryann W. Williams, Andrea C. Pryce, Susan Pollard, and Stephen G. Scott A new method of oocyte separation and preservation for fish reproduction studies, by Susan K. Lowerre-Barbieri and Luiz R. Barbieri Vertical and horizontal movements of adult chinook salmon Oncorhynchus tshawytscha in the Columbia River estuary, by Alan F. Olson and Thomas P. Quinn 179 Occurrence of Echeneibothrium (Platyhelminthes, Cestoda) in the calico scallop Argopecten gibbus from North Carolina, by Lynda S. Singhas, Terry L. West, and William G. Ambrose Jr. 91(2) 183 A comparison of tests for detecting trends in abundance indices of dolphins, by Alejandro A. Anganuzzi 195 A comparison of towed nets, purse seine, and light-aggregation devices for sampling larvae and pelagic juveniles of coral reef fishes, by J. H. Choat, Peter J. Doherty, B. A. Kerrigan, and J. M. Leis 210 Identification and distribution of Urophycis and Phycis (Pisces, Gadidae) larvae and pelagic juveniles in the U.S. Middle Atlantic Bight, by Bruce H. Comyns and George C. Grant 224 Maturation, reproductive seasonality, fecundity, and spawning frequency in Lutjanus vittus (Quoy and Gaimard) from the North West Shelf of Australia, by Tim L. O. Davis and Grant J. West 237 Inferring demographic processes from size-frequency distributions: Effect of pulsed recruitment on simple models, by Thomas A. Ebert, Stephen C. Schroeter, and John D. Dixon 244 Ovarian development, fecundity, and spawning frequency of black drum Pogonias cromis in Louisiana, by Gary R. Fitzhugh, Bruce A. Thompson, and Theron G. Snider III 254 Swimbladder inflation of the Atlantic menhaden Brevoortia tyrannus, by Richard B. Forward, Leslie M. McKelvey, William F. Hettler, and Donald E. Hoss 821 822 INDEX TITLES Fishery Bulletin 9 1 1 1 -A ). 1 993 260 Seasonal variation of reproductive investment of the tropical loliginid squid Loligo chinensis and the small tropical sepioid Idiosepius pygmaeus, by George D. Jackson 271 Estimating von Bertalanffy growth parameters of sablefish Anoplopoma fimbria and Pacific cod Gadus macrocephalus using tag-recapture data, by Daniel K. Kimura, Allen M. Shimada, and Sandra A. Lowe 281 Vertical distribution patterns and diel migrations of larval and juvenile haddock Melanogrammus aeglefinus and Atlantic cod Gadus morhua on Georges Bank, by R. Gregory Lough and David C. Potter 304 Habitat-specific density of adult yelloweye rockfish Sebastes ruberrimus in eastern Gulf of Alaska, by Victoria M. O'Connell and David W. Carlile 310 Detection of contaminant and climate effects on spawning success of three pelagic fish stocks off southern California: Northern anchovy Erigraulis mordax, Pacific sardine Sardinops sagax, and chub mackerel Scomber japonicus, by Michael H. Prager and Alec D. MacCall 328 Length-based analyses of yield and spawning biomass per recruit for black sea bass Centropristis striata, a protogynous hermaphrodite, by Gary R. Shepherd and JosefS. Idoine 338 Larval distribution patterns: Early signals for the collapse/recovery of Atlantic herring Clupea harengus in the Georges Banks area, by Wallace G. Smith and Wallace W. Morse 348 Relative abundance of pelagic resources utilized by the California purse-seine fishery: Results of an airborne monitoring program, 1962-90, by James L. Squire Jr. 362 Differences in haplotype frequencies of mtDNA of the Spanish sardine Sardinella aurita between specimens from the eastern Gulf of Mexico and southern Brazil, by Michael D. Tringali and Raymond R. Wilson Jr. 371 Estimated drift gillnet selectivity for albacore Thunnus alalunga, by Norman Bartoo and David Holts 379 Stage-I zoeae of laboratory-hatched Loplwlithodes mandtii (Decapoda, Anomura, Lithodidae), by Evan B. Haynes 382 Optimal sampling design for using the age-length key to estimate age composition of a fish population, by Han-Lin Lai 91(3) 397 414 Development of larval and early juvenile pygmy poacher, Odontopyxis trispinosa, and blacktip poacher, Xeneretmus latifrons (Seorpaeniformes: Agonidae), by Morgan S. Busby and David A. Ambrose Modeling the potential of fishery reserves for managing Pacific coral reef fishes, by Edward E. DeMartini 428 Allometry of energetics parameters in spotted Dolphin {Stenella attenuata) from the eastern tropical Pacific Ocean, by Elizabeth F. Edwards 440 Harbor Porpoise, Phocoena phocoena (L.), in the coastal waters of northern Japan, by David E. Gaskin, Satoru Yamamoto, and Akito Kawamura 455 Use of a monogenean gill parasite and feasibility of condition indices for identifying new recruits to a seamount population of armorhead Pseudopentaceros wheeleri (Pentacerotidae), by Robert L. Humphreys Jr., Mark A. Crossler, and Craig M. Rowland 464 Variability of zooplankton biomass and dominant species abundance on Georges Bank, 1977-1986, by Joseph Kane 475 Spatial structure and temporal continuity of the South Georgian Antarctic fish community, by James E. McKenna Jr. 491 Annual prey consumption by harbor seals (Phoca vitulina) in the Strait of Georgia, British Columbia, by Peter F. Olesiuk 516 Ontogenetic shift in habitat by early juvenile queen conch, Strombus gigas: patterns and potential mechanisms, by Veronique J. Sandt and Allan W. Stoner 526 Age, growth, maturity, and spawning of Spanish mackerel, Seomberomorus maculatus (Mitchill), from the Atlantic Coast of the southeastern United States, by David J. Schmidt, Mark R. Collins, and David M. Wyanski 534 A comparison of alternative methods for the estimation of age from length data for Atlantic coast bluefish (Pomatomus saltatrix), by Mark Terceiro and Jeffrey L. Ross 550 A statistical method for evaluating differences between age-length keys with application to Georges Bank haddock, Melanogrammus aeglefinus, by Daniel B. Hayes 558 Direct validation of black drum (Pogonias eromis) ages determined from scales, by Gary C. Matlock, Robert L. Colura, and Lawrence W. McEachron 389 Recruitment of bluefish Pomatomus saltatrix to estuaries of the U.S. South Atlantic Bight, by Richard S. McBride, Jeffrey L. Ross, and David O. Conover 564 Incorporation of between-haul variation using boot- strapping and nonparametric estimation of selection cur- ves, by Russell B. Millar INDEX: TITLES Fishery Bulletin 91 (1-4), 1993 823 573 Biological observations from the Cobb Seamount rockfish fishery, by Donald E. Pearson, David A. Douglas, and Bill Barss 577 Effects of body size on probability of predation for juvenile summer and winter flounder based on laboratory experiments, by David A. Witting and Kenneth W. Able 582 Movements of transplanted lingcod, Ophiodon elongatus, determined by ultrasonic telemetry, by Lynne K. Yamanaka and Laura J. Richards 91(4) 587 Developmental instability as an indicator of environmental stress in the Pacific hake (Merluccius productus), by Consuela L. Alados, Juan Escos, and John M. Emlen 594 Analysis of fishery resources: potential risk from sewage sludge dumping at the deepwater dumpsite off New Jersey, by Sukwoo Chang 611 Comparison of age at sexual maturity and other reproductive parameters for two stocks of spotted dolphin, Stenella attenuata, by Susan J. Chivers and Al C. Myrick Jr. 619 PCR-RFLP analysis on thirteen western Atlantic snappers (subfamily Lutjaninae): a simple method for species and stock identification, by Seinen Chow, M. Elizabeth Clarke, and Patrick J. Walsh 628 Effects of dolphin group type, percent coverage, and fleet size on estimates of annual dolphin mortality derived from 1987 U.S. tuna-vessel observer data, by Elizabeth F. Edwards and Christina Perrin 641 Method and precision in estimation of dolphin school size with vertical aerial photography, by James W. Gilpatrick Jr. 649 Determination of age and growth of South Pacific albacore (Thunnus alalunga) using three methodologies, by Marc Labelle, John Hampton, Kevin Bailey, Talbot Murray, David A. Fournier, and John R. Sibert 664 Reproductive biology of the false southern king crab (Paralomis granulosa, Lithodidae) in the Beagle Channel, Argentina, by Gustavo A. Lovrich and Julio H. Vinuesa 676 The assumption of constant selectivity and the stock assessment for widow rockfish, Sebastes entomelas, by David B. Sampson 690 Genetic analysis of the population structure of yellowfin tuna, Thunnus albacares, from the Pacific Ocean, by Daniel R. Scoles and John E. Graves 699 Planktonic and benthic feeding by the reef-associated vermilion snapper, Rhomboplites aurorubens (Teleostei, Lutjanidae), by George R. Sedberry and Nicole Cuellar 710 Latitudinal variation in the birth timing of captive California sea lions and other captive North Pacific pinnipeds, by Jonathan L. Temte 718 Variations on a simple dynamic pool model, by Grant G. Thompson 732 Relationships between otolith microstructure, microchemistry, and early life history events in Dover sole, Microstomus pacificus, by Christopher L. Toole, Douglas F. Markle, and Phillip M. Harris 754 Geographic variation and sexual dimorphism in the skull of the dusky dolphin, Lagenorhynchus obscurus (Gray, 1828), by Koen Van Waerebeek 775 Estimation of historical population size of the eastern spinner dolphin (Stenella longirostris orientalis), by Paul R. Wade 788 Spring distribution and abundance of ichthyoplankton in the tidal Delaware River, by Stephen B. Weisberg and William H. Burton 798 Effects of entanglement and escape from high-seas driftnets on rates of natural mortality of North Pacific albacore, Thunnus alalunga, by Richard W. Brill and David B. Holts 804 Weight change of pink shrimp, Pandalus jordani, after commercial harvest and handling, by Robert W. Hannah and Neil T. Richmond 808 Occurrence and seasonal variations of spiny lobsters, Panulirus argus (Latreille), on the shelf outside Bahia de la Ascension, Mexico, by Enrique Lozano-Alvarez, Patricia Briones-Fourzan, and Fernando Negrete-Soto 816 Growth and maturation of winter flounder, Pleuronectes americanus, in Massachusetts, by David B. Witherell and Jay Burnett Fishery Bulletin Index Volume 91 (1-4), 1993 List of Authors Able, Kenneth W. 577 Alados, Consuela L. 587 Ambrose, David A. 397 Ambrose, William G. Jr. 179 Anganuzzi, Alejandro A. 183 Bailey, Kevin 649 Barbieri, Luiz R. 151, 165 Barss, Bill 573 Bartoo, Norman 371 Brill, Richard W. 798 Briones-Fourzan, Patricia 808 Bruce, Barry D. 76 Burnett, Jay 816 Burton, William H. 788 Busby, Morgan S. 397 Carlile, David W, 304 Chang, Sukwoo 594 Chittenden, Mark E. Jr. 151 Chivers, Susan J. 611 Choat.J. H. 195 Chow, Seinen 619 Clarke, M. Elizabeth 619 Collins, Mark R. 526 Colura, Robert L. 558 Comyns, Bruce H. 210 Conover, David O. 97, 389 Conquest, Loveday L. 107 Crossler, Mark A. 455 Cuellar, Nicole 699 Davis, Tim L. O. 224 Davis, Melinda F. 65 DeMartini, Edward E. 1, 414 Dixon, John D. 237 Doherty, Peter J. 195 Douglas, David A. 573 Ebert, Thomas A. 237 Edwards, Elizabeth F. 428, 628 Ellis, Denise M. 1 Emlen.JohnM. 587 Escos, Juan 587 Fitzhugh, Gary R. 244 Flores-Coto, Cesar 8 Forward, Richard B. 254 Fournier, David A. 649 Francis, Malcolm P. 159 Fuiman, Lee A. 23 Gaskin, David E. 440 Gilpatrick, James W. Jr. 641 Gitschlag, Gregg 129 Grant, George C. 210 Graves, John E. 690 Haldorson, Lewis 36 Hampton, John 649 Hannah, Robert W. 804 Harris, Phillip M. 732 Hayes, Daniel B. 550 Haynes, Evan B. 379 Hettler, William F. 254 Holts, David 371 Holts, David B. 798 Honda, Victor A. 1 Hoss, Donald E. 254 Hostetter, E. Brian 45 Humphreys, Robert L. Jr. 455 Idoine, Josef S. 328 Jackson, George D. 260 Jearld, Ambrose Jr. 65 Jones, Cynthia M. 151 Jordan, Alan R. 76 Kane, Joseph 464 Kawamura, Akito 440 Kerrigan, B. A. 195 Kimura, Daniel K. 271 Klima, Edward 129 Kohler, Nancy E. 138 Koi, Dennis 129 Krieger, Kenneth J. 87 Labelle, Marc 649 Lai, Han-Lin 382 Leis, J. M. 195 Lough, R. Gregory 281 Lovrich, Gustavo A. 664 Lowe, Sandra A. 271 Lowerre-Barbieri, Susan K. 165 Lozano-Alvarez, Enrique 808 MacCall, Alec D. 310 Markle, Douglas F. 732 Marks, Rick E. 97 Matlock, Gary C. 558 McBride, Richard S. 389 McConnaughey, Robert A. 107 McEachron, Lawrence W. 558 McKelvey, Leslie M. 254 McKenna, James E. Jr. 475 Millar, Russell B. 564 Morse, Wallace W. 338 Munroe, Thomas A. 45 Murray, Talbot 649 Myrick, Al C. Jr. 611 Nance, James 129 Negrete-Soto, Fernando 808 O'Connell, Victoria M. 304 Olesiuk, Peter F. 491 Olson, Alan F. 171 Ottey, David R. 23 Pearson, Donald E. 573 Perrin, Christina 628 Pollard, Susan 159 Potter, David C. 281 Powell, Allyn B. 119 Prager, Michael H. 310 Pritchett, Marc 36 Pryce, Andrea C. 159 Quinn, Thomas P. 171 Renaud, Maurice 129 Richards, Laura J. 582 Richmond, Neil T. 804 Ross, Jeffrey L. 389,534 Rowland, Craig M. 455 Sampson, David B. 676 Sandt, Veronique J. 516 Sass, Sherry L. 65 Schmidt, David J. 526 Schroeter, Stephen C. 237 Scoles, Daniel R. 690 Scott, Stephen G. 159 Sedberry, George R. 699 Shah, Arvind 129 Shepherd, Gary R. 328 Shimada, Allen M. 271 Sibert, John R. 649 Singhas, Lynda S. 179 Smith, Wallace G. 338 Snider, Theron G. Ill 244 Squire, James L. Jr. 348 Sterritt, David 36 Stillwell, Charles E. 138 Stoner, Allan W. 516 Temte, Jonathan L. 710 Terceiro, Mark 534 Thompson, Bruce A. 244 Thompson, Grant G. 718 Toole, Christopher L. 732 Tringali, Michael D. 362 Van Waerebeek, Koen 754 Vinuesa, Julio H. 664 Wade, Paul R. 775 Walsh, Patrick J. 619 Warlen, Stanley M. 8 Watts, John 36 Weisberg, Stephen B. 788 West, Grant J. 224 West, Terry L. 179 Williams, Maryann W. 159 Wilson, Raymond K. Jr. 362 Witherell, David B. 816 Witting, David A. 577 Wyanski, David M. 526 Yamamoto, Satoru 440 Yamanaka, Lynne K. 582 024 Fishery Bulletin Index Volume 91 (1-4), 1993 List of Subjects Abundance — see also Population studies anchovy, northern 310, 348 bonito, Pacific 348 crab, Dungeness 107 dolphin eastern spinner 775 eastern tropical Pacific 183, 775, 641 fisheries stocks 594 flounder, summer 594 goosefish 594 hake red 594 silver 594 herring, Atlantic 338 ichthyoplankton, Delaware River 788 juvenile reef fish 195 killer whale, pygmy 641 larvae gadids 210 hakes, Atlantic 210 marine fish 36 reef fish 1 95 lobster, Caribbean spiny 808 mackerel chub 310,348 jack 348 pilot whale 641 rockfish 87 widow 573 yelloweye 304 sardine, Pacific 310,348 sea bass, black 594 sewage outfall and fishes 594 spatio-temporal differences in fish populations 594 tuna, bluefin 348 zooplankton, Georges Bank 464 Aerial survey anchovy, northern 348 bonito, Pacific 348 mackerel chub 348 jack 348 sardine, Pacific 348 tuna, bluefin 348 Africa, Southwest 754 Age at sexual maturity dolphin, spotted 611 Age composition of catches 382 Age determination age-length keys bluefish 534 haddock 550 albacore 649 bluefish 534 drum, black 558 Age determination (continued) flounder, winter 816 haddock 550 mackerel, Spanish 526 porpoise, harbor 440 otoliths mackerel, Spanish 526 Pagrus auratus 159 rockfish, widow 573 spot 8 tautog 45 Pagrus auratus 159 rockfish, widow 573 spot 8 statoliths, squid Idiosepius pygmaeus 260 Loligo ehinensis 260 tautog 45 vertebrae, albacore 649 Age-length key 382 bluefish 534 haddock 550 Age-size estimation albacore 649 bluefish 534 flounder, winter 816 haddock 550 mackerel, Spanish 526 Pagrus auratus 159 spot 8 Age validation drum, black 558 Agonidae — see Poacher Alaska, southeastern marine fish larvae 36 rockfish, yelloweye 304 Albacore 371,649,798 Anchovy, northern 310, 348 Antarctic fish communities 475 Argentina crab, false southern king 664 Argopecten gibbus — see Scallop, calico Armorhead, pelagic 455 Atlantic Bight, south bluefish 389 Atlantic Ocean bluefish 389,534 crab, false southern king 664 dumpsite, 106-mile deepwater 594 flounder, summer 594 goosefish 594 haddock 550 hake red 594 silver 594 mackerel, Spanish 526 sardine, Spanish 362 Atlantic Ocean (continued) Sardinella aurita 362 Sardinella brasiliensis 362 sea bass, black 594 snappers 619, 699 zooplankton, Georges Bank 464 Australia Great Barrier Reef 195 Lutjanus vittus 224 roughy 76 Behavior burrowing conch, queen 516 diel migration, cods 281 drum, red 23 flounder, winter 65 Bahama Islands conch, queen 516 Bertalanffy growth 237, 271, 534, 649, 718 albacore 649 bluefish 534 cod, Pacific 271 flounder, winter 816 sablefish 271 sea urchin purple 237 red 237 Beverton-Holt recruitment 718 Biological indices age at sexual maturity 611 parasites in pelagic armorhead 455 Biological rhythm menhaden, Atlantic 254 Birth timing seal harbor 710 northern fur 710 sea lion California 710 northern 710 Bluefish 97, 389, 534 Bonito, Pacific 348 Bothidae — see Flounder, summer Brazil sardine, Spanish 362 Sardinella aurita 362 Sardinella brasiliensis 362 Brevoortia patronus — see Menhaden, gulf Brevoortia tyrannus — see Menhaden, Atlantic California Bight, southern anchovy, northern 310 mackerel, chub 310 sardine, Pacific 310 purse-seine fishery anchovy, northern 348 bonito, Pacific 348 mackerel chub 348 jack 348 sardine, Pacific 348 825 826 INDEX: SUBJECTS Fishery Bulletin 91(1-4). 1993 California purse seine fishery (continued) tuna, bluefin 348 southern sewage outfall and fish abundance 594 Callorhinus ursinus — see Seal, northern fur Canada, British Columbia seal, harbor 491 Cancer magister — see Crab, Dungeness Cancridae — see Crab, Dungeness Captive breeding seal harbor 710 northern fur 710 sea lion California 710 northern 710 Carangidae — see Mackerel, jack Caribbean conch, queen 516 lobster, Caribbean spiny 808 Catch-at-age estimation 382 bluefish 534 haddock 550 rockfish, widow 676 Catch estimation — see also Population studies albacore 371 crab, Dungeness 107 scallop, Iceland 564 shrimp pink 804 ocean 804 Catch-per-unit-effort anchovy, northern 348 bonito, Pacific 348 flounder, summer 594 goosefish 594 hake red 594 silver 594 mackerel chub 348 jack 348 sardine, Pacific 348 sea bass, black 594 tuna, bluefin 348 Catch rates turtle excluder devices 129 Centropristis striata — see Sea bass, black Cestoda — see Echeneibothrium Cetaceans — see also Dolphin; Pilot whale; Killer whale, pygmy energetics 428 porpoise, harbor 440 school size 641 Chlamys islandica — see Scallop, Iceland Chesapeake Bay mackerel, Spanish 151 Chile dolphin, dusky 754 Classification — see Taxonomy Climate effects 310 anchovy, northern 310 Climate effects (continued) mackerel, chub 310 sardine, Pacific 310 Clupea harengus — see Herring, Atlantic Clupeidae — see Herring; Menhaden, Atlantic; Menhaden, gulf; Sardine, Pacific; Sardine, Spanish; Sardinella Cod Atlantic 281 hake red 594 Pacific 587 silver 594 Pacific 271 Columbia River estuary 171 Community structure fish, Antarctic 475 Conch, queen 516 Contaminant effects 310 anchovy, northern 310 mackerel, chub 310 sardine, Pacific 310 Coral reef fishes 414 Crab Dungeness 107 false southern king 664 lithodid 379,664 Lopholithodes mandtii 379 Crangonidae — see Shrimp, brown Crangon crangon — see Shrimp, brown Cynoscion regalis — see Weakfish Delaware River ichthyoplankton 788 Delphinidae — see Dolphin Delphinus delphis — see Dolphin, common Density dependence 611 reef fishes, larval and juvenile 195 rockfish 87 yelloweye 304 Depth distribution Atlantic cod 281 haddock 281 rockfish, yelloweye 304 salmon, chinook 171 zooplankton, Georges Bank 464 Development hake, Pacific 587 poacher blacktip 397 pygmy 397 Developmental instability El Nino effects 587 hake, Pacific 587 Diet bluefish, young of the year 97 Distribution cod, Atlantic 281 cods 210,281 conch, queen 516 haddock 281 hakes, Atlantic 210 herring, Atlantic 338 ichthyoplankton, Delaware River 788 Distribution (continued) mackerel, Spanish 151 poacher blacktip 397 pygmy 397 porpoise, harbor 440 rockfish 87, 573 harlequin 573 rosy 573 yelloweye 304 widow 573 salmon, chinook 171 zooplankton, Georges Bank 464 Dolphin bottlenose 64 1 common 183,628,641 dusky 754 eastern tropical Pacific spp. 183, 628, 641 gray grampus 64 1 killer whale, pygmy 641 little blackfish 641 melonheaded whale 641 pilot whale 641 roughtoothed 641 spinner 641 eastern 183 pantropical 183 whitebelly 183,628 spotted 183, 428, 611, 628, 641 striped 641 Dredges size selectivity for Iceland scallop 564 Driftnets 371,788 Drum black 244,558 red 23 Dumping ocean 594 sewage sludge 594 Dumpsite 106-mile deepwater 594 Early-life-history studies bluefish 97 cod, Atlantic 281 cods 210 drum, red 23 flounder summer 577 winter 65, 577 haddock 281 hakes, Atlantic 210 menhaden, Atlantic 107 menhaden, gulf 107 sole, Dover 732 spot 8 Echeneibothrium 179 Egg studies lobster, spiny 1 menhaden, Atlantic 107 menhaden, gulf 107 weakfish 165 Embryos — see Larval studies Energetics seal, harbor 491 INDEX: SUBJECTS Fishery Bulletin 91 ( 1-4), 1993 827 Engraulidae — see Anchovy Engraulis mordax — see Anchovy, northern Environmental effects 310 anchovy, northern 310 mackerel, chub 310 sardine, Pacific 310 Estuarine fishes bluefish, young of year 389 salmon, chinook 171 Eumetopias jubatus — see Sea lion, northern Fecundity crab, false southern king 664 drum, black 244 lobster, spiny 1 Lutjanus vittus 224 mackerel, Spanish 526 menhaden, Atlantic 107 menhaden, gulf 107 weakfish 165 Feeding — see Food habits Feresa attenuata — see Killer whale, pygmy 641 Fishery albacore 371 Cobb Seamount 573 coral reef fishes 414 gillnet, drift 371 lobster, spiny 1 mackerel, Spanish 151 scallop, calico 179 shrimp 129 ocean 804 pink 804 Fishery interactions tuna-dolphin 628, 775 Fishery management aggregate size limit 804 shrimp ocean 804 pink 804 coral reef fishes 414 models 718 shrimp ocean 804 pink 804 Fishery reserves coral reef fishes 414 lobster, spiny 1 Fishes, coral reef 1 Flatfish 65, 577, 594, 732, 816 Florida sardine, Spanish 362 Sardinella aurita 362 Sardinella brasiliensis 362 Flounder summer 577, 594 winter 65, 577, 816 Fluctuating asymmetry hake. Pacific 587 Food habits bluefish 97 drum, red 23 Food habits (continued) larvae, red drum 23 porpoise, harbor 440 seal, harbor 491 snapper, vermilion 699 Gadidae — see Cod; Haddock; Hake Gadus maeroeephalus — see Cod, Pacific Gadus morhua — see Cod, Atlantic Genetic studies sardine, Spanish 362 Sardinella aurita 362 Sardinella brasiliensis 362 snappers 619 species identification, snappers 619 stock identification, snappers 619 tuna, yellowfin 690 Geographic variation dolphin, dusky 754 sardine, Spanish 362 Sardinella aurita 362 Sardinella brasiliensis 362 Georges Bank cod, Atlantic 281 haddock 281 herring, Atlantic 328 Gillnet, drift albacore 371, 798 selectivity 371 Globieephala maerorhynehus — see Pilot whale Goosefish 594 Grampus griseus — see Dolphin, gray grampus Growth — see also Age-size estimation albacore 649 bluefish 534 cod, Pacific 271 conch, queen 516 flounder, winter 816 Pagrus auratus 159 porpoise, harbor 440 sablefish 271 sea urchin purple 237 red 237 sole, Dover 732 spot 8 squid, gonads 260 tautog 45 Growth increment analysis cod, Pacific 271 sablefish 271 Gulf of Alaska rockfish 87 yelloweye 304 Gulf of Mexico drum, black 244, 558 sardine, Spanish 362 Sardinella aurita 362 Sardinella brasiliensis 362 Gulf of Mexico fishery sardine, Spanish 362 Sardinella aurita 362 Sardinella brasiliensis 362 Habitat conch, queen 516 rockfish 87 yelloweye 304 Haddock 281,550 Hake Carolina 210 gulf 210 longfin 210 Pacific 587 red 210,594 silver 594 southern 210 spotted 210 white 210 Haddock 281 Hakes, Atlantic 210 Harvest refugia coral reef fishes 414 lobster, spiny 1 Hawaiian Islands fishery lobster, spiny 1 Hermaphroditism sea bass, black 328 Herring, Atlantic 338 Hexagrammidae — see Lingcod Ichthyoplankton 195, 281, 788 Identification dolphin, dusky 754 poacher blacktip 397 pygmy 397 snappers, stock and genetic 619 Impact assessments ocean dumping 594 Japan porpoise, harbor 440 Juvenile studies bluefish 97,389 cod, Atlantic 281 cods 210 drum, red 23 flounder summer 577 winter 577 haddock 281 hakes, Atlantic 210 poacher blacktip 397 pygmy 397 reef fish 195 sole, Dover 732 spot 8 Killer whale, pygmy 641 Labridae — see Tautog Lagenorhynchus obscurus — see Dolphin, dusky Larval studies bluefish 97 cod, Atlantic 281 cods 210 828 INDEX: SUBJECTS Fishery Bulletin 91(1-4). 1 993 Larval studies (continued) crab, Lopholithodes mandtii 379 drum, red 23 haddock 281 hakes, Atlantic 210 herring, Atlantic 338 menhaden, Atlantic 254 poacher blacktip 397 pygmy 397 reef fish 195 roughy 76 sole, Dover 732 spot 8 survey, Delaware River 788 swimbladder inflation 254 Latitudinal variation birth timing in pinnipeds 710 Leiostomus xanthurus — see Spot Length-based sampling 382 Length frequency analysis albacore 649 bluefish 534 haddock 550 lobster, Caribbean spiny 808 Length studies — see Age-size estimation Life history cod, Atlantic 281 haddock 281 sole, Dover 732 Light swimbladder inflation, effects on 254 Light traps 195 Lingcod 582 Lithodidae — see Crab, false southern king; Lopholithodes mandtii Lobster Caribbean spiny 808 spiny 1 Lophiidae — see Goosefish Lophius americanus — see Goosefish Louisiana drum, black 244 Lutjanidae — see Lutjanus vittus Lutjanus — see Snapper analis — see Snapper, mutton 619 apodus — see Schoolmaster 619 buceanella — see Snapper, blackfin 619 campechanus — see Snapper, red 619 cyanopterus — see Snapper, cubera 619 griseus — see Snapper, gray 619 jocu — see Snapper, dog 619 mahogoni — see Snapper, mahogany 619 synagris — see Snapper, lane 619 vivanus — see Snapper, silk 619 vittus 224 Mackerel chub 310,348 jack 348 Spanish 151, 526 Management — see Fishery management Mathematical methods abundance estimation 107, 183 age-length key comparison 550 ANOVA, repeated measures 641 Mathematical methods (continued) lognormal theory central tendency 107 nonpar ametric isotonic regression 564 size selectivity in dredges 564 simulations 183 trend analysis 107, 183 Maximum sustainable yield simple dynamic pool 718 Melonheaded whale 641 Melanogram mus aeglefinis — see Haddock Menhaden Atlantic 119,254 gulf 119 Merlucciidae — see Hake Merluccius bilinearis — see Hake, silver Merluccius productus — see Hake, Pacific Metamorphosis sole, Dover 732 Methods age-length key 382, 534, 550 density estimation, rockfishes 304 growth increment analysis 271 ichthyoplankton sampling 195 length-based 649 length-frequency analysis 534, 550 albacore 649 bluefish 534 haddock 550 oocyte separation and preservation, weakfish 165 photography of dolphin schools 641 stock assessment 676 Microchemistry sole, Dover 732 Microcotyle macropharynx 455 Microstomas pacificus — see Sole, Dover Microstructure 732 Mid-Atlantic Bight bluefish 97 cods 210 hakes 210 Migration — see Movements Mitochondrial analysis DNA sardine, Spanish 362 Sardinella aurita 362 Sardinella brasiliensis 362 tuna, yellowfin 690 RNA snappers, Atlantic 619 Models distributed delay sea bass, black 328 dynamic pool 718 population 310 anchovy, northern 310 mackerel, chub 310 sardine, Pacific 310 recruitment 310 anchovy, northern 310 fishes, marine 310 mackerel, chub 310 sardine, Pacific 310 Models (continued) simple dynamic pool 718 yield-per-recruit coral reef fishes 414 Monogenea — see Microcotyle macropharynx Morphology bluefish jaws 97 cranial variations, dusky dolphin 754 cods 210 crab lithodid, larvae 379 Lopholithodes mandtii 379 hakes, Atlantic 210 otolith 159 poacher blacktip 397 pygmy 397 roughy larvae 76 Morphometries dolphin, spotted 428 Mortality albacore 798 dolphin common 628 eastern tropical Pacific species 628 spinner, whitebelly 628 spotted 628 flounder summer 577 winter 577,816 sea urchin purple 237 red 237 Movements cod, Atlantic 281 conch, queen 516 haddock 281 lingcod 582 porpoise, harbor 440 salmon, chinook 171 New Zealand Pagrus auratus 159 New York Bight ocean dumping 594 Nonparametric methods isotonic regression 564 Ocean dumping 594 Ocyurus chrysurus — see Snapper, yellowtail Odontopyxis trispinosa — see Poacher, pygmy Oncorhynchus tshawytscha — see Salmon, chinook Ophiodon elongatus — see Lingcod Osteology poacher blacktip 397 pygmy 397 Otoliths ageing, rockfish, widow 573 flounder, winter 65 growth 1 59 mackerel, Spanish 526 INDEX: SUBJECTS Fishery Bulletin 91 (1-4), 1993 829 Otoliths (continued) microchemistry 732 microstructure 732 morphology 1 59 rockfish, widow 573 size 159 sole, Dover 732 Pacific Ocean albacore 371,649,798 anchovy, northern 310, 348 armorhead, pelagic 455 bonito, Pacific 348 coral reef fishes 414 crab, false southern king 664 dolphin abundance 183 dolphin 641 bottlenose 64 1 common 183, 641 gray grampus 64 1 melonheaded whale 641 roughtoothed 64 1 spinner 183,641 spotted 183,428,641 striped 64 1 gillnet, drift 371 lobster, spiny 1 mackerel chub 310,348 jack 348 porpoise, harbor 440 rockfish 87 roughy 76 sardine, Pacific 310 seamount 455, 573 Cobb 573 tuna bluefin 348 yellowfin 690 Pacific Ocean, eastern tropical dolphin abundance 775 dolphin 183,611,628,641,775 bottlenose 641 common 183, 641 spinner 183, 641 spotted 183,611,641 striped 64 1 Pagrus auratus 159 Pandalidae — see Shrimp Pandalus jordani — see Shrimp, ocean; Shrimp, pink Panuliridae — see Lobster, spiny Panulirus argus — see Lobster, Caribbean spiny Panulirus marginatum — see Lobster, spiny Paralichthys dentatus — see Flounder, summer Paralichthys lethostigma — see Flounder, southern Paralomis granulosa — see Crab, false southern king Parasites — see Echeneibothrium; Microcotyle macropharynx Parasitism armorhead, pelagic 455 scallop, calico 179 Pectinidae — see Scallop Penaeus duorarum — see Shrimp, pink Pentaeerotidae — see Armorhead, pelagic Peponocephala electro — see Melonheaded whale Phoca vitulina — see Seal, harbor Phocoena phocoena — see Porpoise, harbor Phocoenidae — see Porpoise, harbor Photography dolphin school size 641 Photoperiod effects seal harbor 710 northern fur 710 sea lion California 710 northern 710 Phycis chesteri — see Hake, longfin Pinnipedia — see Sea lion; Seal Pleuronectes americanus — see Flounder, winter Pleuronectidae — see Flounder; Sole, Dover Poacher blacktip 397 pygmy 397 Pogonias cromis — see Drum, black Pollution effects on spawning success 310 ocean dumping 594 Pomatomidae — see Bluefish Pomatomus saltatrix — see Bluefish Population dynamics albacore 371 Population studies anchovy, northern 310 armorhead, pelagic 455 dolphin, eastern spinner 775 fishes, Antarctic 475 genetic structure 619, 690 herring, Atlantic 338 lobster, Caribbean spiny 808 mackerel, chub 310 porpoise, harbor 440 sardine, Pacific 310 sea urchin purple 237 red 237 tuna, yellowfin 690 zooplankton, Georges Bank 464 Porpoise harbor 440 Predation — see also Mortality flounder summer 577 winter 577 Pseudopentaceros wheeleri — see Armorhead, pelagic Pseudopleuronectes americanus — see Flounder, winter Purse seine 195 Recruitment 237 anchovy, northern 310 armorhead, pelagic 455 Recruitment (continued) bluefish 389 mackerel, chub 310 sardine, Pacific 310 sea urchin purple 237 red 237 spot 8 Reef fishes larvae and juveniles 195 Reproduction spot 8 Reproductive biology crab, false southern king 664 dauphin, spotted 611 drum, black 244 flounder, winter 816 lobster, spiny 1 Lutjanus vittus 224 mackerel, Spanish 526 porpoise, harbor 440 seal harbor 710 northern fur 710 sea lion California 710 northern 710 spot 8 squid Idiosepius pygmaeus 260 Loligo chinensis 260 weakfish 165 Rhomboplites aurorubens — see Snapper, vermilion Rockfish 87,573 harlequin 573 rosy 573 widow 573, 676 yelloweye 304 Salmon chinook 171 Salmonidae — see Salmon, chinook Sampling design age-length key 382 Sarda chiliensis — see Bonito, Pacific Sardine Pacific 310, 348 Sardinella aurita 362 Sardinella brasiliensis 362 Spanish 362 Sardinella aurita 362 Sardinella brasiliensis 362 Sardinops sagax — see Sardine, Pacific Scales, ageing drum, black 558 Scallop calico 179 Iceland 564 Schoolmaster 619 School size, dolphin 641 Sciaenidae — see Drum; Spot; Weakfish Sciaenops ocellatus — see Drum, red Scomber japonicus — see Mackerel, chub Scorn beromorus maculatus — see Mackerel, Spanish 830 INDEX SUBJECTS Fishery Bulletin 91(1-1), 1 993 Scombridae — see Albacore; Bonito; Mackerel Scorpaenidae — see Rockfish Sea bass, black 328, 594 Seagrass conch, queen 516 Seal harbor 491, 710 northern fur 710 Sea lion California 710 northern 710 Steller — see northern Seamounts armorhead, pelagic 455 Seasonal studies lobster, Caribbean spiny 808 squid Idiosepius pygmaeus 260 Loligo chinensis 260 Sea urchin purple 237 red 237 Sebastes — see Rockfish Sebastes entomelas — see Rockfish, widow Sebastes rosaceus — see Rockfish, rosy Sebastes ruberrimus — see Rockfish, yelloweye Sebastes variegatus — see Rockfish, harlequin Selectivity 676 rockfish, widow 676 Sepioidea — see Squid Serranidae — see Sea bass, black Sexual dimorphism, dolphin, dusky 754 Sexual maturity — see Reproductive Biology Shrimp 129 brown 577 ocean 804 pink 804 Size estimation — see Age-size estimation Size limits, aggregate 804 Size-selectivity dredges, Iceland scallops 564 incorporating between-haul variation 564 Snapper Atlantic species 619 blackfin 619 cubera 619 dog 619 gray 619 lane 619 mahogany 619 mutton 619 red 619 schoolmaster 619 silk 619 vermilion 619, 699 yellowtail 619 Sole, Dover 732 Sparidae — see Pagrus auratus Spatial closures coral reef fishes 414 lobster, spiny 1 Spawning — see also Reproductive Biology anchovy, northern 310 fishes, marine 310 mackerel, chub 310 sardine, Pacific 310 Spawning biomass herring, Atlantic 338 sea bass, black 328 Spot growth 8 reproduction 8 spawning time 8 Stenella attenuata — see Dolphin, spotted Stenella longirostris — see Dolphin, spinner Stenella longirostris longirostris — see Dolphin, pantropical spinner or Dolphin, whitebelly spinner Stenella longirostris orientalis — see Dolphin, eastern spinner or Dolphin, whitebelly spinner Steno bredanensis — see Dolphin, roughtoothed Stock assessment dolphin, eastern spinner 775 rockfish, widow 676 Stock identification sardine, Spanish 362 Sardinella aurita 362 Sardinella brasiliensis 362 snappers, Atlantic 619 Stress developmental instability in Pacific hake 587 El Nino effects 587 fluctuating asymmetry in Pacific hake 587 Strombidae — see Conch, queen Strom bus gigas — see Conch, queen Strongylocentridae — see Sea urchin Strongylocentrotus franciscanus — see Sea urchin, red Strongylocentrotus purpuratus — see Sea urchin, purple Submersible surveys rockfish 87 yelloweye 304 Survey, aerial anchovy, northern 348 bonito, Pacific 348 mackerel chub 348 jack 348 sardine, Pacific 348 tuna, bluefin 348 Survival — see Mortality Swimbladder inflation 254 Tagging studies albacore 649 cod, Pacific 271 effect on annulus formation 558 lingcod 582 recapture data analysis 271 reporting rate, sablefish 271 ultrasonic telemetry, lingcod 582 Tautog 45 Tautoga onitis — see Tautog Taxonomy Aulotrachichthys sp. 76 dolphin, dusky 754 Hoplostethus atlanticus 76 Optivus sp. 76 Paratraehiehthys sp. 76 sardine, Spanish 362 Sardinella aurita 362 Sardinella brasiliensis 362 Trachichthyidae 76 Temperature distribution Atlantic cod 210 hakes, Atlantic 210 Spanish mackerel 151 zooplankton, Georges Bank 464 effects drum, red 23 Thunnus alalunga — see Albacore Thunnus albacares — see Tuna, yellowfin Thunnus thynnus — see Tuna, bluefin Towed nets 195 Trachichthyidae 76 Trachurus symmetricus — see Mackerel, jack Tracking, ultrasonic 171 Transplant, lingcod 582 Tuna albacore 371,798 bluefin 348 bonito, Pacific 348 mackerel, chub 348 yellowfin 690 Tursiops truneatus — see Dolphin, bottlenose Turtles, sea 129 Turtle Excluder Device 129 Trawling Efficiency Device 129 Ultrasonic telemetry, lingcod 582 Urophyeis chesteri — see Hake, longfin Urophycis chuss — see Hake, red Urophyeis eirrata — see Hake, gulf Urophycis earlli — see Hake, Carolina Urophycis floridiana — see Hake, southern Urophycis regia — see Hake, spotted Urophycis tenuis — see Hake, white Vertebrae albacore 694 Virginia mackerel, Spanish 151 tautog 45 Weakfish 165 Yield per recruit sea bass, black 328 Zalophus californianus — see Sea lion, California Zooplankton, Georges Bank 464 u S Postal Service STATEMENT OF OWNERSHIP, MANAGEMENT AND CIRCULATION Required b\ 39 V S C 3685) 1A Tme ot Publication FISHtRY BULLETIN IB PUBLICATION NO -370 2 Date ol Filing U-lb-93 3 Frequency of Issue QUARTERLY 3A No. of Issues Published Annually 3B Annuel Subsc 24. 00 dome iption Price StlC 30.00 foreign 4 Complete Mailing Address of Known Office of Publication (Street City, Count\; Suae and ZIP + 4 Code) (Not printers) NMFS Scientific Publications Office 7b00 Sand Point Way NE, Seattle, WA 98115-0070 5 Complete Mailing Address of ihe Headquarters of General Business Offices of the Publisher (Not printer) National Marine Fisheries Service, . NOAA/D0C 1335 East-West Highway, Silver Spring, MD 20910 6 Full Names and Complete Mailing Address ot Publisher, Editor, and Managing Editor (This item .MUST SOI be blank) Publisher (Name and Complete Mailing Address) Willis Hobart, NMFS Scientific Publications Office 7600 Sand Point Way NE, Seattle, WA 98115-0070 Editor (Same and Complete Mailing Address) Sharyn Matriotti, NMFS Scientific Publications Office 7600 Sand Point Way NE, Seattle, WA 98115-0070 Managing Editor (Name and Complete Mailing Address) same as above 7 Owner (If owned by a corporation, us name and address must be stated and also immediately thereunder the names and addresses of stockholders owning or holding I percent or more of total amount of stock. If not owned by a corporation, the names and addresses of the individual owners must be given If owned h\ a partnership oi other unincorporated firm, its name and address, as well as that of each individual must be given If the publication is published by a nonprofit organization, its name and address must be stated ) (Item must be completed ) Complete Mailing Address U.S. Department of Commerce 14th St.. NW. Washington DC 20250 Known Bondholders. Mortgagees, and Other Security Holders Owning or Holding 1 Percent or More of Total Amount, of Bonds, Mortgages or Other Securities llf there are none, so state) Complete Mailing Address 9 For Completion by Nonprofit Organizations Authorned To Mail ai Special Rates (DMM Section 423.12 only) The purpose, function, and nonprofit status of this organization and the exempt status for Federal income tax purposes (Check one) [YU Has Not Changed During izJ Preceding 12 Months □ Has Changed During Preceding 12 Months /// changed, publisher must submit explanation of change with this statement ) Extent and Nature of Circulation iSee instru Average No Copies Each Issue During Preceding 1 2 Months Actual No Copies of Single Issue Published Nearest to Filing Date A Total No Copies (Net Press Run) 2045 2038 b Paid andor Requested Circulation Handled by U.S. Superintendent of Documents, Washington, DC 20240 1 Sales through dealers and carriers, street vendors and counter sales 2 Mail Subscription (Paid and'or requested) C Total Paid and/or Requested Circulation l Sum or I0B1 and I0B2) 600 0 Free Distribution by Mail, Carrier or Other Means Samples. Complimantary, and Other Free Copies 143d E Total Distribution (Sum of C and D) 2038 F Copies Not Distributed 1 Office use. left over, unaccounted, spoiled after printing 2 Return from News Agents G TOTAL (Sum of E. Fl and 2— should equal net press ran shown in A) I certify that the statements made by me above are correct and complete PS Form 3526, Feb 1989 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 |~ Z!-[Z (Company or Personal Name) (Please type or print) (Additional address/attention line) (Street address) (City, State, ZIP Code) (Daytime phone including area code) (Purchase Order No.) v ; YES NO May we make your name/address available to other mailers? |~ | | VISA or MasterCard Account rrr (Credit card expiration date) Thank you for your order! (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 pro^tcpteriie us"d pr purchased because of this NMFS publication. Fishery Bulletin Guide for Contributors Preparation Title page should include authors' full names and mailing addresses and the senior author's telephone and FAX number. Abstract should not exceed one double-spaced typed page. It should state the main scope of the research but emphasize its conclusions and relevant findings. Because abstracts are circulated by abstracting agencies, it is important that they represent the research clearly and concisely. Text must be typed double-spaced throughout. A brief introduction should portray the broad significance of the paper; the remainder of the paper should be divided into the fol- lowing sections: Materials and methods, Results, Discussion (or Conclusions), and Acknowledgments. Headings within each section must be short, reflect a logical sequence, and follow the rules of multiple subdivision (i.e., there can be no subdivision without at least two items). The en- tire text should be intelligible to inter- disciplinary readers; therefore, all acronyms, abbreviations, and techni- cal terms should be spelled out the first time they are mentioned. The scientific names of species must be written out the first time they are mentioned; subsequent mention of scientific names may be abbreviated. Follow the U.S. Government Printing Office Style Manual, 1984 ed., the CBE Style Manual, 5th ed., and the most current issue of the American Fisheries Society's Common and Scientific Names of Fishes from the United States and Canada for fish nomenclature. Dates should be writ- ten as follows: 11 November 1991. Measurements should be expressed in metric units, e.g., metric tons as (t); if other units of measurement are used, please make this fact explicit to the reader. The numeral one (1) should be typed as a one, not as a lower-case el (1). Text footnotes should be numbered with Arabic numerals and typed on a separate sheet of paper. Footnote all personal communications, un- published data, and unpublished manuscripts with full address of the communicator or author, or, as in the case of unpublished data, where the data are on file. Authors are advised to avoid references to nonstandard (gray) literature, such as internal, project, processed, or administrative reports, wherever possible. Where these references are used, please in- clude whether they are available from NTIS (National Technical Information Service) or from some other public depository. Literature cited comprises published works and those accepted for publica- tion in peer-reviewed literature (in press). Follow the name and year sys- tem for citation format. In the text, cite Smith and Jones (1977) or (Smith and Jones, 1977). If there is a se- quence of citations, list chronological- ly: Smith, 1932; Green, 1947; Smith and Jones, 1985. Abbreviations of serials should conform to abbrevia- tions given in Serial Sources for the Biosis Previews Database. AUTHORS ARE RESPONSIBLE FOR THE AC- CURACY AND COMPLETENESS OF ALL CITATIONS. Tables should not be excessive in size and must be cited in numerical order in the text. Headings should be short but ample enough to allow the table to be intelligible on its own. All unusual symbols must be explained in the table heading. Other incidental comments can be footnoted with italic numerals. Use asterisks for probability in statis- tical data. Because tables are typeset, they need only be submitted typed and formatted. Zeros should precede all decimal points for values less than Figures include line illustrations and photographs (or slides) and must be cited in numerical order in the text. Unless photographs are submitted on glossy paper with good contrast, we cannot guarantee a good final printed copy. Figures are to be labeled with author's name and number of figure. Use Times Roman font (upper and lowercase letters) to label within figures. Avoid vertical lettering except for y-axis labels. Zeros should precede all decimal points for values less than one. Figures may be submitted in computer software form along with a laser-printed copy, as photomechanical tranfers (PMT), or as high quality photographic prints. Send only xerox copies of figures to the Scientific Editor; original figures will be re- quested later when the manuscript has been accepted for publication. Copyright law does not cover govern- ment publications; they fall within the public domain. If an author repro- duces any part of a government publi- cation in his work, reference to source is considered correct form. Submission Send printed copies (original and two copies) to the Scientific Editor: Dr. Ronald Hardy, Scientific Editor Northwest Fisheries Science Center, F/NWC3 National Marine Fisheries Service, NCAA 2725 Montlake Boulevard East Seattle, WA 98112-2097 Once the manuscript has been ac- cepted for publication, you will be asked to submit a software copy of your manuscript to the Scientific Editor. The software copy should be submitted in Word Perfect text format (or in standard ASCII text format if Word Perfect is unavailable) and should be placed on a 5.25-inch or 3.5- inch disk that is double-sided, double or high density, and that is compatible with either DOS or Apple Macintosh systems. Copies of published articles and notes are available free of charge to the senior author (50 copies) and to his or her laboratory (50 copies). Addi- tional copies may be purchased in lots of 100 when the author receives page proofs. MBL WHOI LIBRARY UH 1TUR