MBL U.S. Department of Commerce Volume 100 Number 1 January 2002 Fishery Bulletin U.S. Department of Commerce Donald L Evans Secretary National Oceanic and Atmospheric Administration Scott B. Gudes Acting Under Secretary for Oceans and Atmosphere National Marine Fisheries Service William T. Hogarth Acting Assistant Administrator for Fisheries Scientific Editor Dr. John V. Merriner Editorial Assistant Sarah Shoffler Center for Coastal Fisheries and Habitat Researcln, 101 Pivers Island Road Beaufort, NC 28516 NOS ^ATES O^ The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fish- eries Service, NOAA, 7600 Sand Point Way NE, BIN C 15700, Seattle. WA 98 1 15-0070. Periodicals postage is paid at Seattle, WA, and at additional mailing offices. POST- MASTER; Send address changes for sub- scriptions to Fishery Bulletin. Superin- tendent of Documents, Attn.: Chief. Mail List Branch, Mail Stop SSOM, Washing- ton, DC 20402-937.3. Although the contents of this publica- tion have. not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this peri- odical is necessary according to law for the transaction of public business of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $45.00 domestic and $56.25 foreign.. Cost per single issue: $28.00 domestic and $35.00 foreign. See back for order form. Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C15700 Seattle, Washington 98115-0070 Editorial Committee Dr. Andrew E. Dizon Dr. Harlyn O. Halvorson Dr Ronald W. Hardy Dr Richard D. Methot Dr. Theodore W. Pietsch Dr Joseph E. Powers Dr. Harald Rosenthal Dr. Fredric M. Serchuk National Marine Fisheries Service University of Massachusetts, Boston University of Idaho, Hagerman National Marine Fisheries Service University of Washington, Seattle National Marine Fisheries Service Universitat Kiel, Germany National Marine Fishenes Service Fishery Bulletin web site: fishbull.noaa.gov The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. It began as the Bulletin of the United States Fish Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office. Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions. State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 100 Number 1 January 2002 Fishery Bulletin Contents JAN 3 1 mi The conclusions and opinions expressed in Fishery Bulletin are solely those of the authors and do not represent the official position of the National Manne Fisher- ies Service (NOAA) or any other agency or institution. The National Marine Fisheries Service (NMFS) does not approve, recommend, or endorse any proprietary product or pro- prietary matenal mentioned in this puh- lication. No reference shall be made to NMFS. or to this publication furnished by NMFS, in any advertising or sales pro- motion which would indicate or imply that NMFS approves, recommends, or endorses any propnetary product or pro- prietary matenal mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication- Articles 1-10 Blick, D. James, and Peter T. Hagen The use of agreement measures and latent class models to assess the reliability of classifying thermally marked otoliths 11-25 Carmona-Suarez, Carlos A., and Jesus E. Conde Local distribution and abundance of swimming crabs (Calllnectes spp. and Arenaeus cribrarius) on a tropical arid beach 26-34 Crabtree, Roy E., Peter B. Hood, and Derke Snodgrass Age, growth, and reproduction of permit (Trachinotus falcatus) in Florida waters 35-41 Denson, Michael R., Wallace E. Jenkins, Arnold G. Woodward, and Theodore I. J. Smith Tag-reporting levels for red drum (Saaenops ocellatus) caught by anglers in South Carolina and Georgia estuaries 42-50 Faunce, Craig H., Heather M. Patterson, and Jerome J. Lorenz Age, growth, and mortality of the Mayan cichlid (Cichlasoma urophthalmus) from the southeastern Everglades 51 -62 Hastings, Kelly K., and William J. Sydeman Population status, seasonal vanation in abundance, and long-term population trends of Steller sea lions (Eumetopias jubatus) at the South Farallon Islands, California 63-73 McBride, Richard S., Michael P. Fahay, and Kenneth W. Able Larval and settlement periods of the northern searobin (Prionotus carollnus) and the striped searobin (P. evolans) 74-80 Pennington, Michael, Liza-Mare Burmeister, and Vidar Hjellvik Assessing the precision of frequency distributions estimated from trawl-survey samples Fishery Bulletin 100(1) 81-89 Potts, Jennifer C, and Charles S. Manooch III Estimated ages of red porgy (Pagrus pagrus) from fishery-dependent and fishery-independent data and a comparison of growth parameters 90-105 Romanov, Evgeny V. Bycatch in the tuna purse-seine fisheries of the western Indian Ocean 106-116 Sainte-Marie, Bernard, and Denis Chabot Ontogenetic shifts in natural diet during benthic stages of American lobster (Homarus americanus), off the Magdalen Islands 117-127 Zug, George R., George H. Balazs, Jerry A. Wetherall, Denise M. Parker, and Shawn K. K. Murakawa Age and growth of Hawaiian seaturtles (Chelonia mydas): an analysis based on skeletochronology Notes 128-133 DiNardo, Gerard T., Edward E. DeMartini, and Wayne R. Haight Estimates of lobster-handling mortality associated with the Northwestern Hawaiian Islands lobster-trap fishery 134-142 Graves, John E., Brian E. Luckhurst, and Eric D. Prince An evaluation of pop-up satellite tags for estimating postrelease survival of blue marlin (Makaira nigricans) from a recreational fishery 143-148 Hazin, Fabio H. V., Paulo G. Oliveira, and Matt K. Broadhurst Reproduction of blacknose shark (Carcharliinus acronotus) in coastal waters off northeastern Brazil 149-152 Porch, Clay E., Charles A. Wilson, and David L. Nieland A new growth model for red drum (Sciaenops ocellatus) that accommodates seasonal and ontogenic changes in growth rates 153 Subscription form Abstract-Otolith thermal marking is an I'llii'it'nt method for mass mark- ing hatehcry-rcared salmon and can be used to estimate the proportion of hatchery fish captured in a mixed-stock fishery. Accuracy of the thermal pattern classification depends on the promi- nence of the pattern, the methods used to prepare and view the patterns, and the training and experience of the per- sonnel who determine the presence or absence of a particular pattern. Esti- mating accuracy rates is problematic when no secondary marking is avail- able and no error-free standards exist. Agreement measures, such as kappa I K). provide a relative measure of the reliability of the determinations when independent readings by two readers are available, but the magnitude of k can be influenced by the proportion of marked fish. If a third reader is used or if two or more groups of paired read- ings are examined, latent class models can provide estimates of the error rates of each reader. Applications of K and latent class models are illustrated by a program providing contribution esti- mates of hatchery-reared chum and sockeye salmon in Southeast Alaska. The use of agreement measures and latent class models to assess the reliability of classifying thermally marked otoliths* D. James Blick Peter T. Hagen Alaska Department of Fish and Game Division of Commercial Fisheries 10107 Bentwood Place Juneau, Alaska 99802-5526 E mail address ((or P T Hagen, contact author) peter hagenmifishgame state ak us Manuscript accepted 16 April 2001. Fish. Bull. 100:1-10(2002). The ability to induce patterns in salmon otoliths by manipulating water temper- atures has proved to be an efficient means for marking large numbers of salmon (Volket al., 1990). Wlien salmon embryos or alevins are exposed to a rapid drop in temperature, otolith growth is temporarily disrupted, and this results in a discontinuity in the otolith "s microstructure. When viewed under transmitted light microscopy, this discontinuity appears as a dark ring. By controlling the number of tem- perature drops and the timing between drops, a coded pattern of dark rings can be recorded on the otolith and this pattern can be recovered from otoliths of older fish by removing the overlay- ing material and exposing the otolith core. For hatcheries that release a large number of fish, this type of marking method has shown to be particularly cost effective for marking 100% of the releases (Munk et al. 1993). Several fisheries management pro- grams in Alaska use thermal marking to estimate hatchery contributions to commercial fisheries (Hagen et al., 1995). Typically, several hundred salm- on otoliths are systematically collected during each two- or three-day com- mercial opening during the fishing sea- son. The otoliths and sampling data are shipped to a processing laboratory where a subsample of otoliths (generally 50 to 100) are processed immediately to meet in-season management needs; a portion of the remaining otoliths are processed later to provide an overall es- timate of hatchery contribution to the fisheries. The process by which a reader de- termines the presence or absence of a thermal mark in an otolith can be char- acterized as one of pattern recognition and image matching. Prior to examin- ing otoliths of unknown origin, the read- ers gain familiarity with the patterns likely to be encountered by carefully examining fry otoliths that were ob- tained after thermal marking but prior to their release into the wild. Because there can be wide variation in the ap- pearance of the thermal marks within a mark group (due in part to differenc- es in developmental stages at marking), a single mark group may be represent- ed by a variety of patterns. As a result, secondary characteristics and measure- ments of the patterns are sometimes necessary to identify an otolith to a mark group. The examination is also used to confirm that all the hatchery fish have been successfully marked. The process of making a determina- tion on otoliths from returning adult salmon can become problematic be- cause wild salmon may also contain otolith patterns that can mimic the fea- tures imposed through thermal mark- ing. Referred to as "noisy patterns," their presence can increase the rate of false positives. Conversely, if the hatch- ery employs poor temperature control or unintended disruptions occur around the period of marking, it may be diffi- cult to identify the otolith as that of a * Contribution PP-184 of the Alaska De- partment of Fish and Game, Commercial Fisheries Division, Juneau, Alaska 99802- 5526. Fishery Bulletin 100(1) hatchery fish, and this would increase the rate of false negatives. Differences between readers in skill and train- ing level, and how they process otoliths, can add to the un- certainty in estimating the accuracy of the readings and the rates of false positives and negatives. Otolith marking generally takes place without any sec- ondary marking, such as fin-clipping or coded-wire-tag- ging; therefore the accuracy of a reading cannot directly be determined through conventional methods that make use of a "gold standard" (known origin sample) or other error-free classification methods. To ensure that the in- formation provided to the Alaskan fisheries managers is accurate, each otolith is independently examined by two readers, and a third reading is used to resolve differenc- es between the first two readings. The resolved readings are used to estimate the contribution of hatchery fish, and the presumption of accuracy is based on the premise that, through multiple readings, all marked fish are ei- ther correctly identified or that errors, if present, are in- consequential. Developing the analytical tools to deter- mine the veracity of that assumption is the objective of this investigation, and by establishing such tools, quality control standards for recovering thermal marks can be developed. In developing the tools to measure the quality of otolith readings, three questions are addressed: 1 How to assess the reliability of otolith readings when no standards are available. 2 How to estimate the proportion of hatchery marks when there is disagreement between two or more readers. 3 How the precision of the estimate of the proportion is influenced by classification error We discuss two approaches: 1 ) indices of agreement typi- cally used in reliability studies, and 2) latent class models where classification errors are estimated for each reader even though the true error rate is considered unknown. The data requirements and their attendant assumptions are presented for each approach. The methods are illus- trated by examining among-reader comparisons of chum salmon (Oncorhynchus keta) and sockeye (Oncorhynchus nerka) salmon otoliths collected from programs that moni- tor inseason contributions of hatchery fish in several com- mercial fisheries in Southeast Alaska (Hagen et al., 1995). The results are used to provide recommendations for mon- itoring the quality of otolith readings for thermal marking programs. Table 1 Notation used to show the cross-classification of a sample of fi otoliths by two readers to either hatchery (H) or wild stock (W) assignment. Row and column sums are indicated by the subscript "." Reader 1 H Reader 2 "h- H W "hh "hh W "WH "ww «w "•H "w /; 2 is infallible (or is considered a "gold standard"), unbiased estimates of the accuracy and error rates of reader 1 and the proportion of hatchery stocks (p) are given by '^HlH ~ "hh/"h- '^WjH ~ "\VH ^ " H - 1 ■'''hIH '^wjw ~ "vv\\7" w- '^Hiw ~ "hw I " w = 1~ ■'^w|w P = "h/". (where, for example, ;r\v|n refers to the probability that reader 1 classifies an otolith as W when its true state is H). These estimates reflect the fact that reader 2 is infallible; the accuracy rates CThih' '^wiw' ^"d the error rates CTwifi- Tc■^^,^^) are conditional on the numbers of hatchery or wild stock otoliths as determined by reader 2. No standard available If a standard is not available, an unbiased estimate of p can be obtained if the accuracy rates for reader 1 are known. The estimate is p* = ("n/«+^wr I'/f'^HI H|H ' W|W 1), where n■^^ is the number of otoliths classified as hatchery otoliths. If the accuracy rates are estimated, thenp* will no longer be unbiased, but will be much less biased than the estimator n■^^ln and will in general have a much smaller mean-squared error (Rogan and Gladen, 1978). For a Bayesian approach to this problem, see Viana et al. ( 1993 ) and Joseph et al. ( 1995). Methods Standard available A sample of /i otoliths, which are examined by two readers, can be cross-classified as hatchery (H) or wild stock (W) as in Table 1. Suppose we wish to estimate the accuracy rate (probability of making a correct classification) or con- versely, the error rate ( probability of making a wrong clas- sification). If we know nothing about reader 1, but reader Agreement measures When accuracy rates are unavail- able, statistics that measure "agreement" between readers are often calculated (e.g. Fleiss, 1981). One such index is simply the proportion of observed agreement (P„), defined as :(/(. )ln. Another index, called kappa (k), corrects P„ for the degree of agi'eement that is expected by chance alone. It is defined as Blick and Hagen Use of agreement measures and latent class models to assess the reliability of classifying ttiermally marked otoliths 3 K = iP„-P,.)/(l-P^.), where P,, = expected agreement = ('!h"h + "w"w'^"^- ^^^ divisor, 1 - P., constrains k to be less than or equal to one, and if all agreement is due to chance {P^=PJ, then k: equals zero. Note that with k; independence between readers is assumed in order to calculate expected agreement. An example of how agreement indices can be used to monitor readings is shown in Figure 1, which displays k and its standard error for 2874 chum otoliths readings di- vided into 27 groups based on different reader pairs and capture locations. Included are P,'s for four of the groups. The results indicate that v levels were similar between the different groups, suggesting overall consistency in read- ings, although some of the groups had lower values, which in practice would invite further investigation. The Pg's in Figure 1 have a different rank order than the ic values. This apparent discrepancy highlights a potential problem in interpretation when using agreement indices to draw conclusions. To help illustrate this point, consider the following examples (Table 2). Table 2A is generated as the expected counts, given ;rj,|j^ = 0.9 and %|w = 1-0 for both readers, and p = 0.1. In this case, P, = 0.98 and k: = 0.89. On the other hand. Table 2B is generated under the same assumptions except that rt^n = 0.5. In this case P„ drops only slightly to 0.95, whereas v drops to 0.47. Be- cause the hatchery stock is rare, the inability of the read- ers to detect the mark is not well reflected by P„ whereas k reflects it better by correcting for the high level of chance agreement. Now let K, HIH 0.9 and /Twiw = 0.9 for both readers, and 0.64. On P= 0.5 (Table 2C). In this case, P, = 0.82 and k the other hand. Table 2D is generated under the same as- sumptions except that P= 0.05. In this case, P, remains unchanged at 0.82, but \' drops to 0.25. In none of the above examples is the index "wrong." Rather, as is the case with most indices, interpretation is affected by the values of the underlying parameters. In the latter example (Table 2, C-D), even though P, is the same for C and D, the scale it is being compared with has changed, thus changing the value of k. This increases the difficulty of comparing k across populations with differ- ent underlying proportions. Note also that Table 2D could have been derived from %|h = 0.5 and ttwiw = 0.944 for both readers, andp = 0.19. Thus, without additional infor- mation, it is impossible to draw reliable conclusions about reader accuracies or the proportion of hatchery marks. Although agreement measures can be ambiguously in- terpreted, in practice they can still sei've a useful moni- toring role during routine comparisons when the circum- stances of the readings are fairly well characterized. The interpretive difficulties with indices such k and P, become apparent when trying to translate agi"eement measures into statements about the accuracy of different readers and about the influence of reading error on the contribu- tion estimates. Latent class models An alternative approach is to try to estimate tTj^, j^ and tt^viw f""" each reader, along with p. Although at first thought this may seem impossible, it can 1 ra 0.6 04 02 T -^ 8si 920 tl J_ J_ 10 20 Group number 30 Figure 1 The values of k{±1 SE) from 27 gi'oups of paired read- ings of chum salmon otoliths (total=2874). The groups are based on pairs of different readers examining oto- liths collected at different times and locations. The pro- portion of agreement (P,) is shown next to group 4, 7, 9, and 12 for comparison with the value of k. be shown that either by setting a few constraints or by col- lecting additional information, estimation is indeed pos- sible. This problem falls into the category of latent class modeling (e.g. Everitt, 1984; Bartholomew, 1987; McCutch- eon, 1987; Clogg, 1995). Latent class models (LCMs) belong to a family of latent variable models that hypothesize the existence of unobservable "latent" variables, about which information can be obtained only though measurements on observable "manifest" variables. LCMs specifically restrict the latent and manifest variables to be categorical. In the present situation, the latent variable is the true class (H or W) to which the otolith belongs, whereas the mani- fest variables are the readers' classifications. Such models have been used for assessing reliability of diagnostic tests in the medical field over the last 20 years (see Walter and Irwig, 1988; Formann, 1996, for reviews). Returning to the problem with two readers, neither of which is a standard, there are five essential parameters to estimate: s-i)|H,^H|H'^w|w.'fw|'w ' andp, with only 3 df (four pieces of data, /i^H' "hw "wH' "ww- minus one because the sample size, n, is fixed). Thus, the model is overparameter- ized, and either constraints on the parameters or more da- ta are needed. Possible constraints include 1) considering that two of the parameters are known (e.g. /r^vjw = Tw|w = ^• i.e. both readers always call a wild stock correctly, there are no "false positives"), or 2) considering that two sets of parameters are equal (e.g. t1|'|'h , 7r|f|H , ;r\v|'\v ='fwi'w' i-^- the accuracy rates are the same for both readers). Although there may be times when such constraints are realistic, in general they will not be; therefore more infor- Fishery Bulletin 100(1) mation will be necessary. One way to generate more in- formation is to have a third independent reader (Walter, 1984). With three readers, there are seven essential pa- rameters; 7i'ii;^"-'''\r^'^\i,''-"" and p. There is also 2^ - 1 = 7 df, so that all the parameters are estimable. Estimation is most commonly done by the method of maximum likeli- hood. If readings are assumed to be independent among read- ers and among otoliths, the likelihood function is i = H,\V ) = H,\V*:H,VV This likelihood function must be maximized numerically and methods for this computation will be discussed later If more than three readers are used, there are extra de- gi-ees of freedom that can be used to assess goodness-of-fit. For example, with four readers there will be nine param- eters with 15 df leaving 6 df for goodness-of-fit. Pearson chi- square or likelihood ratio G'-^ tests would both be applicable. Another way to generate additional information was proposed by Hui and Walter ( 1980). Suppose there are two or more strata with different hatchery proportions in each strata. For example, catch could be stratified temporally or spatially. If it is assumed that ;r||||| and /Ty^iw remain constant over strata, then a solution for just two readers may be obtained. For example, if there are two readers and two strata, then there are six parameters; 'rH|H"'>'i'w|w' > Pj, and p.,, with 2(2'^ - 1) = 6 df Increasing the number of strata increases the degrees of freedom; e.g. three strata for two readers gives 3(2^ - 1) = 9 df for 7 parameters. The likelihood function for two readers and S strata is fin ni^^'^^iH'^^^+'i-^. (1) 12) 1" 'Iw'TjIWl g=l (=H,W_/ = H.W Table 2 Examples from cross-classification data generated as expected counts from a sample of 1000 otoliths based on different accuracy rates for identifying hatchery fish < tt,., | ,^ I and wild fish (/Twiw' under different mark proportions tp). The examples used illustrate differences between obsei"ved agreement IP, i and chance-corrected agi-eement U') under different underlying conditions. A H Reader 2 90 ^Hl H ~ 0.9 P„ = 0.98 ' H W Reader 1 81 9 W 9 901 910 % \V ~ 1.0 K- 0.89 Total 90 910 1000 /' = 0.1 B H Reader 2 50 ^11 n = 0.5 P, = 0.95 H W Reader 1 2.5 25 W 25 925 950 % w = 1.0 K- = 0.47 Total 50 950 1000 P = 0.1 C Reader 1 H Reader 2 500 '^ll|H = 0.9 P„ = 0.82 H W 410 90 W 90 410 500 %■ w - 0.9 V = 0.64 Total 500 500 1000 P = 0.5 D Reader 1 H Reader 2 140 'fH H = 0.9 P, = 0.82 H W 50 90 W 90 770 860 Tu |\V = -0.9 K = 0.25 Total 140 860 1000 P = 0.05 A third way to supply additional information is to take a Bayesian approach (see "Discussion" section). By speci- fying prior distributions of the model parameters, unique estimates can be obtained (Joseph et al., 1995). A critical assumption in the above models is that read- ings are independent. Specifically, the reading of each oto- lith by a given reader is independent of any other reading by the same reader, and each reading by various readers on a given otolith is independent given the true state of the otolith. In principle, the latter assumption may be dif- ficult to meet especially if all readers examine the same otolith. The fact that the otolith is not prepared indepen- dently by each reader could induce a dependence among the readers. Also, variability in the readability of the mark due to the marking process can induce a dependence. Such dependence can bias the estimators of n and p (Vacek, 1985). Note that this latter assumption of independence is also required for v. One remedy for the problem of dependence due to prepa- ration is to require independent preparations. This however, requires additional otoliths and with only two otoliths per fi.sh, this would limit the number of readers to two. But in practice, this may not be a large concern. Typically, the second reader has the option to provide additional process- ing effort to the first otolith or, if needed, to process the second otolith. In almost all cases additional preparation is not done and readers feel they are able to extract suf- ficient information about the presence or absence of a mark from each other's preparations. In addition, reader accura- cy rates obtained by LCM do not appear to vary systemati- cally with the reading order, which also suggests that prep- aration-induced dependency is not a significant factor Dependency associated with variability in the appear- ance of the mark may be harder to address. A general so- lution is to model the dependence with additional param- eters (e.g. Vacek, 1985; Qu et al., 1996; Yang and Becker, 1997; Qu and Hagdu; 1998; Albert et al., 2001). Modeling dependence requires either more readers or more strata. These modeling approaches are complicated and are cur- rently evolving (see Albert et al., 2001). Alternatively, ad- Blick and Hagen: Use of agreement measures and latent class models to assess the reliability of classifying tfiermally marked otolitfis ditional latent classes may be added (Christenson ct al., 1992; Forniann, 1994), e.g. a third class of otoliths from ambiguous sources. In the previous discussion concerning three or more readers, we implied that readers were different individu- als. This need not be so; what is required are three or more independent readings. If it were possible for the same in- dividual to read the same otolith more than once, indepen- dently, then the number of different readers could be re- duced. If independence could not be met, the dependence could be modeled, as discussed above. Another critical assumption, but one that should be met most of the time, is that the individual accuracy rates are known to be either greater than or less than the error rates (e.g. %|h > ^wm ^^'^ %-|W ^ %|W' which im- plies that ^Tj^iH and JT^,-^ are either greater than or less than 0.5) because of an inherent symmetry in the problem that results in the same likelihood function being gener- ated when the error rates are switched with the accuracy rates. Computation Formulas for estimating \'and its standard error are straightforward (Fleiss, 1981). Estimates can also be obtained from several software packages including PROC FREQ in SAS (SAS Institute, 1989). Maximizing either of the likelihood functions for the LCMs requires a numerical procedure. The most straight- forward is to use an optimization routine such as "Solver" in Excel (Microsoft Corporation, 1993) or "nlminb" in S- PLUS (Statistical Sciences, 1995). Alternatively, the EM algorithm (Dempster et al., 1977; Dawid and Skene, 1979; McLachlan and Krishnan, 1997) can be easily used. The simplicity of the EM algorithm follows from the recogni- tion that the LCM is an example of a finite mixture prob- lem, specifically, in this case, a mixture of multivariate Bernoulli distributions with mixing parameter p (Everitt, 1984). Use of the EM algorithm for such mixture prob- lems in fisheries is well documented, e.g. for stock compo- sition estimates (Millar, 1987; Pella et al., 1996) and for age-length keys (Kimura and Chikuni, 1987). A more ef- ficient alternative to the EM algorithm is to use iteratively reweighted least squares (Agresti, 1990). This method is relatively easy to implement in software such as PROC NLIN in SAS (SAS Institute, 1989). Perhaps the most di- rect and efficient way would be to use LCM software. We are not aware of any routines for LCMs in any major statistical package at present, but several independent LCM packages exist (for a review, see Clogg, 1995; and for an Internet listing see http://oui-world.compuserve.com/ homepages/jsuebersax/index.htm). As with many maximum likelihood problems, where nu- merical methods must be used, complications can arise. Constraints may at times be needed to ensure that pa- rameter estimates fall in acceptable intervals (e.g. [0,1] for p and [0.5,1] for the ;r's). Also the likelihood function may have local maxima, which means that several runs with varying starting values may be necessary to identify the global maximum. Finally, estimates of standard er- rors may entail additional computing. PROC NLIN in SAS provides asymptotic (i.e. large-sample) standard errors. Jackknife and bootstrap estimates are relatively easy to program, the jackknife being much less computationally intensive. Finally, the Bayesian programs discussed in Joseph et al. (1995) can be found at http://www.epi.mcgill.ca/Josepli/ software. html. Examples The first example analyzes the results of three readers examining 570 chum otoliths. The samples were taken from a common location, and the readers were familiar with the patterns. Each reading was made without knowl- edge of prior readings. The data, along with pairwise k estimates and the LCM parameter estimates (using PROC NLIN in SAS; see appendix for code) are presented in Table 3. These results indicate that the third reader is signifi- cantly (a=0.05) less able to correctly identify a hatchery mark when it is present and that there are no significant differences among readers in their ability to detect a wild mark when it is present. These conclusions are readily ap- parent from the table of results, and although the pairwise K"'s are consistent with these results, they are more dif- ficult to interpret. With the variance due to sampling es- timated to be (0.7379X1 - 0.7379)/(570 - 1) = 0.0003399, misclassification error contributes only 0.36% to the total variance. The second example consists of two readers with four spatial strata. Samples were obtained from sockeye salm- on caught in four neighboring Alaskan gillnet fisheries in central Southeast Alaska. The data and the LCM esti- mates are shown in Table 4. These estimates indicate that the readers are not statistically different in their ability to detect hatchery marks, whereas the second reader is bet- ter able to distinguish wild marks. With eight parameters and 12 df there are 4 df available for a goodness-of-fit test. Pearson's chi-square yields 4.83, which with 4 df, has a p-value of 0.306, thus indicating an acceptable model fit. Misclassification error contributes from about 8% to 14% to the total variance in the estimates of the proportion of hatchery stock. Design considerations Design of an otolith reading program is complicated by misclassification error. An important consideration is the precision of the estimates, in particular the precision of the estimate ofp. Table 5 shows the asymptotic standard error of p for various combinations ofp, /r^iH' ^^^ ^wiv! f'"' '-^e three-reader model with unknown accuracies, and the one-, two-, and three-reader models with accuracies assumed known. Although this table is derived for a sample of 1000 otoliths, the ratio of any two standard errors within the table would be the same for any sample size (assuming the sample size is large enough to approximate the asymptotic conditions). It is evident that misclassification inflates the standard error over the usual binomial case (right-most column). The table also makes clear the increase in the uncertainty of estimating p when the accuracies also have Fishery Bulletin 100(1) Table 3 Cross-classification data and results for 570 chum otoliths examined bv three readers showing the parameter estimates and stan- dard errors from the latent class model, followed by a comparison of the differences among reader pan's by jsing kapp 3 and the latent class model (LCM) accuracy rates. The data show that the high agreement among read ers as to hatcher V and wild ( lassifica- tion (e.g. HHH= 406 and WWW= = 135) is reflected in the overall high accuracy rates estimated from the LCM However the model also shows that reader 3 has a significantly lower accuracy rate in detecting hatchery marks (;rij5'|H=0.969) than the other readers. Reading Count LCM Parameter Estimate SE HHH 406 'Thih 0.998 0.002 HHW 13 'f'&IH 0.998 0.002 HWH 1 '^'^IH 0.969 0.008 WHH 1 f'^'jW 0.958 0.017 HWW 6 t'w/|w 0.986 0.010 WHW 2 *rl3t 0.957 0.017 WWH 6 P 0.738 0.018 WWW 135 Reader pairs K SE Difference in tTj^ih SE Difference in ^-^v SE land 2 0.954 0.014 0.000 0.004 -0.028 0.020 lands 0.882 0.022 0.029 0.009 0.000 0.024 2 and 3 0.901 0.021 0.029 0.009 0.028 0.020 Table 4 Cross-classification data for 2340 sockeye otoliths e.xamined bv two readers and stratified by four fishing districts showing the estimates of the latent class parameters and their standard errors. Between- reader comparison is based on whether the difference in accuracy estimates are significantly different th an zero. The result s indicate that the readers were not statistical! V different in detecting hatchery marks ' "^H 1 H ' ^"^'- were statistically different in detecting wild marks (;rw|w'LCM = latent class model. Fishing districts 108-30 108-50 106-41 106-30 HH 152 127 85 20 HW 11 9 21 5 WH 2 6 5 1 WW 271 382 832 411 n 436 524 943 437 LCM parameter Estimate SE Reader difference SE '^hih'" rr <2> "HjH 0.980 0.964 0.013 0.021 0.017 0.025 IT 11' "w 1 W TT 12' ''W|W 0.984 0.997 0.005 0.003 -0.013 0.006 P108-30 0.366 0.024 Pi 08-50 0.257 0.020 P1O6--H 0.096 0.010 P1O6-3O 0.047 0.011 to be estimated in the three-reader case. For example, if = 0.8 for all three readers, one would have to '^HlH ''wlw have almost twice (0.035/.019=1.84) the sample size to esti- mate ap of about 0.5. Once accuracy estimates for the read- ers are obtained, dropping one or even two readers may be appropriate, although the assumption must be made that the accuracy rates will be constant for the remainder of the program. Maintaining two readers will allow for that Blick and Hagen: Use of agreement measures and latent class models to assess the reliability of classifying thermally marked otoliths Table 5 AsyniptolK' Uand ard errors Ibr the cs timalcd ])! opor'tion of marked fish p, for various combinat ons of accuracy rates in identify- | iiifj; halclu'rv fish. ;r|,|„,and wild fisli, %|W'"«1 mark proportion p, for a sample of 1000 otoliths Val ues are reported foi the cases u liero accur icy r ites, K. are the same and assumed known or one, two. or three readers and for the case w lere ;r's are estimated lor three readers. Table illustrates how misclassification will increase standard errors in the estimate of hatchery proportion. '''n 1 11 0.8 0.9 1.0 '^W 1 w 0.8 0.9 1.0 0.8 0.9 1.0 0.8 0.9 1.0 ,'i readers P 0.1 0.032 0.016 0.011 0.023 0.013 0.010 0.018 0.011 0.009 1 rfs estimated) 0.3 0.034 0.021 0.017 0.024 0.017 0.015 0.020 0.015 0.014 0.5 0.035 0.023 0.019 0.023 0.018 0.016 0.019 0.016 0.016 0.7 0.034 0.024 0.020 0.021 0.017 0.015 0.017 0.015 0.014 0.9 0.032 0.023 0.018 0.016 0.013 0.011 0.011 0.010 0.009 3 readers 0.1 0.013 0.011 0.010 0.011 0.010 0.009 0.010 0.010 0.009 iffs known 1 0.3 0.018 0.016 0.015 0.017 0.015 0.015 0.015 0.015 0.014 0.5 0.019 0.018 0.016 0.018 0.017 0.016 0.016 0.016 0.016 0.7 0.018 0.017 0.015 0.016 0.015 0.015 0.015 0.015 0.014 0.9 0.013 0.011 0.010 0.011 0.010 0.010 0.010 0.009 0.009 2 readers 0.1 0.015 0.013 0.010 0.013 0.011 0.010 0.011 0.010 0.009 ( ;r's known ) 0.3 0.020 0.018 0.015 0.018 0.016 0.015 0.015 0.015 0.014 0.5 0.022 0.019 0.016 0.019 0.018 0.016 0.016 0.016 0.016 0.7 0.020 0.018 0.015 0.018 0.016 0.015 0.015 0.015 0.014 09 0.015 0.013 0.011 0.013 0.011 0.010 0.010 0.010 0.009 1 reader 0.1 0.023 0.017 0.011 0.020 0.015 0.010 0.018 0.014 0.009 (.It's known) 0.3 0.026 0.021 0.017 0.022 0.019 0.016 0.020 0.017 0.014 0.5 0.026 0.022 0019 0.022 0.020 0.017 0.019 0.017 0.016 0.7 0.026 0.022 0.020 0.021 0.019 0.017 0.017 0.016 0.014 0.9 0.023 0.020 0.018 0.017 0.015 0.014 0.011 0.010 0.009 assumption to be checked because there will now be extra degrees of freedom to assess goodness-of-fit (there are 3 df. but only one parameter. p, needs to be estimated). Esti- mates of p can still be obtained with one reader, but there can be no check of the assumptions. Also, there can be a significant increase in uncertainty in the estimate in using only one reader. Discussion There are numerous classification problems in fisheries that require the judgment of trained individuals. In many of those situations no "gold standard" is available to test those judgments, and it becomes necessary to apply other methods to determine the veracity of the classifications. Reading thermally marked otoliths is a particularly good example of this problem because thousands of classifica- tion decisions are needed each year to provide estimates of hatchery contributions. The common approach for assessing the quality of the readings, in the absence of having samples of known origin, has been to collect independent and multiple readings on the samples, and to presume that agreement between read- ings can serve as a proxy for reading accuracy. Agreement indices such as k" are very easy to compute, and they have utility in that they can serve as flags to indicate reading problems. However, as was shown here, they also suffer dif- ficulties in interpretation. Also, the indices in themselves do not provide inferences about the relative skill of differ- ent readers in pulling out a particular set of patterns. Latent class models provide an approach with readily interpretable quantities for a modest computational cost. Classification accuracies or errors are direct, meaningful parameters unlike an index of agreement. In addition, es- timates of p are available. These models can be readily ex- tended to the case of more than two outcomes, e.g. multiple hatchery marks. These models could also be useful in oth- er applications, such as in aging fish or in the identifica- tion of any character for which there is no "gold standard" (e.g. field identification of species or sex). A somewhat sim- ilar analysis has been proposed for aging (Richards et al., 1992), although the link to LCMs was not discussed. LCMs can handle fairly complicated situations, including ordered classes (Croon. 1990), continuous manifest vari- ables, and parameter constraints (see Clogg, 1995, and Krzanowski and Marriott, 1995. for reviews). We have not discussed the Bayesian approach to these problems in great detail, but we believe it has much to offer in that it can incorporate prior information, either Fishery Bulletin 100(1) in the form of expert opinion (e.g. Demissie et al., 1998) or in the form of results of earher analyses (e.g. Viana et al., 1993). Rather than assuming that estimated accu- racies are "known," one can incorporate the uncertainty in the estimates into the prior distributions. In addition, the Bayesian approach does not rely on asymptotic results that may behave poorly with small samples. We have also not assessed the possible bias due to the lack of indepen- dence in the readings. When suitable software becomes available, this assumption should be checked. In our examples above, misclassification error contribut- ed relatively little to the overall uncertainty. In these ap- plications, where estimates of hatchery contribution were used to make management decisions, the accuracy of read- ings were within an acceptable range. However, the criteria used to establish quality control standards in any program need to be developed in the context of how the information is to be used along with other sources of uncertainty. In conclusion, we believe that the use of agreement mea- sures in combination with latent class models can con- tribute significant information about both the proportions of interest and the quality control aspects of an otolith- marking program. Furthermore these approaches could have application to similar areas in fisheries which re- quire judgments that are not free of error. Acknowledgments We thank Bob Wilbur for editorial comments and three anonymous reviewers for valuable suggestions. Literature cited Agresti, A. 1990. Categorical data analysis. John Wiley, New York, NY, 576 p. Albert. P. S., L. M. McShane, and J. H. Shih. 2001. Latent class modeling approached for assessing diag- nostic error without a gold standard: with applications to p53 immunohistochemical assays in bladder tumors. Bio- metrics 57:610-619. Bartholomew, D. J. 1987. Latent variable models and factor analysis. Oxford Univ. Press, New York. NY'. 427 p. Christen.sen A. H., T. Gjorup, J. Hilden, C. Fenger. B. Henriksen, M. Vyberg, K. Ostergaard, and B. F. Hansen. 1992. Observer homogeneity in the histologic diagnosis of Helicobacter pylori: latent class analysis, kappa coefficient, and repeat frequencv Scand. J. Gastroenterol. 27:933-939. Clogg, C. C. 1995. Latent class models. Chapter 6 ;;; Handbook of sta- tistical modeling for the social and behavioral sciences (G. Arminger, C. C. Clogg, and M. E. Sobel, eds.), p. 311-359. Plenum Press, New York, NY. Croon, M. 1990. Latent class analysis with ordered classes. Brit. J. Math. Stat. Psych. 43:171-192. Dawid, A. P., and A. M. Skene. 1979. Maximum likelihood estimation of observer error- rates using the EM algorithm. Appl. Statist. 28:20-28. Demissie, K., N. White, L. Joseph, and P. Ernst. 1998. Bayesian estimation of asthma prevalence, and com- parison of exercise and questionnaire diagnostics in the absence of a gold standard. Ann. Epidemiol. 8:201-208. Dempster, A.P., N.M. Laird, and D.B. Rubin. 1977. Maximum likelihood from incomplete data via the EM algorithm (with discussion). J. Royal Stat. Soc. B 39: 1-38. Everitt, B. S. 1984. An introduction to latent variable models. Chapman and Hall. London. 107 p. Fleiss, J. L. 1981. Statistical methods for rates and proportions, 2"'' ed. John Wiley, New York, NY, 352 p Formann, A. K. 1994. Measurement errors in caries diagnosis: some further latent class models. Biometrics 50:865-871. 1996. Latent class analysis in medical research. Stat. Meth. Med. Res. 5:179-211. Hagen, P., K. Munk, B. Van Alen, and B. White. 1995. Thermal mark technology for inseason fisheries man- agement: a case study Alaska Fishery Res. Bull. 2:14.3- 158. Hui.S.L, and S.D.Walter 1980. Estimating the error rates of diagnostic tests. Bio- metrics 36:167-171. Joseph, L., T. Gyorkos, and L. Coupal. 1995. Bayesian estimation of disease prevalence and the parameters of diagnostic tests in the absence of a gold standard. Am. J. Epidemiol. 141:263-72. Kimura, D. K.. and S. Chikuni. 1987. Mixtures of empirical distributions: an iterative appli- cation of the age-length key. Biometrics 43:23-35. Ki-zanowski, W. J., and F. H. C. Marriott. 1995. Multivariate analysis, part 2: classification, cova- riance structures and repeated measurements. Arnold, London, 280 p. McCutcheon, A. L. 1987. Latent class analysis. Sage, Beverly Hills, CA, 96 p. McLachlan, G. J., and T. Ki'ishnan. 1997. The EM algorithm and extensions. John Wiley, New York, NY, 304 p. Microsoft Corporation. 1993. Microsoft Excel user's guide. Microsoft Corporation, Redmond, WA. Millar, R. B. 1987. Maximum likelihood estimation of mixed stock fish- ery composition. Can. J. Fish. Aquat. Sci. 44:583-590. Munk, K. M., W W. Smoker, D. R. Beard, and R. W. Mattson. 1993. A hatchery water-heating system and its application to 100'7f thermal marking of incubating salmon. Progi'es- sive Fish-Culturist 55:284-288. Fella, J., M. Masuda, and S. Nelson. 1996. Search algorithms for computing stock composition of a mixture from traits of individuals by maximum like- lihood. U.S. Dep. Commerce, NOAA Tech. Memo. NMFS- AFSC-61. Qu, Y., M. Tan, and M.H. Kutner 1996. Random effects models in latent class analysis for evaluating accuracy of diagnostic tests. Biometrics 52: 797-810. Qu.Y.andA. Hagdu. 1998. A model for evaluating sensitivity and specificity for correlated diagnostic tests in efficacy studies with an imperfect reference test. J. Am. Stat. Assoc. 93:920-928. Blick and Hagen: Use of agreement measLires and latent class models to assess the reliability of classifying tfiermally marked otolitfis 9 Richards, L. J., J, T. Schnute, A. R. Ki-onlund. and K. J. Beamish. 1992. Statistical models for the analysis of ageing error. Can. J. Fish. Aquat. Sci. 49:1801-1815. Regan, W. J., and B. Gladen. 1978. Estimating prevalence from the results of a screening test. Am. J. EpidemiologN' 107:71-76. SAS Institute. 1989. SAS/STAT user's guide, version 6, 4"' ed. SAS Insti- tute, Gary, NC. Statistical Sciences. 1995. S-PLUS guide to statistical and mathematical analy- sis, version 3.3. StatSci. Seattle, WA. Vacek, P. M. 1985. The effect of conditional dependence on the evalua- tion of diagnostic tests. Biometrics 41:959-968. Viana, M. A. G., V. Ramakrishnan, and P. S. Levy. 1993. Bayesian analysis of prevalence from the results of small screening samples. Commun. Statist. Theory Melh. 22:57,5-.585. Volk, E. C., S. L. Schroder, and K. L. PVesh. 1990. Inducement of unique otolith banding patterns as a practical means to mass-markjuvenile Pacific salmon. Am. Fish. Soc. Symp. 7:203-215. Walter, S. D. 1984. Measuring the reliability of clinical data: the case for using three observers. Rev. Epidem. et Sante Publ. 32:206-211. Walter, S. D., and L. M. Ii-wig. 1988. Estimation of test error rates, disease prevalence and relative risk from misclassified data: a review. J. Clin. Epidemiol. 41:923-937. Yang, I., and M. P. Becker 1997. Latent variable modeling of diagnostic accuracy. Bio- metrics 53:948-958. 10 Fishery Bulletin 100(1) Appendix The following SAS (version 6.12) code was used to estimate parameters in the three-reader model discussed above. This program makes use of iteratively reweighted least squares to maximize the likelihood function. Observed values (e.g. the number of HHH) are equated with the corresponding expected value from the model and a weighted least squares fit is computed by using PROC NLIN. This computation is iterated to convergence of the parameter estimates. Weights are inverses of the predicted values at each iteration. Indi- cator variables for each possible outcome are generated so that a model in typical regi'ession form can be written. Bounds on the parameter estimates may be needed to con- strain the estimates to the appropriate intervals. Note that the asymptotic standard errors provided by SAS will be correct if the option SIGSQ=1 is specified. However, the printed degrees of freedom and the associated confidence intei-vals are not correct for this application. The residual weighted sum of squares listed by SAS is the chi-squared goodness-of-fit-statistic. The option, OUTEST, outputs point estimates and the the estimated covariance matrix for the parameters. SAS code for the multistrata model used in the second example is also available from the authors. /* SAS Code for estimating 3-reader, 1 -stratum model 7 data a; array x{8} x1-x8; input y, ntot-i-y. /■ accumulating sample size V if n =8 ttien call symput('ntot'.ntot); /" put total into macro var 7 do 1=1 to 8: if l=_n _ ttien x{i}=1 ; else x{i)=0, /" set up indicator variables 7 end; cards, 406 13 1 1 6 2 6 135 /' H H H 7 /• H H W 7 /• H W H 7 /•WH H 7 /* H W W 7 /• W H W 7 /• W W H 7 /• W W W 7 proc nlin data=a nohalve sigsq=1 outest=esti /' sigsq=1 for correct se's 7 parms a1= 9 a2= 9 a3= 9 b1 = .9 b2= 9 b3= 9 p=,6; /" starting values 7 /* a IS accuracy for H 7 /■ b is accuracy for W 7 el =a1 •a2-a3-p-i-(1 -b1 )71 ■b2)-(1 ■b3)'(1 -p) e2=ara2*(1-a3)'p-i-(1-b1)*(1-b2)*b3*(1-p) e3=a1 •(! -a2)*a3*p+(1 -b1 )*b271 -b3)*(1 -p) e4=(1-a1)*a2'a3-p-i-br(1-b2)*(1-b3)*(1-p) e5=a1 -(1 -a2)-(1 -aSj-p-fll-bl )-b2'b3'(1 -p) e6=( 1 -al )'a2'(1 -a3)'p-fb1 '(1 -b2)*b3'(1 -p) e7=(1-a1)*(1-a2)*a3*p-i-brb2"(1-b3)*{1-p) e8=(1-a1)*(1-a2)'(1-a3)*p-i-b1'b2'b3*(1-p) model y=(e1■x1^-e2■x2-l-e3■x3-^e4■x4+e5*x5+e6■x6-^e7■x7-^e8■x8)■&ntot: bounds 5<=a1<=1, 0.5<=a2<=1, 0-5<=a3<=1, 0,5<=b1<=1, 0-5<=b2<=1 5<=b3<=1, 0<=p<=1: weigtit_=1/model y; run; Abstract— Distriliution. abundance, and .s('\rral [icipulation features were stud- ied in Ensenada de La Vela (Vene- zuela) between 1993 and 1998 as a first step in the assessment of local fisheries of swimming crabs. Arenaeua cribrarius was the most abundant spe- cies at the marine foreshore. Callinectes danae prevailed at the estuarine loca- tion. Callinectes hocourti was the most abundant species at the offshore. Abun- dances of A. cribrarius and C. danae fluctuated widely and randomly. Oviger- ous females were almost absent. Adults of several species were smaller than pre- viously reported. This study suggests that fisheries based on these swimming crabs probably will be restricted to an artisanal level because abundances appear too low to support industrial exploitation. Local distribution and abundance of swimming crabs (Callinectes spp. and Arenaeus cribrarius) on a tropical arid beach Carlos A. Carmona-Suarez Jesus E. Conde Centre de Ecologia, institute Venezolano de investigaciones Cientificas AP 21827 Caracas 1020-A, Venezuela E mail address (for C. A, Carmona-Suarez) ccarmona(a)oil 6 - n - Id Mean s e Range Dl ssolved oxygen Station 1 7 Bl 35 6-10.2 r- Station 2 8 14 46 6-12 _ Slot ion 3 7 02 43 4 1-95 p - Station 4 7 74 42 5-108 : fc i*,tV^ l**Y ~ 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 F M A M JJASONDJF I I 1993 M J J -1994 - a S N D Sampling months Figure 3 Surface temperature, salinity, and dis.solved oxygen at the foreshore and estuarine stations in Ensenada de La Vela (Falcon, Venezuela). at the marine stations was C. danae (Table 2). The highest number of species, six, was recorded at the estuarine site, where C. danae clearly dominated with a relative abun- dance of 75. 29f, followed by C. bocourti ( 14.1%), C. exaspera- tus (4.5%), C. sapidus (2.5%), C. maracaiboensis (2.0%) and C. laruatus (1.5%). Arenaeus cribrarius was absent from the estuarine site. Overall, the highest diversity (Shannon- Weaver index) was registered at the estuarine station, fol- lowed by station 2, station 1, and finally .station 4, the most exposed tract. Hills diversity number 1 (Nl), which indi- cates abundant species, was also highest at the estuarine station 3, followed by stations 2, 1, and 4 (Table 2). In a comparison of the two main biotopes (all three marine sta- tions vs. the estuarine station ) for the most frequent species (A. cribrarius and C. danae ), their abundance was dependent on salinity ( G=306; df=l, P<0.005 ). However, their abundance was independent of wave exposure, when only the stations in the marine biotope were considered (G=5.624; df=2, 0.05 >P>0.1). Offshore A total of 173 swimming crabs were caught witli crab pots. Abundance was highest at the seaward- most station, followed by the inshore and midshore sta- tions (Table 3). The average number of individuals per pot at each site followed a similar sequence (offshore: 5.9 individuals/pot; inshore; 1.75 ind/pot; midshore: 1.5 ind/pot). Differences in abundance between offshore and inshore and between offshore and midshore stations were 16 Fishery Bulletin 100(1) Table 2 Overall abundance (no of crabs) and diversity indexes for swmimmg crabs at seaside in Ensenada de La Vela (Venezuela). Station 1 Station 2 Station 3 Station 4 Totals C*! A. cribrarius 86 70 64 220 (46) C. danae 19 22 149 7 197 (41,2) C. bocourti 2 28 1 31 (6.5) C. maracaiboensis 4 4 (0.8) C. sapidus 6 1 5 1 13 (2.7) C. exasperalus 9 1 10 (2.1) C. larvatus 3 3 (0.6) Number of specimens 111 95 198 74 478 Number of species 3 4 6 5 Simpson (A') 0.6292 0.5928 0.5876 0.7542 Shannon-Weaver (W) 0.6580 0.6930 0.8660 0.5230 Hill's numbers Nl 1.9300 2.0000 2.3780 1.6870 N2 1.5894 1.6868 1.7018 1.3260 Table 3 Overall abundance (no. of crabs) and diversity indexes for | swimming crabs captured with crab pots in Ensenada de La Vela (Venezuela). Inshore Midshore Offshore Totals C. bocourti 30 30 69 129 C. maracaiboe/isia 2 4 13 19 C. danae 3 1 8 12 C. oniatus 6 3 2 11 C. sp. (unidentified! 2 2 Number of specimens 41 40 92 Number of species 4 5 4 Simpson (A') 0.5537 0.5705 0.5860 Shannon-Weaver (W) 0.8485 0.8823 0.7879 Hills numbers Nl 2.3361 2.4164 2.1987 N2 1.8062 1,7528 1.7065 significant, whereas the (difference between inshore and midshore was not. Overall, four species were caught, but C. bocourti prevailed at all the stations with at least 73.2'^f of the total quantity ( Table 3 ). Frequency of crabs by spe- cies varied significantly with the distance of the station to the shore (G=17.024. 0.05 >P>0.01. df=8). C. bocourti maintained a constant presence through the three sta- tions, ranging from 73.2 to 75. 09^ of total crabs at each station, the abundance of C. maracaiboensis increased sea- ward, and the abundance of C. cjrnatus decreased. Calli- nectes danae did not show any trend. In this set of samples, taken at a distance from the shoreline, the highest diversity (Shannon-Weaver index) was registered at the midshore station, closely followed by the inshore station and the offshore station. Hill's diver- sity number 1 (Nl), which indicates abundant species, was also highest at the midshore station, followed by inshore and offshore stations (Table 3). Temporal variability Because data for C. bocourti, C. sapidus, C. e.xasperatus, C. maracaiboensis and C. larvatus were too scarce to allow useful analysis, temporal variability of the abundance at the surf and at the estuarine pond was examined only for A. crihrxii-ius. C. danae, and for total crabs. Temporal variability of the abundances of these species are shown in Figure 4. Abundances fluctuated widely and randomly throughout our study. The density of A. cribrarius peaked in April, July, and October 1993, as well as in February and October 1994 (Fig. 4). In the estuarine site, C. danae abundance peaked in May and October 1993, as well as in February, April, August, and November 1994 (Fig. 4). In the marine sites, C. danae abundance was considerably lower and maxima occurred in June, September, and Octo- ber 1993, and in May and October 1994 (Fig. 4). No sig- nificant correlations were found between abundances of these two species and rainfall, water temperature, salin- ity, and dissolved oxygen (Table 4). However, when total crabs were regressed against rainfall at the estuarine site and oxygen at the foreshore, correlations were significant (Table 4). The negative correlation of this latter factor reached almost significant levels for both species at the marine ecotope. Diel variations Surf zone A total of 196 crabs were caught with hand seines at the foreshore during September 1997-February 1998 samplings: 82 crabs at night and 114 during the day (Table 5). Six species appeared in the diurnal samples (A. cribrarius, C. danae, C. bocourti, C. larvatus, C. mara- caiboensis and C. sapidus), one of which (C. sapidus) did Carmona Suarez and Conde: Distribution and abundance of Callinectes spp and Arenaeus cribrarius 17 Arenaeus cribrarius |/>=I56| T — I — I — I — I — I — I — I — I — I — I — I — V — I — I I I I r JFMAMJJ ASONOJFMAMJJASOND I 1993 II 19 94 1 Sampling months Figure 4 Abundance of Arenaeus cnbranus, Callineclcx danae, and total captured crabs in Ensenada de La Vela (Falcon. Venezuela). not appear at night. Arenaeus cribrarius was the domi- nant species followed by C. danae, during both diurnal and nocturnal samplings, whereas C. maracaiboensis, C. bocour-ti. C. larvatus, and C. sapidus were present in very low numbers. Guild composition did not differ significantly between day and night (G=1.630; 0.90>P>0.50; df=5), nor did the average number of individuals per trawl (0.86 vs. 0.62; <=1.702; 0.10>P>0.05: df=262 ). In both dominant spe- cies, A. crib>-ariu!i and C. danae. the average size of crabs caught during daylight hours (Table 5) did not differ sig- nificantly from those collected at night, nor did the size frequency distributions (G=3.820; df=4; 0.50>P>0.10). Sex ratios of these two species did not show significant diel differences either (G=0.030: 0.90>P>0.50; df=l; G=2.750; 0.50>P>0.10: df=l. respectively). Offshore A total of 64 crab pots were deployed. 32 during each period. Four species were caught during both day and night (C. hocourti, C. maracaiboensis, C. ornatus, and C. danae I. A total of 89 crabs were caught during the day and 84 at night (Table 6). Callinectes bocourti comprised 73.8'7( and 75. 3*/? of the abundance during the day and night, respectively, followed by C. maracaiboensis (15.5% and 6.7%). No differences in guild composition or sex ratios were found between day and night samples at each of the sites (Table 6). Crab species did not show differ- ences in carapace length between day and night captures (C. bocourti, ^=0.704, P>0.05, df=155; C. maracaiboensis, /=1.355, P>0.05, df=13; C. ornatus, ^=0.881, P>0.05, df=9; C. danae, ^=1,811. P>0.05, df=10). Kolmogorov-Smirnov tests run for normality of carapace length distribution for species at the offshore station compared between day and night samples, were statistically nonsignificant. Sex ratios and ovigerous females At the foreshore, all the species had male-biased overall sex ratios (Table 7), although only yi. cribrarius, C. danae 18 Fishery Bulletin 100(1) QJ t— * O o d 1 d A Q, A in o crv C Cfi C C/-J C c c CO d 01 he >, O c: lO lO lO in in O] c^ (N (M C^J +-» <» o lO CO Oi o m IN in in 00 ^ , lO to I— 1 CD CO 1 ■^ ^ o CO .—1 CO d d d d d 1 1 1 "^i o o c u II C ra CO o CO c« Cfi Cfi cfj C :C C c C c C O) 1 tn 3 3 W CJ -M ■c > CI, tie s CO CO CO CO CO CO CO CO CO CO X o Tf ■^ CO 00 ,— < T3 CO t^ 00 t^ CM 0; L. CO c^ c^ !N CO > o o (M o O d d d d d m Xfi -5 0^ -o o J2 C CO < tj cfi m cfi to cfi qJ yn C C c C C tx 3 1 >-. en Ie4 nipe a o "rt rf 'f ^ -* -* i2 *"- C/D " CO CO CO CO CO *J 'c ;r^ ^ 00 E> .-H ^ CO CO o o CO eg t/j L. CD ^ OJ to , 7 O o o o o CO d d d d d c: 1 1 c CO ^ CD -a c C CO CC o d CO Cfl c/: C A tt. CO c c/j OJ A o C 1 to o CO -13 [« M d C 3 _D 'ct3 CO p:^ lO lO in in in J2 CO "- oi O) (M tN CJ t- CJ C 02 c CO c CO c CO '5 'C 3 C 3 u 3 o o '11 o a CO CO o "cO e2 CO 2 Ci> 1 c/i 0.5; df=14). All Kolmogorov-Smirnov tests run for normality of carapace length distribution for species that were compared at the foreshore and estuarine sta- tions, were statistically nonsignificant. Discussion Of the nine species of Callinectes that have been reported for the tropical Western Atlantic (Williams, 1984), seven appeared at the foreshore of Ensenada de La Vela during our study. The species with the widest distributions were C. danae and C. sapiduf;. which were the only ones to appear at all the stations by the sea margin. Callinectes maracai- boensis and C. larvatus had the most restricted distribu- tion, occurring only in the estuarine site, and C. exasperatus was present only in the estuary and at one of the marine stations. At the marine foreshore stations, A. crihrarius was the dominant species, with a share of 78% of the total catch in this ecotope, whereas C. danae (19%) was the second most abundant species. Meanwhile, in the estuarine site, where A. cribrarius was absent, C. danae was the prevail- ing species, followed by C. bocourti. Overall, the highest diversity was registered at the estuarine station, whereas at the foreshore the highest diversity was recorded in the Carmona Suarez and Conde Distribution and abundance of Calllnectes spp and Arenaeus cribmrius 19 Table S Body size carapace le igth in mm ) and species abundance during diel observations at the foreshore of Ensenada de La Vela | (1997-98). Percentages are ?iv en in parentheses in "Abundance ' column. Species Period Abundance Mean body size SE A. cribrarius day 91 (79.8) 19.14 0.94 night 59 (72.0) 19.38 1.24 C. danae day night 17 (14.9) 16 (19.5) 21.92 23.09 1.89 2.45 C. hocourti day night 1 (0.9) 2 (2.4) 21.3 33.83 11.4 C. maracai boensis day night 2 (1.8) 4 (4.9) 47.45 37.28 9.10 6.05 C. larL'otus day night 2 (1.8) 1 (1.2) 30.55 15.65 2.25 C. sapid us day night 1 (0.9) 38.4 — Total day night 114 82 Table 6 Distribution of species abundance (no of crabs found) at the offshore ^ tations during diel samplings and comparisons of sex ratios | (all sites poo cdl. C. bncuurti C. maracaiboensis C. danae C. ornatus C sp.' Totals G(df=4) Significance Inshore night 12 2 1 1 16 5.238 0.50>P>0.10 day 18 2 5 25 Midshore night 18 3 2 2 25 4.226 0.50>P>0.10 day 12 1 1 1 15 Offshore night 32 8 1 2 43 8.506 0.10>P>0.05 day 37 5 7 49 Total night 62 13 2 5 2 84 7.940 0.10>P>0.05 day 67 6 10 6 89 Overall total 129 19 12 11 2 173 Sex ratios G Significance C. bocourti 0.01 0.975>P>0.9 C. maracaibo ensis 0.642 0.5>P>0.1 C. danae 0.07 0.9>P>0.5 C. ornatus 2.864 0.1>P>0.05 ' Unidentified species. most protected marine stations (2 and 1) followed by sta- tion 4, which is located at the most exposed tract. The values of Hill's diversity number 1 (Nl) demonstrated a similar pattern and indicated that the number of abundant species was close to two at stations 2 and 3. slightly above this value at the estuarine station, and below at the most exposed station. Offshore guild composition was substan- tially different from that at the sea margin, as shown by pot samplings. Although several species were common to the three biotopes, each habitat had a distinctive dominant species: Aiviiaeus cribrarius at the siu'f zone (stations 1, 2, and 4), C. danae (station 3) in the estuarine pond, and C. bocourti offshore. Because different sampling gears were used at the sea border and offshore because of practical 20 Fishery Bulletin 100(1) reasons, comparisons should be regarded as qualitative. However, artisanal fishermen do hai-vest C. bocoiirti when using beach seines in the areas next to our crab pot stations (senior author, personal obs. ). Inshore-offshore zonations of species at Ensenada de La Vela diverged from the gradients compiled by Norse and Fox-Norse ( 1979) for other areas in the Caribbean. In many localities, C. bocourti, C. sapidiis. and C. maracai- boeusis (the so-called bocourti group) are known to inhabit the waters by the seaside, whereas C. marginatus and C. ornatus are found at the seawardmost zone, and C. daiiae occupies the intermediate area. However, our patterns of Table 7 Sex ratios for portunids captured in Ensenada de La Vela ( 1993- -94). Male:female Ratio G df Significance Marine stations A. crihrarius 155:61 2.5:1 21.607 1 P<0.005 C. clanae 34:11 3.1:1 6.272 1 0.01>P>0.025 C. bocourti 0:1 0:1 — — — C. sapuliis 7:0 7:0 5.232 1 0.01>P>0.025 C. exasjicratiis 1:0 1:0 — — — Estuarine station C. daiiae 79 63 1.3:1 0.900 1 0.5>P>0.1 C. bocourti 13 15 0.9:1 0.070 1 0.9>P>0.5 C. sapidus 2 3 0.7:1 0.088 1 0.9>P> 0.5 C. exaspei-atus 7 2 3.5:1 1.405 1 0.5>P>0.1 C. larvatus 2 1 2:1 — — — C. maracaiboenais 2 2 1:1 — — — Table 8 Carapace length (mm) for the most abundant species at the foreshore and estuarine stat and comparison of carapace sizes, ns = not significant. ions in Ensenada de La Vela (Venezuela), Calliiiectff: daiiac Marine stations Callinectcs danae Estua rine station n Mean Range BE n Mean Range SE Juvenile females Adult females Juvenile males Adult males 8 3 14 20 24.6 39.6 19.5 .39.1 14.92-32.7 36..58-42.0 8.5-32,4 7.42-56.7 2.099 1.619 2 2.869 Juvenile females Adult females Juvenile males Adult males 37 26 43 36 22.6 39.8 15.8 23.7 11.28-35.6 31.45-47.4 7.62-27.4 10.4-48.4 1.069 0.832 0.691 1.897 Arcnacus cribranus Marine stations only Callinectcs bocourti Estuarine station only n Mean Range BE n Mean Range SE Juvenile females Adult females Juvenile males Adult males 61 22 - . Nn 11.3-36.94 0.831 Juvenile females Adult females Juvenile males Adult males 7 10 4 6 23.1 41 19.6 41.8 11.8-34.4 34-45.1 16.6-23 24.4-56.5 3.251 1.126 1.314 5.035 89 66 17.6 21.8 10.25-28.64 9.48-56.55 0.459 1.213 / df Significance C. danae (all crabs) C. danac (adult females) C. danac (adult males) C. danae (foreshore stations) C. danac (estuarine station) foreshore/estuarine foreshore/estuarine foreshore/estuarine females/males females/males 3.799 0.065 4.653 0.766 6.297 185 27 54 43 140 P<0.05 ns P<0.001 ns P<0.001 Caimona Suarez and Conde Distribution and abundance of Callinecles spp and Arenaeus aibraiius 21 abundance for Callincctea species are similar to another Caribbean locality (Buchanan and Sloner. 1988): Laguna Joyuda (Puerto Rico). All the Callinectcs spp. recorded in this coastal estuarine lagoon were also present in the es- tuarine station of Ensenada de La Vela. Callinectcs danae was the dominant species in both sites, whereas C e.v- asperatus and C. larvatus were present in low numbers. Callinecles maracaiboensis was very scarce at Ensenada de La Vela, but it was not reported at all in Lagiina Joyu- da (Buchanan and Stoner, 1988), although Buchanan and Stoner cautioned that specimens of this species might have been misclassified and listed as C. bocoiirti. On the other hand, the high abundance of A. cribrarius at the ma- rine front of Ensenada de La Vela differed from that of other studies in the Caribbean and Gulf of Mexico, where this species has been reported in low numbers. For in- stance, in the SW Gulf of Mexico A. cribrarius was less than 1% of the total portunid community (Garcia-Montes et al., 1988). In Laguna de Terminos (Mexico), a polyha- line coastal lagoon, four species of Callinecles were found in a population sui-\ey conducted during a whole year, but no individuals of Arenaeus were reported (Roman-Contre- ras. 1986). In the same lagoon, Sanchez and Raz-Guzman (1997) caught a single individual of A. cribrarius out of 986 specimens collected over a 17-year span. The differ- ences probably are probably due to the polyhaline con- ditions at these settings, thus restricting the viability of A. cribrarius. However, in temperate sandy beaches, this species can be very common. On Bogue Banks, in North Carolina, Arenoeus cribrarius ranked as the most impor- tant brachyuran in a high-wave-energy sandy beach ( Leb- er, 1982). In the surf zone at Folly Beach, South Carolina, A. cribrarius was one of the dominant brachyuran crabs during the summer and also a key predator of benthic or- ganisms (DeLancey, 1989). A. cribrarius is considered to be well-adapted to marine and slightly hypersaline salinity regimes and to habitats with heavy surf and sand scouring in shallow coastal waters (Fischer, 1978; Williams, 1984). This fact was evident in our study, in which A. cribrarius was abundant and clearly constrained to a narrow strip in the surf zone. Our results suggest the importance of salinity as an ex- eluding axis in the distribution of some species of swim- ming crabs in the surf and estuarine pond of Ensenada de La Vela. In our study, A. ciibrarius was present in salini- ties from SOVcc to 43" i. thus exceeding the upper limits of tolerance commonly reported for this species. The restrict- ed distribution of this species is probably a consequence of its stenohalinity (27.5-36.5"( i (Gunter, 1950; Norse, 1978; Williams, 1984; Pinheiro, 1991; Avila and Branco, 1996), although very occasionally it may show up in estuaries (Williams, 1965) and can tolerate experimental salinities down to 17.25'w (Norse, 1978). This range indicates that A. cribrarius prefers marine or near-marine environments, thus explaining its absence in station 3 (estuarine). In spite of being considered to be well adapted to heavy surf in shallow coastal waters (Fischer, 1978; Williams, 1984), A. cribrarius appeared to be abundant in all three fore- shore stations, independent of water movement, and was most abundant in the more protected stations 1 and 2. Be- cause the salinity did not show any major differences be- tween foreshore and offshore habitats, other factors are at stake in determining the zonation obsei-ved for the oth- er species. One of the main elements to consider is sub- strate composition (Norse and Fox-Norse, 1979; Pinheiro et al, 1997). Pinheiro et al. (1997) stated that distribu- tional patterns of portunids in Fortaleza Bay (Brazil) are driven mainly by the granulometric composition of the sediments. Substrates at the foreshore and estuarine pond differed from offshore bottoms: at the foreshore the sedi- ment was mainly sand; at the estuarine station a muddy bottom prevailed. At the offshore stations, silt was the main substrate. Hence, this difference could influence the distribution of swimming crabs in Ensenada de La Vela. Callinecles danae was found in both biotopes at the fore- shore but was more abundant at the estuarine site. In the marine stations of the surf zone, C. danae appeared more frequently in the most protected areas. The appearance and persistence of this species in both environments probably stems from its euryhalinity. In several Caribbean locations, C. danae has been obsei-ved dwelling in polyhaline environ- ments (Taissoun, 1969; Norse, 1978; Buchanan and Stoner, 1988). Based on this evidence, it is not surprising to find C. danae in the entire range of salinities in Ensenada de La Vela, although it is important to underline that at the estuarine station it appeared when salinity was below the minimum (ll%o) reported by Norse (1978). Also, several of the portunid species in the surf zone in Ensenada de La Vela were found in higher salinities than those reported by Norse ( 1978) in several localities in Jamaica, except C. ma- racaiboensis and C. larvatus. The absence of C. ornatus at the foreshore stations may be due to reasons other than the sampling method, because the same method was used by Carmona-Suarez and Conde (1996), where specimens of C ornatus were frequently captured at different sites in the State of Falcon, Venezuela, including Ensenada de La Vela. Total abundance of all swimming crabs both at the surf zone and at the estuarine station fluctuated widely and randomly through the year. This pattern also emerged when only the temporal abundance variations of the domi- nant species, A. cribr-arius and C. danae, were examined. No significant correlations were found between abundances of these two species and rainfall, dissolved oxygen, water temperature, or salinity fluctuations. However, the inverse correlation of dissolved oxygen and abundance reached al- most significant levels for both species at the marine fore- shore and indeed was significant for the total abundance of crabs in the surf Additionally, there was a positive cor- relation between rainfall and total abundance of crabs in the estuarine zone, possibly due to the increment of organ- ic material washed into this environment from adjacent terrestrial areas. Although bibliogi-aphic evidence supports the adaptation of portunids to low levels of dissolved oxy- gen in their environment (DeFur et al., 1990; Rantin et al., 1996; Manguni, 1997), and the relation between respira- tion rates and salinity in two Callinecles species (Rosas et al., 1989), nothing supports the idea that the increase of swimming crab densities is due to the decrease in dissolved oxygen. It might be possible that augmenting food resourc- es would increase populations of fishes and invertebrates 22 Fishery Bulletin 100(1) or planktonic blooms, which in turn would require a higher oxygen demand in the area, subsequently provoking drops in oxygen and causing mortahties of high-oxygen-demand- ing invertebrates. In any event, these results suggest that fluctuations in oxygen levels might be a key element in regulating portunid populations at Ensenada de La Vela and merit further research efforts. Berried females were remarkably scarce during our study. Only two, both belonging to C danae, were caught throughout the first period at the estuarine site, and none were caught during day and night samplings in the surf zone nor offshore. Nonetheless, scarcity of ovigerous fe- males of swimming crabs in these coasts is not exceptional. During a 2-year survey of crustaceans along 700 km of Fal- con's shoreline, Carmona-Suarez and Conde (19961 caught very few berried females of several of the littoral portunid species. They caught only one berried female of A. cribrari- us and no ovigerous females of C. sapidus, C. larvatiis, C. or- natus, or C. danae. However, in estuarine areas, substantial numbers of berried females of C. bocourti. C. maracaihoen- sis, and C. exasperatus were caught in the tidal zone. The scarcity or sheer absence of egg-bearing females in some species of swimming crabs might be the result of habitat partitioning by sex. Differential distributions by sex have been reported for C. sapidus (Williams, 1965; Perry, 1975; Archambault et al, 1990), C. maracaiboensis (Norse, 1977), and C. bocourti (Taissoun, 1969; Norse, 1978). However, for the dominant species in the surf, A. cribrarius. ovigerous females do not seem to be segregated into deeper waters. In southern Brazil, ovigerous females of this species appeared in shallow waters close to the coast (Pinheiro et al., 1996). Similarly, many ovigerous females were collected in very shallow water, at the sui-fs edge in North Carolina (Wil- liams, 1984). Likewise, for C. bocourti and C. maracaiboen- sis egg-bearing females have also been reported in marine shallow waters (Norse, 1977). Furthermore, adult females of most species inhabiting the surf zone at Ensenada de La Vela were observed in this area year-round. Thus, alter- native explanations should be considered, such as lack of estuarine habitats or sustained harsh environmental con- ditions that do not allow energy to be invested in reproduc- tion. For instance, a highly seasonal reproductive pattern, with periods without berried females, has been observed in populations of the mangrove tree crab, Aratus pisonii, liv- ing in hypersaline lagoons in this area (Conde, 1989); this pattern contrasts with the pattern for populations inhabit- ing other localities, where these crabs reproduce continu- ously throughout the year (Conde and Diaz, 1989a; Diaz and Conde, 1989). Also, undergi-own or stunted specimens of various species of crustaceans have been reported in this area (Conde and Diaz 1989b, 1992a, 1992b; Carmona, 1992; Carmona-Suarez and Conde, 1996). Thus, it is possible that this arid coast lacks the necessary resources for these crabs to reproduce, except in a few estuarine spots. This hypothe- sis is also supported by the fact that the body size of several species of swimming crabs collected in our study was small- er than that reported in other locations (Fischer, 1978; Wil- liams, 1984). The only river near the Ensenada de La Vela is the Coro River, which lies approximately 2 km westwards. Because of current direction (east-west), it cannot influence estua- rine conditions to the sampled area. The small estuary in the Ensenada de la Vela could be a possible local nutrient supplier But its influence is restricted to a few days dur- ing the end of each year, when the estuary opens to the sea. The setup of an untreated sewage discharge in the small estuarine basin at the beginning of 1997 could in fact have a long-term impact, but it is possible that vari- ous species of swimming crabs may not be affected nega- tively, because of their capacity to live in polluted areas. Such is the case with C. bocourti (Taissoun, 1972; Wil- liams, 1974), and C, sapidus, the main species in the Lake Maracaibo crab industry (Oesterling and Petrocci, 1995), where contamination due to several sources (i.e. sewage and oil) has reached high levels (Rodriguez, 2000). Because trawl studies have shown greater abundances of blue crabs (C. sapidus) and, in general, other decapods at night (Wilson et al., 1990), we ran a series of day and night samplings at the marine front over a six-month pe- riod and later also undertook diel offshore pot sampling on several occasions. Although Fischer ( 1978) has stated that A. cribrarius burrows into the bottom during the day and emerges at night, we collected A. cribrarius in the same range of abundances and sizes during both day and night samplings. Similar results were achieved by DeLancey ( 1989) in South Carolina, where no significant differences were obtained from samples collected at day and night. Wilson et al. ( 1990) ascribed the lack of differences in day and night abundances of C. sapidus to the use of more effective sampling devices than previously employed. In our study, no major differences were obsei'ved in diversity, abundance, body size, or sex ratios for most species, even though two kinds of collecting gears were used; thus, it is feasible to consider that if daily cycles exist in the spe- cies, they do not have a significant impact on daily density variations. In turn, these findings may have practical con- sequences for the decisions regarding sampling schemes to assess fisheries in this area. The exploitation of swimming crabs at Ensenada de La Vela must be considered only at the artisanal level be- cause of the low abundance of all species treated in our work and their wide and random density fluctuations. In fact, local fisheries are currently limited to the arti- sanal capture of portunids by hanging nets or hand-driven trawling nets. The most captured species by fishermen is C. bocourti (senior author, personal obs.), but C. danae is also a promising staple to be hai-vested because it appears in all three major biotopes (marine inshore, offshore, and estuarine). Arenaeus cribrarius, a species commercially exploited in Brazil (Pinheiro and Franzoso, 1998) and re- garded to have an excellent fiavor (Fischer, 1978), may al- so be considered a target species because of its great abun- dance, although its small size may make it less desirable commerciallv. Acknowledgments We heartily thank Sebastian Trompiz, whose involvement in most of the phases of the project was instrumental to Carmona Suarez and Conde Distribution and abundance of Calllnectes spp and Arenaeus cribrarius 23 its completion. We thank Oniegar Cespcdes, Angel Lopez, and (Jregorio (lotopo for occasional assistance during field work. Our gratitude extends to Maria Rondon Medicci, and Nicanor Cifuentes for their critical reading of the manuscript and for assistance during field work in August 1998, and to Eloy Conde, Eloina de Conde and Enrique Molina for logistical support. Last but not least, we thank Kate Rodriguez-Clark for her help with the English text. This study was supported in part by grants CTI 90-068 from the UNEFM and BTA-08 from CONICIT-BID. This work was partially undertaken while C. Carmona-Suarez was a staff member of the Centre do Investigaciones Mari- nas (UNEFM I. Literature cited Ai-chanihault. J. A., E. L. Wenner. and .J. D Whitaker. 1990. Life history and abundance of blue crab, Calliiwctefi Kapidus Rathbun, at Charleston Harbor, South Carolina. Bull. Mar Sci.46:145-1.'58. Arnold, W. S. 1984. The effects of prey size, predator size, and sediment composition on the rate of predation of the blue crab, Calll- nectes sapidus Rathbun, on the hard clam, Mercenaria mer- cenaria (Linne). J. Exp. Mar Biol. Ecol. 80:207-219. Avila, M. G., and J. O. Branco. 1996. Aspectos bioecologicos dc Arenaeus cnbrnniis (La- marck) iDocapoda: Portunidael de Praia de Barra da Lagoa, Florianopolis, Santa Catarina, Brasil. Rev. Bras. Zool. 13; 165-174. Buchanan, B. A., and A. W. Stoner 1988. Distributional patterns of blue crabs iCaHinectes sp.) in a tropical estuarine lagoon. Estuaries 11:231-239. Carmona, C. 1992. Interpopulation size variation in the spider crab, Microphrys bieornutus Latreille 1824 (Crustacea, Brachy- ura, Majidael. Crustaccana 63:319-322. Carmona, C, and J. E. Conde. 1989. Caracterizacion de las costas del estado Falcon. Bol. Inst. Oceanog. Venezuela 28:127-133. Carmona-Suarez, C, and J. E. Conde. 1996. Littoral brachyuran crabs (Crustacea: Decapoda) from Falcon, Venezuela, with biogeogi-aphical and ecological remarks. Rev. Bras. Biol. .56:72r3-747. Conde, J. E. 1989. Ecologia poblacional del cangj-ejo de mangle Aratiis pisonii (H. Milne-Edwards) (Brachyura: Grapsidae) en habitats extremes. Ph.Sc. diss.. Institute Venezolano de Investigaciones Cientificas, Caracas, xxxvi -i- 212 p. Conde. J. E., and H. Diaz. 1989a. The mangi"ove tree crab Ara/!/.s p/.s-o/;;/ in a tropical estuarine coastal lagoon. Est. Coast. Shelf Sci. 28:639- 650. 1989b. Productividad del habitat e historias de vida del can- grejo de mangle Aratus pisonii ( H. Milne-Edwards I ( Brach- yura: Grapsidae). Bol. Inst. Oceanog. Venezuela 28:113- 120. 1992a. Variations in intraspecific relative size at onset of maturity (RSOM) in Aratus pisonii (H. Milne Edwards, 1837) (Decapoda, Brachyura, Grapsidae). Crustaceana 62:214-216. 1992b. Extension of the stunting range in ovigerous fe- males of the mangrove crab Aratus pisonii (H. Milne Edwards,1837) (Decapoda, Brachyura, Grapsidae). Crus- taceana 62:319-322. Conde, J. E., and G. Rodriguez. 1999. Integrated coastal zone management in Venezuela: the Maracaibo system. In Integrated coastal zone man- agement: principles and practice (W. Salomons, K. Turner, A. Spoudiolus, L. D. Lacerda, and R. Ramachandram, eds.). Springer- Verlag, Berlin, xviii + 386 p. DeFur, P. L., C. P. Mangum, and J. E. Reese. 1990. Respiratory responses of the blue crab Callinectes sapidus to lonfj-term hypoxia. Biol. Bull. 178:46-54. DeLancey, L. B. 1989. Trophic relationship in the surf zone during the sum- mer at Folly Beach, South Carolina. J. Coast. Res. 5:477- 488. Diaz, H., and J. E. Conde. 1989. Population dynamics and life history of the mangrove crab Aratus pisonii (Brachyura: Grapsidae) in a marine environment. Bull. Mar Sci. 45:148-163. Ferrer-Montano, O. J. 1997. Effectiveness of two pots and other factors for har- vesting hard blue crabs Callinectes sapidus in Lake Mara- caibo, Venezuela. Ciencia, Univ. Zulia 5:111-118. Fischer, W.(ed.) 1978. FAO species identification sheets for fishery purposes. Western Central Atlantic (fishing area 31), vol. VI.. 278 p. FAO, Rome. Garcia-Montes, J. F, L. A. Soto, and A. Gracia. 1988. Cangrejos portunidos del suroeste del golfo de Mexico: Aspectos pesqueros y ecologicos. An. Inst. Cien. Mar y Limnol. UNAM (Universidad Nacional Autonoma de Mexico) 15:135-150. Goddard, D., and X. Picard. 1976. Geomorfologia y sedimentacion en la costa del estado Falcon, cabo San Roman a Chichiriviche. Mem. II Cong. Latin. Geol.,Tomo II. Bol. Geol., Caracas 7:1157-1180. Gunter, G. 1950. Seasonal population changes and distributions as related to salinity, of certain invertebrates of the Texas coast, including commercial shrimp. Publications of the Institute of Marine Science, Univ. Texas 1:7-51. Hsueh, R W., J. B. McClintock, and T S. Hopkins. 1992. Comparative study of the diets of the blue crabs Cal- linectes similis and C. sapidus from a mud-bottom habitat in Mobile Bay Alabama. J. Crust. Biol. 12:615-619. Lahey J. F 1973. On the origin of the dry climate in northern South America and the southern Caribbean. In Coastal deserts: their natural and human environments (D. H. Amiran and A. W. Wilson, eds. ), p. 7.5-90. Univ. Ai'izona Press, Tucson, AZ. Leber, K. M. 1982. Seasonality of macroinvertebrates on a temperate, high wave energy sandy beach. Bull. Mar Sci. 32:86-98. Lin, J. 1991. Predator-prey interactions between blue crabs and ribbed mussels living in clumps. Estuarine Coast. Shelf Sci. 32:61-69. Ludwig, J. A., and J. F Reynolds. 1988. Statistical ecology. Wiley New York, NY, 337 p. Mangum, C. P. 1997. Adaptation of the oxygen transport system to hy- poxia in the blue crab, Callinectes sapidus. Am. Zool. 37:604-611. Moncada, F G., and O. Gomez. 1980. Algunos aspectos bioldgicos de tres especies del gen- 24 Fishery Bulletin 100(1) ero CciUinectey: ( Crustacea, Decapoda ). Rev, Cubana Invest. Pesq, .5:1-3.5, Norse. E. A, 1977. Aspects of the zoogeographic distribution ofCallinectes (Brachyura: Portunidaei. Bull. Mar Sci. 27:440-447. 1978. An experimental gradient analysis: hyposalinity as an "upstress" distributional detei-niinant for Caribbean portu- nid crabs. Biol. Bull. 155:.586-598. Norse, E. A., and M. Estevez. 1977. Studies on portunid crabs from the Eastern Pacific. I. Zonation along environmental stress gradients from the coast of Colombia. Mar Biol. 40:365-373. Norse, E. A., and V, Fox-Norse. 1979. Geographical ecology and evolutionary relationships in Callinectes spp. (Brachyura: Portunidaei. Proceedings of the blue crab colloquium. Gulf States Mar. Fish. Comm. 7 (19821:1-9. Oesterling, M. J. 1984. Manual for handling and shedding blue crabs {Calli- nectes sapidiif: >. Special Report in Applied Marine Science and Oceanography 271, Virginia Inst. Mar Sci., 76 p. Oesterling, M. J., and C. Petrocci. 1995. The crab industry in Venezuela, Ecuador and Mexico. Virginia Sea Grant Resource Advisory 56, 32 p. Orth, R. J., and J. van Montfrans. 1987. Utilization of a seagi-ass meadow and tidal marsh creek by blue crabs Callinectes sapicliis, 2: Seasonal and annual variations in abundance with emphasis on post-set- tlement juveniles. Mar Ecol. Prog. Ser. 41:283-294. Perry, H. 1975. The blue crab fishery in Mississippi. Gulf Res. Rep. 5:39-57. Perry, H. M., and. W. A. Van Engel (eds.). 1979. Proceedings of the blue crab colloquium. Gulf States Mar. Fish. Comm. 7 ( 19821:1-251. Pinheiro, M. A. A. 1991. Distribu(;ao e Biolugia Populacional de Arcnaeus crihranus (Lamarck, 1818) (Crustacea, Brachyura, Portun- idaei, na Ensenada da Fortaleza, Ubatuba, ,SP. Tesis de Maestria en Zoologia. Universidade Estadual Paulista, Botucatu, Brasil, 175 p. Pinheiro, M. A. A., and A. Fransozo. 1993. Relative gi-owth of the speckled crab A/-c()oe;;,s cribrar- ;'us (Lamarck, 1818 1 ( Brachyura, Portunidae I, near Ubatuba, State of Sao Paulo, Brazil. Crustaceana 65:377-389. 1998. Sexual maturity of the speckled swimming crab Are- naeus crihrarius (Lamarck. 18181 (Decapoda, Brachyura, Portunidaei, in the Ubatuba littoral, Sao Paulo State, Brazil. Crustaceana 71:434-452. 1999. Reproductive behavior of the swimming crab Are- naeus crihranus (Lamarck, 18181 (Crustacea, Brachyura, Portunidaei in captivity. Bull. Mar. Sci. 64:243-253. Pinheiro, M. A. A., A. Fransozo, and M. L. Negi-eiros-Fransozo. 1996. Distribution patterns of Arenaeus crihrarius (Lam- arck, 18181 (Crustacea, Portunidae) in Fortaleza Bay, Uba- tuba (SP), Brazil. Rev. Bras. Zool. 56:705-716. 1997. Dimensionamento c sobreposigao de nichos dos por- tunideos (Decapoda, Brachyura), na Enseada da Fortaleza, Ubatuba, Sao Paulo, Brasil. Rev Bras. Zool. 14:371-378. Prager, M. H, J. R. McConaugha, C. M. Jones, and P. J. Geer. 1990, Fecundity of blue crab, Callinectes sapidns, in Chesa- peake Bay: biological, statistical and management consid- erations. Bull. Mar. Sci. 46:170-179. Rantin, F T, A. L. Kalinin, and J. C. de Freitas. 1996. Cardio-respiratory function of swimming blue crab Callinectes danae Smith, during normoxia and graded hypoxia. J. Exp. Mar Biol. Ecol. 198:1-10. Rodriguez, G. 1980. Crustaceos decapodos de Venezuela. Instituto Vene- zolano de Investigaciones Cientificas, Caracas, Venezuela, 444 p. 2000. El manejo de los recursos naturales del sistema de Maracaibo. Chapter 7 in El sistema de Maracaibo (G. Rodriguez, ed.), p. 991-109. Instituto Venezolano de Investigaciones Cientificas, Caracas, Venezuela. Roman-Contreras, R. 1986. Analisis de la poblacion de Callinectes spp. ( Decapoda: Portunidae) en el sector occidental de la lagunadeTerminos, Campeche, Mexico. An. Inst. Cien, Mar. y Limnol. UNAM (Universidad Nacional Autonoma de Mexico) 13:315-322. Rosas, C, G. Barrera, and E. Lazarc-Chavez. 1989. Efecto de las variaciones de la salinidad y de la temperatura estacional sobre el consume de oxigeno de Callinectes rathbunae, Contreras y Callinectes similis (Crustacea: Portunidaei. Trop. Ecol. 30:193-204. Ryer, C. H., J. van Montfrans, and R. J. Orth. 1990. Utilization of a seagrass meadow and tidal marsh creek by blue crabs Callinectes sapidus. II. Spatial and temporal patterns of moulting. Bull. Mar Sci. 46:95-104. Sanchez, A. J., and A. Raz-Guzman. 1997. Distribution patterns of tropical estuarine brachyuran crabs in the Gulf of Mexico. J. Crust. Biol. 17:609-620. .Scelzo, M. A., and R. Varela. 1988. Crustaceos decapodos litorales de la isla de la Blan- quilla, Venezuela. Mem. Soc. Ven Cien. Nat. 47:33-54. Schubart, C. D., J. E. Conde, C. A. Carmona-Suarez, R. Robles, and D. L. Felder. 2001. Lack of divergence between 16S mtDNA sequences of the swimming crabs Callinectes hocourti and C. mara- caiboensis ( Brachyura: Portunidae I from Venezuela. Fish. Bull. 99:475-481. Sholar, T M. 1979. Blue crab fisheries of the Atlantic Coast. Proceedings of the blue crab colloquium. Gulf States Mar. Fish. Comm. 7(19821:111-127. Smith, D. E., R. J. Orth, and .J. R. McConaugha (conference steering committee). 1990. Proceedings of the blue crab conference held in Virginia Beach, Virginia. May 1.5-17, 1988. Bull Mar. Sci. 46:1-251. Sokal, R, R.,andFJ. Rohlf 1995, Biometry. 3''<' ed. Freeman, New York, NY, 887 p. StatSoft. 1992. Statistica/Mac. StatSoft, Tulsa, Oklahoma. Stuck, K. C, and F M. Truesdale. 1988. Larval development of the speckled swimming crab, Arcnaeus crihranus (Decapoda: Brachyura: Portunidae) reared in the laboratory Bull. Mar Sci. 42:101-132, Taissoun, E. 1969. Las especies de cangrejos del genero Callinectes (Brachyura) en el golfo de Venezuela y lago de Maracaibo. Bol. Centro Invest. Biol., Univ Zulia 2:1-103. 1972. Estudio comparative, taxonomico y ecologico entre los cangrejos (Decapoda: Brachyura: Portunidae), Calli- nectes maracaiboensis (nueva especie), C. hocourti (A. Milne Edwards) y C. rathhunae (ContrerasI en el golfo de Ven- ezuela, lago de Maracaibo y golfo de Mexico. Bol. Centro Invest. Biol., LIniv Zulia 6:7-46. 1973a. Biogeogi'afia y ecologia de los cangrejos de la familia "Portunidae" (Crustaceos Decapodos Brachyura) en la costa atlanticade America. Bol. Centro Invest. Biol., LIniv Zulia 7:7-23. Carmona Suarez and Conde: Distribution and abundance of Callinectes spp and Arenaeus cribrarius 25 1973b. Los canffic'jos do hi famili;i Portunidac (Crustaceos Docapodos Brachyura) en i-l Otcidente de Venezuela. Bol. Centro Invest. Biol., Univ Zulia 8:1-78. van Montfrans, J., J. Capelli. R. J. Orth, and C. H. Ryer. 1986. Use of microwiro tags for tagging juvenile blue crab.s {Callinectes sapidi/t^ Rathbunl. J. Crust. Biol. 6:370-376. Warner, G. F. 1977. The biology of crabs. P^lck Science. London, 202 p. Williams, A. B. 196."). Marine decapod crustaceans of the Carolinas. Fish. Bull. 65:1-298. 1974. The swimming crabs of the genus Callinectes (Dccap- oda: Portunidaei. Fish. Bull. 72:685-798. 1984. Shrimps, lobsters, and crabs of the Atlantic Coast of the Ea.stern United States, Maine to Florida. Smithson- ian Institution Press, Washington D.C., 550 p. Wilson. K. A., K. L. Heck Jr., and K. W. Able. 1987. Juvenile blue crab, Callinectes scipidi/s, sui-vival: an evaluation of celgrass, Zostera marina, as refuge. Fish. Bull. 85:5.3-.58. Wilson, K. A.. K Q. Able, and K. L. Heck Jr 1990. Habitat use by juvenile blue crabs: a comparison among habitats in southern New Jersey. Bull. Mar Sci. 46:105-114. 26 Abstract— We examined 536 permit {Tiachinotus fatcatus. 65-916 mm FL) collected from the waters of Florida Keys and from the Tampa Bay area on Florida's Gulf coast to describe their growth and reproduction. Among permit that we sexed, females ranged from 266 to 916 mm in length (mean=617) and males ranged from 274 to 855 mm (mean=601). Ages of 297 permit ranging from 102 to 900 mm FL were estimated from thin-sectioned otoliths (sagittae). The large proportion of oto- liths with an annulus on the margin and an otolith from an OTC-injected fish suggested that a single annulus was formed each year during late spring or early summer Permit reach a maximum age of at least 23 years. Permit gi-ew rap- idly until an age of about five years, and then growth slowed considerably. Male and female von Bertalanffy growth models were not significantly differ- ent, and the sexes-combined growth model was FL=753.1(l-e-" ■'■•»' •^«>'"^s'^' I. Gonad development was seasonal, and spawning occurred during late spring and summer over artificial and natural reefs at depths of 10-30 m. Ovaries that contained oocytes in the final stages of oocyte maturation or postovulatory fol- licles were found during May-July. We estimated that SO'J'r of the females in the population had reached sexual maturity by 547 mm and an age of 3.1 years and that 50% of the males in the population had reached sexual maturity by 486 mm and an age of 2.3 years. Because Florida regulations restrict the maximum size of permit caught in recreational and com- mercial fisheries to 20-inch (508-mml, most fish harvested are sexually imma- ture. With the current size selectivity of the fishery, the spawning stock bio- mass of permit could decrease quickly in response to moderate levels of fish- ing mortality; thus, the regulations in place in Florida to restrict harvest levels appear to be justified. Age, growth, and reproduction of permit (Trachinotus falcatus) in Florida waters Roy E. Crabtree Peter B. Hood Derke Snodgrass Florida Marine Research Institute Florida Fish and Wildlife Consen/ation Commission 100 Eighth Avenue SE St Petersburg, Florida 33701 5095 Present address (for R, E Crabtree) Division of Marine Fisheries Florida Fish and Wildlife Conservation Commission 620 Meridian St Tallahassee, Florida 32399 1600 E mail address (for R E Crabtree) crabtrno'gfc stale fl us Manuscript accepted 19 July 2001. Fish. Bull. 100:26-34 (2002). The family Carangidae supports a di- verse array of economically important fi.sheries in tropical and subtropical waters worldwide. In the southeastern United States, many carangid stocks are managed at both the state and Fed- eral level. Recently, the National Marine Fisheries Service determined that the Gulf of Mexico greater amberjack stock is overfished, but the status of most carangid stocks is unknown (Anony- mous'). For most carangid stocks, no quantitative stock assessments have been completed, in part, because little biological information exists regarding carangid growth rates and reproduc- tion. As a result, the adequacy of cur- rent management measures to prevent overfishing of many carangid stocks is unclear. Permit, Trachinotus falcatus, are the basis of an important recreational fish- ery and a small commercial fishery in Florida. Estimates of Florida recreation- al landings are unreliable but may ex- ceed 100,000 fish per year (Armstrong et al.'-). Commercial landings of permit peaked in 1991 at 200,000 pounds and then decreased to 50,000 pounds in 1995 (Aj-mstrong et al.-). Current reg- ulations in Florida include a 10-inch (254-mm FL) minimum size limit and a 20-inch (508-mm FL) maximum size limit on both the recreational and com- mercial harvest. In addition, recreation- al anglers are permitted daily to take 10 permit per bag of combined permit and Florida pompano {Trachinotus car- olinus). Many anglers pursuing permit do so with professional guides on a char- ter vessel. In addition to being popular in South Florida, permit are targeted by numerous fishing tourists and recre- ational anglers in the Bahamas and at locations throughout the Caribbean. De- spite the economic importance of permit in these regions, there are no published reports describing gi-owth, longevity, or length and age at sexual maturity. Such information is needed to evaluate the ef- fects of fishing mortality on permit pop- ulations. Previous studies of permit life history by Fields (1962) and Finucane ( 1969) were based only on an examina- tion of larvae and young-of-the year per- mit. Our study describes growth, lon- gevity, and the length and age at which fish become sexually mature. In addi- tion, we document spawning of permit in South Florida waters based upon a histological examination of ovaries and seasonal patterns in the abundance of juveniles. ' Anonymous. 2001. Report to Congress: status of fisheries of the United States, 122 p. National Marine Fisheries Sei-vice, 1315 East-west Highway, Silver Spring, MD, 20910. - Armstrong, M. P., P. B, Hood, M. D. Murphy, and R. G. Mullen 1996. A stock assess- ment of permit, Tracltinotiix falcatus. in Florida waters. Unpubl. rep. to the Flor- ida Marine Fisheries Commission. Flor- ida Marine Research Institute, 100 Eighth Avenue SE, St. Petersburg. Florida 33701- 5095. Crabtree et al : Life history of Trachinotus fakatus 27 Methods Collections Permit that we examined were collected from the Florida Keys (/i=308; between 25°40'N. SOnO'W and 24°30'N, 82''20'W) during 1995-97 and the Tampa Bay area (» =228; 27°40'N, 82°45'W) during 1990-95. Most F^lorida Keys permit were caught with hook-and-line gear (;!=215) or speared («=58) over artificial and natural reefs in the waters off the lower and middle Keys in depths ranging from 10 to 30 m. Other, usually smaller, permit were cap- tured with gill nets (/i = 16), seines (/i = 18). and bottom trawls (n=l) over or near shallow banks adjacent to the Keys. Most of the permit sampled in the Tampa Bay area were small (<400 mm FL) and were captured with seines along sandy beaches; some larger Tampa Bay permit were captured with gill nets (/i=53) or trammel nets (/( = 14). Standard length (SL), fork length (FL), and total length (TL) were measured to the nearest millimeter (mm) and weight was measured to the nearest gram. Unless other- wise indicated, all lengths reported in our study are fork lengths. Otoliths (sagittae) were removed, rinsed in water, and stored dry until sectioned; they were later weighed to the nearest 0.01 mg. Gonad weight was recorded to the nearest gram (g), and gonad samples were removed from the fish and preserved in 10'/( buffered formalin; they were later soaked in water for 24 hours and stored in 70^7^ ethanol. Collections of juvenile permit from sandy beaches off Tampa Bay and the Florida Keys were made with a 21.3 x 1.8-m bag seine (6.4-mm mesh in the wings and 3.2-mm mesh in the bag). Seine hauls were made perpendicular to the beach for distances up to 50 m, depending on water depth. Lengths of up to 50 fish from each sample collec- tion were measured to the nearest millimeter. Near Tam- pa Bay, we collected fish at the Gulf of Mexico beaches of Treasure Island (November 1992 -October 1994; 27°46'N, 82°46.5'W) and Indian Shores (August 1993-November 1994; 27°50'N. 82''50'W). Sampling at each site consisted of five seine hauls every two weeks. Six sandy beaches were sampled monthly in the Florida Keys from July 1994 to July 1997: Lower Matecumbe Beach (July 1994-April 1996; 24°50.95'N, 80°4415'W), Coco Plum Beach (July 1994-April 1996; 24°43.65'N, SFOO.lO'Wi. Clarence P Higgs Beach. Key West (July 1994-July 1997; 24°32.79'N, 81°47.26'\V), Bahia Honda State Park (October 1994-May 1997; 24°39.81'N, 81°15.44'W). Boca Chica Beach (Oc- tober 1994-April 1996; 24°33.60'N, 81°41.65'W), and Sugarloaf Beach (January 1995-May 1996; 24°36.57'N, 81°33.49'W). Age and growth The left sagitta was usually used for age estimation; how- ever, if the left otolith was broken, lost, or destroyed during processing, the right otolith was substituted. We prepared otoliths for age estimation by embedding them in Spurr, a high-density plastic medium (Secor et al.. 1992). A 1-mm to 2-mm-thick transverse section containing the otolith core was cut with a Buehler Isomet low-speed saw with a diamond blade. The section was mounted on a microscope slide with thermoplastic glue (CrystalBond 509 adhesive) and was polished with wet or dry sandpaper (grit sizes ranging from 220-2000) until annuli were visible. Sec- tions were then polished on a Buehler polishing cloth with 0.05-gamma alumina powder to remove .scratches. With- out knowledge of fish size or capture date and using a compound microscope equipped with transmitted light, two readers independently counted annuli on each otolith twice. If three of the four readings agreed, then this mode was accepted as the annulus count. If three of the four readings did not agree, each reader again counted annuli independently and without knowledge of previous counts. If three of the resulting six readings agreed, then this mode was accepted as the annulus count. If there were not three readings that agreed, the otolith was excluded from further analysis. In six cases, two sets of three readings that were in agreement occurred. For these six otoliths the two sets of readings differed by only one annulus; there- fore the mean was accepted as the annulus count. The percentage of otoliths with an annulus on the edge was then plotted by month so that we could look for a sea- sonal pattern in annulus formation. We did not attempt to measure marginal increments because the margin of per- mit otoliths is highly sculptured and easily broken; how- ever, we did believe that we could discern the presence of an annulus on the otolith's edge. The von Bertalanffy (1957) growth equation FL, = L,, (1-e '"'"'"') was fitted to observed age-length data with nonlinear regression procedures. Age was esimated as the annulus count because permit both spawn and form annu- li at about the same time of year. Our estimates of length at age include some seasonal growth that occurred after the formation of the final annulus. Length-weight regres- sions were calculated by linear regression of logju-trans- formed data. Sex-specific growth models were compared with an ap- proximate randomization test described by Helser (1996). This test is based on the premise that when the null hy- pothesis of no sex-specific differences in growth is true, a test statistic derived by random assignment of fish to one of two populations will not be different from that observed between sexes. The test statistic is calculated as the re- sidual sums of squares for the sexes-combined von Berta- lanffy growth model minus the residual sums of squares for the two sex-specific models. A probability distribution of the test statistic was generated by a randomization rou- tine with 1000 iterations of the nonlinear models. Only sexed fish were included in the statistical comparison. Age validation Permit used in the age-validation experiments were cap- tured in waters off the Florida Keys with hook-and-line gear After capture, permit were tagged with dart-type tags and injected with Liquamycin LA-200 (200-mg oxy- tetracycline |OTC|/mL) in the dorsal musculature at a dosage of about 100-mg OTC per kg fish weight. Permit were then held in a 33.5-m-long by 5.5-m-wide by 0.75-m- 28 Fishery Bulletin 100(1) deep pond at the Florida Fish and Wildhfe Consei-vation Commission's Keys Marine Laboratory in Long Key. Fish were held at ambient temperatures and were fed frozen shrimp and fish until satiated at least three times a week. Although several permit were injected and held for vari- ous periods, only one fish survived long enough to have formed an annulus after the OTC injection. The otolith section from this fish was examined with a compound microscope (40-lOOx) equipped with ultraviolet light so that the fluorescent OTC mark could be detected. Reproduction Histological sections of gonads were prepared and assessed for reproductive state. Gonad samples were prepared for histological examination with a modification of the peri- odic acid Schiff's (PAS) stain for glycol-methacrylate sec- tions and with Weigerts iron-hematoxylin as a nuclear stain and metanil yellow as a counterstain (Quintero- Hunteret al., 1991). Developmental stages of oocytes were determined and oocytes were counted from histological preparations at lOOx with a compound microscope attached to a digital im- age-processing system. Four oocyte stages were recognized in permit ovaries: primary growth, cortical alveolar, vitel- logenic, and oocvtes in the final stages of maturation (Wal- lace and Selman, 1981). The final stages of oocyte matu- ration (FOM) included yolk coalescence, germinal vesicle migi-ation, germinal vesicle breakdown, and hydration. We also counted postovulatorv follicles (POFs) and PAS-pos- itive melanomacrophage centers (Ravaglia and Maggese, 1995; Crabtree et al., 1997), which were present in many ovaries. When stained with the PAS stain, these PAS-pos- itive structures are brilliant purple. Melanomacrophage centers are thought to be active in degrading atretic oo- cytes, postovulatory follicles, and residual cells of the sper- matogenic cycle (Chan et al., 1967; Ravaglia and Maggese, 1995). The developmental stage of at least 300 oocytes and other structures on each slide was determined and count- ed in arbitrarily chosen fields, and frequencies were ex- pressed as a percentage of the total count. We counted all oocytes that had at least 50*^* of their area visible in a field before moving to the next field. We examined seasonal reproductive patterns by plotting monthly juvenile length frequencies and monthly mean go- nadosomatic indices (GSIs). Gonadosomatic indices were calculated for 129 sexually mature female permit ranging in length from 476 to 916 mm and for 122 sexually mature male permit ranging in length from 449 to 855 mm as GSI = (GW / (7W - GW )) 100, where GW = total gonad weight 0.05). Neither the slopes (P=0.464) nor the elevations (P=0.063) of the length-weight equations for male and female permit were significantly different. The pooled length-weight equation for sexed and unsexed fish was logi„Wr = 2.803 log,,, FL - 4.078, {n=488, 7--=0.996) where WT = weight in grams; and FL - fork length in mm. Age and growth When viewed with transmitted light, permit otoliths have opaque (dark) annuli that alternate with translucent (light) zones (Fig. 2). Proceeding from the otoliths core towards the otoliths proximal margin, annuli are regu- larly spaced along the sulcal ridge. In some individuals, the annuli are indistinct and irregular in appearance, which made age estimation difficult. We considered 51 oto- liths (17.3%) from permit ranging in length from 243 to 916 mm to be unreadable. The length-frequency distri- bution of fish whose otoliths were considered unreadable was not significantly different from that offish whose oto- liths were considered readable (Kolmogorov-Sniirnov two- sample test, Z)=0.144, P=0.32); thus, no particular length Crabtree et a\ Life histoid of Tmchlnotus falcatus 29 200 400 600 800 1000 1200 25^ females 20 ; r^187 ^5'- 10 n -in 5 in j L 200 400 600 800 1000 1200 Fork length (mm) E 2 20 15 10 5 25 20 15 10 5 males n=124 10 15 20 25 females n^127 10 15 20 25 Age (yr) Figure 1 Fork lengths (mm) and ages (years) of male and female permit. Trachinotus falcatus, sampled from South Florida waters. group of fish was systematically excluded from the age- and-grovvth analysis. Annulus formation in permit occurs during spring and early summer. The percentage of permit with an annulus on the otolith's margin was greatest during summer and least during October-March, suggesting that annulus for- mation is seasonal and that annuli first become visible during late spring or early summer (Fig. 3). A single OTC-injected permit was successfully held for a sufficient length of time to be useful in age validation. This fish was captured and injected with OTC on 17 June 1993. The fish was sacrificed on 30 January 1996 and measured 600 mm in length. After 31 months in captivity, which included two spring-summer periods, the fish had formed two annuli, a number that is consistent with our hypothesis that a single annual mark forms annually during late spring or early summer. Also visible immediately before the OTC mark was an annulus that was probably formed during late spring of 1993, just prior to capture and OTC injec- tion. Moreover, there was a wide margin subsequent to the last annulus that is consistent with the six or more months of otolith growth after formation of the final an- nulus in late spring or early summer of 1995. Estimated ages of 298 permit ranged from to 23 years for fish 102 to 900 mm long. Permit grew rapidly until about age five, and then growth slowed considerably (Ta- ble 1, Fig. 4). Most of the fish in our sample were less than 10 years old, although fish 10-15 years old were common. The oldest permit examined was a 23-year-old I781-mm) male (Table 1). Estimates of von Bertalanffy growth model parameters are presented in Table 2. The growth models for male and female permit were not significantly differ- ent (approximate randomization test, P=0. 059). Sexual maturation We estimated that 50'7f of the males in the population reached sexual maturity by 486 mm and an age of 2.3 years, and 509^ of the females in the population reached sexual maturity by 547 mm and an age of 3.1 years (Table 3). The smallest sexually mature male in our sample was 449 mm long, and the youngest sexually mature male was 3 years old. Our estimate of the age at 50'7f maturity for males was less than the age of the youngest mature male observed. This knife-edge maturity curve could be an artifact of our small sample size. The smallest sexually mature female in our sample was 476 mm long, and the youngest sexually mature female was 3 years old. All of the ovaries we examined contained primary-growth-stage oocytes. Cortical alveolar-stage oocytes occurred only in ovaries from permit larger than 450 mm and older than 2 years and were common only among permit larger than 500 mm and older than 3 years. Vitellogenic oocytes were found only in ovaries from fish larger than 550 mm and older than 3 years and were common only among permit larger than 600 mm. The length and age at which vitel- logenic oocytes were commonly found agrees well with our estimate of the length and age at which 50% maturity was 30 Fishery Bulletin 100(1) B Figure 2 (A) Sectioned sagitta from a 1-year-old (363-mni-FL) permit. Ti-achinotus falcatiis. collected in the Florida Keys on 27 Fet^ruary 1995. showing the location of the core (white arrow! and the first annulus (black arrow). Scale bar = 200 microns. (B) Sec- tioned sagitta from a 23-year-old male permit (781-mm-FL) collected in the Florida Keys on 4 June 1996. Scale bar = 500 m. reached, suggesting that we misclassified few gonads with regression. Spawning seasonality Permit spawning appeared to be seasonal in the areas we sampled and occurred at least during May^uly. We examined 15 permit ovaries that contained either oocytes in the final stages of maturation or POFs, structures indicative of imminent or recent (<24 h) spawning. We usually did not know the time of day when fish were caught, but all fish were captured during daylight hours (0700-1700 h). Oocytes in the final stages of maturation were found during June and July, and POFs were found Crabtree et a\ : Life histoi"y of Tiachinotus fakotus 31 Table 1 Average obsci-\ed and prediclcd l'( irk lengths (mini of permit, 'I'nn IiiikiIiis fat catus. Till average obsei-ved length at age includes some seasona growtli that occuitl d after the format ion ofthe linal anniilus. \ allies in p; irentheses are standard error and sample | size. Age Sexes combined Females Males Average Average Average (yr) observed Predicted observed Predicted observed Predicted 160(12.7:17) 139 277(1) 212 149 1 301 (5.9;56) 319 310(14.5:8) 353 334(12,5:17) 337 2 476(7.6;10) 447 479(6.8:8) 458 465(33.5:2) 464 3 564(10.1:27) 537 555(12.1:13) 538 572(15.9:14) 550 4 612(6.3:28) 601 620(9.8:14) 599 604(7.6:14) 608 5 643 (8,7:29) 645 664(13.8:10) 644 632(10.5:19) 647 6 663 (8.7:26) 677 658(10.7:20) 679 680l9.3;6i 673 i 687(12.0:17) 699 687(10.8:9) 704 687(23.6:81 691 8 703(16.9:12) 715 695(22.7:7) 724 715(27.4:5) 703 9 713(15.1:21) 726 710(26,8,11) 743 717(13.3:10) 711 10 743(30.0:6) 734 754(34.2:5) 750 688(1) 717 11 746(21.8:12) 740 811(25.3:4) 758 714(23,2:8) 721 12 738(35.3;5) 744 797 (0,5:2) 765 698(46,9:3) 723 13 787(19.4:9) 746 803 (26.2:6) 769 754(15,7:31 725 14 762(19.5:13) 748 783(12.8:7) 773 737(38.9:6) 726 1.5 753(13.4:4) 750 737(1) 776 759(17.3:3) 727 16 751 778 727 17 751 779 727 18 745(48.5:2) 752 793(1) 781 696 ( 1 ) 728 19 752 781 728 20 667(1) 753 782 667(11 728 21 687(1) 753 916(1) 783 687(1) 728 22 753 783 728 23 781(1) 753 783 781(1) 728 Table 2 Parameter estimates ofthe von Bertalanffy growth model for permit. Trachinotus falcatus. from South Florida waters. Values in parentheses are standard errors. FL = fork length. Sex n L (mm FL) K 'o adjusted r- Females 127 784.2 0.28 -1.12 0,833 (13.79) (0.027) (0.249) Males 123 728.2 0.39 -0,58 0,855 (9.52) (0.034) (0.168) Combined 297 753.1 0.35 -0.59 0,921 (7.12) (0.015) (0.065) during May-July (Fig 5), Vitellogenic oocytes were most plentiful during March-July and were absent during October-December, No samples were available for histo- logical examination in January or February, but it seems 100 h c h 29 21 O) , 03 E t c 75 o 21 CO 3 8 , 10 1 50 _ ^ 34 5 12 ^ 25 - o 21 o 6 10 °" - * 4 FMAMJ JASOND Month Figure 3 Mean percentage and standard error of permit {Trachi- notus falcatus) otoliths with an annulus on the margin plotted by month. Numbers above the lines are the monthly sample sizes. 32 Fishery Bulletin 100(1) unlikely that spawning occurred during these months. Females with the greatest GSIs (>4%) were captured during March-August, and GSIs were least ( <1.5% ) during October-December (Fig. 6). Male GSIs were generally sim- ilar in magnitude to female GSIs and followed the same pattern. In the Tampa Bay area, small permit (<40 mmi were present from June to November, suggesting that spawning extends into the fall. In the Florida Keys, small fish (<40 mm) were present all year, suggesting an extended spawn- ing season, recruitment from other areas with different seasonal spawning patterns, or variable juvenile growth rates. 1000 750 500 . I 250 >,\ oL xiWW 10 15 All n=297 20 25 1000 750 500 1: i!" ' 250 OL females n=127 5 10 15 20 25 Age (yr) Figure 4 Observed and predicted fork lengths (nimi from the von BertalanfTy growth model for sexed and unsexed permit, Trachinotiis falcatus. Discussion We obtained permit from a variety of fishery-dependent and fishery-independent sources; consequently, our sample is biased towards certain size classes, and the bimodal size-frequency distribution of our sample probably does not reflect that of the population or the Florida harvest. All the small fish (<300 mm) we examined were from fishery-inde- pendent sources; most large fish were from fishery-depen- 25^ • 20 ,- Frequency O Ol t 5 ^ ^ • 1 • : : f • M A M J J A S O N D Month Figure 5 The percen t frequency of occurrence of oocytes in the | final stages of oocyte maturation (FOM) and postovu- latory follicles (POF) in individual permit iTrachino- \ tus fa lea tun ) ovaries plotted by month. Table 3 The relationship of percentage mature and fork length (mm I and the relationship of percentage mature and age (years) for permit, Trachiriotus falcatus. from South Florida waters. FL = fork length (mm) and AGS = age (years). Pr,„is the absolute value of ((a +6 )/c). is the inflection point of the curve, and is the length or age predicted by the logistic regression at which 50''r of the permit in our sample were sexually mature. Sex is a dummy variable equal to 1 for males and for females. PD is the adju.sted percentage of deviance explained by the model. Percent female '■'/(1+e" X PD FL 314 -30.41 3.34 0.056 (6.336) (1.087) (0.0114) AGE 233 -6.71 1.71 2.14 (0.878) (0..'576) (0.238) 0.84 0.09 486 mm (males) 547 mm ( females i 2.3 years (males) 3.1 years (females) Crabtree et al,: Life history of Tmchtnotus lakatus 33 dent sources, such as charterboats. Wo did not sample any permit from the commercial fishery, which principally tar- gets smaller fish as a result of the maximum size limit of 20 inches (508 mm FL) for permit caught by commercial ves- sels. Ai'mstrong et al.- reported that most han-ested permit in Florida were <440 mm. In contrast, our sample contained many fish larger than 600 mm. The high proportion of large permit in our sample could reflect a tendency for charter- boats in the Florida Keys to select larger permit than those selected by more typical anglers statewide. Ai-mstrong et al.'s^ assessment was based on more systematic and state- wide sampling than ours, and the differences between their sample and ours probably reflects our attempt to obtain a sample of all available size classes rather than a represen- tative sample of the Florida hai-vest. Age and growth The oldest permit in our sample was estimated to be 23 years old. Although we examined many relatively large permit, larger fish than those we examined have been caught. Robins ( 1992) reported that permit can reach 1100 mm FL and a weight of 23 kg; consequently, permit lon- gevity probably exceeds our estimate of 23 years. There are no other estimates of age and growth of permit for comparison, but our longevity estimates are similar to those determined from sectioned otoliths for other caran- gids. Manooch and Potts (1997) aged greater amberjack and found fish as old as 17 years. The oldest carangid yet studied is the trevally, Caranx georgianus, reported to reach an age of 46 years (James, 1984). The much smaller Florida pompano has been reported to reach an age of 7 years (Hood et al.-^). 8 - females 6 . n=129 4 ! i - 2 i 1 Tr i 1 1 V^.-^ M A M J J A S N D 8 males 6 - A • n=122 4 2 :/ / 1: i i i s N D ri J ■r t A M A M J Month Figure 6 Gonadosor na tic indice.s i GSI. • anc means (-(-) for sexually mature fer na le and male permit. Trachinotus fatcatus. plot- ted by mor ith. Our estimates of the von Bertalanffy growth model pa- rameters are within the range of those reported for other carangids (James, 1984; Sudekum et al., 1991; Manooch and Potts, 1997). We found no significant differences be- tween male and female von Bertalanffy growth models, but the significance level (P=0.059) was close enough to 0.05 to cause us to suspect that a difference might exist. Hood et al.'^ also found no sex-specific differences in growth models for pompano. Sexual maturation We sampled relatively few permit between 300 and 500 mm long, the size at which sexual maturity is reached. The lack offish in this critical size range resulted in the knife- edge maturity curves. Larger sample sizes are needed to derive more precise estimates of age and size at sexual maturity. In an assessment of the status of permit stocks in Florida, Armstrong et al.- assumed that permit mature at about 440 mm FL on the basis of limited biological data available at the time. Our estimates of length at 50% maturity are larger: 486 mm for males and 547 mm for females. As a consequence of Florida's 20-inch (508-mm) recreational and commercial maximum size limit, most of the permit harvested are sexually immature (Armstrong etal.2). Spawning We believe that permit spawn over artificial and natural reefs in the waters of the middle and lower Florida Keys because ovaries of fish caught over these structures con- tained oocytes in the final stages of maturation and POFs. Other researchers have inferred that permit spawn in nearshore waters from the capture of early-stage lai-vae (Fields. 1962; Finucane, 1969). Permit ovaries that con- tained fresh POFs and oocytes in the final stages of mat- uration also contained clutches of early- and mid-stage vitellogenic oocytes, suggesting that permit are multiple- batch spawners. Spawning occurred at least during May-June in the Florida Keys during 1995-97. Juvenile length frequen- cies in the Keys suggest a more prolonged spawning sea- son — perhaps even year-round spawning; however, the prolonged presence of small juveniles could also be attrib- uted to variable juvenile growth rates rather than extend- ed spawning. This question could be resolved by direct ag- ing of juveniles to evaluate growth rates. Our sample of adult permit may have been too small to reveal low levels of spawning outside of spring and early summer, and no mature permit were collected during January or February. On the basis of seasonal occurrence of juveniles, Finucane ( 1969) suggested that permit spawn during April-June in the Tampa Bay area, but Fields (1962) found juveniles 3 Hood. P. B.. D. T Menyman. and D. J. Harshany 1999. Age, growth, mortality, and reproduction of the Florida pompano, Trachinotus carolinus, from Florida waters. Unpubl. manu- script. Florida Marine Research Institute, 100 Eighth Avenue SE, St. Petersburg, FL. 34 Fishery Bulletin 100(1) year round suggesting a prolonged spawning period. Oth- er carangids spawn during spring and summer: Caranx tgnobilis and Caranx melampygus spawn during May-Au- gust in Hawaii (Sudekum et al., 1991) and T. carolinus spawns during January-August in Florida (Hood et al.-^). Our data suggest that maturation occurs at greater lengths than assumed by Armstrong et al.-; however, even using our maturation data, their observation that most permit landed are sexually immature remains true. With the current selectivity of the fishery, permit spawning stock biomass could decrease quickly in response to mod- erate levels of fishing mortality; thus, the regulations in place in Florida to restrict harvest levels appear to be jus- tified. Significantly better estimates of the magnitude and age structure of the catch would be required to complete a comprehensive age-structured stock assessment. Acknowledgments We thank Capt. J. C. Wells, who provided us with most of the permit examined in this study and whose efforts made this work possible, and Don DeMaria, who also pro- vided specimens. We thank John Swanson, Bill Gibbs, and the staff at the Keys Marine Laboratory for their assis- tance; Jim Colvocoresses, John Hunt, and others at the South Florida Regional Laboratory for their cooperation; and David Harshany, Heather Patterson, Dan Merryman, Graham Gerdeman, and Connie Stevens for their assis- tance. We also thank Jim Colvocoresses, Rich McBride, Jim Quinn, Judy Leiby, and Llyn French for helpful com- ments on the manuscript. This work was supported in part under funding from the Department of the Interior, U.S. Fish and Wildlife Service, Federal Aid for Sportfish Resto- ration F-59. Literature cited Chan, S. T. H., A. Wright, and J. G. Phillips. 1967. The atretic structures in the gonads of the rice-field eel (Monopterus albus) during natural sex-reversal. J. Zool. (Lend.) 153:527-539. Crabtree, R. E., D. Snodgrass, C. W. Harnden. 1997. Maturation and reproductive seasonality in bonefish, Albula vulpes. from the waters of the Florida Keys. Fish. Bull. 95:456-465. Fields. H. M. 1962. Pompanos iTrachinotus spp. ) of south Atlantic coast of the United States. Fish. Bull. 62:189-222. Finucane, J. H. 1969. Ecology of the pompano iTrachinotus carolinus) and the permit (Trachinotiis falcatus) in Florida. Trans. Am. Fish. Soc. 95:478-486. Reiser, T.E. 1996. Growth of silver hake within the U.S. continental shelf ecosystem of the northwest Atlantic Ocean. J. Fish. Biol. 48:1059-1073. Hunter, J. R., and B. J. Macewicz 1985. Measurement of spawning frequency in multiple spawning fishes. In An egg production method for esti- mating spawning biomass of pelagic fish: application to the northern anchovy, Engraulis mordax (R. Lasker, ed. ), p. 79- 94. NOAA Tech. Rep. NMFS 36. James, G. D. 1984. Trevally, Caranx georgianus Cuvier: age determina- tion, population biology, and the fishery. N. Z. Ministry Agr. Fish. Fish. Res. Bull. 25, 50 p. Manooch, C. S., IH, and J. C. Potts. 1997. Age, growth and mortality of greater amberjack from the southeastern United States. Fish. Res. 30:229-240. Quintero-Hunter, L, H. Grier, and M. Muscato. 1991. Enhancement of histological detail using metanil yellow as counterstain in periodic acid SchifTs hematoxylin staining of glycol methacrylate tissue sections. Biotech- nol. Histochem. 66:169-172. Ravaglia, M. A., and M. C. Maggese. 1995. Melano-macrophage centers in the gonads of the swamp eel, Synbranchus marmoratus Bloch, (Pisces, Syn- branchidae): histological and histochemical characteriza- tion. J. Fish Dis. 18:117-125. Robins, C.R. 1992. American nature guides to saltwater fish. Smith- mark Publ., Inc., New York, NY. 192 p. Secor, D. H., J. M. Dean, and E. L. Laban. 1992. Otolith removal and preparation for microstructural examination. In Otolith microstructure examination and analysis (D. K. Stevenson and S. E. Campana, eds. ), p. 19-57. Can. Spec. Publ. Fish. Aquat. Sci. 117. Sudekum, A. E., J. D. Parrish, R. L. Radtke, and S. Ralston 1991. Life hustory and ecology of large jacks in undisturbed, shallow, oceanic communities. Fish. Bull. 89:493-513. von Bertalanffy, L. 1957. Quantitative laws in metabolism and growth. Q. Rev. Biol. 2:217-231. Wallace. R. A., and K. Selman. 1981. Cellular and dynamic aspects of oocyte gi-owth in tele- osts. Am. Zool. 21:325-343. 35 Abstract-A total of 1784 legal-size (>35G nun TL) hatchery-produced red drum (Sciaenops ocellatus) were tagged and released to estimate tag-reporting levels of recreational anglers in South Carolina (SC 1 and Georgia ( GAl. Twelve groups of legal-size fish (-150 fish/ group) were released. Half of the fish of each group were tagged with an external tag with the message "reward" and the other half of the fish were implanted with tags with the message "$100 reward."These fish were released into two estuaries in each state (n=4); three replicate groups were released at different sites within each estuary (/i = 12). From results obtained in previ- ous tag return experiments conducted by wildlife and fisheries biologists, it was hypothesized that reporting would be maximized at a reward level of $100/tag. Reporting level for the "reward" tags was estimated by dividing the number of "reward" tags returned by the number of "$100 reward" tags returned. The cumulative return level for both tag messages was 22.7 (±1.9)9; in SC and 25.8 (±4.1)% in GA. These return levels were typical of those recorded by other red drum tagging pro- grams in the region. Return data were partitioned according to verbal survey information obtained from anglers who reported tagged fish. Based on this partitioned data set, 14.3 (±2.1)9; of "reward" tags were returned in SC, and 25.5 (±2.3)9, of "$100 reward" tags were returned. This finding indicates that only 56.79; of the fish captured with "reward" tags were reported in SC. The pattern was similar for GA where 19.1 ( + 10.6)9, of "reward" mes- sage tags were returned as compared with 30.1 (±15.6)9; for "$100 reward" message tags. This difference yielded a reporting level of 639; for "reward" tags in GA. Currently, 509; is used as the estimate for the angler reporting level in population models for red drum and a number of other coastal finfish species in the South Atlantic region of the United States. Based on results of our study, the commonly used reporting estimate may result in an overestimate of angler exploitation for red drum. Tag-reporting levels for red drum (Scioenops ocellatus) caught by anglers In South Carolina and Georgia estuaries* Michael R. Denson Wallace E. Jenkins Marine Resources Research InsKtute South CaroNna Department of Natural Resources 217 Ft Johnson Rd Charleston, South Carolina 29422-2559 E mail address (for W. E Jenkins, contact autlior) lenkinswigimrd dnr.slale sc.us Arnold G. Woodward Coastal Resources Division Georgia Department of Natural Resources 1 Consei^ation Way Brunswick, Georgia 31523 Theodore I. J. Smith Marine Resources Research Institute South Carolina Department of Natural Resources PO Box 12559 217 Ft. Johnson Rd. Charleston, South Carolina 29422-2559 Manuscript accepted 1 August 2001. Fish. Bull. 100:35-41 (2002). There are major marine recreational fisheries along the south Atlantic and Gulf of Mexico coasts of the United States that target red drum, Sciaenops ocellatus (Matlock, 1986a; 1986b). Dur- ing the late 1980s, overexploitation of red drum in many states resulted in the closure of commercial fisheries in most states and in the imposition of creel and size limits on catch of rec- reational anglers (McGurrinM Concur- rently, studies were initiated in a num- ber of coastal states to gain a better understanding of red drum life history and to attempt to estimate exploita- tion rates. These investigations relied heavily on the use of fishery-dependent, mark-recapture studies to obtain the data necessary for creating a robust population model (McGurrin^). Generic population models have been developed by using mark-recapture studies to estimate expected number of animals that survive and are re- captured from a year class within a giv- en year (Brownie et al., 1985). Pollock et al. (1991) emphasized the need to modify tag recovery models in which data from multiyear tagging studies were used and suggested incorporat- ing variables for postmarking survival and for reporting to estimate the re- capture component of the model more accurately. The current model used to estimate recovery (recapture) rates of tagged fish (0) includes a number of variables in an attempt to accurately account for what happens in nature {9=5 km). At each site, fish were released individually approximate- ly every 20 meters along the edge of the salt marsh to min- imize the possibility of schooling behavior and subsequent multiple captures by individual anglers. A total of 1774 fish were tagged and released during the project. Approximately 150 fish were released at each stocking site within each estuary (Table 1 1. Equal num- bers of fish released at each site contained "reward" or "$100 reward" tags. Fish were released into Charleston Harbor, SC, and St Simons Sound, GA, during the fall of 1996 and into Calibogue Sound. SC, and Wassaw Sound, GA, during late spring and early summer 1997 (Table 1, Fig. II. The expiration date for "$100 reward" tags de- ■vy y Charleston Harbor Calibogue Souna Wassaw Sound "^ Atlantic Ocean '\'^' St, Simons Sound L. FL + 40 80 Kilometers Figure 1 Map of coastal South Carolina (.SO, Georgia (GAi, and north Florida (FL) showing the location of each estuary where tagged red drum were released during the reward study. ployed in fall 1996 was 31 March 1997, and for spring and summer 1997 releases, the expiration date was 31 Decem- ber 1997. Neither the study nor the releases were publi- cized in any way other than by the normal information provided by ongoing tagging programs in each state. Cap- tured tagged fish were reported directly to the respective Department of Natural Resources in each state. Partici- pants who returned tags inscribed with "reward" received a prize that would normally be awarded by each agency (e.g. T-shirt or hati and those reporting a "$100 reward" tag received a state-issued check for that amount. Our study was based on two assumptions: 1) $100 was an adequate incentive to maximize reporting (assumed -lOC^'f ) of captured tagged fish; 2) the quotient of returns (the number of "reward"-inscribed tags divided by the re- turns of "$100 reward" tags) would yield the angler report- ing level (A) for the standard "reward" tag. Tags were re- turned in either of two ways; phone message or mail. All anglers who reported tags were later interviewed. During the interviews respondents were asked to confirm their reporting information and to express their attitudes and 38 Fishery Bulletin 100(1) Table 1 Cumulative data on release locations and stocking dat es for fish , and both number of tags released and returned for each reward | message. Release location Stocking date Tag " Reward" "$100 reward" No. released No. returned No released No. returned Charleston Harbor site 1 31 Oct 1996 75 16 75 21 site 2 31 Oct 1996 75 18 75 21 site 3 31 Oct 1996 75 18 75 16 St. Simons Sound site 1 13 Nov 1996 75 10 75 17 site 2 13 Nov 1996 74 11 74 11 site 3 13 Nov 1996 75 10 75 15 Wassaw Sound site 1 8 May 1997 73 31 73 42 site 2 8 May 1997 75 23 75 29 sites 8 May 1997 68 10 68 18 Calibogue Sound site 1 5 Jun 1997 75 9 75 21 site 2 9 Jul 1997 73 19 73 23 site 3 10 Jul 1997 74 9 74 12 opinions about the reporting procedure. All participants were asked the same questions from a standardized sur- vey script. During the interview no information was pro- vided to the anglers about the study design. For statistical analysis each release site was treated as a replicate. By nesting site within estuary, within state, differences associated with each site, estuary, and state could be treated in the analysis to assess influence of the reward messages. The study design was a 2x2 factorial de- sign (state and reward) with three levels of nesting (state, estuary, and site) (Table 1). Owing to differences in growth rates, insufficient numbers of legal-size fish were available to stock all estuaries during the same month. Thus one estuary in each state was stocked in the fall of 1996 and the remaining estuaries were stocked the following spring and summer However, each stocking group was available for capture during the fall season when fishing pressure is heaviest (Wenner'). Percent return data were arcsine square-root transformed prior to analysis. Return data were analyzed by using a two-way analysis of variance ( ANOVA) with significance determined at P<0.05. The ini- tial analysis examined all reported or "cumulative" data. The data were then partitioned in two additional ways: by single returns and survey data. 'Wenner, C. 1997. Personal commun. South Carolina Depart- ment of Natural Resources, 217 Ft. Johnson Rd. Charleston, SC 29422-2559. Single returns This data set was the most restrictive. The assumption was that the partitioned data would be free of any poten- tial bias associated with captures of multiple fish, or with monetary rewards or interactions with project staff Survey data The data were partitioned according to the angler's answers during the interview to determine whether the inducement of a $100 dollar reward changed his or her reporting behavior This data set included all tags reported individually, all tags of the same message reported as mul- tiples, and all $100 tags. However, it excluded "reward" tags in instances where answers during the interview sug- gested that the angler's behavior had been changed by capturing a fish with a "$100 reward" tag. Mean data for each of these analyses were reported with standard errors. Results Nearly 95% of tags that were returned were reported within 160 days after release of fish. More fish with "reward" tags were reported than those with "$100 reward" tags in one of the 12 release sites. Overall in SC, 151 anglers reported capture of 203 fish with tags. Anglers reported capture of 1-9 red drum per trip. One hundred Denson et a\ Tag-reporting levels for Sciaenops ocellatus in Sorith Carolina and Georgia estuaries 39 and nineteen anglers in SC (79.0'7r of total anglers) reported only one tagged fish during the study. In GA, 184 anglers reported capture of 226 tagged fish. Single reports in GA represented 80.4''; (;( = 148i of the total catch of tagged fish. The overall return level for all fish reported in SC (22.7 [±1.8]%) was not significantly different from that in GA (25.8 [±4.1]%) (P=0.8129. F=0.67) (Table 2). For the cumulative data, no significant differences were detected between "$100 reward" (27.8 [±3.3]%l and "reward" tags (20.8 [±2.7]%) (P=0.0724, F=12.33l (Table 2). There were also no statistical differences in the cumulative data among the estuaries within states (P=0.0604. F=4.07) (Table 2) and no detectable interaction between state and reward or reward and estuary within states, from the high variability in the cumulative data among estuaries and sites (52.5% and 47.5% of total variation, respectively). Single returns To further restrict the potential for bias caused by inter- action of different reward messages or caused by the project biologist, capture reports were partitioned to include instances where an angler returned only one tag during the entire study. Overall, no significant differ- ences (P=0.1215,F=6.76) were detected between the single returns of "reward" (11.6 [±1.11% ) and "$100 reward" ( 15.0 [±2.5]%) treatments within SC. This was also the case in GA (P=0.1215, F=6.760 where 15.1(±2.9)% of "reward" tags were returned, as compared with 17.6 (±2.7)% for "$100 reward" tags (Table 3). In addition, when data were compared between states, no differences were detected (P=0.6152, F=0.35). However, when single returns among estuaries were compared, Wassaw Sound in GA (Fig. 1) yielded significantly higher returns (P=0.0126, P=7.95) than any of the other estuaries where fish were released (Table 3 1. Survey data In SC, 52% of respondents indicated that they had previ- ously caught tagged fish. Of those, several anglers admit- ted that they had not routinely reported tags. Additionally, others ( 16% ) indicated that they would not have reported the tag if it had not been worth $100. In one extreme case an angler who reported six "$100 reward" tags and an equal number of "reward" tags at once, indicated that he would not have turned in an individual "$100 reward" tag because in his words "he did not need the money." In GA, 29% of anglers had caught a tagged fish prior to the study; however only 7 ( 5% ) said that they would not turn in tags worth less than $100. In light of this infor- mation, the return data were partitioned to eliminate po- tential bias that would result from encountering a "$100 reward" tag. This partitioned data set revealed that sig- nificantly fewer (P=0.0310, F=30.81) unbiased "reward" tags (14.3 [+2.1]%) were returned in SC than "$100 re- ward" tags (25.5 [±2.3]%) (Table 4). This was also true in GA, where 19.1(±4.3)% of "reward" tags were unbiased re- turns, as compared with 30.1 (±6.4)*^"^ of "$100 reward" tags (P=0.0310, F=30.81) (Table 4). Table 2 Cumulative mean return level C/r) and standard error for red drum tagged with one of two reward messages ("reward" or "$100 reward"). No significant differences were detected between reward message, estuary, or state. Release location Charleston Harbor Calibogue Sound South Carolina (mean) St. Simons Sound Wassaw Sound Georgia (mean) Overall mean Return level "Reward" "$100 reward" (9f) CJl 23.1 ±0.9 25.8 ±2.2 16.7 ±4.7 25.2 ±4.6 19.9 ±2.6 25.5 ±2.3 13.9 ±0.6 19.2 ±2.2 29.3 ±8.0 40.9 ±9.0 21.6 ±5.0 30.1 ±6.4 20.8 ±2.7 27.8 ±3.3 Table 3 Mean tag return level (% ) and standard error for red drum tagged with one of two reward messages ("reward" or "$100 | reward"). There were no significant differe nces in return levels by reward message within or among estuaries with the e.xception of those from Wassaw Sound which were sig- nificantly higher (P<0.05 noted by *) for both reward mes- sages than any other estuary. SC = South Carolina; GA = Georgia. Tag message "Reward" '$100 reward" Release location ('7c) (%) Charleston Harbor, SC 13.3 ±1.6 15.1 ±5.3 Calibogue Sound, SC 9.9 ±1.0 14.8 ±2.0 South Carolina (mean) 11.6 ±1.1 15.0 ±2.9 St. Simons Sound, GA 9.9 ±1.6 13.0 ±1.6 Wassaw Sound, GA 20.2 ±3.8* 22.1 ±3.7* Georgia (mean) 15.1 ±2.9 17.6 ±2.7 Overall mean 13.3+1.6 16.3 ±1.8 Discussion Overall return levels for the tagged fish released in our study were similar to levels of angler return for red drum in each states fishery-dependent tagging programs (Wenner^, Woodward''). Because of high variability within estuaries, there were no significant differences between returns of "reward" and "$100 reward" according to the analysis of cumulative return data. The high variability Woodward. A. G. 1997. Personal commun. Georgia Depart- ment of Natural Resources, 1 Conser\-ation Way, Brunswick, GA 31523. 40 Fishery Bulletin 100(1) Table 4 Mean return level {'?/ ), standard error, and range for un biased data lad ustments based on verbal interviews) fo red drum tagged with one of two reward messages ( "reward" or "$100 rew ard"). Return data for the "$100 rewai d" message were s gnificantly higher (P<0.05 ) for each estuary, state, and overall than those for the "reward' message. Release location Tag message Unbiased reporting Mean level' (rn Range "Reward" Ci I $100 reward"!'*) Charleston Harbor 17.3 ±1.3 2.5.8 ±3.9 67.1 57-78 Calibogue Sound 11.3 ±6.3 25.2 ±8.0 44.8 19-67 South Carolina (mean) 14.3 ±5.2 25.5 ±5.6 56.7 — St. Simons Sound 11.7 ±2.1 19,2 ±3.9 60.9 41-82 Wassaw Sound 26.5 ±10.4 40.9 ±15.7 64.8 56-79 Georgia (mean) 19.1 ±10.6 30.1 ±15.6 63.4 — Overall mean 16.7 ±6.2 27.8 ±11.5 60.1 19-82 ' Example: Charleston Harhnr: -$100 reward" tags reported - 00' r: 17. ■3/2.'") 8 = STl'r reportinj^ level for "rew; ird" ta^s. between sites within the same estuary was unexpected. In addition, variation between estuaries in the same state made comparisons between states difficult. However, "reward" tags were returned less often than "$100 reward" tags from 11 of the release sites in the unpartioned data set. After identifying and excluding possible sources of bias, we found that there were statistically significant dif- ferences between reporting level of "reward" and "$100 reward" tags in all areas (Table 4). The range of 19-82''i in levels of reporting between sites was more variable than anticipated (Table 4). Removal of the suspected biased anglers from the data set resulted in a mean unbiased reporting level of 67.1'~f in Charleston Harbor and 44.8'f in Calibogue Sound (Table 4). Unbiased reporting in GA was somewhat higher than in SC (63.4'7f vs. 56.7'"*). The fact that significant differences were found only after biased angler data were removed from the data set illus- trates that a small number of skilled anglers can have an effect on fisheries-dependent data. Their failure to report tags may be due to a lack of novelty in encountering tagged fish, or to insufficient reward incentives (having already received a number of t-shirts, fishing caps, etc. I. These data suggest that use of noncash rewards is ben- eficial only for the first time an angler catches a tagged fish and decreases as anglers catch additional tagged fish. Further repeated exposure to tagging programs within each state eventually results in angler ambivalence and reduced cooperation. This indifference is of particular con- cern with the use of a constant regional reporting rate as described by Hocnig et al. (1998). A decreasing rate of tag return could be mistaken for lower hai-vest, reduced fish- ing effort, poor survival, or increased population size. Lack of differences in reporting levels between "reward" and "$100 reward" in the single-return (one fish) parti- tion of data (Table 3) confirms that anglers who capture many tagged fish per trip or per season (who were omitted from this data set) significantly influence reporting. Sin- gle return-data also suggest that anglers who catch fewer fish (tagged or not tagged) are more likely to report cap- tures of tagged fish regardless of reward amount. Consid- ering the impact a few skilled anglers can have on tag re- porting estimates, these results demonstrate the need for further evaluation of the interaction between tagging pro- grams and angler behavior. The 50*7^ reporting level cur- rently used by managers is approximately a IT^'i under- estimate (.50/60=0.83) of actual reporting recorded for the red drum fishery in SC and GA. Continuing to use the 50'7( reporting estimate for this fishery will be more conserva- tive than using the actual reporting level (A) to calculate angler recovery rate (H). Reporting was also extremely site specific, and application of data from one site to a broader area may not be appropriate. Ideally tag-recapture models should be weighted by site-specific reporting information to account for this variability which could be accomplished by regular deployment of high value (>$100) reward tags within each system to gauge angler reporting. Even if of- fering a $100 does not result in lOO*^? reporting, as Nichols et al. ( 1991) suggested, it may yield the highest reporting possible with monetary incentives, meaning that our unbi- ased reporting may have been slightly overestimated. Re- gardless, this approach is still more accurate than that of adopting a regional average. Our results emphasize that researchers need to conduct controlled tag reward studies regularly and also to offer sufficient rewards in order to avoid under reporting. Furthermore, tag reports must be followed up with angler interviews to determine attitudes and give managers an opportunity to remove bias from the data (Reinecke et al., 1992. Zale and Bain, 1994, Pegg et al.,1996). Acknowledgments We would like to thank the staff of the Inshore Fisheries Sections of the SCDNR and GADNR for tagging, distri- bution of fish and tag collection and processing. We espe- cially thank John Fortuna and Carolyn Belcher for their assistance with statistical design and data analysis. We Denson et al : Tag reporting levels for Sdaenops ocellatus in South Carolina and Georgia estuaries 41 also thank Charlie Bridghain and Allan Hazel, for produc- tion, maintenance, and transportation offish, and Charlie Wenner, for reviewing this manuscript and providing valu- able insights during the project. The study was funded in part by USDOC, NMFS the Saltonstall-Kennedy Pro- gram gi-ant #A67FD0031 and NA77FD0062 and the state of South Carolina. Literature cited Brownie, C, D. R. Anderson, K. P. Burnham, and D. S. Robson. 1985. Statistical inference from band recovery data: a band- book. 2nd ed. U.S. Fisb and Wildl. Sei-v. Resour. Publ. 156, 305 p.. Butler. L. 1962. Recognition and return of trout tags by California anglers. Oalif Fish Game 48:5-18. Conroy, M. J., and W. W. Blandin. 1984. Geographical and temporal differences in band report- ing rates for American black ducks. J. Wildl. Manage. 48:23-36. Henny. C. J., and K. P. Burnham. 1976. A reward band study of mallards to estimate band reporting rates. J. Wildl. Manage. 40:1-14. Hoenig. J. M., N. J. Barrowman. K. H. Pollock, E. N. Brooks, W. S. Hearn, and T. Polacheck. 1998. Models for tagging data that allow for incomplete mixing of newly tagged animals. Can. J. Fish. Aquat. Sci. 55:1477-1483. Jenkins, W. E., M. R. Denson, and T. I. J. Smith. 2000. Determination of angler reporting level for red drum (Sciaenops ocellattif:) in a South Carolina estuary. Fish. Res. 44:273-277. Matlock, G. C. 1981. Non-reporting of recaptured tagged fisb by saltwater recreational boat anglers in Texas. Trans. Am. Fish. Soc. 110:90-92. 1986a. Estimate of the number of red drum anglers in Texas. N. Am. J. Fish. Manage. 6:292-294. 1986b. Estimating the direct market economic impact of sport angling for red drum in Texas. N. Am. J. Fish. Manage. 6:490-493. Murphy. M. D., and R. G. Taylor 1991 Preliminary study of the effect of reward amount on tag-return rate for red drum in Tampa Hay, Florida. N. Am. J. Fish. Manage. 11:471-474. Nichols. J. D.. R. J. Blohm. R. E. Reynolds, J. E. Mines, and J. P Bladen. 1991. Band reporting rates for mallards with reward bands of different dollar values. J. Wildl. Manage. 55:119-126. Pegg, M. A., J. B. Layzer, and P. W. Bettoli. 1996. Angler exploitation of anchor-tagged saugers in the lower Tennessee River N. Am. J. Fish. Manage. 16:218- 222, Pollock. K. H., J. M. Hoenig and C. M. Jones. 1991. Estimation of fishing and natural mortality when a tagging study is combined with a creel survey or port sam- pling. Am. Fish. Soc. Symp. 12:423-434. Rawstron. R. R. 1971. Non-reporting of tagged white catfish, largemouth bass, and bluegills by anglers at Folsum Lake, California. Calif Fish Game. 57:246-252. Reinecke. K. J., C. W. Shaiffer. and D. Delnicki. 1992. Band reporting rates of mallards in the Mississippi alluvial valley J. Wildl. Manage. 56:526-531. Roberts Jr., D. E., B. V. Harpster. and G. E. Henderson. 1978. Conditioning and induced spawning of the red drum (Sciaenops osellatiis ) under varied conditions of photoperiod and temperature. Proceed. World Aqua. Soc. 9:311-332. Ross. J. L.. T. M. Stevens, and D. S. Vaughan. 1995. Age, growth, and reproductive biology of red drums in North Carolina waters. Trans. Am. Fish. Soc. 124:37-54. Yeager. D. M.. and M. J. Van Den Avyle. 1979. Rates of angler exploitation of largemouth. small- mouth, and spotted bass in Central Hill Reservoir. Ten- nessee. Proc. Annu. Conf Southeast. Assoc. Fish Wildl. Agencies 32:449-458. Zale.A. v., andM.B. Bain. 1994. Estimating tag-reporting rates with postcards as tag surrogates. N. Am. J. Fish. Manage. 14:208-211. 42 Abstract— Mayan cichlids ^Cichlasoma urophthalniiis) were collected monthly from March 1996 to October 1997 with hook-and-line gear at Taylor River. Flor- ida, an area within the Crocodile Sanc- tuary of Everglades National Park, where human activities such as fish- ing are prohibited. Fish were aged by examining thin-sectioned otoliths, and past size-at-age information was gen- erated by using back-calculation tech- niques. Marginal increment analysis showed that opaque gi'owth zones were annuli deposited between January and May The size of age-1 fish was esti- mated to be 33-66 mm standard length (mean=45.5 mm) and was supported by monthly length-frequency data of young-of-year fish collected with drop traps over a seven-year period. Mayan cichlids up to seven years old were observed. Male cichlids grew slower but achieved a larger size than females. Growth was asymptotic and was mod- eled by the von Bertalanffy growth equa- tion L,=263.6( l-exp[-0. 166( ?-0.001 1] ) for males (/•'''=0.82, ;i=581 ) and Z,,=21.5.6 (l-e.\p|-0.197(r-0.058ll I for females !;■-= 0.77, n =639). Separate estimates of total annual mortality were relatively con- sistent 1 0.44-0.60 ( and indicated mod- erate mortality at higher age classes, even in the absence of fishmg mortality. Our data indicated that Mayan cichlids grow slower and live longer in Florida than previously reported from native Mexican habitats. Because the growth of Mayan cichlids in Florida periodi- cally slowed and thus produced visible annuli, it may be possible to age intro- duced populations of other subtropical and tropical cichlids in a similar way. Age, growth, and mortality of the Mayan cichlid (Cichlosoma urophthalmus) from the southeastern Everglades Craig H. Faunce Estuanne and Marine Research Group Tavernier Science Center, Audubon of Florida 115 Indian Mound Trail Tavernier, Florida 33070 E-mail address cfaunceaaudubonorg Heather M. Patterson Florida Manne Research Institute Florida Fish and Wildlife Conservation Commission 100 Eighth Avenue SE St Petersburg, Flonda 33701-5095 Jerome J. Lorenz Estuanne and Manne Research Group Tavernier Science Center, Audubon of Flonda 115 Indian Mound Trail Tavernier. Florida 33070 Manuscript accepted 1 August 2001. Fish. Bull. 100:42-50 (2002). The Mayan cichlid, Cichlasmna uroph- thalmus (Giinther),is native to the fresh and brackish waters of the Atlantic slope of Central America from Mexico to Nicaragua (Miller, 1966), where it is exploited commercially in artesanal fisheries and aquaculture (Martinez- Palacios and Ross, 1992). The first collections of the Mayan cichlid in the United States were made in 1983 from a freshwater habitat and a man- grove creek within Everglades National Park, Florida ( Loftus, 1987 ). Although it remains unknown how or where Mayan cichlids first entered Florida waters, there is evidence that the discovery of this exotic fish was made shortly after their introduction (Loftus, 1987). Since their discovery, Mayan cichlids have expanded their range to include a variety of habitats from Naples (26°05'N, 81°48'W) to West Palm Beach (26°45'N,80''04'W). The species remains abundant in the man-made freshwater canals and estuarine mangrove habi- tats of the region (Trexler et al., 2000). The introduction of the Mayan cich- lid into southern Florida has had both economic and ecological significance. This species supports a small sport fish- ery because it is edible, attractive, and aggressively takes baits and artificial lures (Shafland, 1996). Anglers, howev- er, have mixed feelings towards this fish because it readily takes artificial baits and fights hard on light tackle, and it can interfere with the pursuit of larger gamefishes, such as the common snook (Centropomus undecinialis). In some ar- eas, the Mayan cichlid is the most com- mon fish caught by recreational anglers and is targeted by subsistence anglers. There is concern, however that the in- teraction between Mayan cichlids and native fishes could alter the ecology of the Everglades and Floi'ida Bay region. Although the role of Mayan cichlids as food for higher trophic-level fishes has not been quantified, they themselves are omnivorous and prey upon native fish- es (Martinez-Palacios and Ross, 1988; Howard et al.^). Previous studies of the Mayan cichlid have focused almost entirely on its suit- ability for aquaculture in Mexico (e.g. ' Howard, K. S., W. F Loftus, and J. C. Trexler. 199.5. Seasonal dynamics of fishes in arti- ficial culvert pools in the C-111 basin, Dade County, Florida. Final Rep. CA5280- 2-9024. 34 p. and append. South Florida Research Center, Everglades National Park, Homestead, FL. Faunce et a\ Age, growth, and mortality of Cichlasoma uiophtha/nnis 43 Map of southeastern HC'=Highway Creek i. Martinez-Palacios and Ross, 1986; Flores-Nava et al., 1989; Ross and Bt'veridge, 1995) and on the po- tential for range expansion in the United States I e.g. Stauffer and Boltz, 1994). Few studies have ad- dressed the life history of the Ma- yan cichlid, and only scant infor- mation e.xists on the age structure and growth rate of this species. From the seasonal length-frequen- cy distributions for Celestun La- goon, Mexico, Martinez-Palacios and Ross (1992) concluded that Mayan cichlids from 70 to 130 mm standard length had complet- ed their first spring and were re- productively active, whereas in- dividuals from 131 to 200 mm standard length had entered their second reproductive year. Observ- ing no fish >200 mm, Martinez-Pa- lacios and Ross (1992) concluded that the population of Mayan cich- lids in the lagoon comprised fast- growing fish with one, or two (rarely), reproductive seasons in their lifetimes. Aging of Mayan cichlids using a validat- ed method is needed to determine the accuracy of previ- ously reported age and growth estimates and to compare the age structure between populations from Mexico and Florida. Here we provide a first account of the age, growth, and mortality of the Mavan cichlid from Florida waters. Methods Mayan cichlids were collected from the dwarf mangrove forests of southeastern Florida. This habitat is dominated by small (0.5-2.0 m tall) red mangrove trees (Rhizophora mangle) in an expansive, seasonally inundated wetland of typically shallow water (average maximum depth=30 cm). These mangroves increase in canopy width and height nearer to continuously inundated deeper creeks. The system is inundated mostly by fresh water during July-February but becomes more saline ( 10-35"^^? ) during the dry season (March-June). Cichlids <65 mm standard length (SL) were collected by using drop traps (Lorenz et al., 1997) to determine when Mayan cichlids first recruit. Drop-trap samples were col- lected every six weeks from August 1990 to September 1996 at Highway Creek, Joe Bay, and Taylor River (Fig. 1 ). Larger cichlids (>65 mm SL) were collected by using hook- and-line gear comparable to that used in other studies (Martinez-Palacios and Ross, 1992). Hook-and-line collec- tions were conducted monthly from March 1996 to October 1997 in Taylor River, a major freshwater distributary of the Everglades emptying into northeastern Florida Bay. Each fishing effort continued until approximately 40 fish were obtained. Fish collected by both methods were measured (SL and total length |TL|, mm), weighed to the nearest Figure T Florida showing sampling locations (TR=Taylor River, .JB=Joe Bay, 0.1 gram, and their sex was determined macroscopically when possible (Faunce and Lorenz, 2000). Fish captured during 1994—97 were used for age-and-growth analyses. All lengths reported hereafter are standard lengths. Sagittal otoliths were removed, blotted dry, and stored in vials until they were sectioned. The left sagitta, unless broken, was used for age determination. Otoliths were sec- tioned by using a low-speed Beuhler Isomet saw with dia- mond blade. Three or four 0.5-mm thick transverse sections, one through the core, were cut and mounted on microscope slides with Histomount '■'^' adhesive and allowed to dry. Sag- ittae from fish <100 mm were embedded in Spurr, a high- density plastic medium (Secor et al, 1992) and a 1-2 mm thick transverse section containing the otolith core was then cut. The sections were mounted on a microscope slide with Crystal Bond^-^' 509 adhesive, and polished with wet and dry sandpaper of grit sizes 220-2000 until growth rings were visible. A polishing cloth with 0.05-gamma alu- mina powder was used to remove scratches. A standardized protocol for interpreting otolith growth zones was followed. When viewed with reflected light, the transverse sections of Mayan cichlid otoliths had two dis- tinct regions; 1) an "inner region" extending from the core to the first visible opaque zone (ring), and 2) an "outer re- gion" extending from the first visible opaque zone to the edge of the otolith (Fig. 2). The inner region was typically more opaque than the outer region and sometimes con- tained a visible growth zone or numerous check marks, or both. LTnfortunately, these marks were difficult to inter- pret, inconsistent between sections from individual fish, and in many cases absent altogether Consequently, we did not count any marks from the inner region in our age es- timations. However, the translucent appearance of the out- er region of the otolith made it possible to count distinct, separate, opaque rings when present. The number of rings 44 Fishery Bulletin 100(1) Figure 2 Transverse section of a six-year-old Mayan cichlid tCichlasuma urophthalmiis) otolith showing the outer region (ORl, inner region ( IRi, and five visible annuli ( 1-51. Note that the first annulus (1) corresponds to the fish's second year of growth. A ring correspond- ing to the first year of growth was not consistently visible and was therefore not counted. Measurements for marginal-increment analysis were made on an axis adjacent to the sulcal ridge from the core (C) to the dorsolateral margin (DLM). Scale bar=500 /im. on each otolith section was counted independently by two readers using compound microscopes, and the results were compared. If there was a discrepancy in the counts be- tween readers, the section was re-examined. If a consensus could not be reached between the readers after the third reading, the otolith was excluded from the study. Linear regi-ession was used to determine the relation- ship of otolith radius to standard length and marginal- increment analysis was used to determine the periodicity of ring formation. Distance from the core to the proximal edge of each ring and to the dorsolateral margin of the oto- lith (otolith radius) was measured (Fig. 2). Measurements were made with a digital-image processing system along an axis adjacent to the sulcal ridge. The distance from the outermost ring to the dorso-lateral margin (i.e. mar- ginal increment=MI) was plotted by month (marginal in- crement analysis). Because the majority of Mayan cichlids in Taylor River spawn during May and June (Faunce and Lorenz, 2000), and ring formation occurred during Janu- ary-May, we assigned each fish a biologically realistic me- dian hatching date of 1 June. Fish collected prior to 1 June that had not yet formed a new opaque ring (=high MI), and all fish collected after 1 June, were assigned a yearly age equal to their ring count. Fish collected before 1 June that had already formed a new opaque ring (i.e. an "early" ring) were assigned a yearly age of one less than their ring count. To compare the timing of ring formation between age groups, marginal-increment analysis was performed on pooled ages 0-3 and 4-7 because our monthly sample sizes for individual age classes were insufficient for this analysis. We used linear regression to determine the relationship between standard length and total length for all hook- and-line caught fish. The relationship between standard length and total weight was calculated separately for each sex with logjy-transformed data. Analysis of covariance (ANCOVA) was used to test for significant differences be- tween the slopes and intercepts of male and female length- weight relationships. Length-frequency distributions for males and females caught with hook-and-line were com- pared by using the Mann-Whitney rank sum t-test. Non- linear least squares procedures (SAS, 1989) were per- formed on final obsen'ed age-at-length data to estimate parameters for the von Bertalanffy gi'owth equation L, =L..(l-exp\-Kit-t„)]), where L, = the standard length (mm); L = the asymptotic length; K = the Brody growth coefficient; t = the age (years); and tfy = the age at zero length (von Bertalanffy, 19.57). Faunce et al.: Age, growth, and mortality of Cichlasoma urophthalmus 45 To increase the number of observations used for fitting the prowtli model, we back-calculated past size-al-age information for each sexed fish using the Fraser-Lee melh(ul lollowing Devries and Frie (1996); L, =[(L,, -a)/S,]S, +a, where L, = the back-calculated length of fish when the /"' increment was formed; L^ = the length of fish at capture; S^ = the otolith radius at capture; and S, = the otolith radius at the ;''' increment. The slope, {L^-a)IS^. was calculated for each fish as the slope of a line connecting two points: (S^ , L, ) and (0, a). The \'-intercept parameter a was determined from the relationship between otolith radius and standard length for all fish, and should approximate the fish length at which otolith radius equals zero (Devries and Frie, 1996). Because we could not accu- rately determine the sex of each fish <70 mm. fish whose sex could not be determined were included in the fitting of both male and female growth cui-ves. Catch cui-\'es were analyzed with two methods to determine annual mortality rates for Mayan cichlids. Sun'ival rate (S) and its respective variance were es- timated by using the empirical abundance data (Rob- son and Chapman, 1961) and the regression of the natural logarithm of year-class abundance (Ricker, 1975). The instantaneous rate of mortality (Z) was derived from the relationship Z=-ln(e^). Total annual mortality (A) was computed as A = 1-S. The age at full recruitment to the hook-and-line gear based on our catch cur\'e was determined to be four years. Results The fragile nature of Mayan cichlid otoliths caused a high proportion (54'f ) to be lost during the cutting process. However, only five of the 391 successfully sectioned otoliths were discarded because a consen- sus between readers could not be reached. A newly formed opaque ring was generally obsei'V'ed in fish captured Jan- uary-May, and the mean monthly marginal increment reached a single yearly minimum in June for all age classes examined (Fig. 3). These data indicate that the opaque rings observed were annuli. The growth of young-of-year Mayan cichlids collected with drop traps could be followed by the progression of modal length from monthly length frequencies. Newly re- cruited fish were present in August (mode=10 mm) and grew to a size of 50 mm by June (Fig. 4). An early spawn- ing event in 1993 (senior author, unpubl. data) produced a smaller-size (20 mm) cohort that was obsen'ed in June. Fish with one annulus were much larger (50-149 mm, mean=97 mm) than the size of age-1 fish suggested from our drop-trap data (June mode=50 mm). This information, combined with the presence of marks in the inner region All individuals Figure 3 Monthly mean marginal increment and range for all Mayan cich- lids and pooled age classes 1-3 and 4-7. Note the consistent annual minimum in June. Numbers indicate sample size. of the otolith, led us to conclude that the first annulus in our age estimations was laid down between January and May of the fish's second year of growth, and we added a year to each individual's total age. Length-length and length-weight relationships are giv- en in Table 1. As i-equired by the Fraser-Lee method for back-calculation of size-at-age information, otolith ra- dius and standard length were closely related (SL=131.2x 0R+A.Q2A1, n=37l, r^=0.80). Analysis of covariance did not detect differences between the slopes of length-weight relationships for males and females (F, y.j(,=0.15, P=0.696) but did reveal significant differences between the re- spective intercepts for males and females (F, g3g=4.10, P=0.043). The length of Mayan cichlids at a given age was mod- eled by the von Bertalanffy growth equation (Fig. 5). Pre- dicted lengths fitted well with the final adjusted observed 46 Fishery Bulletin 100(1) 40 30 20 10 40 30 20 10 ' " ^f^T^T^T August (n=522) September (n=491) 40 30 - 20 - 10 - I ' I ' I ' I October (0=193) i MMM '!' i^T T -i T 40 30 20 10 - November (n=772) r f M l l lf'T'T-T T T 40 -1 30 20 10 December (0=714) fiHIv 40 30 20 10 m- January (n=597) -T* M*i T-l^ 40-] 30 - 20 10 February (n=471) 4llM^M- T-T '!■ 40 -, 30 20 - 10 March (n=713) 40 30 - 20- 10 April (n = 587) m^^ 40 30 20 10 - I ' I ' I ' I ' T^T'T'I ' I ' I IVIay (n=430) I t MU ' I ^T I IT 40 -J 30 - 20 - 10 June (n=193) i MM 'f' H 'TT I T 40 -| 30 - 20 - 10 - -■ July No samples I ' I ' I ' I ' I ' I ' I ' I ' I ' I 10 20 30 40 SO 60 70 80 90 100 10 20 30 40 50 60 70 60 90 100 Standard length (mm) Figure 4 Pooled length-frequency histograms for Mayan cichlids collected with drop traps from 1990 to 1996. and back-calculated length-at-age data for males (r'-=0.82, n=581) and females (;'-^=0.77, /!=639). Our observed and back-calculated size of age- 1 fish (mean ±1 standard er- ror=45.5 ±10.11 mm, range=33-68 mm, ti=22} correspond- ed well with the modal length of age-1 fish collected in drop-trap samples (50 mm). Differences in the parameter estimates for the von Bertalanffy growth equation were observed for each sex. Males were larger than females for all ages (Table 2). Although males exhibited a slower growth rate (A") and larger maximum attainable size iL J than females, the von Bertalanffy growth model param- eters were not significantly different between sexes (95% CI, Table 3). Male and female Mayan cichlids up to seven years old were observed. The size of fish examined ranged from 21 to 210 mm (median=127 mm, interquartile range=98 mm, « = 1046). Males ranged from 69 to 210 mm (median=137 mm, inter- quartile range=119 mm, ?;=400), and females ranged from 58 to 190 mm (median=132 mm, interquartile range=115 mm, 71=449) (Fig. 6). The length-frequency distribution for males was significantly larger than that for females (P<0.001). The overall ratio of males to females was 1:1.1. Age-frequency distributions of Mayan cichlids collect- ed by hook-and-line gear suggest that these fish are fully recruited to the fishery at age four (Fig. 7). The majority of males (85.1'7r ) and feinales (81.5%) were 3-5 years old, and there was a significant difference (Mann- Whitney rank sum ^test, P<0.001) in the age-frequency distribution of males (median=3.67 years, interquartile range=2.12) and females (median=4.78 years, interquar- tile range=3.86). Total instantaneous mortality (Z), annu- al survival (S), and annual mortality iA), based on the re- gression of our catch-cui"v'e data, were estimated at 0.57, 0.56, and 0.44, respectively (/■-=0.91, n=3). Robson and Chapman ( 1961) estimates were Z=0.91, S=0.40 (±0.035), andA=0.60. Discussion Transverse otolith sections can be used to precisely age Mayan cichlids from Florida waters. There was a high congruence (98.7%) between the age estimations of each reader Annuli corresponding to years 2-7 were clearly Faiince et a\ Age, growth, and mortality of Cich/nsomn iimphthalrmis 47 Table 1 LeiiKtli-lcngtli. lenKth-vvi'ight, and ololith-r ad us- -standard-len gth regressions for the Mayan cichlid Cichldsonia urophih aintus, from Taylor Slough, Florida. Regi-essions are in the form V = a + bX. TL = = total length (mm); SL = standard length (mm); WT = total weight (g); OR = otolith radius (mini: range = sample standar d length range in r egressions. Values n parentheses are standard | errors. Y a b X n Range (mm) /■- Sexes combined TL 0.6220 1.3067 SL 961 40-210 0.997 (0.2875) (0.00221 SL -0.1281 0.7631 TL 961 40-210 0.997 (0.2202) (0.0013) SL 4.0247 131.2092 OR 371 33-210 0.800 (3.2740) (3.4214) logi„\VT -4.2490 2.9314 log,„SL 847 58-210 0.986 (0.0257) (0.0122) Males log,„UT -9.7958 2.9329 log.oSi 395 84-210 0.986 (0.0874) (0.0178) Females log,„UT -9.7387 2.9232 log„-,SL 444 89-182 0.984 (0,0856) (0.0176) visible on the otoliths. The annulus corresponding to the first year's growth was not consistently clear to the read- ers, which has been observed in thin-sectioned otoliths of other fish species in Florida. Murphy and Taylor (1994) found that the first annulus was visible only in certain individuals of spotted seatrout, Cynoscion nebulosus. Sim- ilarly, Murphy and Taylor (1990) found that the annulus corresponding to the first winter or spring was absent in red drum, Sciaenops ocellatiis. Direct validation of marked otoliths is needed to confirm the presence and location of the first annulus on the otolith of Mayan cichlids. We obsen'ed differences in the growth patterns of males and females that are likely linked to reproduction. Males were larger than females and did not appreciably slow their growth with age. The nearly linear growth of males resulted in a theoretical maximum size (L. i of 263.6 mm, well above the -200 mm commonly observed for this spe- cies (Loftus, 1987; Martinez-Palacios and Ross, 1992; pres- ent study). Larger males are common in riverine and la- goonal populations of tilapias (Cichlidae) and may have a selective advantage during the reproductive season if they can defend a spawning pit or brood against potential pred- ators (Lowe-McConnell, 1982). Because sperm production requires less energy than egg production ( Jalabort and Zo- har, 1982), the slowed growth observed in females com- pared with that for males is likely due to differences in energy budgets during the reproductive season. No significant differences were found by ANCOVA in the slopes of sex-specific length-weight relationships, but there were significant differences in the intercepts of those lines. Because the actual difference between the y-in- tercepts (weight) of each length-weight relationship was <0.001g, we attribute no biological meaning to the statis- Table 2 Average predicted and observed standard lengths ( mn )for male and female Mayan cichlids. Average Standard Age (yrl Predicted observed error n Males 1 40.3 45.5 2.2 22 2 74.4 74.3 1.12 148 3 103.4 102.0 1.24 152 4 127.9 127.1 1.36 139 5 148.7 147.8 1.66 88 6 166.3 173.0 2.56 28 7 181.1 206 1.32 4 Females 1 36.4 45.5 2.16 22 2 68.4 70.6 1.05 156 3 94.7 96.8 1.31 160 4 116.3 119.8 1.4 151 5 134.0 137.0 1.59 102 6 148.6 148.1 1.94 45 7 160.6 151.0 3.22 3 tical difference and consider the length-weight relation- ships for both sexes to be equal. Mayan cichlids in Florida were much smaller at a given age than those reported by Martinez-Palacios and Ross (1992) in Mexico. One-year-olds were 33-66 mm in Florida vs. 70-130 mm in Mexico, and age-2 fish were 44-130 mm 48 Fishery/ Bulletin 100(1) Table 3 Parameter estimates for the von BertalanfTv growth model ( 19571 for Mayan cichlids and associated standard error i SE ) and con- fidence intei'vals (CI). Sex L imm) A' 'i, 'yri U l'~ Males 263.66 0.165 0.001 581 0.82 SE 25.15 0.027 0.124 959, CI 49.28 0.053 0.243 95'^i CI range 214.34-312.95 0.112-0.218 0.242-0.244 Females 215.63 0.197 -0.058 639 0.77 SE 17.33 0.031 0.142 959t CI 33.96 0.061 0.278 95';; CI range 181.67-249.59 0.136-0.258 -0.336-0.220 in Florida and 131-200 mm in Mexico. We found a maxi- mum age of seven yeai's. vvherea.s two years was suggest- ed by Martinez-Palacios and Ross (1992). We offer three explanations for these observed differences in the length- at-age data. First, exploitation rates may differ between 250 Males n = 581 L = 263-6 K = 166 'o = 0007 (•2 = 82 Figure 5 Obsen'ed and predicted lengths at age for male and female Mayan cichlids from the von Bertalanffy gi'owth model. /)=obsei'ved and back-calculated size-at-age information from 148 males and 157 females. study areas. Fish in our study came from the Crocodile Sanctuary within Everglades National Park and had not been exposed to fishing mortality. Fishing for Mayan cich- lids occurs outside of our study area, and heavy exploita- tion can select for faster-growing fish with a shorter life- span (Ricker, 1975 1. Martinez-Palacios and Ross (1992) suggested that their population was over- fished. Second, differences in temperature impact fish growth. Colder winter temperatures in Florida were sufficient to form seasonal marks in the oto- liths of Mayan cichlids and may have caused slower growth than in Mexican populations. Third, the sea- sonal length frequencies of Martinez-Palacios and Ross (1992) were insufficient to accurately identify older year classes. Because growth slows with age, the length-frequency of cohorts corresponding to older age classes can overlap significantly, resulting in erroneously lower age estimates. Future efforts to age Mayan cichlids in Mexico should include thin-sectioned otoliths to evaluate the findings of Martinez-Palacios and Ross (1992). Although the Mayan cichlid has proliferated for over a decade in the natural and man-made habitats surrounding the Everglades, studies are only recent- ly becoming available (e.g. Trexler et al., 2000 1. More introduced fish species are found in Florida than in any other state in the United States, and 13 of the 18 species with established populations are cich- lids (Shafland, 1996). The impact of exotic species on native Florida fishes has been debated: Shafland (1996) proposed no demonstrable effect on native fishes, whereas Courtenay ( 1997 ) argued that lack of available data precludes a determination. Trexler et al. (2000) provided empirical data that support Shafland (1996), concluding that although exotics have been credited with native species extinctions in other ecosystems, native Florida fishes are not specialized or restricted to certain habitats and thus are able to cope with the invasion of exotics. Finding no drastic changes in the native ichthyofauna does not necessarily mean that exotic species do not af- fect indigenous fishes. Exotic species can introduce FaLince et a\ Age, growth, and moitalily of Cichlasoma umphthalmus 49 numerous stresses not easily quantified, e.g. nest prcdation, direct predation, and competition for space (Trexler et al., 2000; senior author, un- puhl. data). These stresses may afTect the pop- ulation dynamics of native fishes by altering their growth rate, increasing mortality, or de- creasing reproductive success. During 1990-99, till' Mayan cichlid population underwent a cycli- cal "boom and bust" pattern of yearly abundance typical of invasive species (Trexler et al., 2000). Why these patterns occur requires a better un- derstanding of the parameters of reproduction, gi'owth. and moi'tality that drive the population dynamics of this species. The presence of the Ma- yan cichlid in the Everglades and Florida Bay es- tuary warrants further research and monitoring efforts in the region to firmly understand the life history of exotics, native fishes, and their role in the ecosystem. Acknowledgments 80 60 40 I 20 n I 20 z 40 Males n = 140 60 • Females n = 449 80 t ^ — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I 50 100 150 200 250 Standard length (mm) Figure 6 Length-frt'quency histoprrams for male and female Mayan cichlids col- lected with hook-and-line gear. We would like to thank Roy Crabtree, Daniel Merryman, Connie Stevens, and Rich McBride for their assistance. Ron Taylor and Mike Murphy shared their technical expertise with us, and with their editorial suggestions, Joe Serafy, two anonymous reviewers, and John Merriner greatly improved the manuscript. This project was funded by the U.S. Ai-my Corps of Engineers through cooperative agi-eement 970092 between Everglades National Park and the National Audubon Society. Literature cited Courtenay, W. A., Jr. 1997. Nonindigenous fishes. In Strangers in paradise (D. S. Siberloff, D. C. Schmitz, and T. C. Brown, eds.), p. 109-122. Island Press, Wash- ington, DC. DeVries, D. R.. and R. V. Frie. 1996. Determination of age and growth. In Fisheries techniques, 2nd ed. (B. R. Murphy and D. W. Willis (eds.), p. 483-512. Am. Fish. Soc, Bethesda, MD. Faunae, C. H., and J. J. Lorenz. 2000. Reproductive biology of the introduced Mayan cichlid, Cichlasoma iirophthalmus, in an estuarine mangrove habi- tat of southern Florida. Environ. Biol. Fish. 58:215-22.5. Flores-Nava, A., M. A. Olvera-Novoa, and A. Garcia-Cristiano. 1989. Effects of stocking density on the growth rates of Cichlasoma umphthalmus (Gunther) cultured in floating cages. Aqua. Fish. Manage. 20:73-78. Jalabort, B., and Y. Zohar 1982. Reproductive physiology in cichlid fishes, with par- ticular reference to Tilapia and Sarothe/'odon. In The biology and culture of tilapias (R. S. V. Pullin and R. H. Lowe-McConnell, eds.), p. 129-140. ICLARM Conference Proceedings 7. 80 60 40 "§ 20 20 40 60 80 Males n = 152 41 52 1 1 9 I 1 33 12 4 L 1 Females n = 160 6 3 22 30 45 53 1 (- 1 1 1 1 1- -1 012345678 Age (years) Figure 7 Age-frequency distributions for male and female Mayan cichlids col- lected with hook-and-line gear Numbers indicate sample size. Loftus, W. F. 1987. Possible establishment of the Mayan cichlid, Cichlaso- ma urophthalnius (Gunther) (Pisces: Cichlidae), in Ever- glades National Park, Florida. Florida Scientist 50:1-6. Lorenz, J. J., C. C. Mclvor, G. V. N. Powell, and P C. Frederick. 1997. A drop net and removable walkway used to quantita- tively sample fishes over wetland surfaces in the dwarf man- groves of the southern Everglades. Wetlands 17:346-3.59. Lowe-McConnell, R. H. 1982. Tilapias in fish communities. In The biology and cul- ture of tilapias (R. S. V. Pullin and R. H. Lowe-McConnell, eds.), p. 83-113. ICLARM Conference Proceedings 7. Martinez-Palacios, C. A., and L. G. Ross. 1986. The effects of temperature, body weight and hypoxia on 50 Fishery Bulletin 100(1) the oxygen consumption of the Mexican niojarra, Cich- lasonia urophthcilmus (Giinther). Aqua. Fish. Manag. 17: 243-248. 1988. The feeding ecology of the Central American cichlid Cichlasoma urophthalmus (Giinther). J. Fish. Biol. 33: 665-670. 1992. The reproductive biology and growth of the Central American cichlid Cichlasoma urophthalmus (Giinther). J. Appl. Ichthyol. 8:99-109. Miller, R. R. 1966. Geographical distribution of Central American fresh- water fishes. Copeia 4:773-802. Murphy, M. D.. and R. G. Taylor 1990. Reproduction, gi'owth, and mortality of red drum Sciae- naps ocellatus in Florida waters. Fish. Bull. 88:531-542. 1994. Age, growth, and mortality of spotted seatrout in Flor- ida waters. Trans. Am. Fish. Soc. 123:482-497. Ricker,W. E. 1975. Computation and interpretation of biological statistics offish populations. Bull. Fish. Res. Board Can. 191, 382 p. Robson, D. S., and D. G. Chapman. 1961. Catch curves and mortality rates. Trans. Am. Fish. Soc. 90:181-189. Ross, L. G., and M. C. M. Beveridge. 1995. Is a better strategy necessary for development of native species for aquaculture? A Mexican case study. Aquaculture Res. 26:539-547. SAS Institute Inc. 1989. SAS/STAT users guide, version 6, 4th ed. Gary, NC, 943 p. Secor, D. H., J. M. Dean, and E. H. Laban. 1992. Otolith removal and preparation for microstructural examination. In Otolith microstructure examination and analysis (D. K. Stevenson and S. E. Campana, eds.), p. 19-57. Can. Spec. Publ. Fish. Aquat. Sci. 117. Shafland, P. L. 1996. Exotic fishes of Florida-1994. Rev. Fish. Sci. 4:101- 122. Stauffer, J. R., and S. E. Boltz. 1994. Effect of salinity on the temperature preference and tolerance of age-0 Mayan cichlids. Trans. Am. Fish. Soc. 123:101-107. Trexler, J. C, W. F. Loftus, F Jordan, J. Lorenz, J. Chick, and R. M. Kobza. 2000. Empirical assessment of fish introductions in a sub- tropical wetland: an evaluation of contrasting views. Bio- logical Invasions 2:265-277. von Bertalanffy, L. 1957. Quantitative laws in metabolism and growth. Q. Rev. Biol. 2:217:-231. 51 Abstract— We examinod seasonal ami annual variation in numbers of StcUcr (northern! sea lions iEumetopias juba- tiis) at the South Farallon Islands from counts conducted weekly from 197-4 to 1996. Numbers of adult and sub- adult males peaked during the breeding season (May-July), whereas numbers of adult females and immature indi- viduals peaked during the breeding season and from late fall through early winter (September-December). The seasonal pattern varied signifi- cantly among years for all sexes and age classes. From 1977 to 1996, num- bers present during the breeding season decreased by 5.99r per year for adult females and increased by 1.9% per year for subadult males. No trend in numbers of adult males was detected. Numbers of immature individuals also declined by 4.5'^r per year during the breeding season but increased by S.O't per year from late fall through early winter Max- imum number of pups counted declined significantly through time, although few pups were produced at the South Faral- lon Islands. The ratio of adult females to adult males averaged 5.2:1 and declined significantly with each year, whereas no trend in the ratio of pups to adult females was discernible. Further stud- ies are needed to determine if reduced numbers of adult females in recent years have resulted from reduced sur- vival of juvenile or adult females or from changes in the geographic distri- bution of females. Population status, seasonal variation in abundance, and long-term population trends of Steller sea lions (Eumetopias jubatus) at the South Farallon islands, California* Kelly K. Hastings William J. Sydeman Point Reyes Bird Observatory 4990 Shoreline Highway Stinson Beach, California 94970 Present address (for K K Hastings): Alaska Department of Fish and Game Division of Wildlife Conservation 333 Raspberry Rd Anchorage, Alaska 99518 Email address (for K K Hastings) kellyhaslingsiSfishgame slate ak us Manuscript accepted 1 August 2001. Fish. Bull. 100(11:51-62(20021. Steller sea lions (Eunwtopias jubatus) range from southern California along the West Coast of North America through the Aleutian and Pribilof Islands to the Kuril Islands and Okhotsk Sea, Japan (Kenyon and Rice, 1961). Major haulouts and rookeries have his- torically been centered at the Aleutian Islands and at islands and mainland sites around the Gulf of Alaska, where over 70% of the world population was located in the 1950s and 1960s (Lough- lin et al., 1984). In 1990, the species was listed as threatened throughout its range under the Endangered Spe- cies Act owing to declines of over 50% from an estimated world population of 240.000-300.000 in the early 1960s to 116,000 individuals in 1989 (Loughlin et al., 1992). Numerically the decline was most severe in the western Gulf of Alaska where 50-80'% declines occurred (Loughlin et al., 1992). Reduced juve- nile sui-vival appears to be the prox- imate cause for the decline (York, 1994); ultimate causes, however, are unknown. Effects of long-term environ- mental change and pollutants on Steller sea lions, and interactions or compe- tition of these sea lions with commer- cial fisheries are potential contributing causes of this decline (NMML'). In contrast to rookeries in the west- ern Gulf of Alaska, southeast Alaska rookeries have increased by more than 60% over the past three decades ( Lough- lin et al., 1992). Based on differences in population trends and genetics (Bick- ham et al., 1996), a distinction has been made between two separate stocks: 1) the eastern stock, ranging from south- east Alaska to California, and 2) the western stock, ranging from the Gulf of Alaska, Aleutian Islands, and Prib- ilof Islands to Russia (LIS. Federal Register 62:24345-24355). In 1997, the National Marine Fisheries Service list- ed the western stock as endangered, whereas the eastern stock remained listed as threatened. However, differ- ences in trends between rookeries in southeast Alaska and those in Cana- da, Oregon, and California may indi- cate that these areas deserve separate management considerations. For example, rookeries in Canada and California suffered 40% and 80% declines respectively, from the early 1900s to 1970 (Bigg, 1988; Ainley et al.-); declines continued over the past * Contribution 790 of the Point Reyes Bird Observatory, Stin.son Beach, CA 94970. ' NMML (National Marine Mammal Labora- tory). 1995. Status review of the United .States Steller sea lion [Eumetopias juba- tuf) population. Report of the National Marine Mammal Laboratory. National Marine Fisheries Service, Seattle, WA, 61 p. [Available from National Marine Mammal Laboratory, 7600 Sand Point Way N.E., Seattle, WA 98115-0070.] ' Ainley, D. G., H. R. Huber, R. R Henderson, and T. J. Lewis. 1977. Studies of marine mammals at the Farallon Islands, Califor- nia. 1970-1975. Final report to the Ma- rine Mammal Commission. Washington D.C. I NTIS publication number PB274046. Avail- able from Point Reyes Bird Observatory, 4990 Stinson Beach, CA 94970.1 52 Fishery Bulletin 100(1) 37''42'4.'S'-N .17'42'(l(l" _17"41'I5"- Sugorlodt' Kiel SOUTH FARALLON ISLANDS Nonh Landing. LiyhlhauscHill Lion /■- A .if Cove r^,'?^^^" \jf^.- ■^ 'ir- Y-'A- (»'^***^'- Vl / \ rv Nonh Farjlli.i Isl,.n,i4 ^ Snulh h jullon Klands . Indian Head SOUTHEAST FARALLON ISLAND \ Piicific Occiin 1 1 i:.? IKI'^d" 12.^ IKI'OII"VV Figure 1 Map of the South Farallon Islands, including Southeast F'arallon Island and West End Island. Steller sea lions were counted weekly from 1974 to 1997 from Lighthouse Hill, and several gi'ound areas: North Land- ing. Cormorant Blind Hill, Sewer Gulch, and Garbage Gulch. several decades at several California rookeries (Westlake et al., 1997; Sydeman and Allen. 1999; Le Boeuf et al.'). Wliereas over 2000 Steller sea lions used the Channel Is- lands in the late 1930s, only 50 animals were obsei^ed there in 1959 (Bartholomew and Boolootian, 1960). Pup- ping at San Miguel Island, an historical rookery, has not been observed since 1981 (NMMLM. Therefore to better understand patterns and causes of the population decline, trends and status of the eastern stock at southern rooker- ies deserve further investigation. The Farallon Islands (:37°42'N. 123°00'W). 40 km off the coast of San Francisco, California, are currently one of the most southerly haulout and breeding areas for Steller sea lions; Aiio Nuevo Island is the only rookery farther south. The Farallon Islands consist of three groups of islands: South Farallones (two islands. Southeast Farallon and West End, separated by a small surge channel). Middle Farallon (an intertidal rock), and North Farallones (four large sea stacks; Fig. 1). Although the status of Steller sea lions in California prior to 1800 was poorly documented. Steller sea lions bred at the Farallon Islands in the 1800s 3 Le Boeuf. B. J., K. A. Ono. and J. Reiter. 1991. History of the Steller sea lion population at Ano Nuevo Island. 1961-1991. Final report to National Marine Fisheries Service, Southwest Fisheries Science Center. La Jolla.CA. Admin, report LJ9145C, 9 p. [Available from National Marine Fisheries Service. South- west Fisheries Science Center. P.O. Box 271. La Jolla. CA 92038.) and early 1900s (Allen. 1880; Rowley. 1929) and were the most abundant sea lion in California and at the Farallon Islands from the early to mid 1900s (Rowley, 1929; Bon- not and Ripley, 1948). A large amount of data is now avail- able to examine seasonal variation and long-term trends at the Farallon Islands from historical surveys conduct- ed from 1927 to 1970 by the California Department of Fish and Game (CDFG; Bonnot and Ripley, 1948; Ripley et al., 1962; Carlisle and Aplin, 1971) and from surveys conducted weekly by Point Reyes Bird Obsei-vatory (PR- BO) from 1971 to 1996. Although maximum numbers de- clined significantly from 1974 to 1997 for the total popula- tion ( 1.6% per year) and for adult females (3.6% per year; Sydeman and Allen, 1999), it is unknown whether num- bers of other age classes also declined and in which sea- sons declines occurred. To understand proximate causes and consequences of the decline, several questions have yet to be addressed: have reduced pup production and re- duced reproductive rates also occurred in recent years?; and what effect has the decline had on the adult sex-ra- tio? Finally, seasonal variation in counts for different sex- es and age classes and variation in the seasonal pattern among years also have not been examined in detail at the Farallon Islands. The objectives of our study were to ex- amine 1) seasonal variation in numbers among sexes and age classes; 2) trends in numbers from 1974 to 1996 by age class, sex, and season; and 3) averages and trends in pup production, reproductive rate, and adult sex-ratio. Hastings and Sydeman Population status of Eumetopias jubctus at the South Farallon Islands, California 53 Methods Survey methods PRBO began conducting surveys of pinnipeds at the South Farallon Islands in 1971. In June 1973 surveys were stan- dardized and all Steller sea lions visible on or in the water near the coast of the South Farallon Islands were counted weekly from standard vantage points on Southeast Far- allon Island: 1) atop Lighthouse Hill (110 m) with bin- oculars or a 20-60x spotting scope, 2) from Cormorant Blind Hill (35 m) with binoculars, and 3) from North Land- ing, Sewer Gulch, and Garbage Gulch with no optical aids I Fig. 1). Most surveys were conducted between 1000 and 1800 hours on Thursdays if visibility was adequate. Begin- ning in 1977, animals were classified by age class (adult male, subadult male, adult female, immature, yearling, or pup) when possible, primarily by body size. Adult males were distinctive as very large animals with large muscular necks bearing well-developed manes of long, coarse hair on the chest, shoulders, and back. Subadult males were dis- tinguished from adult males by their smaller size and less developed mane. Immature individuals included animals of distinctly smaller size, such as young-of-the-year (after November) and animals likely one to four years of age. The adult female category included animals smaller than sub- adult males but larger than immature individuals. Pups were distinguishable from June until late November by their thick, dark brown coats, which were later molted and replaced with a lighter brown coat after five to six months of age. Counts were conducted by numerous observers over the years; several observers conducted surveys for over a decade and all observers were trained in identifica- tion of sea lions by age class. Counts represent minimum estimates of numbers of sea lions hauled out because only SS"* to 90'7f of the islands were visible from the study's vantage points. We compared maximum counts taken during the breed- ing season (June-July i in recent years (1974-97) with counts from surveys conducted a single time during the breeding season (once annually) and intermittently over the years by CDFG from 1927 to 1970. From 1927 to 1938, counts of subadult or adult sea lions (i.e. excluding pups) made by at least two obsei-vers from boats were averaged (Bonnot, 1931, 1937: Bonnot et al., 1938). Meth- ods of counting changed after 1938, such that counts af- ter 1938 could only be compared cautiously with earlier years. Surveys were conducted by airplane, blimp, or boat in 1946 and 1947 and by airplane only from 1958 to 1970 (Bonnot and Ripley. 1948: Ripley et al, 1962; Carlisle and Aplin, 1971). Counts from 1946 to 1970 were likely over- estimates because observers assumed that all sea lions north of Point Conception were Steller sea lions (many sea lions may have been California sea lions, Zalophus califoi-niaruis) and because pups were likely included in these counts (Bonnot and Ripley, 1948; Ripley et al., 1962). Counts conducted by PRBO since the 1970s targeted only the South Farallon Islands, whereas CDFG counts includ- ed the South and North Farallon Islands. Monitoring of the North Farallon Islands since 1970. however, has been sparse. Although the North Farallon Islands are a known haulout area for Steller sea lions, pupping rates are un- known. The North Farallon Islands were surveyed during the breeding season by PRBO in 1977 (when 17 adult fe- males and 1 pup were counted) and in 1983 (when 92 adults but no pups were counted; PRBO, unpubl. data^). Because of the exclusion of the North Farallon Islands in recent counts, comparisons with earlier CDFG data were made cautiously. Under the direction of D. G. Ainley and H. R. Huber, pup production and pup mortality were monitored intensively from 1973 to 1986, when animals were breeding in accessi- ble areas. Breeding areas were checked daily for new pups, and prematurely born and dead pups were noted. Breed- ing areas shifted from accessible to inaccessible areas over the years. From 1973 to 1975, all full-term pups were born on the more accessible Saddle Rock, a small islet one- quarter mile offshore (Fig. 1), and a few premature pups were born on the mainland. From 1976 to 1983, females pupped in the equally accessible Sea Lion Cove (Fig. 1), perhaps because of reduced disturbance on Southeast Far- allon Island, although one pup was obser\'ed on Saddle Rock in 1981. Although photogi-aphs from the 1930s show large numbers of Steller sea lions on West End (Huber^), they were not obsei-ved there in recent years until 1983 (one female in the spring). The first pup was born on West End in 1985 ( Huber"' ). Cun-ently, the majority of the popu- lation is found and all pupping occurs at Indian Head and Shell Beach on West End (Fig. 1); both of these areas are inaccessible and difficult to monitor. Statistical analyses Statistical models have been developed that account for effects of obsei-ver, and environmental and survey-related covariates on counts of birds and marine mammals (Link and Sauer, 1997, 1998; Calkins et al., 1999; Frost et al., 1999; Forney, 2000). These models can increase accuracy in estimating and power in detecting population trends by reducing variability in counts and correcting biases in trend that result from methodological changes in sui-vey design over time (such as changes in survey dates), par- ticularly when few surveys are conducted during a stan- dard survey window each year (Calkins et al., 1999; Frost et al., 1999). However, environmental covariates could not be included in the statistical models in our study when the full data set was used because observers recorded the times that sui-veys began and ended on only a few occa- sions prior to 1983 (41 of 569 surveys, or 7.2'^^^f of surveys), such that the majority of data during the first decade of the time series would be excluded. To include the entire •» PRBO (Point Reyes Bird Observatory). 1988. Unpubl. data. (Available from W. J. Sydeman, Point Reyes Bird Observatory, 4990 Shoreline Hwy, Stinson Beach. CA 94970.1 ' Huber, H. R. 1985. Reproduction in northern sea lions on Southeast Farrallon Island, 1973-1985. Final report to the Gulf of the Farallones National Marine Sanctuary, San Fran- cisco. CA, 22 p. [Available from Point Reyes Bird Observatory, 4990 Shoreline Hwy., Stinson Beach, CA 94970.) 54 Fishery Bulletin 100(1) time series, standard regression models, including survey date but excluding effects of environmental covariates, were used to examine seasonal patterns and trends. We believe the exclusion of other covariates during statistical modeling had little effect on trend estimates because sur- veys were conducted consistently over years and over the entire year interval, resulting in large sample sizes (?! = 1134 surveys conducted; range among years 1974- 96: 45 to 52 surveys/year). It is unlikely that population trend estimates were confounded by changes in environ- mental conditions because no obvious annual trends in environmental conditions over the 22 years of the study (weather and tide data were collected daily [at 1000 hours] at Southeast Farallon Island) were apparent, except for a potential increasing annual trend in sea surface tempera- ture (PRBO. unpubl. data^). Seasonal abundance patterns To examine seasonal abun- dance patterns, polynomial regression (Kleinbaum et al.. 1988) was used to fit a cui-ve to counts pooled over years, 1974 to 1996. Data from 1971 to 1973 were excluded because survey methods were not standardized until the end of 1973. We fitted the regression model by first con- verting Julian date to orthogonal polynomial variables (linear combinations of the natural polynomial variables that contain the same information as the natural polyno- mial variables but are uncorrelated to each other) to avoid problems of multicollinearity when using higher-order terms (POeinbaum et al., 1988). Higher-order terms were then added sequentially until the last term was not signif- icant in the model (forward stepwise procedure, P>0.05). We then added year as a variable to the model and tested the year x date interaction to determine if the seasonal pattern varied significantly among years. To examine sea- sonal patterns by sex and age class, polynomial regression curves were fitted separately to counts of adult females, males (adults and subadults pooled), and immature indi- viduals as described above. We excluded surveys in which not all individuals were identified by sex and age class (i.e. all surveys before 1977). Annual abundance trends Because high-order polynomial models were used to address seasonal haulout patterns, annual abundance trends were examined in a separate analysis to simplify results. Seasonal variability in abun- dance was accounted for in annual trend models by using residuals from the regression of Julian date on counts. Assuming e.xponential rates of change, we log-transformed (log^,) the residuals (centered about the mean count) and regressed the transformed residuals against the variable year. Annual rates of change were calculated as el\-^,^^ - 1 X 100%, where Pycar 's ^^^ regression coefficient for annual trend (Caughley, 1977). The following groups were ana- lyzed: 1) all animals, by pooling data over all 12 months and sex and age classes; and 2) each sex and age class, by pooling over a) all months, and b) two periods when peaks in counts were observed for some age classes (the breed- ing [May-July] and late fall through early winter [Sep- tember-December] seasons). Nonlinearity in trend was assessed by using orthogonal polynomials as described earlier in this article. Assumptions of the regression model were verified by visual inspection of residuals. Trends In pup production, reproductive rate, and adult sex ratio during the breeding season We used linear regres- sion to test if the decline in maxnnum pup counts during sui-veys presented in Sydeman and Allen (1999) was sig- nificant. We used only data after 1977, when counts by age class were conducted consistently. Only data from sur- veys conducted from June to July were included because during the fall, the ability to distinguish young-of-the year from immature individuals was difficult and because an influx of nonnative pups may have occurred. For example, in November 1978, five times the number of pups known to have survived the breeding season and an increased number of adult females were observed (PRBO, unpubl. data^). The origin of these young-of-the-year is unknown, but the nearest known pupping areas are Aiio Nuevo Island and the North Farallon Islands. Although Steller sea lions are present at Point Reyes, no pups have been obsei-ved there in the past two decades (Sydeman and Allen, 1999). To examine averages and trends in adult sex ratio and reproductive rate, we used maximum counts of adult fe- males, adult males, and pups during June and July in each year and linear regi'ession to test for annual trends. Re- productive rate was calculated as the maximum count of pups divided by maximum count of adult females. Because not all pups born were observed during surveys, we in- creased the maximum count of pups by 57%, the average amount that maximum pup counts underestimated true pup production from 1973 to 1986 (range: 33-90''^ among years). This average was determined from unpublished data of pup production as determined from daily observa- tions of breeding areas (Huber et al.'^). Results Seasonal abundance patterns When data from all sexes and age classes were pooled, the seasonal abundance pattern was bimodal; one peak in numbers occurred before and during the breeding season (April-July) and another peak occurred from late fall through early winter (October-December; Fig. 2A). The regression model was complex with significant date and higher-order terms (variables date- through date'^); all P<0.001; adjusted r-=0:28. ;i = 1134); the variable date-' was not significant (P>0.65). Counts varied significantly with year (P<0.001) and the seasonal pattern varied sig- nificantly among years (datexyear through date^xyear; P<0.001; adjusted r-=0.61 ). Total numbers during the peak '^ Huber, H. R., D. G. Ainley, R. J. Boekelheide, R. R Henderson, and T. J. Lewis. 1988. Annual and seasonal variation in num- bers of pinnipeds on the Farallon Islands. California (Table 3). Final report to the National Marine Mammal Laboratory, National Marine Fisheries Service. Seattle, WA, 3.5 p. [Avail- able from Point Re.yes Bird Observatory, 4990 Shoreline Hwy., Stinson Beach, CA 94970.1 Hastings and Sydeman Population status of Etimetopias /ubahis at the South Farallon Islands, California 55 JIH) - A ' 200 - '; ■• 100 - - -/.•*-'"v. ■.•>■■. ;■•:•'" ''■.— l-^rf J- '■•,'•■;''■*■■'. ^ f2 30 60 W 120 150 180 210 240 270 300 330 360 30 60 90 120 150 ISO 210 240 270 300 330 360 Julian J.ile S 100 -- B 30 60 90 120 150 180 210 240 270 300 330 360 1 50 100 -- 50 -- D 30 60 90 120 150 180 210 240 270 300 330 360 Julian date Figure 2 Seasonal variation in counts of Steller sea lions at the South Farallon Islands for (A) both sexes and all age classes; (B) adult females; (C) subadult and adult males; and (Dl immature mdividuals and yearlings. Data from 1974 or 1977 to 1996 were pooled. Black dots indicate counts: black lines indicate predicted values from the regression model. Divisions on the .v-axis approximate months. The best regression model for each group included the variables date and date'^ through (A) date^ for total counts (adjusted r-=0.28); (B) date^ for adult females (adjusted r-'=0.22); (C) date^'~ for males (adjusted r'^=0.77); and (D) date*^ for immature individuals (adjusted r2=0.13l. breeding season averaged approximately 100 animals, ranging from 50 to 200 animals; whereas numbers from the late fall through early winter peak were more vari- able, averaging slightly less than 100 and ranging from less than 10 to 300 animals (Fig. 2A). Seasonal patterns varied among sexes and age classes (Fig. 2, B-D). Counts of adult and subadult males peaked only during the breeding season (Fig. 2C), whereas counts of adult females and immature sea lions were bimodal (Fig. 2, B and D). When models including the variables date through date'^ were fitted to data for adult females and immature individuals separately, the seasonal pat- tern differed significantly between the two groups (age class, year, and all interaction terms; all P<0.001). Counts of immature sea lions were less peaked during the breed- ing season than those of adult females and, in contrast to the average adult female pattern, numbers during winter peaked on average slightly higher than during the breed- ing season (Fig. 2, B and D). The seasonal pattern varied significantly among years for all age classes (year and yearxdate interactions for adult females and immature individuals; P<0.001, and for subadult and adult males; P<0.05). Variation in seasonal pattern among years was complex but several general pat- terns could be noted. A gradual shift in the peak breeding season count from the beginning of May in 1974 to the be- ginning to middle of June in 1979 was evident (Hastings and Sydeman"). The late fall-winter peak was very pro- nounced from 1984 to 1986, with maximum counts of 200 to 300 animals (Hastings and Sydeman'), most of which were immature individuals. From 1992 to 1996, the sea- sonal abundance pattern was muted with equal or higher numbers in the winter than in the breeding season (Hast- ings and Sydeman'). Hastings, K. K., and W. J. Sydeman. 1998. Status, seasonal variation and long-term trends in numbers of Steller sea lions, Eumetopias jubatus. at the South Farrallon Islands, California: 1927-1996. Final report to the National Marine Fisheries Ser- vice, Southwest Fisheries Science Center, La Jolla, CA, 30 p. (Available from Point Reyes Bird Observatory, 4990 Shoreline Hwy., Stinson Beach, CA 94970.) 56 Fishery Bulletin 100(1) Table 1 Linear rates of change in counts of Steller sea lions on the South P^arallon Islands by season, sex. and age class. Rate of change per year was calculated by 1) removing the effect of date on counts (i.e. by using residuals from regi'ession of date on counts), and 2) log-transforming (log^ ) the sum of the residuals added to the mean count, mdicates significant trends (P<0.05) from regi'essions. ;! = sample size. Age class change /^> SEi/3,,.J All animals All months Adult females All months Breeding season (May-Jul) Late fall through early winter Males (breeding season! All males Bulls Subadult males Immature sea lions All months Breeding season Late fall through early winter -0.44 -0.0044 0.0027 0.03* 1134 3.16 -0.0321 0.0032 <0.001* 866 .5.89 -0.0607 0.0081 <0.001' 217 2..50 0.0247 0.1637 0.80 280 1.12 0.0111 0.0050 0.03* 217 0.16 0.0016 0.0022 0.63 217 1.94 0.0192 0.0086 0.03* 217 0.55 0.0055 0.0019 0.004* 866 4.51 -0.0461 0.0168 0.007* 217 4.97 0.0485 0.0078 0.60; Table 1). Counts of immature indi- viduals also increased slightly (0.6% per year; Table 1) but significantly when counts from all months were pooled (variables yea/- and year-: P<0.01; ;?=866; Fig. 3D). The in- crease was due to the greater numbers of immature in- dividuals from late fall through early winter in recent years (linear trend=5.0%' per year, Table l;yea7- and year^: P<0.01; ;;=280; Fig. 4D). However, numbers of immature individuals present during the breeding season declined at a rate of -4.5% per year (Table l;year: P<0.01; ;;=217; Fig. 3D). Trends in pup production, reproductive rate, and adult sex ratio during the breeding season Maximum pup count from surveys declined significantly from the mid-1970s to the mid-1980s from 15 to 2-4 pups and has remained low in recent years (year and year-: P<0.003; Fig. 4). After adjusting for pups not seen during surveys, reproductive rates of adult females ranged from 2.0% to 21.2'7( among years, with an average rate of 10.7% (Fig. 4). Although reproductive rate appeared to decline in the 1980s and recovered to 1970s levels in the 1990s, no trend was discernible (year and higher-order terms: P>0.20; Fig. 4). The ratio of adult females to adult males during the breeding season ranged among years from 10.3:1 to 1.8:1, with an average of 5.2:1 (Fig. 4). The ratio of adult females to adult males declined significantly and linearly with vear (P<0.001). Discussion Although the Farallon Islands are an important haulout area for Steller sea lions in California, numbers of ani- Hastings and Sydeman Population status of Eumetopim juhahis at tlie South Faiallon Islands, California 57 3 'J "2 -!< ! I : i ! : 1 n h I i 1 Pi I : rprrrrr 73 75 77 79 81 83 85 87 89 91 93 95 97 £ 2 ? 76 78 80 82 84 86 88 90 92 94 96 Year 4 - ^ 2 ^ oc .4 B 76 78 80 82 84 86 88 90 92 94 96 6 5 4 -5 •) -4 - D 76 78 82 84 86 88 90 92 94 96 Year Figure 3 Annual trends in counts of Steller sea lions at the South Farallon Islands from 1974 or 1977 to 1996 for (A) both sexes and all age classes; (B) adult females; (C) subadult males; and (D) immature individuals and yearlings. Significant trends in counts, after accounting for survey date ( residual centered about the mean count from Figure 2. square-root transformed ). are shown for; all months (light dashed line); only counts during the breeding season (May - July; solid black line); and only counts from late fall through early winter (September-December; solid light black line). Results of significance tests using square-root and log-transformed counts were identical; Linear rates of change from log.-transformed counts are shown in Table 1. mals at the Farallon Islands are currently lower (0.06 of the 1989 statewide count, 0.09 of the count from four major sites) than at the other three major California sites (Alio Nuevo Island, St. George Reef and Sugarloaf Island) which ranged from 0.16 to 0.18 of the 1989 statewide count, and from 0.26 to 0.37 of the count from four major sites (Loughlin et al., 1992). A smaller proportion of the statewide Steller sea lion population has used the Farallon Islands in recent years, compared with population counts in the 1927-30 data, when Farallon animals accounted for 0.11 to 0.14 of the statewide count (Bonnet and Ripley, 1948). Historical pup production at the Farallon Islands is unknown, but both the Farallon Islands and Ano Nuevo Island were identified as the two largest and most impor- tant Steller sea lion rookeries in the state in the early 1920s (Rowley, 1929). Pup production at the South Faral- lon Islands over the past two decades has been very low at <30 pups per year and in the last 10 years, at <10 pups per year. Pup production since the mid-1980s, how- ever, may be underestimated owing to the reduced prob- ability of sighting pups since 1984 when pupping areas shifted to West End Island, which is farther away from the survey vantage points. Pup production at other major sites in California included 117-137 pups at Sugarloaf Island and Cape Mendocino in the early 1980s, 115 pups at St. George Reef in 1994, and 230-243 pups at Ano Nuevo in 1993-94 (Westlake et al., 1997; NMML'). Reproductive rates of Steller sea lions at the South Far- allon Islands were also low; an average of only 0.11 of fe- males present during the breeding season produced pups. This number may be biased low because some immature males may have been included in the adult female count. This ratio is much lower than that for rookeries in Brit- ish Columbia (>0.70, Pike and Maxwell, 1958), Afio Nue- vo, California (average of 0.40 to 0.50 from 1962-1990; Le Boeuf et al.-') and Ugamak Island, Alaska, where ratio of 58 Fishery Bulletin 100(1) pups to females increased from 0.75 to >1.00 from 1968 to 1986 (Merrick et al., 1987). The South Farallon ratio is more typical of pe- ripheral areas of rookeries in Alaska where only 0.01 to 0.09 of females had pups com- pared with main areas of rookeries where ra- tios averaged 0.63 to 0.74 (Withrow, 1982). Similarly, high pup mortality rates observed at the Farallon Islands (average of 0.49 of pups born from February to August, range of 0.33 to 0.90 among years; Huber et al.^) are more characteristic of peripheral areas of rookeries where pup mortality ranged from 0.30 to 1.00 compared with 6.10 to 0.12 at main rookery sites (Withrow, 19821. Rooker- ies had much lower pup mortality rates dur- ing the first two months of life than those ob- served at the Farallon Islands, including Ario Nuevo Island, California (0.10, Gentry, 1970), and sites in Alaska (0.03-0.14, Merrick et al., 1987). The frequency of premature pupping (0.40 of those born; Huber et al.'') is also very high compared with the frequency at rook- eries in Alaska (0.09; Pitcher and Calkins, 1981), Oregon (0.04; Mate, 1973), at Aho Nue- vo Island (0.02; Gentry, 1970). As at Aho Nue- vo, most premature pups are born from F'eb- ruary to May at the Farallon Islands (0.65 born in April with a range of February to May), whereas full-term pups are born from mid-May to late July (Gentry, 1970; Huber^). Causes of the high rate of premature pupping at the Farallon Islands are unknown but may be due to several factors known to cause reproductive failure in pinnipeds, including disease or exposure to pollutants (Gilmartin et al., 1976; Huber''), or a prevalence of young, inexperienced, or malnourished females (Pitcher et al., 1998). A high frequency of abortions has been observed at haulout sites rather than at rooker- ies in Alaska (Pitcher and Calkins, 1981 ). Low pup produc- tion and reproductive rates, coupled with high pup mortal- ity and premature pupping rates, support characterization of the Farallon Islands in recent years as a haulout site or peripheral rookery for this species. Seasonal patterns in counts Seasonal haulout patterns varied significantly among sexes and age classes. Adult and subadult male attendance was highly seasonal and males were present only during the breeding season. In contrast, adult females and immature individuals were present year-round and their numbers peaked twice (breeding season and from late fall through early winter). Many studies reported the absence of adult and subadult males at California rookeries outside the breeding season, including Ano Nuevo Island (Orr and Poulter, 1967) and San Miguel Island ( Bartholomew, 1967), and the presence of females and immature individuals at rookeries year-round (Rowley, 1929; Bartholomew, 1967). At Canadian rookeries, males were also generally absent in the winter, but small numbers of females and young OReproduclive rale D Sex-ratio • Maximum pup count -1 — ' — 1 — ' — 1 — I — I- 76 78 80 82 84 86 88 90 92 94 96 98 "lear Figure 4 Annual variation in reproductive rate, adult sex ratio, and ma.ximum pup count during tlie breeding season (June-.Julyi at the South Farallon Islands, 1977-97. Reproductive rate was defined as the ma.ximum pup count divided by the maximum count of adult females per year. Adult sex ratio was calculated as the maximum number of adult females divided by the maximum number of adult males (bulls) counted per year Signifi- cant (P<0.05) trends are shown for adult sex ratio (dashed black line) and maximum pup count (bold black line). of the year usually remained at rookeries throughout the year (Bigg, 1988). Circumstantial evidence suggests males from California migrate northward or males from South- east Alaska move southward in winter, or both movements take place. Large numbers of males have been seen outside the breeding season off northern California (Fry, 1939), Oregon, Washington (Mate, 1973), and southern Vancouver Island (Bigg, 1988). Total numbers of Steller sea lions are also higher in the winter than in the summer off the Cana- dian coast (Bigg, 1988); some winter haulouts in Canada consist almost exclusively of males (Bigg, 1988). The earli- est evidence for sea lion migrations was provided by the recovery of north-coast native American spearheads from several sea lions killed off southern California in the late 1800s; and in June 1870, a spearhead used by native Alas- kans was found in a large male sea lion at Point Arena, California (Scammon, 1874). Seasonal northward move- ment has also been documented in male California sea lions, which were similarly absent from southern sites out- side the breeding season but which ranged up into Wash- ington and British Columbia during winter (Starks, 1921; Fry, 1939; reviewed by Bartholomew, 1967). In contrast to animals on the Farallon Islands, animals of all age classes and both sexes on Aho Nuevo Island were present in significant numbers only during the breeding season from 1967 to 1990 (Le Boeuf and Bonnell, 1980; Le Boeuf et al.^). Data from 1962 and 1963 indicated a sub- stantial presence of Steller sea lions at Aho Nuevo through the fall and winter (Orr and Poulter, 1965) and therefore the lower numbers and, more recently, near absence of all Hastings and Sydeman Population status of Eumetopias jubatus at the South Farallon Islands, California 59 2()()() -1 1 750 c 3 1500 •r. = I2?0 i 1000 i 750 I 500 - o 250 • Siellcis - hiskinca! icnint-. O Slcllcrs iiiui Cajitoinuins ciinilimcJ - hislonciil coiinls Slcllciv I'RBOcounls D Slclk-rs Miul C.ililomums coinhincd I'RHC.) cmmls - LiniitcJ lumrsr 1 1920 1930 1940 1950 19(i0 1970 19X0 1990 2000 \c.,r Figure 5 Counts of sea lions at the Farallon Islands during the breeding seasons, from 1927 to 1997. Historical counts, 1927-.38: total count from single census per year conducted by boat; includes North and South Farallones and adults and sub- adults only (Bonnet et al., 1938). Historical counts, 1946-70: total count from a single census conducted each year by airplane. Wimp, or boat; includes North and South Farallon Islands and may include pups, subadults and adults. Steller and California sea lions were not distinguished during these surveys. Instead all sea lions north of Point Conception were considered Steller sea lions and those south of Point Conception were considered California .sea lions (Bonnot and Ripley. 1948; Ripley et al., 1962; Carlisle and Aplin, 1971 ). Point Reyes Bird Obsei-vatory counts, 1974-97: maximum total counts during June and July from weekly censuses at South Farallon Islands only (North Farallones excluded); includes pups, immature individuals, subadults, and adults. Means or trends over years are shown for Steller counts only (solid black lines) and for counts of Steller and California sea lions combined (dashed linesA age classes after the breeding season may be a recent phe- nomenon. Similarly. Steller sea lions of various sexes and age classes were present off Humboldt County, California, only from mid-April to September (Sullivan, 1980). Diverse seasonal patterns among sites were also evi- dent in Canada and Alaska. In Canada, animals were usually present year-round on rookeries and numbers peaked during July, whereas year-round haulouts showed no marked seasonal variation and a variety of sexes and age classes were present in winter (Bigg. 1988). Winter haulouts were occupied only in the winter and consisted of either only males or a variety of sexes and age classes (Bigg, 1988). In Alaska, many rookeries were abandoned and some haulouts were occupied only in winter; other haulouts and rookeries were occupied year-round (Ken- yon and Rice, 1961; NMMLM. Major seasonal shifts in distribution were not evident in Alaska, although winter counts were substantially lower than summer counts and there was a greater proportion of animals at haulouts than at rookeries in winter ( NMML' ). The diversity in sea- sonal patterns observed among sites (including rookeries and haulouts) in California and elsewhere has confounded generalizations concerning seasonal haulout patterns, al- though a general shift from rookeries to haulouts in win- ter seems to occur throughout most of the species range. Population status of Steller sea lions in southern and central California Decline from historical numbers Substantial declines in Steller sea lions at the Farallon Islands have been evident since the 19'20s and in recent decades. Numbers declined approximately 75-809( from an average of 600-790 ani- mals from 1927 to 1947 to an average of 150 animals (maximum count) from 1974 to 1997 (Fig. 5). This decline may be overestimated because animals on the North Far- allon Islands have not been included in sui-veys since 1970 and because more animals are likely visible by boat or air than from island-based vantage points (Westlake et al., 1997). However, 85% to 90% of the island is visible from vantage points and therefore effects of incomplete cover- age should be small. Although the decline in numbers was severe between 1938 and 1974, the rate of decline cannot be determined for this period because surveys from this period did not distinguish Steller from California sea lions (Fig. 5). These surveys assumed that all sea lions north of Point Conception were Steller sea lions and that all sea lions south of Point Conception were California sea lions (Carlisle and Aplin, 1971). Assessing the status of Steller sea lions from the 1946-70 CDFG counts has been con- founded by growth in the California sea lion population 60 Fishei7 Bulletin 100(1) over the same period. For example, California sea lions made up only 20% of the total sea hon count at the Faral- lon Islands in 1938 (Bonnot and Ripley, 1948); but by the mid 1970s, California sea lions were twice as numerous as Steller sea lions during June and July (Fig. 51. The role of commercial hai-vest and direct take or ha- rassment of sea lions by humans in this decline is un- certain. Large numbers of sea lions were hunted in Cal- ifornia in the late 1800s for oil, hides, and "trimmings" (which included the whiskers, genitalia, and gall bladder of adult males) that were sold to Chinese markets (Scam- mon, 1874). Hunting sea lions for oil became unprofitable around 1900 because of the reduction in sea lion numbers and the wide-spread availability of petroleum products (Rowley, 1929). A reduced sea lion hai-vest for hides, trim- mings, and (in Mexican waters) pet food, continued until the end of the 1930s when Chinese markets disappeared with the onset of the Japanese-Chinese war and protests were successful in stopping Mexican harvests (Bonnot, 1951). During the same period, although fewer sea lions were taken by sportsman, fisherman, and collectors for museums and zoos, rookery abandonments and population declines still persisted in Oregon and southern California (Rowley, 1929; Bonnot, 1931). An additional cause for these population declines may have been the sea lion hunts that were introduced by commercial fisheries around 1900 to reduce competition for fish (Bonnot, 1937). For example, a bounty was offered for Steller sea lions in the early 1900s in areas north of California (Rowley, 1929; Bonnot, 1931; Bonnot, 1951). Although numbers hai-vested in California are not well documented and the role of harvest in the decline is not obvious, several arguments can be made that declines in Steller sea lions from the 1940s to 1970s were likely not due to effects of hai-vest alone. During the period of com- mercial harvest, Steller numbers appeared stable (Bonnot and Ripley, 1948), whereas the 75-80% decline was evi- dent after 1947, after commercial hunting and collections had ended, although harassment by fisherman continued. After 1947, the California sea lion population increased exponentially throughout the state from 3050 in 1947 to a minimum of 18,047 in 1970 (Bonnot and Ripley, 1948; Carlisle and Aplin, 1971), whereas numbers of Steller sea lions on the Channel Islands and at the Farallon Islands declined from 80% to 100% during this period. Large in- creases in California sea lions were evident after commer- cial hai-vesting ended, even though many more California than Steller sea lions were likely hunted commercially, poached, or captured because of difficulty hunting in the steep, rocky intertidal areas frequented by Steller sea li- ons (Rowley, 1929; Bonnot, 1951). This reasoning suggests that factors in addition to hai-vest have influenced the population decline. Proposed causes include reduction of the prey base due to overexploitation by commercial fish- eries (Ainley and Lewis, 1974), shifts in prey composition due to ocean warming, and competition for food with grow- ing numbers of California sea lions (Bartholomew, 1967). Human disturbance, however, likely played some role in the decline, in respect of which Steller sea lions may be more affected by human disturbance than California sea lions. For example, the large Steller sea lion rookeries at San Miguel Island and at Seal Rocks, just off San Fran- cisco, were abandoned permanently because of harassment and shooting by hunters for sea lion trimmings or by fisher- man (Rowley, 1929). Southeast Farallon Island was inhab- ited by fair numbers of lighthouse keepers and their fami- lies (since the mid- 1800s) and egg hunters (men collecting seabird eggs for sale in commercial markets for human consumption) from the mid-1800s to the mid-1900s. High- est human occupancy occurred during World War II, when over 50 military personnel wore added to the island's pop- ulation (Ainley and Lewis, 1974). Families were removed in 1965 and the lighthouse was automated in 1972, after which time only PRBO researchers remained on the island (Ainley and Lewis, 1974). Despite the designation of the North and Middle Farallon Islands in 1909 and the South Farallon Islands in 1969 as a national wildlife refuge, ha- rassment by fisherman and disturbance from low-flying helicopters was common into the 1970s (Ainley and Lewis, 1974). Heightened human presence in the mid-1900s likely increased the abandonment of Steller sea lions from the is- lands during the period of dramatic decline. Recent population trends Over the last 20 years, the numbers of Steller sea lions on the South Farallon Islands has continued to decline significantly. Numbers of adult females present during the breeding season declined by 5.9% per year from 1977 to 1996, although the rate of decline has lessened since the mid to late 1980s (Fig. 3B). This rate of decline is much higher than the 3.6% per year decline reported for adult females by Sydeman and Allen (1999), who used maximum counts and data from all sea- sons, although rates are similar between the two studies when similar data were used (3.2% per year estimated from our study, when data from all seasons were pooled). These findings demonstrate the importance of accounting for seasonal effects when investigating population trends. The rate of decline of 5.9% per year is similar to the rate of decline obsei-ved during the breeding season in the area of greatest decline in Alaska (from Kiska Island to the Kenai Peninsula), where rates of decline varied from approxi- mately 5% (1975-85 and 1990-94) to 16% (1985-90; York etal., 1996). Numbers of immature individuals present during the breeding season have also declined by 4.5% per year over the past several decades, but an overall net increase in immature individuals on the islands has been apparent owing to increased numbers in the late fall and early win- ter. Numbers of immature individuals on the Farallon Is- lands in the winter were particularly high from 1984 to 1986. Immature individuals have continued to be present in significant numbers during winter in recent years. It is uncertain where these young animals originated from, but overall declines in juvenile counts, coupled with sig- nificant declines in juvenile counts during the breeding season, suggest that increased numbers in winter may represent changes in movement and haulout patterns of juveniles rather than improved juvenile sui-vival in recent years. Increased numbers of subadult males hauled out on the South Farallon Islands during the breeding season in Hastings and Sydeman Population status of Eumetopias jubatus at the South Farallon Islands, California 61 recent years may have resulted from increased emigration or movement of subadult males from Ano Nuevo Island due to increased competition for the declining number of females there. A stable number of adult males, couplfxl with declines in numbers of adult females, has resulted in a significant reduction in the adult male-to-female ratio on the South Farallon Islands during the breeding season in recent years. These results demonstrate that reduced numbers of Steller sea lions on the Farallon Islands in recent years have been driven by reduced numbers of adult females during the breeding season, although reproductive rate and pup mortality rate were stable at this peripheral rook- ery. Patterns were similar at Ano Nuevo, where there were sharp declines in numbers of females and pups during the breeding season but where no trend in reproductive rate w-as apparent from 1962 to 1990 (Le Boeuf et al.'l. How- ever, unlike the Farallon Islands, number of males at Ano Nuevo during the breeding season also declined sharply during the same time period (Le Boeuf et al.'). Although the rate of decline at the Farallon Islands has lessened in recent years, large declines of 9.9*^^ per year for pups and 31.5'~r per year for older animals may have occurred at Ano Nuevo from 1990 to 1993. when negative effects of the 1992 El Nino may have affected estimates from this short time series (Westlake et al., 1997). It is unknown whether reduced numbers of adult fe- males and immature individuals present during the breed- ing season have resulted from reduced survival or chang- es in geographic distribution. Because significant declines in Steller sea lions from historical numbers and over the past several decades have occurred at San Miguel Island. Alio Nuevo Island, and the South Farallon Islands, gi-eater monitoring and protection by state or federal agencies of the southern populations are warranted. Estimates of age- class specific sui-\'ival rates of females are needed to deter- mine if reduced numbers of females are due to increased juvenile or adult mortality. More intensive studies track- ing individual Steller sea lions in California are required to determine if declining numbers indicate a northward shift in the breeding range and to document migratory movements of males and females. Population dynamics and movements of prey of Steller sea lions, dietary overlap with California sea lions, and interactions of sea lions with commercial fisheries in California must be examined to determine natural and anthropogenic causes for changes in sea lion numbers or distribution. Acknowledgments H. R. Huber deserves special recognition for her contri- butions during the early years of our study. Financial support for manuscript preparation was provided by the National Oceanic and Atmospheric Administration. National Marine Fisheries Ser\ice. Southwest Fisheries Science Center under contract 40JGNF600336 to W. J. Sydeman. The Friends of the Farallones. Homeland Foun- dation. Roberts Foundation, Bradford Foundation, and Exxon Corporation also provided funds for data prepara- tion and fieldwork. We are particularly grateful to D. G. Ainley for initiating pinniped studies on the Farallon Islands in 1971. We also sincerely thank Nadav Nur and (irey Pendleton for statistical advice and reviews of the manuscript. We also thank the many obsei-\-ers who have conducted surveys over the past three decades: D. Ainley, G. Ballard, B. Boekelheide, H. Carter, S. Emslie, P. Hen- derson, M. Hester, H. Huber, S. Johnston, J. Lewis, E. McLaren, S. Morrel, J. Nusbaum, J. and T. Penniman, P. Pyle, T. Schuster, J. Walsh, and others. General studies of marine mammals at the South Farallon Islands have been graciously supported over the years by the Marine Mammal Commission. LI.S. Fish and Wildlife Service (USFWS), and the Gulf of the Farallones National Marine Sanctuary. In particular. USFWS and the San Francisco Bay National Wildlife Refuge have provided 28 years of financial, logistical, and moral support; to those involved, we offer sincere gratitude. We also thank the Farallon Patrol for transport to and from Southeast Farallon Island. Michael Rehburg assisted with creating the Far- allon Island map. This manuscript benefited greatly by suggestions from Andrew Trites and several anonymous reviewers. Literature cited Ainley, D. G.. and T. J. Lewis. 1974. The history of Farallon Island marine bird popula- tions. 1854-1972. The Condor 76:432-446. Allen, J. A. 1880. History of the North American pinnipeds. A mono- graph of the walruses, sea lions, sea bears, and seals of North American. Publ. U.S. Geol. Geogr. Surv. 12. 785 p. Bartholomew. G. A. 1967. Seal and sea lion populations of the California Islands. In Proceedings from the symposium on the biolog>' of the California Islands ( R. N. Pliilbrick, ed. i, p. 229-244. Santa Barbara Botanic Garden. Santa Barbara, CA. Bartholomew, G. A., and R. A. Boolootian. 1960. Numbers and population structure of the pinnipeds on the California Channel Islands. J. Mammal. 41:366-375. Bickham, J. W., J. C. Patton, and T. R. Loughlin. 1996. High variability for control-region sequences in a marine mammal: implications for conservation and bio- geography of Steller sea lions (Eumetopias jubatus). J. Mammal. 77:95-108. Bigg, M. A, 1988. Status of the Steller sea lion, Eumetopias jubatus. in Canada. Can. Field-Nat. 102: 315-336. Bonnot, P. 1931. The California sea lion census for 1930. Calif Fish Game 17:150-1.55. 1937. California sea lion census for 1936. Calif Fish Game 23:108-112. 1951. The sea lions, seals and sea otter of the California coast. Calif Fish Game 37:371-389. Bonnot, R, G. H. Clark, and S. R. Hatton. 1938. California sea lion census for 1938. Calif Fish Game 24:415-419. Bonnot. P.. and W. E. Ripley. 1948. The California sea lion census for 1947. Calif. Fish Game 34:89-92. 62 Fishery Bulletin 100(1) Calkins, D. G., D. C. McAllister, K. W. Pitcher, and G.W.Pendleton. 1999. Stellar sea lion status and trend in Southeast Alaska: 1979-1997. Mar Mammal Sci. 15:462^77. Carlisle, J. G., and J. A. Aplin. 1971. Sea lion census for 1970, including counts of other California pinnipeds. Calif Fish Game 57:124-126. Caughley, G. 1977. Analysis of vertebrate populations. John Wiley and Sons, London, England, 2.34 p. Forney, K. A. 2000. Environmental models of cetacean abundance: reduc- ing uncertainty in population trends. Conserv. Biol. 14: 1271-1286. Frost, K. J., L. F. Lowry, and J. M. Ver Hoef 1999. Monitoring the trend of harbor seals in Prince Wil- liam Sound, Alaska, after the Exxon Vatdez oil spill. Mar Mammal Sci. 15:494-506. Fry, D. H. 1939. A winter influx of sea lions from lower California. Calif Fish Game 25:245-250. Gentry, R. L. 1970. Social behavior of the Steller sea lion. Ph.D. diss., Univ. California, Santa Cruz, CA, 113 p. Gilmartin, W. G., R. L. Delong, A. W. Smith, J. C. Sweeney, B. W. De Lappe, R. W. Risenbrough, L. A. Griner, M. D. Dailey, and D. B. Peakall. 1976. Premature parturition in the California sea lion. J. Wildl. Dis, 12:104-115. Kenyon, K. W., and D. W. Rice. 1961. Abundance and distribution of the Steller sea lion. J. Mammal. 42:223-234. Kleinbaum, D. G., L. L. Kupper, and K. E. Muller 1988. Applied regression analysis and other multivariable methods. 2nd ed. Duxbury Press, Belmont, CA, 718 p. Le Boeuf B. J., and M. L. Bonnell. 1980. Pinnipeds of the California Islands: abundance and distribution. In The California Islands: proceedings of a multidisciplinary symposium (D. M. Power, ed. I, p. 475-493. Santa Barbara Museum of Natural History Publications, Santa Barbara, CA. Link, W. A., and J. R. Sauer 1997. Estimation of population trajectories from count data. Biometrics 53:488^97. 1998. Estimating population change from count data: appli- cation to the North American breeding bird survey. Ecol. Appl. 8:258-268. Loughlin, T. R., A. S. Perlov, and V. A. Vladimu-ov. 1992. Range-wide survey and estimation of total number of Steller sea lions in 1989. Mar Mammal Sci. 8:220-239. Loughlin, T. R., D. J, Rugh, and C. H. Fiscus. 1984. Northern sea lion distribution and abundance: 1956-80. J. Wildl. Manag. 48:729-740. Mate, B. R. 1973. Population kinetics and related ecology of the northern sea lion, Eumetopias jubatus. and the California sea lion. Zalophus califormanus, along the Oregon coast. Ph.D. diss., LTniv. Oregon, Eugene, OR. 94 p. Men-ick. R. L., T. R. Loughlin. and D. G. Calkins. 1987. Decline in abundance of the northern sea lion.i?(/me/o- pws jubatus, in Alaska, 19,56-86. Fish. Bull. 85:351-365. Orr, R. T., and T. C. Poulter. 1965. The pinniped population of Aiio Nuevo Island. Cali- fornia. Proc. Cal. Acad. Sci. 32:377-404. 1967. Some observations on reproduction, growth, and social behavior in the Steller sea lion. Proc. Cal. Acad. Sci. 35: 193-226. Pike, G. C, and B. E. Maxwell. 1958. The abundance and distribution of the northern sea lion {Eumetopias jubatus) on the coast of British Columbia. J. Fish. Res. Board Can. 15:5-17. Pitcher, K. W., and D. G. Calkins. 1981. Reproductive biology of Steller sea lions in the Gulf of Alaska. J. Mammal. 62:599-605. Pitcher, K. W., D. G. Calkins, and G. W. Pendleton. 1998. Reproductive performance of female Steller sea lions: an energetics-based reproductive strategy? Can. J. Zool. 76:2075-2083. Ripley W. E., K. W. Cox, and J. L. Baxter 1962. California sea lion census for 1958, 1960 and 1961. Calif Fish Game 48:228-231. Rowley, J. 1929. Lifehistoryofthe sea-lions on the California coast. J. Mammal. 10:1-36. Scammon, C. M. 1874. The marine mammals of the north-western coast of North America. Dover Publications, Inc., New York, NY ( re- print!, 319 p. Starks, E. C. 1921. Notes on the sea lions. Calif Fish Game 7:250-253. Sullivan, R. M. 1980. Seasonal occurrence and haulout use in pinnipeds along Humboldt County, California. J. Mammal. 61:754- 760. Sydeman, W J., and S. G. Allen. 1999. Pinniped population dynamics in Central California: correlations with sea surface temperature and upwelling indices. Mar Mammal Sci. 15:446-461. Westlake, R. L., W. L. Perryman, and K. A. Ono. 1997. Comparison of vertical aerial photographic and gi'ound censuses of Steller sea lions at Ano Nuevo Island, July 1990-1993. Mar Mammal Sci. 13:207-218. Withrow, D. E. 1982. Using aerial surveys, ground truth methodology, and haul out behavior to census Steller sea lions, Eumeto- pias jubatus. M. S. thesis, Univ. Washington, Seattle, WA, 102 p. York. A. E. 1994. The population dynamics of northern sea lions, 1975- 85. Mar Mammal Sci. 10:38-51. York. A. E., R. L. Merrick, and T. R. Loughlin. 1996. An analysis of the Steller sea lion metapopulation in Alaska. In Metapopulations and wildlife conservation (D. R. McCullough, ed.), p. 259-292. Island Press, Washing- ton D.C. 63 Abstract— Tins study rcporis new nilormatioEi about soarobiii iPnoiKitiis spp. ) early life history from samples col- lected with a Tucker trawl (for plank- tonic stat;es) and a beam trawl (for newly settled fishi from the coastal waters of New^ Jersey. Northern scaro- bin, Prionoliis caroliiuis. were much more numerous than striped searobin, P. evotans, often by an order of mag- nitude. Larval Prionotus were collected during the period July-October and their densities peaked during Septem- ber For both species, notochord fle.\ion was complete at 6-7 mm standard length (SLi and individuals settled at 8-9 mm SL. Flexion occurred as early as 13 days after hatching and set- tlement occurred as late as 25 days after hatching, according to ages esti- mated from sagittal microincrements. Both species settled directly in conti- nental shelf habitats without evidence of delayed metamorphosis. Spawning, larval dispersal, or settlement may have occurred within certain estuar- ies, particularly for P. evolans; thus col- lections from shelf areas alone do not permit estimates of total larval produc- tion or settlement rates. Reproductive seasonality of P carolinus and P. evo- tans may vary with respect to latitude and coastal depth. In this study, hatch- ing dates and sizes of age-0 P. caro- linus varied with respect to depth or distance from the New Jersey shore. Older and larger age-0 individuals were found in deeper waters. These varia- tions in searobin age and size appear to be the combined result of intraspecific variations in searobin reproductive sea- sonality and the limited capability of searobin eggs and larvae to disperse. Larval and settlement periods of the northern searobin (Prionotus carolinus) and the striped searobin {P. evolansY Richard S. McBride Marine Field Station Institute of Marine and Coastal Sciences Rutgers University 800 Greal Bay Blvd Tuckerton, New Jersey 08087 Present address: Flonda Marine Research Institute 100 Eighth Avenue SE St Petersburg, Florida 33701 5095 E-mail address rictiard mcbnden fwc stale 11 us Michael P. Fahay Sandy Hook Laboratory Northeast Fisheries Science Center National Manne Fisheries Service, NCAA Highlands, New Jersey 07732 Kenneth W. Able Marine Field Station Institute of Marine and Coastal Sciences Rutgers University 800 Greal Bay Blvd Tuckerton, New Jersey 08087 Manuscript accepted 30 July (2001). Fish. Bull. 100:6.3-73 (2002). Although adult fish assemblages off- shore of the middle Atlantic states are fairly well known (e.g. Edwards, 1976; Colvocoresses and Musick, 1984; Gabriel, 1992), the early life history of many of these same species and the function of shelf habitats as nurs- ery grounds are poorly understood (e.g. Fahay, 1983, 1993; Able and Fahay, 1998). Because year-class strength is believed to stabilize prior to the early juvenile stage, information about the transition from the plankton to ben- thic (i.e. settlement) habitats should contribute to our understanding of the population processes of benthic fishes (Gushing and Harris, 1973; Gampana et al., 1989; Myers and Cadigan, 1993). Settlement is regarded as a dynamic period of early development because mortality rates can differ between pre- and postsettlement life stages (Sale and Ferrell, 1988), dramatic morpho- logical and physiological transforma- tions occur ( Youson, 1988; Markle et al., 1992; McCormick, 1993), and behav- iors become evident that allow for delay- ing settlement until suitable juvenile habitat is found (Cowen, 1991; Sponau- gle and Gowen, 1994). Ultimately, an understanding of the life cycle of any benthic species is constrained if the set- tlement period is not viewed as an inte- gral transition from the planktonic to the adult period. Our study contributes to an under- standing of how fishes use continental shelf habitats as nurseries with an ex- amination of the early life history of the northern searobin, Prionotus caro- linus, and the striped searobin, P. evo- lans. Both are common species in the coastal region between Gape God and Gape Hatteras, but relatively little is known about their early life history ow- ing largely to their low economic impor- tance in relation to the heavily exploit- ed fisheries of this region (McBride et * Contribution 2001-28 of the Institute of Marine and Coastal Sciences, Rutgers Uni- versity, New Brunswick, NJ 08901. 64 Fishery Bulletin 100(1) al., 1998). Both species are known to begin spawning as early as May and to continue spawning into Octo- ber as determined by maturity indices (e.g. Richards et al., 1979; Wilk et al., 1990). Prionotus spp. eggs and larvae are known to be seasonally abundant above the continental shelf and within some estuaries (e.g. Rich- ards et al., 1979; McBnde and Able, 1994) but eggs and lai-vae are difficult to identify to species on a rou- tine basis. Therefore we took advantage of recently reported morphological information (Able and Fahay, 1998) to examine ichthyoplankton collections. Our study was designed to examine how spawning patterns varied between two congeners, but intraspe- cific spawning variation also became evident. A sec- ond goal of our study was to examine settlement — to date not reported for either species. Both species un- dergo flexion and complete fin-ray development at about 6-8 mm SL ( Yuschak and Lund, 1984; Yuschak, 1985; Able and Fahay, 1998). Separation of prehensile rays on the pectoral fin. a major adaptation for ben- thic feeding (Morrill, 1895; Bardach and Case, 1965; Finger and Kakil, 1985), occurs in fish as small as 12 mm SL (Yuschak, 1985). Yet settled juveniles <25 mm SL are rare (Lux and Nichy, 1971; Richards et al., 1979; McBride and Able, 1994), which raises the question of whether Prionotus spp. are competent to settle after completing fin-ray development or whether they common- ly delay settlement. Using a novel combination of sam- pling gears, we collected a continuum of late lai-val and early juvenile Prionotus spp. to examine settlement di- rectly. We report for the first time species-specific larval abundances, distributions, ages, sizes, growth rates, and descriptions of early benthic existence. Materials and methods Collections were made in coastal waters of New Jersey, specifically near Beach Haven Ridge (Fig. 1, Table 1), a prominent sand ridge formation that rises to about 8 m depth and is surrounded by depths of 14-16 m (Stahl et al., 1974). Sampling frequency at two stations, one land- ward and the other seaward of the ridge, was every two to six weeks from July 1991 to November 1992. Two tows of a Tucker trawl ( 1 m-i were made at each station in a double, stepped-oblique fashion. One tow was made from the sur- face to the bottom (three minutes duration) and the other tow was fished from the bottom back to the surface (six minutes). Newly settled juveniles and older fishes were sampled with a 2-m beam trawl in Great Bay estuary, near Beach Haven Ridge, as well as in other habitats (Fig. 1). The data from these stations were arranged in the follow- ing gi'oups: 1) the two principal ridge stations (described above); 2) miscellaneous stations scattered on top of and around the ridge; 3) stations along a transect leading directly offshore from the ridge; and 4) a cluster of stations within nearby Great Bay. Generally, three tows were com- pleted at stations immediately landward and seaward of the ridge, but only two tows were completed at other sta- tions. Beam trawl tows offshore of Little Egg Inlet took 1 2 U4=J kilometers -/ MULLICA "" ,5' RIVER £*■ i\ -, Areata / A' ^ BAY A LITTLE / lil A A EGG / / / BEACH HAVEN/ ^ ridge/ v./ / 39 30' - / O^ JL. A. / / A 39 25' - 74 20' / ' 74" 10' Figure 1 Map of sampling station locations in southern New Jersey, including the main stations at Beach Haven Ridge (landward and seaward: filled circles), other ridge stations (open circles), continental shelf transect stations (filled triangles), and estuarine stations (open triangles). The state of New Jersey, and the study location, are shown in the inset. one minute to complete, but estuarine tows were reduced to 20 or 30 seconds to avoid collecting large volumes of macroalgae, detritus, shell, etc. Sampling occurred during daylight unless otherwise stated. Details of sampling pro- cedures are provided by Hales et al.' Volume or area sampled was calculated by using a flow-meter for ichthy- oplankton collections or a meter wheel for beam trawl collections. Larval density is presented as the geometric mean number of fislVm' for Tucker trawl collections. Juve- nile density is presented as the geometric mean number of fislVm- of sea bottom. Calculations of geometric means follow Sokal and Rohlf ( 1981). The standard length (SL) of all, or at least 20 fish per tow, was measured after the fish were presei-ved in 95'^?- ETOH. The term "lai'va" was used in reference to individu- als collected in Tucker trawl tows. Preflexion larvae were distinguished from flexion lai-vae by the absence or pres- ence, respectively, of cartilaginous urals on the ventral edge of the notochord tip; the development of these urals accompanied flexion of the notochord tip (Kendall et al., 1984 ). Larvae were characterized as postflexion stage once the notochord tip moved anterior to the posterior edge of the hypurals. Daily age was estimated from counts of sagittal otolith mici'oincrements, which were validated as daily by Mc- Bride.- Otoliths with a maximum length less than about Hales. L. S., Jr., R. S. McBride, E. A. Bender, R. L. Hoden, and K.W.Abie. 1995. Characterization of non-target inverte- brates and substrates from trawl collections during 1991-1992 at Beach Haven Ridge (LEO-1.5) and adjacent sites in Great Bay and on the inner continental shelf ofT New Jersey. Techni- cal report (contribution 95-09). 34 p. Institute of Marine and Coastal Sciences, Rutgers, The State University of New Jersey, New Brunswick, NJ. ' McBride, R. S. In review. Spawning, growth, and ovei-wintering size of searobins (Triglidae: Prionotus carolinuK and P. evolans). McBiide et al Ldivdl dnd settlement peiiods of Pnonotus catolinus and P cvolans 65 500 \\n\ were removed and mounted whole on glass slides in immersion oil. Otoliths longer than about 500 pm were mounted in nail polish on a glass slide, sanded with 1500 grit sandpaper along the sagittal plane, and polished with 0.3-pm grinding powder. Immersion oil was used liberally to enhance the clarity of all otoliths, and polarized light aided the viewing of microincrement structure. Micro- increment counts were made with a compound microscope, typically at 400x. Slides were coded and microincrements were counted by one reader on three separate occasions. A constant of 4 days, representing the period between hatch- ing and deposition of the first ring, was added to the mean microincrement count to estimate age since hatching (Mc- Bride2). Preserved (95% ETOH) P. carolinus and P. evo- lans were selected in a stratified i0.5-mm intervals), ran- dom manner to compare ages and lengths. Microincrement counts from this comparative material ranged, based on all individuals, between 07i and 32% of the mean micro- increment count for each otolith (mean=12.0'~f ; 11=41). Pnonotus carolinus were collected in far greater num- bers than P. evolans and they were examined in greater detail. Size and age distributions were initially defined from collections made during the period of peak seasonal abundance (i.e. late September 1991), when a random sample of 34 larvae was selected from a Tucker trawl sam- ple for 23 September. Another sample of juvenile P. caro- linus was selected from a 2-m beam trawl tow on 23 Sep- tember 1991 at a station near the above plankton tow (Table 2). Four final samples were selected from 2-m beam trawl tows set one month later (21-22 October) at four sta- tions along a transect of varying depths. Otoliths from all juveniles collected at these stations were analyzed (i.e. on- ly fish that were mutilated or that had cracked otoliths or otoliths sectioned beyond the core were e.xcluded). Gener- al methods of measuring and staging individual fish, and preparing otoliths, followed that described above. Sagittal microincrements were counted on two (for lai-vae) or three (for juveniles) separate dates by one reader. The range of these microincrement counts, for all individuals, was from 0.0% to 25.0% (mean=9.6%; n = 127) of each mean count. Results Interspecific comparisons Prionotus carolinus were more numerous and occurred more frequently than P. evolans in nearly all collections, typically by an order of magnitude (Table 1). Spawning by both species occurred from at least July to October off- shore of southern New Jersey (Fig. 2). Modal size of larvae generally increased with time, but there were exceptions that indicated a pattern of multiple spawning events. For example in August 1991, modal size for Prionotus spp. and P. carolinus was notably smaller (3-4 mm) than the pre- vious month (5-6 mm) (Fig. 3). Peak larval abundances varied somewhat between years but were highest from July to September. Prionotus carolinus was the smaller but older congener at each developmental stage. Size and stage were com- ■- £ d. U] a ° c- (ft §3 hi CO 3 -r I- o o o o o o O .S ''- O X ^^ T3 Oj T3 o a o O i _C "^ -i X c Cj > P u y-j !-§ o =; a OJ 5 _ C a-, -a o 00 o CO o CD 00 CO '^ o _ o 5 t^ lO CM 00 CO t-- >= !- rM 't (M •II a; 0; p =C S CD ^ '5- Z i CTi c "^ ^ to CT> CM CO c- 00 CO a " 3 2 C^ lO CD 00 C^ Tf IM CO >•! •z !- rt ra o ^ G a> a > > > .,-> o o Q> o CJ o o u 2 t x: Q 2 Q z o Z Z o O C ^ a -^ c "3 c bfi c *-> ^ M .2.S o •^ CO CO 3 CO -3 O < 3 < ^=S 7 « - t) a ^ ^ r- £ o c i_ ,—1 0-1 ,-H (N ,—1 C-1 ,—1 CM o-l r- o c rt (71 (35 (35 05 Oi 05 Ci (T> 05 :2i (35 Oi OJ (T> Ol 05 O^ 05 (35 S3 o 3 rt i 2 « 3 5 -J -^ -2- = 2: 3 X 1 CO _ J ci _ 1 _ C3 - ^ O g £ CO (ft £ u *-• £ (ft £ £ -C Cfi E i3 ^ 1 e g 3. .s g ^ g J £ ^ s ■5 ^ X o g c £ g £ £ e Qj Qj [/} lO (N CO CN CO iTJ CD ci ? C J2 rt e = 1 X tUD n m C ~ data gener ampli c lO ^ I> -+ 00 D -C OQ 1 f- -K 1 1 1 1 ^ bjD C cfl o Ol t— ' t- * ■S s^ ^^ CD *"* 1 o fc B S C t. *j S *" a; M ^ f: ^ t- C 0) o >-, 0) Tj n and b arating Figure ] ~ Si. CO c o 3 9- o ^ Oj n, -a (ft -a _o C3 lis of planl acters for s ntheses). S be c a; > C3 c CO CO -a c b£ u u ■s (LI X c CO i- Qi >> CO m TO t_ ni s CJ CO 0) bo CO t^ ra ^ CO -C T3 p U-1 O O. « ZJ ■1.^ i- c/: CQ o S a 66 Fishery Bulletin 100(1) Table 2 Daily age. size, and hatching dates for planktonic (flexion and postflexion stages) larvae and benthic (settled stage) juveniles of P. caroHnus collected in September and October 1991, offshore of southern New Jersey (see Fig. 8 for station locations). Data are presented as means (±1 standard error), and the range of values is given in parentheses. Larvae were collected with a 1 « 1-m Tucker trawl (0.505-mm mesh) and juveniles with a 2-m beam trawl (6-mm mesh). Date Stage Station n Age (days) Length (mm) Hatching date 23 Sep flexion 0T5 9 15.4 ±0.73 (12-17.5) 5.3 ±0.20 (4.1-6.2) 8 Sep ±0.73 (6Sep-ll Sep) 23 Sep postflexion OT5 25 17.7 ±0.53 (12-23.0) 7.0 ±0.20 (5.7-9.5) 5 Sep ±0.53 (31 Aug-U Sep) 23 Sep settled OT5 23 37.4 ±1.91 24-61.3) 12.2 ±0.44 (8.5-15.8) 17 Aug ±1.91 (24 Jul-30Aug) 21 Oct settled OT2 15 60.9 ±3.06 (46-94.7) 17.5 ±1.42 (12.8-30.4) 19 Aug ±3.06 (18Jul-5Sep) 21 Oct settled 0T5 13 62.2 ±2.81 (52-90.7) 16.8 ±1.. 58 (12.8-35.3) 20 Aug ±2.81 (22Jul-30Aug) 22 Oct settled Sta. C 29 75.1 ±2.71 (54-134.0) 21.9 ±1.44 (13.1-59.4) 8 Aug ±2.71 (10Jun-29Augl 22 Oct settled Sta. E 13 90.0 ±3.20 (69.3-105.3) 26.3 ±0.96 (20.2-33.7) 24 Jul ±3.20 (8 Jul-13Aug) 1 g '*^ ^ '2^ ^ -73.2) 35 30 - 2.5 20 - 15 10 0,5 0.0 ■X - Prionolus spp. — P evolans — P carolinus I • I — I — '—] — — — r-" — • 1 Jul 1 Oct 1 Jan 1 Apr 1 Jul 1 Oct Figure 2 Density (geometric mean number of larvae per 100 m^ |±1 standard error, SE]) of Pnonotiis spp.. P. carolinus, and P. evolans larvae for each cruise near Beach Haven Ridge, based on daylight tows of a Tucker trawl at the land- ward and seaward stations (see Fig. 1). Note break in scale (range of SE bars are given in parentheses). Pnonotus carolinus 50 40 30 20 50] 40 30 20 10 50 40 30 20 10 50 40 30 20 10 July n=64 Pnonotus evolans Pnonotus spp. July ffeSI \L^ 10 100 August 80 n=46 60 EtU 20 September 50 f7=553 40 30 20 10 11^ 2 4 6 8 1012 September rtll 8 10 12 Standard length (mm) Figure 3 Size frequency of Prionolus carolinus, P. evolans, and Pri- onotus spp. from Tucker trawl collections near Beach Haven Ridge during July-September 1991. n = total number of larvae collected. pared for 534 P. carolinus and 81 P. evolans collected with the Tucker trawl during both day and night. Flexion was complete at a larger size for P. evolans than for P. ca/-- olinus (range: 6.7-7.5 mm versus 5.4-6.8 mm SL), and planktonic postflexion P. evolans larvae were captured at larger sizes than postflexion P. carolinus (range: 6.7-11.9 versus 5.4-9.8 mm SL). Pnonotus evolans completed flex- ion at a younger age than P. carolinus (approximately 13 McBnde et a\ Larval and settlement periods of Pnonotus carolinus and P cvolans 67 versus hS days after hatching) \V\g. 4). Both species set- tled as early as 18-19 days after hatching, but this was more characteristic of P. cvolans: most P. canilinus did not settle until 24-25 days old. Both species grew relatively slowly, and approximately linearly, during the larval and early juvenile period (i.e. <0.3 mm/d; Fig. 4, and next subsection). These slow growth rates, combined with the late peak in spawning (i.e. around August), resulted in small body sizes by the onset of win- ter. These smaller body sizes were particularly true for P. carolinus, for which the most pronounced size mode was 10-15 mm SL in autumn 1991 and 1992 (Fig. 5). At this time (i.e. September-December), individuals <50 mm SL constituted 85% of P. carolinus and 52% of P. evolans from all beam trawl tows combined; during autumn a majority of Prionotus spp. were <25 mm SL. At beam trawl stations, densities of P. carolinus were consistently higher than those for P. evolans in both 1991 and 1992 (Fig. 6). Geometric mean densities of age-0 P. carolinus during the peak period of settlement (Septem- ber-October) were much higher in 1991 (8.98 fish 100/m-) than in 1992 (1.56 fish 100/m-'). Geometric mean densities of age-0 P. evolans during September-October were also higher in 1991 (0.32 fish 100/m2) than in 1992 (0.09 fish 100/m-'). These interannual differences were consistent with higher larval densities of both species in 1991 versus 1992 (Fig. 2). Maximum densities of age-0 searobins at a single station reached 28.9 P. carolinus 100/m- and 3.2 P. evolans lOO/m^, both in September 1991. Searobins larger than 150 mm SL were collected infre- quently from June to October; occasionally they were found together with age-0 conspecifics in the same beam trawl tows. Age-0 searobins of both species were collected primarily in continental shelf versus estua- rine habitats during July-December (Fig. 7). Settlement of Prionotus carolinus The seasonality of settlement by P. carolinus, although lasting from at least July to October, 1991, was punc- tuated by a 2-3 week period in September when the vast majority of larvae appeared to settle near Beach Haven Ridge (Fig. 8). Densities of age-0 P. carolinus near Beach Haven Ridge were very low during both July and August (geometric means ranging from 0.0 to 1.1 fish 100/m-). During September, densities increased dra- matically (range: 0.8-7.3 and 0.8-28.9 fish lOO/m^ on September 12 and 23-24, respectively). Individuals were collected at all stations along a depth transect, from 6 to 16 m, in late September Settled, age-0 P. carolinus were still widespread and abundant in late October (0.0-13.3 fish lOO/m^), but they were not collected on 2 December 1991, and on 28 January and 10 March 1992. Collections for 23 September 1991 demonstrate a wide range of P. carolinus developmental stages and ages present at Beach Haven Ridge (Fig. 9A). All flex- ion stages were present (6.5% preflexion, 26.0% flex- ion, and 67.4% postflexion; /? = 169). Planktonic larvae subsampled randomly from a Tucker trawl tow (;!=34; Table 2) had hatched during a two-week period from 9 preflexion/flexion A pelagic/postflexion a settled luveniles □ 15. □ ° - 10 5- - D a -"^--'i-a ° - tU'^^ • CP 6«* 0- 5 10 15 20 25 30 35 40 45 Age (days after tiatctiing) Figure 4 Relationship between daily age and length for Pnon- otus carolinus (open symbols) and P. evolans (filled symbols) for preflexion and flexion stages collected with a Tucker trawl (circles), postflexion stages col- lected with a Tucker trawl (triangles), and settled juveniles collected with a beam trawl (squares). The upper dashed line indicates the approximate size at settlement, and the lower dashed line indicates size at completion of flexion for both species. 10 Pnonotus carolinus Summer. 1991 r7=63 .rprm Pnonotus evolans 20 10 69.7 Autumn, 1991 n=267 Summer, 1991 n=^ Autumn, 1991 n=39 Q- 10 Summer, 1992 n=52 T1 , ,n r|ff][|l1lr]lln}i 85.7 10 Autumn, 1992 n=42 50 100 150 200 50 Standard length (mm) Figure 5 Size frequency of Prionotus carolinus and P. evolans for beam trawl collections near Beach Haven Ridge (daylight tows at the landward and seaward stations (Fig. 11). Data were pooled by season: summer (May-August) and autumn (September-Decem- ber), n = total number of fish collected. No data for January- April 1992 are shown because only a single fish (P. carolinus; 4.5 mm SL) was collected at these stations during this period. 68 Fishery Bulletin 100(1) August 31 to September 11. Individuals collected by beam trawl on the same day (23 September 1991) had hatched about 2 weeks earlier (from 24 July to 30 August) than the above larvae (Fig. 9). These juveniles appeared to settle as young as 24 days after hatching and at sizes as small as 8.5 mm SL (Table 2). The total hatchmg date distribu- tion for both larvae and newly settled juveniles collected on September 23 reflected a spawning period that ranged from late July to early September and that peaked in late August and early September. Settled juveniles with a similar hatching date distribu- tion were identifiable one month later at stations near Beach Haven Ridge, but not at stations farther offshore (Fig. 9, B and C). Fish collected near Beach Haven Ridge on 21 October 1991 had a hatching date distribution with a mode from late August through early September and the overall distribution was skewed to the left. This period was similar to the hatching date distributions for larvae and newly settled fish collected on 23 September 1991. In contrast, fish collected from offshore stations (i.e. stations C and E) on 22 October 1991 were 2-4 weeks older and 5-10 mm larger on average (Table 2, Fig. 9). Plots of P. carolinus size versus age did not indicate any abrupt change at settlement, specifically for postflex- ion lai^vae and settled juveniles collected on 23 Septem- ber 1991 (Fig. 10). Growth rates for this September collec- tion fitted a hnear model (SL=3. 24-1-0. 229[age|; r2=0.77). Because Prionotus lai-vae hatch at about 3 mm SL (Yus- chak, 1985), this model's y-intercept is biologically realis- tic. Growth rates offish collected in October did not differ significantly between stations (ANCOVA:p7-o6.^, .^=0.13, pro6.,,,,p ,=0.51); therefore the data were pooled. Linear, least squares regression of all data produced an unreal- istic y-intercept (SL=-7.01-H0.382lage]: SE^=2.0; ;--=0.74). This model was rerun after restricting the y-intercept to 3 mm and the resulting equation indicated that age-0 P. ca/'olinus continued to grow at about 1 mm every 4 days (SL=3 +0:25l[age\: ;-=0.65) as they had during the lai-val and settlement period. Size and age of P. caro/i>H^s juveniles varied significantly along a 12-km transect ( 12-20 m depths; Fig. 1 ). The linear relationship: Hatching age = 17.8 -i- 3.43 x depth; r-=0.35, P<0.01; ri=69) showed that for every two meters change in depth offshore the fish collected were about one week older on average (Fig. 11). Sampling in both 1991 and 1992 showed a consistent trend for larger (and presumably old- er) fish to be collected in deeper water in October and No- vember (Fig. 12). After accounting for the effects of depth, or possibly the distance from shore, it appeared that fish reached a larger size in October of 1991 than in 1992 or that larger fish in 1992 were not found in the sampling area. Discussion Spawning grounds and seasonality of spawning Prionotus caro/inus are more abundant than P. evolans in continental shelf habitats whether they are measured as eggs, larvae, juveniles, or adults (Keirans et al., 1986; 1601 120- 8,0 4.0 00 Prionotus carolinus Age-0 Age-1 + ^ u - Prionotus evolans r —•- Age-0 1 5- —A — Age-1 + T 1 0- <► \ ' T 0,5- J n m UU- •I* 1 ' ' 1 • 1 • 1 • 1 1 Jul 1 Oct Figure 6 Density (geometric mean number of fish per 100 m- |±1 standard en-or | ) of different cohorts of postsettlcnient Prionotus carolinus and P. evolans during daylight tows at the landward and seaward stations near Beach Haven Ridge. Note scale differences for each species. McBride and Able, 1994; Able and Fahay, 1998; McBride et al., 1998; our present study). The low numbers of P. evo- lans observed in our study may be biased somewhat by our focused effort to sample the continental shelf rather than estuaries. Prionotus evolans reside in shallower, warmer habitats than do P. carolinus during the spawning season (McBride and Able, 1994). If P. evolans spawn to some degi'ee in shallower waters or estuarine habitats, then this would at least partly explain the generally low abundance of P. evolans early life stages in our collections. In general, we expect that larval distributions are good predictors of spawning locations for both Prionotus spe- cies because of the short (i.e. about three weeks) lai-val dispersal periods of these species (e.g. Houde and Zastrow 11993] reported several shelf species with planktonic du- ration >100 days). In some coastal areas, the distribution of Prionotus spp. eggs and larvae indicates that spawning may be limited to estuaries; however, the abundance of Pri- onotus larvae offshore of New Jersey suggests that spawn- ing by these species occurs outside estuaries as well. For example, Merriman and Sclar ( 1952) did not find Prionotus McBride et al : Larval and settlement periods of PnonottK camlinus and P evolans 69 03 3.0 20 1 0.0 100- 90 80' 70 60 50 40 30 2.0 1 00 Jul-Aug 1991 _Qd_ -OiL Sep-Dec 1991 i 30 2,0 1,0 00 100 90- BO- ZO 60 50 40 30 20 1 00 Main ridge Other ridge Transect Estuary stations stations liw. Mam ridge May-Aug 1992 Sep-Nov 1992 K. Ottier ridge Transect Estuary stations stations Figure 7 Density (geometric mean number offish per 100 m^ |±1 standard error]) of age-0 Pnonotus carnlinuf: (open bars) and age-0 P. evolans (filled bars) collected with a beam trawl from four major station groups ( nd= no data; 0=sampling occurred but no Pnonotus were collected). See Figure 1 and Table 1 for station groupings and locations. 74°20' / 74°10' 1 24 July 1991 '/I \ - 39°30' / E 20. V ^/ - 39 25' / o o • 10. 10 1 /.- 74°20' # / ! 1 / 74°10' 1 14-15 August 1991 / / / - 39°30' / / / / / ./ - 39 25' / - 39°30' 1 — /T- 74°20' /74°10' 21-22 October 1991 ^ fl 's 1 w / 1 1 74°20 / 74 "1 0' 12 September 199/ / - 39° -■ i / / y>sf ^ /■ / 'K 1 # / «t ; / ^ / J -39° 25' / / Figure 8 Densities (geometric mean [±1 standard error) ) of agc-0 Pnonotus carolmus collected with a beam trawl during six consecutive cruises (July-December 1991 1. Number of stations varied between cruises. The scale bar indicates 10 fish/100 m-. 70 Fishery Bulletin 100(1) 100 80. 60 40 20 \ Pelagic larvae & benthic luveniles (at OT5) 23 September 1991 flexion larva (n=9) E3 postflexion larvae (n=25) n benttiic luveniles (n=23) n n n Jl _ 1 ' f ^° -■ B Benttiic juveniles - near Beacti Haven Ridge 21 October 1991 ^ Station 0T2(n=1 5) nStationOTS (n=^3) 40 30- 10. 50-, 40 30 20. 10. Q Benttiic juveniles - offstiore of Beacti Haven Ridge 22 October 1991 Station C (n=29) D Station E{n=13) H m Jun 1 Jul 1 Aug 1 Sep 1 Ocl 1 Hatctiing date Figure 9 Hatching-date distributions for larval and newly settled juvenile Pnonotus carolinus collected 23 September 1991 (A) and for juve- niles collected on 21 October 1991 (B) and 22 October 1991 IC). See Figures 1 and 8 for locations of sampling stations. 60-, + Sept - Flexion larvae » Sept - Postflexion larvae 50- • Sept - Benthic luveniles A Oct -Benthic juveniles 1 40. i 30. ■D en 1 20. CO 10- 0- ^ A . .v.. ■• ( ) 20 40 60 80 100 120 140 Age (days after hatching) Figure 10 Age (days after hatching) and standard length (mm) for flexion, post- flexion, and juvenile Pnonotuf; carolinus collected seaward of Beach Haven Ridge on 23 September 1991 and 21-22 October 1991. spp. eggs or larvae in Block Island Sound, and Able and Fahay ( 1998) did not observe Prionotus larvae above the continental shelf north or east of Hud- son Canyon, New York. Instead there are many reports of Prionotus eggs, larvae, and juveniles in southern New England estuaries, specifically in Long Island Sound (Wheatland, 1956; Richards, 1959; Williams, 1968; Richards et al., 1979) and Narragansett Bay (Herman, 1963; Bourne and Go- voni, 1988; Keller et al, 1999). Thus, the relative importance of estuaries versus shelf habitats as spawning grounds for Prionotus may vary in other regions compared with our results for New Jersey. Nonetheless, Prionotus spawning seasonality ap- pears to follow a pattern similar to that of other species with a wide latitudinal range that have a shorter spawning season at higher latitudes (e.g. Conover, 1992) and that spawn later in the south (e.g. Barbieri et al., 1994). An important departure from this general trend is that Prionotus repro- ductive seasonality may vary not only with respect to latitude but along an estuary-shelf gradient as well. Because adults of both Prionotus species enter estuaries early in the spring and migrate back out to the shelf in summer (McBride and Able, 1994), we postulate that spawning occurs first in estuaries at a given latitude. In support of this hypothesis are the collective results from our study and other published reports. After the summer spawning peak within estuaries such as Chesapeake Bay and Long Island Sound, Priono- tus spawn during August and September offshore of Chesapeake Bay and New Jersey. In contrast, spawning does not continue into late summer off- shore of southern New England (Pearson, 1941; Richards et al., 1979; Able and Fahay 1998). To explain this potentially novel spawning pat- tern does not require any new controlling mecha- nism other than that used to explain spawning by other coastal fishes of the region. Temperature and photoperiod are known to influence spawn- ing activity in fishes (Burger, 1939) and may in- fluence spawning seasonality of searobins. Water temperatures offshore of the middle Atlantic sea- board are known to fluctuate widely both tem- porally and spatially (Colvocoresses and Musick, 1984) and this fluctuation affects the spawning pattern of many species. For example, a simple south to north progression of spawning activity above the shelf is evident for Centropristis stri- ata (Able et al., 1995) and Scophthalmus aquo- sus (Morse and Able, 1995). For Prionotus, how- ever, we propose that spawning seasonality is controlled by an interaction between latitudinal and estuarine gradients of temperature (i.e. earli- er spawning in estuaries occurs because of earlier warming of these shallow embayments). Temper- ature has already been shown to affect the distri- bution of Pr-ionotus adults along both latitudinal and estuarine gi-adients (McBride and Able, 1994; McBride et a\ Larval and settlement periods of Pnonotiis carolinwi and P evolam McBride et al.. 1998). Because few other .species use both estuarine and shelf habitats lor spavvninj^. such pallcrns arc not commonly obser\i'(l. Linking metamorphosis and settlement Prionotus carollniis are otten ranked as among the most abundant species in regional trawling surveys for adult fish or plankton sui-veys for larval fish (McBride and Able, 1994). The results from the small-mesh beam trawl used in our experiment demonstrate that the juvenile stages of P. caroliniift are also very abundant offshore. Age-0 Pri- onotus spp. are found in estuaries, as discussed above for the southern New England region, but our observation of high densities of juvenile Prionotus in shelf habitats off- shore of New Jersey suggest that neither species requires estuarine nursery habitats during their life cycle. Most searobins complete their life cycle in continental shelf hab- itats (Hoff, 1992), with the notable exception of P. scitulus whose young are concentrated in lower salinity, estuarine habitats (Ross, 1978). Our findings of age and size at settlement largely agree with Yuschak and Lund's (1984) and Yuschak's (1985) de- scriptions of early development of cultured specimens. The developmental rate of cultured specimens of P. carolinus and P. evolans-^ did not differ notably from our obsei-va- tions of field-collected individuals, which further supports our conclusion that neither species delays settlement. Pri- onotus evolans. cultured at 20°C, were all at prefiexion and flexion stages after 11-13 days; they were at flexion and newly postflexion stages after 18-20 days; and all were at the postflexion stage at 25 days. All available P. carolinus specimens, cultured at 15°C, were less than 20 days old; they followed a similar, if slightly slower, rate of development compared with P. evolans. Fin-ray development, which greatly facilitates locomotion, is complete in postflexion individuals. Prehensile, chemosen- sory pectoral rays, which would facilitate benthic feeding, are completely separated by 11.5 mm SL. Thus, on the basis of cultured and field-caught specimens, both species are well developed (i.e. are similar to adults) and well-suited for a bottom- feeding and swimming life style as they complete flexion. Other species delay metamorphosis to set- tle during favorable lunar phases (Sponaugle and Cowen, 1994), but settlement by Prionotus was so concentrated in a single month (i.e. September) that we could not test spawning or settlement cy- cles in more than one month. Nevertheless, be- cause larvae o{ Prionotus species commonly bury themselves in loose substrate (Bardach and Case, 1965), and this material was common to all our sampling stations (Hales et al.M, competent lar- vae are likely not habitat limited (one mechanism identified with the delay of settlement). The variation of P. carolinus sizes and ages along a depth gradient could be caused by one or a combination of three processes. There could be differential larval survi- vorship, juvenile movements, or adult reproduction rates a 100- c u 90 Z 70. S 60. A i < 50- 12 13 14 15 16 17 18 19 20 30-, Standard length (mm) B I 1 1 12 13 14 15 16 17 18 19 20 Depth (m) Figure 11 Ago (A) and size (mean ±1 standard error) (B) of juvenile Prionotus carolinus plotted in relation to depth. Data are for fish collected on 21-22 October 1991 (near and offshore of Beach Haven Ridge: sta- tions 0T2, 0T5, C, and E) and aged by using sagittal microincrements (see Table 2 and Fig. 9 for sample sizes). 30-, E 25. o September 23-24, 1991 o October 21-22, 1992 October 27, 1992 ▲ November 10, 1992 , 1 i en .§ 20. ■D ro •o |1 T [ ro or, 15. 10 « (-. 1 ) 10 15 20 25 Depth (m Figure 12 Standard length (mm; mean [±1 standard error]) of juvenile Pnono- | tus carolinus in relation to depth for four separate cruises in 1991 (open symbols) and 1992 (filled symbols). AH benthic juveniles col- lected were included in these calculations This material was cultured by P. Yuschak (see Yuschak and Lund 119841 and Yuschak 11985]) and has been examined by R.S.M. 72 Fishery Bulletin 100(1) across the shelf to account for this spatial pattern of older and larger juveniles farther offshore. The last process was identified earlier as potentially important. Testing and eliminating these hypotheses, however, requires spatially explicit lan'al distribution data, in addition to the benthic data that we collected, which would allow a comparison of pre-and postsettlement distributions with abundance of Prionotus propagules. Spatially explicit environmental data would also be useful because we obsei-ved dynamic changes in the physical parameters in our sampling area, and we suspect that these could affect Prionotus sui-vival. Vertical stratification of the water column was noted near Beach Haven Ridge on 14 August 1991, but not during cruises in September or October 1991. Low dissolved oxy- gen levels near the bottom of the ridge, at about 3 ppm (in contrast to >8 ppm in the upper water column) could have negatively affected settlement rates near the ridge in 1991. Stratification near the ridge was also noted in 1992 with similarly depressed levels of dissolved oxygen (Hales et al.^). Low dissolved oxygen offshore of New Jersey is not uncommon (Falkowski et al., 1980; Glenn et al., 1996) and may be another process that can contribute to geographic variations in size and age of Prionotus species offshore of the middle Atlantic states. We postulate that spatially ex- plicit patterns of reproductive seasonality and age-0 fish size for P. carolinus and P. evolans within coastal waters offshore of the middle Atlantic states are related to each other because the short planktonic larval durations for both species limit larval dispersal. Interannual variations in water temperature or vertical stratification of oxygen concentrations may be proximate causes for these geo- graphic variations of reproductive seasonality and age-0 size. These patterns could be somewhat unique to searo- bins, compared with other regional fishes, because searo- bins use both estuarine and shelf habitats for spawning. Acknowledgments R. Cowen, J. Hare, J. P. Grassle, R. Loveland, and C. L. Smith contributed thoughtful discussions and helpful com- ments on earlier drafts. S. Richards provided cultured specimens of searobins and miscellaneous data that had been used in P. Yuschak's research. This study was part of a doctoral dissertation (R. S. M.) and was supported by the Institute of Marine and Coastal Sciences (IMCS), Rutgers University Marine Field Station, Anne B. and James H. Leathem Fund, Manasquan Marlin and Tuna Club, NOAA National Undersea Research Program for the Middle Atlantic Bight, and NOAA Sea Grant Program. Final preparation and revision of this manuscript were made while the senior author was employed by the Florida Marine Research Institute. We thank all of the above. Literature cited Able, K. W.. and M. P. Fahay. 1998. The first year in the life of estuarine fishes in the Middle Atlantic Bight. Rutgers Univ. Press, New Bruns- wick, NJ, 342 p. Able. K. W, M. P Fahay, and G. R. Shepherd. 1995. Early life history of black sea baas, Centropristis stri- ata, in the mid-Atlantic Bight and a New Jersey estuary. Fish. Bull. 93:429-445. Barbieri, L. R., M. E. Chittenden Jr., and S. K. Lowerre-Barbieri. 1994. Maturity, spawning, and ovarian cycle of Atlantic croaker, Micropof;oniaf> undulatus. in the Chesapeake Bay and adjacent coa.stal waters. Fish. Bull. 92:671-685. Bardach, J. E., and J. Case. 1965. Sensory capabilities of the modified fins of squirrel hake (Urophycis chuss) and searobins iPnoiidtus carolinus and P. evolans). Copeia 1965:194-206. Bourne, D. W., and J. J. Govoni. 1988. Distribution of fish eggs and lai-vae and patterns of water circulation in Narragansett Bay, 1972-1973. Am. Fish. Soc. Symp. 3:132-148. Burger, J. W. 1939. Some experiments on the relationship of the external environment to the spermatogenic cycle of Fiinduhis het- eroclttus. Biol. Bull. 77:96-103. Campana, S. E.. K. T Frank. P. C. F Hurley. R A, Koeller, F H. Page, and P C. Smith 1989. Survival and abundance of young Atlantic cod ( Gadus morhua ) and haddock (Melanogrammus aeglefinus) as indi- cators of year-class strength. Can. J. Fish. Aquat. Sci. 46(suppl. 11:171-182. Colvocoresses. J. A., and J. A. Musick. 1984. Species associations and community composition of Middle Atlantic Bight continental shelf demersal fishes. Fish. Bull. 82:29.5-313, Conover, D. O. 1992. Seasonality and the scheduling of life history at dif- ferent latitudes J. Fish Biol. 41lsuppl. B):161-178. Cowen, R. K. 1991. Variation in the planktonic lai-val duration of the tem- perate wrasse Semicossyphus pulchcr. Mar. Ecol. Prog. Ser 69:9-15. Cushing, D. H., and J. C. K. Hams. 1973. Stock and recruitment and the problem of density de- pendence. Rapp. P.-V. Reun. Cons. Perm. Int. Explor Mer 164:142-155. Edwards, R. L 1976. Middle Atlantic Fisheries: recent changes in popula- tions and outlook. In Middle Atlantic continental shelf and the New York Bight: proceedings of the symposium, 3-5 November 1975. Special symposium, vol. 2 (M. G. Gross. Lawrence, ed.), p. 302-311. Am. Soc. Lininol. Oceanogr. Lawrence, KS Fahay, M. P 1983. Guide to the early stages of marine fishes occun-ing in the western North Atlantic Ocean, Cape Hatteras to the southern Scotian Shelf J. Northwest Atl. Fish. Sci. 4:1-423. 1993. The early life history stages of New Jersey's saltwa- ter fishes: sources of information. Bull. N.J. Acad. Sci. 38:1-16. Falkowski, R G., T S. Hopkins, and J. J. Walsh. 1980. An analysis of factors affecting o.\ygen depletion in the New York Bight. J. Mar. Res. 38:479-506. Finger, T. E., and K. Kakil. 1985. Organization of motorneuronal pools in the rostral spinal cord of the sea robin, Prionotus carolinus. J. Comp. Neurol. 239:384-390. Gabriel, W.L. 1992. Persistence of demersal fish assemblages between Cape Hatteras and Nova Scotia, Northwest Atlantic. J. Northwest Atl. Fish. Sci. 14:29-46. McBride et a\ : Larval and settlement periods of Pnonotus carollnus and P evolans 73 Glenn. S. M., M. F. Crowley, D. B. Haidvogel, and Y. T. Song. 1996. Underwater observatory captures coastal upwclling events off New Jersey. Eos 77:233-236. Herman. S. S. 1963. Planktonic eggs and larvae of Narragansett Bay. Lininol. Occanog. 8:103-109. Hoff.J. G..Jr. 1992. Comparative biology and population dynamics of scarobins (Genus Prionotus) with emphasis on populations in the northwestern Gulf of Me.xico. Ph.D. diss.. Virginia Institute of Marine Science. College of William and Mary, Gloucester, VA, 229 p. Houde, E. D., and C. E. Zastrow. 1993. Ecosystem- and taxon-specific dynamic and energet- ics properties of larval fish assemblages. Bull. Mar. Sci. 53:290-335. Keirans. W. J.. Jr., S. S. Herman, and R. G. Malsberger 1986. Differentiation of Prionotus carolinus and Prionotus evolans eggs in Hereford Inlet Estuary, southern New Jersey, using immunodiffusion. Fish. Bull. 84:63-68. Keller, A. A., G. Klein-MacPhee, and J. St. Onge Burns. 1999. Abundance and distribution of ichthyoplankton in Narragansett Bay Rhode Island, 1989-1990. Estuaries 22:149-163. Kendall, A. W. Jr., E. H. Ahl.strom, and H. G. Moser 1984. Early life history stages of fishes and their character- istics. In Ontogeny and systematics of fishes (H. G. Moser. ed.), 11-22 p. Am. Soc. Ichthyol. Herpetol. Lux, F. E., and F. E. Nichy 1971. Number and lengths, by season, of fishes caught with an otter trawl near Woods Hole. Massachusetts. September 1961 to December 1962. Spec. Sci. Rep. Fisheries 622. Wash- ington. D.C., National Marine Fisheries Service. NOAA. Markle. D. F. P. M. Harris, and C. L. Toole. 1992. Metamorphosis and an ovei-view of early-life-history stages in Dover sole Microstomus pacificus. Fish. Bull. 90:285-301. McBride. R. S., and K. W. Able. 1994. Reproductive seasonality, distribution, and abundance ofPriono/u.s caro/(n us and P. er!o/ans( Pisces: Triglidae) in the New York Bight. Estuarine Coastal Shelf Sci. 38:173-188. McBride. R. S.. J. B. O'Goi-man. and K. W. Able. 1998. Seasonal movements, size-structur'e. and interannual abundance of searobins (Triglidae: Pnonotus) in the tem- perate, northwestern Atlantic. Fish. Bull. 96:303-14. McCormick, M. I. 1993. Development and changes at settlement in the barbel structure of the reef fish. Upeneus tragula (Mulli- dae). Environ. Biol. Fishes 37:269-282. Merriman. D., and R. C. Sclar. 1952. The pelagic fish eggs and larvae of Block Island Sound. Bull. Bingham Oceanog. Coll. 13:16.5-219. Morrill. A. D. 1895. The pectoral appendages of Pnonotus and their innci'- vation. J. Morphol. 11:177-192. Morse. W. W.. and K. W. Able. 1995. Distribution and life history of windowpane. Scoph- thalmus aquosus. olf the noi'theastern United States. Fish. Bull. 93:675-693. Myers. R. A., and N. G. Cadigan. 1993. Density-dependent juvenile mortality in marine de- mersal fish. Canadian J. Fish. Aquat. Sci. 50:1576-1590. Pearson, J. C. 1941, The young of some marine fishes taken in lower Ches- apeake Bay, Virginia, with special reference to the gray sea ti-Qut, Cynoscion regal is iBloch). Fish. Bull. 50:97. Richards, S. W. 1959. Pelagic fish eggs and larvae of Long Island Sound. Bull. Bingham Oceanogr. Coll. 17:95-124. Richards, S. W., J. M. Mann, and J. A. Walker. 1979. Comparison of spawning seasons, age, growth rates, and food of two sympati'ic species of searobins, Prionotus carolinus and Prionotus evolans, from Long Island Sound. Estuaries 2:255-268. Ross, S. T 1978. Trophic ontogeny of the leopard searobin, Prionotus scitulus (Pisces: Triglidael. Fish. Bull. 76:225-234. Sale, P F., and D. J. Ferrell. 1988. Early survivorship of juvenile coral reef fishes. Cor- al Reefs 7:117-124. Sokal. R. R..andRJ. Rohlf 1981. Biometry W.H. Freeman and Co., New York, NY, 859 p. Sponaugle, S.. and R. K. Cowen. 1994. Larval durations and recruitment patterns of two Caribbean gobies (Gobiidae): contrasting early life histo- ries in demersal spawners. Mar Biol. 120:13.3-143. Stahl. L.. J. Koczan. and D. Swift. 1974. Anatomy of a shoreface-connected sand ridge on the New Jersey shelf implications for the genesis of the shelf surficial sand sheet. Geology. 2:117-120. Wheatland. S. B. 1956. Oceanography of Long Island Sound. 1952-54. VII. Pelagic fish eggs and larvae. Bull. Bingham Oceanogr Coll. 15:234-314. Wilk. S. J., W. W. Morse, and L. L. Stehlik. 1990. Annual cycles of gonad-somatic indices as indicators of spawning activity for selected species of finfish collected fi-om the New York Bight. Fish. Bull. 88:775-786. Williams. G. C. 1968. Bathymetric distribution of planktonic fish eggs in Long Island .Sound. Limnol. Oceanogr. 13:382-385. Youson, J. H. 1988. First metamorphosis. In Fish physiology, vol. lib (W. S. Hoar and D J. Randall, eds.), p. 13.5-196. Academic Press, San Diego, CA. Yuschak, P. 1985. Fecundity, eggs, larvae and osteological development of the striped searobin, Prionotus evolans (Pisces, Trigli- dael. J. Northwest Atl. Fish. Sci. 6:65-85. Yuschak, P., and W. A. Lund. 1984. Eggs, larvae and osteological development of the northern searobin, Prionotus carolinus (Pisces, Triglidae). J. Northwest Atl. Fush. Sci. 5:1-15. 74 Abstract— In trawl surveys a duster of fish are caught at each station, and fish caught together tend to have more similar characteristics, such as length, age, stomach contents etc., than those in the entire population. When this is the case, the effective sample size for estimates of the frequency distri- bution of a population characteristic can, therefore, be much smaller than the number of fish sampled during a survey. As examples, it is shown that the effective sample size for estimates of length-frequency distributions gen- erated by trawl surveys conducted in the Barents Sea, off Namibia, and off South Africa is on average approxi- mately one fish per tow. Thus many more fish than necessary are measured at each station (location). One way to increase the effective sample size for these sui"veys and, hence, increase the precision of the length-frequency esti- mates, is to reduce tow duration and use the time saved to collect samples at more stations. Assessing the precision of frequency distributions estimated from trawl-survey samples Michael Pennington Institute of Marine Research Department ol Marine Resources Nordnesgaten 33 N-5005 Bergen, Norway E mail address michaeliaimrno Liza-Mare Burmeister Ministry of Fishenes and Marine Resources of Namibia, NatMIRC PO Box 912 Swakopmund, Namibia Vidar Hjellvik Institute ol Manne Research Department of Marine Resources Nordnesgaten 33 N-5005 Bergen, Norway Manuscript accepted 21 August 2001. Fish. Bull. 100:74-80 (2002). Survey-based assessments often appear to provide a more accurate prognosis of the status of a fish stock than catch- based assessments (Nakken, 1998; Pen- nington and Stromme, 1998: Kors- brekke et al., 2001). Aji advantage that sui-vey-based assessments have over those based on commercial catch sta- tistics is that the uncertainties asso- ciated with survey estimates can be studied and quantified, and based on such research, survey methods, and ultimately stock assessments, can be improved (Godo, 1994). In contrast, it is generally difficult to determine either the accuracy or the precision of esti- mates based on commercial catch data, and it is not clear how to improve, at a reasonable cost, the collection of catch data so that these data would more accurately reflect the mortality caused by fishing (Christensen, 1996). Trawl surveys provide estimates of the abundance or relative abundance of a fish stock and estimates of the relative frequency of various population charac- teristics, such as length, age, and stom- ach contents. In our study we examined the precision of survey-based estimates of the length-frequency distributions of cod and haddock in the Barents Sea, hake off South Africa, and hake off Namibia. The focus was on length, but the results are relevant for es- timating the frequency distribution of other population characteristics. Materials and methods Survey length data Bottom trawl survey length data for Northeast Artie cod ^Gadun mohua) and Northeast Arctic haddock (Mela- nogrammus aeglefinusY were collected during the Institute of Marine Research (Norway) winter and summer surveys in the Barents Sea. The surveys were stratified systematic surveys and at each station the trawl was towed for 30 minutes. - Also known as "Atlantic cod" and "had- dock," respectively, according to Common and scientific names of fishes from the United States and Canada. 1991. Am. Fish. Soc. Spec. publ. 20. Besthesda, MD, 183 p. Aglen, A. 1999. Report on the demersal fish sui-veys in the Barents Sea and Sval- bard area during summer/autumn 1996 and 1997. Unpubl. manuscr. Fisket og Havet NR. 7-1999. Institute of Marine Research. PO Box 1870 Nordnes. N-5817 Bergen. Norway. Pennington et a\ Assessing the precision of frequency distributions from trawl survey samples 75 The data for the Naniihian deepwater hake iMt'lurciiis paradoxus) were collected during bottom trawl sui-\-cys off Namibia conducted by the Ministry of Fisheries and Ma- rine Resources of Namibia in conjunctioTi with the Noi-we- gian Agency for Foreign Aid (NORAD). For these sui-v-eys, tows of 30-minute duration were made at stations along transects perpendicular to the coast. ' The data for the deepwater hake for South Africa, were collected from during bottom trawl surveys off the west coast of South Africa. The sui-veys were conducted by the Marine and Coastal Management Centre, South Africa, by using a stratified random design. Tows of 30-minute dura- tion were made at each station (see Payne et al., 1985). Assessing the precision of length-frequency estimates The sample offish of a particular species measured during a survey is not a random sample of individual fish from the entire population but a sample of?! clusters, one cluster from each station. Because fish caught together are usually more similar than those in the general population, a total of M fish collected in 7i clusters will contain less informa- tion about the population length distribution than M fish sampled randomly. One way to measure the information contained in a sample of length measurements is to esti- mate the number of fish that one would need to sample at random (the effective sample size) to obtain the same infor- mation on length contained in the cluster samples. The effective sample size for cluster sampling can be defined and calculated as follows (Pennington and Vols- tad, 1994: Folmer and Pennington. 2000). First estimate the population mean fish length and its variance based on the clusters of fish caught at n stations. Because both the lengths and the number of fish at a station are ran- dom variables, a ratio estimator is appropriate (Cochran, 1977). The ratio estimator, R. of the mean length is given by R = ^, (1) where M, = the number of fish caught (either actual or estimated) at station /; and fi, = an estimate of the average length of fish at station i. var( «-I (Af, Mrit-t, -/?) nUi-1) (2) ^■here M = M. In. Next estimate the variance, a'~, of the population length distribution. If/?;, fish are randomly selected at each sta- tion (or if all fish are measured), then y V li.e. fish of similar length tend to be caught together), then the terms in the parentheses can greatly increase the variance and thus drastically reduce the effective size. In particular, the term ap M is relatively large for trawl surveys. Finally, if p < 0, which is rarely if ever the case for trawl surveys, then the effec- tive sample size will be larger than M. The precision of estimates of other population charac- teristics, such as age distribution, can also be relatively low compared with the number of fish sampled if the par- ticular attribute or measurement is more similar for fish caught together than for those in the general population. For example, the precision of estimates of mean stomach contents (Bogstad et al. , 1995) or diet composition (Tirasin and Jorgensen, 1999) can be relatively low because of in- trahaul correlation. An effective sample size of one fish per tow does not mean only one fish should be measured at each station, but it implies that the only way to improve survey pre- cision significantly is to increase the number of stations, i.e. to sample fish from as many locations as possible. The bootstrapped estimates of precision and the sampling sim- ulations showed that reducing or increasing the number of Winter 1995 ll Jiilii, Winter 1999 n. ~^si«....._. 20 40 60 100 120 140 20 40 60 80 100 120 140 Lengtti (cm) Figure 3 Bootstrapped estimates of the 95'7r confidence intervals for the proportion of cod in the Barents Sea in each 5-cm length bin. for winter 1995 and for winter 1999. The inner brackets denote the confidence intervals if the estimates are based on all the cod measured during the surveys and the outer brackets denote the confidence intervals if 10 fish arc measured for each subsample. fish measured (or caught and measured) at a station will not significantly affect the precision of length-distribution estimates. In general, if intracluster correlation is positive for an attribute, then it is usually best to take a small sample from as many locations as possible (e.g. Bogstad et al.. 199.5; McGarvey and Pennington, 2001 ). It has been shown that tows of short duration are in general more efficient for estimating stock abundance than long tows (Godo et al., 1990; Pennington and Vols- tad, 1991; Gunderson. 1993; Carlsson et al.. 2000). There- fore one way to collect samples from more locations and improve overall survey efficiency without increasing sur- vey cost is to reduce tow duration and use the time saved to increase the number of survey stations (Pennington and Volstad, 1994). For example, if tow duration were re- duced from 60 minutes to 15 minutes for a trawl survey of shrimp off West Greenland, then 44^^^ more stations could be surveyed (Carlsson et al., 2000). Likewise, a reduction in tow duration from 30 minutes to 10 minutes for a trawl survey on Georges Bank would increase the number of survey stations by about 30*7^ (Pennington and Volstad, 1994). The total number of fish caught would be fewer, on av- erage, if tow duration was reduced, but estimates of fish density would be more precise and the resulting sample of individuals would be more representative of the entire population (Pennington and Volstad. 1994). 80 Fishery Bulletin 100(1) Acknowledgments We thank Rob Leslie (Marine and Coastal Management Centre, South Africa) for providing us with the South Afri- can survey data, and Jon HelgeVolstad ( Versar, Inc., USA) and two anonymous referees for their constructive com- ments and suggestions. Literature cited Bhattacharyya, G. K., and R. A. Johnson. 1977. Statistical concepts and methods. John Wiley and Sons, New York, NY, 639 p. Bogstad, B., M. Pennington, and J. H. Volstad. 1995. Cost-efficient sui-vey designs for estimating food con- sumption by fish. Fish. Res. 23:.37-46. Carlsson, D., R Kanneworff, O. Folmer, M. Kingsley, and M. Pennington. 2000. Improving the West Greenland trawl sui-vey for shrimp iPandalus borealis). J. Northwest Atl. Fish. Sci. 27:151-160. Christensen, V. 1996. Virtual population reality. Rev. Fish Biol Fish 6: 243-247. Cochran, W. G. 1977. Sampling techniques, 3'''' ed. John Wiley and Sons, New York, NY, 428 p. Efron, B. 1982. The jackknife, the bootstrap, and other resampling plans. Society for Industrial and Applied Mathematicians (SLAM), Conference Board of the Mathematical Sciences iCBMSi- National Science Foundation ( NSF i regional conference series in applied mathematics 38. Philadelphia, PA, 92 p. Folmer O., and M. Pennington. 2000. A statistical evaluation of the design and precision of the shrimp sui-vey off West Greenland. Fish. Res. 45: 16.5-178. Godo, O. R. 1994. Factors affecting the reliability of gi-oundfish abun- dance estimates from bottom trawl surveys. In Marine fish behaviour in capture and abundance estimation (A. Feme and S. Olsen, eds. ), p. 166-199. Fishing News Books, Farn- ham, UK. Godo, O. R.. M. Pennington, and JH. Volstad, 1990. Effect of tow duration on length composition of trawl catches. Fish. Res. 9:165-179. Gunderson, D. R. 1993. Surveys of fisheries resources. John Wiley and Sons, New York. NY, 248 p. Korsbrckke, K., S. Mehl, O. Nakken, and M. Pennington. 2001. A sui-vey-based assessment of the Northeast Ai'ctic cod stock. ICES J. Mar Sci. 58:76.3-769. McGai-vey, R., and M. Pennington. 2001. Designing and evaluating length-frequency surveys for trap fisheries with application to the southern rock lob- ster Can. .J. Fish. Aquat. Sci. .58:254-261. Nakken, O. 1998. Past, present and future exploitation and manage- ment of marine resources in the Barents Sea and adjacent areas. Fish. Res. 37:23-35. Payne, A. I. L., C. J. Augustyn. and R. W. Leslie. 1985. Biomass index and catch of Cape hake from random stratified sampling cruises in division 1.6 during 1984. Colin. Scient. Pap. Int. Com. SE Atl. Fish. 12:99-123. Pennington, M., and T. Stromme. 1998. Surveys as a research tool for managing dynamic stocks. Fish. Res. 37:97-106. Pennington. M., and J. H. Volstad. 1991. Optimum size of sampling unit for estimating the density of marine populations. Biometrics 47:717-723. 1994. Assessing the effect of intra-haul correlation and vari- able density on estimates of population characteristics from marine surveys. Biometrics 50:725-732. Research Triangle Institute. 2001. SUDAAN users manual, release 8.0. Research Tri- angle Institute, Research Triangle Park, NC, 886 p. Tirasin, E. M., and T. Jorgensen. 1999. An evaluation of the precision of diet description. Mar Ecol. Prog. Ser 182:243-252. 81 Abstract— Tlie red porgy, Pagrus pag- Ills. IS ;in important roof fish in sovoral offshore fisheries along the southeastern United States. We examined samples from North Cai-olina through south- east Florida from recreational i head- boat) and commercial (hook and line) fisheries, as well as samples from a fishery-independent source. Red porgy attain a maximum age of at least 18 years and 733 mm total length. The weight-length relationship is repre- sented by the In-ln transformed equa- tion: VV = 8.85 X 10-"(L)-!'"^, where W = whole weight in gi'ams, and L = total length in mm. The von Bertalanffy growth equation fitted to the most recent, back -calculated lengths from all the samples is L, = 644( 1 - e-"'^-'" * "-'S'). Our study revealed a difference in mean length at age of red porgy from the three sources. Red porg\' in fishery- independent collections were smaller at age than specimens examined from fishery-dependent sources. The differ- ence in length-at-age may be related to gear selectivity and have important consequences in the assessment of fish stocks. Estimated ages of red porgy (Pagrus pagrus) from fishery-dependent and fishery-independent data and a comparison of growth parameters Jennifer C. Potts Charles 5. Manooch III Center for Coastal Fisheries and Habitat Research Beaufort Laboratory National Manne Fishenes Service, NOAA 101 Pivers Island Road Beaufort, North Carolina 28516 9722 Email address (lor J C Potts) Jennifer pottsidinoaa gov Manuscript accepted 20 August 2001. Fish. Bull. 100:81-89(2002). Red porgy, Pagrus pagrus, inhabit con- tinental shelves in temperate and trop- ical waters throughout the Atlantic Ocean and Mediterranean Sea. The spe- cies supports fisheries in many coun- tries and is heavily exploited. Since 1992, red porgy has ranked relatively high (38 of 200) in value among all fin- fish landed commercially in the south- eastern United States.' Red porgy form a substantial part of overall reef fish landings, especially in North Carolina and South Carolina, although there is little directed fishing for the species. Commercial landings of red porgy from the southeastern U.S. peaked in 1982 at 535 metric tons (t) and declined to 134 t in 1993 (Potts and Burton-). Red porgy ranked second by weight for reef fish landed by recreational headboat^ anglers through the early 1980s. Since then, headboat landings of red porgy have declined, and landings of vermil- ion snapper, Rhomhoplites aurorubens, which are also declining, have now sur- passed red porgy. White grunt, Hae- mulon plumieri, and gray triggerfish, Balistes capriscus, which were less pre- ferred than other members of the snap- per grouper complex, have increased in landings and now surpass red porgy.^ Mean weight of red porgy from the commercial and recreational fisheries has declined from 1.06 kg in the 1970s to 0.66 kg in 1997.- Minimum size regu- lations ( 305 mm total length ) for recre- ational and commercial fisheries enact- ed in 1992 did little to increase mean weight in catches, although the head- boat fishery did show a slight increase from 0.48 kg in 1991 and 1992 to 0.60 kg in 1997. Additionally, population bio- mass estimates for red porgy in the southeastern United States have plum- meted from a peak of 3.27x10*" kg in 1978 to 0.43xl0« kg in 1992 (Huntsman et al.''). These trends suggest that red porgy stocks are being overexploited. Age determination studies have been conducted throughout the range of red porgy. Manooch and Huntsman (1977) conducted the first comprehen- sive study using scales (n = 1777) and whole otoliths {n=222) to age red porgy that were caught by recreational fisher- men using hook-and-line gear off North Carolina and South Carolina when the species was lightly exploited ( 1972-74). Harris and McGovern (1997) aged red porgy from whole otoliths (;!=4281) of ' General canvas. 1998. Unpubl. data. Miami Laboratory, National Marine Fish- eries Sei-vice, 75 Virginia Beach Dr, Miami, Florida 33149. - Potts, J. C, and M. L. Burton. 1999. Trends in catch data for fifteen species of reef fish landed along the southeastern United States. Unpubl. data. South At- lantic Fishery Management Council, 1 Southpark Circle, Charleston, SC 29407. ' A "headboat" is a fishing vessel that car- ries more than six passengers who pay per person lor by the "head") to go offshore fishing. ■• Headboat annual summaries. 1998. Un- publ. data. Center for Coastal Fisheries and Habitat Research. Beaufort Labora- torv, 101 Pivers Island Rd.. Beaufort. NC 28516-9722. '■ Huntsman, G. R., D. S. Vaughan, and J. C. Potts. 1994. Trends in population status of red porgy, Pagrus pagrus. in the Atlantic Ocean of North Carolina and South Carolina, USA, 1971-1992. Unpubl. data. SouthAt- lantic Fisherv Management Council, 1 South- park Circle, Charieston, SC 29407. 82 Fishery Bulletin 100(1) fish caught from North Carolina to Florida with fishery- independent gear during 1979-81 and 1988-94. Nelson ( 1988) aged red porgy with scales (?! = 126) from fish caught in the northwestern Gulf of Mexico with fishery-indepen- dent hook-and-line gear and trap gear during 1980-82. In the eastern Gulf of Mexico, Hood and Johnson (2000) aged red porgy from sectioned otoliths (^=852) collected from headboat and commercial catches during 1995-96. Vassi- lopoulou and Papaconstantinou (1992) used scales from 138 red porgy that were taken with fishery-independent hook and line and trammel nets in the Mediterranean Sea during 1985-86, and Serafim and Krug (1995) aged red porgy from whole otoliths (?!=358) that were collected by using commercial longlines and fishery-independent gear in the Azores during 1991-93. Researchers in the Canary Islands aged 1505 red porgy from commercial trap and longline samples during 1985-86 and 1991-93 (Pajuelo and Lorenzo, 1996), and researchers off the Argentinian coast used trawl-caught samples during 1972-81 to obtain 5859 red porgy that were aged from scales (Cotrina and Raimondo, 1997). Predictions offish populations from models rely heavily on input data sets, including age and growth. If samples used in the aging study are not representative of the en- tire population (i.e. the entire geogi'aphic range of the stock, full range of fish size, and different gear types), model predictions (e.g. spawning potential ratio |SPR|) can mislead management decisions. A comparative stock assessment of red porgy was done by using growth pa- rameters and age-length keys generated from two stud- ies: 1) fishery-independent data (Harris and McGovern, 1997) and 2) fishery-dependent data (Manooch and Hunts- man, 1977; Potts et al.'M. Each set of age and growth data was applied to fishery-dependent landings and length fre- quencies. The fishery-independent age and growth data produced a static SPR of 46^^^, which is well above the overfished definition (SPR<30'7f) as set forth by the South Atlantic Fishery Management Council (SAFMC). The fish- ery-dependent age and growth data produced a static SPR of 19*7^ (Potts et al.''), which makes red porgy, by definition, overfished and which necessitates that stringent manage- ment measures be put in place to protect the stock. The purpose of our study was to update the age and growth information on red porgy caught in the recreation- al and commercial fisheries operating along the southeast- ern United States. We present the von Bertalanffy growth model, weight-length relationship, and age-length keys for red porgy collected from the headboat hook-and-line fishery, commercial hook-and-line fishery, and fishery-in- dependent samples. We also compare mean age at length of red porgy collected from recreational fisheries, commer- cial fisheries, and fishery-independent sources. We discuss how data source selection affects the growth parameters. Materials and methods Sagittal otoliths were collected from red porgy landed by hook-and-line fishermen from the headboat (recreational) fishery («=249) between 1989 and 1998 (59% from 1996 to 1998) and the commercial fishery (n=264) between 1997 and 1998 operating from North Carolina to southeast Flor- ida. From the two fisheries, 64% of the samples came from North Carolina, 14% from South Carolina, and 22% from the east coast of Florida. Because of minimum size limit regulations (305 mm total length), the South Caro- lina Department of Natural Resources (SCDNR) Marine Monitoring and Prediction (MARMAP) Program supplied us with otoliths from red porgy that were smaller than those available from the fisheries (n=59) and an additional 62 samples ranging from 300 to 425 mm total length. These fish were caught primarily with Chevron traps off South Carolina during 1996 and 1997. Total length, whole weight, port of landing, and date of capture were recorded for each sample. Tlie otoliths were stored dry in coin envelopes. For age analysis, three transverse (dorsoventral) sec- tions from the left otolith of each fish were taken by us- ing a low-speed saw. One section was made on either side of the core, and the other encompassed the core. The sections were mounted on glass slides with thermal ce- ment, and examined through a microscope at 80x and illuminated with reflected light. Clove oil was applied to each section to enhance the legibility of the growth zones on the section. The samples were put in sequential order from smallest to largest, and one reader counted the number of opaque zones in the otolith section. A sec- ond reader examined a random sample of the otoliths. If the readers disagreed on the age of a sample, they exam- ined it again. If consensus was reached, the sample was retained; otherwise, the sample was discarded. Measure- ments from the core to the outer edge of each successive opaque zone and the otolith margin were taken along the lateral plane on the dorsal lobe of the section by using an ocular micrometer. Analysis of the marginal increment (the distance be- tween the last opaque zone and otolith margin) was used to validate the annual deposition of the opaque zones in the otoliths. For each age and month, the mean of the rela- tive marginal increment, the ratio of the marginal incre- ment to the distance between the last two opaque zones, was plotted. An opaque zone was considered an annulus if a minimum ratio was recorded for one month or season. The relationship of fish length and otolith radius was described by regi'essing the obsei-ved total length on oto- lith radius (/?(.). The linear equation was L = a +blR^.). where L = total length in mm. 6 Potts, J. C, M. L. Burton, and C. S. Manooch, III. 1998. Trends in catch data and estimated static SPR values for fifteen spe- cies of reef fish landed along the southeastern United States. Unpubl. data. South Atlantic Fishery Management Council, 1 Southpark Circle, Charleston, SC 29417. The back-calculated total lengths at each age were deter- mined from the body proportional equation (Francis, 1990): L^ =[(a+hR^)/{a + bRc)]Lc, Potts and Manooch Estimated ages of Pagrus pagtvs 83 where L^ = back-calculated total length to aniuilusA; a = intercept from the linear total lengtii-otohth radius regression; b = slope from the linear total length-otolith radius regression: L(. = total length at time of capture; /?, = otolith radius to annulus A; and /?(. = total otolith radius at time of capture. The von Bertalanffy ecjuation. L, = L |1 - e.\p(-A'(^-/(,)|, was fitted to back-calculated lengths-at-ages for the most recently formed annuli (Ricker, 1975; Everhart et al., 1981; Vaughan and Burton, 1994). Growth parameters were es- timated by using SAS PROC NLIN with the Marquardt Option (SAS Institute, 1982) for all aged fish and for fish obtained from fishery-dependent sampling. Differences in mean back-calculated length at age for the most recently formed annulus for the three sample sources, i.e. recreational, commercial, and fishery-indepen- dent, were tested by using the general linear model analy- sis of variance. To estimate the whole weight of gutted red porgy landed in the commercial fishery and to estimate stock biomass from assessment models, a regression of hii fisli tveif^ht) on \n(fish length ) was performed and transformed to W = aiL)^, where W = weight in g, and L = total length in mm. Age-length keys were constructed from observed age at length by sample source in which the ages were unadjust- ed for time of year. Fish that were aged were assigned to 25-mm length intei-vals. Results Red porgy sampled for our study ranged from 176 to 733 mm TL and from 1 17 to 5895 g in whole weight. Ages were determined for 631 of 634 (99'7f) sectioned sagittal oto- liths. Of those aged, 603 idd'v'c) otoliths were considered legible to record measurements from the core to each suc- cessive opaque zone and the otolith margin. On twenty additional samples, we were able to measure only the oto- lith radius. Sectioned otoliths exhibited a recurrent pat- tern of alternating wide translucent zones and thin opaque zones. Estimated ages ranged from 1 to 18 years. Analyses of marginal increment data indicated that the opaque zones were annular in nature and were formed in the spring (Fig. 1). Mean relative marginal increments for ages 2 through 8 were lowest in March through May and were the only months that had marginal increments equal to zero. They then steadily increased from June through October and remained high through February. Back-calculated total lengths at age of red porgy were estimated from the parameters from the regi-ession equa- tion of total length (L) on otolith radius (R^.). The plot of length on radius was linear, and the linear regression equation that best fitted the data was L = -132.84 -i- 10.87(fl ) (;--=0.91, «=623). Using the Francis (1990) body proportional hypothesis, we found that weighted mean back-calculated lengths ranged from 103 mm for age-1 fish to 721 mm for age-18 fish (Table 1). 1.2 iOBp\ .b^^ CD ^ 0.6 0.2 Mean n porgy fi w^ ^ - • + - Age 7 ■-Q--Age8 2 3 4 5 6 7 8 9 10 11 12 IVlonth Figure 1 lonthly relative marginal increment (MI) of red om the southeastern United States plotted by age. The back-calculated lengths at the last annulus forma- tion were used to estimate the von Bertalanffy equation. The equation parameters (±1 SE) were L„ = 644.72 ±17.93, A' = 0. 15 ±0.01, and /„ = -0.76 ±0. 10. The theoretical lengths at age ranged from 149 mm at age 1 to 605 mm at age 18. Theoretical lengths closely fitted the observed and back- calculated lengths through age 14 (Table 1). When we used fishery-dependent samples only to generate the von Berta- lanffy growth equation, the resulting parameters (±1 SE) were L,= 773.73 ±39.49, A' = 0.09 ±0.01 and t„ = -1.96 ±0.21. The fishery-dependent theoretical lengths at age ranged from 181 mm at age 1 to 646 mm at age 18. We used ages 2 through 6 and data years 1996 through 1998 to compare length at age of the three data sources be- cause the three sets overlapped for those ages and years. The ANOVA on the mean back-calculated length at age of the most recently formed annulus between sample sources indicated a significant difference in age at size between the MARMAP, headboat, and commercial red porgy sam- ples (r-=0.88; F-value=522.21; P=0.0001 for all combina- tions) and were represented by the model TL = a„ + Ycc + y,,h + ^ /3, A, , where c = 1 if fishery = commercial, or c = if fishery i^ commercial; h = 1 if fishery = headboat, or /i = if fishery ^ headboat; A, = 1 if age = 2, or A, =0 if age ^ 2, etc.; and J = age categories. Average TL = a^ for fishery = fishery-independent and age = 6; average TL = a,, -i- y^.c for fishery = commercial and age = 6; etc. The model indicated no year effect, and no interaction between fishery and age. MARMAP (fishery- independent) samples were smaller at age than those from the commercial and headboat fisheries. Although mean back-calculated lengths at age between headboat and com- mercial data sources were statistically different, the dif- ferences were slight (<15 mm) (Fig. 2). 84 Fishery Bulletin 100(1) Table 1 Mean back-calculated. mean observed, and theoretical total lengths Immt of red porgy from the sou theastern United States. Age (yrl 11 An nulus numl er 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1 15 134 2 33 96 206 3 136 98 208 274 4 158 104 212 274 327 5 136 105 212 274 326 363 6 56 104 215 275 329 367 399 7 31 107 216 284 342 381 413 441 8 26 100 213 278 332 371 400 427 449 9 6 106 221 283 337 377 410 438 461 483 10 1 121 221 278 334 379 412 457 491 513 536 11 2 112 215 275 329 362 395 427 449 476 498 520 12 1 108 219 274 330 374 407 440 462 485 507 529 551 14 1 111 226 294 351 385 419 442 465 487 510 533 556 567 579 18 1 116 247 306 365 413 448 484 506 531 555 567 591 614 638 662 686 709 721 Total 603 Weighted mean TL 103 211 275 329 368 404 436 455 489 517 534 566 591 608 662 686 709 721 Incremental growth 103 108 64 54 39 36 32 19 34 28 17 32 25 17 54 24 24 12 Observed TL 198 236 303 350 386 423 459 470 508 547 536 562 590 733 Theoretical TL 149 218 278 329 374 410 443 471 495 516 534 549 562 574 583 592 599 605 — 1 2 3 4 5 6 7 8 9 10 11 12 13 14 IE Age (yr) -Fishery-Independent — »— Headboat -h- Commercial Figure 2 Mean back-calculated length at age of red porgy obtained from a fishery-independent source ( MAR- MAPl, headboat operations, and commercial fish- ery operations between 1996 and 1998. The weight-length relationship for red porgy in the southeastern United States was best described by the con- verted In-ln regression equation of IV = 8.85 x 10^''(L)''^"' (r-=0.96, /!=230, MSE=0.01l. Age-length keys by sample source are presented in Table 2. The fishery-independent key is appropriate for fishery-independent length data only The headboat and commercial keys can be used to convert unaged length sam- ples of red porgy from the fisheries operating in the south- eastern United States to aged ones. Annual keys were not available owing to the small sample size from each year. Discussion In two previous aging studies on red porgy from the south- eastern United States, populations were examined at two different levels of exploitation and different structures were used to determine age. Manooch and Huntsman (1977) sampled recreationally caught fish from an almost virgin stock off North Carolina and South Carolina during 1972-74. Although they used scales and whole otoliths, the main focus of their study and the analysis were on ages determined from scales. Of the 3278 scales analyzed, only 54'^'f (1901) were legible enough to record ages. The main problem of aging red porgy with scales was the large number of regenerated scales." Harris and McGov- ern (1997) used whole otoliths (?7=4281) to age fish col- lected between 1979 and 1981 and between 1988 and 1994 from the MARMAP survey, a fishery-independent source. The 1988-94 samples in their study came from the area off North Carolina through northeast Florida, although 73% were collected in the area off Charleston, SC, between 32°N and 33°N. The samples were limited mainly to indi- viduals below 450 mm TL (less than l'~f of samples were Manooch. C. S. 1998. Personal conimun. NOAA. Center for Coastal Fisheries and Habitat Research. 101 Fivers Island Road, Beaufort, NO 28516. Potts and Manooch; Estimated ages of Pagrus pagrus 85 Table 2 Age-lt>ngth keys for red porgy fron the southeastern United States by sample source; fishery- independent headboat. and com- morcial. Total length classes are in 25-mm intervals (i.e. 175 = 17.5-199). TL class n Age (yr) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Fishery-independent 175 11 10 1 200 21 5 16 225 7 1 6 250 9 9 275 10 10 300 16 8 6 2 325 20 11 9 350 11 5 5 1 375 14 6 8 400 1 1 425 450 475 500 525 550 575 725 Total 120 Headboat 175 200 225 3 3 250 6 5 1 275 28 5 23 300 44 36 7 1 325 48 15 31 2 350 41 2 23 16 375 37 11 21 5 400 19 11 5 2 1 425 11 2 5 2 2 450 5 1 3 1 475 2 2 500 3 1 2 525 2 1 1 550 575 725 Total 249 continued >450 mm). The data Harris and McGovern presented for the period between 1979 and 1981 showed that 89; of the samples for the reproductive study were greater than 450 mm TL. We sampled from a heavily exploited stock, and fish ranged up to 733 mm TL, and 14'7r of the fishery- dependent samples were greater than 4.50 mm TL. Addi- tionally, our samples of red porgy were from a broader geographic range (529f from North Carolina, 309^ from South Carolina, assuming all MARMAP samples were from South Carolina, and 18% from east coast Florida) than that of previous studies. Our distribution of samples more closely reflected the landings of red porgy in the southeastern United States. Also, we used sectioned oto- liths, which have been determined to be the best struc- 86 Fishery Bulletin 100(1) Table 2 (continued) TL class n Age (yr) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Commercial 175 200 225 250 275 5 2 3 300 25 22 3 325 34 4 27 3 350 28 24 4 375 37 8 24 5 400 44 29 14 1 425 27 6 17 3 1 450 31 7 16 8 475 20 5 15 500 7 1 6 525 1 1 550 1 1 575 1 1 725 1 1 Total 262 Table 3 Mean observed total I ength (mm) of red porgy from this study and others: 1 = our stud V (all data combmed); 2 = 1 Manooch and Huntsman (1977: scale data); 3 = Manooch and Huntsman (1977: atolith data); 4 = Harris and McGov- | ern (1997: 1994 data year); 5 = Harris and McGovern | (1997: 1988 data year) Age (yr) Study 1 2 3 4 5 1 198 238 229 218 224 2 236 290 288 294 297 3 303 342 331 306 329 4 350 382 374 328 379 5 386 419 402 344 445 6 423 451 425 362 399 7 459 483 453 386 431 8 470 505 474 374 406 9 508 527 496 448 10 547 543 534 449 11 536 558 557 12 562 604 13 595 14 590 15 694 16 17 18 733 ture to age fish in many studies (Beamish, 1979; Boehlert, 1985; Smale and Punt, 1991). Red porgy form their opaque zones during the spring along the southeastern United States. Both Manooch and Huntsman (1977) and Harris and McGovern (1997) re- ported that annulus formation occurred in red porgy dur- ing March and April. We found that the opaque zone formed from March through May Springtime formation has also been reported from the Gulf of Mexico (Nelson, 1988; Hood and Johnson. 2000), from the Azores (May; Se- rafim and Krug, 1995) and from Argentinian waters (Oc- tober; Cotrina and Raimondo, 1997). Pajuelo and Lorenzo, in their (1996) study of red porgy off the Canary Islands, found that the opaque zone was formed during the sum- mer (June through October). A comparison of the mean length at age from our study and that from Harris and McGovern's ( 1997 ) study from 1994 (Table 3) clearly reveals that red porgy caught with fishery-independent gear are much smaller at age for fish 3 years and older than fish caught with fishery-dependent gear. Mean size at age in our study was similar to data for fish 5 years and older reported by Manooch and Hunts- man (1977) using otoliths, although our mean sizes were smaller that those reported for their scale data. Ages 1-5 of our study were a mix of MARMAP samples and fishery- dependent samples, which may explain why red porgy were smaller for those ages than those reported by Manooch and Huntsman (1977). The mean obsei-ved length at age from our study were on average 23 mm smaller than the corre- sponding lengths from Manooch and Huntsman (1977) for ages 6 through 12 as assigned by scales. The differences may Potts and Manooch Estimated ages of Pagnis pagnis 87 -M&H: Southeastern US H&M 79-81 Soulheaslern US • HSM ee 90 Soulheaslern US H&M 91-94 Soulheaslern US P&M Soulheaslern US P&L Canary Islands V&P Easlern Medilerranean S&K Azores C&R North Buenos Aires C&R South Buenos Aires H&J Eastern Gull ol Mexico N: Western Gull ol Mexico 1 3 5 7 9 11 13 15 17 Age (yr) Figure 3 Comparison of von BertalanfTy growth curves from various locations in the Atlantic Ocean: M&H = Manooch and Huntsman 1 1977); H&M = Harris and McGovern (1997) from three separate data year sets; P&M = Potts and Manooch (this studyi; P&L = Pajuelo and Lorenzo 11996); V&P = Vassilopoulou and Papaconstantinou (1992); S&K = Serafim and Ki'ug ( 1995); C&R = Cotrina and Raimondo ( 19971 from two study areas; H&J = Hood and Johnson (2000); N = Nelson (1988). 100 1 3 5 7 9 11 13 15 17 Age (yr) -»-M &H -iic -P&M - Fishe ry-dependent data Figure 4 Comparison of the von Bertalanffy growth curve from Manooch and Huntsman (1977) using headboat data versus the resulting growth curves from headboat and commercial fisheries data from this study. have been due to heavy fishing on the population, the small sample size of older age fish, or differences in assigned ages due to the structure used for aging, because the comparison of ages determined from otoliths was very close. Harris and McGovern (1997) and Hood and Johnson (2000) have put forth the theory that heavy fishing pres- sure may cause a shift in the size and age structure of a population to smaller, slower growing fish. Our study does not support this theory. Cotrina and Raimondo ( 1997 ) dem- onstrated the differences in growth of red porgy caught from two areas off Argentina. Because we feel that our samples encompassed the full range of red porgy along the southeastern United States., our data more truly rep- resents that population than the data from Harris and McGovern's (1997) study. We also feel that the perceived changes in length at age reported in Harris and McGov- ern's (1977) study may be confounded by the changes in sampling strategy of the MARMAP program between 1979 and 1994 (e.g. sampling gear used, locations of sampling, personnel, etc). Differences in growth of red porgy in the Gulf of Mexico between Hood and Johnson's (2000) study and Nelson's (1988) study may certainly be affected by the different locations the samples came from for the two studies. We suggest that further investigation into sample design and effects of habitat, temperature, fishing pres- sure, weather, fishing gear, etc. on fish populations is need- ed to help resolve differences between these studies. The von Bertalanffy growth parameters L . and K are integral components of stock assessment models. Samples for age determination should be representative of the tar- get population (i.e. entire geographic range, all habitats, and all gear types used in the fisheries). Because back- calculated length-at-age information was unavailable to us from all previous studies, we compared the various von Bertalanffy growth curves with direct comparisons of length at age (Fig. 3). The growth curve of red porgy es- timated by Manooch and Huntsman (1977), L, = 763(1 - g-uo96i( -1- 1 88i)_ jg similar to our equation for all samples. Our calculated von Bertalanffy equation with fishery-de- pendent samples only was almost identical to that of Ma- nooch and Huntsman (1977): L, = 774(1 - g-ooaa/ * i96i) (Fig. 4). Although aging structures for the two studies were very different, ages were validated in both studies, and overall size ranges were similar In comparison with our study and that of Manooch and Huntsman (1977), the growth curves estimated by Harris and McGovern ( 1997) from their 1979-94 data had L, val- ues ranging from 411 to 451 mm TL. Red porgy in their study exhibited theoretical growth only from 58 to 69 mm, between ages 5 and 11. Red porgy from our study (all data combined) and Manooch and Huntsman's ( 1977) study the- oretically grew from 164 mm to 172 mm, between ages 5 and 11. Also, the growth coefficients (0.27 to 0.34: 1979-94 data) from Harris and McGovern's ( 1997) data seemed high for red porgy in relation to those reported in other red por- gy age and growth studies (Table 4) (Manooch and Hunts- man, 1977; Vassilopoulou and Papaconstantinou, 1992; Se- rafim and Krug, 1995; Pajuelo and Lorenzo, 1996; Cotrina and Raimondo, 1997; Hood and Johnson, 2000). Studies using fishery-independent sources for red porgy showed smaller L^. and higher A' values than those from most studies where a combination of commercial, recre- ational, or fishery-independent samples were used (Table 4). The differences in growth curves are likely a result of smaller fish in the fishery-independent samples and a con- sequent truncated upper-length range. Although Hood and 88 Fishery Bulletin 100(1) n u a e CO 5 r ^ (^ s: rfi i- K 6- Si " II ■n tL. < w z a, II E H -n o !^ X o 3 o [/J C/J II •n 3 1.; C/D w X ffl :S ^ fci rt T) o C X C8 o o T3 3 3 o 3 O CO CO 3 CO 3 CO -a o CTj ^ 3 03 o o •c CO 02 1 en _o 'cfi CO 11 S > 1.0 ,_ o Ol 00 CO 1 1 o K o CO ^ CO ^ e:^ c — ti. 5 ^ Potts and Manooch Estimated ages of Pagrus pagrus 89 .Johnson (2000) used fishery-dependent samples, the larg- est fish in that study was 489 mm TL, and their samples came from primarily one location. Selectivity of fishing gear may also explain differences in growth parameters. For example, hook-and-line gear may catch faster growing and more aggi'essive fish, whereas traps and trawls may catch slower growing fish and overall smaller fish. We rec- ommend that in future age and growth studies of any fish species, samples represent the full range of fish sizes in a population, including fish caught by many different gear types, and are obtained from the entire geographic range of the stock. Where practical, landings should be stratified and sampled accordingly. Acknowledgments We would like to thank the port agents of the National Marine Fisheries Service and the South Carolina Depart- ment of Natural Resources MARMAP personnel for pro- viding us with otolith samples. Jim Waters and Doug Vaughan, NOAA's Center for Coastal Fisheries and Habi- tat Research, were instrumental in the statistical analysis of the length-at-age data and provided thoughtful com- ments on the manuscript. Dean Ahrenholz and Joe Smith also of NOAA's Center for Coastal Fisheries and Habitat Research provided a critical review of the manuscript. Literature cited Beamish. R. J. 1979. Differences in the age of Pacific Hake (Mcrluccius productus) using whole otoliths and sections of otoliths. J. Fish. Res. Board Can. 36: 141- 151. Boehlert, G. W. 1985. Using objective criteria and multiple regression models for age determination in fishes. Fish. Bull. 83:103-117. Cotrina, C. P.. and M. C. Raimondo. 1997. Study on the age and gi'owth of the red porgy Pagrus pagrus from the Buenos Aires coa.stal shelf Rev. Invest. DesaiT. Pesq. 11:95-118. Everhart. W. H.. A. W. Eipper, and W. D. Youngs. 1981. Principles of fishery science, 2nd ed. Cornell Univ. Press, Ithaca, NY. 288 p. Francis, K. 1. C. C. 1990. Back-calculation offish Iciigtlis: a critical review. J. Fish. Biol. 36:883-902. Harris. P. J., and J. C. McGovern. 1997. Changes in the life history of red porg>-. Pagrus pagrus. from the southeastern United States. Fish. Bull. 95:732-747. Hood, P. B., and A. K. Johnson. 2000. Age, growth, mortality, and reproduction of red porgy, Pagrus pagrus, from the eastern Gulf of Mexcio. Fish. Bull. 98:723-7.35. Manooch, C. S., Ill, and G. R. Huntsman. 1977. Age, growth, and mortality of the red porgy. Pagrus pagrus. Trans. Am. Fish. Soc. 106:26-33. Nelson, R. S. 1988. A study of the life history, ecology, and population dynamics of four sympatric reef predators iRhombnplites aurnrube/is. Lutjanus campechanus, Lutjanidae; Haeniu/on luctanurum. Haemulidae: and Pagrus pagrus. Sparidae) on the East and West Flower Garden Banks, northwestern Gulf of Mexico. Ph.D. diss.. North Carolina State Univ, Raleigh, NC, 197 p. Pajuelo, J. G., and J. M. Lorenzo. 1996. Life history of the red porgy, Pagrus pagrus (Teleostei: Sparidae I. off the Canary Islands, central east Atlantic. Fish. Res. 28:163-177. Ricker. W. E. 1975. Computations and interpretations of biological sta- tistics of fish populations. Bull. Fish. Res. Board Can. 191:l-.382. SAS Institute, Inc. 1982. SAS user's guide: statistics. SAS Institute, Gary, NC, 1028 p. Serafim, M. P R. and H. M. Ki'ug. 1995. Age and growth of the red porgy, Pagrus pagrus (Lin- naeus, 1758) (Pisces. Sparidae). in Azorean waters. Arqui- pelago (Life Mar Sci.) 13A:ll-20. Smale. M. J., and A. E. Punt. 1991. Age and growth of the red steenbras Pctrus rupcslns I Pisces: Sparidae) on the southe-east coast of South Africa. S. Afr J. Mar .Sci. 10:131-139. Vassilopoulou, v., and C Papaconstantinou. 1992. Age, growth and mortality of the red porgy. Pagrus pagrus. in the eastern Mediterranean Sea (Dodecanese, Greece). Vie Milieu 42:51-55. Vaughan, D. S., and M. L. Burton. 1994. Estimation of von Bertalanffy gi-owth parameters in the presence of size-selective mortality: a simulation exam- ple with red grouper Trans. Am. Fi.sh. Soc. 123:1-8. 90 Abstract— Bycatch taken by the tuna purse-seine fishery from the Indian Ocean pelagic ecosystem was estimated from data collected by scientific observ- ers aboard Soviet purse seiners in the western Indian Ocean (WIO) during 1986-92. A total of 494 sets on free- swimming schools, whale-shark-associ- ated schools, whale-associated schools. and log-associated schools were ana- lyzed. More than 40 fish species and other marine animals were recorded. Among them only two species, yellow- fin and skipjack tunas, were target spe- cies. Average levels of bycatch were 0.518 metric tons (t) per set, and 27.1 t per 1000 t of target species. The total annual purse-seine catch of yellowfin and skipjack tunas by principal fishing nations in the WIO during 1985-94 was 118.000-277,000 t. Nonrecorded annual bycatch for this period was estimated at 944-2270 t of pelagic oce- anic sharks, 720-1877 t of rainbow runners, 705-1836 t of dolphinfishes, 507-1322 1 of triggerfishes, 1 13-294 t of wahoo, 104-251 t of billfishes, .53-112 t of mobulas and manias, 35-89 t of mackerel scad, 9-24 t of barracudas, and 67-174 t of other fishes. In addi- tion, turtle bycatch and whale mortal- ities may have occurred. Because the bycatches were not recorded by some purse-seine vessels, it was not possible to assess the full impact of the fish- eries on the pelagic ecosystem of the Indian Ocean. The first step to solving this problem is for the Indian Ocean Tuna Commission to establish a pro- gram in which scientific observers are placed on board tuna purse-seine and longline vessels fishing in the WIO. Bycatch in the tuna purse-seine fisheries of the western Indian Ocean Evgeny V. Romanov Southern Scientific Research Institute of Marine Fisheries and Oceanography (YugNIRO) 2, Sverdlov St 98300, Kerch, Crimea, Ukraine E mail address islande'cnmeacom Manuscript accepted 20 March 2001. Fish. Bull. 100(1): 90-105 (2002). One of the most inipoitaiit require- ments of the UN Convention on the Law of the Sea of 1982, which determines strategies for exploitation of marine living resources (Article 119, b), is to take into account the impact of fish- eries on ". . . species associated with or dependent upon harvested species with a view to maintaining or restor- ing populations of such associated or dependent species above levels at which their reproduction may become seri- ously threatened. . ." (United Nations, 1983). Estimating the magnitude of bycatch is one of the first steps to deter- mine the impact of fisheries on associ- ated species. Tuna purse-seine fisheries probably apply the most intensive direct htnnan impact on the tropical epipelagic eccsys- tems in all oceans. Because of the world- wide scale of purse-seine fisheries, an assessment of their impact on associat- ed and dependent species is essential. Two tunas, yellowfin Thunnus alba- cares (Bonnaterre, 1788) and skipjack Katsiiwonus pe/amis (Linnaeus, 1758), are the target species of most purse- seine fisheries. In this study bycatch is defined as the fraction of the catch that consists of nontarget species (including other species of tuna) that are encircled by the fishing gear and are unable to escape by themselves. Bycatch of asso- ciated and nonassociated species dur- ing purse-seine fishing for tropical tu- nas may be rather high, and generally depends on fishing tactics. The species composition of bycatch in purse-seine fisheries depends on the structure, behavior, and spatial organi- zation of siu'face multispecies aggrega- tions. Schools of different tuna species and other pelagic fishes, marine mam- mals, and other marine animals have aggregated distributions. From our ob- servations and in the opinion of other researchers (Au and Ferryman, 1985; Au and Pitman, 1986; Au, 1991; Cort, 1992), marine birds arc also an inte- gral component of the majority of these multispecies groups. The tunas, as a rule, prevail by bio- mass and abundance in such groups. Tuna schools are traditionally classi- fied by the visually distinctive part of the group or by whether they associate with floating objects or marine mam- mals (Scott, 1969; Petit and Stretta, 1989). "Free-swimming schools" may include associations between different species of tuna. For each type of school, its various components occur in differ- ent ratios. Some epipelagic species that occur in the purse-seine bycatches are not mem- bers of multispecies aggregations. They, instead, may comprise members of the flotsam community or are tuna forage. Several associated components, such as whales and birds, usually escape or avoid the nets and do not become by- catch. Therefore, the composition of the catch often does not represent the actu- al species composition of the multispe- cies associations. Assessments of bycatches have been made for the eastern Pacific Ocean purse-seine tuna fishery (Joseph, 1994; Garcia and Hall, 1995; Hall, 1996, 1998; Anonymous, 1997, 1998, 1999). where the bycatch problem attracted attention because of dolphin mortality during sets on dolphin-associated tuna schools. The economic, political, and ecological implications of this problem produced wide international attention (Charat- Levy, 1991; Jcseph, 1991, 1994; Hall, 1998). Bycatch estimates for the west- ern Pacific purse-seine tuna fisheries have been published also (Bailey et al., 1996). Romanov Bycatch in the tuna purse seine fisheries of the western Indian Ocean 91 111 th(_' \v(-stern Indian Ocean iW'IOi. lima-clolpliin as- sociations are well known in coastal pelagic zones, e.g. Gulf of Aden (Deniidov') and Sri-Lanka (do Silva and Bon- iface-). They are often used in small-scale troll and pole- and-line fisheries for locating yellowfin tuna. In offshore regions of the WIO tuna-dolphin associations are rare, purse seining for them is not practiced, and there is no dol- phin bycatch problem. Perhaps for this reason, the magni- tude of bycatch in the WIO is unknown, except for recent information on species composition (Santana et al.. 1998). Bycatches are not recorded for tuna seiners operating in the WIO, except bycatches of nontarget tuna species. This paper represents a first attempt to estimate catches of as- sociated species by tuna purse seiners in the WIO, based on scarce information collected bv scientific obsen'ers. Materials and methods Bycatch assessments were based on data collected by Yug- NIRO scientific observers aboard Soviet (since 1992 — Rus- sian) tuna purse seiners in the WIO, during 1987, and 1990-91. The vessels were the "Rodina" type.-^ In addition, observer data collected in the same area aboard sister- ships by AtlantNIRO^ and "Zaprybpromrazvedka""' during 1986-90 and data by TINRO'^ and TURNIF' during 1990 and 1992 were used. The fishing vessels all used purse seines of 1800 m in length, 250-280 m in depth, and 90-100 mm mesh size in the bunt. The principal goal of the observer sampling program was an estimation of the species composition of catches in this fisheries, biological analysis of the principal species, and estimates of the length and weight compositions of these principal species in the catches. The observers were placed on board opportunistically (i.e. if a vessel had a free sleeping bed and if there was available funding), without a sampling scheme and without preference to any vessel type. Thus, the sampling could be considered as random. ' Demidov, V. F. 1998. Personal commun. Southern Scien- tific Research Institute of Marine Fisheries and Oceanography (YugNIRO), 2. Sverdlov St., 98300, Kerch, Crimea Ukraine. -' de Silva, J, and B. Boniface. 1991. The study of the handline fishery on the west coast of Sri Lanka with special reference to the use of dolphin for locating yellowfin tuna I Thimnuti albacares I. /;i Indo-Pacific Tuna Development and Management Pi'ogramnie (IPTPl Coll. Vol. Work. Doc TWS/90/18., Vol. 4, p. 314-324. Food and Aginculture Organization of the United Nations (FAO), Viale delle Terme di Caracalla, 00100, Rome, Italy. ■'* Length overall: 85 m; CRT (gross tonnage): 2634; carrying capacity: -1600 mV ■> AtlantNIRO— The Atlantic Scientific Research Institute of Marine Fisheries and Oceanography, 5 Dmitry Donskoi St., 2.36000 Kaliningrad. Russia. ■'' The Department of Searching and Scientific Research Fleet of the Western Basin "Zaprybpromrazvedka," ^" Dmitry Donskoi St., 236000 Kaliningi-ad. Russia. " TINRO— The Pacific Scientific Research Institute of Marine Fisheries and Oceanography, 1 Shevchenko Alley, 690600 VHad- ivostok, Russia. ' TLIRNIF — The Pacific Department of Fish Searching and Sci- entific Research Fleet, 2 Pervogo Maya St., 690600 Vladivostok, Russia. Two other types of Soviet fishing vessels, "Tibiya"*' and "Kauri,""' which took part in the Indian Ocean fisheries during 1985-87 and since 1991 (under the Liberian fiag), were not sampled. In this study coverage rate was esti- mated as percentage of sampled catch to total catch. The obsei-vers recorded the results of each set. The type of school, according to Scott ( 1969) and Petit and Stretta ( 1989), of each set was recorded. I considered sets ftir which an ob- server recorded catch in any quantity as positive sets. The average bycatch level was estimated for all positive sets. For the positive sets, species composition, total weights, and numbers of each species in the catch were recorded. In the vessels of the "Rodina" type, the retained catch was frozen and stored separately. The retained catch was weighed after freezing while being moved to the ship's holds. In nine cases, the weight of some of the catch was es- timated by the ship masters because the holds were over- loaded and some catch was stored in the freezers till land- ing. Therfore estimates of retained catch are presented in this study as frozen weights rather than wet weights. The bycatch was estimated as wet weight. CJnly bycatch taken on board was sampled. The sets when bycatch was not taken onboard but discarded alive (usually with negligible target species catch) and malfunction sets, which do not produce any catch, were not analyzed in this study. Large species, sharks and billfishes generally, were weighed and counted. The weights of specimens heavier than 200 kg (i.e. Mobulidae) were estimated. When the bycatch was more than 200-300 kg, species composition and weight were estimated by using representative samples. Sometimes the obsei-ver recorded the bycatch in num- bers. In these rare cases, the total weights of the fishes were estimated from the average weights of these species in previous catches. The obsei-vers had free access to every fish in the catch. Nevertheless, some obsei-vers had difficulties identifying some billfishes, sharks, and Mobulidae species. Therefore, I pooled the records with doubtful species identification into these three groups for my analysis. These are marked by "?" in the tables. The data were gi'ouped and analyzed by free-swimming schools (including associations between schools of differ- ent species of tuna) and associated schools. The latter in- cluded whale-associated schools and log-associated schools (associated with floating objects). Schools caught in the area of seamounts and shoals — at the peaks of the Equator Seamount and at Saya-de-Malha bank — were considered free-swimming schools. Some ob- servers did not record the type of floating objects that were set on; therefore the sets on natural floating objects (509f to 90% of the log sets sampled) and on fish aggregation devices (FADs) (10-50%) were grouped. Several log sets were made in areas with surface evidence of water masses or current interactions (rips). A set that could not be clearly identified as to set tjqpe was made in such an area and was treated as a log set because of the species composition of the catch and the occurrence of small scattered debris in the rips. ^ Length overall: 55.5 m, GRT: 736, carrying capacity: -361 m-'. '' Length overall: 79.8 m. GRT: 2100. carrying capacity: -1200 m-'. 92 Fisher-y Bulletin 100(1) Table 1 Numbers of sets sampled by year. Positive sets are sets m which an obsei-ver registered catch in any quantity . 1986 1987 1988 1989 1990 1991 1992 Total Total number of sets Number of positive sets Percentage of sets with catch 115 102 30 41 113 54 39 494 68 62 28 41 92 53 33 377 59* 63% 93% lOO'-i 81-"/ 98^; 85'-'f 76 Table 2 Numbers of sets sampled by season and type of school. Type of school Seasoi s Total/positive Winter Spring Summer Autumn Free-swimming 136 35 27 8 206/121 Whale-shark-associated 2 2/2 Whale-associated 23 21 1 45/37 Log-associated 46 50 80 65 241/217 Total 207 106 108 73 494/377 Because tuna purse-seine fishing in the WIO is clearly seasonal (monsoons governing fishing techniques and op- erations), the data were analyzed by season. I followed Romanov's (1982) seasonal divisions, in accordance with long-term average seasonal variations in the monsoon atmospheric circulation for the WIO. The winter season (northeastern monsoon) lasts from December to March. the spring intermonsoon period falls during April and May, the summer (southwestern monsoon) lasts from June to August, and the autumn intermonsoon period lasts from September through November. The wind regime de- termines the onset and duration of the hydrological sea- sons, which do not quite coincide with seasons of atmos- pheric circulation owing to a considerable time lag of the processes occurring in the ocean. However, the wind re- gime is instrumental in determining the tactics of purse seining for tuna; therefore I used seasonal strata based on atmospheric rather than on hydrological processes. The spatial and temporal distribution of catch and ef- fort for the Soviet tuna purse-seine fishery in the Indian Ocean was determined from data in the YugNIRO data- base, a collection of daily radio reports from vessels fishing in the area from 1983 until the mid- 1990s. i" The catches reported by the author's estimates varied by 96-99'7f dur- ing 1985-91, decreasing to 71% in 1992. This study did not take into account reflagging of some Soviet (from 1992 — Russian) vessels with the Liberian flag, and the vessels' nationality was defined in this study by the loca- tion of their shipowners. Analysis of fleet activity and ex- '" Daily information on fishing activity of these vessels in the Indian Ocean in 1983-84 and since 1995 is not available. trapolations of results were made on the assumption that the operations and procedures on vessels that did not car- ry observers did not differ from the operations and proce- dures on vessels with an obsei"ver aboard; similarly it was assumed that the species composition of the catch from these vessels did not differ. Some of the bycatch was retained on board the fishing vessels. Unused bycatch was discarded in the ocean. The observers usually did not record the levels of discards, and it was not possible to assess quantitatively the discards of tuna and associated species. Average values are presented as arithmetic means, plus or minus 95% confidence intervals for estimated values. Estimates of unrecorded bycatches for all fishes, except tu- nas, are provided in numbers and metric tons per positive set and per 1000 t of target species. Results Primary data and adequacy of samples A total of 494 purse-seine sets were sampled and 377 posi- tive sets were analyzed. The total catch in the sets that were sampled amounted to 7713 t. The distribution of sets sampled by years, seasons, and the types of schools is given in Tables 1 and 2. The catch sampled by type of school is presented in Table 3. The obsei-ver coverage rate varied from 0% (no obsei-v- ers at sea) to 75% and averaged 14% during 1986-92. Dur- ing the periods when observers were on board, the cover- age rate averaged 30% and varied from 5% to 75%. The spatial distribution of sampled sets agi-eed quite well with Romanov Bycatch In the tuna purse seme fisheries of the western Indian Ocean 93 S20 35E 40E 45E 50E 55E 60E 65E 70E 35E 40E 45E 50E 55E 60E 65E 70E Figure 1 (A) Fishing effort distribution I0=noon positions of vessels on fishing days with sets) of the Soviet tuna purse- seine fishery in 1985-94; (B-Di sampled set positions: (B) on free-swimming schools that were sampled; (C) on whale-shark (A) and whale-associated schools (x); (D) on log-associated schools. The shaded area represents the region of the main international tuna purse-seine fishing activity in the WIO. according to Ardill." Table 3 Sampled catch I metric tons) by season and type of school. Type of school Seasons Total Winter Spring Summer Autumn Free-swimming 1884 249 73 24 2230 Whale-shark-associated 28 28 Whale-associated 584 467 4 1055 Log-associated 925 785 1156 1534 4400 Total 3421 1501 1213 1558 7713 the distribution of the total fishing effort of the Soviet fleet in the WIO (Fig. 11. Sampled sets were distributed throughout the region of the principal international tuna purse-seine fishing activity in the WIO (Aj-dill"). Thus, I Ardill, J. D. 1995. Atlas of industrial tuna fisheries in the Indian Ocean ( IPTP/95/AT/3 ). IPTP, Colombo, Sri Lanka, 138 p. FAO. Viale delle Terme di Caracalla, 00100, Rome. Italy 94 Fishery Bulletin 100(1) Total number of sampled sets and average annual fishing effort by seasons 250 200 150 - Two whale-shark assocrated sets ] Log-associaled ] Whale-associated I Whale-shark-associated 9 Ffee-swimming - Average effort (fishing days) - Average effort (sets) E 100 Winter Spring Summer Autumn Total sampled catch and average annual catch (t) by seasons Winter Spnng Summer Autumn 500 450 400 350 300 250 200 150 -I- 100 50 B 3 500 -T 1 - 4 000 \ 1 1 Log-associated 1 1 Whale-associated Whale-shark-associated BIB Free-swimming „,4.„ Average catch - 3 000 - 2 500 - 28 1 caught in whale- shark-asscx;iated sets - 3 500 - 3 000 sz u S 2 000 - ^V < - - 2 500 > \ t caught in whale- associated set * - - 2 000 :uras iixyniuhu.'i Rafinesque, 1809 + hiiriit: spp. 9 Carcharhinidae Cai-charhmiix falcifnrmis {Bihron. 1839) + + + C. longimaniis (Poey, 1861 ) + + + r'C. obsciirus (LeSueur, 1818) + ? +7 Ccirchaihinus spp. 9 9 Sphyrnidae Sphyrna lewini i Griffith & Smith. 1834) -h Sphyrna spp. + Exocoetidae sp. + Belonidae sp. + TylosuruK cruciidilus (Peron & LeSueur. 1821) + Lampidae Lcimpris giittatiis (Brunnich, 1788) + SphjTaenidae Sphyraena barracuda (Walbaum, 1792) + Sphyraena spp. + Carangidae Caranx spp. ■f Decapterus macarelliis Cuvier, 1833 + Decapterus spp. + Elagatis bipmnulata (Quoy & Ganiiard, 1824) + + Seriola spp. + + Naucratea ductor (Linnaeus. 1758) + Coryphaenidae Corypliaena hippuriis Linnaeus, 1758 + + Coryphaena spp. + Kyphosidae Kyphofiiis cinerasccns (Forsskal, 1775) + Gempylidae Gempy/iis serpens Cuvier, 1829 + Ruvettus pretiosus Cocco, 1829 + Ephippididae Platax spp. + + Scomberomoridae Scomberomortisconiiiicrsoii (Lacepedc, 1800) + Scomberomonis spp. ■^ Scombridae AcanthocybiiiDi solaitdri (Cuvier. 1831) + continued Romanov: Bycatch in the tuna purse-seine fisheries of the western Indian Ocean 97 Table 4 (continued) Family and species School type Free-swimmins Whale-associated Log-associated Pisces 1 continued 1 Aiixis rochvi (Risso, 1810) + Aiixis thazard (Lacepede, 1800) + -1- + Euthynnus affinis (Cantor, 1849) + Katsiiironiix pplamis (Linnaeus, 1758) + + + Tluiiinut: alalunga (Bonnaterre, 1788) + -f- Thuiiniis albavarea (Bonnaterre. 1788) + + + Tht/nrius obcsiis (Lowe, 1839) + + + Istiophoridae laliophorus platypterus iShaw & Nodder. 179 2) + Makaira indica (Cuvier, 1832) + + M. mazara (Jordan et Snyder, 1901) + + Makaira spp. + + Tctrapti/riis audax (Philippi, 1887) + Xiphiidae Xiphias g/adius (Linnaeus, 1758) + Nomeidae Cuhiceps paiuiradiatus Gunter, 1872 + Balistidae Canthidennis inaculatus (Bluch, 1786) + + Monacanthidae Alutcnis monoceros (Linnaeus, 1758) + Alutcnit: spp. + Diodontidae Diudon spp. + + > Mammalia Balaenopteridae Balacnoptera borealis Lesson, 1828 + Salpae + Ctenophora + Chelonidea ^■ Number of species (taxa) 19 17 45 ' Recorded in whale-shark-a.ssociated schools. Table 5 Average tuna catch per positive set (t) by "Rodina -type Soviet vessels in the western Indian Ocean (total and by species). YFT = yellowfin tuna, SKJ = skipjack tuna, BET = bigeye tuna, ALB =albacore, FRI = frigate tuna. KAW = kawakawa. + = catch was <0.001 t. Type of school Total Species YFT SKJ BET ALB FRI KAW Free-swimming 18.4 ±5.2 14.7+4.9 2.8 ±1.7 0.8 ±1.0 0.03 ±0.03 0.05 ±0.06 — Whale-associated 31.0 ±9.3 9.8 ±4.3 18.3 ±8.5 2.0 ±2.4 — 0.2+0.2 — Log-as.sociated 20.6 ±3.2 4.9+0.9 13.9 ±2.7 0.6 ±0.2 0.04 ±0.04 0.3 ±0.3 0.001 +0.001 Total 20.6 ±2.7 8.6 ±1.8 10.5 ±1.9 0.8 ±0.4 0.03 ±0.03 0.2 ±0.2 + 98 Fishery Bulletin 100(1) Table 6 Estimates of the bycatch (t) of various species (groups) of marine animals by school type The numerator is the average values per a positive set, the denominator is the average values per lOOOtoftai get species. + = catch was <0.001 t. School type' Free- Whale- Log- All types Species or group of species swimming associated associated of schools Billfishes (Istiophoridae, Xiphiidae) 0.016/0.89.5 0.006/0.218 0.019/1.008 0.017/0.880 Wahoo (A. solandn) — — 0.031/1.621 0.018/0.934 Sharks (Lamnidae, Carcharhinidae, Sphyrnidael 0,02.3/1.296 0.289/10.302 0.17.5/9.288 0.151/7.938 Rainbow runner (£. hipinnulata) 0.001/0.0.54 — 0.19.5/10.314 0.114/5.962 Dolphinfishes (C. hippurus) +/0.027 0.001/0.0.51 0.191/10.098 0.111/5.836 Barracuda (S. barracuda) — — 0.002/0.132 0.001/0.076 Triggerfishes (C. maculatua.Alutcrus spp.l +/+ — 0.137/7.277 0.080/4.195 Mackerel scad (£). macarf/lus) — — 0.0093/0.491 0.00.5/0.283 Mantas, mobulas (Mobulidae) 0.020/1.128 0.009/0.318 0,002/0.126 0.009/0.455 Sea turtles — — +/0.025 +/0.014 Other bycatch +/0.002 +/0.003 0.018/0.958 0.011/0.553 r For positive set I For 1000 t of target species 0.060+0.031 3.403 +2.770 0.306 ±0.344 10.891 ±15.787 0.780 ±0.144 41.337 ±14.281 0.518 ±0.099 27.127 ±8.869 ' Because of the small sample size, estimates of bycatch for whale-shark-associated schools are not presented in the Table, 70 -| 60 - 50 - 40 I 30 - 20 10 1 i Free- Whale- Log- swimming assoc ;iated assoc ;iated All types Figure 4 Bycatch (t) per 1000 t of target species by school type for the Soviet tuna purse-seine fishery. Dots are means, bars are 95'>'f confidence intervals. sible. Whales often remain in the net until the end of purs- ing and then escape from the purse seine by either diving under the purse line, by ramming through the net wall, or by sinking the corkline (a rare occurrence). Observers registered a single case of entanglement in the net and subsequent death of a young sei whale about 10 m in length and about 12 t in weight. The dead animal was taken up on the vessel's deck, released from the purse seine, and discarded into the ocean. It is not possible to as- sess the frequency and probability of whale mortality by the purse-seine fishery in the WIO. There were 1 7 species ( or groups ) of marine animals iden- tified in the catches of whale-associated schools (Table 4). Salps, ctenophores, and batfish (Platax spp.) were consid- ered accidental bycatch, whereas long-finned fathead (Cu- biceps pauciradiatus) was a prey item of both tunas and whales. Nontuna bycatch in this type of association aver- aged 0.306 ±0.344 t for a positive set or 10.891 ±15.787 t per 1000 t of target species (Figs. 3 and 4). Sharks of the genus Cairharhiniis and Isui-iis made up the bulk of the bycatch in whale-associated school .sets (0,289 t/10.302 t) (Tables 4 and 6). Log-associated schools Log-associated schools are one of the predominant school types found in the WIO all year round (Table 2, Fig, 2, A and B). Sets on log-associated schools were made through- out the sampling area as far south as 15°S (Fig. ID). In log-associated schools the bulk of the catch were skipjack, yellowfin, and bigeye tunas — 6Ty( , 2A'7i . and y?( . respec- tively (Table 5). Log-associated schools in all cases con- sisted of several fish species. Bycatch was found in 93% of the sets, and nontuna bycatch in 87%. The absence of bycatch was rare, observed only during successive sets on the same floating object. The species composition associated with floating objects was the most diverse of any set type and included 45 spe- cies (or higher taxa of fishes) (Table 4). Nontuna bycatch was at its highest in log-associated sets, as much as 0.780 ±0.144 t per positive set or 41.337 ±14.281 t per 1000 t of target species (Figs. 3 and 4). The bulk of the bycatch in sets on log-associated schools was made up of rainbow runner. Romanov: Bycatch in the tLina purse seine fisheries of the western Indian Ocean 99 Elcii^atis hipinnulata (0.195 t/10.314 t), common dolphin- fish. Coryphaena hippiirus (0.191 1/10.098 t), triggerfish of the genus Canthidermis (0.137 t/7.277 t), sharks of the ge- nus Carcliarhinits (O.n.'i t/9.28cS t), wahoo. Acaiithocyhi- uni solandri (0.031 t/1.621 t), billfishes of the genera AUik- aira and Tetrapturus (0.019 t/1.008 t), and mackerel scad, Decapterus macarclliis (0.0093 t/0.491 kg). One capture of a sea turtle (unknown species) was recorded (Tables 4 and 6). All types of schools Considering all school types in the aggregate, skipjack, yel- lowfin, and bigeye tuna prevailed in the catch — Sf/r , 429r, and 4'?; by weight, respectively (Table 5). Albacore repre- sented a mere 0.2'7(, frigate tuna 0.9%, and kawakawa, Etithyninis affinls. less than 0.1%. Nontuna bycatch accounted for less than 3'^f of the catch. On the average, there was 0.518 ±0.099 t of nontuna by- catch caught per positive set, or 27.127 ±8.869 t per 1000 t of target species (Fig. 3). Bycatch levels by species (groups) are given in Table 6. Discussion The lowest fish bycatch in the WIO tuna purse-seine fish- ery was taken from free schools (mainly carcharhinid sharks and Mobulidae rays) (Figs. 3 and 4, Tables 4 and 6). Bycatch of fishes was highest and most diverse from catches on log-associated schools. Rainbow runner, common dolphinfish. triggerfish. carcharhinid sharks, wahoo, bill- fishes, and mackerel scad were predominant. Whale-asso- ciated schools were characterized by an intermediate level of bycatch (mainly carcharhinid and lamnid sharks) (Figs. 3 and 4, Tables 4 and 6 1. It is interesting to compare the bycatch rates obtained in this study with those published for other regions. The principal bycatch fishes in the Pacific (Bailey et al., 1996; Hall, 1996, 1998; Anonymous, 1997) are the same as those presented here. Bycatch levels are known to vary consid- erably by year, area, fleet (Bailey et al, 1996; Hall, 1996; Anon., 1997), and school type; this variability hampered direct comparisons of the results from the present study with those from published data. However, for the purpose of comparison, I pooled my estimates by gi'oups in accor- dance with the published data (Bailey et al., 1996; Hall, 1996, 1998; Anonymous, 1997). Bycatch levels per set and per 1000 t of target species for various regions of the Pa- cific and my estimates for the Indian Ocean are on the same order of magnitude for most groups in similar types of associations (Figs. 5 and 6). I also attempted to estimate the unrecorded bycatch by the purse-seine fleets of the principal fishing nations of the WIO by a comparison of fishing tactics. The Soviet fleet in the WIO made an equal proportion of sets on free-swim- ming schools and on log-associated schools during the year (Table 2). Seasonally they switched effort from sets on free-swimming schools to those on log-associated schools (Fig. 7, A and B).The fishing practices of French and Span- ish tuna seiners showed similar seasonality until the niid- 1990s (Anonymous;'"'"' 1*^ Planet;'^'" Moron'-'). The fishing tactics of the Japanese (Hallier;^'' Okamoto and Miyabe-') and Mauritian (Norungee et al.;-- No,.yn. gee and Lim Shung-') purse-seine fleets differed consider- ably from that described above. Japanese and Mauritian vessels made sets on log-associated schools all year round, with single instances of sets on other schools types. Only two school types (log schools and free schools) have been described by Hallier;-" Hallier;-^ Parajua Aranda;^'' 'Anonymous. 1992. Report of the workshop on stock assess- ment of yellowfin tuna in the Indian Ocean, Colombo, Sri Lanka. 7-12 October 1991, 90 p. |IPTP/91/GEN/20.| FAG, Viale delle Termc di Caracalla, 00100, Rome, Italy. • Anonymous. 1994a. Report of the expert consultation on Indian Ocean tunas, .5th session. Mahe, Seychelles, 4-8 Octo- ber 199.3, 32 p. IIPTP/94/GEN/22.1 FAO. Viale delle Terme di Caracalla. 00100, Rome, Italy. ' Anonymous. 1994b. National report of Spain. In Proceed- ings of the expert consultation on Indian Ocean tunas, 4-8 October. 1993 (J. D. Ardill. ed. i. p. 44-47. IPTP Coll. Vol. 8., TWS/93/1/14. FAO. Viale delle Terme di Caracalla. 00100, Rome. Italy. ' Pianet. R. 1994a. Purse seine fishery trends in the western Indian Ocean from data collected in Victoria (Seychelles), 1984-1992. //( Proceedings of the expert consultation on Indian Ocean tunas, 4-8 October, 1993 (J. D. Ardill. ed.). p. 41-44. IPTP Coll. Vol. 8.. TWS/93/1/13. FAO. Viale delle Terme di Caracalla, 00100. Rome. Italy. ' Pianet. R. 1994b. National report of France. //! Proceedings of the expert consultation on Indian Ocean tunas. 4-8 October. 1993 (J.D.Ai-dill,ed.i.p.48-.52. IPTP Coll. Vol. 8.TWS/93/1/16. FAO, Viale delle Terme di Caracalla, 00100, Rome. Italy ' Moron. J. 1996. National report of Spain. In Proceedings of the expert consultation on Indian Ocean tunas. 6th session, Colombo. Sri Lanka. 2.5-29 September. 1995 (A. A. Anagnuzzi, K. A. Stobberup, N. J. Webb, eds. I, p. 63-69. IPTP Coll. Vol. 9. FAO, Viale delle Terme di Caracalla, 00100, Rome, Italy ' Hallier. J.-P. 1991. Tuna fishing on log associated schools in the Western Indian Ocean: an aggregation behaviour. /;; IPTP Coll. Vol. Work. Doc, Vol. 4. p. 325-342 [TWS/90/66.1 FAO, Viale delle Terme di Caracalla. 00100. Rome. Italy. Okamoto. H., and N. Miyabe. 1996. Review of Japanese tuna fisheries in the Indian Ocean. In Proceedings of the expert consultation on Indian Ocean tunas. 6th session, Colombo. Sri Lanka. 25-29 September. 1995 (A. A. Anagnuzzi, K. A. Stobb- erup, N.J. Webb, eds.), p. 15-21. IPTP Coll. Vol. 9. FAO, Viale delle Terme di Caracalla, 00100, Rome, Italy ' Norungee, D., A. Venkatasami, and C. Lim Shung. 1994. Catch and landing statistics of the Mauritian tuna fisheries (1987-1992) and an analysis of the skipjack tuna catch of the Mauritian purse seine fishery (1987-1993). In Proceed- ings of the expert consultation on Indian Ocean tunas. 5th ses- sion. Mahe. Seychelles. 4-8 October. 1993 (J. D. Ai'dill. ed.), p. 266-273. IPTP Coll. Vol, 8. TWS/93/4/5. FAO. Viale delle Terme di Caracalla. 00100. Rome. Italy. ' Norungee. D.. and C. Lim Shung. 1996. Analysis of the purse seine fishery of Mauritius, 1990-1994, and comparison of catch rate and species composition of catches of Mauritian purse seiners to those of French fleet. In Proceedings of the expert consultation on Indian Ocean tunas, 6th session, Colombo, Sri Lanka. 25-29 September, 1995 (A. A. Anagnuzzi, K. A. Stobb- erup. N.J. Webb. eds.). p. 1.5-21. IPTP Coll. Vol. 9. FAO. Viale delle Terme di Caracalla, 00100. Rome. Italy Hallier. J.-P 1994. Purse seine fishery on floating objects: What kind of fishing effort? Wliat kind of abundance indices? In continued 100 Fishery Bulletin 100(1) Table 7 Bycatch estimates in tons in the western Indian Ocean pu ■se-SfUie fisheries during 1985-94. MIX = fleets targe ted all t\ pes of schools (France Spain. USSR ,LOG = fleets targeted log-associated schools (Japan an d Mauriti us). Species, a groui: of species 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 MIX 913 1047 1257 1674 1622 1503 1471 1793 1796 1925 Pelagic oceanic sharks LOG 31 30 61 81 108 ISO 278 477 451 143 Total 944 1077 1318 1755 1730 1683 1749 2270 2247 2068 MIX 686 786 944 1257 1218 1129 1105 1347 1349 1446 Rainbow runners LOG 34 33 68 90 120 199 309 530 500 159 Total 720 819 1012 1347 1338 1328 1414 1877 1849 1605 MIX 671 770 925 1231 1193 1105 1082 1318 1320 1415 Dolphinfishes LOG 34 33 67 88 117 195 303 518 490 1.56 Total 705 803 992 1319 1310 1300 1385 1836 1810 1571 MIX 483 554 665 885 857 794 778 948 949 1017 Triggerfishes LOG 24 24 48 64 84 141 218 374 353 113 Total 507 578 713 949 941 935 996 1322 1302 1130 MIX 108 123 148 197 191 177 173 211 211 227 Wahoo LOG 5 5 11 14 19 31 49 83 79 25 Total 113 128 159 211 210 208 222 294 290 252 MIX 101 116 139 185 180 167 163 199 199 213 Billfishes LOG 3 3 7 9 12 20 30 52 49 16 Total 104 119 146 194 192 187 193 251 248 229 MIX 52 60 72 96 93 86 84 103 103 110 Mobulas and in intas LOG <1 <1 1 1 1 2 4 6 6 2 Total 53 60 73 97 94 88 88 109 109 112 MIX 33 37 45 60 58 54 53 64 64 69 Mackerel scad LOG 2 2 3 4 6 10 15 25 24 8 Total 35 39 48 64 64 64 68 89 88 77 MIX 9 10 12 16 16 14 14 17 17 19 Barracudas LOG <1 <1 1 1 1 3 4 7 6 2 Total 9 11 13 17 17 17 18 24 23 21 MIX 64 73 88 117 113 105 102 125 125 134 Other fishes LOG 3 3 6 8 11 18 29 49 47 15 Total 67 76 94 125 124 123 131 174 172 149 MIX 3120 3576 4295 5718 5541 5134 5025 6125 6135 6574 Total nontuna bycatch LOG 137 134 273 360 479 799 1239 2121 2004 638 Total 3257 3710 4568 6078 6020 5933 6264 8246 8139 7212 Anonymous;'^ ^'' "' Planet;'"- 1- Hastings and Domingue;-'' and Moron'" for the tuna purse-seine fishery in the In- 24 i""itinui'di Proceedings of the expert consultation on Indian Ocean tunas, 5th session, Mahe, Seichelles, 4-8 October. 1993 (J. D. Ai-dill,ed.), p. 192-198. IPTP Coll. Vol. 8.,TWS/9.3/2/25, FAO. Viale delle Terme di Caracalla. 00100, Rome, Italy 2= ParajuaAranda.J. I. 1991. Spanish status report of vellowfin tuna fishery 1984-1990. In IPTP Coll. Vol. Work. Doc, Vol. 6, TWS/91/13, p. 99-130. FAO, Viale delle Terme di Caracalla, 00100, Rome, Italy '^^ Hastings, R. E., and G. Domingue. 1996. Recent trends in the Seychelles industrial fishery. In Proceedings of the expert consultation on Indian Ocean tunas, 6th session, Colombo, Sri Lanka, 25-29 September, 1995 (A. A. Anagnuzzi, K. A. Stob- berup, N. J. Webb, eds), p. 97-109. IPTP Coll. Vol. 9. FAO, Viale delle Terme di Caracalla, 00100, Rome, Italy. dian Ocean. Free schools in these analyses included all types of associations with marine animals. The propor- tion of sets of the French fleet on other types of schools and on resulting catches is not known. Cort (1992) pre- sented such data for Spanish vessels, based on fishing logbooks. Therefore, I used the observers data of the Sey- chelles Fishing Authority (SFA) (Cort, 1992) for the ves- sels of France, Spain, Japan, and USSR to assess these values in the WIO. The percentage of sets on whale-asso- ciated schools varied from 1.7% to 8.8% in 1986-90, the percentage among positive sets was from 1.2% to 9.1%, and the catch from such schools was 1.6% to 7.8% (cited from Cort, 1992). These values are slightly lower than the observer data I report in the present study (9%, 10%, and Romanov Bycatch in the tnna purse seine fisheiies of the western Indian Ocean 101 1,000 1 900 800 - 700 S 600 f 500 z 400 300 200 100 Billfisties A DThis study I-ATTC1993 Dl-ATTC 1994 Free-swimming Log-associated Marine mammals Small fishes 2 8 DThis study l-ATTC 1993 Dl-ATTC 1994 00 42 24 Free-swimming Log-associated H/larine mammals 450 400 350 300 250 200 150 100 50 Stiarl^s Free-swimming Log-associated IVIarine mammals Dolptiinfisties, wahoo, rainbow runners D 07 DThis study l-ATTC 1993 Dl-ATTC 1994 02 1 Of Free-swimming Log-associated Manne mammals Sea tunies DThis study l-ATTC 1993 Dl-ATTC 1994 Free-swimming Log-associated t^anne mammals Figure 5 Bycatch levels in numbei's per set by groups of species and by types of schools in the western Indian Ocean and eastern Tropical Pacific (Anonymous, 1997). 14^'r . respectively), which is explained by the fact that the SFA data included Japanese vessels known to fish on log- associated schools only. Nevertheless, the SFA values and those from our obser\'ers were on the same order of mag- nitude. Proceeding from this, I estimated the ratio of sets on various school types and the magnitude and species composition of bycatch by the French and Spanish ves- sels. These values were close to those for the Soviet fleet employing similar fishing tactics.-^ Thus, the average bycatch estimates presented in this study can be extrapolated for this period to the total WIO purse-seine catch of principal fishing nations targeting all types of schools.-^** Estimates of bycatch from log-associat- ed schools, I believe, can be extended, with some caution, to the pooled purse-seine catch of Japan and Mauritius in the WIO. The annual purse-seine catches of yellowfin and skip- jack tunas by fleets targeting all types of schools (France, Data from logbooks (Cort, 1992) show a lower proportion of sets and of catches on whale-associated .schools for Spanish vessels, but in the author's view a comparison of data collected in the same way (by observers) is preferable. France and Spain (along with catch from the vessels from these two countries flying "flags of convenience" [Panama, Cote d'lvoire, and recently Belize] and applying the same fishing tactics), and USSR (recently Russia or Liberia). 102 Fishery Bulletin 100(1) Billfishes 1 ^ 1 — 1 ^ 1 — DThis study Hall, 1996 Sharks and rays 700 -, 600 500 300 - 200 100 B J DThis study [■^Hall, 1996 I Free-swimming Log-associated Marine mammals Dolphinfishes Free-swimming Log-associated Marine mammals Wahoo 5,000 4,000 23,000 a) n E ^2,000 1,000 2,500 2,000 2! 1,500 OJ E z 1,000 500 165 nThis study Hall, 1996 76 24 Free-swimming Log-associated Marine mammals Rainbow runners E DThis study m^all, 1996 24 5 36 5 00 00 Qi 2,500 2,000 >. 1,500 ) ) ' 1,000 500 10,000 9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 D 26 7 DThis study Hall, 1996 1 00 06 Free-swimming Log-associated Marine mammals Tnggerfishes F 5 75 6 DThis study Hall, 1996 1 00 74 Free-swimming Log-associated Manne mammals Sea turtles Free-swimming Log-associated Marine mammals Total by-catch Free-swimming Log-associated Marine mammals Free-swimming Log-associated Marine mammals Figure 6 • A-G), Bycatch levels, in numbers per 1000 t of target species, by groups of species and by types of schools in the western Indian Ocean and eastern Tropical Pacific; (H) bycatch levels, in tons per set by types of schools in the western Indian Ocean and western Pacific Ocean. Romanov: Bycatch in the tuna purse-seine fisheries of the western Indian Ocean 103 Spain, and I'SSRV-''' In the WIO ranged between 115,000 and 242.000 t in 1985-94 (Anonymous'"). Japanese and Mauritian catches varied from 3000 to about 51,000 t. Based on these vakies, the estimated bycatcli was 3257 to 8246 t of various fishes during the same period (Table 7), These fishes could serve as food for the coastal countries of the area. Estimat- ed bycatch in numbers is presented in Table 8. Turtle bycatch and whale mortality in purse seines are also possible in the WIO, but the probability of the latter is very low. No instances of whale mortal ity have been recorded earlier for tuna purse-seine fisheries in other areas (Northridge, 1984, 1991a. 1991b: Medina-Gaertner and Gaertner, 1991; San- tana et al., 1991; Cort, 1992; Cayre et al.. 1993; Bai- ley et al.. 1996). No avian mortality by the Soviet tuna purse-seine fishery has been noted by observ- ers. A similar fact was reported for the western Pa- cific (Bailey et al.. 1996). Target fishing for rainbow runner, dolphinfish, triggerfishes, wahoo, mackerel scad, and barracuda is not conducted in the WIO, and these fish are taken only as bycatch. Their bycatch levels, estimated in this study, do not seem to endanger the populations of these species. Estimated bycatch of billfishes ( 104-251 1 annual- ly) was less than I'Ti of the total catch for these spe- cies (14,000-33,000 t during 1985-94) in the WIO (Anonymous^"). The bycatch by the purse-seine fish- ery was unlikely to substantially affect the billfish stocks. Many pelagic sharks are taken as bycatch by the longline, trawl, coastal driftnet, and other fisheries, but are not recorded. The total shark catch by all fisheries may be considerable. Many shark species are characterized by low abundance, low fecundity, long life span, and conse- quently, by high vulnerability to overfishing. Underesti- mation of the removal through fisheries of a number of pe- lagic shark species, and the impact of the fisheries on their populations, may lead to a reduction in their abundance to critical levels, diminishing the biodiversity of the pelagic ecosystem of the Indian Ocean. Some part of the bycatch is released into the ocean alive, although subsequent survival rates are unknown. The lack of bycatch and discard records and estimates of survival rates of discarded animals prevents assessment of the impact of the fishery on the Indian Ocean pelagic ecosystem. Fishing tactics in the WIO have changed considerably by all principal purse seine fleets toward the extensive use of FADs in recent years (generally from 1995). The majority of Japanese vessels have left the area and have moved to the eastern Indian Ocean. Therefore estimates presented here for total WIO purse-seine fisheries are ap- 100 1 , ^ —•— Free swimming 75- -e- Log associated ; ^,_^ Percentage U1 o y<:^^^Z ^"~~~* Winter Spring Summer Autumn lOOn B —•—Free swimming Jd- 75- 0) en m 1 50- o Q) CL 25- -e- Log associated y^ Xl 0- ^~~~~~~~-~~-~,»-________ Winter Spring Summer Autumn Figure 7 (A) Percentages of free-school and log-schools sets; (B) percent- ages of free-school and log-school catch in the .Soviet tuna purse- seine fishery. plicable for a limited time span only (pre- 1995). Recent de- velopment of the WIO fisheries warrants further investi- gation of bycatches through extensive observer sampling by time-area strata. Establishing a scientific program by the Indian Ocean Tuna Commission to monitor the principal tuna fisheries in the region, by placing international scientific observers on purse-seine and longline vessels, might be the first step to- ward a more accurate assessment of the impact of bycatch- es on the epipelagic ecosystem of the Indian Ocean. This program might also lead to developing technical and man- agement measures to reduce the bycatches or to use them. The solution to the bycatch problem should take two di- rections: 1 1 an effort to reduce or eliminate bycatches of un- desired species; or 2) to use bycatch animals to make them target species. The former involves developing gear modi- fications or changes in fishing tactics. The latter involves management regulation of the fishery so that bycatch spe- cies are treated in the same way as other target species. Acknowledgments '^ Including vessels flying flags of convenience. ■"' Anonymous. 1998. Indian Ocean tuna fisheries data sum- mary, 1986-1996. Indian Ocean Tuna Commission (lOTC) data summary 18, 180 p. lOTC, P.O. Box 1011, Victoria, Seychelles. I am giateful to AtlantNIRO scientists V. F Bashmakov, G. A. Budylenko, V. Z. Gaikov, M. E. Grudtsev, to TINRO sci- entist K. A. Karyakin for their data made available to the author and to V. F. Bashinakov and G. A. Budylenko for their personal sampling efforts. I sincerely thank masters of the 104 Fishen/ Bulletin 100(1) Table 8 Bycatch estimates in numbe ■s in the western Ir dian Ocean pursu-seine fisheries during 1985-94. Codes are same as table 7. MIX = fleets targeted all types of schools (Fr ance, Spain. USSR ; LOG = fleets targeted log-associated schools (Japan and Mauri iusi. Species, a gi-oup of species 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 MIX 4.X600 52,273 62,780 83,581 80,993 75,052 73,455 89,529 89,676 96,094 Pelagic oceanic sharks LOG 2161 2100 4280 5653 7531 12,.546 19,456 33,320 31,488 10,030 Total 47,761 54,373 67.060 89,234 88,524 87,598 92,911 122,849 121,164 106,124 MIX 162,4.57 186,232 223.664 297,770 288,5,50 267,386 261,694 318,961 319,485 342,351 Rainbow runners LOG 8112 7883 16.065 21,218 28.267 47,090 73,029 125,065 118,190 37,646 Total 170,.569 194,115 239,729 318,988 316.817 314.476 334,723 444,026 437,675 379,997 MIX 107,711 123,473 148,291 197,424 191.312 177,280 173,.505 211,474 211,821 226,982 Dolphinfishes LOG 5373 5221 10,641 14,0.53 18,723 31,190 48,370 82,835 78,282 24,934 Total 113.084 128.694 158,932 211,477 210.035 208,470 221,875 294,309 290,103 251,916 MIX 621.823 712,823 856,096 1.139.747 1.104.4.58 1,023,4.50 1,001,661 1.220.857 1,222,862 1,310,387 Triggerfishes LOG 31.215 30.334 61,820 81,646 108.774 181,205 281.018 481.252 4.54,799 144,863 Total 653.038 743.156 917.916 1.221. .393 1,213,232 1,204,655 1.282.679 1.702.109 1,677,661 1,4.55,2.50 MIX 17.444 19,996 24.016 31.973 30.983 28.710 28,099 34.248 34,304 36,760 Wahoo LOG 876 851 17.34 2290 3051 5083 7883 13..501 12,7.59 4064 Total 18.320 20,847 25,750 34,263 34.034 33,793 35,982 47.749 47,063 40,824 MIX 750 859 1032 1374 1332 1234 1208 1472 1474 1580 Billfishes LOG 26 25 51 68 90 151 233 400 378 120 Total 776 884 1083 1442 1422 1385 1441 1872 1852 1700 MIX 250 286 344 458 444 411 403 491 491 527 Mobulas and manias LOG 3 2 5 7 9 15 23 39 37 12 Total 253 288 349 465 453 426 426 530 528 539 MIX 45,1.34 51,739 62,138 82.726 80.164 74,285 72.703 88,613 88,758 95.111 Mackerel scad LOG 2266 2202 4487 5926 7895 13,1.53 20,398 34,931 33.011 10,515 Total 47.340 53,941 66,625 88,652 88,059 87,438 93,101 123,.544 121.769 105,626 MIX 1350 1.547 1858 2474 2397 2221 2174 2650 2654 2844 Barracudas LOG 68 66 1.34 177 236 393 610 1044 987 314 Total 1418 1613 1992 2651 2633 2614 2784 3694 .3641 31.58 KUTF tuna seiners A. G. Burlyko. V. N. Volvach, A. A. Kiry- anov for their assistance rendered to observers in sampling. The author is grateful to V. F. Demidov, N. N. Kukharev, M, A. Pinchukov, L. K. Pshenichnov. S, T, Rebik, B, G, Trot- senko for useful discussions when preparing the manu- script and to two anonymous reviewers for their comments and suggestions. The author wishes to thank L V. Charova for translating the paper into English. Revisions and an edition of the pa- per by R. J. Olson (I-ATTC) and his corrections of English were extremely valuable. Literature cited Anonymous. 1997. Annual report of the Inter-Ainerican Tropical Tuna Commission. 1995. lATTC, La Jolla, CA. 334 p. 1998. Annual report of the Inter-American Tropical Tuna Commission. 1996. lATTC. La Jolla. CA. 306 p. 1999. Annual report of the Inter-American Tropical Tuna Commission. 1995. lATTC, La Jolla, CA, 310 p. Au. D. W. K. 1991. Polyspecific nature of tuna schools: shark, dolphin and seabird associations. Fish. Bull. 89:34.3-354. Au. D. W. K., and W. L. Ferryman. 1985. Dolphin habitats in the Eastern Tropical Pacific. Fish. Bull. 83(4 ):62,3-643. Au. D. W. K.. and R. L. Pitman. 1986. Seabird interactions with dolphins and tuna in the Eastern Tropical Pacific. The Condor 88:304-317. Bailey. K., P. G. Williams, and D. Itano. 1996. Bycatch and discards in Western Pacific tuna fisher- ies: a review of SPC data holdings and literature. South Pacific Comm. Tech. Rep. 34. Noumea, New Caledonia. 171 p. Cayre. P. J. B. Anion Kothias, T. Diouf and J. M. Stretta. 1993. Biology of tuna. In Resources, fishing and biology of the tropical tunas of the Eastern Central Atlantic, p. 147-244. FAO Fish. Tech. Pap. 292. FAO, Rome. Charat-Levy, F. 1991. The consequences of the tun;Vdolphin issue in the Romanov Bycdlch in the tLina pLiise seme fishenes of the western Indian Ocean 105 Eastern Pacific. In Tiiiia 91 Bali papers of the '2'"' world tuna trade conference Bali. Indonesia, l,'!-!!) May, 1991 iHenri dc Saram, cd.), p. 19-22. INKOFISH, Kuala Lumpur, Malaysia. Cort, J. L. 1992. Estudio de las asociaciones de tunidos, en especial la denominada "atun-delfin." Su integracion en la biologia de estos peces migi'adores. //) International Commission for the Consei-\-ation of Atlatic Tunas (ICCAT) Coll. Vol. Sci. Pap. 39(1 ):358-384. Garcia. M.. and M Hall. 1995. Spatial and temporal distribution of bycatches of yel- lowfin, skipjack, mahi-mahi and wahoo in the eastern Trop- ical Pacific's purse seine tuna fishery. In Proceedings of the 46th annual tuna conference. (A. J. Mullen, and J. Suter, eds.), p. 54. lATTC, La JoUa. CA. Hall, M. A. 1996. On bycatches. Rev. Fish Biol. Fish.. 6:319-352. 1998. An ecological view of the tuna-dolphin problem: im- pacts and tradeoffs. Rev. Fish Biol. Fish. 8:1-34. Joseph. J. 1991. The consei"vation ethic and its impact on tuna fisher- ies. /'( Tuna 91 Bali papers of the 2'"' world tuna trade con- ference Bah, Indonesia. 13-15 May. 1991 (Henri de .Saram. ed.), p. 12-18. INFOFISH, Kuala Lumpur. Malaysia. 1994. The tuna-dolphin controversy in the Eastern Tropical Pacific Ocean: biological, economic, and political impacts. Ocean Development and International Law 25:1-30. Medina-Gaertner, M. and D. Gaertner. 1991. Factores ambientales y pesca atunera de superficie en el Mar Caribe. ICCAT Coll. Vol. Sci. Pap. 36:523-550. Northridge. S. P. 1984. World review of interactions between marine mam- mals and fisheries. FAG Fish. Tech. I'a[). 251, 190 p. FAO, Rome. 1991a. An updated world review of interactions between marine mammals and fisheries. FAO Fish. Tech. Pap. 251, suppl. 1, 58 p. V\0, Rome. 1991b. Driftnet fisheries and their impact on non-target species: a worldwide review. FAO Fish. Tech. Pap. 320, 115 p. FAO, Rome. Petit M., and J. M. Stretta. 1989. Sur le comportement des bancs de thons observers par avion. ICCAT Coll. Vol. Sci. Pap. 30( 1 1:488-490. Romanov, Yu. A. 1982. Climate features. In The Indian Ocean (series: the world ocean geography) (V. G. Kort and S. S. Salnikov, eds.), p. 43-62. Nauka, Leningrad Santana, J. C, J. Ariz, and A. Delgadu de Molina. 1991. Nota sobre la presencia de mamiferos marinos en la pesquera de tunidos al cerco en el Atlantico este intertropical. ICCAT Coll. Vol. Sci. Pap. 35(1):196-198. Santana. J. C, A. Delgado de Molina, R. Delgado de Molina. J. Ariz, J. M. Stretta, and G. Domalain. 1998. Lista faunistica de las espccics asociados a las capturas de atun de las flotas de cerco comunitarias que faenan en las zonas tropicales de los oceanos Atlantico e Indico. ICCAT Coll. Vol. Sci. Pap. 48(3):129-137. Scott, J. M. 1969. Tuna .schooling terminology. Calif Fish Game 55(2): 136-140. LInited Nations. 1983. The law of the sea. Official text of the United Nations convention on the law of the .sea with annexes and index/ final act of the third United Nations conference on the law of the sea/introductory material on the convention and the conference. United Nations, New York, NY, 224 p. 106 Abstract-Thf natural diet of 506 American lobsters iHomarus america- niis) ranging from instar V (4 mm cephalothorax length. CLi to the adult stage (112 mm CD was determined by stomach content analysis for a site in the Magdalen Islands, Gulf of St. Lawrence, eastern Canada. Cluster and factor analyses determined four size groupings of lobsters based on their diet: <7.5 mm, 7.5 to <22.5 mm, 22.5 to <62.5 mm, and >62.5 mm CL. The onto- genetic shift in diet with increasing size of lobsters was especially appar- ent for the three dominant food items: the contribution of bivalves and animal tissue (flesh) to volume of stomach con- tents decreased from the smallest lob- sters (2871 and 399r, respectively) to the largest lobsters (2'* and 11'^, respec- tively), whereas the reverse trend was seen for rock crab Cancer irroratus il'.i in smallest lobsters to 539* in largest lobsters). Large lobsters also ate larger rock crabs than did small lobsters. This study is the first to examine the natural diet of shelter-restricted juveniles (SP{Js, <14.5 mm CLi, which were thought to be principally suspension feeders and to a lesser degree browsers or ambush pred- ators in or near their shelter However, at our study site no planktonic organ- isms were identified from the stom- achs of SRJs, whereas formaniferans, crustacean meiofauna, and macroalgal debris that could be derived by brows- ing, together represented only 10-14'(^ by volume of stomach contents. We infer that SRJs obtained bivalves by pre- dation and flesh by exploiting larger lobsters' meal scraps or food resei-ves. Some implications of these findings for lobster arti,ficial reef programs and for the conservation of lobster stocks are discussed. Ontogenetic shifts in natural diet during benthic stages of American lobster iHomarus amen'canus), off the Magdalen Islands Bernard Sainte-Marie Denis Chabot Division des invertebres et de la biologie expenmentale Institut Maurice-Lamontagne Peches et Oceans Canada 850 route de la Mer Mont-Joh (Qc), G5H 3Z4 Canada E-mail address (for B Sainte Mane) Sainte Maneadfo mpo gc ca Manuscript accepted 3 October 2001. Fish. Bull. 100(l):106-116i2002). American lob.ster, Homaiiis amcrica- niis, is a long-lived, dominant predator in temperate coastal waters of eastern North America (Elner and Campbell, 1991; Ojeda and Dearborn, 1991). After the lar\'al phase, lobsters settle and spend much of their time in burrows or natural shelters (Cobb, 1971; Lawton, 1987; Barshaw and Bryant-Rich, 1988). However, laboratory and in situ obser- vations indicate that benthic lobsters pass through successive life-history phases as they grow in size, changing from a shelter-restricted habit to a more overt lifestyle involving daily forays and seasonal migrations away from shelter (Cooper and Uzmann, 1977; Cobb and Wahle, 1994). A variety of classifica- tions have been proposed for these suc- cessive ontogenetic phases. The latest scheme, by Lawton and Lavalli ( 1995), recognizes five life-history phases: shel- ter-restricted juvenile (SRJ, -4-14 mm cephalothorax length, CL), emergent juvenile (-15-25 mm CL), vagile juve- nile (-25 mm CL to size of physiological maturity), adolescent, and adult. In several decapod crustaceans, diet changes as individuals grow and be- come more mobile and their chela size and strength increases (e.g. Lee and Seed, 1992; Freire et al., 1996). Such dietary shifts should occur in the lob- ster as well, especially considering this species' changing dependency on shel- ter which, in turn, has implications for foraging range and accessibility of prey types (Elner and Campbell, 1987; Lawton, 1987). Some studies of the nat- ural diet of lobsters 12-125 mm CL have found little or no differences in the identity or in the frequency of food items that were ingested by different size groups (Weiss, 1970; Ennis, 1973; Hudon and Lamarche, 1987). However, other studies have pointed to changes in the identity and especially in the fre- quency of food items ingested by dif- ferent lobster size groups. Carter and Steele ( 1982b). using their own results and data from nonconcomitant studies conducted at different sites in New- foundland (Squires, 1970; Ennis, 1973), have suggested that lobsters of 12-73 mm CL consume sea urchins, ophi- uroids, and mussels more frequently than larger (adult) lobsters. Scarratt ( 1980) reported that lobsters consumed more crabs, mussels, and fish, but fewer echinoderms. as they grew in size and approached maturity. This trend was attributed to differential accessibility of prey. Elner and Campbell (1987) in- dicated that the stronger chelae of larg- er lobsters would enable them to crush prey that are protected by heavy shells, such as gastropods and bivalves, more so than the chelae of smaller lobsters. The natural diet of SRJ lobsters has not been examined to date (Lawton and Lavalli, 1995). e.xcepting rare spec- imens of 12-14 mm CL. The feeding appendages of SRJs are capable of capturing and processing both plank- tonic and benthic organisms ( Lavalli and Factor, 1995). From laboratory ob- servations, several authors have pro- posed that SRJs may live primarily as suspension-feeders, and to a lesser de- gree as browsers, within the shelter or as ambush predators at the shelter's entrance (Barshaw and Brvant-Rich, Sainte Mane and Chabot Natural diet of Homarus americanus off the Magdalen Islands 107 1988; Barshaw, 1989; Lavalli and Barshaw, 1989; Lawton and Lavalli, 1995 ). Wahle ( 1992 ) offered a conceptual mod- el suggesting that lobsters shift from a cryptic to a wide- roaming behavior as predation risk becomes offset by the need for a high-energ\' diet that cannot be satisfied through shelter-restricted feeding. Our study was conducted at the Magdalen Islands, east- ern Canada, to resolve the natural diet of SRJ lobsters and to compare it with that of larger lobsters by using stomach content analysis. We found a gradual ontogenetic shift in lobster diet over the size range of 4 to 112 mm CL. SRJs were carnivorous and probably derived their meals main- ly through predation and scavenging. We also determined the predator-prey size relationship for one of the lobster's preferred and most important prey, i.e. Atlantic rock crab. Cancer irroratus (Reddin, 1973; Evans and Mann, 1977; Carter and Steele, 1982a I. Materials and methods The study site was a narrow 2-km rocky section (47°14.5'N. 6r50..5' to erSl.S'W) of the south shore of Baie de Plai- sance, Magdalen Islands, eastern Canada. This site corre- sponds to the Butte-a-la-Croix location that Hudon (1987) determined to be a settlement ground for lobster Divers collected lobsters by hand or by suction-sampling at depths of 1 to 7 m. Lobsters were processed live usually within minutes and at most two hours after collection. The sex of collected specimens was determined and their CL was measured to the nearest 0.1 mm with a vernier caliper. Lobsters that were not berried and that were judged to be intermolt, based on criteria of shell hardness, coloration, and fouling in Aiken (1980), were dissected to remove the stomach which was preserved in buffered formalin diluted to 4'7(- in seawater Stomachs with calcified gastroliths were subsequently disregarded, thereby effectively eliminating from the present study all premolt lobsters from stage D'-5 (=Dq) on (Aiken, 1980). The resulting sample con- sisted of 471 stomachs from lobsters of 7-112 mm cepha- lothorax length (CL) collected from 24 July to 31 October 1996, and of 35 stomachs from lobsters of 4—12 mm CL col- lected between 4 August and 13 September 1997. The 1997 lobsters were added to improve coverage of stomach con- tents of the early juveniles because very little settlement occurred in 1996 iSainte-Marie et al., 2001). There was no commercial fishery during the sampling periods; therefore items in lobster stomachs were not discards or bait. In the laboratory, stomachs were opened and their con- tent was emptied into dishes for examination under a Wild M8 compound microscope (10-50x). Identity of food items was determined to the lowest taxonomic level possible, based on comparisons with illustrations in literature and samples of benthic and pelagic fauna from our study site. Particular care was taken when examining the stomach contents of lobsters <12 mm CL; for these stomach con- tents we often resorted to higher magnification (>100x) with a Leitz Dialux 20 microscope. The contribution of each food item, exclusive of miner- als and nylon debris, to the volume of stomach contents of each lobster was visually scored from to 10, by 10% increments (0=0% of volume, 1=1-10%, 2=11-20%, etc.). The total for all food items could exceed 10, for example, if more than two minor food items each were scored 1 in addition to one predominant food item that was scored 8. In such cases, the corrected contribution of each food item was obtained by dividing its score by the sum of scores for all organic food items in a given stomach. Corrected volu- metric contribution of each food item was expressed as a proportion of stomach content volume. To obtain information on the size spectrum of rock crab consumed by lobsters, we established predictive (least squares) linear regressions (Sokal and Rohlf 1995) be- tween 30 measurements of distinctive hard body parts and cephalothorax width (CW) of 26 crabs ranging from 7 to 62 mm CW (following the approach in Lovrich and Sainte-Marie, 1997). All the predictive regressions were highly significant (/■-=0.970-0.999. P<0.001). When rock crab remains were encountered in lobster stomachs, dis- tinctive hard body parts were measured with an eyepiece micrometer to estimate crab CW from predictive regres- sions. When more than one body part could be measured, the crab's CW was determined as the mean of the various estimates unless it was obvious that multiple crabs had been ingested. Such was considered to be the case when more than two similar fragments of a paired structure (e.g. eyes or claws) were found in one lobster stomach or when there was considerable divergence among crab CW estimates based on different body parts. The functional re- lationship between the CW of rock crab prey and the CL of lobster predators was established with a model II regi'es- sion (Laws and Archie, 1981; Sokal and Rohlf 1995). The stomach contents, once identified and scored for volume, were transferred separately to preweighed trays, dried to constant mass at 60°C, and weighed to the near- est mg. Dry mass was not obtained for eight stomach con- tents because of manipulation errors. The allometric rela- tionship between the dry mass of stomach contents and lobster CL was established by least squares linear regres- sion, after logarithmic transformation of both variables. Diet was described by occurrence, volumetric contribu- tion, and the specific abundance of food items in the stom- achs of lobsters grouped into 5-mm CL size classes (2.5 to <7.5 mm, 7.5 to <12.5 mm, etc.). Percent occurrence (PO) was the percentage of stomachs in one size class that con- tained a given food item. Volumetric contribution ( VC ) was the average of corrected contributions of each food item to the stomachs of all lobsters in a given size class. Spe- cific abundance (SA) was the average volumetric contribu- tion of a food item determined only for lobsters that had this food item in their stomach. This index is useful for food items with a low average volumetric contribution be- cause it allows the distinction between the case when few animals consume large quantities of a given food item or when many animals consume small quantities of the same food item (Amundsen et al., 1996). The mathematical rela- tionship of the three indices is SA = VC x 100/PO. To assess how the overall diet varied with lobster size, and thus whether or not there were size-related shifts in diet supporting the ontogenetic phases of lobster, we 108 Fishei-y Bulletin 100(1) performed a cluster analysis (Ward's minimum variance method) on the volumetric contribution of food items per lobster size class, after standardization. A sudden increase in the joining distance of the clustering sequence repre- sented by the dendrogram represents a natural cutting point for the determination of meaningful clusters (SAS Institute. 1995). In addition, a factor analysis (VARIMAX rotation of the first three principal components) was per- formed on the correlation matiix of the volumetric contri- bution of food items for each 5-mm-CL size class of lob- sters. Cluster and factor analyses were done with JMP statistical software (SAS Institute, 1995). Relationships between volumetric contribution and lob- ster CL were described by least squares linear regression for bivalves, rock crab, and flesh. Relationships between percent occurrence of bivalves and rock crab were described by locally weighted (lowess) regi"ession with a SOf smootii- ing factor, and by least-squares regression for flesh. Results Sample composition, stomach fullness, and types of food items The 506 lobsters retained for analyses varied in size from 4.3 to 112.4 mm CL (median=35.6 mm CL). Most size classes contained more than 25 lobsters (Table 1). The smallest size class (2.5 to <7.5 mm CL) contained only 16 lobsters with a median of 7.0 mm CL; therefore we refer to this group of lobsters as the 7-mm-CL size class. The 21 lobsters >67.5 mm CL were pooled together into a single size class, which we refer to as the 77-mni-CL size class in reflection of their median CL. Females and males accounted respectively for 43.2'^fi and 44.1''r of all lobsters examined; the remainder were too small to deter- mine sex. Lobsters were pooled for analyses irrespective of sex because Weiss ( 1970) and Ennis ( 1973) concluded that diet was the same for both sexes. Only two lobsters had empty stomachs and they be- longed to the 10-mm size class. With these two empty stomachs excluded, there was a highly significant relation- ship between the dry mass of stomach contents and lob- ster CL (Fig. 1). Identifiable food items included macroal- gae or benthos that were grouped into broad taxonomic or ecological categories (Table 2). No planktonic organisms were identified from the stomachs, even of the smallest lobsters. However, the crustacean meiofauna group includ- ed the remains of very small crustaceans, some like the harpacticoids and ostracods, known to be bottom-dwell- ing, whereas unidentified minute crustacean lemains may have originated from holo- or mero-planktonic forms or from juvenile amphipods, isopods. or carideans. Sand, silt, and infrequently bits of nylon rope were also found in the stomachs. "Flesh" refers to tissue bolus composed of an- imal soft parts that could not be attributed to a taxon, generally because no distinctive part was found in the stomach along with the tissue or less commonly because distinctive parts from several prey types were present in the stomach but none was attached to the tissue. Table 1 Nunibci nf lobster stom achs sample d by classes r f cepha- lothorax length (CL, in iim). Size cl asses represent 5-nim groupin ?s except the smallest i7 mm CL) and lai gest (77 mm CL , which include all 1 obsters <7.5 mm CL and all lobsters >67.5 mm CL, respectively. Numbei of stom achs Cfphalo thorax leii Kth isizc classes) 1996 1997 Total 7 1 15 16 10 17 20 37 1.5 28 28 20 38 38 ■'F, 56 56 30 45 45 35 52 52 40 51 51 45 45 45 50 45 45 55 31 31 60 25 25 65 16 16 77 21 21 Total 471 35 506 Ontogenetic shifts in diet A cluster analysis on the volumetric contribution of food items to lobsters by size class yielded four groups: 7 mm, 10-20 mm, 25-60 mm, and 65-77 mm CL lobsters (Fig. 2). These same groups could be seen on a plot of the fii'st three factors of a factor analysis of the correla- tion matrix of the volumetric contribution of food items (Fig. 3). The three factors explained 68. 87^ of the vari- ance (39.9%, 18.2%, and 10.7% for factors 1, 2, and 3). The first factor had strong loadings for crustacean meiofauna (0.96), foraminiferans (0.96), bivalves (0.84), macroalgae (0.82), amphipods (0.78). and rock crab (-0.71). Because lobsters in the 7-mm-CL size class had little rock crab in their stomachs, but relatively high proportions of the other food items, they stood out with a very large score (3.1) on this factor. The next two size classes, 10 and 15 mm CL, scored 0.8 and 0.6. respectively. All other size classes scored between and -0.6 on the first factor. The second factor had strong loadings for flesh (0.73), lobster (-0.82), and barnacles (-0.73). Lobsters of the two largest size classes (65 and 77 mm CL) had strong negative scores on this factor (-2.5 and -1.5, respectively), whereas lob- sters of the 10-35 mm size classes scored between 0.5 and 1.1. The smallest size class (7 mm CL) and size classes of 40-60 mm CL had scores close to 0. Finally, the third factor had a high loading for carideans (0.74) and some- what smaller loadings for isopods (0.67), coralline algae (-0.57), and pagurids (-0.54). This third factor separated Sainte Mane and Chabot: Natural diet of Homarus amencanus off the Magdalen Islands 109 10' O °°° / O) 10" ° ,rS&?r c ^.^^^W% c tomach co o if) o 10'^ iy^^ "° if) i ' V** °° Q 10' /tf "o o o IC o o 4 10 100 200 Carapace length (mm) Figure 1 Relationship of dry mass of stomacli contents (DMi to cephalothorax length (CL) of lobsters from the Magdalen Islands. Two lobsters of the 10-nim class had an empty stomach and are not shown. Model II regression: DM - 7.9fi7 . 10 «xCZ.-^''-'^ |;-'=0,51.P<0.001|. the 25- and 35-mni-CL size classe,s from the 10-20 mm CL size classes. For each grouping, Figure 4 shows the specific abun- dance of each food item plotted against its percent occur- rence. Bivalves and flesh accounted for a large proportion of stomach contents of the smallest lobsters (7-mm-CL size class) and were found in >751 of stomachs, making them the most important food items for this grouping. Rock crabs, amphipods, and polychaetes contributed 0.2 to 0.4 of stomach volume when they were ingested, but were found in fewer than 30'* of the stomachs. Macroalgae and gastropods, on the other hand, were eaten by >50'; of small lobsters but were ingested in small volumes. All other prey categories contributed little to stomach volume and were found in a small proportion of stomachs. Flesh and bivalves were also the most important food items for the 10-20 mm CL lobster grouping (Fig. 4). They accounted for 0.46 and 0.22 of stomach volume, respec- tively, when they were ingested, and were found in 90' ^ of stomachs. Rock crab was another important prey, with a specific abundance of 0.32 and an occurrence of 41'r. Pagurids, carideans, and echinoderms had high specific abundances but were found in less than 5'^r of stomachs. Gastropods and polychaetes were found in about 40'5c of stomachs, but accounted for a small fraction of stomach volume. All other prey categories constituted a small frac- tion of the volume of very few stomachs. The two main food items of lobsters measuring 25-60 mm CL were rock crab and flesh: specific abundance was high (0.34 and 0.38, respectively) and these food items Table 2 Major categories of liiod ilc nis, divic ed into specific food items when possible, and t heir overall volumetric contribu- tion (total=l ) to stomach coi tents of; ill examined lobsters from Baie de Plaisance, Ma gdalen Islands. Abbreviations | for major categories of food i tems are shown in brackets. Volumetric Categories of food items contribution Formaniferans |For| 0.0031 Macroalgae (Algl 0.0394 Coralline algae iCiirnllinn o fi(in(dis 1 ICorl 0,0178 Hydrozoans |Hyd| 0.0207 Bivalves [Biv| 0.1657 Mytihis ediilis 0,0202 Modiolus modiolus 0.0992 Unidentified Pelecypoda 0.0463 Gastropods |Gasl 0.0.585 Lacuna vincta 0.0028 Unidentified tiastropoda 0.0057 Polychaetes |Pol| 0.0597 Ncreidae 0.0318 Polynoidae 0.0271 Unidentified Polychaeta 0.0008 Barnacles iBalanuN sp.l [Bar] 0.0012 Crustacean meiofauna ICru 0.0053 Harpacticoida 0.0003 Ostracoda 0.0021 Unidentified minute Crus t acea 0.0029 Amphipods lAmpl 0.0054 Coropltium sp. 0.0004 Gamniarus sp. 0.0003 Caprellidea 0.0004 Gammaridae 0.0016 Unidentified amphipods 0.0027 Isopods llsoj 0.0067 Idotea sp. 0.0013 Idoteidae 0.0019 Unidentified valvif'eran isopods 0.0034 Carideans [Carl 0.0024 Crangon septemspinoaa 0.0010 Unidentified carideans 0.0013 Pagurids [Pag] 0.0416 Pagurus acadianus 0.0051 Paguridae 0.0365 Rock crab 'Cancer irroratufi [Cral 0.2637 American lobster iHuniarus anicricanuK) [Lobl 0.0076 Echinoderms [Ech] 0.0222 Strongylocentrotus drueha chiensis 0.0102 Ophiuroidea 0.0012 Unidentified echinoderms 0.0109 Fish [Fisl 0.0066 Flesh [Flel 0.2724 no Fishen/ Bulletin 100(1) Figure 2 Dendrogj-am resulting from a cluster analysis on the mean volumetric contribution of major food categories by size class of lobsters from the Magdalen Islands. The bottom graph shows the joining distance at each step. The vertical dashed line indicates the cut-off value for clusters, selected because of the sudden increase in joining distance. were found in more than TCS of stomachs (Fig. 4). Bi- valves were still found in a large proportion of stomachs {8T7c) but accounted for a low proportion (0.18) of volume. Gastropods, polychaetes, and macroalgae also occurred frequently but accounted for only a small fraction of stom- ach volume. Pagurids and lobsters were found in few stom- achs but contributed >0.'2 of stomach volume. The grouping of the largest lobsters, 65-77 mivi CL, had rock crab as the most important food item (specific abun- dance=0.55; occurrence=86' 7 ). Lobsters, pagurids and fish contributed a large proportion of stomach volume when they were eaten, but these prey were ingested by <20'^( of lobsters. Gastropods, flesh, bivalves, polychaetes, and macroalgae were found in a large proportion of stomachs but occupied a small proportion of the volume of these stomachs. Overall, bivalves, rock crab, and flesh were the only food items that each accounted for >0.1 of stomach volume for the whole sample (Table 2). For these food items, a signifi- cant linear relationship existed between volumetric con- tribution and lobster CL, the latter explaining 68*7? to 929c of the variability in volume (Fig. 5). Regi-ession of volumet- ric contribution on lobster CL produced a negative slope for bivalves and flesh, and a positive slope for rock crab. Similarly, strong linear or nonlinear relationships existed between percent occurrence of these three food items and lobster CL (Fig. 5). Furthermore, large lobsters tended to eat larger rock crabs than small lobsters, as evidenced by the significant positive linear relationship between the CW of rock crabs found in lobster stomachs and lobster CL (Fig. 6). Figure 3 Results of the factor analysis on the correlation matrix ol volumetric contribution of major food categories by size class of lobsters from the Magdalen Islands. See text for factor loadings. Three clusters identified in Figure 2 are shown inside ellipses; the other size classes constitute the fourth cluster. Discussion Data Stomach content analysis is a useful method for the inves- tigation of the natural diet of animals, even though the lack of distinctive hard parts in some prey and differential digestibility of soft and hard body parts limits the spec- trum of food items that can be recognized and can lead to biased perception of the relative importance of the food items. We took care to process lobsters as quickly as pos- sible after collection, thus attenuating the effects of differ- ential digestibility, and we examined only intermolt and nonovigerous lobsters, thus reducing sources of diet vari- ability associated with molt cycle and female reproduc- tive status (e.g. Weiss, 1970: Ennis, 1973). In addition, our study was conducted over a small area where the various lobster size classes were evenly distributed; therefore all lobsters potentially could access the same food. We rec- ognize that our volumetric contribution index underesti- mates the importance of predominant food items, owing to correction for stomachs with multiple food items and total scores >10. However, this was a minor problem because analyses using uncorrected values revealed that the vol- umetric contribution of the three main food items was underestimated by no more than 2-5'^< and that relation- ships to lobster size class were unchanged. Therefore, we are confident that the dietary differences among the lob- ster size classes that we detected are real and that they Sainte Marie and Chabot NatLiial diet of Haniaiui ameiicaniis off tlie Magdalen Islands 111 1 A 7 mm 1 H Id 20 mm 08 08 ■.,>- 06 08 i» ** / '.-' ..-^- 04 - jC- «> 04 <.-» ^°'~ J^ .-^^ .f- • .o'- <^ 02 f ^<^ 02 .o^S? .s • bundance o o Bal Ech ISO ^ .c^ ,CarF,s LobV O* .'^ ||CorHydPag • 9 . . . . 00 /•' .*" 20 40 60 80 100 20 40 60 80 100 3Cific a o C 25-(i() mm 1 D 65 -77 mm Q. (/2 08 08 J' 06 06 r"" .o"^ 04 04 . [Amp • Bal J Car 02 «^ 2 \Cru .<^» .^<^:> .o^^ 00 ^1^ t 1 1 1 00 1 ■t 20 40 60 80 100 20 40 60 80 100 Occurrence (%) Figure 4 Relationship between specific abundance and percent occurrence for the major food catego- | ries in relation to clusters for size classes ofthe(Al 7 mm.(Bl 10- -20 mm. (C) 25-60 mm, and (D) 65-77 mm CL (see Fig. 2) for lobste ■s from the Magdalen 1 slands. Refer to Table 2 for abbreviations of major food categories. reflect mainly changing lobster preferences and differen- tial accessibility of prey types. Ontogenetic shifts in diet There was clear evidence of a progressive dietary shift with increasing lobster size at our study site. Smaller lobsters relied to a greater extent than larger lobsters on soft or easily acquired food items (flesh, sessile juvenile bivalves, macroalgae, meiobenthic crustaceans, and foraminiferans). Larger lobsters fed on bigger, more mobile and also more nutritious prey, including crustaceans that were protected by heavy shells, and fish. Fishes were probably taken by predation (see Weiss, 1970) because there was no fishing activity at or near our study site that might have provided lobsters with fish bait or discards. The most striking ontogenetic changes in volumetric contribution of prey types occurred for rock crab and bi- valves, the former increasing from 0.07 to 0.53 and the lat- ter decreasing from 0.28 to 0.02 from the smallest to the largest lobster size class, respectively (Fig. 5). Only lim- ited comparisons with other studies are possible, given the differences in methods and in the size range of lobsters examined. However, the observed trends of increasing im- portance of rock crab and of decreasing importance of bi- valves with increasing lobster size were consistent with the analyses of Scarratt (1980) and of Carter and Steele ( 1982b), and they suggest that lobsters are not simply op- portunistic or unspecialized feeders (see Elner and Camp- bell, 1987). Multivariate analysis of lobster diet resulted in size groupings that are quite consistent with Lawton and La- valli's (1995) size classification of the early life-history phases based on a broad set of behavioral and ecological criteria. Major shifts in diet in the present study occurred at about 7.5, 22.5, and 62.5 mm CL (Fig. 2). The two clas- sifications differ in the smaller size for the transition from the first to second group (7.5 mm in our diet-based classifi- cation compared with 14.5 mm CL in Lawton and Lavalli's scheme), but the size for transition from the second to the 112 Fishery Bulletin 100(1) 1 A bi\al\Os Q 100 08 80 06 \ ■^ 60 04 • l/C = 304-0 004 CL r- = 72 40 02 "^^^^^^^*^^^^^_ 20 00 10 *~~ —-* . 100 ' B Hcsh □ ^ ^^ ^ ^ PO = 90 506 - 467CL c 08 ~"~--_n r^ = 59 - 80 o ~~" -— S" ~ 06 O^ ^ - ^ ^ ^ n CD 60 3 o o o o o • • o c S 04 — ^_____^ 40 g E " ~~— i^ • 13 13 O o • *^-— -__• (D > 02 • """^ — •- — •-_ \/C = 441 -0 004CL ^"— » ^ r- = 68 20 00 1,0 1 .... 1 ..,, 1 .... 1 .... 1 .... 1 .... ~ 100 C iiick crab ^ ^ n___,_n-- ,, 08 80 0,6 60 04 40 0.2 " / _»,-— r"'*'^^ ;/c = 02i +0 007C/. .X-"*"^ /-' = 92 20 00 • c 10 20 30 40 50 60 70 8 Cephalothorax length (mm) Figure 5 Relation between percent occurrence IPO, Di or volumetric contribution (VC.») and lobster cephalothorax length for the three main food items of lobsters from Magdalen Islands: (Al bivalves, (Bi flesh bolus, and (C) rock cral). All linear regi-essions are highly significant (P<0.001 1. third group is the same in both studies (22.5 and -25.0 mm CD. Comparison of the size threshold for transition from the third to the fourth group is less appropriate be- cause Lawton and LavaUi (1995) considered this thresh- old to be determined by physiological maturity, which is a temperature-dependent trait that varies among regions. Natural diet of shelter- restricted juveniles This first investigation of the diet of SRJ lobsters does not support the view that these juveniles derive a substan- tial portion of their diet by suspension feeding and brows- ing in their shelters, at least at our study site and during the two years we sampled. With respect to suspension feeding, there was no evidence of planktonic organisms in stomachs, although some of the unidentified prey of the crustacean meiofauna category may have been planktonic. Foraminiferans, harpacticoids, ostracods, and macroalgal debris represented food items that potentially could be browsed within shelters. However, these taxa together contributed relatively little to stomach volume of lobsters in the 7-mm size class (0.14 for the combined categories. Sainte Mane and Chabot Natural diet of Homanis amencamis o\\ the Magdalen Islands 113 60 r / 50 / D n / D / E E a / — 40 / ^ S On/ ° S X 2 30 □ /o° ° o o o. 20 0) o °d/ d n " 10 a /na4D ct, cP 7?^ D □/ Oan o d/ Q 20 40 60 80 100 120 Lobster cephalothorax length (mm) Figure 6 Relationship of predator (lobster) size to prey (rock crab) size, based on rock crab cephalothorax width (CWl esti- mated from measurements of indicator fragments and lob- ster cephalothorax length (CL), for the Magdalen Islands. Model II regression: CW = -12.341 -i- 0.677CL |/-=0.34, /'<(). 001 1. in spite of the fact that one or the other category occurred in S8'/( of stomachs) and even less to stomach volume of lobsters in the 10-mm size class (0.10, 86'^). During our study, lobsters settled in August at sizes of 4.3-5.2 mm CL and grew to 12-14.5 mm CL by October (Sainte-Marie et al., 2001). Thus, we sampled the lobster population during the only period of time when SRJs were present and sea- sonal sampling bias cannot be invoked to explain the lack of plankton in their diet. The other food items in the stomachs of SRJs, and es- pecially the predominant bivalves and flesh (Figs. 4 and 5 ), probably were derived by predation and scavenging. Bi- valves in the stomachs of SR.J lobsters were represented by recently settled Modiolus and Myfilus. Mussel spat may settle aggregatively and quite synchronously, forming dense patches that can provide a short-term prey pool requiring little or no search time (e.g. Auster, 1988). Furthermore, be- cause mussel spat often settle in crevices or under rocks (e.g. Nair et al., 1975), SRJs could access them with little or no risk of exposure to predators. Lawton ( 1987 ) argued that dominance and territoriality were likely to exist early in the ontogeny of lobsters, as demonstrated subsequently ( James- Pirri and Cobb. 1999; Paille and Sainte-Marie, 2001). and that prolonged occupation and defense of shelters located close to a food patch would be advantageous for juveniles. Exploitation of mussel patches, inferred from the present study, is consistent with that hypothesis. Flesh (tissue boluses) that could not be attributed to a particular animal for lack of indicator fragments was a very important food item in the diet of SRIs, both in terms of percent occurrence and of volumetric contribution (Figs. 4 and 5). Elner and Campbell ( 1987) also found that uniden- tified animal tissue was one of the most frequent and most volumetrically important foods in the stomachs, however, of larger lobsters. Weiss (1970) observed that adolescent and adult lobsters often captured crabs or other shelled prey, cracked them open, and then selectively ingested only soft tissue. Interestingly, the percent occurrence and volu- metric contribution of flesh to diet was greater in lobsters of size classes <30 mm CL (i.e. SRJs and emergent juve- niles) than in larger lobsters (Fig. 2). It is unlikely that the smallest of lobsters could find (within the confines of their shelter) and subdue prey sufficiently large to provide tis- sue boluses devoid of hard parts. Furthermore, claws are not differentiated into cutter and crusher forms in SRJs (Govind and Lang, 1978; Costello and Govind, 1984) and early juveniles may be incapable of breaking open shelled prey (Costello and Lang, 1979; Lawton and Lavalli, 1995). Therefore, flesh ingested by SRJs and emergent juveniles probably was obtained by scavenging animal remains. Con- sidering that larger lobsters may hoard and bury food in or nearby their dens (Herrick, 1895; Smith, 1976; Lawton, 1987; Wickins et al., 1996), we propose that early juveniles exploit the meal scraps or food resei-ves of larger lobsters. Indeed, we obsei-ved that small lobsters often occupied gal- eries beneath, or in rock pilings nearby, the dens of larger lobsters. This is consistent with reports that odor from con- specific adults is a proximate cue for lobster settlement (Boudreau et al.. 1993). Cohabitation of small lobsters with large lobsters would offer the former protection from pred- ators and a potentially abundant, high-quality, sheltered food source, and would therefore represent a form of com- mensalism. The risk of cannibalism for small lobsters liv- ing in the vicinity of larger lobsters probably does not off- set the benefits. Few lobster remains were found in lobster stomachs in this (Fig. 4) as in other studies (Weiss, 1970; Carter and Steele, 1982b; Elner and Campbell, 1987), and an unknown proportion of those remains may have been exuviae. Some other rarer food items found in the stomachs of SRJ lobsters were probably taken by predation, possibly within, but more likely in the neighborhood of, the lob- sters" shelters. The most important of these prey by volu- metric contribution were polychaetes, comprising juvenile nereids and polynoids that are frequently found in soft sediment or on the underside of rocks, and recently settled rock crab. Similarly, amphipods and gastropods found in the stomachs of SRJs were juveniles or small species that may abound in crevices and in spaces beneath rocks. A carnivorous, high-energy diet such as the one demon- strated for SRJs in our study would promote growth from settlement time. By contrast, Lavalli ( 1991 ) demonstrated that a diet of only diatomous algae was insufficient for ex- tended growth and sui-vival of early juvenile lobster A diet of mesozooplankton sustained growth of juvenile lobsters, at least for some time after settlement (e.g. Daniel et al., 1985; Barshaw, 1989; Lavalli, 1991). However, Lawton and Lavalli ( 1995) pointed out that intermolt periods tended to be longer and molt increments smaller in laboratory-held. 114 Fishery Bulletin 100(1) juvenile lobsters reared on mesozooplankton than in wild lobsters, suggesting that the latter incorporated more nu- tritious foods into their diet. The finding that early juvenile lobsters are primarily predators or scavengers, if confirmed by studies at other sites, has implications for the development and implemen- tation of artificial reefs. Such structures are increasingly being considered as a means to enhance lobster produc- tivity on traditional grounds or to expand lobster habitat onto less hospitable grounds (e.g. Gendron, 1998). The car- nivorous benthic feeding mode of SRJs and of emergent ju- veniles at our site implies that successful reefs will have to be designed, localized, and weathered so that they are ini- tially well colonized and subsequently regularly colonized by benthic prey that are easily accessible and of high nu- tritional value to juvenile lobsters. Additionally if SRJs and emergent juveniles derive some protective and nutri- tional benefits from the presence of larger conspecifics, reefs designed to offer shelter to a full suite of lobster sizes may prove to be more productive in the long term than reefs offering shelter only to small lobsters. Importance of rock crab to lobster Several previous studies have noted the importance of rock crab in the diet of lobster (Reddin, 1973; Evans and Mann, 1977; Carter and Steele, 1982al. Boghen et al. (1982) found that juvenile lobsters survived and grew better on a diet containing crab protein alone than on a diet of live brine shrimp iAiienua salina) or of protein extracts from urchin iStrongylocentrotus droebachiensi.';). mussel (Mytilus ediilis). or shrimp iPenaeus sp.). Gendron et al. (2001) found that condition, somatic gi-owth, and gonadal development of lobster increased with increasing amount of rock crab in diet. In nature, even SRJs may ben- efit from a diet including large amounts of rock crab pro- tein because they preyed directly on very small rock crabs (Figs. 4 and 5), and the tissue boluses they contained may have been that of rock crab (see above). We were able to establish a positive size relationship for lobster preying on rock crab (Fig. 6). The smallest rock crab prey were 2-6 mm CW and belonged to the first ben- thic instars of this species. In our study, apparently no rock crabs larger than 50 mm CW were consumed by lob- sters, and the maximum ratio of crab CW over lobster CL was 0.90, even though rock crabs up to 120 mm CW were seen (own personal diving obsei-vations). In the labo- ratory Weiss (1970) observed that lobsters of 60-80 mm CL attacked crabs offered in the size range of 62-78 mm CW. Lawton and Lavalli (1995) reported that juvenile lob- sters can subdue juvenile intermolt rock crabs up to ap- proximately 0.40 times their own body size. Their obser- vation was based on the comparison of predator and prey wet masses; when expressed in terms of crab CW over lob- ster CL, the maximum ratio was equivalent to about 1.27.' This ratio of prey CW to predator CL is somewhat larger than that derived from our stomach analvses. Because lob- Lawton, P. 2000. Personal comniun. Fisheries and Oceans Canada, St. Andrews, New Brunswick, Canada. sters probably ingest only soft tissue when the prey-pred- ator size ratio is sufficiently high (Weiss, 1970; and see above ), our analysis of rock crab prey-size frequencies may correctly estimate the minimum prey size but underesti- mate the maximum prey size and the volumetric contribu- tion and occurrence of rock crab in the diet of any given lobster size class. Nevertheless, the present study clearly shows that all lobster size classes rely on rock crab as food and that the size spectrum of rock crab that is used by lob- sters is broad and includes even those at the settlement stage. Given the much greater economic value of lobster in relation to rock crab, and the trophic dependency of the former on the latter, caution should be exercised in devel- oping rock crab fisheries (Gendron and Fradette, 1995). Acknowledgments We thank our diving partners F. Hazel. J.-G. Rondeau, J. A. Gagne, K. Gravel, R. Larocque, J.-F. Lussier, N. Faille, and A. Rondeau. We are particularly gi-ateful to J. Hudon for her major contribution to the identification of stomach contents and to three anonymous reviewers for construc- tive comments. This is a contribution to the Canadian Atlantic-Wide Lobster Studies (CLAWS) research initia- tive of Fisheries and Oceans Canada. Literature cited Aiken, D. E. 1980. Molting and gi-owth. In The biology and manage- ment of lobsters iJ. S. Cobb, and B. F. Phillips, eds. ), p. 91-163. Academic Press, New York. NY. Amundsen, P. -A., H.-M. Gabler, and F. J. Staldvik. 1996. A new approach to graphical analysis of feeding strategy from stomach contents data — modification of the Costello (19901 method. J. Fish Biol. 48:607-614. Auster, P. J. 1988. Response of megafaunal predators to synchronous settlement of sessile prey. J. Shellfish Res. 7:108-109. Barshaw, D. E. 1989. Growth and survival of early juvenile American lob- sters. Homarua ODiencanuK. on a diet of plankton. Fish. Bull. 87:366-370. Barshaw, D. E., and D. R. Bryant-Rich. 1988. A long-term study on the behavior and survival of early juvenile American lobster, Homarus americantis, in three naturalistic substrates: eelgrass, mud, and rock. Fish. Bull. 86:789-796. Boghen, A. D., J. D. Castell, and D. E. Conklin. 1982. In search of a reference protein to replace "vitamin- free casein" in lobster nutrition studies. Can, J. Zool. 2:2033-2038. Boudreau, B., E. Bourget, and Y. Simard. 1993. Behavioural responses of competent lobster postlar- vae to odor plumes. Mar. Biol. 117:63-69. Carter, J. A., and D, H. Steele. 1982a. Attraction to and selection of prey by immature lob- sters iHomai'us omencaniis). Can. J. Zool. 60:326-336. 1982b. Stomach contents of immature lobsters (Homarus americanus'i from Placentia Bay, Newfoundland. Can. J. Zool. 60:337-347. Sainte Marie and Chabot Natural diet of Homaiw, ameiicanus off the Magdalen Islands 115 C.hh. .1 s 1971. 'Plu' slu'ltiT-ii'lati'tl bchavHir (il'thc l(il)sl('i', Hnniinus anicnviiniis. F>ology 52:108-115. Cnl)l).-I. S..and K. A. Wahliv 1994. Karly life history and rccniilnirnt |>roc(.'.ssi.'s olclauud lobsters. Crustaceana 67:1-25. Cooper, R. A., and J. R. Uzmann. 1977. Ecolofiy of'juvenile and adult clawed lobsters, Homa- rus aniericaniis. Hiiniarus ganimaruti. and ^Vcp/i/op.s nor- wgiciis. Div. Fish. OceanogT. Circ. ( Aust. C.S.I.R.O. I 7: 187- 208. Costello, W. J., and t'. K. (lovind. 1984. Contractile proteins of fast and slow fibers during dif- ferentiation of lobster claw muscle. Dev. Biol. 104:434- 440. Costello, W. J., and F. Lang. 1979. Development of the dimorphic claw closer muscles of the lobster Hoinarus americanus. IV. Changes in functional morphology during growth. Biol. Bull. iWoods Hole, MAi 156:179-195. Daniel, P. C. R. C. Bayer, and S. Chapman. 1985. Barnacle larvae (Batanus spp. ) as a potential diet for juvenile lobsters {Homarus americantis). Aquaculture 46:67-70. Elner. R. \V., and A. Campbell. 1987. Natural diets of lobster Homarus americamia from barren gi-ound and macroalgal habitats off southwestern Nova Scotia, Canada. Mar Ecol. Prog. Ser 37:131-140. 1991. Spatial and temporal patterns in recruitment for American lobster, Homarus americanus, in the northwest- ern Atlantic. Mem. Queensl. Mus. 31:349-363. Ennis, G. P 1973. Food, feeding and condition of lobsters, Homarus amer- icanus. throughout the seasonal cycle in Bonavista Bay, Newfoundland. J. Fish. Res. Board Can. 30:190.5-1909. Evans, P D., and K. H. Mann. 1977. Selection of prey by American lobsters [Homarus amencanus) when offered a choice between sea urchins and crabs. J. Fish. Res. Board Can. 34:2203-2207. Freire, J., M. Paz Sampedro, and E. Gonzalez-Gurriaran. 1996. Influence of morphometry and biomechanics on diet selection in three portunid crabs. Mar Ecol. Prog. Ser 137:111-121. Gendron, L. 1998. Proceedings of a workshop on lobster stock enhance- ment held in the Magdalen Islands (Quebec) from October 29 to 31, 1997. Can. Ind. Rep. Fish. Aquat. Sci. 244. Gendron, L., and P. Fradette. 199.5. Revue des interactions entre lecrabecommun [Cancer irroratus) et le homard americain (Homarus americanus), dans le contexte du developpenient d'une peche au crabe commun au Quebec. Can. MS Rep. Fish. Aquat. Sci. 2306. Gendron, L., Fradette, P., and G. Godbout. 2001. The importance of rock crab iCancer irroratus) for growth, ovary development and condition of adult Amer- ican lobster (Homarus americanus). J. Exp. Mar. Biol. Ecol. 262:221-241. (iovind, C. K., and F. Lang. 1978. Development of the dimorphic claw closer muscles of the lobster Homarus americanus. III. Transformation to dimorphic mu.scles in juveniles. Biol. Bull. (Woods Hole, MAl 1.54:5,5-67. Herrick, F. H. 1895. The American lobster: a study of its habits and devel- opment. Bull. U.S. Fish. Comm. 15:1-252. lliidon, C. 1987. Ecology and growth of postlarval and juvenile lobster, Homarus americanus, off lies de la Madeleine (Quebec). Can. J. Fish. Aquat. Sci. 44:185.5-1869. Hudon, C, and G. Lamarche. 1987. Niche segi-egation between American lobster Homa- rus americanus and rock crab Cancer irroratus. Mar. Ecol. Prog. Ser 52:155-168. James-Pirri, M. J., and S. J. Cobb. 1999. Behavioral interactions of postlarval and fifth instar lobsters {Homarus americanus) in a simulated cobble envi- ronment. Mar Freshwater Behav. Physiol. 32: 207-222. LavalH, K. L. 1991. Sui-vival and growth of early-juvenile American lob- sters Homarus americanus through their first season while fed diets of mesoplankt.on, microplankton, and frozen brine shrimp. Fish. Bull. 89: 61-68. Lavalli, K. L., and D. E. Barshaw. 1989. Post-larval American lobsters (Homarus americanus) living in burrows may be suspension feeding. Mar. Behav. Physiol. 15:255-264. Lavalli, K. L., and J. R. Factor. 1995. The feeding appendages. In Biology of the lobster Homarus americanus (J. R. Factor, ed.), p. 349-393. Aca- demic Press, San Diego, CA. Laws, E. A., and .J. W. Archie. 1981. Appropriate use of regression analysis in marine biol- ogy. Mar. Biol. 65:13-16. Lawton, P. 1987. Diel activity and foraging behavior of juvenile Ameri- can lobsters. Homarus americanus. Can. J. Fish. Acjuat. Sci. 44:119.5-1205. Lawton, P.. and K. L. Lavalli. 1995. Postlai-val, juvenile, adolescent, and adult ecology. In Biology of the lobster Homarus americanus (J. R. Factor, ed. ), p. 47-88. Acad. Press, San Diego. CA. Lee, S. Y., and R. Seed. 1992. Ecological implications of cheliped size in crabs: some data from Carcinus maenas and Liocarcinus holsatus. Mar. Ecol. Prog. Ser. 84:151-160. Lovrich, G. A., and B. Sainte-Marie. 1997. Cannibalism in the snow crab, Chionoecetes opilio (Brachyura: Majidac), and its potential importance to recruitment. J. Exp. Mar Biol. Ecol. 211:225-245. Nair, R. V., K. N. Nayar, and S. Mahadevan. 1975. On the large-scale colonisation of the spat mussel, Mv^ ilus viridis, in Cochin region. Indian J. Fish. 22:236-242. Ojeda, F. P.. and J. H. Dearborn. 1991. Feeding ecology of benthic mobile predators: experi- mental analyses of their influence in rocky subtidal com- munities of the Gulf of Maine. J. Exp. Mar. Biol. Ecol. 149:1.3-44. Paille, N., and B. Sainte-Marie. 2001. Effects of crowding and shelter limitation on the behaviour and survival of the first benthic stage of lobster. In Proceedings of the Canadian lob.ster Atlantic wide stud- ies (CLAWS) symposium, Moncton, March 2000 (M. J. Tremblay and B. Sainte-Marie, eds. ), p. 30-33. Can. Tech. Rep. Fish. Aquat. Sci. 2328. Reddin, D. 1973. The ecology and feeding habits of the American lobster (Homarus americanus) (Milne-Edwards, 1837) in Newfoundland. M.Sc. thesis. Memorial LIniv., St. John's, Newfoundland, 101 p. Sainte-Marie, B., D. Chabot, F. Hazel, and L. Gendron. 2001. Preliminary analysis of settlement intensity and 116 Fishery Bulletin 100(1) growth oljuvenile lobster in the shallows of Baie de Plai- sance, Maplalon Islands. In Proceedings of the Canadian lobster Atlantic wide studies i CLAWS) symposium, Monc- ton, March 2000 (M.J. Tremblay and B. Sainte-Marie, eds. I. p. 27-29. Can. Tech. Rep. Fish. Aquat. Sci. 2328. SAS Institute. 1995. JMP statistics and graphics guide, version 3. SAS Institute Inc., Cary NC. 593 p. Scarratt,D. J. 1980. The food of the lobster. //( Proceedings of the work- shop on the relationship between sea urchin gi-azing and commercial plant/animal hai-vesting (J. D. Pringle, G. J. Sharp, and J. F. Caddy eds.), p. 66-91. Can. Tech. Rep. Fish. Aquat. Sci. 954. Smith, E. M. 1976. Food burial behavior of the American lobster, Hn/ua- rus americanus. M.Sc. thesis, Univ. Connecticut, Storrs, CT, 52 p. SokalR. R.and F J. Rohlf 1995. Biometry; the principles and practice of statistics in biolog- ical research, 3"' ed. W.H. Freeman, New York, NY, 887 p. Squires, H .J. 1970. Lobster iHomarus cinwricaniis) fishery and ecology in Port au Port Bay Newfoundland, 1960-65. Proc. Natl. Shellfish Assoc. 60:22-39. Wahle, R.A. 1992. Body-size dependent anti-predator mechanisms of the American lobster. Oikos 65:52-60. Weiss, H. M. 1970. The diet and feeding behavior of the lobster, Honiarus americanus, in Long Island Sound. Ph.D. diss., Univ. Con- necticut, Storrs, CT, 80 p. Wickins, J. F, J. C. Roberts, and M. S. Heasman. 1996. Within-burrow behaviour of juvenile european lob- sters Homariis gammaruft (L.). Mar Freshwater Behav. Phvsiol. 28:229-253. 117 Abstract— Skeletochronological data on fjniu ih changes in humerus diam- eter were used to estimate the age nf Hawaiian gi'een seaturtlos ranging from 28.7 to 96.0 cm straight carapace length. Two age estimation methods, correction factor and spUne integration, were compared, giving age estimates ranging from 4.1 to 34.6 and from 3.3 to 49.4 yr, respectively, for the sample data. Mean growth rates of Hawaiian green seaturtles are 4-5 cm/yi- in early juveniles, decline to a relatively con- stant rate of about 2 cm/yr by age 10 yr. then decline again to less than 1 cm/yr as turtles near age 30 yr. On average, age estimates from the two techniques differed by just a few years for juvenile turtles, but by wider mar- gins for mature turtles. The spline-inte- gration method models the curvilinear relationship between humerus diame- ter and the width of periosteal gi'owth increments within the humerus, and offers several advantages over the cor- rection-factor approach. Age and growth of Hawaiian green seaturtles (Chelonia mydas): an analysis based on skeletochronology George R. Zug Division of Amphibians and Reptiles Department of Systematic Biology National Museum of Natural History Washington, DC. 20560-0162 E-mail address: zuggeorgeiSnmnhsiedu George H. Balazs Jerry A. Wetherall Honolulu Laboratory Southwest Fisheries Science Center National Manne Fishenes Service, NOAA 2570 Dole St, Honolulu Hawaii 96822-2396 DenJse M. Parker Shawn K. K. Murakawa Joint Institute for Manne and Atmosphenc Research 2570 Dole St Honolulu, Hawaii 96822-2396 Manuscript accepted 20 August 2001. Fish. Bull, 100:117-127 (20021. The Hawaiian population of the green seaturtle {Chclonia mydas) provided some of the first published gi-owth data (Balazs, 1979, 1980. 1982) for this spe- cies. These early data showed how slowly seaturtles grow and how long a female must survive simply to lay her first clutch of eggs. Twenty or more years to reach sexual maturity seemed biologi- cally unrealistic, yet slow growth and late maturity has been repeatedly con- firmed for some seaturtle species, e.g. Bahamian C. mydas (Bolten et al,, 19921, West Atlantic Caretta caretta (Parham and Zug. 1998). Slow growth and the resulting delayed maturity greatly affect the demography of a population (Grouse et al., 1987; Chaloupka and Musick, 19961. An understanding of growth and gi-owth-pattern variation in seaturtles is an important prerequisite to the devel- opment of population models that are required to guide seaturtle population recovery and conservation. Green seaturtles within the Hawai- ian Islands contribute to the larger In- do-Pacific C. mydas gene pool, yet the nesting females of the Hawaiian popu- lation comprise a distinct genetic unit and contain a unique intDNA haplotype (Bowen et al., 1992), Except during their posthatching pelagic phase, the gi'eat majority of Hawaiian green seaturtles reside in coastal waters, primarily around Hawaii. Kauai, Maui, Molokai, Oahu, and other islands in the south- eastern part of the Hawaiian chain. Most reproduction takes place at French Frigate Shoals in the Northwestern Ha- waiian Islands (Balazs, 1980; Wether- all et al,. 1999), The population of Ha- waiian Chelonia mydas has benefited from over two decades of intense con- servation management (Balazs, 1998). Despite important early research on the Hawaiian population of C. mydas (Balazs, 1982), the patterns of growth within and among the geographic habi- tat components of this population have remained incompletely documented ( Balazs et al„ 1994, 2000 ), The chief rea- son for this has been a lack of methods to age sea turtles. This deficiency has been overcome recently by the devel- opment of skeletochronological tech- niques that estimate age from the num- ber of growth increments formed on the humerus (Parham and Zug, 1998), 118 Fisher/ Bulletin 100(1) Our goal in this study was twofold. First, we used humerus growth-increment data to estimate ages of a sample of Hawaiian green seaturtles from various locations in the ar- chipelago and developed a growth model for the general Hawaiian population; geograph- ic variation in growth will be addressed in a subsequent paper. Second, we compared two different methods of deriving the age esti- mates, the so-called "correction-factor" meth- od described by Parham and Zug ( 1998) and a newer approach, the "spline-integi'ation" method, introduced in the present study. Materials and methods Our sample consisted of 104 individuals of C. mydas, collected from the islands of Hawaii, Kauai, Lanai, Maui, Oahu, and the Northwestern Hawaiian Islands; the Oahu sample predominated with 64 individuals. All individuals were measured to the near- est 0.1 cm straight carapace length (SCL). The smallest individual was a 5.3-cm-SCL hatchling. The smallest posthatchling was a pelagic juvenile (of assumed Hawaiian origin) recovered from the former squid driftnet fishery north of the island chain. All other posthatchling turtles were from coastal Hawaiian waters, found stranded dead and re- trieved by the National Marine Fisheries Service's Hawai- ian Islands Seaturtle Stranding Network. The salvaged turtles ranged from 28.7 to 96.0 cm SCL. The sample was divided into eight 10-cm size classes; representation was roughly equivalent for the middle six classes (Fig. 1). The 30-39 cm sample contained only turtles in the upper quar- tile of this size class. Each turtle was necropsied and its right humerus removed for skeletochronological examina- tion. The necropsy data included a complete set of carapace measurements, organ condition evaluations, and informa- tion on fibropapilloma tumor occurrence and severity; see Work and Balazs (1999a, 1999b) for details on the entire data set. In addition to the skeletochronological data, we used only the SCL measurements and tumor-evaluation observations in the present study. Where possible, we selected tumor-free individuals for the present analysis because our goal was to examine the overall growth pattern for normal Hawaiian Chelonia my- das. Because the prevalence of fibropapillomatosis is high in the wild Hawaiian population (Murakawa et al., 2000), we included in our sample individuals with fibropapillo- mas. but otherwise appearing normal, in order to ensure adequate representation in the larger size classes. Healthy animals were those showing no evidence of weight loss or other indicators of illness and no evidence of disruption or retardation of normal growth. Individuals with tumors represented 27'7r of the 50-59 cm, 507^ of 60-69 cm, 76% of 70-79 cm, 73% of 80-89 cm, and 17% of 90-99 cm SCL size classes (Fig. 1). 20-29 30-39 40-49 50-59 60-69 70-79 80-89 90-99 Straight carapace lengtti (cm) Figure 1 Size (straight carapace length) distribution of the Hawaiian Chelonia mydas skeletochronological sample. The members of each class are segre- gated into individuals without (shaded bar) and with (black bar) fibropapil- loma tumors. Tumors in our sample are present only in larger turtles. Our skeletochronological data derived from cross-sec- tions (0.6-0.8 mm thick) from the middle of the humeral shaft just distal to the deltopectoral crest and at the nar- rowest diameter of the diaphysis (Zug et al., 1986). On each specimen, we counted the number of visible growth layers and measured the widths (long-axis diameters) of the humerus at each successive growth cycle and the width of the resorption core. Bone sections were taken from mid-shaft, the narrowest location of the bone, be- cause the humerus retains the gi'eatest number of perios- teal growth layers there, and hence this location permits the most accurate estimation of the number of growth cy- cles (periosteal layers) and the relative rates of growth (=successive humerus diameters). We used two procedures for estimating the total num- ber of growth layers, and hence age, of each turtle. In the correction-factor (CF) method, as described in Parham and Zug ( 1998), the turtle's age is estimated as the number of growth layers observed in the outer region of the humerus section plus the predicted number of resorbed growth lay- ers represented in the remodeled core of the humerus. The latter, unobservable component is estimated as C (i? - R[^), where R is the radius of the absorption core, Rf^ is the ra- dius of a hatchling's humerus (before the beginning of in- crement formation), and C is the so-called correction fac- tor The correction factor is a constant "aging rate" (yr/mm) assumed to apply to the resorption core, and calculated as the reciprocal of the mean growth layer width in small turtles. The mean growth layer width was estimated from 129 periosteal growth layer widths observed in 34 turtles Zug et al Age and growth of Hawaiian Chelonia 119 witli S(^Ls <60 cm and resorption core diameters <19.0 mm. Selecting only small turtles with minimum core di- ameters reduces the frequency of the narrower periosteal layers found in the outer margin of the humerus in larger turtles; hence, it reduces the possibility of overestimating the number of layers in the resorption core. The resulting correction factor was used to estimate the number of re- sorbed periosteal layers. A second method, spline integration (SI), is introduced here. The SI method uses a scatterplot smoothing spline (Hardle, 1990; Hastie and Tibshirani, 1990) to model the relationship between the aging rate and humerus diam- eter. Once the aging function is estimated, a turtle's age is estimated by integrating the spline over the total diam- eter of the turtle's humerus section. The method of model- ing increment width patterns in hard parts and of estimat- ing age by the integration of the resulting aging function was first formalized by Ralston and Miyamoto ( 1983) for a Hawaiian snapper and first applied to seaturtles by Zug et al. ( 1995). In those applications, the aging rate was a para- metric function of size. In our analysis, we modeled the ag- ing rate nonparametrically by fitting a smoothing spline to pairs of obsei-\-ations of growth-layer width and humer- us diameter The SI approach uses the same source of data as the CF method but without selection. Lines of arrested growth (LAG) delimit each observable growth layer Incre- ment width is measured as the difference between the hu- merus diameters at the outer LAG and the inner LAG. As- suming an increment represents one year of growth, each increment width measurement provides a measure of the humerus growth rate (nini/yr) and its reciprocal, a mea- sure of the aging rate (yr/mm) at the obsein-'ed humerus di- ameter (the mean diameter of the pair of LAGs). The skel- etochronological sample yielded 269 such observations of aging rate and humerus diameter. The aging rates were grouped in 1-mm intei-vals of humerus diameter and aver- aged. A cubic smoothing spline was fitted to the mean ag- ing rates by usnig S-PLUS (MathSoft, Inc., 1999). The age (yr) of each turtle was estimated by integrating the aging spline from its origin to the observed outside diameter of the humerus section. To assess the effect of estimation method on age esti- mates, the data were divided into 10-cm SCL groups. With- in each gi'oup a Student's t statistic was used to test the hypothesis that the two methods give equal age estimates. Nonparametric growth models were estimated based on the CF- and Sl-derived age estimates and associated carapace lengths, by using the same S-PLUS procedure employed for the Sl-method aging spline. The validity of the growth models was judged qualitatively by comparing growth predicted by the models with obsei"ved giowth in a sample of 171 Hawaiian gi-een turtles tagged and recap- tured in waters around Molokai (Balazs et al, 1999). To assess uncertainty in the Sl-based growth curve, the 269 pairs of aging rate and mean humerus diameter data were resampled 100 times, and the SI procedure applied to each bootstrap replicate data set. The 100 aging curves derived in this manner generated a bootstrap distribution of estimated age for each turtle. Nonparametric growth curves were then fitted to each derived data set, produc- ing bootstrap distributions of predicted mean length at age. Empirical confidence intervals for the predicted mean length at age were approximated by using percentiles of the latter bootstrap distributions. A linear regression of S('L on outside humerus diam- eter was estimated for the 104 sample turtles. The slope of the linear predictor was applied to the 269 humeral in- crements to estimate a corresponding set of carapace in- crements, presumed to represent annual growth. These growth rate estimates were summarized in box plots over 10-cm intervals of SCL. Mean growth rate as a function of estimated age was also estimated by computing finite dif- ferences of the Sl-based growth model. Results Patterns in humerus growth and aging Carapace length has a strong linear association with humerus diameter (7=0.643 + 2.326A: (where y=SCL cm; A''=humerus diameter mm), r- =0.98, P<0. 001, « = 104 includ- ing the hatchlingl. Humerus growth-increment width, on the other hand, is nonlinearly associated with humerus diameter at the point of growth (Fig. 2 ). Specifically, growth increments tend to be larger when the turtles are smaller (i.e. at smaller humerus diameters) and decline as the tur- tles grow. Variation in humerus increment width (growth rate) shows the same pattern. The estimated aging rate, as the reciprocal of growth rate, increases as the turtles grow. The aging rate does not increase uniformly (Fig. 3). Rather, it increases gradually in small turtles, plateaus over a broad range of length for mid-size turtles, increases abruptly as turtles approach maturity, and maintains an increased rate as the mature turtles grow. Age and growth-rate estimates In the CF-method analysis, the correction factor, C, was estimated as 1.14 yr/mm. The resulting age estimates range from 4.1 to 34.6 yr (/;=70; excluduig the hatchling with age zero). The smallest turtle in the sample had the lowest age estimate and the two largest turtles, the high- est estimates. Only 68% could be aged by the CF method. Skeletochronology requires a pattern of distinct layering within the bony element examined. Such patterns are most evident in the smaller, presumably younger, individ- uals, and the frequency of individuals with distinct perios- teal layers decreases as body size increases. In selecting specimens for the CF analysis, growth layers were suffi- ciently distinct to estimate the number of resorbed layers, and hence the age, in decreasingly fewer turtles: 899c of turtles in the 30-69 cm SCL group, 729;^ in the 70-79 cm group, 18% in the 80-89 cm group, and 29% in the >89 cm gi'oup were used in the CF analysis. Of the individ- uals for which we were unable to obtain an estimate of resorbed layers, a nearly equal number (48%) had fibro- papillomas. The prevalence of tumors for the CF-aged sub- sample (31%) was somewhat less than in the total sample (37%). Importantly, the tumor prevalence in the subsam- 120 Fishei-y Bulletin 100(1) 5-| 4- width (mm) CJ 1 Increment 1 1 m ■■■§■ ^m m m ^ 10 20 30 40 50 Humerus diameter (mm) Figure 2 Humerus growth increment width (the increase in humerus diameter) in relation to humerus diameter (mean of the inner and outer diameters i. Mean diameter better reflects the size of the humerus during the entire growth inten'al than does outer diameter e-. 5- / f '- / ■>^^ / aging rate CO 1 / ■D CD E 7 . / 111 >^-^*^,«>^ 1 - ^r^^^^ " ^■^''^mm n ^^-^ ^ ! - 1 1 ' 1 1 10 20 30 40 50 Humerus diameter (mm) Figure 3 Estimated "aging rate" (average of the reciprocals of the humerus growth- increment widths) in relation to humerus diameter and a fitted smoothing spline. The expected age at a given humerus diameter is obtained by inte- grating the spline up to the specified diameter Not all members of the sample could be aged by the CF method; sec explanation in ■■^hlterials and methods" section. pie retained for CF analysis and the sub- sample e.xcludecl from analysis were not significantly different ichi-square test of homogeneity Z"=--80. 1 df). A nonparamet- ric growth model was estimated by fitting a scatterplot smooth to the carapace length and estimated age data (Fig. 4A). Because the CF method could not be applied to the largest turtles in the sample, the model provided no information about growth in mature turtles. Age estimates for posthatchling turtles based on the SI method range from 3.3 to 49.4 yr (n = 103). The nonparametric growth model for these data fitted very well (Fig. 4B ), a result iiot unexpected giv- en the dependence of the SI age estimate on humerus diameter and the tight linear relationship between humerus diameter and carapace length. Variation in carapace length around the SI growth model reflect- ed only variance in the linear relationship between carapace length and humerus di- ameter: it did not incorporate variance in the age of turtles at a fi.xed humerus di- ameter or estimates of their age. The boot- strap distributions of age estimates for the SI method suggested that age estimates for a turtle can vary by up to 10 years I Fig. 5 A I. Nevertheless, the nonparamet- ric confidence inten'als for predicted mean length at age were narrow (Fig. 5B). Although the CF and SI methods yield- ed different estimates of age iFig. 4). over a wide range of carapace lengths, the dif- ferences were not striking. The magnitude and direction of differences in age esti- mates depend on SCL. Judging from the fitted smooth curves, for turtles up to about 81 cm SCL. the correction factor method was expected to give estimates of age up to 2 years higher than those produced by the spline-integration meth- od. For larger turtles, however, age esti- mates derived from the correction-factor method were predicted to be as many as 10 years (or more) lower than those generat- ed by spline integration iFig. 6). The /-tests (Table 1) indicated there were highly sig- nificant differences in age estimates be- tween methods within the 40-50 cm size group (/=4.36, 9 df) and the 50-60 cm group ( f =4.66, 12 df). Differences for the 30-40 cm. 60-70 cm, and 70-80 cm size groups were nonsignificant; samples were too small for comparisons in other size groups. The Molokai mark-recapture data al- lowed a visual comparison of observed growth and growth predicted by the CF and SI models derived from skeletochro- Zug el a\ Age and growth of Hawaiian Chelonia 121 - lOOi w 80 ■^J^T-"-^ Spline integration 20 30 Estimated age (yr) 40 50 Figure 4 Relationship betwoen observed carapace length and estimated age (squares) for the two methods of age determination: correction factor lAi and spline integration (B). Fitted gi'owth models (cui-ves) are cubic smoothing splines. Table 1 Mean age estimation for correction-factor and splmc-integi-ation method significantly between estimation methods for turtles in the middle length s for 10-cm length groups, groups. * " indicates a sign Estimates of mean age vary ficant difference. Length group (SCL.cm) Sample size Mean age estimate lyri / Idfl.P Correction-factor method I OF I Spline-integration method (SI) 30-40 14 7.3 6.9 1.09 1131,0.296 40-50 10 12.9 11.0 4.36 19). 0.002 50-60 13 19.1 15.8 4.66 1121,0.001 60-70 14 22.1 22.1 -0.06 1131.0.952 70-80 13 25.3 25.6 -0.67 1121,0.514 nological data (Fig. 7). In plotting the growth vectors for marked (tagged) turtles, we fixed the origin of each vector by assuming that the carapace length at time of first capture was given exactly by the CF- or Sl-based growth curve. The obsei-ved length at recapture and time at liberty then determined the cndpoint coordinates of the growth vector. Despite considerable variation in the ob- served gi'owth of the marked turtles, the growth vectors were generally concordant with the growth model predic- tions (Fig. 7). 122 Fishery Bulletin 100(1) 100 100 10 20 30 40 Estimated age (yr) 50 Figure 5 (A) Bootstrap distributions of estimated age at observed SCL generated by resam- pling the data for aging rate to humerus diameter 100 times and by applying the spline-integration method to each bootstrap replicate. iBi Corresponding distribu- tion of growth curves, described by the bootstrap mean growth cui-ve (line) and the approximate 50'^f and 95^c confidence intervals for predicted mean length at age (outer and inner edges of solid region). Age- and length-specific growth rates The box plots of the Sl-estimated carapace growth rates (Fig. 8) indicated relatively fast growth for smaller tur- tles, a reduced growth rate remaining fairly constant over intermediate length classes, and declining growth rates in larger, mature turtles. Mean carapace gi'owth rate declined from 4.4 cm/yr for turtles in the 20-30 cm SCL group to less than 1 cm/yr for mature turtles in the 90-100 cm group, and remained around 2.0-2.5 cm/yr for imma- ture juveniles in the intermediate length groups (Table 2, Fig. 8). Differences in mean growth rate among the intermediate length groups were not significant. The aver- age growth rates predicted by first differences of the SI- based growth model (Fig. 9) indicated a similar pattern, as expected, showing a decrease in growth rate during the first decade of life (when Hawaiian gi-een turtles are still foraging in the open ocean or in the early years of their residence in inshore habitats), relatively constant growth during the next 15-20 year interval, and a further decline in growth rate as the turtles approach 30 years of age. The gi-owth rate appears to remain low in older turtles. Discussion Age and growth-rate estimates Age estimates by both methods indicate an age range of 4-49 years for the Hawaiian green seaturtles in their coastal habitats. The pelagic juvenile (28.7 cm SCL) in our sample was about four years old (4.1 and 3.3 yr by CF and SI methods, respectively). The smallest juveniles (35-37 cm SCL) in coastal waters were 6 to 9 years old by the CF Ziig et al Age and growth of Hawaiian Chclonia 123 40 60 Straight carapace length (cm) 80 100 Figure 6 Difference in estimated age (yr) between the correction-factor method and the spline-integration method within the comparable range of carapace lengths (SCL). For positive values (black region!, the CF method gives older ages and for negative values (striped region), younger ages. methoci and 4 to 10 years old by the SI method. These age estimates for Hawaiian greens in the last years of their pelagic developmental stage are similar to those reported for C. mydoK populations in the southern Great Barrier Reef (SGBR) (5-6 yr; Chaloupka et al., in press) and for the Atlantic coast of central Florida (3-6 yr; Zug and Glor, 1999). The smallest C. mydas turtle in the Florida sample was 28 cm SCL (several others were less than 35 cm), whereas the smallest SGBR turtle was 38.5 cm CCL (Limpus and Chaloupka, 1997) and the smallest Hawai- ian specimen was 34.8 cm SCL, indicating an earlier shift from pelagic to benthic life in Florida greens. The growth of the juvenile turtles predicted by both CF and SI models is consistent with the growth observed in the Molokai mark-recapture sample, but predictions of the CF model depart from the tag-recapture results in older turtles (Fig. 6). Mean growth rates for smaller (30-60 cm) turtles estimated from our transformed humerus incre- ment data (Table 2) were about half as high as growth rates reported for turtles of the same size in most Atlan- tic and Caribbean locales (based on tagging and skeleto- chronology: Tables 2 and 3 in Zug and Glor, 1999). Our estimates were similar to the observed growth rates of tagged turtles in Kiholo Bay, Hawai'i (Balazs et al., 2000) and nearly twice as high as rates observed in some other Pacific samples (Galapagos. Heron Island; Tables 2 and 3 in Zug and Glor, 1999). Subsequent studies of Australian populations (Limpus and Chaloupka, 1997; Chaloupka et. al., in press) have shown a mid-juvenile growth rate more similar to our estimates; however, growth rate is associated with a gi-owth surge in the Australian turtles over a narrow mid-juvenile length range (50-60 cm SCL). Such a spurt Table 2 Growth rate f tati ^tics by length c ass Rates were derived from spline-in tegi ation data. Grow th ■ate (cm/yr) Length gi'oup (SCL. cm) Sample size Mean SD 20-30 9 4.4 2.2 30-40 37 3..5 2.7 40-.50 67 2.1 1.2 .50-60 53 2.3 1.0 60-70 62 2.2 0.9 70-80 21 2.1 1.0 80-90 12 1.3 0..5 90-100 7 0.6 0.3 in growth was not evident in our Sl-based growth cui"ve, and growth rates were fairly constant in the 40-80 cm size classes (Fig. 8). Our age and growth estimates pertain to the Hawaiian population as a whole, because the sampled turtles orig- inated from locations throughout the archipelago. Some variation in age and growth between island foraging groups is likely, given the extensive latitudinal range of the habitats and the associated variation in physical and biological parameters affecting growth (Balazs, 1982). A geographic analysis of age and growth will be the subject of a future study. Future study will also investigate the 124 Fisheiy Bulletin 100(1) UU ' A .l^r 80- ^^ 60- ^ 40- Correction factor 20- n- 1 1 10 20 30 40 50 UU - B > 80 ^^ 60- 7\^^^P wr^ 40- ^ Spline integration 20 - / - 1 1 ; 10 20 30 Estimated age (yr) 40 50 Figure 7 Obsei-i'ed gi'owth vectors of marked Molokai green seaturtles between release and recapture (vectors) in relation to the gr-owth predicted by non- parametric models of SCL and age (cui-ves); age was estimated by the two methods: correction factor (A) and spline integration (Bl. The turtle's length at release is assumed to fall on the growth cur\'e. effects of fibropapillomatosis and gender on growth rates among the Hawaiian population. Age-estimation methods A key assumption of both the CF and SI methods is that each estimated humerus growth layer represents 1 year of gi-owth. This assumption has been validated only recently (Hohn and Snover' ). Hawaiian turtles tagged and injected with tetracycline have been recaptured, and these turtles show the appropriate number of LAGs for the years since their receipt of tetracycline. Furthermore, strong support is provided by the consistency of the growth model pre- dictions with obsei-ved gi-owth in tagged Molokai turtles. Additional justifications have been advanced in other stud- ies of seaturtle humerus LAG formation (e.g. Zug and Glor, 1998; Coles etal., 2001). ' Hohn, A., and M. Snover. 2001. Personal commun. Beaufort Laboratory, Southeast Fisheries Science Center. Beaufort, NC. Both the CF and SI methods require histological prepa- ration and analysis of humerus sections. But they use the same skeletochronological data in independent and differ- ent ways to estimate the total number of humerus growth layers. The CF method assumes that a constant humerus growth rate (the "correction factor") applies to the resorp- tion core regardless of the diameter of the core and despite the fact that periosteal increment width decreases with length of the turtle (Fig. 2). The correction factor, C. is esti- mated from a subset of the skeletochronological data tak- en from juvenile turtles only, i.e. excluding larger turtles likely to have narrower increments in the outer region of the humerus. Even so, the CF estimates of age for juve- nile turtles appear to be biased upward (Fig. 5), suggest- ing that the constant correction factor also failed to reflect the effect of wider increments deposited in the early years of life. In age estimation, the CF method can be applied only to turtles displaying a complete set of periosteal lay- ers, i.e. distinct LAGs, from the resorption core to the outer margin of the humerus. Zug el a\ Age and growth of Hawaiian Chelonia 125 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100 Straight carapace length (cm) Figure 8 Variation of the growth rate estimates (Sl-based) by carapace length class. Each box plot shows median (white bar), limits of 2'"' and 3"' quartiles (solid notched box), range (brackets), and outliers (solid lines). The SI metho(3 models increment-width variation over the entire distance from humerus center to outer margin by using a nonlinear, nonparametric smoother. The model is estimated from all available sample data with two or more LAGs and associated diameter measurements with- out regard to size of the turtle. The SI model can be ap- plied to age all turtles for which the outside diameter of the humerus section has been measured. In the skeletochronological sample we studied, the CF method gave age estimates significantly higher than the SI method for turtles of intermediate length (Table 1), but expected differences for such turtles were no greater than about 2 years. On the other hand, based on current data the CF-based model gives much lower age estimates than the Sl-based model for turtles longer than about 86 cm SCL (Fig. 5). Thus although either method may suffice for aging juvenile Hawaiian green seaturtles, only the SI method appears to provide support for inferences about growth in turtles larger than about 80 cm. The CF method is computationally simpler, because it involves only linear regression rather than fitting and integrating a nonpara- metric smoother, and this consideration may recommend it to some users. Based on our experience with Hawaiian green seaturtle data, the main issue in judging the two techniques ap- pears to be the assumption with the CF method that hu- merus growth is linear. In reality, it is curs-ilinear, and the SI method explicitly models this cun'ilinearitv. Moreover, the SI method makes fuller use of available humerus in- crement data than the CF method. Further comparisons of the methods with additional skeletochronological data sets are recommended. Ecological precis 1 Chelonia myclas within the Hawaiian Islands is a com- ponent of the larger Indo-Pacific C. inydas gene pool, yet the nesting females of the Hawaiian population comprise a distinct genetic unit and contain a unique mtDNA haplotype (Bowen et al., 1992). 2 Except for the posthatching pelagic phase, the green seaturtles of the Hawaiian coastal waters are year- round residents, and all known individuals reproduce within the Hawaiian island chain, predominantly on the beaches of the Northwestern Islands at French Frigate Shoals (Balazs, 1998; Wetherall et al, 1999), 3 Skeletochronological age estimates indicate that Hawai- ian juveniles exit the pelagic phase between the ages of 4 to 10 years. 4 In coastal waters, juveniles 10 years and older possess a relatively constant growth rate until about 28 to 30 years (approximately 80 cm SCL), then growth begins to slow as individuals attain sexual maturity. 5 The mean SCL of nesting females is 92 cm (range 81- 106 cm; Balazs, 1980), suggesting ages of 30 or more years at first nesting for some individuals. 126 Fishery Bulletin 100(1) 6 5 4H 3 2 1 20 30 Estimated age (yr) 20 40 60 80 100 Straight carapace length (cm) Figure 9 Growth rates of Hawaiian Clielonia inydas in relation to age and carapace length, estimated by taking first differences of the Sl-based growth model. Acknowledgment We thank the following individuals and organizations for their valuable contributions to this research; A. Aguirre, D. Akaka, G. Antonelis, S. Eames, W. Gilmartin. S. Hau. D. Heacock, J. Kendig, R. Morris, W. Puleloa, L. Hallacher, W. Dudley, J. Coney, S. Patton, R. Sparks, M. Rice, G. Watson, M. Wisner, T. Work, the State of Hawaii Depart- ment of Land and Natural Resources, the Hawaii Prepara- tory Academy, Makai Animal Clinic, and the Marine Option Program of the University of Hawaii (Manoa and Hilo). George Zug thanks the NOAA-NMFS-SWFSC Honolulu Laboratory and SI/NMNH Department of Systematic Biol- ogy for encouraging and supporting his skeletochrono- logical research. Literature cited Balazs, G. H. 1979. Growth, food sources and migrations of immature Hawaiian Chelonia. Marine Turtle Newsl. 10:1-1.3. 1980. Synopsis of biological data on the gi-een turtles ni the Hawaiian Islands. U.S. Dep. Commer., NOAA Tech. Memo. NOAA-TM-NMFS-SWFC-7, p. i-ix, 1-141. 1982. Growth rates of nnmature gi-een turtles in the Hawai- ian Archipelago. In Biology and conservation of sea tur- tles (K. A. Bjorndal, ed.). p. 117-125. Smithsonian Inst. Press, Washington, D.C. 1998. Sea turtles, 3'"'' ed. /;; Atlas of Hawaii (S. P. Juvik and J. O. Juvik, eds.), p. 115. Univ. Hawaii Press, Honolulu, HI. Balazs, G. H., W. C. Dudley, L. E, Hallacher, J. P. Coney, and S. K. Koga. 1994. Ecology and cultural significance of sea turtles at Punalu'u, Hawaii. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFSC-35 1:10-13. Balazs, G. H., W. Puleloa, E. Medeiros, S. K. K. Murakawa. and D. M. Elhs. 1999. Growth rates and incidence of fibropapillomatosis in Hawaiian green turtles utilizing coastal foraging pas- tures at Palaau, Molokai. U.S. Dep. Commer, NOAA Tech. Memo. [1998] NMFS-SEFSC-415:130-132. Balazs. G. H., M. Rice, S. K. K. Murakawa, and G. Watson. 2000 Growth rates and residency of immature green tur- tles at Kiholo Bay. Hawaii. U.S. Dep. Commer, NOAA Tech. Memo. NMFS-SEFSC-436:283-285. Bolten, A. B., K. A. Bjorndal. J. S. Grumbles, and D. W. Owens. 1992. Sex ratio and sex-specific growth rates of immature green turtles, Chelonia mydas. in the southern Bahamas. Copeia 1992:1098-1103. Bowcn, B. W., A. B. Meylan, -J. R Ross, C. J. Limpus. G. H. Balazs, and J. C. Aviso. 1992. Global population structure and natural history of Zug et a\ Age and growth of Hawaiian Chclonia 127 the p'een turtle [Chclonia mydcia) in terms of matriarchal phylogeny. Evolution 46:865-881. ( 'haloupka, M. Y., C. J. Limpus, and J. D. Miller. In press. Sea turtle growth dynamics in a spatially struc- tured population. Can. J. Zool. < 'halou|)ka, M. Y., and .J. A. Musick. 1996. Age. growth, and population ilynamus, /;; The liiol og>' of sea turtles (P. L. Lutz and J. A. Musick. eds.). p. 233- 276. CRC Press, Boca Raton. Fl. Coles. W. C, J. A. Musick, and L. A. Williamson 2001. Skeletochronology validation from an adult logger- head {Caretta caretta). Copeia 2001:240-242. Cnnise, D. T., L. B. Crowdcr, and H. Caswell. 1987. A stage-based population model for loggerhead sea tur- tles and implications for conservation. Ecology 68:1412- 1423. Hardle, W. 1990. Applied non-parametric regi'ession. O.xford Univ. Press. New York. NY. 333 p. Hastie. T. J., and R. J. Tibshirani. 1990. Generalized additive models. Chapman and Hall. New York, NY, 335 p. Limpus, C. J., and M. Y. Chaloupka. 1997. Nonparametric regi-ession modelling of green sea turtle growth rates (southern Great Barrier Reef). Mar Ecol. Prog. Ser 149:23-34. MathSoft. Inc. 1999. S-PLUS 2000 professional: modern statistics and advanced graphics [software]. MathSoft. Seattle. WA. Murakawa. S. K. K., G. H. Balazs, D. M. Ellis. S. Hau. and S. M. Eames. 2000. Trends in fibropapillomatosis among gi"een turtles stranded in the Hawaiian islands, 1982-98. U.S. Dep. Conimer., NOAATech. Memo. NMFS-SEFSC-443:239-242. Parham.J. F.andG. R. Zug 1998. Age and growth of loggerhead sea turtles iCaivtta caretta) of coastal Georgia: an assessment of skeletochrono- logical age-estimates. Bull. Mar. Sci. 11997] 61:287-304. Ralston. S.. and G. Miyamoto. 1983. Analyzing the width of daily otolith increments to age the Hawaiian snapper. PristiponiDttlcn filamcntosus. Fish. Bull.81:.52.3-535. Wetherall, J. A.. G. H. Balazs. and M. Y Y Yong, 1999. Statistical methods for gi'ccn turtle nesting surveys in the Hawaiian Islands. U.S. Dep. Commer., NOAATech. Memo. 11998] NMFS-SEFSC-415:278-280. Work. T. M., and G. H. Balazs. 1999a. Causes of gi'een turtle iChclnnin niydafi) morbidity and mortality in Hawaii, U.S. Dep. Commer. NOAA Tech. Memo. 119981 NMFS-SEFSC-415:291-292. 1999b. Relating tumor score to hematology in green turtles with fibropapillomatosis in Hawaii. J. Wildl. Disease 35: 804-807. Zug. G. R.. G. H. Balazs, and J. A. Wetherall. 1995. Growth in juvenile loggerhead seaturtles iCaretta caretta ) in the North Pacific pelagic habitat. Copeia 1995: 484-487. Zug. G. R.,andR. E. Glor 1999. Estimates of age and gi-owth in a population of green sea turtles tChelouia mydas) from the Indian River lagoon system. Florida: a skeletochronological analysis. Can. J. Zool. 11998] 76:1497-1506. Zug, G. R., A. H. Wynn. and C. Ruckdeschel. 1986. Age determination of loggerhead sea turtles, Caretta caretta, by incremental growth marks in the skeleton. Smithson. Contrib. Zool. (427):l-34. 128 Estimates of lobster-handling mortality associated with the Northwestern Hawaiian Islands lobster-trap fishery Gerard T DiNardo Edward E. DeMartini Honolulu Laboratory, Southwest Fisheries Science Center National Marine Fisheries SeiA/ice, NOAA 2570 Dole Street Honolulu, Hawaii 96822 E-mail address (for Gerard T DiNardo) gdinardoShonlab nmfs hawaiiedu Wayne R. Haight Joint Institute of Marine and Atmosphenc Research School of Ocean and Earth Science and Technology University of Hawaii, 1000 Pope Road Honolulu, Hawaii 96822 The commercial lobster fishery in the Northwestern Hawaiian Islands ( NWHI ) is a distant-water trap fishery that tar- gets the Hawaiian spiny lobster (Pan- ulirus marginatus) and slipper lobster (Scyllarides squammosus). The ISTWHI are an isolated group of islands, atolls, islets, reefs, and banks that extend 1500 nmi west-northwest of the main Hawai- ian Islands from Nihoa Island to Kure Atoll (Fig, 1). Reported landings in the NWHI peaked at about 2,000.000 lob- sters (spiny and slipper combined) in 1985, and then declined to about 38.000 lobsters from 1986 to 1995 (Fig, 2). The NWHI lobster fishery is man- aged under the Fishery Management Plan for the Crustaceans of the West- ern Pacific Region (Crustaceans FMP) implemented in 1983 and developed by the Western Pacific Regional Fish- ery Management Council (WPRFMC), The National Marine Fisheries Sei-vice (NMFS) is responsible for stewardship of the resource and review and im- plementation of proposed management measures, A variety of management measures have been adopted in re- sponse to declining catches: a limited- entry fishing regime that limited the number of permit holders to 15; a prohi- bition on fishing from January through June when lobsters spawn; an annual catch quota system; a minimum legal tail width (TW) of 50 mm for spiny lob- ster and 56 mm TW for slipper lobster. which are close to the sizes at first ma- turity for these species in the NWHI; a prohibition on landing berried (ovig- erous) females; and a requirement that traps be equipped with escape vents to reduce capture of undersize lobsters (WPRFMCM, Prior to 1996, fishermen were required to discard all berried and undersize lobsters, which were not counted against the catch quota. The management plan assumed that escape vents allowed substantial num- bers of undersize lobster to escape cap- ture and that undersize and berried lobsters do not die during the discard process. Although research on lobster fisheries has found that escape vents effectively reduce the capture of un- dersize lobsters (Ki'ouse. 1978; Fogarty and Borden, 1980; Harris, 1980; Ever- son et al.. 1992; Skillman et al.-), con- siderable numbers of undersize (hence- forth termed "sublegal") and berried lobsters are caught in the NWHI lob- ster fishery Between 1983 and 1995 the reported lobster discard rate in- creased from 2Q"c to 62^?^ (Fig. 3), re- sulting from changes in the size- and age-structures of the populations and in the areas fished. The average size of spiny lobsters generally increased northwestward from Nihoa along the Hawaiian Archipelago (Uchida et al., 1980). Although as many as 16 banks within the NWHI have been fished, the spatial distribution of fishing effort has shifted to banks in the southeast of the Hawaiian Archipelago where there is a higher concentration of spiny lobsters. Qualitative data collected during the early days of the fishery suggested that mortality associated with the handling and discarding practices of the NWHI commercial lobster-trap fishery might be high (Gooding. 1985; Gooding'^). Un- less discard mortality is explicitly con- sidered, fishing policy decisions can be suboptimal, or worse. Where catch quo- tas are used, the total fishing-induced mortality of the population is greater than expected and can even result in recruitment overfishing. Using an equi- librium yield-per-recruit (YPR) model, Kobayashi^ found that the reproduc- tive potential of the NWHI lobster population more than doubled, and mean weight per individual increased by 22% in a retain-all fishery (all lob- sters brought on deck were retained as catch) if the mortality rate of discard- ed lobsters was high (>75'7f ). Based on these results, the observed high discard rate of sublegal and berried lobsters (62%), and the presumption that the 1 Western Pacific Regional Management Council. 199.5. Fisheiy management plan for the crustacean fisheries of the Western Pacific region, amendment 9. Western Pa- cific Regional Fishery Management Coun- cil. Honolulu. Hawaii. 227 p. - Skillman. R. A.. A. R. Everson, and G. L. Ki-amer. 1984. Prospectus escape vent experimental procedure for the spiny lob- ster fishery under management of the Magnuson Fishery Conservation and Man- agement Act. Southwest Fish. Sci. Cent, Admin. Rep. H-84-1.3, unpubl. report, lip. Honolulu Lab.. Southwest Fish. Sci. Cent., Natl. Mar Fish Serv., NOAA. Honolulu. HI 96822-2396. * Gooding, R. M. 1979. Obsorv-ations on surface-released, sublegal spiny lobsters, and potential spiny lobster predators near Necker and Nihoa. Southwest Fish. Sci. Cent. Admin. Rep. H-79-16. unpubl. report, 8 p. Honolulu Lab.. Southwest Fish. Sci. Cent., Natl. Mar Fish Sei-v., NOAA, Hono- lulu, HI 96822-2396. ■t Kobayaslii, D. R. 2001. Southwest Fish. Sci. Cent. Admin. Rep., in prep. Simu- lated effects of discard mortality on spiny lobster iPanulirus mai-ginatus) sustain- able yield and spawning stock biomass per recruit in the Northwestern Hawaiian Islands. Honolulu Lab., Southwest Fish. Sci. Cent., Natl. Mar Fish Serv., NOAA. Honolulu. HI 96822-2396. Manuscript accepted 11 Mav 2001. Fish. Bull. 100:128-133 (2002). NOTE DiNardo et al.: Estimates of lobster mortality in the Northwestern Hawaiian Islands 129 Hancock . Bank Kure Atoll '^"^"ay Island Salmon Bank ' Pearl & Hermes Reef Laysan Island Lisianski ^ Island Maro Reef Northwestern Hawaiian Islands Ralta Bank Gardner '-■' Pinnacles Necker Island Nihoa 180 Pacific Oci'ciu 175 30 25' Kauai A Molokai Oahu .;'J^^Maui 170" 165 Mam Hawaiian Islands 160 155" Figure 1 Map of the Hawaiian Archipelago. 2500 2000 ^ r-. i / \ 1 1500 / \ ■o (U T3 C / \ ^^\ ^ 1000 J \ /*^ \ Numbe / \ 500 / \ ^^^^^^ ^ 1982 1984 1986 1988 1990 1992 1994 1996 Year Figure 2 Annual commercial landings for the NWHI lobster fishery, 1983-96. mortality rate of discarded lobsters is greater than 75"*, the WPRFMC amended the Crustaceans FMP in 1996 to allow the retention of all lobsters caught in the NWHI com- mercial lobster-trap fishery subject to the quota on total catch (WPRFMC). The WPRFMC also recommended that NMFS conduct experiments to assess mortality associated with possible handling and discarding practices. This study reviews research conducted in the NWHI on fishery-induced mortality of sublegal and berried lobsters. The authors examined on-deck mortality of sublegal and 130 Fishery Bulletin 100(1) 70 60 50 ? "D S 30 a 20 10 1982 1984 1986 1988 1990 Year 1992 1994 1996 Figure 3 Annual estimates of the reported discard rate of lobsters in the NWHl lobster fishery. 1983-95 berried spiny and slipper lobsters that occurred within two days after capture under handling methods known to have been used in the NWHI lobster-trap fishery. Methods Studies of handling-induced mortahty ( referred to as "han- dling mortahty" in the remainder of this note) for spiny lobster were conducted at Necker Island onboard the NOAA ship Towitsend Cromwell during the NMFS Hono- lulu Laboratory's 1996 NWHI lobster sui-vey from 21 to 26 June 1996 and for slipper lobster at Maro Reef from 4 to 11 July 1996 (see Fig. 1). These locations were chosen because of their historically high lobster catches and their overall importance to the commercial lobster fishery over the past 13 years. Handling mortalities were estimated for two handling methods: "dry" and "wet." The effects of on-deck exposure time on handling mortality were also estimated for spiny lobsters handled by the dry method. We defined a particu- lar combination of factors (method and exposure time) as an experimental treatment. For dry treatments, lobsters were held on deck in 30-gallon containers without water and in direct sunlight for 1, 2, or 3 hours. In the wet treat- ment, lobsters were held on deck for 3 hours in shaded 30-gallon containers with circulating seawater. The 2- and 3-h dry treatments represent prevailing commercial fish- ing practices (Anderson'^), whereas the 1-h dry treatment ^Anderson,?. 1997. Personal commun. University of Hawaii, Marine Option Program, 1000 Pope Road, Honolulu. HI 96822. Kazama, T. 1997. Personal commun. Honolulu Laboratory, Southwest Fisheries Science Center, 2.570 Dole Street, Hono- lulu, HI 96822-2396. and the 3-h wet treatments represent possible handling alternatives or mitigative measures. Spiny and slipper lobsters were collected in baited com- mercial lobster traps and sorted (legal, sublegal, and ber- ried). All sublegal and berried lobsters were held in tanks of circulating water until a sufficiently large paired exper- imental treatment sample («=200 lobsters) was collected. Experiments with the 3-h wet and dry treatments were done first, followed by experiments involving 1- and 2-h dry treatments. For each treatment, 100 lobsters (sublegal and berried) were randomly chosen from the tank and placed into two 30-gaUon treatment containers (dry or wet). 50 lobsters/container After the exposure time (1, 2, or 3 hours), lobsters were removed from the treatment containers, their condition recorded as active (i.e. lobsters were capable of tail flexion), weak (incapable of tail flexion but able to move ap- pendages when prodded), or dead. Five lobsters were placed in each of 20 lobster-holding traps (commercial lobster traps with sealed entrances) and held in two 1320-gallon bait- wells with recirculating seawater (21 gallons/minute) for 2 days. It was assumed that the 2-d holding period would al- low ample time for latent effects from a treatment to man- ifest and that each lobster holding trap was a replicate (7! =20 replicates/treatment). To check for possible baitwell effects, five control traps, each containing five previously untreated lobsters, were placed in the baitwells (two traps in the port baitwell and three traps in the starboard bait- well) at the beginning of the holding phase, and their condi- tion was recorded at the beginning and end of the holding phase. At the conclusion of the holding period, the condition of the treated lobsters was again assessed. Handling mortality for each treatment was computed as the arithmetic mean of the percent mortality (dead/ total or (dead -i- weak)/total — see below) observed in the 20 NOTE DiNardo et al Estimates of lobster mortality in the Northwestern Hawaiian Islands 131 100 80- ? 60 ..■■' ■^^^^ ^' 03 t: o 5 40 20 1 1 2 3 Exposure lime (hours) Figure 4 Handling mortality (dead/total) estimates for spiny lobster subjected to 1-, 2 - . and .3-h drv treatments. Dashed lines indicate the 95'^^ confidence limits. treatment replicates. Randomization resampling was used to evaluate estimates of handling mortality. Approximate lower and upper QS'/f confidence limits for handling mor- tality were computed as the 2.5'7f and 97 .59i percentiles of the bootstrap distributions by using the computer pro- gram RT (Manly, 1994). Results Spiny lobster Spiny lobster sample sizes by discard category were suble- gal («=84-88) and berried (n = l2-16) (Table 1 ). All spiny lob- sters subjected to the .3-h wet treatment were active at the conclusion of the 2-d baitwell holding period. For dry treat- ments, the number of active spiny lobsters decreased with increasing exposure time, whereas the number of dead lob- sters increased with increasing exposure time. All control lobsters were active, indicating no baitwell holding effects. Handling mortality (dead/total) for the dry treatment ranged from 12'^ for the 1-h treatment to 707r for the 3-h treatment (Fig. 4) and did not differ between berried and unberried lobsters (P>0.05). Pooled dead and weak lobsters resulted in a handling mortality that ranged from 16% for the 1-h treatment to 77% for the 3-h treatment (Fig. 5). Slipper lobster Handling mortalities for slipper lobster were estimated only for the 3-h dry and wet treatments. Mechanical failure of the baitwell recirculating water pumps during the first 3-h paired dry and wet experiments forced us to reduce the baitwell holding time to 1 day. After the baitwell pumps Table 1 Spiny and slipper lobster sample sizes by discard category for each treatment. NB = nonberried and sublegal; B = berried. Treatment Spiny lobster Slipper lobster NB B NB B 1 hour dry 84 16 — 2 hours dry 88 12 — — .3 hours dry 87 13 87-89 11-13 3 hours wet 88 12 81 19 failed, we repeated the 3-h dry treatment with a 1-d hold- ing period, suspending the holding and control traps 3 m below the sea surface from the Townsend Cromwell. Slipper lobster sample sizes by discard category (suble- gal and berried) for each treatment are shown in Table 1. Most slipper lobsters subjected to the 3-h wet treatment were active at the end of the 1-d holding period. For the 3-h dry treatment, the number of dead lobsters ranged from 14 to 39, and the number of active lobsters ranged from 56 to 83. All control lobsters were active, indicating no baitwell holding effects. Slipper lobster handling mortalities for the 3-h dry treat- ment ranged from 14% to 39% with an average estimate of 27% and were unrelated to berried condition (P>0.05). Esti- mates of handling mortalities with weak and dead lobsters combined ranged from 17% to 44% , with an average of 31%. Handling mortality for the 3-h wet treatment was 1%. 132 Fishery Bulletin 100(1) 100 1 80 „.......-. ■■■;^ S" 60 ....-•■■■" ^^^^^^^'^^ _.■■■■■" 2 40 ,..•••■■' ^^^^'^^^ ...■■••■' 20 .:,--^.-----""" ^ 1 2 3 Exposure time (hours) Figure 5 Pooled handling mortality ((dead + weakl/total) estimates for spiny lobster subjected to 1-, 2-, and 3-h dry treatments. Dashed lines indicate the 95''; confidence limits. Discussion Many factors affect the survival of lobsters discarded in commercial trap fisheries, including capture, handling, and discarding processes. In our experiment we focused on handling factors to assess their impacts on mortality in the NWHI commercial lobster fishery. If the 2- and 3-h dry experimental treatments typify commercial handling practices, then the mortality of dis- carded spiny lobster from handling practices on commer- cial vessels is extreme, ranging from an average of 25% to 45% and 70% to 77%, respectively, depending on how da- ta are pooled. Handling mortality for slipper lobster also appears high (estimated at 31%) but is considerably less than that for spiny lobster; thus spiny lobsters may have a lower handling tolerance than slipper lobsters. Although there are no published estimates of handling mortality for P. i7iarginatiis and S. squa?nmosus, studies on other lobster species suggest that handling mortality in the NWHI lobster fishery is high. Lyons and Kennedy (1981), reporting on P. argiis. estimated that 12.3% of lob- sters died after 30 minutes of exposure to direct sunlight and an average 24.1% after 1-4 hours of exposures. They also found that lobsters exposed for 2-4 hours tended to die within a week following exposure, whereas those ex- posed for only 1 hour survived longer. Laboratory experi- ments by Brown and Caputi ( 1983) on small western rock lobster, Paniiliri/s cygniis, exposed to direct sunlight re- sulted in an expected time to 50%f. mortality that decreased with increasing temperature, ranging from 233 minutes at 27°C to 99 minutes at 31-35°C. Time to 50% mortality was 387 minutes for lobsters held in the shade at 26.5-32°C. Handling mortality, however, represents only a portion of the total mortality of discarded lobsters resulting from their capture, shipboard processing, and subsequent re- lease in the NWHI. Additional mortality resulting from habitat displacement, predation, and other factors associ- ated with discarding might result in total discard mortal- ity estimates that approach 100%. Qualitative evidence suggests that discarded lobsters are subject to high preda- tion from the giant trevally, Caranx ignohilis, which ag- gregate around vessels during fishing operations (Good- ing, 1985). Sorting and discarding lobsters immediately after they are placed on deck appears to reduce total discard mortal- ity. The discarded lobsters would need to be returned to the general vicinity of their capture and as close to the sea floor as possible to avoid the gauntlet of predators in the water column. Brown and Caputi (1986) reported a reduction in recapture rates of displaced undersized rock lobsters compared with nondisplaced rock lobsters and re- lated the reduction directly to predation mortality. Adoption of the retain-all fishery by the WPRFMC in 1996 significantly reduced fishery-induced handling mortality. Although sporadic discarding occurs, current discard rates are less than 1% and will have no detect- able consequence at the population level. The research does, however, provide insight into past fishery-induced handling impacts, which likely contributed to the de- cline in NWHI lobster catches. If future management again reverted to mandatory discarding practices, this research provides information to assess its impact on fishery-induced mortality. However, to fully understand the synergistic effects of catching, handling, and dis- carding practices on mortality in the NWHI lobster fish- ery, additional research to assess the impacts associat- ed with shipboard sorting and releasing (e.g. postrelease predation) is required. NOTE DiNardo et a\ Estimates of lobster mortality in the Northwestern Hawaiian Islands 133 Acknowledgments Robert Moffitt, Donald Kobayashi, Jerry Wethi'iall. and the anonymous referees provided helpful comments on the draft, for which we are very gi-ateful. We also acknowledge the crew of the Townsend Cromnell and the scientific staff who worked to ensure that good biological information is available for analyses such as this. Literature cited Brown. R. S., and N. Caputi. 1983. Factors affecting the recapture of undersized western rock lobster Pomiliru.s cygnuf: George returned by fisher- men to the sea. Fish. Res. 2:103-128. 1986. Conservation of recruitment of the western rock lob- ster (Panulirus cygnus) by improving survival and growth of undersize rock lobster captured and returned by fisher- men to the sea. Can. J. Fish. Aquat. Sci. 43:2236-2242. Everson. A. R., R. A. Skillman, and J. J. Polovina. 1992. Evaluation of rectangular and circular escape vents in the Northwestern Hawaiian Islands lobster fishery. N. Am. Fish. Manag. 12:161-171. Fogarty. M. J., and D. V. D. Borden. 1980. Effect of trap venting on gear selectivity in the inshore Rhode Island American lobster. Homarus americanus. fish- ery. Fish. Bull. 77:925-933. Gooding, R. M. 1985. Predation on released spiny lobster, Paniihriit: nuirgin- atus, during tests in the Northwestern Hawaiian Islands. Mar Fish. Rev. 47:27-35. Harris, K. 1980. Escape gaps in rock lobster pots. Fish. Ind. News Tasmania 2:40-43. Krouse, J, S. 1978. Effectiveness of escape vent shape in traps for catch- ing legal size lobsters, Homarus americanus, and harve.st- able-sized crabs. Cancer borealis and Cancer irroralus. Fish. Bull. 76:425-432. Lyons, W. G., and F. S. Kennedy. 1981. Effects of harvest techniques on sublegal spiny lob- sters and on subsequent fishing yield. Proc. Gulf Caribb. Fish. In.st. 33:290-300. Manly, B. F. J. 1994. RT: a program for randomization testing., version 1.02C. Centre for Applications of Statistics and Mathe- matics, Univ. Otago, P.O. Box 56, Dunedin, New Zealand. Uchida, R. N., J. H. Uchiyama, D. T. Tagami, and P. A. Shiota. 1980. Biology, distribution, and estimates of apparent abun- dance of the spiny lobster, Panultrus marginatus I Quoy and Gaimard) in waters of the Northwestern Hawaiian Islands. Part II: size distribution, legal and sublegal ratio, sex ratio, reproductive cycle, and morphometric characteristics. /;; Proceedings of the symposium on status of resource inves- tigations in the Northwestern Hawaiian Islands, Honolulu, Hawaii, April 24-25, 1980 (R. W. Grigg and R. T. Pfund, eds.), p. 131-142. Univ. Hawaii Sea Grant College Pro- gram MR-80-04. Honolulu. Hawaii. 134 An evaluation of pop-up satellite tags for estimating postrelease survival of blue marlin {Makaira nigricans) from a recreational fishery* John E. Graves Virginia Institute of Marine Science College of William and Mary Rt. 1208 Greate Road Gloucester Point, Virglna 23062 E-mail address gravesfaJvims edu Brian E. Luckhurst Division of Fisheries Department of Agnculture and Fistienes PO Box CR 52 Crawl CRBX, Bermuda Eric D. Prince National Marine Fistieries Service 75 Virginia Beach Dnve Miami, FL 33146 Blue marlin [Makaira nigricans) rep- resent an important commercial and recreational resource throughout trop- ical and subtropical oceanic waters. In the Atlantic Ocean, blue marlin are managed as a single, oceanwide stock. In the most recent assessment of Atlantic blue marim ( ICCAT, 2001), the Standing Committee for Research and Statistics (SCRS) of the International Commission for the Conservation of Atlantic Tunas (ICCAT) estimated the current biomass of blue marlin to be about 40% of that required for maxi- mum sustainable yield (MSY). Further- more, the assessment indicated that the current level of fishing mortality (F) was about four times higher than F,y,gY and that catch levels in recent years were more than twice the equi- librium yield, contributing to a fur- ther decline of the overexploited stock. Based on the most recent stock assess- ment, fishing-induced mortality must be reduced by about 60% to halt the decline of the stock (Goodyear, 2000). The greatest source of billfish (Istio- phoridae) mortality occurs as a result of incidental catches by longline gear deployed for tunas and swordfish (IC- CAT, 1997, 2001). These highly migra- tory species co-occur in the subtropical and tropical epipelagic environment, and all are vulnerable to the pelagic longline. Not all billfish are dead at the time of capture (haulback) on longline gear; data from observers on vessels in the Venezuelan industrial longline fishery indicate that about 49% of blue inarlin caught on pelagic longline gear are alive at the time of capture (Jack- son and Farber, 1998). To reduce billfish mortality ICCAT in 1997 required nations to reduce their landings of Atlantic blue marlin by 25% from 1996 levels. Furthermore, the IC- CAT SCRS has recommended that the Commission consider requiring the re- lease of all live billfish taken on long- line gear (ICCAT 1997, 2001). It is believed that such a management mea- sure would be more acceptable to mem- ber nations than an overall reduction in longline effort that would also re- duce catches of target species. How- ever, representatives from several na- tions have pointed out that there are not sufficient data to estimate postre- lease survival of billfish; therefore the conservation impact of a recommenda- tion requiring live released fish cannot be evaluated. In fact, low recovery rates of billfish tagged and released with conventional tags by recreational and commercial fishermen ( <2%; Jones and Prince, 1998; Ortiz et al, 1998) are con- sistent with high postrelease mortality. However, factors such as tag shedding and failure to report tag recaptures could also account for low rates of tag returns ( Bayley and Prince, 1994; Jones and Prince, 1998). Clearly, data are needed to support or refute the hypothe- sis that the release of live billfish would significantly eliminate fishing mortali- ty of blue marlin (Graves et al.'). Acoustic tracking studies designed to investigate billfish physiology and behavior have provided insights into the postrelease survival of billfish tak- en on recreational gear Specifically, ob- sei-ved and inferred mortalities during the course of the acoustic tracks indi- cate that not all released billfish sur- vive (reviewed in Pepperell and Davis, 1999). Unfortunately, it is not possible to estimate levels of postrelease mor- tality of billfish from previous acoustic tracking studies for several reasons. First, owing to the high cost of ship and personnel tiine, relatively few ani- mals have been investigated in acous- tic tracking studies. Second, because ocean conditions can deteriorate quick- ly, many of the acoustic tracks were for less than 12 hours duration, pro- viding a limited opportunity to ob- ser\'e mortality after 12 hours. Third, billfish were caught and subsequently tracked under a variety of conditions, making cross-study comparisons diffi- cult. Finally, an estimate of postrelease mortality rates resulting from acoustic studies may be biased because in soine cases only healthy fish were selected to carry acoustic transmitters. The development of pop-up satellite tag technology may present a possible means to estimate postrelease mortali- ty of billfish. Although relatively expen- sive pop-up satellite tags reduce the need to use a tracking vessel to follow * Contribution 2416 of the Virginia Insti- tute of Marine Science, College of William and Mary, Gloucester Point, VA 2.3062. ' Graves, J. E., G. Skomal. and E. D. Prince. 1995. Report of the billfish mortality workshop. Contribution rep. MIA-94/ 95-49, 7 p. Southeast Fisheries Science Center, Miami, FL. Manuscript accepted 30 July 2001. Fish, Bull. 100434-142 (2002). NOTE Graves et al An evakialion of iatellite tags for estimating postrelease srirvivai of Mnkaim nigiican^ 135 hillfish on the liigli seas. Pop-up satellite tags are capable (if recording environmental variables over predefined in- ten-als, of detaching from an animal at a designated time, lioating til the surface, and of transmitting the stored data to a satellite. Until now, these tags have been deployed pri- marily on bluefin tuna for relatively long durations (up to nine months) to determine movement patterns (Block et al.. 1998a; Lutcavage et al., 19991. Recovery of tag data has been very good in most cases, with some reported rates in excess of 90% (Block et al., 1998a; Lutcavage et al., 1999). These results suggest that the technology may be well suit- ed for shorter term studies, including the determination of postrelease sui-vival. In this paper we present the results of a preliminary study to evaluate the feasibihty of applying pop-up satellite tag technology to estimate short-term sur- vival of blue marlin. We also include a brief analysis of the movement and behavior of blue marlin that we inferred from the pop-up tagging results. Materials and methods Pop-up satellite tags The Microwave Telemetry. Inc. PTT-100 pop-up satellite tag was used in this study. The tag can withstand a pressure of 1000 psi (equivalent to a depth of about 650 meters) and is sufficiently small (38 cm by 4 cm diameter) that it would not appear to impose a major drag on a large marine tele- ost. such as blue marlin (Block et al.. 1998a). Tags were progi-amed to measure water temperature every hour and record the mean value for each two-hour period for a total of 61 cycles (122 hours). Inclinometer values were taken every two minutes and summed for the periods before tag detachment (pre-pop-up) and after tag detachment (post- pop-up). For each period the inclinometer started with an initial value of 128. If at the time of measurement (every two minutes) the tag was oriented below 30 degrees above horizontal, a value of one was subtracted from the total. If the tag was above 30 degrees above horizontal at the time of measurement, the inclinometer total was increa.sed by one, but could not exceed 255. Final values below 255 indi- cated sufficient forward propulsion such that the positively buoyant tag was depressed below 30 degrees above horizon- tal for certain periods, demonstrating forward propulsion. All nine tags were progi-amed to detach from the fish 122 hours after activation, at which time the memory within each tag would contain 61 direct temperature measure- ments and the pre-pop-up inclinometer value. The five-day attachment period of the pop-up tag was chosen, in part, as a result of a review of data from conventional tag-re- captured blue marlin in the Cooperative Tagging Center (CTC) database (E. Prince, unpubl. data). Of the 160 blue marlin tag returns in CTC that have been validated, ten individuals were recaptured within five days of release, suggesting that some blue marlin are able to survive the catching and tagging event and commence feeding again within a few days. In addition, acoustical tagging studies have shown high sui"vival rates of different marlin species in the first 1-2 days following release, demonstrating that mortality, when it occurred, generally happened within the first 48 hours of release (Pepperell and Davis, 1999). With these considerations in mind, we assumed that the five- day period of tag attachment was an adequate period for catch and release mortality to be expressed. As indicated by Goodyear (in press), the duration of this type of experi- ment should be the minimum number of days necessary to account for postrelease mortality events. Longer periods would allow for greater influence of tag shedding, tag mal- function, and natural mortality, all of which could compro- mise estimates of postrelease survival. Tag deployment Pop-up satellite tags were activated and tested at the start of each fishing day. Blue marlin were caught south- west of Bermuda in the vicinity of Challenger and Argus Banks on standard recreational gear for the blue marlin fishery in Bermuda ( 130 lb test line) by using trolled high- speed lures or skirted dead baits (in most cases with two hooks). All hooks employed in this study were "J" hooks (no. 16/0-20/0). We tagged the first nine fish available to us. Six blue marlin were caught on the vessels we were aboard. Three individuals were taken on other vessels and transferred to the tagging vessel after the fish were brought to leader (brought to the boat): one blue marlin was caught and attached to a drifting buoy until the tag- ging vessel, which was several miles away, could gain access to the fish; and two fish were directly transferred after capture from the fishing vessel to the tagging vessel by using a procedure described in Block et al. ( 1998b). Once fish were brought to leader (reeled to the side of the boat), quieted and secured, the pop-up satellite tag and a conventional (streamer) tag were deployed. Pop-up satellite tags were attached to one end of a 400-lb ( 182 kg) test monofilament leader about 18.5 cm in length, with an outside diameter of 1.8 mm. The other end of the leader was attached to a double barb nylon anchor (about 33 mm long and 10 mm wide) made of medical-grade nylon. The anchors were implanted by using a stainless steel tag ap- plicator modified to accomplish placement to a depth of about 10 cm into the dorsal musculature, about 10 cm pos- terior and 5 cm below the base of the peak of the first dor- sal fin (Fig. 1). Hook location, as well as observations on foul-hooking (tissue damage, bleeding, etc.), were noted at the time of hook removal. Analyses After detachment from the animal, the positively buoyant tags floated to the surface and began transmitting data to satellites of the Argos™' system. Position information and sections of the temperature and inclinometer data were captured with each satellite pass and transmitted to a ground station and ultimately to the investigators by the internet. Data were analyzed to determine net movement from the point of detachment to the point when the tag popped-up (usually the first tag transmittal; however, if the first satellite pass was near the horizon, the location of the second transmittal was used to obtain greater accu- 136 Fishery Bulletin 100(1) Figure 1 Blue marlin with pop-up satellite tag attached. The tag is located below the anterior portion of the dorsal fin (within outlined area). racy). Water temperature was determined from tempera- ture sensor readings by using a calibration provided by the manufacturer Depth was estimated from water tempera- ture values by using temperature and depth relationships provided by the Bermuda Biological Station for Research, which maintains an oceanographic sampling station (sta- tion S) about 13 miles (24 km) to the southeast of the island. Results and discussion Nine blue marlin. with estimated weights ranging from 150 to 425 lb (68 to 193 kg), were tagged between 25 July and 11 August 1999 (Table 1). Four specimens were below the minimum size for tagging recommended by the tag manufacturer (200 lb or 90.9 kg). Fight times ranged between 15 and 35 minutes. Seven fish were initially hooked in the jaw, and two were "foul-hooked" (i.e. outside the jaw and mouth): one in the operculum and one in the dorsal musculature. After tag placement, but before release, one fish that was originally hooked in the jaw became foul-hooked in the ventral musculature. Three of the nine fish were transferred to the tagging vessel after capture. Fish generally quieted down shortly after being brought to the side of the vessel, which maintained a head- way of 4-5 km/li during the tagging operation. Only a few minutes were required to implant the satellite and con- ventional tags, photograph the fish, estimate weight, mea- sure lower jaw fork length (most individuals), and remove the hook. Condition of the fish varied, and three individu- als required resuscitation prior to release. Eight of the nine tags became detached from their re- spective host fish after five days, floated to the surface, and transinitted to the Argos™"" satellite system. Based on the first accurate location of the tags, net displacements ranged from 40 to 134 nmi (72-248 km) with a mean lin- ear displacement of 90 nmi (167 km) for each individual (Fig. 2). These values are in the range reported for blue marlin by Block et al. ( 1992) who followed six blue marlin with acoustic transmitters for periods of one to five days. They noted individual total movements (as opposed to net displacements) of 253 km in about three days, 100 km in five days, and four animals with movements of less than 100 km over the course of the respective tracking periods. Individual marlin in our study dispersed in all directions from their point of release ( Fig. 2 ). The blue marlin tracked by Block et al. ( 1992) and Holland et al. (1990) in Hawaiian waters moved away from the point of capture in several dif- ferent directions. However, the authors noted an orientation of movements to the coastline of the Hawaiian Islands. Our r-eleases were farther offshore and an affinity to the Bermu- da coastline was not evident from the net movement data. Depending on the time of tag activation and the time of tag deployment, up to 61 direct water temperature read- ings, taken every two hours, were obtained for each blue NOTE Graves et al : An evakiation of satellite tags for estimating postrelease survival of Makaiia nigncam 137 Table 1 I'op-up s tagging. under its itellite tag deploynicnt information. "Tran.s ■Resuscitation" indicates the time spent to own power. Tag no. 24040 failed to report. er" mda move a ales hlue whether a ti.sh was moved b marlin through the water ..'tween vessels aftei ifter capture until capture to allow t could swim off Tag no. Deployment Fight time Transfer (yes/no) Hook location E;stimated weight (lb) Resuscitation (no/yes — time) date hour min. 24519 25 Jul 1999 1610 25 yes jaw 400 no 24059 28 Jul 1999 1110 35 yes jaw and ventral musculature 200 no 24522 lAug 1999 1030 20 yes jaw 175 no 24520 2 Aug 1999 1015 15 no jaw 180 no 24033 2 Aug 1999 1245 30 no jaw 425 yes-10 min. 24040 2 Aug 1999 1500 17 no jaw 200 no 24523 3 Aug 1999 1550 15 no operculum 150 yes-8 min. 24527 11 Aug 1999 1255 15 no jaw 350 no 24029 11 Aug 1999 1340 23 no dorsal mu.sculature 150 yes-3 min. marlin (Fig. 3). Temperature readings demonstrated that tagged individuals spent the majority of their time at tem- peratures above 26°C (Fig. 4). The maximum temperature range recorded for any of the eight individuals was 9°C (22-3rC, tag no. 24033). Block et al. ( 1992), using acoustic tracking, determined that the six blue marlin which they tracked spent half of their time in the upper 10 m of the water column in water temperatures 2.5-27°C, and Hol- land et al. (1990) reported that blue marlin in waters off Hawaii remained at temperatures of 26° or greater. Differences in thermal histories were evident among the individuals in our study. Blue marlin no. 24029 (Fig. 3G) spent the vast majority of time at temperatures equal to that of the surface waters (30-31°C). In contrast, individu- al no. 24527 (Fig. 3H) spent much less time in the warmer surface waters and repeatedly moved up and down in wa- ters between 23° and 31°C. Several individuals appeared to remain at or very near the surface for extended peri- ods, evident in Figure 3 as a continuous string of tempera- ture readings at or slightly above 30°C. An analysis of the data examining diurnal-nocturnal periods with tempera- ture (inferred depth) indicated a high level of variability between individuals and no clear pattern was apparent (Fig. 3). In contrast, Holland et al. ( 1990) determined that blue marlin spent a higher proportion of their time (-50%) in the upper 10 m at night than during the day (-25% ). It was possible to infer swimming depths of blue marlin by comparing water temperature values with the temper- ature-depth profiles at station "S" provided by the Bermu- da Biological Station for Research.- All blue marlin en- Although this station is situated 24 km to the southeast of the island, similar temperature-depth profiles would be expected for the general region (Johnson, R. 2000. Personal commun. Bermuda Biological Station for Research, 17 Biological Lane, Ferry Reach, St. George's GEOl Bermuda). This allowed us to use the station S profiles to infer swimming depth, realizing that, depending on when an animal was tagged and where it moved, there would be some differences for which we could not account. tered cooler waters at various times during the five-day period, with excursions to depths as great as 40 meters. Temperature records were consistent with the tagged blue marlin actively undertaking vertical movements in the upper 40 meters of the water column. However, six of the eight fish spent >75% of their time in the upper 10 m of the water column for the five-day duration of the study. If the data from all eight fish are pooled, this yields a mean value of 79.9% (SD 15.8%) of the time spent in this zone. This is a higher percentage of time spent in the upper 10 m than that obsei-ved by Block et al. ( 1992), who reported that fish spent about half of the time in this zone. How- ever, this comparison should be viewed with some caution because the Block et al. (1992) data were based on con- tinuous tracking, whereas each data point in our analysis was the average of two hourly measurements. All post-pop-up inclinometer values were 254 or 255, where 255 represented the maximum (vertical) inclinom- eter value expected for an upright, floating tag. Pre-pop- up inclinometer values ranged from 203 to 251, with three individuals at 233 and four between 247 and 251. These values indicate tags were inclined at an angle below 30 degrees above horizontal for more than 40% of the 1830 sampling times for each individual, and are consistent with sufficient forward propulsion to depress the positive- ly buoyant tag more than 60 degrees from vertical. There was no correlation between pre-pop-up inclinometer values and net displacement. The fish with the largest net dis- placement (no. 24059) had the second highest inclinometer value. This was not unexpected because the difference be- tween the lowest and highest pre-pop-up inclinometer val- ues represents a minor difference in the time the tag was depressed below 30 degrees above horizontal. Also, the re- lationship between total movement and net displacement could be quite different for different individuals. Three different lines of evidence provided by the pop- up satellite tags (net movement, water temperature, and tag inclination) each suggested that at least eight of the 138 Fishery Bulletin 100(1) Sft-W 63°W ws Jl^N 13^2405^1134 4 Nml Tag 2-10:>) 1113 1 Nm) \ Tae24033(')S,6Nm) Tag 24520 (94. R NmJ Ias2«22(8S2Nm) 'Tjg245l9(55 2 Nm) ^^/T^g 24523 (J")*) Nm) Bermuda Islands \Tjji24527(89 8Nm) 32°N 63°W Figure 2 Map showing points of release (squares) and points of recovery (end of straight lines) for eight of nine blue marlin equipped with pop-up satellite tags near Bermuda, 25 July-11 August 1999. The pop-up satellite tag number for each fish and straight line distance between point of release and geolocation where tag transmitted data to the Argos satellite (given in parentheses) are provided. nine blue marlin caught on recreational gear survived for five days following capture, tagging, and release events. The net movement data indicated a broad dispersal of the eight fish in different directions that cannot be explained by local currents. In contrast to the differing direction of movement, the net displacements of the eight fish were fairly similar. The mean displacement of 89.25 nmi over a five-day period, compares favorably with blue marlin swimming velocities of 1-2 nmi/h reported from acoustic tracking studies (Holland et al., 1990) and is consistent with the constant slow swimming of the individuals. Al- though currents could have accounted for some of the net displacement, inclinometer values indicated that all eight individuals were actively swimming. The water temperature measurements indicated that each blue marlin actively undertook dives into cooler wa- ter throughout the course of the five days. All eight in- dividuals spent the vast majority of their time in waters with temperatures of 26°C or greater, and no readings be- low 22°C were recorded. The successful data recording, tag detachment, and transmission of eight of the nme pop-up satellite tags begs the question of what happened to the one tag that failed to report. It is not possible to distinguish between the postre- lease mortality of a tagged blue marlin and the mechani- cal failure of a pop-up satellite tag. If a marlin dies and sinks in deep water, the attached pop-up satellite tag even- tually will be crushed by increasing hydrostatic pressure. NOTE Graves et a\ An evaluation of satellite tags for estimating postrelease survival of Makaim nigricans 139 c tag number 24519 10 20 30 40 50 60 70 80 90 100 110 12 B tag number 24059 10 20 30 40 50 60 70 80 90 100 110 tag number 24522 50 60 70 Time (hours) Figure 3 Temperature and inferred depth for eight of nine blue marlin equipped with pop-up satellite tags near Bermuda. 25 July-11 August 1999. See text for explanation of inferred depth and Table 1 for release information correspond- ing to each tag number Hours of darkness are shaded on the time line. The blue marlin whose tag did not report (tag no. 24040. Table 1 ) was hooked in the jaw, caught in less than 20 min- utes, did not require resuscitation, was quickly tagged, and actively swam away from the boat when released. Shark predation on released billfish has been reported (Holland et al., 1990; Pepperell and Davis, 1999); therefore mortal- ity cannot be excluded despite the apparent vigorous con- dition of the fish. Failures in component subsystems could account for the failure of reporting from a pop-up tag. A detailed analysis of the reliability of each tag component could be undertaken, but several factors external to the tag could also result in a failure of reporting. Tag man- ufacturer innovations and upgrades of the systems will allow researchers to better identify mortalities, but they will not completely solve the problem of discriminating between tag failure and fish mortality. Nonreporting tags would have significant consequences for efforts to make ocean-wide estimates of postrelease sui-vival. The ability to account for all pop-up satellite tags deployed is directly related to the accuracy of the resulting estimates of postre- lease survival (Goodyear, in press). Nonreporting satellite tags introduce uncertainty that cannot be quantified in the estimates of postrelease survival, thus compromising meaningful conclusions. Excluding nonreporting tags from the analysis decreases precision of the estimate, and in- cluding mortalities biases the survival estimate down- ward. Further, any extension of the 5-day pop-up period to allow study of possible delayed effects of tagging should involve careful consideration of the benefits and the li- abilities that longer durations might have on estimating postrelease survival (Goodyear, in press) Successful tagging and reporting of pop-up tags from four fish under 200 lb (90.9 kg) indicate that the size and design of the PTT-100 tag is tolerated by smaller blue mar- 140 Fishery Bulletin 100(1) tag number 24033 60 70 80 Time (hours) Figure 3 (continued) 100 110 120 NOTE Graves et al,: An evaluation of satellite tags for estimating postrelease survival of Makaira nigricans 141 % 15 31+ 30 29 28 27 26 25 24 23 22 Temperature ( C) Figure 4 Distribution of time at temperature for eight of nine blue marlin equipped with pop-up satellite tags near Bermuda. 25 July-11 August 1999. Histogram represents combined data for the five-day period of data transmission for each individual. lin than that recommended by the manufacturer-, at least in the short term. Thus a pop-up tag of this size might be tested on even smaller specimens or other target species to expand the study of behavior in a wider size range of species than was originally thought possible. Acknowledgments The authors would like to thank Captains Alan DeSilva, Andrew Card, Gilbert Amaral and Bobby Rego for pro- viding assistance in deploying the pop-up tags. We also express our appreciation to the anglers and to the orga- nizing committee of the Bermuda Billfish Tournament for their support. Paul Howey and Molly Lutcavage provided expert technical advice. Phil Goodyear made helpful com- ments on the manuscript. Support for this project was obtained from the Bermuda Department of Agriculture and Fisheries, the ICCAT Enhanced Research Program for Billfish, and Hewit Family Foundation. Literature cited Bayley, R. E., and E. D. Prince. 1994. A review of tag release and recapture files for Istiophoridae from the Southeast Fisheries Science Cen- ter's Cooperative Gamefish Tagging Program. Int. Comm. Cons. Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap. 41:,527-548. Block, B. A., D. T Booth, and F. G. Carey. 1992. Depth and temperature of the blue marlin. Mahaira nigricans, observed by acoustic telemetry. Mar Biol. 114: 17.5-183. Block, B. A.. H. Dewar, C. Fai-well, and E. D. Prince. 1998a. A new satellite technology for tracking the move- ments of Atlantic blucfin tuna. Proc. Natl. Acad. Sci. 95:9384-9389. Block, B. A., H. Dewar, T D. Williams, E. D. Pnnce, C. Farwell, and D. Fudge. 1998b. Ai'chival tagging of Atlantic bluefin tuna iThunnus lliynniis tliynniis). Mar Tech. Soc. J. 32:37-46. Goodyear, C. P. 2000. Biomass projections for Atlantic blue marlin: potential benefits of fishing mortality reductions. Int. Comm. Cons. Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap. 52:1502-1506. In press. Factors affecting robust estimates of the catch and release mortality using pop-up tag technology. In Sympo- sium on catch and release in marine recreational fisheries (A. Studholme, E. D. Prince, and J. Lucy, eds). Spec. Pub. Am. Fish. Soc. Holland. K., R. Brill, and R. K. C. Chang. 1990. Horizontal and vertical movements of Pacific blue marlin captured and released using sportfishing gear. Fish. Bull. 88:397-402. ICCAT (International Commission for the Conservation of Atlantic Tunas). 1997. Report for biennial period. 1996-97. Part 1 (1996), vol. 2, 204 p. ICCAT, Madrid, Spain. 2001. Report of the fourth ICCAT billfish workshop. Int. Comm. Cons. Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap 53:1- 22. Jackson, T L., and M. I. Farber 1998. Summary of at-sea sampling of the western Atlantic Ocean, 1987-1995, by industrial longline vessels fishing out of the port of Cumana, Venezuela: ICCAT Enhanced Research Program for Billfish 1987-1995. Int. Comm. Cons. Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap. 47:203-228. Jones, C. D., and E. D. Prince. 1998. The cooperative tagging center mark-recapture data- 142 Fishery Bulletin 100(1) base for Istiophoridae (1954-1995) with an analysis of the west Atlantic ICCAT billfish tagging program. Int. Comm. Cons. Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap.47:31 1-322. Lutcavage, M. E., R, W. Brill, G. B. Skomal, B. Chase, and P. Howey. 1999. Results of pop-up satellite tagging of spawning size class fish in the Gulf of Maine: Do North Atlantic bluefin tuna spawn in the mid-Atlantic? Can. J. Fish. Aquat. Sci. 56:173-177. Ortiz, M., D. S. Rosenthal, A. Venizelos, M.I. Farber, and E. D. Prince. 1998. Cooperative Tagging Center Annual Newsletter: 1998. U.S. Dep. Commer . NOAATech. Memo., NMFS-SEFSC-423, 23 p. Pepperell, J. G., and T. L. O. Davis. 1999. Post-release behavior of black niarlin Makaira indica caught and released using sportfishing gear off the Great Barrier Reef (Australia). Mar. Biol. 135:369-380. 143 Reproduction of the blacknose shark (Corcharhinus acronotus) in coastal waters off northeastern Brazil Fabio H. V. Hazin Paulo G. Oliveira Matt K. Broadhurst Depaitamento de Pesca, Laboratorio de Oceanografia Pesqueira Universidade Federal Rural de Pernambuco Av Dom Manuel de Medeiros, s/n Dois Irmaos, Recife PE, Brasil, CEP: 52.171-900 E mail address (for F Hazin) fhvhazin gelogica com br The blacknose shark, Carcharhiin/s acronotus, is a relatively small car- charinid, typically inhabiting continen- tal shelf areas in the western Atlantic Ocean, from North Carolina through- out the Gulf of Mexico (Bigelow and Schroeder. 1948) and along the South American coast to Rio de Janeiro (Com- pagno, 1984). The abundance of this shark in nearshore areas throughout its distribution makes it accessible to commercial fishing, mainly from in- shore hook-and-line and gill-net fish- eries (Trent et al.. 1997; Mattes and HazinM. Aspects of the biology of C. aci-ono- tus have been reported by Springer 11938); Bigelow and Schroeder (1948); Clark and von Schmidt ( 1965); Dodrill (1977); Branstetter (1981); Schwartz ( 1984 ); Castro ( 1993 ); and Carlson et al. (1999). Schwartz (1984) provided the most comprehensive synopsis, includ- ing information on their reproduction and life cycle off North Carolina. Many of the other studies were based on rel- atively few specimens collected off the southeastern United States and cor- roborate much of Schwartz's (1984) work, including patterns in spatial and temporal abundances, size at maturi- ty, fecundity, and time of parturition. However, some inconsistencies exist with respect to the duration of the ovarian, gestation, and breeding cycles: i.e. Dodrill (1977) proposed a biennial breeding cycle with gestation taking between 10 and 11 months, Schwartz (1984) suggested a gestation of ap- proximately 9 months, and Branstet- ter ( 1981) observed two gravid females with large ovarian eggs (having con- current ovarian and gestation cycles, copulation having occurred shortly af- ter parturition). Because C. acronotus are represent- ed in catches from various inshore fish- eries and carcharhinids typically are characterized by low rates of popu- lation increase, adequate information about their reproductive cycle is re- quired to facilitate management of stocks. Given the uncertainty of knowl- edge about reproduction in C. acrono- tus and the lack of information for the southern part of their range, our aims in the present study were to provide a preliminary overview of the repro- ductive biologv' and life cycle of the blacknose shark in coastal waters off northeastern Brazil, using available fishery-dependent data. Material and methods Fishing gear used and data collected Carcharhinus acronotus (79 females and 45 males) were collected from the catches of commercial gillnetters and vessels using bottom longlines off the coast of northeastern Brazil (approx. 7°30' to 9°30'S — near Recife) between August 1994 and January 1999. The configuration of fishing gears used remained similar over this period. Gill nets were monofilament, 900 m in length, and had a stretched mesh size of 17 cm and a depth of 70 meshes. Nets were set perpendicular to the beach at depths between 5 and 10 m. Bottom longlines consisted of a multifilament mainline (6 mm in diameter) with up to 100 secondary lines, each approx. 5 m in length and constructed from 3-mm diameter monofilament attached to a wire snood ( 1 m in length). Types of hooks varied among brands, but rela- tive sizes (i.e. 9/0 ) remained similar The main baits were sardine (SardineUa hrasiliensis) and mackerel (Scomber spp. ), although some other species, including sting ray (Aetobarus nari- nari) and skipjack tuna iKatsuwonus pelamis) were occasionally used. Long- lines were set on the continental shelf at depths between 10 and 60 m, but most sets were shallower than 40 m. All fishing gears were set at dusk and hauled the following morning at dawn. All specimens were measured (total length ITLj in the natural position ) and dissected. Reproductive organs were removed and stored in a solution of lO^-'f formalin in seawater prior to be- ing transported to the laboratory. Data collected from females included weight and width of the oviducal gland and the functional right ovary, maximum ovar- ian follicle diameter (MOFD), width of the largest uterus, and, if present, the TL, sex, and number of embryos. Using the methods described by Pratt ( 1993), we examined the oviducal glands of 10 mature females for the presence of spermatozoa. Length and calcification stage of claspers, width of epididymi- des, and the presence of seminal fluid in the ampullae of the ductus deferens were recorded from males. Reproduc- tive organs were measured to the near- est 0.1 mm with vernier calipers. Inferences on stages of reproduction were made according to definitions pro- vided in previous studies on carcha- rinids (e.g. Pratt, 1979; Hazin et al., 2000). Females were categorized into six stages, mainly based on develop- ' Mattos, S. M. G. and F. H. V. Hazin. 1997. Analise dc viabilidade economica da pesca de tubaroes no litoral do estado de Per- nambuco. Boletim Tecnico-cientifico do Cepene. 5: 89-114. IBAMA (Instituto Brasiliero de Meio Ambiente e dos Recur- sos Naturals Renovaveis), Rua Samuel Hardman, s/n Tamandare - PE, Brazil. Manuscript accepted .5 July 2001| Fish. Bull. 100:143-148(2002). 144 Fishery Bulletin 100(1) ment of the oviducal gland, ovary, and uterus. Specimens were considered juvenile if they had undeveloped sexual organs, filiform uteri, and no vitellogenic activity in their ovaries. Pre- ovulatory females had relatively larger ova- ries with orange vitellogenic follicles, but no uterine eggs. Ovulating females had uterine eggs and ripe ova still in the ovary, whereas in gravid specimens ovulation was complete. Postpartum females showed similar-size ova- ries, oviducal glands, and ovarian follicles as those of gravid specimens, but had flaccid uteri that were still slightly enlarged (compared with other nongravid females) indicating re- cent parturition. Individuals that had simi- lar-size uteri as those of pre-ovulatory and ovulating individuals, but smaller ovarian fol- licles with little or no vitellogenic activity were termed "resting" females. Male maturation was evaluated according to development of the tes- tes and claspers. Individuals with relatively short, flexible claspers, and filiform ampullae of the ductus deferens were considered juve- niles. Adults were characterized by elongate and calcified claspers and relatively large epi- didymides (compared with juveniles). Statistical analyses Size-frequency distributions of males and fe- males were compared by using a two-sample Kol- mogorov-Smirnov test (P=0.05). Chi-squared goodness-of-fit tests were used to examine the hypothesis of an equal sex ratio between num- bers of juvenile and adult C. acronotus sampled and among embryos in gravid females. To help define time of parturition, analysis of variance (ANOVA) was used to investigate the hypoth- esis of an increase in TL of embryos from near- term females captured in November and December. To provide a balanced analysis, three gravid females captured on 10 December 1998 were analyzed with three individu- als captured almost one month earlier on 11 November 1998. Data were assessed for normality by using the Sha- piro and Wilk procedure (Zar, 1996), tested for heterosce- dasticity with Cochran's test, and then analyzed in the appropriate one-nested factor ANOVA (Undei-wood, 1981). Gravid females were considered a random-effects factor nested in months and the four embryos for each gravid female were the replicates. Results The Kolmogorov-Smirnov test detected significant differ- ences in size-frequency compositions between males and females; proportionally more larger-size females ( 121-131 cm TL) were captured. Both sexes showed two distinct cohorts. The first consisted of juveniles (46-65 cm) cap- tured by gill nets at depths between 5 and 10 m, and I "1 A A -D 30- A 033 D 1 "- ■■' o°^^ 3 \ To *^ P 1 5- > A ^ D ° 10- o • : £ 05- ^M. ^^ > 0- • ^%^te 30- 25- B ^^ Juvenile n - 21 ^ ^1 Pre-ovulatory n = & J^ Ovulating n ^ 4 I 20- Q Gravid n = 21 D ,5- 1 1 Postpartum n = 8 ^ A S ,0- 05- - • 16- — 14 - C 0° O B 12- r. 9fi^ ^ 10- „^^ % B 8- 3 o O 6- £ 4- 1^°^ ■o 5 2- . „i.^ ^^A A - - # >— 45 55 65 75 85 95 105 115 125 13! Total length (cm) Figure 1 Rt'lationshi p between total length and (A) width of oviducal gland. (B)MOFD maximum ovarian follicle diameterl. and (C) width of uterus. the second included larger specimens (i.e. 85-131 cm), and mostly adults, captured on bottom longlines in deeper water ( 10 to 60 m). The ratio of juvenile males to females was not significantly different from 1:1 (X"=0.027,P>0.011, whereas significantly fewer adult males than females were sampled (ratio of 1:2.34) (x-=14.08.P<0. 01). Female maturation and reproduction Juveniles ranged in size from 46 to 101 cm TL and had narrow oviducal glands, light ovaries with undeveloped white follicles, and thin uteri (Fig. 1, A-C, Table 1). Pre- ovulatory specimens ranged in size from 103 to 129.5 cm TL, had well-developed oviducal glands (width of between 1.9 and 3 cm), and mature ovaries with thick orange fol- hcles (1.5 to 2.8 cm in diameter) (Fig. 1, A and B, Table 1). The shortest pre-ovulatoi\v specimens (e.g. 103 and 106 cm TL), had oviducal glands that were substantially wider (i.e. >2x) than those in the longest juveniles ( 101 cm TL) (Fig. lA, Table 1); therefore sexual maturity probably was approached within this size range. This should be con- NOTE Hazin et a\. Reproduction of Caicharinus acronotus 145 Table t Characti'ristics of fi'inale C. acronotus in each matura ion stage and the total 1 -■ngth (TL) range and number ol specimens. All widths and lengths are in cm , weights in g. MOFD = maximum ovarian follicle diameter. Characteristic Juvenile Pre-ovulatory Ovulating Gravid Postpartum Resting Width of oviducal gland <1.1 1.9- -3.0 2.3-3.5 1.7- -3.0 1.5- -3.0 1.4-2.6 Weight of oviducal gland 2.2-3.0 5.7- -8.7 8.6-10.0 2.3- -5.0 3.0- -5.0 3.0-7.7 Width of uterus <0.7 1.9-4.7 2.6-3.9 6.3 -16 4.2- -5.6 1.9-3.5 MOFD 0.1-0.4 1.5- -2.8 2.8-3.0 0.2- -1.0 0.2- -1.2 0.4-1.4 Width of ovary 0.4-1.7 2..3- -6.0 4.9-5.7 1.5- -3.8 2.2- -3.5 1.7-4.0 Weight of ovary 2.7-4.6 12.2- -18.2 20.0-24.6 4.5- -16.5 12.0- -23.0 4.97-16.5 TL of specimens 46.0-101.0 103.0- -129.5 108.5-123.8 114.0- -130.0 123.0- -131.0 107.0-130.0 Number of specimens 21 8 4 21 8 17 sidered a preliminary estimate of size at sexual maturity because few individuals were caught between 95 and 105 cm TL and none between 75 and 95 cm TL. Compared with pre-ovulatory females, those in the process of ovulation had slightly wider oviducal glands and heavier ovaries with thicker follicles (2.8-3 cm in diameter) (Fig. 1, A and B, Table 1 ). Ovulation appeared to occur only when MOFD was at least 2.8 cm (Fig IB). The majority of gravid, post- partum, and resting females ranged in size from 111 to 131 cm TL and had similar-size oviducal glands (Fig. lA, Table 1 ). Gravid females had undeveloped ovaries with no vitellogenic activity (MOFD was less than 1 cm). Uteri in postpartum females were flaccid and slightly distended (4.2-5.6 cm in width), whereas uteri of resting females were comparable to pre-ovulatory and ovulating individu- als (Fig. IC, Table 1). No mating scars were observed on any of the females. There was no evidence of spermatozoa stored in the oviducal glands of 10 females examined. The temporal abundance of females according to stages of reproduction showed that juveniles were present in catches between January and November, but mainly from February to June (Fig. 2A). Pre-ovulatory and ovulating individuals were sampled between February and April and April and May, respectively (Fig. 2A). Except for one gravid female caught during August, all gravid, postpar- tum, and resting females were caught between November and January (Fig. 2A). MOFD of mature females was low- est in August and between November and January, after which it steadily increased to May (Fig. 2B). Examination of the uterine contents of the 22 gravid fe- males (Table 2) revealed that individual litter size was al- ways four with both sexes present in varying proportions, although the total pooled ratio of males to females (1:1.25) was not significantly different from 1:1 (x^=1.13, P>0.05). All embryos from individual gravid females were at simi- lar stages of development and showed small variation in TL. ANOVA detected significant differences in mean TLs of embryos among near-term females (F=20.51, P<0.01) and for the main effect of months (F=10.45, P<0.05). ANOVA showed that the mean (±SE) TL of embryos in three near- term females caught on 10 December 1998 (45.96 ±0.38 cm) was significantly longer than in three near-term females Table 2 Date of capt ure and total length (TLl of gravid C. acrono- tus and the mean TL(±E) and sex ratio of embryos. Date of TL of gravid Embryos Mean TL Males: capture specimen (±SE) "emales 17 Aug 94 118.6 29.15(1.23) 1:3 1 1 Nov 98 121 42.50(0.14) 1:3 11 Nov 98 122 45.32 (0.60) 2:2 1 1 Nov 98 123 40.75(0.25) 3:1 1 1 Nov 98 125 43.25(0.43) 3:1 11 Nov 98 129 45.12(0.43) 3:1 13 Nov 98 130 38.50 (2.06) 2:2 17 Nov 98 114 34.27 (0.19) 3:1 17 Nov 98 127.7 42.50(0.15) 3:1 21 Nov 98 115.2 41.37(0.37) 3:1 22 Nov 98 127.2 45.47 (0.62) 2:2 25 Nov 98 126 43.87(2.01) 2:2 25 Nov 98 127.5 46.07 (0.22) 1:3 6 Dec 98 117 44.75(0.14) 2:2 6 Dec 98 123 43.37(0.85) 2:2 6 Dec 98 123 48.50 (0.28) 2:2 7 Dec 98 125 46.87(0.12) 2:2 10 Dec 98 124 45.62 (0.37) 2:2 10 Dec 98 124 47.50(0.28) 3:1 10 Dec 98 122 44.75(0.25) 2:2 12 Dec 98 123 42.50(0.35) 3:1 5 Jan 95 114,5 45.37 10.24) 2:2 caught on 11 November 1998 (42.31 ±0.07 cm), suggesting that embryos continued to develop between these periods. Male maturation and reproduction Of the 45 males examined, 23 were juvenile with thin epi- didymides (Fig. 3A) and filiform ampullae of the ductus deferens without seminal fluid. Juveniles ranging from 49 146 Fishery Bulletin 100(1) to 63 cm TL had relatively short flexible claspers (Fig. 3B) and testes that were not fully differentiated from the epigonal organ. No males between 64 and 87 cm TL were caught. Three juveniles (88, 93. and 94 cm TL) had claspers that were enlarged and beginning to calcify (Fig. 3B) and two larger specimens also had thick epididymi- des (Fig. 3A). Males appear to approach sexual maturity before 104 cm TL because all specimens longer than this had completely developed sexual organs, including elon- gate and calcified claspers, thick epididymides, and cir- cumvoluted ampullae of the ductus deferens (Fig. 3). Discussion Prevalence of adult females in catches was similar to that from observations made by Schwartz (1984) for stocks off the southeastern United States and can be attributed to a reproductive migration involving relatively large num- 24 22 20 18 16- 14- 12 ^■Juvenile Ig^ Pre-ovulalory ^Sovulaling ^Gravid ^Postpartum ^^ Resting 3.0- Q O 00 B Jan Feb Mat Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 2 Monthly distribution of (A) females in catches according to their stage of reproduction and (B) MOFDs (maximum ovarian follicle diameters) for mature females. hers of gravid individuals (114 to 130 cm TL) into the sampled area. Because there was no evidence of disequi- librium between sexes (i.e. there were equal numbers of male and female embryos and juveniles), large numbers of adult males were probably segi'egated. It is unlikely males were in the sampled area and not caught because females, although proportionally larger, were captured across the same size range (Fig. 1), implying that the selectivity of the gear encompassed the range of sizes of males. Evaluation of the maturation stages of males and fe- males showed delineation between juveniles and adults. All females longer than 103 cm TL had enlarged oviducal glands and developed ovaries (Fig. 1, A and B, Table 1), and males longer than 104 cm TL had elongate and cal- cified claspers and thick epididymides (Fig. 3), indicating that sexual maturity was probably approached at these lengths. Although these estimates were derived from few individuals, they are comparable to those proposed by most researchers for specimens collected off the southeast- ern United States (e.g. Springer, 1938; Clark and von Schmidt, 1965; Compagno, 1984), with the ex- ceptions of Branstetter ( 1981) who suggested 110 cm TL for both sexes and Schwartz ( 1984 ) who suggested 110 cm TL for males. Size at birth, number of embryos, sex ratio of embryos, and the time of parturition of C. acrono- tus in our study are consistent with correspond- ing data from earlier works (Springer, 1938; Clark and von Schmidt, 1964; Dodrill (1977); Brans- tetter, 1981; Compagno, 1984; Schwartz, 1984). Size at birth has been suggested to be between 45 and 50 cm TL (e.g. Branstetter, 1981; Castro, 1993) and litter sizes commonly range from 3 to 6 (Springer, 1938; Dodrill, 1977; Compagno, 1984). We observed gravid females caught in November and December (late spring to early summer) with embryos longer than 45 cm TL (Table 2). Given the significant increase in size of embryos between these months (indicating that embryos were still developing) and the capture of several neonates (46-51 cm TL) in February and March (late sum- mer to early autumn), we conclude that parturi- tion off northeastern Brazil probably occurs from December to January (mid to late summer). A sim- ilar seasonal timing has been proposed for stocks off the southeastern LInited States (e.g. during June — Schwartz, 1984) and is generally typical for the majority of carcharhinids (e.g. Castro, 1993). In contrast to the inference made by Branstet- ter (1981) but in agi'eement with obsei-vations of Schwartz (1984), we showed that vitellogenisis and gestation occur consecutively in C. acronotus. Ovaries of adult females off northeastern Brazil begin to mature (pre-ovulatory stage) in Febru- ary (late summer) with ovulation occurring two to three months later (Fig. 2, A and B). This se- quence of events is illustrated by a rapid increase in MOFD from February (1.5-2 cm in diameter) to May (3 cm in diameter) (Fig. 2B). Given the proposed summer parturition (December to Janu- NOTE Hazin et a\: Reproduction of Carchannus aaonotus 147 arv>, mating and fertilization during April and May (autumn ) would result in a gestation peri- od of approximately 8 months — slightly short- er than that suggested by Schwartz ( 1984) and Dodrill 1 1977) (9 and 10 to 11 months, respec- tively) for stocks off the southeastern United States. Further, the periods required for vi- tellogenisis and gestation indicate that repro- duction of C acronotiis off northeastern Brazil could be completed within 10 to 11 months. With the simultaneous capture of nongravid and gi-avid females off Florida, Dodrill (1977) proposed a biennial reproductive cycle for C. ac- ronotus. Although we also collected gravid and nongi-avid (i.e. postpartum and resting) females together (i.e. mostly during December and Jan- uary), the latter individuals could have given birth 3 or 4 weeks earlier If these females sub- sequently ovulated 2 months later (in April), then reproduction could conceivably be annu- al. A fast vitellogenic period, combined with clear reproductive progress, is also supportive of a 1-year cycle. An alternative hypothesis that supports biennial reproduction is that the rest- ing females represented some proportion of seg- regated nongravid females that moved with the relatively large numbers of gravid females into the parturition area. Given the lack of adult fe- males in catches between June and November (winter to late spring), it is likely, however, that they frequent other areas after copulation. Without additional fishery-indepen- dent data, it is impossible to determine these locations and whether some proportion of the female population consists of nongravid or resting individuals throughout the yean Our evidence suggests a shorter reproductive cycle for C. acronotus than that previously noted in the literature. Given the 6-month difference between times of ovulation and parturition for females off northeastern Brazil and those off the southeastern United States, our results also indicate the existence of at least two separate stocks. Not- ing temporal differences in the sizes of embryos from fe- males, Schwartz (1984) proposed two partially separated populations off the southeastern United States: one off North Carolina and the other comprising individuals from Florida and the Gulf of Mexico. Additional research is needed to determine if the population in the southwest- ern Atlantic is separate from those in the north because the existence of a unit stock off northeastern Brazil would require separate management measures according to the status of that stock. Acknowledgments This study was funded by the Comissao Interministerial para os Recursos do Mar (CIRM) through the Programa Nacional de Avalia^ao dos Recursos Vivos da Zona Economica Exclusiva ( RE VIZEE ), the Fundagao de Amparo a Ciencia e Tecnologia do Estado de Pemambuco ( FACEPE ), and the Conselho Nacional de Ensino e Pesquisa (CNPq). E 25 o S 20 ■D 1 .5 ■o TD Q- 10 0) ° 05 ■g 5 00- 16- ^ 12- ^, '1' where t = age; / . = asymptotic length; /?,,/;., = instantaneous growth rate coefficients; and f J, t., = age intercept parameters. The second, dubbed the "linear" von Bertalanffy curve (Hoese et al., 1991; Vaughan, 1996), expresses the asymp- totic length as a linear function of age: /, =(f)fi +bit){l-e -kit-t„). (2) Of course other generalizations of the von Bertalanffy cui-ve may also be appropriate, such as the Richards (1959) equation /, =La-de- ')' where 5 5^0. (3) tp=ik.,t.,-k,ti)Hk.2-k,), The double von Bertalanffy curve ac- commodates the possibility that older, larger fish might grow more slowly in proportion to their length than young- er, smaller fish. (The linear von Ber- talanffy curve has no biological inter- pretation.) In reality, one might expect the gi'owth rate in proportion to length to decrease gradually with the age of the fish rather than at some abrupt piv- otal point. Moreover, the growth pattern of juvenile red drum seems to have a : ^ + /i,e"^'' + k^e'^'' sin(2;r(f (5) where A, and A., are damping coeffi- cients and t^ is a shifting parameter for the sine wave valued between and 1. Substituting Equations 5 and 4 and integrating with / = when ^ = ?o gives ' Condrey, R., D. W. Beclcman and C. A. Wilson. 1988. Management implications of a new growth model for red drum. Appen- dix D. In Louisiana red drum research, J. A. Shepard (ed.), 26 p. U.S. Dept. Commerce Cooperative Agi'eement NA87- WC-H-06122. Marine Fisheries Initiative IMARFIN) Program. Louisiana Depart- ment of Wildlife and Fisheries, Seafood Division, Finfish Section. Baton Rouge, Louisiana 70803-7.503. - Goodyear, C. P. 1996. Status of the red drum stocks of the Gulf of Mexico: report for 1996. Rep. MIA 95/96-47, 21 p. Miami Lahoratorv, .Southeast Fish. Sci. Cent., Natl. Mar" Fish. Serv., NOAA. 75 Virginia Beach Dr, Miami, Fl. 33149. Manuscript accepted 17 May 2001. Fish. Bull. 100:149-152 (2002). 150 Fishery Bulletin 100(1) ^1 Po fc 4;r-+(A,)' 2KCOs{2!t{t^.'t)}-' A.,sin{2nit^ -t)] '2;rcos{2;r(/, -^o)}-' A._,sin{2;r(*,, -<„)} (6) Assuming the animal will not shrink with age, i.e., dl I dt >0, implies the constraint /fo + V"^''+ V~ 'sin(2;r(?-r))>0. (7) Equation 6 may appear formidable, but typically re- quires only a minute or two more to enter into standard statistical fitting packages. It reduces to a form similar to the Gompertz equation when k., = and to the von Berta- lanffy equation when ^j = ^2=0- Fitting the model to data Equations 1, 2, 3, and 6 were fitted to observations of length-at-age from red drum collected in the northern Gulf of Mexico between September 1985 and October 1998 (see Beckman et al.. 1988. or Wilson et al.'^ for further details regarding the data collection and aging procedures). The fitting was accomplished by ordinary least squares by using a Nelder-Mead simplex search^ and, as a check, proc NLIN of SAS ( 1990). The least-squares solution is equiva- lent to the maximum likelihood solution when the distri- bution of length at age is normal with constant variance, which seems to be approximately true of this particular data set (Porch, unpubl. data). Akaike's (1973) information criterion (AIC) was used to rank the gi-owth models in terms of their ability to provide statistically parsimonious explanations of the data. The formula for the AIC may be written A/C=-21og(L) + 2p, where L is the likelihood function and p is the number of parameters (see Buckland et al., 1997). In this case, -21og(L) is equal to the residual sums of squares. Wilson, C. A., D. L. Nieland and A. L. Stanley. 1993. Varia- tion of year-class strength and annual reproductive output of red drum Sciaenops ocellatus and black drum Pogonias cromis from the northern Gulf of Mexico. Final Report 1991-1992. 31 p. U.S. Dept. Commerce Cooperative Agi-eement NA90AA- H-MF724. Marine Fisheries Initiative (MARFINi Program. Coastal Fisheries Institute. Louisiana State University, Baton Rouge, La 70803-7503. Shaw, D. E., R. W. M. Wedderburn, and A. Miller 1991. A Program for function minimization using the simplex method. CSIRO, Division of Mathematics and Statistics, P.O. Box 218, Lmdfield, N.S.W. 2070. Australia. Table 1 Akaike's information criteria (AIC) quantifying the fit of the various growth models to the red drum length-at-age data. Smaller AIC values indicate statistically better fits. Model Number of parameters Negative log- likelihood AIC (-25,000) von Bertalanffy 3 16045.9 7098 Richards 4 14169.8 3348 linear von Bertalanffy 4 12883.8 776 double von Bertalanffy 5 12876.6 763 damped (Eq. 6, ^^=0) 5 12651.3 313 seasonal + damped (Eq. 6) 8 12584.9 186 The parameter estimates for the Richards equation tended to be unstable unless good initial estimates were provided. This was accomplished by conducting the esti- mation in two stages. In the first stage the exponent 5 was fixed to 1 and the other parameters were estimated, reduc- ing the Richards equation to the von Bertalanffy form. In the second stage the initial guesses were set equal to the final estimates from the first stage and then all four pa- rameters were estimated simultaneously. Results and discussion All five alternative growth models fitted the data signif- icantly better than the von Bertalanffy equation accord- ing to the AIC statistic (Table 1). The Richards equation, however, did not fit the data nearly as well as the other alternative formulations and suffered from well-known instability problems (Ratkowsky, 1983), therefore it can probably be dropped from any future consideration with respect to red drum. The double von Bertalanffy cui-ve fitted the data better than the linear von Bertalanffy curve, but the comparison is rendered moot by the perfor- mance of the new model. The five-parameter version with- out seasonal oscillations (Eq. 6 with /?.,=0) fitted the data significantly better than either The eight-parameter ver- sion with seasonal oscillations fitted the data significantly better still (see Fig. 1). The estimated seasonal component to the growth rate was fairly substantial initially, having an amplitude at age of 0.301 (k.-,'> and a peak in June, but declined rap- idly with age (Fig. 2). It is possible that an even stronger seasonal signal would have been estimated if age-0 fish, which exhibit the strongest seasonal pattern (Goodyear*), had been adequately represented in the sample. Some of the parameter estimates were highly correlat- ed, as is the case in most growth studies. In particular, the correlations between the estimates for the growth rate and asymptotic length coefficients were typically above 0.8. However, the asymptotic variance-covariance matrix NOTE Porch el a\ A growth model for Sciaenops occllatus 151 ou - A •■'■'>' 40 iiiiMW mm^ 30 iW t||fn?p'r- 20 10- n —> — 1 — — I — 1 — 1 — 1 — 1 — 1— H 1 1 1 1 1 1 10 15 20 25 30 35 40 35 J 30 - 25 20 - 15 -- 10 - 5 B 12 3 4 5 Age Figure 1 Fit of the proposed seasonal growth model to red drum length at age data: (A) the fit for all ages; iBi the fit for the younger ages. g o 2 3 4 5 6 Age Figure 2 Growth rate coefficient from seasonal model as a function of age. suggests that the parameters for all of the models were estimated fairly precisely (Table 2). The new model, either with or without the seasonal component, has both practical and theoretical advantages over the four other models examined in this study. By vir- tue of its greater flexibility, it was able to fit the red drum Table 2 Parameter estimates and correspond ing coefficient h 0.412 1 h 0.0530 23 K 0.114 2 h -8.41 3 damped (Eq. 6. A-.=Oi /, 44.1 1 *0 0.0416 7 to 0.362 3 K 0.667 2 A, 0.464 1 seasonal and damped /,, 43.4 1 Ao 0.0475 5 «o 0.443 3 K 0.695 2 K 0.476 1 K 0.301 16 h 0.344 22 t,-' 0.439 3 data significantly better than the linear and double von Bertalanffy curves (its nearest competitors). Moreover, the Richards and linear von Bertalanffy curves are theoret- ically disadvantaged because their parameters have no physical interpretation. The double von Bertalanffy curve, although it has a physical interpretation, suffers because it allows only a single discontinuous change in the growth rate at one age rather than a continuous change through time. For these reasons, the new model should be more widely applicable than the others, particularly for species that change habitat preferences with age or are subject to strong seasonal environmental fluctuations. Acknowledgments We thank C. Legault, G. Scott, S. Turner, D. Vaughan, and an anonymous reviewer for their helpful comments. Literature cited Akaike, H. 1973. Information theory and an extension of the maximum likelihood principle. In Second international symposium 152 Fishery Bulletin 100(1) on information theory (B. N. Petrov and F. Csaki. eds.), p. 267-281. Akademiai Kiado, Budapest. Beckman, D. W., C. A. Wilson, and A. L. Stanley. 1988. Age and growth of red drum, Sciaenops ocellatus from offshore waters of the Gulf of Mexico. Fish. Bull. 87: 17-28. Buckland, S. T. K. P. Burnham. and N. H. Augustm. 1997. Model selection: an integral part of niference. Bio- metrics 53:603-618. Hoese, H. D., D. W. Beckman, R. H Blanchet, D. Drullinger, and D. L. Nieland. 1991. A biological and fisheries profile of Louisiana red drum Sciaenops ocellatus. Fishery management plan series, number 4, part 1. Louisiana Department of Wildlife and Fisheries, Baton Rouge, LA, 93 p. Murphy, M. D., and R. G. Taylor 1990. Reproduction, growth, and mortality of red di-um, Sciae- nops ocellatus. in Florida waters. Fish. Bull. 88:.531-.542. Ratkowsky, D. A. 1983. Nonlinear regression modeling. Marcel Dekker, New York, NY, 276 p. Richards. F J. 19.59. A flexible growth function for empirical use. J. Exp. Botany 10:290-.300, SAS. 1990. SAS/STATu.sers guide, vol. 2, version 6, 4th ed. SAS Institute Inc., Gary NG, 1686 p. Vaughan, D. S. 1996. Status of the red drum stock on the Atlantic Goast: stock assessment report for 1995. U.S. Dep. Commer. NCAA Tech. Memo. NMFSF-SEFC-380. 50 p. Vaughan, D. S., and T. E. Reiser. 1990. Status of the red drum stock of the Atlantic Goast: stock assessment report for 1989. U.S. Dep. Gommer., NCAA Tech. Memo. NMFS-SEFC-263, 53 p. Fishery Bulletin 100(1) 153 Superintendent of Documents Publications Order Form *5178 I I Yli/O, please send nie the following publ ications: Subscriptions to Fishery Bulletin for $55.00 per year ($68.75 foreign) The total cost of my order is $ . Prices include regular domestic postage and handling and are subject to change. (Company or Personal Name) (Please type or print) (Additional address/attention line) (Street address) (City, State, ZIP Code) (Daytime phone including area code) (Purchase Order No. MastarCaid Please Choose Method of Payment: I I Check Payable to the Superintendent of Documents I I GPO Deposit Account I I VISA or MasterCard Account (Credit card expiration date) -D To fax your orders (202) 512-2250 (Authorizing Signature) Mail To: Superintendent of Documents RO. Box 371954, Pittsburgh, PA 15250-7954 Thank you for your order! Fishery Bulletin Guidelines for contributors Content of papers Articles Articles are reports of 10 to 30 pages (double spaced) that describe original research in one or a combination of the following fields of marine science: taxonomy. hiolog>'. genetics, mathematics (including modeling), statistics, engineering, eco- nomics, and ecology. Notes Notes are reports of 5 to 10 pages without an abstract that describe methods and results not supported by a large body of data. Although all contributions are subject to peer review, responsi- bility for the contents of articles and notes rests upon the authors and not upon the editor or the publisher. It is therefore important that authors consider the contents of their manuscripts care- fully. Submission of an article is un-derstood to imply that the article is original and is not being considered for publication elsewhere. Manuscripts must be written in English. Authors whose native language is not English are strongly advised to have their manuscripts checked for fluency by English-speaking colleagues prior to submission. Preparation of papers Text Title page should include authors' full names and mailing addresses (street address required) and the senior author's telephone, fax number, e-mail address, as well as a list of key words to describe the contents of the manuscript. Abstract must be less than one typed page (double spaced) and must not contain any citations^ It should state the main scope of the research but emphasize the author's con- clusions and relevant findings. Because abstracts are oirculated by abstracting agencies, it is impor- tant that they represent the research clearly and concisely. General text must be typed in double- spaced format. A brief introduction should state the broad significance of the paper; the remainder of the paper should be divided into the following sec- tions: Materials and methods. Results, Discussion (or Conclusions), and Acknowledgments. Headings within each section must be short, reflect a logical sequence, and follow the rules of multiple subdi- vision (i.e. there can be no subdivision without at least two subheadings). The entire text should be intelligible to interdisciplinary readers; therefore, all acronyms and abbreviations should be written out and all lesser-known technical terms should be defined the first time they are mentioned. The scientific names of species must be written out the first time they are mentioned; subsequent mention of scientific names may be abbreviated. Follow Sci- entific style and format: CBE manual for authors, editors, and publishers 1 6th ed.) for editorial style and the most current issue of the American Fish- eries Society's common and scientific names of fishes from the United States and Canada for fish nomenclature. Dates should be written as fol- lows: 11 November 1991. Measurements should be expressed in metric units, e.g. metric tons (t). The numeral one ( 1 ) should be typed as a one, not as a lower-case el ( I ). Footnotes Use footnotes to add editorial comments regarding claims made in the text and to document unpub- lished works or works with local circulation. Foot- notes should be numbered with Arabic numerals and inserted in 10-point font at the bottom of the first page on which they are cfted. Footnotes should be formatted in the same manner as citations. If a manuscript is unpublished, in the process of review, or if the information provided in the footnote has been conveyed verbally, please state this information as "unpubl. data," "manuscript in review," and "personal commun.," respectively. Authors are advised wherever possible to avoid ref- erences to nonstandard literature (unpublished lit- erature that is difficult to obtain, such as internal reports, processed reports, administrative reports, ICES council minutes, IWC minutes or working papers, any "research" or "working" documents, laboratery reports, contract reports, and manu- scripts in review). If these references are used, please indicate whether they are available from NTIS (National Technical Information Service) or from some other public depository. Footnote format: author (last name, followed by first-name initials); year: title of report or manuscript; type of report and its administrative or serial number; name and address of agency or institution where the report is filed. Literature cited The literature cited section comprises works that have been published and those accepted for pub- lication (works in press) in peer-reviewed jour- nals and books. Follow the name and year system for citation format. In the text, write "Smith and Jones (1977) reported" but if the citation takes the form of parenthetical matter, write "(Smith and Jones, 1977)." In the literature cited section, list citations alphabetically by last name of senior author: For example, Alston, 1952; Mannly, 1988; Smith, 1932; Smith, 1947; Stahnsky and Jones. 1985. Abbreviations of journals should conform to the abbreviations given in the Serial sources for the BIOSIS previews database. Authors are responsible for the accuracy and completeness of all citations. Literature citation forrnat: author (last name, followed by first-name initials); year: title of report or article; abbreviated title of the journal in which the article was published, volume number, page numbers. For books, please provide publisher, city, and state. Tables Tables should not be excessive in size and must be cited in numerical order in the text. Headings in tables should be short but ample enough to allow the table tOibe intelligible on its own. All unusual symbols must be explained in the table legend. Other incidental comments may be footnoted (use italic arable numerals for footnote markers). Use asterisks only to indicate probability in statistical data. Place table legends on the same page as the table data. We accept tables saved in most spread- sheet software programs (e.g. Microsoft Excel). Please note the following; • Use a comma in numbers of five digits or more (e.g. 13,000 but 3000). • Use zeros before all decimal points for values less than one (e.g. 0.31). Figures Figures include line illustrations, computer-gener- ated line graphs, and photographs (or slides). They must be cited in numerical order in the text. Line illustrations are best submitted as original draw- ings. Computer-generated line graphs should be printed on laser-quality paper Photographs should be submitted on glossy paper with good contrast. All figures are to be labeled with senior author's name and the number of the figure (e.g. Smith, Fig. 4). Use Helvetica or Arial font to label anatomical parts (line drawings) or variables (graphs) within figures; use Times Roman bold font to label the dif- ferent sections of a figure (e.g. A, B, C). Figure leg- ends should explain all symbols and abbreviations seen within the figure and should be typed in dou- ble-spaced format on a separate page at the end of the manuscript. We advise authors to peruse a recent issue of Fishery Bulletin for standard for- mats. Please note the following: •Capitalize the first letter of the first word of axis labels. • Do not use overly large font sizes to label axes or parts within figures. • Do not use boldface fonts within figures. • Do not create outline rules around graphs. •Do not use horizontal lines through graphs. • Do not use large font sizes to label degrees of longitude and latitude on maps. • Indicate direction of .degrees longitude and latitude on maps (e.g. 170°E). • Avoid placing labels on a vertical plane (except ony axis). •Avoid odd (nonstandard) patterns to mark sections of bar graphs and pie charts. Copyright law Fishery Bulletin, a U.S. government publication, is not subject to copyright law. If an author wishes to reproduce any part oiFishery Bulletin in his or her work, he or she is obliged, however, to acknowledge the source of the extracted literature. * Submission of papers Send four printed copies (one original plus three copies) — clipped, not stapled — to the Scientific Edi- tor, at the address shown below. Send photocopies of figures with initial submission of manuscript. Original figures will be requested later when the manuscript has been accepted for publication. Do not send your manuscript on diskette until requested to do so. Dr John V. Merriner Scientific Editor, Fishery Bulletin Center for Coastal Fisheries and Habitat Research, NOS 101 Pivers Island Road Beaufort, NC 28516 Once the manuscript has been accepted for pub- licatioUj you will be asked to submit a software copy of your manuscript. The software copy should be submitted in WordPerfect or Word format (in Word, save as Rich Text Format). Please note that we do not accept ASCII text files. Reprints Copies of published articles and notes are avail- able free of charge to the senior author (50 copies) and to his or her laboratory (.50 copies i. Additional copies may be purchased in lots of 100 when the author receives page proofs. U.S. Department of Commerce Volume 100 Number 2 April 2002 Fishery Bulletin ?■ ?- mi U.S. Department of Commerce Donald L Evans Secretary National Oceanic and AtmospHieric Administration Vice Admiral Conrad C. Lautenbacher Jr., USN (ret.) Under Secretary for Oceans and Atmosphere National Marine Fisheries Service William T. Hogarth Acting Assistant Administrator for Fisfieries The Fishery Bulletin (ISSN 0090-06561 is published quarterly by the Scientific Publications Office, National Marine Fish- eries Service, NOAA, 7600 Sand Point Way NE, BIN C 15700, Seattle, WA 981 15-0070. Periodicals postage is paid at Seattle, WA, and at additional mailing offices. POST- MASTER: Send address changes for sub- scriptions to Fishery Bulletin, Superin- tendent of Documents, Attn.: Chief Mail List Branch, Mail Stop SSOM, Washing- ton, DC 20402-9373. . Although the contents of this' publica- tion have not been copyrighted and may be reprinted entirely, reference to source Js appreciated. f' The Secretary of Commerce has deter- mined that the publication of this peri- odical is necessary according to law for the transaction of public business of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per -year: $45.00 domestic and $56.25 foreign. Cost per single issue: $28.00 domestic and $35.00 foreign. See back for order form. Scientific Editor Dr. John V. Merriner Editorial Assistant Sarah Shoffler . Center for Coastal Fisheries and Habitat Research, NOS 101 Pivers Island Road Beaufort, NC 28516 Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific, Publications Office 7600 Sand Point Way NE, BIN C15700 Seattle, Washington 98115-0070 Editorial Committee Dr. Andrew E. Dizon Dr. Harlyn O. Halvorson Dr. Ronald W. Hardy Dr. Richard D. Methot Dr. Theodore W. Pietsch Dr. Joseph E. Powers Dr Harald Rosenthal Dr. Fredric M. Serchuk National Marine Fisheries Service University of Massachusetts, Boston University of Idaho, Hagerman National Marine Fisheries Service University of Washington, Seattle National Marine Fisheries Sen/ice Universitat Kiel, Germany National Marine Fisheries Service Fishery Bulletin web site: fishbull.noaa.gov The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. It began as the Bulletin of the United States Fish Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fisher^' Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions. State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 100 Number 2 April 2002 Fishery Bulletin Contents The conclusions and opinions expressed in Fishery Bulletin are solely those of the authors and do not represent the official position of the National Marine Fisher- ies Service INOAA) or any other agency or institution. The National Marine Fisheries Service (NMFSi does not approve, recommend, or endorse any proprietary product or pro- prietary material mentioned in this pub- hcation. No reference shall be made to NMFS. or to this publication furnished by NMFS. in any advertising or sales pro- motion which would indicate or imply that NMFS approves, recommends, or endorses any prnpnetary product or pro- prietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. Articles 155-167 Brill, Richard, Molly Lutcavage, Greg Metzger, Peter Bushnell, Michael Arendt, Jon Lucy, Cheryl Watson, and David Foley Horizontal and vertical movements of juvenile bluefin tuna (Thunnus thynnus) in relation to oceanographic conditions of the western North Atlantic, determined with ultrasonic telemetry 168-180 Chase, Bradford C. Differences in diet of Atlantic bluefin tuna (Thunnus thynnus) at five seasonal feeding grounds on the New England continental shelf 181-192 Comeau, Michel, and Fernand Savoie Movement of American lobster (Homarus americanus) in the southwestern Gulf of St. Lawrence 193-199 Else, Page, Lewis Haldorson, and Kenneth Krieger Shortspine thornyhead (Sebastolobus alascanus) abundance and habitat associations in the Gulf of Alaska 200-213 Hatfield, Emma M.C., and Steven X. Cadrin Geographic and temporal patterns in size and maturity of the longfin inshore squid (Loligo pealeii) off the northeastern United States 214-227 Hesp, Sybrand A., Ian C. Potter, and Norman G. Hall Age and size composition, growth rate, reproductive biology, and habitats of the West Australian dhufish (.Glaucosoma hebraicum) and their relevance to the management of this species 228-243 Kitada, Shuichi, and Kiyoshi Tezuka Longitudinal logbook survey designs for estimating recreational fishery catch, with application to ayu (Plecoglossus altivelis) 244-257 MacFarlane, R. Bruce, and Elizabeth C. Norton Physiological ecology of juvenile chinook salmon (Oncorhynchus tshawytscha) at the southern end of their distribution, the San Francisco Estuary and Gulf of the Farallones, California Fishery Bulletin 100(2) 258-265 McFee, Wayne E., and Sally R. Hopkins-Murphy Bottlenose dolphin (.Tursiops truncatus) strandings in South Carolina, 1992-1996 266-278 Natanson, Lisa J., Joseph J. Mello, and Steven E. Campana Validated age and growth of the porbeagle shark (Lamna nasus) in the western North Atlantic Ocean 279-298 Olson, Robert J., and Felipe Galvan-Magana Food habits and consumption rates of common dolphinfish (Coryphaena hippurus) in the eastern Pacific Ocean 299-306 Powles, Perce M., and Stanley M. Warlen Recruitment season, size, and age of young American eels (Anguilla rostrata) entering an estuary near Beaufort, North Carolina 307-323 Shima, Michiyo, Anne Babcock Hollowed, and Glenn R. VanBlaricom Changes over time in the spatial distribution of walleye pollock (Theragra chalcogramma) in the Gulf of Alaska, 1984-1996 324-337 Starr, Richard M., John N. Heine, Jason M. Felton, and Gregor M. Cailliet Movements of bocaccio (Sebastes paucispinis) and greenspotted (5. chlorostictus) rockfishes in a Monterey submarine canyon: implications for the design of marine reserves 338-350 Steele Philip, Theresa M. Bert, Kristine H. Johnston, and Sandra Levett Efficiency of bycatch reduction devices in small otter trawls used in the Florida shrimp fishery 351-375 Vaughan, Douglas S., and Michael H. Prager Severe decline in abundance of the red porgy (Pagrus pagrus) population off the southeastern United States Notes 376-380 Abookire, Alisa A., John F. Piatt, and Suzann G. Speckman A nearsurface, daytime occurrence of two mesopelagic fish species (Stenobrachius leucopsarus and Leuroglossus schmidti) in a glacial fjord 381-385 Link, Jason S., and Frank P. Almeida Opportunistic feeding of longhorn sculpin (Myoxocephalus octodecemspinosus): Are scallop fishery discards an important food subsidy for scavengers on Georges Bank? 386-389 Romanov, Evgeny V., and Veniamin V. Zamorov First record of a yellowfin tuna (Thunnus albacares) from the stomach of a longnose lancetfish (Aleplsaurus ferox) 390 Subscription form 155 Abstract— Wi' omplin'od ultrasonic trans- mittLT.s lo follow (for up to 48 h) the hor- izontal and vertical movements of five juvenile (6.8-18.7 kg estimated body mass) bluefin tuna [Thunnus thynntis^ in the western North Atlantic (off the eastern shore of Virginia ). Our objective was to document the fishes' behavior and distribution in relation to ocean- ogi'aphic conditions and thus begin to address issues that currently limit pop- ulation assessments based on aerial surveys. Estimation of the trends in adult and juvenile Atlantic bluefin tuna abundance by aerial sun'eys, and other fishery-independent measures, is con- sidered a priority. Juvenile bluefin tuna spent the major- ity of their time over the continental shelf in relatively shallow water (gen- erally less then 40 m deep). Fish used the entire water column in spite of rela- tively steep vertical thermal gradients (=24°C at the surface and =12°C at 40 m depth), but spent the majority of their time (=90'"r) above 15 m and in water warmer then 20°C. Mean swimming speeds ranged from 2.8 to 3.3 knots, and total distance covered from 152 to 289 km (82-156 nmi). Because fish gen- erally remained within relatively con- fined areas, net displacement was only 7.7-52.7 km (4.1-28.4 nmi). Horizontal movements were not correlated with sea surface temperature. We propose that it is unlikely that juvenile bluefin tuna in this area can detect minor horizontal temperature gi-adients (gen- erally less then 0.5°C/km) because of the steep vertical temperature gi-adi- ents (up to =0.6°C/m) they experience during their regular vertical move- ments. In contrast, water clarity did appear to influence behavior because the fish remained in the intermediate water mass between the turbid and phytoplankton-rich plume exiting Ches- apeake Bay ( and similar coastal waters ) and the clear oligotrophic water east of the continental shelf Horizontal and vertical movements of juvenile bluefin tuna (Thunnus thynnus), in relation to oceanographic conditions of the western North Atlantic^ determined with ultrasonic telemetry Richard Brill Pelagic Fisheries Research Program Joint Institute for Marine and Atmospheric Research School of Earth and Ocean Science Technology University ol Hawaii at Manoa Honolulu, Hawaii 96822 E mail address rbnllia'honlab nmfs hawaii.edu. Molly Lutcavage Edgeiton Research Laboratory New England Aquanum Boston, Massachusetts 02110-3399 Greg Metzger Department of Biology Southampton College Long Island University Southampton, New York 11968-4198 Peter Bushnell Department of Biological Science Indiana University South Bend South Bend, Indiana 46634-7111 Michael Arendt Jon Lucy Sea Grant Program Virginia Institute ol Marine Science College of William and Mary Gloucester Point Virginia 23062-1346 Cheryl Watson Department of Biological Sciences Central Connecticut State University New Britain, Connecticut 06053-2490 David Foley NOAA CoastWatch Program, Hawaii Regional Node Pelagic Fisheries Research Program Joint Institute for Marine and Atmospheric Research School of Earth and Ocean Science Technology University of Hawaii at Manoa Honolulu, Hawaii 96822 Manuscript accepted 6 Julv 2001, Fish. Bull. 100:1,5.5-167 12002). Current estimates of spawning biomass for Atlantic bluefin tuna iThiintnis thynnus) remain controversial (Butter- worth and Punt. 1993; Restrepo et al., 1994; Restrepo, 1996), although the most conservative predicts that a pop- ulation eight times the current size would be needed to produce maximum sustainable yields (Sissenwine et al,, 1998), The current strict catch quotas are based on abundance assessments for both adult and juvenile (i.e, "school- ing") fish (age classes 1-5 years, body mass =6-60 kg). Adult abundance is derived from commercial landings data; juvenile abundance has, since 1985, been based on fishing effort and land- ings data obtained from dockside inter- cepts and telephone polling of the largely recreational fishery for juvenile bluefin tuna conducted by the National Marine Fisheries Service's Large Pelag- ics Survey (Turner et al,, 1993, 1997), The usefulness of both data sets can be compromised, however, because the relationship between catch-per-unit-of- effort (CPUE) data and real abundance is not known with certainty (Bakun et al,, 1982; Hilborn and Walters, 1992; Lauck. 1996), This problem is especially critical with highly mobile schooling fishes like tunas because of environ- mental influences on fish distribution and vulnerability to specific fishing gears, as well as the introduction of new fishing techniques (Sharp, 1978; Clark and Mangel, 1979; Brill, 1994; Bertrand and Josse, 2000). Juvenile Atlantic bluefin tuna appear in the surface waters off the east coast 156 Fishery Bulletin 100(2) 40"N United States ^^ *^ 35 30 / N J A 25 V4 A \ "^ -^^ Fish 3 & Fish 4 \5 // ?^ "^^5 ' 85 80 75 70 76 Figure 1 (A> Map of the east coast of the United States. The rectangle shows the area enlarged in panel B. (B) Movements of five juvenile bluefin tuna tThunnun thynnui^). The limit of the continental shelf is shown by the 50-, 100-, and 200-m isobath lines. The topographic features considered by local fisher- men to aggregate juvenile bluefin tuna are shown by the shaded areas. Place names are taken from local fishing charts. The rectangles show the areas enlarged in Figures 3 and 4. of the United States, from North Carohna to Rhode Island, usually during June and July (Rivas, 1978; Sakagawa, 1975; koffer, 1987; Lucy et al.," 1990; Mather et al, 1995 1. Their presence provides an opportunity for direct popu- lation assessments with aerial surveys similar to those conducted on adult Atlantic bluefin tuna (Lutcavage and Ki'aus, 1995; Lutcavage et al., 1997), southern bluefin tu- na tTluininis maccoyii).^ and other fish species (e.g. Lo et al., 1992). Assessments of juvenile bluefin tuna abundance are considered particularly crucial for effective stock man- agement because these will allow the forecasting of re- cruitment and long-term population trends iPolacheck et al., 1996; Sissenwine et al., 1998). There is, however, a need to establish the probability of detecting schools and estimating school size before aerial survey data can pro- vide robust population assessments. This need is present regardless of whether the census techniques are simple photography (Lutcavage and Kraus, 1995; Lutcavage et al., 1997) or new laser-based digital remote sensing tech- niques (Oliver et al., 1994; Lo et al., 1999). As with tra- ditional CPUE-based abundance estimates, knowledge of the effects of oceanographic conditions on depth distribu- tion, surfacing frequency, travel speeds, and residence pat- terns is critical because these conditions will affect vul- nerability to "capture," either on photographic film or as digital data. To obtain the necessary data, we undertook a study of the horizontal and vertical movements of juvenile Atlan- Cowling, A., C. Millar, and T. Polacheck. 1996. Data analysis of the aerial surveys (1991-19971 for juvenile southern bluefin tuna in the Great Australian Bight. Rep. RMWS/96/4, 87 p. Recruitment Monitoring Program, CSIRO Division of Marine Research, GPO Box 1538, Hobart 7001, Australia. tic bluefin tuna using depth sensitive ultrasonic telemetry devices. LHtrasonic telemetry is a proven technique for ac- quiring the required precise and detailed data on the be- haviors of pelagic fishes in relation to oceanographic condi- tions (e.g. Holland et al., 1990; Dagorn et al., 1999, 2000a; Lutcavage et al., 2000). Besides being useful for improv- ing stock assessments (Brill and Lutcavage, 2001), the re- sultant data can also help clarify basic ecological relation- ships and provide inferences on physiological abilities and species-specific behaviors (Carey, 1983; Brill. 1994; Brill et al., 1999). Materials and methods Fishing operations were conducted from a 16-m commer- cial fishing boat (FV Gruryipy) in the western North Atlan- tic off the eastern shore of Virginia (Fig. lA) during June and July 1998. Bluefin tuna were captured with standard recreational trolling gear. The fish were brought aboard with a plastic sling and detached from hooks. Straight line fork length was measured, and a Vemco (Halifax, Nova Scotia, Canada) ultrasonic transmitter (model V32) was attached near the second dorsal fin with nylon straps as described by Holland et al. ( 1986, 1990). The transmit- ted signal was detected with a Vemco VR-60 ultrasonic receiver connected to a directional hydrophone mounted on the end of an aluminum pipe. The pipe was clamped to the side of the vessel with a custom designed alumi- num bracket that allowed the hydrophone to be rotated to find the relative bearing to the transmitter Fish depth, encoded by the interval of the transmitters pulsed signal, was decoded by the receiver and the resultant digital data recorded by an attached laptop computer. Geographic Brill et a\: Horizontal and vertical movements of iiivenlle Thunnw, thynniK 157 Table 1 Summary of tracks Body masses were of'fiv :alcu] ? juvenile Atlantic bluofin ated from fork lengths wi tuna th the Tlumiuts thynnus) equipped with ultrasonic depth sensitive transmitters, weight-length regression equation provided by Coull el al. (1989). Kish 11(1. Fork loiigth cm Body mass kg Dates of trai (1988) k Duration of t rack h Total distance covered km (nmi) Distance between start and end points km (nmi) Mean(±SEM) swimming speed knots 1 74 6.7 17-19 Jun 47.2 217(117) 52.7(28,4) 2.8 ±0.03 2 91 12.1 23-25 Jun 47.8 267(144) 11.4(6.1) 3.0 ±0.03 3 79 8.0 2-3 Jul 30.0 152(82) 7.7(4.1) 2.7 ±0.04 4 99 15.4 6-7 Jul 31.2 192(104) 14.7(7.9) 3.3 ±0.03 5 106 18.8 10-12 Jul 47.9 289(156) 32.2(17.4) 3.2 ±0.03 positions were obtained by using a GPS satellite receiver and were recorded on a second laptop computer every minute. The tracking vessel's position was assumed to be the same as that of the fish. Sea surface temperature and bottom depth were recorded manually every 15 minutes by using a hull-mounted electronic temperature sensor and color fathometer, respectively. Depth-temperature profiles were taken approximately every four hours with a Sippican (Marion, MA) portable XBT system (model MK12). Aggregate time-at-depth and time-at-temperature dis- tributions were calculated fromlO-m and 1°C bins (respec- tively), as described by Holland et al. (1990). These data were subsequently expressed as a fraction of the total time each fish was followed, and the fractional data bins were averaged across all fish. Speed over ground (henceforth re- ferred to simply as "speed") was calculated by assuming that the fish moved in a straight line between successive geographic locations. Sea surface temperature (SST) data were recorded by the advanced very high resolution radiometer (AVHRR) carried onboard the NOAA-14 polar orbiting operational environmental satellite. High resolution picture transmis- sion (HRPT) data were obtained from the National Coast Watch Active Access System at the National Oceano- graphic Data Center and had a spatial resolution of 1.25 x 1.25 km pixels. Ocean color data were recorded by the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) car- ried onboard the Orbview-2 spacecraft (Orbimage, Inc., Dulles. VA). The level-2 global area coverage (GAC) data were obtained from the NASA Goddard Space Flight Cen- ter's Distributed Active Archive Center. These 4-km reso- lution data sets included chlorophyll-a surface concentra- tion and the diffuse attenuation coefficient at 490 nmi, m vacuo. We calculated the occupancy of waters with specific chlorophyll-a concentrations and light attenua- tions from values corresponding to and coincident with the tracks of fishes derived from satellite images. These data were subsequently expressed as a fraction of the to- tal number of observations for each fish, and the fraction- al data bins were averaged across all fish. For illustrative purposes, we also generated composite images using data from the 21-day period over which all tracking operations were conduced. Results The bottom topography in the areas where the fish were tracked is generally featureless, except for small areas where the vertical relief is approximately 2 m above the surroundings. Local fishermen have named these features (Fig. IB and subsequent figures), and the names used in this study are taken from local fishing charts. Size of fish, starting and ending dates of tracks, dura- tion of tracks, distances covered, distance between start- ing and ending points, and mean (±SEM) swimming speed offish are listed in Table 1. With the exception offish num- ber 4 (referred to simply as "fish 4"), individuals tended to follow highly irregular courses that often repeatedly covered the same areas (Fig. IB). The mean distance be- tween starting and ending points for all fish was only ll'7f (range: 4— 25'/f ) of the total distance covered (Table 1). From tracking studies of yellowfin tuna [Thitnniis alba- cares) in the Pacific, Dagorn et al. (2000a) concluded that such frequent directional changes might be characteristic of foraging behavior. The frequency of observed swimming speeds is shown in Figure 2. Although all fish reached maximum speeds of =7 knots for brief periods, over 90% of the observed speeds were less than 3.6 knots. Horizontal movements Fish 1 was captured and released at 1340 h, approximately 1.8 km (1 nmi) west of the "26 Mile Hill" (Fig. 3). It pro- ceeded on a southerly course for about 33 h, a direction that carried it over the "Hot Dog" and "Southeast Lumps." After sunset on the second day, the fish reversed its course and eventually recrossed both features. The fish was approxi- mately 5.6 km (3 nmi) south of the "Southeast Lumps," and moving south, when the track was terminated at 1300 h. Fish 2 was captured approximately 5.6 km (3 nmi) south of the "Southeast Lumps" (Fig. 3) at 1547 h, adjacent to where the track of fish 1 was completed four days earlier. It 158 Fishery Bulletin 100(2) 2 - I 1 I U I I H I I I ]\ 1 I 11 L I, jW^. O- Q O- N- N- nk^T^- T/- 'b-^ t«- tx- tK- vnA^ ■n*^ ■^ v / — 1 — V y H 1 — H 1 — 1 — \— 1X:>-- 1 H — 1 — -1 1- — 1 — — h- 4 c -- 3 -- 2 in 1400 1800 2200 0200 0600 1000 1400 1800 8 12 16 20 24 28 Temperature ( 'C) Figure 6 Swimming speed (solid line, upper panel) and vertical movements (lower panel) of fish 4. The change in tempera- ture in the horizontal direction (expressed as sea surface temperature, SST) is shown by the broken line in the upper panel (Brill and Lutcavage, 2001). As in Figure 5. the change in temperature in the vertical direction (mean ±SEM) is shown to the right of the vertical movement plot. Note that changes in swimming speed are not cor- related with changes in SST, and that the steepest temperature change the fish could experience moving horizon- tally (generally less then 0.5°C/km) is several orders of magnitude less then that experienced moving vertically (=0.6°C/m). occur throughout the water column during daylight, are abundant in the areas where we tracked the fish, and dominate the diet of tunas in this area (Mason, 1976; Egg- leston and Bochenek, 1989). The nature offish 4's descents up to =160 m while off the continental shelf (Fig. 6) re- main unclear, although they too may be related to foraging (Dagorn et al., 2000a, 2000b). Their brevity is most likely due to the inability of Atlantic bluefin tuna to withstand temperatures below 10°C for long periods of time, rather than to an intolerance of low ambient oxygen conditions. Although no depth-oxygen profiles were obtained during our study, available data- suggest that juvenile tuna did not encounter ambient oxygen levels that were likely to be stressful (Bushnell and Brill, 1991, 1992). The behavior pattern we observed of short oscillatory dives near the surface is similar to that of both juvenile bluefin tuna in the eastern Pacific (Marcinek et al., 2001) and adult bluefin tuna tracked in the Gulf of Maine (Lut- cavage et al., 2000). In all cases, fish spent the majority of their time in the surface layer, although in the Gulf of Maine and eastern Pacific, the temperature of the warm- est water available was lower (=13-22°C) and more vari- able. As shown in Figure 11, when expressed as the rel- ative change in temperature with depth (i.e. in relation to the surface water temperature occurring during each track), time-at-temperature distributions of juvenile and adult Atlantic bluefin tuna become essentially identical. Moreover, the limiting effects of temperature change on vertical movements are independent of body size. Simi- larly, yellowfin tuna tracked near the main Hawaiian Is- lands and off the coast of California occupy the warmest water available, regardless of body size, even though sur- face water temperature in the two areas differs by more than 5°C (Holland et al., 1990; Block et al., 1997; Brill et al., 1999). Atlantic bluefin tuna, however, are more eury- thermal than yellowfin tuna. The latter will rarely expose themselves to more than an 8°C change in temperature, whereas the former regularly subject themselves to a tem- perature change of up to 13°C (Fig. 11). Surprisingly, the behavior of juvenile bluefin tuna observed by Marcinek et al., (2001) was more like that of yellowfin tuna in that these juvniles would not expose themselves to more then an 8°C temperature change. It still remains to be conclusively demonstrated, how- ever, whether the vertical movement patterns of tunas and other large pelagic fishes are (as suggested by Brill et al., 1993, 1999) limited by the effects of ambient temperature on cardiac function. Or whether, as suggested by Mar- cinek et al. (2001), that depth distributions "... may have more to do with the location of prey, and the physiological limitations of the prey, than physiological limitations of the bluefin Itunaj." Moreover, months-long observations 162 Fishery Bulletin 100(2) 30 20 10 10 Time at deptti (%) 20 10 10 20 30 40 Time at temperature (%) Figure 7 Vertical distribution of five juvenile bluefin tuna ex- pressed as percent time (mean ±SEMi spent at specific deptbs I Ai and at specific temperatures iBi. Shaded bars indicate nighttime and open bars indicate daytime. of juvenile bluefin tuna in the western Pacific recently ob- tained with archival (i.e. electronic data recording) tags have shown that the vertical movements of juvenile blue- fin tuna can have strong seasonal and geographic com- ponents (Kitagawa et al., 2000). In areas and at times (e.g. winter) when there was a strong thermocline, bluefin tuna remained in the uniform-temperature surface layer and demonstrated vertical movement behaviors similar to those observed during the short-term ultrasonic tele- metry studies. In the summer, when the themocline was less pronounced, the fish showed very distinct diel peri- odicity in their vertical movement patterns. They would remain at the stirface at night and make rapid vertical movements (from surface to =120 m and from =21''C to 14°C) during the day. Kitagawa et al. (2000) concluded that the differences in behavior patterns were related to foraging. It is also still an open question as to what ex- tent bluefin tuna's ability to conserve metabolic heat and maintain elevated muscle temperatures (Carey and Teal, 1966) enhances vertical mobility. Roffer (1987) was apparently the first to propose that movements and abundance of juvenile Atlantic bluefin tuna are controlled by the depth and thickness of the 18.5-20.5°C "preferred habitat" temperature layer Likewise, Inagake et al. (2001), using archival tags implanted into juvenile 37,8 37.6 - 75 8 75.6 75 4 75.2 75.0 74 8 74 6 74 4 74 2 74.0 Figure 8 Composite satellite sea surface temperature image (17 June-10 July 1998) and movements of the five juvenile bluefin tuna (Brill and Lutcavage, 2001 ). Figure reprinted with permission of Ameri- can Fisheries Society. bluefin tuna in the western Pacific, found evidence that this temperature range is indeed always "preferred." Dur- ing the periods of our observations, juvenile bluefin tuna spent the majority of their time (=809c) in water greater then 22°C. A plausible explanation is that under the condi- tions of our observation period, juvenile bluefin tuna simply occupy the warmest water available, although a relatively uniform temperature surface layer was evident only dur- ing tracks of fish 3, 4, and 5. We also did not find any con- clusive indication that juvenile bluefin tuna avoided sur- face water temperatures above 26°C. Although fish spent less than 'ZO'^i of time at these temperatures (Fig. 7), less than 20% of the recorded sea surface temperatures (i.e. the warmest water available) were above 26°C. We also found no relationship between sea surface tem- perature and horizontal movements (Figs. 6 and 8), al- though this relationship has been demonstrated for other tuna species in other areas (e.g. Laurs et al., 1977; Fiedler and Bernard, 1987; Uda, 1973). We argue that our results are due to the differences in the vertical and horizontal temperature gradients occurring along the Virginia coast. Juvenile bluefin tuna routinely traveled through the ther- mocline, moving from the relatively warm surface layer into the mid-Atlantic cold-pool water (Houghton et al., 1982; Houghton and Marra, 1983) underlying it. The fish thus experienced temperature gradients of up to =0.6°C/m (Figs. 5 and 6). In contrast, the steepest horizontal tem- perature gradient in the area where the fish were tracked was approximately three orders of magnitude smaller (=0.5°C/km). In other words, the frequent vertical move- ments of juvenile bluefin tuna probably prevent them from detecting and responding to SST gradients. Brill et al.: Horizontal and vertical movements of juvenile Thunnus thynnus 163 diffuse attenuation poefficienl. 490 nm ^^002 [comoos !e r^age A^^^^fi July 1998) ^ |B>>Bin 75.8 75 6 75 4 ,■. . - 74 8 74.6 74 4 /4 ^' 74.0 Figure 9 Composite satellite images (17 June-lO July 1998) showing (A) chloro- phyll-Q concentrations (mg/m') and (B) water clarity measured as the diffuse attenuation coefficient (1/m, at an in vacuo wavelength of 490 nmi) and movements of the five juvenile bluefin tuna. Locations of juve- nile bluefin tuna schools recorded during aerial surveys conducted in 1997 are shown by filled circles. (Lutcavage, M. 1998. Aerial survey of school bluefin tuna off the Virginia Coast, July 1997. Report to the National Marine Fisheries Service, cooperative agreement NA77fm0.533. (Available from the author, Edgerton research Laboratory. New England Aquarium, Central Wharf Boston, MA 02110].) are shown by filled cir- cles. The edge of the continental shelf is indicated by the .50-, 100-. and 200-m isobath lines (Brill and Lutcavage, 2001). Figure reprinted with permission of American Fisheries Society. Carey ( 1992) was one of the first to appreciate the impor- tance of vertical thermal structuring and stated "Temper- ature gradients of 15° to 20°C are not uncommon within the depth ranges of pelagic fish. By moving a few hundred meters vertically, an animal may encounter a greater tem- perature change than it experiences seasonally or in mov- ing thousands of miles horizontally." As with bluefin tuna, the vertical movements of yellowfin tuna and swordfish also result in their experiencing vertical temperature gra- dients orders of magnitude greater than horizontal tem- perature gradients (Carey and Robison, 1981; Carey, 1990; Holland et al., 1990; Cayre and Marsac, 1993). The inabil- 164 Fishery Bulletin 100(2) «D N "^ 1 "^ C> ^ Si- C) O O C) O C) S 60 1 Log (chlorophyll-a) (mg/m) Diffuse attenuation coefficient (490 nmi/m) Figure 10 Frequency histogram (mean ±SEM) of chlorophyll-o concentrations ( mg/ni'^ i and water clarity measured as the diffuse attenuation coefficient ( 1/m, at an in vacuo wavelength of 490 nmi ) in waters along the track lines of five juvenile bluefin tuna (Brill and Lutcavage, 2001). The span of the horizontal axes show the approximate range of these variables present off the eastern shore of Virginia. ity of fish to sense shallow horizontal temperature gra- (Jients in the face of the steep vertical temperature gra- dients they routinely experience may explain, therefore, why Power and May (1991) and Podesta et al. (1993) could find no correlation between SST "fronts" and the appar- ent abundance of yellowfin tuna in the Gulf of Mexico and swordfish in the western north Atlantic. In contrast to SST, water clarity and phytoplankton abundance appear to have a strong influence on the hor- izontal movements of juvenile bluefin tuna (Figs. 9 and 10). Tunas are sight hunters, and possess the highest vi- sual acuity of any teleost (Nakamura, 1968). We suspect that juvenile bluefin tuna remain in water masses with a standing phytoplankton biomass sufficient to support con- centrations of prey, but where turbidity is low enough that visual prey detection and prey capture abilities are not impeded. Our conclusion is further supported by the lo- cations of juvenile bluefin tuna schools detected in aerial 50 40 30 20 10 ■! adult Atlantic bluefin tuna JjjWL..^ -? 60 -I o I 50 I 40 Cl I 30 t 20- E P 10 60 50 40 30 20 10 T> f>' ■?> ^*• <^' N*' ■i^' -y v>' H' :S > 53 fe ^ ,* ?l kS> N^ 0- v> n"* <0^ C^ <^ <$- <^ C^ <^ '^ <^^^^^ ^ Temperature interval ( C) Figure 11 Frequency histograms (mean +SEM) showing time spent at specific temperatures by adult bluefin tuna tracked in the Gulf of Maine (western North Atlantic) with tem- peratures expressed as water temperature (A), and with temperatures expressed in relation to surface layer tem- perature (B), data taken from Lutcavage et al., 2000). Equivalent data for juvenile bluefin tuna are presented in panel C. Shaded bars indicate nighttime and open bars indicate daytime. surveys conducted in 1997.^ Although satellite data show- ing diffuse attenuation coefficients and chlorophyll-a con- centrations are not available for 1997, bluefin tuna schools were located in the areas where the fish carrying ultra- sonic transmitters remained (Fig. 9). Olson and Podesta ( 1987), Olson et al. ( 1994), and Humston et al. (2000) have also concluded that aggregations of highly mobile species Lutcavage, M. 1998. Aerial survey of school bluefin tuna off the Virginia Coast, July 1997. Report to the National Marine Fisheries Service (cooperative agreement NA77FM0533). (Avail- able from the author, Edgerton Research Laboratory, New Eng- land Aquarium, Central Wharf Boston. MA 02110,1 Brill et a! Horizontal and vertical movements of luvenlle Thunnus thynnus 165 at fronts result from cues other than SST, such as changes in the photic environment associated with phytoplankton distribution, changes in prey abundance, or enhanced for- age opportunities. Aerial survey techniques and population assessments of juvenile bluefin tuna Techniques for interpretation of aerial survey data with respect to population assessments are complex (e.g. Lo et al.. 1999; Newlands and Lutcavage, 2001), and a thorough discussion is beyond the scope of our present study. We can, however, use our data on juvenile bluefin tuna's verti- cal movements and distribution patterns to provide some inferences as to how often they are likely to be visible at the ocean's surface or detectable at a specific depth. Juve- nile bluefin tunas spent less than 13'7c of daylight hours at depths of 0-3 m (Fig. 7), where visual or photographic detection is possible. The depth distribution of juvenile fish was similar to that of adult bluefin tuna tracked in the Gulf of Maine (12*^ of daylight hours at depths of 0-4 m; Lutcavage et al.. 2000). Abundance estimates based solely on photographic data will, therefore, have to be corrected to account for the significant number offish that maybe be present, but that are beyond detection range. Fish detec- tion systems that use lasers (the so called "light detection and range" or "LIDAR" systems) are expected to have a depth detection zone of up to 60 m (Oliver et al., 1994). This detection zone encompasses almost the entire water column over the sections of continental shelf where juve- nile bluefin tuna are likely to be found. Moreover, if the behavior of the fish that moved into deeper water off the continental shelf is assumed typical, then juvenile bluefin tuna would be detected by LIDAR systems even in deep water The relatively small net displacement distance (i.e. distance between start and end points. Table 1) may require the development of filtering algorithms to reduce errors caused by double counting if parallel transects are flown less than =50 km apart, or if the same area is resur- veyed weekly or more often. Conversely, significant fish aggregations could be missed if parallel transects are too widely spaced. Acknowledgments This project was funded by a grant from the National Marine Fisheries Service to the Edgerton Research Labo- ratory. New England Aquarium. RWB's participation was funded through cooperative agreements NA37R.J0199 and NA67RJ0154 from the National Oceanic and Atmospheric Administration with the Joint Institute for Marine and Atmospheric Research, University of Hawaii. We grate- fully acknowledge Mark Luckenbeck and the staff and students of the Virginia Institute of Marine Science's East- ern Shore Laboratory for their gracious hospitality and extraordinary efforts to make this project a success. We also thank Jim Hannon and Sippican Inc. (Marion, MA) for use of their Mark 12 system and generous donation of XBT probes, and Mitch Roffer for access to historical bluefin tuna data. We acknowledge the DAAC at NASA's Goddard Space Flight Center and Orbimage Inc. for ocean color data and the NOAA Coast Watch Program and the National Oceanographic Data Center for SST data. We especially thank Captain Jack Stallings of the FV Grumpy (Virginia Beach, Virginia) for his unflagging enthusiasm and his significant contributions to making our project a success. Literature cited Bakun. A., J. Beyer. D. Pauly, J. G. Pope, and G. D. Sharp. 1982. Ocean sciences in relation to living resources. Can. J. Aquat. Sci. 39:1059-1070. Bertrand. A., and E. Josse. 2000. Acoustic estimation oflongline tuna abundance. ICES (Int. Council for the Exploration of the Sea) J. Mar. Sci. .57:919-926. Block, B. A., J. E. Keen, B. Castillo, H. Dewar, E. V. Freund, D. J. Marcinek. R. W. Brill, and C. Farwell. 1997. Environmental preferences of yellowfin tuna iTInin- nus albacares) at the northern extent of their range. Mar. Biol. 130:119-132. Brill, R.W. 1994. A review of temperature and oxygen tolerance stud- ies of tunas pertinent to fisheries oceanography, movement models, and stock assessments. Fish. Ocean. 3:204-216. Brill, R.. B. Block, C. Boggs. K. Bigelow, E. Freund, and D. Marcinek. 1999. Horizontal movements and depth distribution of large adult yellowfin tuna iThunnus albacares) near the Hawai- ian Islands, recorded using ultrasonic telemetry: implica- tions for the physiological ecology of pelagic fishes. Mar Biol. 133:395-408. Brill R. W., D. B. Holts, R. K. C. Chang, S. Sullivan, H. Dewar. and F. G. Carey 1993. Vertical and horizontal movements of striped marlin {Tetrapturusaudax) near the main Hawaiian Islands, deter- mined by ultrasonic telemetry, with simultaneous mea- surement of oceanic currents. Mar Biol. 117:567-574. Brill, R. W., and M. E. Lutcavage. 2001. Understanding environmental influences on move- ments and depth distributions of tunas and billfishes can significantly improve population assessments. In Island in the stream: oceanography and fisheries of the Charles- ton Bump (G. Sedberry ed.), p. 179-198. Am. Fish. Soc. Symp. 25. Am. Fish. Soc, Bethesda, MD. Bushneil, P.G., and R.W. Brill. 1991. Responses of swimming skipjack [Katsuwonus pela- mis) and yellowfin (Thunnus albacares) tunas exposed to acute hypoxia, and a model of their cardio-respiratory func- tion. Physiol. Zool. 64:787-811. 1992. Oxygen transport and cardiovascular responses in skipjack tuna (Katsuwonus pelamis) and yellowfin tuna iThunnus albacares) exposed to acute hypoxia. J. Comp. Physiol. B 163:131-143. Butterworth D. S., and A. E. Punt. 1993. The robustness of estimates of stock status for the western north Atlantic bluefin tuna population to violations of the assumptions underlying the associated as.sessments models. ICCAT (Int. Commission for the Conservation of Atlantic Tunas) Coll. Vol. Sci. Pap. 43(1): 192-210. 166 Fishery Bulletin 100(2) Carey, F. G. 1983. Experiments with free-swimming fish. In Oceanog- raphy — the present and future ( P. G. Brewer, ed. I, p. 57-68. Springer- Verlag, New York, NY. 1990. Further acoustic telemetry observations of swordfish. In Plannmg the future of billfishes ( R. H. Stroud, ed. ), p. 102- 122. National Coalition for Marine Conservation Inc., Savanna, GA. 1992. Through the thermochne and back again. Oceanus 35:79-85. Carey, F. G., and B. H. Robison. 1981. Daily patterns in the activities of swordfish, Xi'p/j/as gladius. observed by acoustic telemetry. Fish. Bull 79:277- 292. Carey FG., and. J. M. Teal. 1966. Heat conservation in tuna fish muscle. Proc. Natl. Acad. Sci. 56:1464-1469. Cayre P., and F Marsac. 1993. Modelling the yellowfin tuna iThunnus albacares) vertical distribution using sonic tagging results and local environmental parameters. Aquat. Living Resour. 6:1-14 Churchill, J. H., E. R. Levme, D. N. Connors, and P. C. Cornillon. 1993. Mi.xing of shelf slope and Gulf Stream water over the Continental Slope of the Middle Atlantic Bight. Deep Sea Res. 40:1063-1065. Clark, C. W., and M. Mangel. 1979. Aggregation and fishery dynamics: a theoretical study of schooling and the purse seine tuna fisheries. Fish. Bull. 77:317-329. Coull, K. A., A. S. Jermyn, A. W. Newton, G. I. Henderson, and W. B. Hall. 1989. Length/weight relationships for 88 species offish en- countered in the North Atlantic. Scottish Fish. Res. Rep. 43, 80 p. Dagorn, L., P. Bach, and E. Josse. 1999. Movement patterns of large higeye tuna [Tli tin n us obe- sus) in the open ocean, determined using ultrasonic tele- metry Mar. Biol. 1.36:361-371. Dagorn, L., E. Josse, and P. Bach. 2000a. Individual differences in horizontal movements of yellowfin tuna iThunnus albacares) in nearshore areas in French Polynesia, determined using ultrasonic telemetry. Aquat. Living Resour 13:19.3-202. Dagorn, L., F. Menczer, P. Bach, and R. J. Olson 2000b. Co-evolution of movement behaviors by tropical pelagic predatory fishes in response to the environment: a simulation model. Ecol. Model. 134:325-341. Eggleston, D .B., and E. A. Bochenek. 1989. Stomach contents and parasite infestation of school bluefin tuna Thunnus thynnus collected from the Middle Atlantic Bight, Virginia. Fish. Bull. 88:389-395. Fiedler, R C, and H. .J. Bernard. 1987. Tuna aggregations and feeding near fronts observed in satellite imagery Continental Shelf Res. 7: 871-881. Hilborn, R., and C. .J. Waters. 1992. Quantitative fisheries stock assessment: choice, dy- namics and uncertainty. Chapman and Hall. New York, NY, 570 p. Holland, K. N.. R.W. Brill, J. S. Ferguson, R. K. C. Chang, and R. Yost. 1986. 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. Houghton R. W., and J. Marra. 1983. Physical/biological structure and exchange across the thermochne ShelfSlope in the New York Bight. J. Geo- phys. Res. (C Oceansi 88:4467-4480. Houghton, R. W., R. Schultz, R. C. Beardsley B. Butman, and .J. L. Chamberlin. 1982. The middle Atlantic Bight cold pool: evolution of the temperature structure during summer 1979. J. Phys. Oceanogr. 12:1019-1029. Humston, R., J. S. Ault, M. Lutcavage, and D. B. Olson. 2000. Schooling and migration of large pelagic fishes rela- tive to environmental cues. Fish. Oceanogr. 9: 136-146. Inagake, D., H. Yamada, K. Segawa, M. Okazaki, A. Nitta, and Tltoh. 2001. Migration of young bluefin tuna, Thunnus orientalis Temminck et Schlegel, through archival tagging experi- ments and its relation with oceanographic conditions in the western north Pacific. Bull. Natl. Res. Inst. Far Seas Fish. 38:.53-81. Kitagawa, T, H. Nakata, S. Kimura, T Itoh, S. Tsuji, and A. Nitta. 2000. Effect of ambient temperature on the vertical dis- tribution and movement of Pacific bluefin tuna Thunnus thynnus orientalis. Mar Ecol. Prog. Sen 206:251-260. Klimley, A. P. 1993. Highly directional swimming by scalloped hammer- head sharks, Sphyrna lewini. and subsurface irradiance, temperature, bathymetry, and geomagnetic field. Mar. Biol. 117:1-22. Lauck, T 1996. LIncertainty in fisheries management. In Fisheries and uncertainty: a precautionary approach to resource management (D. V. Gordon and G. R. Munro, eds.), p. 91- 105. Univ. Calgary Press, Calgary. Laurs, R. M., H. S. H. Yuen, and .J. H. Johnson. 1977. Small-scale movements of albacore, Thunnus ala- lunga, in relation to ocean features as indicated by ultra- sonic tracking and oceanographic sampling. Fish. Biol. 75:.347-355. Lo, N. C. H., J. R. Hunter, and J. H. Churnside. 1999. Modeling statistical performance of an airborne LIDAR survey system for anchovy. Fish. Bull. 98:264282. Lo, N. C. H., L. D. Jacobson, and J. L. Squire. 1992. Indices of relative abundance from fish spotter data based on delta-log normal models. Can. J. Fish. Aquat. Sci. 49:251.5-2526. Lucy, J. A., E. A. Bochenek, and N. J. Chartier. 1990. Fleet characteristics and boat-owner expenditures associated with Virginia's recreational marlin-tuna fishery. In Planning the future of billfishes: research and manage- ment in the 80's and beyond: proceedings of the second international, billfish symposium, part 2 (R. H. Stroud, ed.) p. 253-262. Marine Recreational Fisheries 13. National Coalition for Marine Conservation, Savannah. GA. Lutcavage, M. E., R. W. Brill, G. B. Skomal, B. C. Chase, J. L. Goldstein, and J. Tutem. 2000. Movements and behavior of adult North Atlantic blue- fin tuna {Thunnus thynnus) in the northwest Atlantic deter- mined using ultrasonic telemetry. Mar. Biol. 137:347-358. Lutcavage M.. and S. Kraus. 1995. The feasibility of direct photographic assessment of giant bluefin tuna in New England waters. Fish. Bull. 93: 49,5-503. Lutcavage, M., S. Kraus, and W. Hoggard. 1997. Aerial assessment of giant bluefin tuna in the Bahama Banks-Straits of Florida, 1995. Fish. Bull. 95:300-310. Brill et a\: Horizontal and vertical movements of luvenile Thunmis thynnus 167 Marcinek, D. J., S. B. Blackwell, H. Dewar, E. V. Freund, C. Farwell, D. Dau, A. C. Seitz, and B. A. Block. 2001. Depth and muscle temperature of Pacific bluefin tuna examined with acoustic and pop-up satellite archival tags. Mar. Biol. 138:869-885. Marsac. F., and P. Cayre. 1998. Telemetry applied to behavior analysis of yellowfin tuna iThunnus alhoca/Vf:. Bonnaterre. 17881 movements in a network of fish aggi-egating devices. Hydrobiolgia 371/ 372:155-171. Mason. J. M. 1976. Food of small, northwestern Atlantic bluefin tuna, Thunnus thynnus (L), as ascertained through stomach con- tent analysis. M.S. thesis, Univ. Rhode Island, Kingston, RI, 27 p. Mather, F. J., J. M. Mason, and A. C. Jones. 1995. Historical document: life history and fisheries of Atlantic bluefin tuna. U.S. Dep. Commer, NOAA Tech. Memo. NMFS-SEFSC-370, 165 p. Nakamura, E. L. 1968. Visual acuity of two tunas, Katsuivonus pelatnis and Euthynnus affinis. Copeia 1968:41-49. Newlands, N., and M. Lutcavage. 2001. From individuals to local population densities: move- ments of north Atlantic bluefin tuna (Thunnus thynnus) schools in the Gulf of Maine/Northwest Atlantic. In Elec- tronic tagging and tracking in marine fish (J. Sibert and J. L. Neilsen, eds.), p. 421-433. Kluwer Academic Pubis.. Dordrecht. Oliver C. W., W. A Armstrong, and J. A. Young. 1994. Development of an airborne LIDAR system to detect tunas in the Eastern tropical Pacific purse-seine fishery U.S. Dep. Commer , NOAA Tech. Memo. NMFS-SWFSC-204, 67 p. Olson. D. B., G. L. Hitchcock, Mariano, A. J.. C. J. Ashjian, G. Peng, R. VV. Nero, and G. P. Podesta. 1994. Life on the edge: marine life and fronts. Oceanogra- phy 7:52-60. Olson, D. B., and G. P. Podesta. 1987. Oceanic fronts as pathways in the sea. In Signposts in the sea (W. F. Hernkind and A. B. Thistle, eds. I, p. 1-15. Florida State Univ. Press, Tallahassee, FL. Polacheck. T. E. Pikitch. and N. Lo. 1996. Evaluation and recommendations for the use of aerial surveys in the assessment of Atlantic bluefin tuna. ICCAT Coll. Vol. Sci. Pap. 68( 1 ):61-68. Podesta, G. P., J. A. Browder, and J. J. Hoey. 1993. Exploring the association between swordfish catch rates and thermal fronts on U.S. longline grounds in the western North Atlantic. Continental Shelf Res. 13:253- 277. Power, J. H., and L. N. May Jr 1991. Satellite obsen'ed sea-surface temperatures and yellowfin tuna catch and effort in the Gulf of Mexico. Fish. Bull. 89:429-439. Reiss, C. S., and J. R. McConaugha. 1999. Cross-frontal transport and distribution of ichthyo- plankton associated with Chesapeake Bay plume dynam- ics. Continental Shelf Res. 19:151-171, Rcsprepo, V. 1996. An implementation of an age-structured production model with application to western Atlantic bluefin tuna fisheries. ICCAT Coll. Vol. Sci. Pap. 66(21:348-353. Restrepo, V. R., C. E, Porch, S. C.Turner, G. P. Scott, and A. A. Rosenberg. 1994. Combination of spawner-recruil. spawning biomass- per-recruit and yield-per-recruit computations for the esti- mation of the long term potential for west Atlantic bluefin tuna. ICCATColl.Vol. Sci. Pap. 42(1 ):214-222. Rivas, L. R. 1978. Preliminary models of annual life history cycles of the north Atlantic bluefin tuna. In The physiological ecology of tunas (G. D. Sharp and A. E. Dizon, eds.), p. 369-393. Academic Press. New York, NY. RofTer. M. A. 1987. Influence of the environment on the distribution and relative apparent abundance of juvenile Atlantic bluefin tuna along the United States east coast. Ph.D. diss. Univ. Miami, Miami. FL, 154 p. Sakagawa, G.T 1975. The purse seine fishery for bluefin tuna in the north- western Atlantic Ocean. Mar Fish. Rev. 37:1-8. Sharp, G. D. 1978. Behavioral and physiological properties of tunas and their effects on vulnerability to fishing gear In The physi- ological ecology of tunas (G. D. Sharp and A. E. Dizon, eds.), p. 397-449. Academic Press, New York, NY. Sissenwine, M. P., P. M. Mace, J. E. Powers, and G. P. Scott. 1998. A commentary on western Atlantic bluefin tuna assessments. Trans. Am. Fish. Soc. 127:838-855. Turner, S. C. C. A. Brown, and H. Huang. 1997. Standardized catch rates of small bluefin tuna, Thun- nus thynnus. from the rod and reel fishery off Virginia- Rhode Island in 1980-1995. ICCAT Coll. Vol. Sci. Pap. 46(l):295-297. Turner, S. C, H, A, Forster, J. Witzig, C. Brown, M. Terceiro, D. A. Van Voorhees. G. P. Scott, J. A. Browder, D. J. Christensen, and C. E. Porch. 1993. Estimation and projection of the 1992 landing of blue- fin tuna in the rod and reel fishery for large pelagic fish off the northeast United States. ICCAT Coll. Vol. Sci. Pap. 40(11:1-16. Uda, M. 1973. Pulsatile fiuctuations of oceanic fronts in association with the tuna fishing grounds and fisheries. J. Fac. Mar. Sci. Technol.. Tokai Univ 7:245-266. Yuen. H. S. H. 1970. Behavior of skipjack tuna. Katsuwonus pelamis, as determined by tracking with ultrasonic devices. J. Fish. Res. Board Can. 27:2071-2079. Walker. M. M. 1984. Learned magnetic field discrimination in yellowfin tuna, Thunnus albacares. J. Comp. Physiol. A. 155:673- 679. 168 Abstract-The stomachs of 819 Atlan- tic bluefin tuna (Thiinnus thynnusi sampled from 1988 to 1992 were ana- lyzed to compare dietary differences among five feeding grounds on the New England continental shelf (Jef- freys Ledge, Stellwagen Bank. Cape Cod Bay, Great South Channel, and South of Marthas Vineyard) where a majority of the U.S. Atlantic commer- cial catch occurs. Spatial variation in prey was expected to be a primary influence on bluefin tuna distribution during seasonal feeding migi'ations. Sand lance iAmmodytes spp.), Atlantic herring (Clupea harengus). Atlantic mackerel (Scomber scombr-us), squid (Cephalopoda), and bluefish (Pomato- mus saltatrix) were the top prey in terms of frequency of occuiTence and percent prey weight for all areas com- bined. Prey composition was uncorre- lated between study areas, with the exception of a significant association between Stellwagen Bank and Great South Channel, where sand lance and Atlantic herring occurred most fre- quently. Mean stomach-contents bio- mass varied significantly for all study areas, except for Great South Channel and Cape Cod Bay. Jeffreys Ledge had the highest mean stomach-contents bio- mass (2.0 kg) among the four Gulf of Maine areas and Cape Cod Bay had the lowest (0.4 kg). Diet at four of the five areas was dominated by one or two small pelagic prey and several other pelagic prey made minor contri- butions. In contrast, half of the prey species found in the Cape Cod Bay diet were demersal species, including the frequent occurrence of the sessile fig sponge iSuberites ficus). Prey size selec- tion was consistent over a wide range of bluefin length. Age 2-4 sand lance and Atlantic herring and age 0-1 squid and Atlantic mackerel were common prey for all sizes of bluefin tuna. This is the first study to compare diet composition of western Atlantic bluefin tuna among discrete feeding grounds during their seasonal migi'ation to the New Eng- land continental shelf and to evaluate predator-prey size relationships. Previ- ous studies have not found a common occurrence of demersal species or a pre- dominance of Atlantic hening in the diet of bluefin tuna. Differences in diet of Atlantic bluefin tuna iThunnus thynnus) at five seasonal feeding grounds on the New England continental shelf* Bradford C. Chase Massachusetts Division of Marine Fisheries 30 Emerson Avenue Gloucester, Massachusetts 01930 E mail address brad chase gistale ma us Manuscript accepted 21 September 2001. Fish. Bull. 100:168-180 (2002). Atlantic bluefin tuna iTIuinmis thyn- nus) are widely distributed throughout the Atlantic Ocean and have attracted valuable commercial and recreational fisheries in the western North Atlantic during the latter half of the twentieth century. The western North Atlantic population is considered overfished by the International Commission for the Conservation of Atlantic Tunas ( NMFS, 1999). Bluefin tuna are the largest scombrid species, and the largest tele- ost occurring in the Gulf of Maine (Big- elow and Schroeder, 1953). Bluefin tuna migrate to coastal waters off New Eng- land during warmer months, feeding on local concentrations of prey. This migration supports a major component of the U.S. Atlantic commercial fishery for bluefin tuna; from 1978 to 1992 New England accounted for between TS'/r and 98% of annual commercial land- ings (Chase, 1992; and NMFS, 1995). Substantial annual variation has been seen in the harvest locations on the New England continental shelf (Chase, 1992). Spatial variation in prey popu- lations is suspected to be the primary influence on these annual aggregations of bluefin tuna. Adaptations in the cir- culatory system of the bluefin tuna allow these fish to retain metabolic heat, facilitating the regulation of body temperature (Carey and Teal. 1969) and assisting in the efficient transfer of energy from consumed prey to fast growth rates and large body size. These adaptive features also allow bluefin tuna to make extensive migrations into cold-temperate waters in search of prey. Diet information is necessary to improve the understanding of sea- sonal movements of bluefin tuna and predator-prey relationships in the New England continental shelf region, and as a baseline for bioenergetic analyses. Information on the feeding habits of this economically valuable species and apex predator in the western North Atlantic Ocean is limited, and nearly absent for the seasonal feeding grounds where most U.S. Atlantic commercial catches occur. Previous food habit studies have shown that western North Atlantic bluefin tuna are opportunistic feeders on a wide variety of finfish, cephalo- pods, and crustaceans (Crane, 19.36; Krumholz. 1959; Dragovich. 1970; Ma- son, 1976; Holliday, 1978; Eggleston and Bochenek, 1990; Matthews et al.M. Pinkas (1971) reported similar results for Pacific bluefin tuna iThunnus thyn- nus orientalis). These bluefin tuna food habit studies either reported on small numbers of samples or were qualitative studies on samples distributed over a broad geographic range. Previous studies have not evaluated size relationships between bluefin tuna and their prey and the amount of food consumed. The present study tar- geted regions of the New England gi- ant tuna (>140 kg) fishery to quan- titatively analyze stomach contents of bluefin tuna among discrete sea- sonal feeding grounds and to investi- gate predator-prey size relationships for this species in the Gulf of Maine. Contribution 3 of the Massachusetts Divi- sion of Marine Fisheries, Gloucester. MA 01930. Matthews, F. D., D. M. Damkaer, L. W. Knapp, and B. B. Collette. 1977. Food of western North Atlantic tunas (Thun- nus) and lancetfish lAlepisaurus). Nat. Oceanic Atmos. Admm. Tech. Rep. NMFS SSRF-706, Washington. D.C., 19 p. Chase: Differencesin diet of Thunnus thynnus at seasonal feeding grounds off New England 169 42 N Materials and methods Study area Five fishery areas (Jeffreys Ledge, Stelhva- gen Bank, Cape Cod Bay, Great South Chan- nel, and South of Martha's Vineyard IFig. 11) were selected for their traditional associa- tion with the bluefin fishery and because geo- graphic and bathymetric differences among areas could result in distinct prey communi- ties. The northernmost area, Jeffreys Ledge, is a major bathymetric feature in the Gulf of Maine and is important for commercial fisheries from northern Massachusetts, New Hampshire, and southern Maine. Stellwagen Bank is a distinct bathymetric ridge located on the eastern boundary of Massachusetts Bay (DOC^) that provides close access for ports from Gloucester to Cape Cod. Cape Cod Bay is a large, relatively shallow Gulf of Maine bay that is semi-enclosed on three sides by the land mass of Cape Cod and the south shore of Massachusetts. The two southern areas. Great South Channel and the area south of Martha's Vineyard, are larger regions with less distinct bathymetric features and are separated by Nantucket Shoals. The Great South Channel covers a wide nearshore region running east of Cha- tham and Nantucket and is bordered by the slopes of Nantucket Shoals on the west. The area south of Martha's Vineyard is located on the continental shelf off southern New England and is distinguished from the other areas by its warmer water, predominance of smaller bluefin tuna (<50 kg), and the sea- sonal occurrence of other large pelagic fish, such as marlins and tropical tunas. Commercial catch records and trawl sur- vey indices of abundance indicate that At- lantic herring (Clupea harengus) and Atlan- tic mackerel (Scomber scombrus) are the principal pelagic fish species for all these study areas (Clark and Brown, 1976; and NOAA, 1998). Atlantic herring occur in lower relative abundance in the region south of New England than in northern regions. Jeffreys Ledge and Great South Channel are primary spawning locations for Atlantic her- ring. Catch records and trawl survey indices indicate that the following prey species have been abundant in the study areas: silver hake (Merluccius bilineahs), butterfish (Peprilus triacanthus), northern shortfin squid (lllex illece- brosus), and longfin inshore squid (Loligo pealeii). 70°W 2 DOC (Department of Commerce). 1991. Stellwagen Bank Na- tional Marine Sanctuary, Draft environmental impact statement/ management plan. U.S. Dep. Commerce (DOC», Nat. Oceanic Atmos. Adm., Sanctuaries and Reserves Division, Washington, D.C., 238 p. Figure 1 Map of the five bluefin tuna feeding grounds used as study areas ( * ) July- October 1988-92. Depths are given in fathoms. Sample collection and analysis Bluefin stomach samples were collected primarily from commercial landings at ports in Massachusetts during the 1988-92 seasons (July-October). Sportfishing tour- naments were a secondary source of stomach samples. Handgear landings (rod and reel, handline, and harpoon) were primarily collected in Gloucester and Cape Cod. Purse-seine landings were also a large source of samples, and were collected in Gloucester, New Bedford, and Cape Cod. A majority of stomach samples was collected during 1989 and 1990. A reduced number of samples after 1990 was influenced by the increasing practice of gutting blue- fin at sea to sustain a quality product for the sashimi export market. Most samples were collected at the docks and process- ing locations of commercial tuna buyers. The vessel cap- tain or tuna buyer was interviewed for location of catch. 170 Fishery Bulletin 100(2) The cui-ved fork length (CFL) and weight of the catch were recorded from the National Marine Fisheries Ser- vice (NMFS) tuna logbook. Stomachs were removed by cutting the esophagus above the pylorus and were stored on ice until analysis later that day or were frozen for anal- ysis at a later date. Stomach samples were typically re- moved from bluefin tuna the day of capture, except those from purse-seine landings, where the catch was often sub- merged in ice for one or two days prior to landing. Stomach contents were identified to the lowest possible taxon and weights and counts were made of individual prey species. Wet weights of prey were measured with Homs tubular scales (±5 g) after contents were rinsed through a standard testing sieve (2.00-mm mesh). Prey counts were made only when all individual prey items could be identified and counted for a given stomach sam- ple. Because bluefin tuna consume relatively large prey items and swallow prey whole, species identification was possible for nearly all contents. Skeletal remains were compared with the skeletons of known specimens to assist with identification. Fish contents that could not be identi- fied were categorized as "unidentified fish." Otoliths, squid beaks, and skeletal traces less than .5 g were rounded up to a minimum weight of 5 g. Stomach-contents data were analyzed by frequency of occurrence, percent composition by number, and percent composition by weight for each prey item (Hyslop 1980; Bowen 1986). Frequency of occurrence can indicate prey composition and availability, and number and weight per- centages can represent the quantity a prey item contrib- utes to a diet. "Stomach-contents biomass" refers to all prey in stomach contents, and "prey weight" refers to the weight of individual prey species. Stomach samples were assigned a status of "empty," "chum," "chum and prey," or "prey only." Chum refers to cut pieces of bait fish that fishermen use to attract bluefin. Both empty and chum stomach samples were eliminated from further data analysis. For "chum and prey" stomachs, chum weight was eliminated from the analysis, and prey weight was included because chum and natural prey were easily distinguished in bluefin tuna stomachs. Wlien the status of contents, or the veracity of catch location could not be re- solved, the samples were eliminated from the analysis. Statistical analysis Stomach-contents biomass data were analyzed to deter- mine differences among the five fishery areas. Prey weight data were tested for normality (Shapiro and Wilk, 1965) and equality of variance (BMDP. 1990). Stomach-contents biomass data were transformed to natural logarithms and tested for area differences by using the Brown-Forsythe test for unequal variances and the Welch test for painvise comparisons of areas (BMDP, 1990). The species composition of stomach contents were test- ed among areas with the Spearman rank correlation test, under the null hypothesis of no association between spe- cies composition and feeding area. The twelve most com- mon prey species were ranked according to frequency of occurrence for each location. The Spearman rank correla- tion test was applied by pairwise ranking for all locations, excluding missing cases. Bluefin tuna size was compared with prey length and stomach-contents biomass data by using the Pearson prod- uct moment correlation coefficient (Sokal and Rohlf, 1981) to test for significant associations between prey and pred- ator length, and to correlate bluefin tuna weight and stom- ach-contents biomass for all samples from the Gulf of Maine. The relationship between bluefin tuna size and food consumed was also evaluated by comparing the ratio of stomach-contents biomass and tuna weight (% kg wet weight of prey biomass/kg wet weight of tuna) to tuna length (cui-ved fork length) (Young et al., 1997). The area south of Martha's Vineyard area was excluded from tests on size relationships because of the limited sample num- bers of juvenile bluefin tuna collected there. Results A total of 819 bluefin tuna stomachs were analyzed during 1988-92 (Table 1) and 568 contained prey; empty stom- achs (206) and "chum only" samples (45) were eliminated from further analysis. Approximately equal quantities of the samples with prey came from purse-seine landings (273) and from rod and reel and handline landings (264). The fishing method used for the hook-and-line landings was recorded for 242 samples and was evenly divided between chumming (where chum was used as bait) (123) and trolling (119). Size composition of sampled bluefin tuna was similar for the four Gulf of Maine study areas; a large majority of fish were large, mature adults, esti- mated to be age 10 and older (Mather and Schuck, 1960). In contrast, nearly all fish sampled from the area south of Martha's Vineyard were small juveniles, ages 2-6. Tuna sampled from Cape Cod Bay had the largest average size <251 cm CFL, and 273 kg). " Prey composition All areas combined Stomach contents comprised at least 21 species of teleosts, two species of elasmobranchs, and at least nine species of invertebrates (Table 2). Stomach- contents biomass in terms of taxonomic composition was dominated by Osteichthyes (Fig. 2A). Of the invertebrates, only squid ("squid" refers to two species, Loligo pealei and lllex illecebrosus) were a consistent component of prey composition. Although squid accounted for only about 2'7c of the stomach-contents biomass. it was the second most common prey, occurring in a third of all stomach samples. The only other common invertebrate was the fig sponge iStiberites ficus) found in Cape Cod Bay. Despite the large diversity of prey items, few species made major contri- butions to overall prey composition. Sand lance (Amino- dytes ssp). squid, Atlantic herring, Atlantic mackerel, and bluefish {Pomatomus saltatrix) exceeded all other prey in terms of frequency of occurrence and accounted as a group for 88% of total stomach-contents biomass (Fig. 2B). In the Gulf of Maine areas, sand lance, and Atlantic herring were the major prey in the diet of bluefin tuna during Chase: Differencesin diet of Thunnus thynnus at seasonal feeding grounds off New England 171 Table 1 Summary of numtx'r of l)lui fin tuna stomach samples col ected at five study areas on thf New England continental shelf. July- October 1988-92. Also included are lengths (cui'ved fork length ICFLl, cm) and weights kg of sampled tuna. "Other" column refers to twelve samples with prey collected at nearshore fishing areas in the Gulf of Maine that were outside the five study areas Sample Jeffreys Stellwagen Cape Cod Great South South of category Ledge Bank Bay Channel Marths's Vineyard Other Total By year 1988 13 30 16 59 1989 13 33 28 88 13 1 176 1990 57 15 26 61 33 2 194 1991 34 8 13 34 2 8 99 1992 6 7 26 1 40 By condition of stomach Number with prey 123 93 109 183 48 12 568 Number with chum 19 17 6 1 2 45 Number empty 5 1 158 27 8 7 206 Total 147 111 273 210 57 21 819 Mean length of tuna 221 240 251 221 124 221 227 SD 32 35 19 38 30 37 44 Mean weight of tuna 186 243 273 196 36 205 215 SD 78 94 58 90 38 83 97 the study period. Twenty-one bluefin tuna stomach sam- ples were collected outside of the five study areas; mostly at two inshore locations north of Gloucester and south of Stellwagen Bank. Jeffreys Ledge Atlantic herring were the dominant prey in the 123 stomach samples from Jeffreys Ledge (Table 3). The frequency of occurrence for Atlantic herring ( 74*7^ ) was the second highest and the percentage of stomach-contents biomass iST^'i ) was the highest for any individual prey item among areas. Squid were in nearly half of all samples, but comprised less than 2% of stomach-contents biomass. Atlantic mackerel were the third most common prey for this region {32"r occurrence). All other prey species occurred inci- dentally. Menhaden iBrevoortia tyrannus) and pollock (Pol- lachius virens) were unique to samples from Jeffreys Ledge. Mean stomach-contents biomass (-2.0 kg) was the highest among study areas and few empty stomachs were found. Stellwagen Bank As at Jeffreys Ledge, a single species dominated the stomach-contentss from Stellwagen Bank. Sand lance were found in nearly 80% of 93 stomachs and accounted for nearly 70'^^ of the stomach-contents biomass (Table 3). Atlantic herring, squid, spiny dogfish (Squalits acanthias), Atlantic mackerel, and bluefish tuna were sec- ondary prey items, with frequencies of occurrence ranging from 9% to 14%. All other prey species found in stomach contents from Stellwagen Bank occurred incidentally, and there were no species from this area that were unique to the overall studv area. dicate a dominant pelagic prey but did include the common occuiTence of demersal prey. Six prey species were only found in Cape Cod Bay. Squid occurred most frequently but accounted for only 2% of stomach-contents biomass biomass (Table 3). The fig sponge was the top prey in terms of per- centage by weight (27%). The diet of bluefin tuna caught in Cape Cod Bay displayed the most diversity among study areas. A total of 16 prey species were identified, of which eight were demersal species. Three species of flounder were identified, representing 9% of the stomach-contents biomass. The occurrence of bluefish tuna as prey in Cape Cod Bay (25%) was the highest among study areas. The amount of food in Cape Cod Bay stomach samples was the lowest for the four Gulf of Maine locations; 60% of the stomachs collected from this area were empty. Great South Channel A large number of stomach sam- ples with prey ( 183 ) were collected from the Great South Channel during 1989-91. The most abundant prey was sand lance, occurring in 62% of the stomach samples and accounting for 28% of the stomach-contents biomass. Atlantic herring was also important, with a frequency of occurrence of 27% and stomach-contents biomass of 48%. As with Stellwagen Bank and Jeffreys Ledge, squid was an important secondary prey item, and bluefish and Atlantic mackerel were secondary prey of lesser impor- tance. Four species were unique to Great South Channel samples: shrimp i Panda! us spp.), finger sponge (Haliclona oculata), silverstripe halfbeak [Hyporhaniphus unifascia- tus), and Atlantic cod (Gadus morhua). Cape Cod Bay Unlike diet composition for the other study areas, diet composition for Cape Cod Bay did not in- South of Martha's Vineyard Only 48 stomachs with prey were analyzed from the area south of Martha's Vineyard. 172 Fishery Bulletin 100(2) Table 2 Stomach contents of bluefin tuna caught off New England during 1988-92. Prey species are combined for all five study areas, including 12 samples from outside of the study areas. Percent frequency of occurrence C^'t O) and percent weight C^^'r W I data were determined from the 568 stomach samples that contained prey. Prey species Sand lance Atlantic herring Atlantic mackerel Bluefish Butterfish Silver hake Windowpane Hake Winter flounder Atlantic menhaden Sea horse Atlantic cod Fourspot flounder American plaice Wrymouth Pollock Filefish Halflieak Longhorn sculpin Unidentified fish Spiny dogfish Skate Skate egg case Squid Octopus Shrimp Lobster Argonaut Crab Salp Fig sponge Finger sponge Frequency of occurrence Total weight (gl Ammodytes (spp. ) Cliipca harengus Scomber scui/ibrus Pom a torn ii s saltatrix Pi-prihis tnacanthiis Mcrlucciua bilineans Scoptha/miiK aqiiosus Urophycia (spp.) P.vuclopleuroncctes americanus Brevoortia ty ran mis Hippocampus erect us Gadus morhua Paralichthys oblongus Hippoglussoidcs platessoides Ciyptacanthodes masculatiis Pullachius virens Monocanlhus luspidus Hyporhamphus unifasciatus Myoxocephalus octodecimspinosus Teleostei Squahis acanthias Raja (spp. I Raja ( spp. 1 Cephalopoda Cephalopoda Pandalus (spp.) Homarus americanus Argonauta argo Cancer (spp.) Salpidae Suberites ficus Haliclona oculata Stomachs with chum and prey Stomachs with chum only Empty stomachs Total stomachs with prey Total stomachs sampled 194 167 108 55 21 16 12 11 8 7 7 6 2 2 2 1 1 1 1 34 13 13 3 186 1 5 2 2 2 2 30 1 95 45 206 568 819 YcO % W 128,240 34.2 22.6 299,550 29.4 .52.8 18,930 19.0 3.3 40,830 9.7 7.2 1685 3.7 0.3 1490 2.8 0.3 1890 2.1 0.3 2500 1.9 0.4 1820 1.4 0.3 5500 1.2 1.0 40 1.2 <0.05 22,840 1.1 4.0 370 0.4 0.1 310 0.4 0.1 1300 0.4 0.2 940 0.2 0.2 30 0.2 <0.05 120 0.2 <0.05 280 0.2 <0.05 605 6.0 0.1 10,490 2.3 1.9 4670 2.3 0.8 20 0.5 <0.05 10,835 32.8 1.9 30 0.2 <0.05 25 0.9 <0.05 230 0.4 <0.05 25 0.4 <0.05 15 0,4 <0.05 60 0.4 <0.05 11,510 5.3 2.0 50 0.2 <0.05 107,430 (chum) 49,040 (chum) 567,230 723,700 Giant bluefin were scarce in this area during the study period and numerous samples could not be used because the bluefin tuna were caught in association with trawler fleet discards. This area is distinguished from the others by a predominance of juvenile bluefin tuna . Four prey species were unique to this area: lined sea horse (Hip- pocampus erectus), argonaut {Argonauta argo), planehead filefish (Monocauthus hispiduxl, and octopus (Cephalop- oda ). The filefish and seahorse were associated with bluefin tuna foraging at sargassum weed communities. Squid and Atlantic mackerel were the two most important prey for this area. The frequency of occurrence and percentage of prey weight for mackerel and butterfish were the highest among study areas. Combined stomach-contents biomass for squid, mackerel, and butterfish represented nearly 809;^ of the stomach contents for this area. Chase: Differencesin diet of Thunnus thynnus at seasonal feeding grounds off New England 173 Comparison of study areas The top 12 prey items, overall, were lanked i'or each area by frequency of occur- rence, and area differences were tested with Spearman rank correhition. Of the ten painvise comparisons, only Stelhvagen Bank and Great South Cliannel showed a significant association in the ranking of prey items (;=0.98. P<0.02). attributable to a high rank of sand lance and a similar ranking of squid. Atlantic herring. and Atlantic mackerel for both areas. Stomach-contents biomass Comparison of study areas Large differences in stom- ach-contents biomass were found: Jeffreys Ledge aver- aged nearly 2 kg, followed by approximately 1 kg for Stellwagen Bank and Great South Channel, and less than 0.5 kg for the remaining areas. Stomach-con- tents biomass data from the five areas were positively skewed and heteroscedastic. The natural logarithm- transformed data for Stellwagen Bank, Cape Cod Bay, and Great South Channel were normal (Wilk-Shapiro test, P>0.05). Transformed biomass data for the other two areas still differed significantly from normality. Transformation of biomass data reduced the inequal- ity of variances, but significant differences (Levenes test, P<0.05) remained, which precluded use of analy- sis of variance. The Brown-Forsythe test for unequal variances showed a significant effect of area on stom- ach-contents biomass (P<0.0001). Painvise compari- sons of the equality of prey weight means were made with the Welch test (Bonferroni corrected significance level of P=0.005). All area paii-wise comparisons of stomach-contents biomass were significantly different except that between Cape Cod Bay and Great South Channel(P=0.816). The similarity in the amount of food found at Cape Cod Bay and Great South Channel was also indicated by the geometric mean of stomach-contents biomass (Table 4). The arithmetic mean of stomach-contents biomass was higher at the Great South Channel than Cape Cod Bay, but this value was biased by a few samples with large amounts of prey. The stomach- contents biomass range for Cape Cod Bay did not exceed 3.0 kg, in contrast to the wider range for the Great South Channel up to 16.0 kg, including 13 samples over 3.0 kg. The use geometric means reduced the bias of skew- ness and indicated that samples from Jeffreys Ledge con- tained the most prey and that the amounts declined mov- ing southward. Effect of tuna size Increased stomach-contents biomass with increasing body size (due to increasing gape and stomach size) was not clearly demonstrated from these data. Correlation of stomach-contents biomass to bluefin weight was not significant for Gulf of Maine samples. A scatterplot of data for the Gulf of Maine samples revealed that a large majority of the samples contained small amounts of food, regardless of body size (Fig. 3). A size- related trend was observed in that only very large bluefin tuna (>250 kg) contained more that 6 kg of food. Size Osteichttiyes 93% Elasmobranchii 3% Cephalopoda 2% Ponfera 2% B other invertebrates sand lance 2% 23% ^^ I f—.. other fish squid 2% Atl mackerel 3% Atlantic herring 53% Figure 2 Percent prey weight composition in stomach contents of bluefin tuna caught off New England during 1988-92 ln=568). The taxo- nomic composition (A) includes 0.05'^ Crustacea and O.Ol^r Uro- chordata. The comparison by major prey type (B) comprises the five most common prey and all remaining fish and invertebrates. effects were also compared by using the ratio of stomach- contents biomass and tuna weight C* kg/kg, wet weight) to tuna length. The percentage of food to body mass declined with increasing bluefin tuna length (Fig. 4). Food obsei-ved in 120-149 cm bluefin tuna averaged over 1** of their body weight, and declined to approximately 0.5% for bluefin tuna over 230 cm. The high ratios primarily resulted from large meals of Atlantic herring or sand lance. Cape Cod Bay ratios were consistently the lowest among the four areas, ranging from 0. 1 to 0.2%. Characteristics of prey species Prey size Prey size was evaluated for 190 stomach sam- ples that contained measurable prey from the four Gulf of Maine areas. A total of 1866 prey items were measured, of which 95% were either sand lance, Atlantic herring, squid, or Atlantic mackerel (Table 5). A significant positive cor- 174 Fishery Bulletin 100(2) Table 3 Stomach contents of bluefin tuna caught off New England durii g 1988-1992, listed by the five study areas Percent frequency of | occurrence (% 0) and percent by tveightC/c W) were calculated for each prey species. South of Prey species Jeffreys Ledge Stellwagen Bank Cape Cod Bay Great South Channel Martha's Vineyard 'i ^■, W '-r O ^,\\ KiO ^i W 't r^w ^'f C^ W Sand lance 3.3 1.8 79.6 69.3 62.3 28.3 Atlantic herring 74.0 87.2 14.0 6.0 8.3 3.1 27.3 48.4 2.1 2.5 Atlantic mackerel 3L7 2.0 10.8 2.6 18.3 12.7 8.2 0.6 33.3 56.2 Blucfish 7.3 3.5 8.6 17.5 24.8 14.7 4.9 5.7 Butterfish 2.4 <0.05 11 <0.05 5.5 1.5 1.6 0.2 16.7 10.4 Silver hake 3.3 0.2 2.2 0.3 10.4 2.9 Window-pane 10.1 4.1 Hake 4.1 0.7 1.1 0.2 0,9 0.2 8.3 9.9 Winter flounder 6.4 3.9 Atlantic menhaden 5.7 2.3 Sea horse 14.6 0.7 Atlantic cod 3.3 13.5 Fourspot flounder 1.8 0.9 American plaice 0.8 0.1 1.1 <0.05 Wrymouth 1.8 0.9 Pollock 0.8 0.4 Filefish 2.1 0.5 Halfteak 0.5 0.1 Longhorn sculpni Unidentified fish 4.9 <0.05 3.2 <0.05 2.8 0.1 8.7 0.1 12.5 2.9 Spiny dogfish 10.8 3.7 2.8 16.6 Skate 9.2 9.3 0.5 0.3 Skate egg case 1.1 <0.05 1.8 <0.05 Squid 48.8 L8 14.0 0.5 35.8 1.8 22.4 2.5 60.4 12.9 Octopus 2.1 0.5 Shrimp 2.7 <0.05 Lobster 1.8 0.5 Argonaut 4.2 0.4 Crab 0.9 <0.05 0.5 <0.05 Salp 1.8 0.1 Fig sponge 27.5 27.2 Fmger sponge n n 0.5 <0.05 relation (P<0.001) between prey and bluefin tuna lengths was found for all prey-size data (Fig. 5). Despite this cor- relation, there appeared to be httle association between predator and prey length for most species. The positive correlation was influenced by 29 larger prey items (>40 cm) all consumed by bluefin tuna larger than 230 cm. The larger prey were spiny dogfish, skate, bluefish, or Atlantic cod. Size data on the four most common prey species pro- vided evidence of the consistency of prey size across a wide range of bluefin lengths in the Gulf of Maine. A signifi- cant positive size relationship was found for sand lance and Atlantic mackerel (both P<0.001), although both may have contained biases. The relationship for sand lance was influenced by smaller bluefin tuna that ate smaller sand lance (/-test, P<0.001) in Great South Channel than at Stellwagen Bank. All mackerel were YOY or age-1, except for two large mackerel consumed by larger bluefin tuna (>200 cm). The predator-prey size relationships for Atlantic herring (P=0.36) and squid (P=0.16) were not sig- nificant. A large majority (939f ) of Atlantic herring prey were 18-27 cm in length, which corresponds to age-2 to age-4 cohorts for the western Gulf of Maine (Penttila et Chase: Differencesin diet of Thunnus thynnus at seasonal feeding grounds off New England 175 Table 4 Summary statistics on stomacli-contonts liioniass 1^,' w( Eiifjland, 1988-92. The geometric mean, confidence into ral logarithms. ■t weight 1 from bluefin tuna caught at five seasonal feeding areas ofTNew n-als (CI), and coefficient of variation ((-V) are backtransformed from natu- Statistic All areas JefTreys Ledge Stellwagen Bank Cape Cod Bay Great South Channel South of Martha's Vineyard Number of stomach samples with prey 568 123 93 109 183 48 Minimum stomach biomass 5 5 5 5 5 5 Maximum stomach biomass 16100 5800 6240 2900 16100 1200 Arithmetic mean 999 1957 1012 389 925 124 Geometric mean 214 8.32 438 133 126 37 Lower 95^f CI 180 608 324 96 91 23 Upper 95% CI 255 1140 593 184 174 60 CV 39 26 24 35 46 44 16 1 m 14 ^£. cn 12 - m H o 100 .Q C 80 - CD c- o CJ 60 a en F 40 n fl 2 - 00- al., 1989). and the youngest were age 2. There were no young-of-the-year (YOY) sand lance. Most sand lance from Great South Channel samples were age 2. in contrast to age 3 or age 4 at Stellwagen Bank (Weston et al, 1979). Most squid were YOY or age 1. Numbers of prey species Few prey species were found in large numbers in a given stom- ach sample. Only sand lance, squid, Atlantic herring, and Atlantic mackerel had sample counts higher than 20 individual fish (Table 6). Data on prey numbers are limited because prey counts were made for only 208 samples. Many samples with large numbers of well-digested sand lance were difficult to count. From this subsample, the mean number of sand lance per stomach was 159, much higher than the next highest mean of 19 for Atlantic herring. Squid and mackerel commonly occurred as prey, although typically only a few individuals were found per stomach. Weight of prey species Sand lance and Atlantic herring (combined) accounted for 75'7i of the total stomach-contents biomass for all areas combined. The next highest species was bluefish at T^i. Despite a high frequency of occurrence, squid accounted a low percentage of overall biomass (2%) and a mean stomach prey weight of only 58 g. The highest mean prey weight was 1794 g for Atlantic herring. Only three other prey items averaged over 500 g in stomach-con- tent weight: spiny dogfish (807 g), bluefish (742 g), and sand lance (661 g). With few exceptions, stomachs that were full or near full, contained only sand lance or Atlantic herring. Only five stomachs contained over 10 kg of prey contents: three with Atlantic herring, and one each with sand lance and Atlantic cod ( 16.0 kg, the largest meal obsei-ved). In summary, five prey items occurred in frequency and mass to be considered important dietary components of n=520, r=0 048. P=0,276 40 cm ) by large bluefin tuna (>230 cm). Prey this large were not common in stom- achs and were probably limited by mouth and esophagus gape. Otherwise, the sizes of prey were consistent across the range of bluefin tuna sampled in the Gulf of Maine. The findings on prey size and numbers (per stomach) provide evidence of three selective foraging strategies used by bluefin tuna in the Gulf of Maine. Feeding on indi- vidual, fast-swimming pelagic prey was evident from con- sumed bluefish and Atlantic mackerel that are abundant pelagic species in the Gulf of Maine, but which occurred much less frequently in stomach contents and with few in- dividuals per stomach. Ram feeding (swimming through a dense school prey with mouth open) of small prey may have resulted in high average number of sand lance found in stomachs and contributed to the similar prey sizes in most sizes of bluefin tuna. Bluefin tuna in Cape Cod Bay displayed a different foraging behavior, selecting larger, individual demersal prey. The variation in prey compo- sition and different feeding strategies is consistent with previous descriptions of bluefin tuna as an opportunistic predator However, the dominance of sand lance and At- lantic herring in the Gulf of Maine diet suggests a depen- dence on these species as an optimal energy source. Trophic influences on bluefin tuna distribution Changes in biomass and spatial availability of forage pop- ulations may affect the distribution of bluefin tuna on the New England continental shelf Major changes in the prey community of the Gulf of Maine have occurred in recent decades. After Atlantic mackerel and Atlantic her- ring stocks off New England were severely overharvested in the 1960s and 1970s (NOAA, 1998), sand lance popula- tions increased, presumably as a result of decreased pre- dation and competition for food (Sherman et al., 1981). Atlantic herring and Atlantic mackerel stocks off New England increased steadily during the 1980s and 1990s ( NEFSC, 1998). By the mid-1990s, the U.S. Atlantic coastal spawning stock biomass for these species increased to the highest levels on record (NEFSC, 1998). Commercial bluefin tuna catches have increased in areas where Atlantic herring abundance has increased (western Gulf of Maine and Great South Channel) and diminished at traditional areas south of the Gulf of Maine (Chase, 1992). The northward shift in bluefin tuna distribution on the New England continental shelf may be influenced by improved foraging opportunities on Atlantic herring in the Gulf of Maine. I suspect that the timing of bluefin tuna migi-ations to the New England continental shelf are as- sociated with seasonal spawning and feeding aggregations of Atlantic herring. Sand lance populations appear to be an important influence on bluefin tuna feeding migrations. Chase: Diffeiencesin diet of Thunnus thyninis at seasonal feeding grounds off New England 179 but they occur on a limited spatial scale in relation to At- lantic herring in the (nilf of Maine, and changes in tlieir population abundance are not well documented. Atlantic herring and sand lance are also important in the diet and distribution of marine mammals in the Gulf of Maine (Payne et al., 1990; Weinrich et al., 1997; Gannon ct al., 1998). Changes in bluefin tuna stock composition and the long- term impact of small-mesh trawling gear on commercially important prey items (squid, silver hake, and butterfish) in southern New England waters and Mid-Atlantic areas where bluefin tuna catches have diminished are two poten- tially confounding factors in this discussion. In the man- agement of Atlantic bluefin tuna, care should be taken to recognize that the fluctuations in the regional abundance of this species can be influenced by more than changes in stock structure. Changes in major prey populations, pro- duced either by environmental features or by fishery prac- tices, can have a profound effect on regional aggregations of bluefin tuna. Investigations should be conducted on in- troducing forage-base information into the interpretation of catch-per-unit-of-effort indices of abundance for Atlan- tic bluefin tuna populations. There is also a need for future research to improve our knowledge on the bioenergetics of this warm-bodied tuna and its associated relationships with prey species. Acknowledgments This study was conducted by the Massachusetts Division of Marine Fisheries (DMF), Commonwealth of Massachu- setts, and funded by the Federal Aid in Sportfish Res- toration Program. I would like to thank the DMF and National Marine Fisheries Service staff who assisted with field sampling and the technical review of this project. I am especially thankful to Steve Cadrin for his review of the manuscript and assistance with the statistical analy- ses. I am very grateful to the many fishermen and tuna buyers who contributed stomach samples and catch infor- mation. Special thanks are due to the following for the contribution of large numbers of samples: Bill Raymond, Mark Godfrey, Rodman Sykes, and the crew of FV Debra Lynn, Great Circle Fisheries, Ralboray, Crocker and Sons, Canal Marine, and Atlantic Coast Fisheries. Literature cited Amundsen, P. A., and A. Klemetsen. 1986. Within-sample variabilities in stomach contents weight of fish- implications for field studies of consump- tion rates. In Contemporary studies on fish feeding (C. A. Simenstad and G. M. Cailliet. eds.), p. .307-314. Dr. W. Junk Publishers. Boston. MA. Bigelow, H. B., and W. C. Schroeder. 19,53. Fishes of the Gulf of Maine. U.S. Fish and Wildl. Serv. Bull. 74, vol. 53, 577 p. BMDP (Biomedical Dynamic Programs). 1990. BMDP statistical software manual, vol. I. UnivCal. Press, Berkeley, CA, 617 p. Bowen, S. H. 1986. Quantitative description of the diet. In Fisheries tech- niques (L. A. Nielson, and I). I,. Johnson, eds.), p. 325-336. Am. Fish. Soc, Bethesda, MD. Butler, M. J., and J. M. Mason Jr 1978. Behavioral studies on impounded bluefin tuna. Int. Comm. Conserv. Atl. Tunas Ooll. Vol. Sci. Pap. 7(2);379-381. Carey, F G., and J. M. Teal. 1969. Regulation of body temperature by the bluefin tuna. Comp. Biochem. Physiol. 28:20.5-213. Chase, B. C. 1992. A profile of changes in the Massachusetts bluefin tuna fishery: 1928-1990. M.S. thesis, Univ. Rhode Island, Kingston, RI, 266 p. Clark, S. H., and B. E. Brown. 1976. Changes in biomass of finfishes and squids from the Gulf of Maine to Cape Hatteras, 1963-1974, as determined from research vessel survey data. Fish. Bull. 75:1-21. Crane, J. 1936. Notes on the biology and ecology of giant tuna Thun- nus thynnus, L., observed at Portland, Maine. Zoologica 212:207-212. Dragovich, A. 1970. The food of bluefin tuna tThunnus thynnus) m the western North Atlantic Ocean. Trans. Am. Fish. Soc. 99: 726-731. Eggleston, D. B., and E. Bochenek. 1990. Stomach contents and parasite infestation of school bluefin tuna Thunnus thynnus collected from the Middle Atlantic Bight, Virginia. Fi.sh. Bull. 88:389-395. Gannon, D. P., J. E. Craddock, and A. J. Read. 1998. Autumn food habits of harbor porpoises, Phocoena phocoena, in the Gulf of Maine. Fish. Bull. 96:428-437. Hodgson, J. R., S. R. Carpenter, and A. P. Gripentrop. 1989. Effect of sampling frequency on intersample variance and food consumption estimates of nonpiscivorous large- mouth bass. Trans. Am. Fish. Soc. 118:11-19. Holhday. M. 1978. Food of Atlantic bluefin tuna, Thunnus thynnus (L.), from the coastal waters of North Carolina to Massachusetts. M.S. thesis. Long Island Univ., Long Island, NY, 31 p. Hyslop, E. J. 1980. Stomach contents analysis — a review of methods and their application. J. Fish. Biol. 17:411-429. Ki'umholz, L. A. 1959. Stomach contents and organ weights of some bluefin tuna, Thunnus thynnus (Linnaeus), near Bimini, Bahamas. Zoologica 44:127-131. Mason, J. M. 1976. Food of small, northwestern Atlantic bluefin tuna, Thunnus thynnus (L.) as ascertained through stomach con- tent analysis. M.S. thesis, Univ. of Rhode Island, Kings- ton, RI, 31 p. Mather, F. J, III, and H. A. Schuck. 1960. Growth of bluefin tuna of the western north Atlantic. Fish. Bull. Fish Wildl. Serv. 179:39-52. NEFSC (Northeast Fisheries Science Center). 1998. A report of the 27"' Northeast regional stock assess- ment workshop. Stock Assess. Rev. Comm. (SARC ), consen- sus summary of assessments. Lab. Ref Doc 98-15, 350 p. [Available from Northeast Fisheries Science Center ( NEFSC), Woods Hole, Woods Hole, MA.] NMFS (National Marine Fisheries Service). 1995. Final environmental impact statement for a regulatory amendment for the western Atlantic bluefin tuna fishery. 180 Fishery Bulletin 100(2) National Marine Fisheries Service i NMFS ), Silver Springs, MD, 142 p. 1999. Final fishery management plan for Atlantic tunas, swordfish. and sharks. In Chapter 3: Rebuilding and main- tainmg HMS fisheries, p. 1-.321. (Available from NMFS, Silver Springs, MD.l NOAA (National Oceanic and Atmospheric Administration). 1998. Status of fishery resources off the northeastern United States for 1998. U.S. Dep. Commer, NOAA Tech. Memo. NMFS-NE-115, 149 p. Payne, P. M., D. N. Wiley, S. B. Young, S. Pittman, P. J, Clapham, and J. W. Jossi. 1990. Recent fluctuations in the abundance of baleen whales in the southern Gulf of Maine in relation to changes in selected prey Fish. Bull. 88:687-696. Penttila, J. A., G. A. Nelson, and J. M. Burnett III. 1989. Guidelines for estimating lengths at age for 18 northwest Atlantic finfish and shellfish species. U.S. Dep. Commen, NOAA Toch. Memo., NMFS-F/NEC-66, 39 p. Pinkas, L. 1971. Bluefin tuna food habits. //; Food habits of albacore. bluefin tuna, and bonito in California waters (L. Pinkas, M. S. Ohphant, and I. K Iverson. eds. i. p. 47-63. Cal. Fish Game Fish. Bull. 1.52:1-105. Shapiro, S. S., and M. B. Wilk. 1965. An analysis of variance test for normality (complete samples). Biometrika 52: 591-611. Sherman. K. C, C. Jones, L. Sullivan, W. Smith, P. Berrein, and L. Ejsymont. 1981 . Congi'uent shifts in sand eel abundance in western and eastern north Atlantic ecosystems. Nature 291:486-489. Smagula, C. M., and I. R. Adelman. 1982. Day-to-day variation in food consumption by large- mouth bass. Trans. Am. Fish. Soc. 111:543-548. Sokal, R. R., and F J. Rohlf 1981. Biometry, 2nd ed. W. H. Freeman and Co., New York, NY, 859 p. Weinrich, M., M. Martin, R. Griffiths, J. Bove, and M. Schilling. 1997. A shift in distribution of humpback whales, Megap- tera novaeangliae. in response to prey in the southern Gulf of Maine. Fish. Bull. 95:826-836. Weston, D. T, K. J. Abernethy, L. E. Meller, and B. A. Rogers. 1979. Some aspects of biology of the American sand lance, Ainnuidytes amcncanua. Trans. Am. Fish. Soc. 108:328- 331. Young, J. W., T. D. Lamb, D. L. Russel, W. Bradford, and A. W. Whitelaw. 1997. Feeding ecology and interannual variations in diet of southern bluefin tuna, Thunnus maccoyii, in relation to coastal and oceanic waters off eastern Tasmania, Australia. Environ. Biol. Fi.sh. .50:275-291. 181 Abstract-A total of 42,445 American lobsters iWomaru.s americanus) wore tagged in thirty-one sites throughout the southwestern Gulf of St. Lawrence betw^een 1980 and 1997. Results from the recapture of 8503 tagged lobsters showed small distances traveled be- tween the release and the recapture position for animals ranging in size from 51 to 152 mm carapace length. The average distance traveled ranged from 2 km in parts of Bale des Chal- eurs and western Cape Breton to 19 km in central Northumberland Strait. Lob- sters moved generally along the shore (939J of the dispersion was in areas between the shore and the 20-m bathy- metric contour). As a result, lobsters traveled longer distances in sites char- acterized by a gradually sloping bottom where the distance between the shore and the 20-m contour line was exten- sive in contrast to areas characterized by rapidly changing depths and by a relatively small amount of habitat shal- lower than 20 m. In the majority of sites (14 of 19) there was no significant difference between males and females in the average distance they traveled. In four of the five sites females moved farther than males. In general, the average distance traveled by berried females was shorter than that traveled by males or nonberried females. No relationship was obsen.'ed between the distance traveled and the size of the animal. There was no strong evidence of a relationship between the average distance traveled and the number of days at liberty. In general, lobsters in the southwestern Gulf of St. Lawrence traveled short distances and dispersion was restricted to the nearshore habi- tat. Further, the distance traveled was not correlated to size, sex, or years at large. These findings show that there is little interaction between American lobsters from different fishing areas at the benthic level and that American lobster movements should have mini- mal consequences for management of the species in the southwestern Gulf of St. Lawrence. Movement of American lobster (Homorus americanus) in the southwestern Gulf of St. Lawrence Michel Comeau Fernand Savoie Department of Fisheries and Oceans 343 University Ave. Moncton, New Brunswick, Canada EIC 9B6 E-mail address (lor M Comeau) comeaumm'dfo mpo gcca Manuscript accepted 14 September 2001. Fish. Bull. 100:181-192 (2002). The American lobster (Homarus ameri- canus Milne-Edwards, 1837) fishery in the southwestern Gulf of St. Lawrence (GSL) and for the entire Canadian Mari- time Provinces has become the most eco- nomically important fishery of coastal communities. Consequently, there is increasing interest by fishermen and the fishing industry to better under- stand the biology of the species and factors that may play a role in the fluctuations of landings, including pos- sible lobster movements between lob- ster fishing areas (LFAs). Fishermen are particularly concerned by lobster movements because the minimal legal size increased in different LFAs in the southwestern GSL throughout the 1980s and 1990s, and they want to know whether lobsters returned at sea in a given area could be recaptured elsewhere. Several tagging projects have been conducted in the past to study lobster movements in the GSL (Table 1). These tagging projects, initiated in the 1930s by Templeman ( 1935), showed that the average distance traveled by lobsters in the GSL was generally less than 15 km and that very few animals traveled up to 70 km (for review see Stasko, 1980; Lawton and Lavalli, 1995). Other tag- ging studies conducted in inshore wa- ters outside the GSL in Nova Scotia (Wilder, 1974; Campbell, 1982, 1989; Campbell and Stasko, 1985; Miller et al., 1989; Tremblay et al, 1998), Bay of Fundy (Campbell, 1986; Campbell andStasko, 1986), Maine (Cooper, 1970; Cooper et al., 1975; Krouse, 1981), New Hampshire (Watson et al, 19991. Mas- sachusetts (Karnofsky et al., 1989) and Rhode Island (Fogarty et al., 1980) have also shown that lobster move- ments were generally similar (4 to 18 km) to those from the GSL. However, long-distance movements of more than 90 km for up to 20% of the animals have also been observed for lobsters tagged inshore (Dow, 1974; Fogarty et al., 1980; Campbell and Stasko, 1985, 1986; Campbell, 1989; Robichaud and Law- ton, 1997); the farthest distance trav- eled reported was 798 km (Campbell and Stasko, 1986). These long distances traveled are more similar to those re- ported for offshore lobsters tagged on the continental shelf and over the off- shore deep canyons (Saila and Flowers, 1968; Cooper and Uzmann, 1971; Uz- mann et al, 1977; Fogarty et al., 1980; Campbell et al., 1984; Campbell and Stasko, 1985). Movements of more than 70 km have itever been reported for lob- sters tagged in the southwestern GSL. Since 1980, forty-six tagging studies have been conducted throughout the southwestern GSL, mostly in areas where information on lobster movements has been unavailable. These tagging studies have covered fishing grounds characterized by a flat bottom and hav- ing a relatively smooth transition from shore to 30 m and a narrow habitat close to shore where changes in depths occur over a relatively short distance. The purpose of our study was to in- vestigate the benthic movement of lob- sters tagged in different locations with- in the southwestern GSL by comparing the distance traveled, number of days at liberty, and size and sex of lobsters. 182 Fishery Bulletin 100(2) Materials and methods The recapture position was documented for 8503 lobsters, ranging in size from 51 to 152 mm carapace length (CL), from a total of 42,445 tagged and released during forty-six Table 1 Locations and authors of lob ster tagging studies on move- | ments of American lobster iHoma rus americanus) con- ducted m the Gulf of St. Lawrence. Location Authors and date Various locations in Gulf of St. Lawrence Wilder (1974) West coast of Newfoundland Templeman (1940) Magdalen Islands (Quebec) Templeman ( 1935 ) Bergeron (1967) Munro and Therriauh(1983) Gaspe Peninsula (Quebec) Corrivault(1948) Malpeque ( Prince Edward Is land) Templeman ( 1935) Western Prince Edward Island Wilder (1963) Northumberland Strait Templeman ( 1935 ) Wilder (1963) tagging studies in thirty-one sites throughout the south- western GSL (Fig. 1) between 1980 and 1997 (Table 2). The animals were captured by traps and tagged between July and November after the commercial fishing seasons (May and June in LFAs 23, 24, 26A, and 26B, and from mid-August to mid-October in LFA 25). The CL, measured to the nearest mm, sex of each animal, and presence or absence of eggs under the female abdomen were recorded. Prior to 1989, all animals were tagged with orange sphy- rion anchor tags, and blue streamer tags were used after 1992. Between 1989 and 1992 inclusively, both type of tags were used. A description of the tags and the tagging tech- nique are presented in Moriyasu et al. ( 1995) and Comeau et al.(1998). The positions of recaptured lobsters came from fisher- men. Because all tagging projects were undertaken after each commercial fishing season, no lobsters were recap- tured during the same year of their tagging. To increase fishermen's participation in reporting tagged animals, a major awareness campaign was conducted. Prior to 1993, representatives of the Department of Fisheries and Oceans (DFO) were present at each wharf to measure and collect information on tagged lobsters that were recap- tured. Beginning in 1993, letters were sent to all lobster fishermen in regions where tagging projects were conduct- ed; in these letters the tagging project was described, in- Gulf of St. Lawrence Figure 1 Locations (*) of American lobster {Hnniaruf; amencanus) tagging studies conducted in the southwestern Gulf of St. Lawrence between 1980 and 1997. The names of each site are presented in Table 2. The lobster fishing areas (LFAs) are indicated on the map. Comeau and Savoie: Movement of Homaivs amencanus in the southwestern Gulf of St Lawrence 183 structions given for returninfi lobster tag information, and the cooperation of the fishermen was sought (Comeau et al., 1998). Similar information was posted at wharves. For each tagged lobster, fishermen were asked to record date of capture, tag number, position of capture, and depth. They were then asked to freeze the animal with the tag still attached and contact an information collection center, where they could leave their names, addresses, and tele- phone numbers. A DFO representative collected tagged lobsters for measurement and reimbursed fishermen ac- cording to the market value of the recovered lobsters. No reward was issued. Fishermen also had the option to bring the lobsters and the information to a fisherman-represen- tative. If a tagged lobster under the legal size or a tagged berried female was captured, fishermen were asked to re- cord the above information and release the lobster to the water with the tag still attached. The distance traveled by each recaptured animal was calculated as the linear distance between release and re- capture positions. The Kruskal-Wallis and Mann-Whitney (t/-test) tests were used to compare the average distance traveled for males, females, and berried females for ani- Table 2 Average distances traveled by American lobsters tagged in the southwestern Gulf of St. Lawrence between 1980 and 1997. The number of each site corresponds to its geographical position on the map in Figure 1. n = number of recaptured lobster with tags indicating release location for which distance traveled to recapture site could be calculated. Average carapace Carapace Average distance length ±SD length range traveled ±SD Tagging site Year of tagging n (mml (mml (km) 1 Belledune 1980 604 70.5 ±4.9 50-115 7.4 +6.5 2 Pointe-Verte 1996 26 78.2 ±6.4 71-90 6.1 +4.0 3 Stonehaven 1994-97 618 73.2 ±5.4 53-111 2.4 ±3.0 4 Anse-Bleue 1994 290 73.5 ±4.8 54-91 3.5 ±6.3 5 Caraquet 1993- 97 1416 74.2 ±9.2 55-133 8.0 ±8.8 6 Petit-Shippagan 1994 134 76.2 ±10.4 54-116 8.7 ±8.8 7 Miscou 1995 58 69.9 ±8.2 66-117 4.5 ±6.5 8 Le Goulet 1996 101 73.8 ±6.2 59-98 5.3 ±3.5 9 Val Comeau 1985. 1996-97 519 76.3 ±9.1 57-117 7.0 ±6.5 10 Neguac 1996 74 78.0 ±8.7 65-105 14.2 ±8.2 11 Escuminac 1997 17 73.9 ±5.2 66-88 9.6 ±9.9 12 Seacow Pond 1996 62 71.1 ±6.0 61-94 5.3 ±7.8 13 Alberton 1996 62 74.0 ±6.9 65-97 5.7 +5.9 14 Malpeque 1989 492 68.1 ±7.2 51-100 10.1 ±7.0 15 Tracadie 1984 27 68.1 ±12.4 55-111 10.4 ±10.1 16 North Lake 1996 78 72.8 ±5.0 61-89 2.5 ±3.4 17 Pointe-Sapin 1996 22 72.5 ±5.0 65-83 6.7 ±4.1 18 Kouchibouguac 1997 53 67.7 ±4.3 60-74 8.0 ±7.3 19 Cap-Pele 1997 41 72.2 ±6.0 60-84 12.2 ±11.9 20 Skinner's Pond 1996 13 71.4+5.8 63-131 5.4 ±6.7 21 Miminegash 1996 37 74.1 ±8.5 64-100 8.6 ±10.8 22 Howard's Cove 1995 79 66.9 ±7.1 54-84 15.3 ±13.2 23 Egmont Bay 1982 15 66.1 ±7.0 58-77 19.4 ±16.0 24 Souris 1996 213 75.4 ±7.1 59-109 4.9 ±6.8 25 Fortune 1996 208 72.7 ±3.9 66-84 3.3 ±3.6 26 Beach Point 1982 243 88.0 ±12.9 61-120 7.4 ±8.9 27 Lismore 1997 51 78.2 ±11.4 58-103 7.8 ±12.4 28 Ballantynes Cove 1986 226 74.0 ±11.2 55-130 9.3 ±13.0 29 Port Hood 1988 339 67.4+11.6 51-150 4.5 ±2.9 30 Margaree 1984, 1988, 1992 1332 70.7 ±8.5 52-152 3.2 ±5.1 31 Pleasant Bay 1988, 1992 1053 69.8 ±8.1 54-130 2.3 ±2.1 184 Fishery Bulletin 100(2) mals recovered during the first recapture period following their tagging. Correlation (/•) was used to determine the relationship between the distance traveled and the size of the animal. To determine the relation between the average distance traveled and days at liberty, sites with recaptures over multiple years were used. The relationship between the average distance traveled and the extent of shallow waters was also established. The extent of shallow waters was quantified as the distance from shore to the closest 30-m bathymetric contour for each tagging site. Results A total of 7565 tagged lobsters were returned during their first recovery period, with size and geogi'aphical position at recapture. Only sites with fifty recaptures or more were con- sidered. There was no evidence of a relationship between the size of the animal and distance traveled in the southwestern GSL because the correlation coefficient ir) ranged from -0.19 Table 3 Correlation coefficient (;l for the relationship of the dis- tance traveled and the carapace length (CL) of American lobsters tagged in the southwestern Gulf of St. Lawrence. The number of each site corresponds to its geogi'aphical position on the map in Figure 1. n = number of recaptured lobster with tags indicating release location for which dis- tance traveled to recapture site could be calculated. Tagging site /) CL range (mm) r 1 Belledune 536 50-115 -0.12 3 Stonehaven 580 53-111 0.04 4 Anse-Bleue 232 54-91 0.02 5 Caraquet 1288 55-133 0.08 6 Petit-Shippagan 117 54-116 0.10 8 Le Goulet 90 59-98 -0.03 9 Val Comeau 500 57-117 -0.05 10 Neguac 72 65-105 0.22 12 Seacow Pond 61 61-94 0.08 13 Alberton 61 65-97 -0.08 14 Malpeque 251 51-100 0.11 16 North Lake 64 61-89 0.18 18 Kouchibouguac 51 60-74 -0,06 22 Howards Cove 73 54-84 -0.19 24 Souris 202 59-109 0.14 25 Fortune 207 66-84 0.04 26 Beach Point 243 61-120 -0.02 27 Lismore 51 58-103 0.13 28 Ballantynes Cove 188 55-130 0.23 29 Port Hood 339 51-150 0.20 30 Margaree 1319 52-152 -0.05 31 Pleasant Bay 1040 54-130 0.04 to 0.23 (Table 3). In tliis study small lobsters (<70 mm CL) traveled as far as the large animals (>90 mm CL) (Fig. 2). There was no significant difference in the average dis- tance traveled between males and females in eleven out of nineteen sites (Table 4). Females traveled significantly farther than males (Table 4) at three sites located in the upper part of Baie des Chaleurs (Fig. 1, sites 1, 3, and 4) and one site on the northeastern tip of Prince Edward Is- land (Fig. 1, site 25). The only site where males traveled significantly farther than females was in Val Comeau (Ta- ble 4, Fig. 1, site 9). No significant differences were ob- served in the average distance traveled by berried fe- males compared with males or nonberried females in four out of nine sites where data were available for berried females (Table 4). The average distance traveled by ber- ried females was significantly shorter than that by both males and females in three sites (Table 4, sites 9, 16, and 25) and significantly farther only in Port Hood (Table 4, site 29). In Souris (Table 4, site 24), no significant differ- ence was obsei-ved for average distance traveled between males and berried females, but they were both significant- ly shorter than the average distance traveled for nonber- ried females. Only seven out of thirty-one sites had a sufficient num- ber of recaptures over multiple years to allow a compari- son between distance traveled and days at liberty. A sub- stantial decrease in the percentage of tags recaptured from the first to the second and third recovery periods was observed (Table 5), reflecting high exploitation rates by the fishery. There is no strong evidence of a positive relationship between the average distance traveled and days at liberty. No significant difference (P>0.05) was ob- sei-ved in the average distance traveled over time in Stone- haven (site 3), Anse-Bleue (site 4), Caraquet (site 5) and North Lake (site 16) (Table 5). The average distance trav- eled decreased significantly (P=0.0002) in Belledune (site 1), whereas it increased significantly in Souris (site 24) (P=0.0119) and Margaree (site 30) (^=0.004) over multi- ple years recovery (Table 5). Even in these areas where distances were significantly different, the differences were not large (2.4, 2.3, and 5.3 km). Also, the longest distance traveled obsei^ved for the second or third (or both) recovery periods was equal or less than the one obsei"ved for the first recovery period for all 7 sites (Table 5). Lobster movements in the southwestern GSL seemed to be restricted to short distances along the coast near shore in areas where the lobster habitat is restricted to a few kilometers from the shore and longer distances over a broader gradually sloping bottom (Fig. 3). In general, 93"^^ of lobster dispersions were limited to the 20-m ba- thymetry contour. The shorter average distances traveled (2.4-4.9 km. Table 2) were observed in part of Baie des Chaleurs (Figs 1 and 4, sites 3 and 4), the northeastern tip of Prince Edward Island (Fig. 1, sites 16. 24, and 25) and Cape Breton (Figs 1 and 5, sites 29, 30, and 31). Relative- ly short average distances (5.3-8.6 km. Table 2) were ob- served around northeastern New Brunswick (Figs 1 and 6, sites 1, 2, 5, 6, 8, and 9), the northwestern tip of Prince Ed- ward Island (Fig. 1, sites 12, 13, 20, and 21), eastern New Brunswick (Fig. 1, sites 17 and 18) and the eastern end of Comeau and Savoie: Movement of Homaius amencanus in the southwestern Gulf of St. Lawrence 185 60 • A 50-- ♦Vt n=1288 ^^ ^ mean distance = 8.0 km ♦«»<♦♦ . SD = 8.8 40- ^ * r=0.08 30 - 20 10 55 75 95 115 135 155 60 B ? 50 n = 73 « mean distance = 15.3 Distance traveled 3 o o o o SD = 13.2 • 50 60 70 80 90 60 - 50 - ♦^ n=1319 mean distance = 3.2 40 - 30 - SD = 5 1 • r=-0.05 • 20 10 - 4 5 65 85 105 125 145 165 Carapace length (mm) Figure 2 Relationship between the distance traveled and the size of the animal for American lobsters recaptured in (A) Caraquet, (B) Howard's Cove, and iC) Margaree. the Northumberland Strait (Fig. 1, sites 26 and 27). The longest average distances traveled (9.3-19.4 km, Table 21 were observed in the immediate vicinity of Miramichi Bay (Figs 1 and 6, sites 10 and ll),Malpeque Bay andTracadie Bay in northern Prince Edward Island (Fig 1, sites 14, and 15), St. Georges Bay (Fig. 1, site 28) and central Northum- berland Strait (Figs 1 and 7, sites 19, 22, and 23). Lobster movements in the Northumberland Strait were related to whether the site was located toward the center or at either extremity (either end) of the Strait. Lobsters at sites close to the external boundaries (Fig. 1, sites 17, 18, 20, 21, 26, and 27) traveled on average shorter dis- tances (5.4-8.6 km) than those located in the center of the Strait (Fig. 1, 12.2-19.4 km, sites 19, 22, and 23). Although tags were recovered in the western portion of Northum- berland Strait (LFA 25) at a different time (a different fishing season) compared with tag recoveries at the other LFAs, movements seemed to be related to the extent of shallow waters (<20 m) rather than the time of the recov- ery period. Discussion Lobster movements in the southwestern GSL are related to the local bottom topography and are depth-dependent, i.e. lobsters traveled on average longer distances in areas where the shallow waters (<20 m) extended farther from shore. We observed that on the narrow coastal shelf of western Cape Breton and in some areas in Baie des Chal- eurs, lobsters traveled on average less than 5 km compared with distances ranging from 9.3 to 19.4 km in the grad- ually sloping bottom of the Northumberland Strait and some shallow bays. Similarly, Templeman (1935) reported 186 Fishery Bulletin 100(2) Table 4 Comparison of the average distance traveled between the average distance traveled by male, female. and beiTied female American lobsters. The number of each site corresponds to its geographical position on the map in Figure 1. ;i = number of recaptured lobster with tags indicating release location for which distance traveled to recapture site could be calculated . Average distance [^-Test or Biological traveled ±SD Ki-uskal-Wallis' Tagging site category ;( (kml P 1 Belledune male female 360 176 7.0 ±5.9 9.0 ±8.3 0.0003 3 Stonehaven male female 316 264 2.1 ±2.8 2.7 ±3.3 0.0248 4 Anse-Bleue male female 127 105 2.1 ±3.3 3.8 ±5.2 0.0056 5 Caraquet male female berried female 746 487 55 8.1 ±9.1 7.7 ±8.6 5.5 ±6.6 0.0937 8 Le Goulet male female 45 45 5.3 ±2.7 5.2 ±2.7 0.4385 9 Val Comeau male female 279 218 7.7 ±7.1 6.3 ±5.6 0.0260^' berried female 5 1.3 ±0.7 0.0018 12 Seacow Pond male female 20 41 5.0 ±7.7 5.5 ±7.9 0.5800 13 Alberton male female 33 29 6.0 ±6.7 5.4 ±5.0 0.8849 14 Malpeque male female berried female 114 95 42 11.0 ±7.6 11.2 ±7.4 10.5 ±9.9 0.7310 16 North Lake male female 26 33 2.1 ±1.4 3.5 ±4.9 0.4700-' berried female 5 0.8 ±0.6 0.0243 18 Kouchibouguac male female 25 26 8.9 ±7.8 8.0 ±7.0 0.5847 22 Howard's Cove male female 43 31 17.3 ±12.3 12.8 ±14.0 0.1333 24 Souris male 131 3.6 ±5.3 0.5826' female 66 7.1 ±9.0 0.0140^' berried female 5 1.6 ±0.7 0.0355 25 Fortune male female 133 62 3.2 ±3.6 4.0 ±3.7 0.015.5- berried female 12 0.5 ±0.2 0.0001 26 Beach Pomt male female 178 65 6.4 ±6.2 10.1 ±13.4 0.4729 28 Ballantynes Cove male female 56 132 8.5 ±10.0 8.0 ±9.7 0.6193 29 Port Hood male female 139 1.58 4.4 ±3.1 4.3 ±2.6 0.7616- berried female 42 5.6 ±3.2 0.0294 30 Margaree male female berried female 449 553 317 3.1 ±4.9 3.5 ±5.0 2.8 ±5.0 0.1478 31 Pleasant Bay male female berried female 322 565 1.53 2.1 ±1.8 2.3 ±2.2 2.5 ±2.4 0.2638 ' The ['-test and the Kru *kal-Wallis test were used to compare t vo and three groups. respectively. ■■' [/-test between male.s and fe nales. ' U-test between males and be rried females. Comeau and Savoie: Movement of Homarus amencanus in the southwestern Gulf of St Lawrence 187 Table 5 The average and the longest distances trav eled bv American lobsters for sites with t igs returned over multiple years. The time between tagging and recapture of the animal (number ofd ays at hberty), and the number of observations («) are indicated. The number of each s te corresponds to its geogi aph ical position on the map in Figure 1. Number Average t/-test or Longest of days distance traveled Kruskal-Wallis' distance traveled Tagging site at hberty n (km) P (km) 1 Belledune 295-378 678-747 1048-1102 497 72 35 7.6 ±6.1 6.7 ±7.5 5.2+3.9 0.0002 44 37 17 3 Stonehaven 231-291 598-648 580 38 2.4 ±3.0 2.4 ±4.1 0.5827 28 17 4 Anse-Bleue 217-270 577-628 942-972 235 48 7 2.9 ±4.3 6.2+11.9 5.0 +7.3 0.2753 32 56 17 5 Caraquet 213-360 584-725 965-1057 1302 97 17 7.9+8.4 7.7 ±11.8 11.8 ±14.5 0.6416 51 49 50 16 North Lake 232-280 605-641 65 13 2.7 ±3.7 1.7 ±1.5 0.2075 24 6 24 Souris 230-285 584-625 202 11 4.7 ±6.9 7.0 ±4.2 0.0119 31 13 30 Margaree 250-308 622-671 147 13 3.3 ±3.7 8.6 ±7.6 0.004 23 21 ' Tlie C'-te. from the 1994 tagging project conducted in Stonehaven, New Brunswick. The release sites are indicated by a star symbol. Recently, a trawl survey conducted at a depth of 40 m in the Caraquet area (Bale des Chaleurs) over a 7-month pe- riod produced lobsters in mid-May, late-October, and No- vember, but not between June and early-October ( Comeau, personal obs.). Further, the recapture positions during the fishing season showed that lobsters tagged during these trawl surveys were recaptured along the coast at depths less than 20 m from Stonehaven to Miscou (Comeau, per- sonal obs.). This finding suggests that there is an inshore- offshore movement on the south shore of Bale des Chal- eurs similar to the one observed by Corrivault (1948) on the north shore of that bay. Unfortunately, our tagging projects were not designed to study this type of movement and did not allow us to speculate more on inshore-offshore movements. The lack of long-range movements across the south- western GSL could be explained by the presence of an extensive cold (<1.5°C) intermediate layer (CIL). In the southwestern GSL, the CIL is a large volume of water sandwiched between the coastal water and the deep wa- ter located in the Laurentian channel. The top of the lay- er ranges from 20 to 40 m depth from June to October and rises to the surface from January to April (Gilbert and Pettigrew, 1997). As it was hypothesized by Stasko (1980), there seems to be no advantage in long-distance movement to deeper water (>40 m) for lobsters in the GSL because it is cold (<1.5°C) in both summer and winter (CIL). Although lobsters can tolerate temperatures rang- ing from -1.5° to 30°C, at temperatures below 0°C they are in a state of hibernation, and below 5°C molt induc- tion is blocked (Waddy et al., 1995). It is clear that lob- sters can "tolerate" cold temperature but to be active they need warmer waters. Lobster movements of more than 40 km were rare in the southwestern GSL and movements exceeding 70 km through deep, colder waters (>40 m) were not observed. In contrast, lobsters from the Bay of Fundy, from coastal waters of southwestern Nova Scotia, and from coastal waters off New England can take advan- tage of warmer temperatures in deep waters in the Gulf of Maine and on the continental shelf during the winter (Campbell and Stasko, 1986). Campbell and Stasko ( 1985) and Campbell (1989) showed that lobsters tagged in the inshore waters of southwestern Nova Scotia traveled up to 240 km to the edge of the continental shelf off Georges Bank to depths below 200 m, seemingly without crossing a wide area of cold water. Similarly, lobsters tagged in the Bay of Fundy were also recaptured in deep waters at the edge of the continental shelf across the Gulf of Maine and along the coastal waters of the United States (Campbell and Stasko, 1986), at a distance of more than 780 km. These types of movements were not observed in the south- western GSL. Lobster movements between mainland New Brunswick, Prince Edward Island, or Cape Breton, and the Magdalen Islands, for example, have not been report- ed. The Magdalen Islands are an archipelago with a sub- stantial lobster fishery located in the middle of the south- western GSL, surrounded by >60 m depths at about 80 to 90 km from Prince Edward Island, and Cape Breton. Lob- ComeaLi and Savoie: Movement of Homarus amencanus in the southwestern Gulf of St- Lawrence 189 ly / -'}^ / ^ '/ w /yy 35. V / 46°53' - m. . , /- . ' w Gulf of St. Lawrence ' ,■ / ' ' •'t^s^ / ' / ; JF Pleasant Bay /-■^ /'■// / - ---^, \ / ■■/ / / i Cape Breton 46°4r — /.://' / / . ; // / / 7 Nova Scotia \^ & /■//■■' / J 2-5 5 Kilometers 60 mm CL) in the southwestern GSL were not sex- or size-dependent, ex- cept for berried females. Similar to our findings, the re- sults of most recent studies do not indicate significant differences between the distance traveled in relation to size or sex for lobsters tagged in coastal waters (Fogarty et al, 1980; Ki'ouse, 1981; Campbell. 1982; Tremblay et al.. 1998). In contrast. Templeman (1935) and Bergeron ( 1967) suggested that lobster movements were sex-depen- dent, but neither author reported whether the differences were supported in a statistical or biological sense. Camp- bell and Stasko ( 1985. 1986) and Campbell ( 1989) indicat- ed that lobster movements were size-dependent because they observed that large mature animals (>95 mm CL) on average traveled significantly farther than small im- mature ones. They explained that mature animals would move more extensively to reach the warmest seasonal temperature to maximize their degree-days (the accumu- lative sum of daily mean temperatures recorded above 0°C) needed for somatic and gonadic development. This was not the case in our study. We observed, however, that on average berried females in the southwestern GSL traveled shorter distances than males and nonberried fe- males. Saila and Flowers (1968) indicated that the dis- tance traveled by berried females was related to their physiological state. In a tagging experiment, they cap- tured berried females on the continental shelf tagged, and released them in the inshore waters off the coast of Rhode Island at about 220 km from their captured posi- tion. When females were carrying eggs, they traveled on- ly short distances within the inshore waters. Once they shed their eggs, however, these females returned to the continental shelf where they were originally captured. Berried females tagged in inshore waters in the Magda- len Islands (Munro and Therriault. 1983). Jeddore Har- bour and Clam Bay. Nova Scotia (Jan'is. 1989). and New Hampshire (Watson et al., 1999) also traveled short dis- tances. In the Cape Cod area, berried females tagged in the inshore waters were reported to have traveled an av- erage distance of 30 km, mostly parallel to the coast (Mor- rissey, 1971; Estrella and Mornssey, 1997). Off Grand Manan Island, Campbell (1986, 1990) reported relatively small inshore-offshore migration, less than 15 km, for 75"^ of the berried females and attributed this migration to an effort to maximize egg development by exposure to warmer water. In general, it seems that the condition of carrying eggs could influence the extent of movements in the southwestern GSL, not the size or sex of the animal. In terms of fishery management, there is relatively lit- tle interaction between lobsters at different LFAs at the benthic level because lobster traveled on average small Comeau and Savoie: Movement of Homaivs amencanus in the southwestern Gulf of St, Lawrence 191 distances. The main concern of fishermen in the south- western GSL was the minimal legal size (MLS) disparity between LFA 24 (MLS of 63.5 mm CD located on the north side of Prince Edward Island and the other LFAs (MLSs from 65.1 to 70.0 mm CL). More precisely, they were interested in lobster movements in relation to time. i.e. if lobsters released in a given area in one season would be recaptured in the same area in future seasons. From our findings, lobsters do not move farther if they are at large for a longer period. Distances traveled by lobsters were not time-dependent for lobsters at large between 200 (with at least one winter season) and 1102 days. Results of our tagging studies were consistent with results from ear- lier studies carried out in the southwestern GSL and dem- onstrated that lobsters in their benthic stages have little long distance interaction. Hence, lobster movements in the southwestern GSL should have minimal consequences in terms of lobster management. Acknowledgments The authors wish to thank all fishermen from the south- western Gulf of St. Lawrence who returned lobster tags. We also want to thank Bruno Comeau, Daniel Ferron, Marc Lanteigne, Wade Landsburg, Manon Mallet. Pierre Mallet. Donald R. Maynard. Gilles Paulin and Guy Robichaud for their technical assistance in the field and during tag col- lection. We especially thank J. Mark Hanson. Marc Lan- teigne. and David Robichaud for critically reviewing the manuscript and three anonymous reviews for thoughtful suggestions that improved the quality of this manuscript. Literature cited Bergeron. T. 1967. Contribution a la biologie du homard {Homarus ameri- canus M. Edw. idesIles-de-la-Madeleine. NaturalisteCan. 94:169-207. Campbell, A. 1982. Movements of tagged lobsters released off Port Mait- land. Nova Scotia, 1944-80. Can. Tech. Rep. Fish. Aquat. Sci. 1136, 41 p. 1986. Migratory movements of ovigerous lobsters. Homa- rus americanus, tagged off grand Manan, eastern Canada. Can. J. Fish. Aquat. Sci. 43:2197-2205. 1989. Dispersal of American lobsters, Homarus americanus. tagged off southern Nova Scotia. Can. J. Fish. Aquat. Sci. 46:1842-1844. 1990. Aggregations of berried lobsters tHomarus america- nus) in shallow waters off Grand Manan, eastern Canada. Can. J. Fush. Aquat. Sci. 47:520-523. Campbell. A., D. E. Graham. H. J. MacNichol. and A. M. Williamson. 1984. Movements of tagged lobsters released on the conti- nental shelf from Georges bank and Baccaro bank, 1971-73. Can. Tech. Rep. Fish. Aquat. Sci. 1288, 16 p. Campbell. A., and A. B. Stasko. 1985. Movements of tagged American lobsters. Homarus americanus. off southwestern Nova Scotia. Can. J. Fish. Aquat. Sci. 42:229-238. 1986. Movements of lobsters [Homarus americanus) tagged in the Bay of Fundy, Canada. Mar. Biol.. 92:393-404. Comeau, M., W. Landsburg, M. Lanteigne, M. Mallet, P. Mallet, G. Robichaud, and F. Savoie. 1998. hohster {Homarus american us ) tagging project in Cara- quet (1993)— tag return from 1994 to 1997. Can. Tech. Rep. Fish. Aquat. Sci. 2216, 35 p. Cooper. R. A. 1970. Retention of marks and their effects on growth, behav- ior and migrations of the American lobster. Homarus amer- icanus. Trans. Am. Fish. Soc. 99:109-417. Cooper, R. A., and J. R. LTzmann. 1971. Migrations and growth of deep-sea lobster, Homarus americanus. Science 171:288-290. Cooper. R. A.. R. A. Clifford, and C. D. Newell. 1975. Seasonal abundance of the American lobster, Homa- rus americanus, in the Boothbay region of Maine. Trans. Am. Fish. Soc. 104:669-674. Corrivault, G. W. 1948. Contribution a I'etude de la biologie du homard {Homarus americanus) des eaux de la province de Quebec. Ph.D. diss., Universite Laval. Quebec. 283 p. Dow. R. L. 1974. American lobsters tagged by Maine commercial fish- ermen. 1957-59. Fish. Bull. 72:622-623. Ennis. G. P. 1984. Small-scale seasonal movements of the American lob- ster.Homarus americanus. Trans. Am. Fish. Soc. 113:336- 338. Estrella. B. T. and T D. Mornssey. 1997. Seasonal movement of offshore American lobster. Homarus americanus, tagged along the eastern shore of Cape Cad, Massachusetts. Fish. Bull. 95:466-476. Fogarty, M. J., D. V. D. Borden, and H. J. Russell. 1980. Movements of tagged American lobster, Homarus americanus, off Rhode Island. Fish. Bull. 78:771-780. Gilbert. D.. and B. Pettigrew. 1997. Interannual variability ( 1948-19941 of the CIL core temperature in the Gulf of St. Lawrence. Can. J. Fish. Aquat. Sci. 54 (suppl. l):57-67. Jarvis. C. 1989. Movement patterns of late-stage ovigerous female lobsters {Homarus americanus Milne-Edwards I at Jeddore, Nova Scotia. M.S. thesis, Dalhousie Univ., Halifax. Nova Scotia, Canada, 148 p. Karnofsky, E. B.. J. Atema, and R. H. Elgin. 1989. Natural dynamics of population structure and habitat use of the lobster, Homarus americanus, in a shallow cove. Biol. Bull. 176:247-256. Krouse, J. S. 1981. Movement, growth, and mortality of American lob- sters, Homarus americanus, tagged along the coast of Maine. U.S. Dep. Commer, NOAA,Tech. Rep. NMFS SSRF- 747, 12 p. Lawton. P. and K. L. Lavalli. 1995. Postlarval. juvenile, adolescent, and adult ecology. In Biology of lobster. Homarus americanus (J. R. Factor, ed.). p. 47-81. Academic Press. New York. NY. 528 p. Miller, R. J., R. E. Duggan. D. G. Robinson, and Z. Zeng. 1989. Growth and movement o{ Homarus americanus on the outer coast of Nova Scotia. Can Tech. Rep. Fish. Aquat. Sci. 1716. 17 p. Moriyasu M., W. Landsburg, and G. Y Conan. 1995. Sphvrion tag shedding and tag induced mortality of the American lobster, Homarus americanus H. Milne Edwards, 1837 ( Decapoda, Nepbropidae 1. Cru.staceana 68: 184- 192. 192 Fishery Bulletin 100(2) Morrissey, T. D. 1971. Movements of tagged American lobsters. Homarus americanus, liberated ofTCape Cod, Massachusetts. Trans Am. Fish. Soc. 100:117-120. Munro, J., and J.-C. Themault. 1983. Migi'ations saisonnieres du homard (Homarus ameri- canus) entre la cote et les lagunes des Iles-de-la-Madeleine. Can. J. Fish. Aquat. Sci. 40:905-918. Robichaud, D. A., and P. Lawton. 1997. Seasonal movement and dispersal of American lob- sters. Homarus americanus, released in the upper Bay of Fundy Can Tech. Rep. Fish. Aquat. Sci. 2153, 21 p. Saila, S. B., and J. M. Flowers. 1968. Movements and behaviour of berried female lobsters displaced from offshore areas to Narragansett Bay, Rhode Island. J. Cons. Int. Explor Mer 31:342-351. Stasko, A. B. 1980. Tagging and lobster movements. Can. Tech. Rep. Fish. Aquat. Sci. 932:141-150. Templeman, W. 1935. Lobster tagging in the Gulf of St. Lawrence. J.Biol. Board Can. 1:269-278. 1936. Local differences in the life history of lobster on the coast of the maritime provinces of Canada. J. Biol. Board Can. 2:41-88. 1940. Lobster tagging on the west coast of Newfoundland 1938. Nfld. Dep. Nat. Resource Fish. Res. Bull. 8, 16 p. Tremblay, M. J., M. D. Eagles, and G. A. P. Black. 1998. Movements of the lobster, Homarus americanus. off northeastern Cape Breton Island, with notes on lobster catchability Can. Tech. Rep. Fish. Aquat. Sci. 2220, 32 p. Uzmann, J. R.. R. A. Cooper, and K. J. Pecci. 1977. Migration and dispersion of tagged American lob- sters, Homarus americanus. on the southern New England continental shelf U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-705, 92 p. Waddy S. L., D. E. Aiken, and D. P V. De Kleijn. 1995. Control of growth and reproduction. In Biology of lobster, Homarus americanus (J. R. Factor, ed. ), p. 217-266. Academic Press, New York, NY, 528 p. Watson III, W. H.. A. Vetrovs, and W. H. Howell. 1999. Lobster movements in an estuary Mar. Biol. 134:6.5— 75. Wilder, D. G. 1963. Movement, gi'owth and survival of marked and tagged lobsters liberated in Egmont Bay, Prince Edward Island. J. Fish. Res. Board Can. 20: 305-318. 1974. Inshore and offshore lobster stocks. Fish. Res. Board Can. MS Rep. 1293, 14 p. 193 Abstract— Short.spino thoniyhead (St-- Ixi.stolohii.-i ulaacanus) abundance was es- timated from 107 video transects at 27 stations recorded from a research sub- mersible in 1991 off southeast Alaska at depths rantrinf; from 165 to 355 m. Num- bers of invertebrates in seven major taxa were estimated, as was substrate type. Thornyhead abundance ranged from to 7.5/100 m-, with a mean of 1.22/100 m'-, and was positively correlated with depth and amount of hard substrate. Invertebrate abundances were not sig- nificantly correlated with numbers of thornyheads. Shortspine thornyhead abundance estimates from this study were several times higher than esti- mates produced by bottom trawl sur- veys off southeast Alaska in 1990 and 1993, the two years of survey that encompassed the submersible tran- sects: however, the trend of increasing abundance with depth was similar in the trawl surveys and in the submers- ible transects, suggesting that trawl surveys systematically underestimate abundance of shortspine thornyheads. Shortspine thornyhead iSebastolobus alascanus) abundance and habitat associations in the Gulf of Alaska Page Else Lewis Haldorson School of Fisheries and Ocean Sciences University of Alaska 11120 Glacier Highway Juneau, Alaska 99801 E-mail address (for L Haldorson, coniaci author) lewhaldorsoniwuafedu Kenneth Krieger Auke Bay Laboratory National Manne Fishenes Service Juneau, Alaska 99802 Manuscript accepted 24 August 2001. Fish. Bull 100:193-199 (2002). Fish distributions are affected by physi- cal conditions such as depth, tempera- ture and substrate type and by biotic variables such as prey distribution, pred- ator presence, and habitat features (e.g. kelp forests). Patterns of habitat use are important factors in resource assess- ments, and stratification based on habi- tat characteristics are common features of survey design. For example, trawl sur- veys in the Gulf of Alaska are stratified by depth and general bottom type (flats, gullies, shelf break, slope) (Stark and Clausen, 1995). Assessment offish popu- lations based on traditional fishing gear, such as trawls or longlines, provide rel- atively economical surveys with broad geographic coverage; however, they pro- vide limited detailed information on spe- cies associations or habitat preferences (Matlock etal., 1991). Shortspine thornyhead (Sebastolobus alascanus) is a member of the family Scorpaenidae, which includes the rock- fishes {Sebastes — over 60 species in the northeast Pacific) and three species of thornyheads {Sebastolobus). Shortspine thornyhead range from Baja California, Mexico, to the Bering Sea and are found at depths to 1500 m (Moser, 1974). Off southeast Alaska, most fish sampled from 200-310 m depths were between 15 and 30 years old and had lengths of 25-35 cm (Miller, 1985). The maxi- mum age observed by Miller ( 1985 ) was 62 years, and age at 509!^ maturity was 12 years for both sexes. These life-histo- ry features are similar to those of rock- fishes, and such long-lived fishes are difficult to manage because they are easily overexploited and recover slowly from overfishing (Adams, 1980). Short- spine thornyhead have been commer- cially valuable in the Gulf of Alaska, where catch (including discards) ranged from 1298 to 2020 metric tons from 1991 tol996 (lanelli and ItoM. Bycatch of thornyhead has the potential of forc- ing closure of other high-value fisher- ies (lanelli and Ito^). Estimates of their abundance and size distribution are currently based on bottom trawl and longline surveys (lanelli and Ito^); how- ever, direct observations from submers- ibles can provide an alternate method for assessing abundance ( Krieger, 1993 ). Assessments of thornyheads and rock- fishes often result in population esti- mates with high variances; resulting in considerable uncertainty for assigning ' lanelli.J. N.andD. H. Ito. 1998. Status of Gulf of Alaska thornyheads iSebastolo- biis sp. ) in 1998. In Stock assessment and fishery evaluation report for the ground- fish resources of the Gulf of Alaska, p. 371-402. [Available from North Pacific Fishery Management Council, 605 West 4'*' St., Anchorage, Alaska 99501.] - lanelli,.:. N., and D.H. Ito. 1994. Thorny- heads. In Stock assessment and fishery evaluation report for the gi-oundfish re- sources of the Gulf of Alaska for 1995. (Available from North Pacific Fishery Man- agement Council. 605 4"' St., Anchorage, Alaska 99501.1 194 Fishery Bulletin 100(2) 1 40'W 55 30'N Figure 1 Locations of stations ( • ) where transects were run off southeast Alaska. The offshore line marks the 200-m isobath near the edge of the continental shelf harvest levels. Submersible obsei-vations with line-transect methods have been used in attempts to improve estimates of rockfish abundance and to understand rockfish habitat associations (Richards, 1986; Pearcy et al., 1989; Krieger, 1992; Stein et al.. 1992; O'Connell and Carlile, 1993). Our goal was to assess abundance and habitat use by shortspine thornyhead in the eastern Gulf of Alaska. based on data from existing video records taken during submersible transects. We estimated abundance of fish and invertebrates and quantified substrate type. We ex- plored the relationships between thornyhead abundance and both physical and biotic environmental variables and compared abundance estimates from submersible tran- sects with those from trawl survevs in the same area. Materials and methods Sources of data were video tapes of 107 bottom transects recorded at 27 stations during submarine dives in June 1991 on the outside coast of northern southeast Alaska from Cape Ommaney to Yakatat (Fig.l). All transects were conducted with the Delta submersible. This battery-pow- ered two-man submersible is 4.7 m long, dives to 365 m, and travels 2-6 kni/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, and underwater telephone and transponder that allowed the submersible to be tracked from the sur- face support vessel. The surface vessel recorded LORAN fixes at the beginning and end of each transect. A pilot and observer formed the crew of the submersible: the pilot attempted to maintain the submersible within 0.5 m above the bottom at 3-4 km/h while the observer made obsei-va- tions through a starboard porthole. Video recordings were made from a downward project- ing external Hi-8 color video camera tilted obliquely for- ward on the starboard side and included a digital read- out of depth, temperature, and height above bottom. Data were collected either in strip transects by using the entire length of each transect, or in quadrats by freezing individ- ual frames from the video. The width of the strip transect was calculated from the height above bottom and field of view of the camera. The field of view formed a trape- zoid on the seafloor beginning almost directly beneath the camera and projecting forward; its maximum width iW) was estimated by calibrations performed on a subsequent cruise that had the same camera and camera configura- tion (Zhou and Shirley, 1997). When the submarine was resting on the bottom, the camera height iH) was 0.93 m and the width of the field of view (i.e. longest side of trap- ezoidal field) was 1.78 m. Transect width i \V) was estimat- ed by W= {1.78/0.93) H. llZhou and Shirley 1997)]. The height above bottom (H) was recorded at one minute intervals during each transect and mean height was used to estimate width ( W) of that transect. All dives were made during daylight between 0600 and 1900 h. At each station a series (usually 4) of parallel transects was run. and spac- ing between transects was about 200 m. Transect lengths Else et a\ Abundance of Sebaslolobus alasconus in the Gulf of Alaska 195 were calculated from the position fixes taken from the sur- face vessel. The entire length of each transect was viewed and all thornyhcads in the field of view were counted. Inverte- brates were also counted in seven categories: starfish, seap- ens, sea urchins, anemones, corals, sponges, and sea cu- cumbers. Depth and temperature were recorded once every minute during each transect and averaged over the com- plete transect. Substrate type was estimated by scoring vid- eo quadrats at one-minute intei-vals during the transect. A mylar grid of about twenty-five 50-mm squares was placed on the screen over the freeze-framed image, and substrate within the squares was scored in three categories of soft (mud, sand, and gravel), cobble, and rock-boulder, the lat- ter two of which were then combined into a single catego- ry, hard-bottom. The proportions of each category (soft and hard) for each transect were estimated as the mean from the one-minute quadrats. Abundance (number/100 m-) of shortspine thornyheads and invertebrate categories was es- timated for each transect by dividing the number counted by the area estimate ( W x Ti-ansect length x 100). To evaluate variables that may have affected thorny- head densities, we assembled a correlation matrix and a partial correlation matrix of three physical variables (depth, substrate, temperature) and transformed (log(.v-i-l)) biotic variables (shortspine thornyheads and seven inver- tebrate categories). Based on those matrices, we selected depth and substrate for further analyses. We used nonparametric procedures (Kruskal-Wallis and Mann-WTiitney) to determine if thornyhead densities var- ied among sampling stations and to evaluate the effects of substrate and depth. Substrate and depth were coded into nominal categories for use as independent variables in those analyses. We used Scheffe's (Zar, 1984) post-hoc test to evaluate between-categorv differences when Krus- kal-Wallis results were significant. An ANOVA factorial model was used to explore the joint effects depth and substrate on shortspine thorny- head abundance (transformed by logtx-i-l)). In addition, we used stepwise linear regression with thornyhead abun- dance as the dependent variable to evaluate the relative importance of all possible independent variables. 0.2 0.3 0.4 0.5 0,6 0.7 0.8 0.9 Distance (km) 50 B 40 c 30 0) o I 20 10- 100 150 200 250 300 350 400 450 Depth (m) 0.4 0.6 O.i Hard substrate Figure 2 Distribution of transects by (A) distance (km). (B) average depth, and (C) the proportion of bottom categorized as hard. Results Transect lengths ranged from 320 to 800 m and had a mean of 580 m (Fig. 2A). The mean transect depth ranged from 165 to 355 meters, and the highest numbers of tran- sects were at 200-250 m (Fig. 2B). Most transects (85'7f ) had substrate that was completely soft (silt, sand, and gravel) or hard (cobble and rock-boulder) (Fig. 2C). Abun- dance of shortspine thornyheads per transect ranged from to 13.6/100 m'-; and the mean abundance at the 27 sta- tions ranged from to 7.5/100 m- i Table 1). Difference in abundance among stations was very highly significant (P <0.0001, Kruskal-Wallis). A correlation matrix of all variables indicated that depth, substrate, and sponge abundance were related to variation in thornyhead abundance; however, the partial correlation matrix indicated that among those three vari- ables, substrate type and depth were most strongly re- lated to shortspine thornyhead abundance (Table 2). The high correlation of sponge and thornyhead abundances was apparently spurious because of the relationship be- tween sponge abundance and substrate type (Table 2). For further analyses of the relationship between thorny- head abundance, depth, and substrate, we coded depth into three nominal categories, <200 m, 200-300 m, and >300m, chosen to correspond to the depth intervals used in National Marine Fisheries Service (NMFS) triennial trawl surveys (Stark and Clausen. 1995). We also coded substrate into two nominal categories, soft bottom (>90'7( sand, mud, and grav- el) and hard bottom (>40% cobble and rock-boulder). Abun- dance increased with depth (Table 3), and differences that were highly significant occurred among three depth catego- ries (P<0.001, Kruskal-Wallis). Significant differences exist- 196 Fishery Bulletin 100(2) Table 1 Mean means numbers, with standard deviation, c of depth, bottom temperature, and f shortspine thornyheads ^Seha^tvlobus alascanus) per 100 proportion of the bottom categorized as hard substrate. m- at transect stations, with Station n No./lOO m- SD Mean depth (ml Temp. Proportion of bottom as hard substrate Lat. Long. 55 4 0.4 0.002 210.1 4.00 0.27 56°37'11" 135°57'50" 56 4 0.5 0.002 209.0 4.33 0.00 56°32'19" 135°54'39" 57 4 2,7 0.007 319.9 3.83 1.00 56°06'06" 135°39'72" 59 4 0.2 0.002 280.1 4.51 0.00 55°46'03" 134°77'97" 60 4 0.0 227.9 4.83 0.00 55°61'28" 134°55'86" 61 4 0.0 202.9 4.90 0.00 55°71'53" 134''76'89" 62 4 0.0 165.5 5.42 1.00 55°86'00" 134°84'00" 63 4 0.9 0.004 355.9 4.04 0.04 55°98'36" 134°93'97" 64 4 3.8 0.011 220.3 4.86 0.78 56°80'97" 135°80'97" 65 4 2.5 0.007 237.9 4.73 0.54 56°69'22" 135°87'69" 67 3 2.6 0.012 247.9 3.95 0.11 57°24'72" 1.36°29'31" 68 4 3.6 0.02 246.2 4.50 0.75 57°21'67" 136°25'03" 69 4 7.5 0.045 242.7 4.53 0.72 57°55'33" 1.36°55'22" 70 4 0.7 0.01 217.7 4.84 0.20 57°95'94" 137°21'61" 71 4 0.1 0.002 177.4 4.95 1.00 58°02'00" 137°10'00" 72 4 0.1 0.001 289.9 4.02 0.00 58°10'92" 136°98'39" 73 4 1.2 0.007 210.8 4.79 1.00 59''03'75" 14n3'75" 75 4 0.1 0.001 208.4 4.25 0.86 59°24'50" 14r62'14" 77 4 2.9 0.015 299.8 3.76 0.71 59°31'44" 142°11'42" 78 4 0.0 205.8 4.25 0.77 59°22'22" 141°70'86" 79 4 0.0 183.9 4.67 0.51 58°00'00" 140°50'00" 80 4 0.4 0.001 251.4 4.43 0.00 58''66'81" 139°37'33" 81 4 0.3 0.002 208.4 5.04 0.00 58°65'61" 139°56'08" 82 4 0.3 0.001 262.8 4.59 0.00 58°53'81" 139°6r.56" 83 4 0.0 218.9 4.81 0.00 58°35'00" 139°32'89" 84 4 0.0 172.4 5.32 0.76 58°22'00" 139°00'00" 85 4 2.4 0.016 215.2 5.26 0,74 .58°10'17" 138°65'3.3" ed between the shallowest depth category and both deeper depth groups (Table 3, Sheffe test). Substrate type also af- fected abundance (Table 3); transects with hard substrate had a significantly higher density of thornyheads than tran- sects on soft bottom (P=0.016, Mann- Whitney). There was a significant interaction (P=0.01) of depth and substrate in the ANOVA factorial analysis (Table 4) due to the sharp in- crease in thornyhead abundance in depths >200 m on the hard substrate than on the soft substrate (Fig. 3). Stepwise multiple regi'ession with all variables resulted in a model that incorporated two variables: substrate type and depth, with substrate type entered first. The final two- variable regression model was A = 0.017 S + 0.00016 D - 0.033, (/•■-=0.2081 where A = thornyhead abundance; S = proportion of bottom that is hard substrate; and D = depth (m). Thornyhead abundance increased with depth and amount of hard substrate, although there was an indication that abundance may have reached maximum levels at depths of200-300m(Fig. 3). Discussion Film (still or motion) and videotape recordings of tran- sects have been used to assess abundance of aquatic or- ganisms (Auster et al., 1989; Butler et al.'l. Potential biases in such data include systematic underestimation Butler, J. L., W. W. Wakefield, P. B. Adams, B. H. Robison, and C. H. Baxter 1991. Application of line transect methods to sui'vcying demersal communities with ROVs and manned sub- mersibles. Proceedings of the IEEE Oceans '91 Conference, p. 689-696. histitute of Electrical and Electronic Engineers, Pis- cataway, NJ. Else et a\ Abundance of Sebastolobus alascanus in the Gulf of Alaska 197 Table 2 Correlation (lower left diagon substrate type, temperature. al) and partial correlation matrices (upper right diagonal) for shortspine thornyhcad (SSTHi, depth, ind invertebrates. Significant correlations are indicated in bold. SSTH Depth Substrate Temperature Sea star Urchin Sea pen Anemone Coral Sponge Cucumber SSTH 0.35 0.37 0.08 -0.13 0.24 0.04 -0.13 0.04 0.27 -0.12 Depth 0.23 -0.34 -0.56 0.02 0.03 0.00 0.09 -0.16 -0.08 0.11 Substrate 0.32 -0.31 0.00 -0.12 -0.16 -0.01 0.39 -0.15 0.23 0.25 Temp. -0.10 -0.62 0.17 -0.02 -0.07 0.24 0.04 -0.05 -0.15 -0.22 Seastar -0.18 0.05 -0.16 -0.04 0.23 0.18 0.24 0.04 -0.06 0.01 Urchin 0.11 0.18 -0.21 -0.12 0.35 0.19 0.04 0.08 -0.13 -0.04 Sea pen -0.01 -0.09 0.04 0.13 0.36 0.33 -0.06 0.35 0.10 0.43 Anemone 000 -0.07 0.38 0.03 0.22 0.04 0.15 0.07 0,07 0.09 Coral -0.05 -0.09 -0.03 0.03 0.30 0.27 0.61 0.16 0.05 0.27 Sponge 0.37 -0.08 0.41 -0.02 -0.18 -0.16 -0.02 0.14 -0.04 -0.17 Cucumber -0.06 0.08 0.13 -0.15 0.27 0.20 0.59 0.24 0.53 -0.09 Table 3 Mean abundance (number/100 ni-) of shortspine thorny- heads in three depth categories ( Kruskal-Wallis, P<0.0001). and in two substrate categories (Mann-Whitney, P=0.016), with sample sizes and standard errors. Results of nonpara- metric analyses are included. Abundance u SE Depth 100-200 0.1 18 0.1 200-300 1.4 80 0.2 >300 1.9 9 0.4 Substrate soft 0.6 57 0.1 hard 1.9 50 0.4 due to gear avoidance or overestimation due to attraction of fish to the submersible. The behavior of shortspine thornyheads obsei^ed in the videotapes from our study indicated that they were not affected by the submersible. They typically remained relatively motionless unless the submersible came very close (almost touching the fish). It is also unlikely that they had moved away, or toward, the submersible, before coming into the camera field. On all dives there was an observer and another video camera recording a broader area, and there was no indication that shortspine thornyheads were responding to the sub- marine. The passive behavior of shortspine thornyheads in response to submersibles has been noted previously (Krieger, 1992). In transect assessments, the probability of detecting or- ganisms typically is not equal over the entire field of view. Detection is affected by such factors as lighting, orienta- tion of the organism and its reflectivity, sea floor relief. Table 4 Factorial analysis of effect of depth and substrate (propor- tion of bottom categorized as hard substrate) on shortspine thornyhead abundance (number/100 m^, log(.v-)-l) trans- formed). Variable df Sum of squares F-value P Depth 2 1.01 10.28 <0.0001 Substrate 1 0.35 7.10 0.009 Depth-substrate interaction 2 0.41 4.17 0.018 Residual 101 4.96 suspended particles, and size of the organism. Butler et al.-^ examined the detection functions for three types of fish (flatfish, hagfish, and thornyhead) from data collected off California. For thornyheads, the probability of detec- tion was relatively constant to a distance of about 180 cm. This distance is larger than the width of the transects in our study, indicating that our use of a constant detection function (probability of 1.0) was appropriate. We found that shortspine thornyheads preferred habi- tat with hard substrate. Submersible transects have been used to identify habitat used by rockfishes and thorny- heads in the northeast Pacific Ocean (Richards, 1986; Pearcy et al., 1989; Stein et al., 1992; O'Connell and Car- lile, 1993; Ki'ieger and Ito, 1999). Off Oregon, thornyheads were included in an assemblage of fishes associated with mud bottom (Stein et al., 1992), and Pearcy et al. (1989) observed that in deep water they occurred on mud bot- tom, but in shallower water were found over both rock and mud. We have no explanation for the differences in hab- itat association between Oregon and southeast Alaska. 198 Fishery Bulletin 100(2) Our obsei-vations are based on a larger number of tran- sects than those surveyed off Oregon, but the consistency of results from the two Oregon studies suggests that the difference is not due to a low sample size there. Our ob- servation that abundance of shortspine thornyheads in- creases with depth is consistent with results from trawl surveys off Alaska, Oregon, and California (Martin and Clausen, 1995, Stark and Clausen, 1995, Jacobson and Vetter, 1996). Off Oregon, their abundance was highest in the 200-400 m depth zone, and decreased sharply between 3- 2.5- DSoft R E 2- BHard £1 o o 1 15- .Q E i 1- 0.5- 0- 48 '■ 1 7ni3 "' <200 20&-300 >300 Depth (m) Figure 3 Mean atiundance of shortspiiiL' thornyhead un soft and hard bottom substrates in three depth intervals. Sample size (number of transects 1 and standard errors are indicated m the number and vertical line over each bar 2.5 2 - E 01990 Trawl Q1991 Sub 01 993 Trawl o ? 1-5 . o 1 - 5 - 18 ^ <200 200-300 Depth (m) >300 Figure 4 Mean abundance of shortspine thornyhead in three depth inter- vals from submersible transects in 1991 and from the 1990 and 1993 trawl surveys off southeast Alaska. Sample size (number of transects) and standard errors of the submersible estimates arc indicated in the number and vertical line with each bar 400 m and 1400 m (Jacobson and Vetter, 1996). Off Califor- nia, the highest abundance of shortspine thornyheads oc- curred at 400-600 m, probably because of the warmer wa- ters off California (Jacobson and Vetter, 1996). Thus, our sampling depths covered the most important depth zones for this species in the waters off southeast Alaska, which are colder than those off Oregon. Bottom trawl sui-veys of fishes in the Gulf of Alaska are conducted triennially, and occurred one year before ( 1990) and two years after ( 1993 ) our submersible sui-vey ( Martin and Clausen, 1995; Stark and Clausen, 1995). For many species, trawl sui^vey results may be biased because some species may be herded by the trawl doors into the path of the net, resulting in overestimates of abundance when the "area-swept" method is applied (Ki'ieger, 1992). Other fish in the water column above the bottom may swim over the net and be underestimated by the survey (Balsiger et al., 1985). Escape routes under the foot rope and through the larger meshes in the trawl wings are also possibilities. The NMFS trawl survey does not cover exceptionally nig- ged rocky habitats that would destroy equipment; conse- quently, species that select high-relief habitats may be un- dei'estimated by the survey, whereas species that select low-relief soft-bottom habitats may be overestimated. Sub- mersible obsei-vations provide a means to quantify the bi- ases inherent in bottom trawl sui-veys. The depth catego- ries we used to analyze the submersible data matched the depth strata used in the NMFS triennial trawl sui-veys. Mean abundance of shortspine thornyheads in submersible sui^veys were several times higher that those in the 1990 and 1993 trawl surveys; however, the ratios of abundance in the three depth zones were very similar ( Fig. 4 ). We suggest that trawl sui-veys underestimate the abundance of short- spine thornyheads in the Gulf of Alaska but may be good indicators of relative abundance, patterns of dis- tribution, and stock trends. Acknowledgments We thank Dan Ito for his encouragement and support for this project. This research was supported by a grant (43ABNF401902) from the U.S. Dept. of Com- merce, National Marine Fisheries Ser\'ice, Auke Bay Laboratory, Juneau, Alaska. Literature cited Adams, P. B. 1980. Life history patterns in marine fishes and their consequences for fisheries management. Fish. Bull. 78:1-12 Auster, P. J., L. L. Stewart, and H. Spunk. 1989. Scientific imaging with ROVs: tools and tech- niques. Mar Technol. Soc. J. 23:16-20. Balsiger J. W.. D. H. Ito. D. K. Kimura. D. A. Somerton. and J. M Terry. 1985. Biological and economic assessment of Pacific ocean perch (Sebastea alutus ) in waters of Alaska. U.S. Dep. Commer, NOAA Tech. Memo. NMFS F/NWC-72, Else et al : Abundance of Sebastolobus alascanus in the Gulf of Alaska 199 NMFS Alaska Fish. Sci. Cent., Seattle, WA. 210 p. Jacobson, L. D.. and R. D. Vetter. 1996. Bathymctric demog^raphy and niche separation of thornyhead rockfish: Sebastolobus alascanus and Sebastol- obus altivelis. Can. J. Fish. Aquat. Sci. 53:600-609. Krieger, K. J. 1992. Shortraker rockfish, Sebastes horealis. observed from a manned submersible. Mar Fish. Rev. .54(4):34-36. 1993. Distribution and abundance of rockfish determined from a submersible and by bottom trawHng. Fish. Bull. 91 : 87-96. Kreiger, K. J., and D. H. Ito. 1999. Distribution and abundance of shortraker rockfish, Se- bastes borealis, and rougheye rockfish, S. aleutianus, deter- mined from a manned submersible. Fish. Bull. 97:264-272. Martin, M. H., and D. M. Clausen. 1995. Data report: 1993 Gulf of Alaska bottom trawl sui-vcy. U.S. Dep. Commor. NOAATech. Memo. NMFS-AFSC-59. 217 p. Matlock, G. C. W. R. Nelson. R. S. Jones, A. W. Green, T. J. Cody E Gutherz, J. Doerzbacher 1991. Comparison of two techniques for estimating tilefish, yellowedge grouper and other deepwater fish populations. Fish. Bull. 89:91-99. Miller, P. 1985. Life history of the shortspine thornyhead. Sebasto- lobus alascanus. at Cape Omnaney, S.E. Alaska. M.S. thesis. Univ. Alaska, Juneau, AK, 75 p. Moser, H. G. 1974. Development and distribution of larvae and juveniles of Sebastolobus (Pisces; Family Scorpaonidac). Fish. Bull. 72:865-884. O'Connell, V. M.. and D. W. Carlile. 1993. Habitat-specific density of adult yelloweye rockfish Sebastes ruberrimus in the eastern Gulf of Alaska. Fish. Bull. 91:304-309. Pearcy W. G.. D. L. Stein, M. A. Hixon, E. K. Pikitch, W. H. Barss, and R. M. Starr 1989. Submersible observations of deep-reef fishes of Heceta Bank, Oregon. Fish. Bull. 87:955-965. Richards, L. J. 1986. Depth and habitat distributions of three species of rockfish [Sebastes) in British Columbia: observations from the submersible PISCES IV. Environ. Biol. Fishes 17: 13-21. Stark, J. W., and D. M Clausen. 1995. Data report: 1990 Gulf of Alaska bottom trawl survey. U.S. Dep. Commer, NOAA Tech. Memo. NMFS-AFSC-49, 221 p. Stein, D. L., B. N. Tissot, M. A. Hi.xon, and W. Barss. 1992. Fish-habitat associations on a deep reef at the edge of the Oregon continental shelf Fi.sh. Bull. 90:540-551. Zar, J. H. 1984. Biostatistical analyses. Prentice-Hall. Englewood Cliffs, NJ. 718p. Zhou, S., and T. C. Shirley. 1997. Distribution of red king crabs and Tanner crabs in the summer by habitat and depth in an Alaskan fjord. Invest. Mar Valparaiso 25:59-67. 200 Abstract— Analysis of 32 years of stan- dardized survey catches ( 1967-98 ) indi- cated differential distribution patterns for the longfin inshore squid iLoligo pealeii) over the northwest Atlantic U.S. continental shelf, by geographic region, depth, season, and time of day. Catches were greatest in the Mid- Atlantic Bight, where there were sig- nificantly greater catches in deep water during winter and spring, and in shallow water during autumn. Body size generally increased with depth in all seasons. Large catches of juve- niles in shallow waters off southern New England during autumn resulted from inshore spawning observed during late spring and summer; large propor- tions of juveniles in the Mid-Atlantic Bight during spring suggest that sub- stantial winter spawning also occurs. Few mature squid were caught in sur- vey samples in any season; the major- ity of these mature squid were cap- tured south of Cape Hatteras during spring. Spawning occurs inshore from late spring to summer and the data suggest that winter spawning occurs primarily south of Cape Hatteras. Geographic and temporal patterns in size and maturity of the longfin inshore squid iLoligo pealeii) off the northeastern United States Emma M.C. Hatfield Steven X. Cadrin Northeast Fisheries Science Center National Manne Fishenes Service, NOAA 166 Water Street Woods Hole, Massachusetts 02543 Present address (for E.M.C, Hatfield) FRS Marine Laboratory Victoria Road Aberdeen ABll 9DB Scotland, United Kingdom E-mail address (for E M C Hatfield) e hatfield a marlab ac uk Manuscript accepted 17 mav 2001 Fish. Bull. 200-213 (2002). ' The longfin inshore squid, Loligo pea- leii. is distributed in the northwest Atlantic from Canada to the Carib- bean (Cohen, 19761. Within its range of commercial exploitation (from southern Georges Bank to Cape Hatteras) the population is considered to be a unit stock ( NEFC ' ), although heterogeneous subpopulations may exist (Garthwaite etal., 1989). North of Cape Hatteras, L. pealeii migrate seasonally. The migration has been described as a movement offshore during late autumn (so that the species can overwinter in warmer waters along the edge of the continental shelf ) and a return movement inshore during the spring and early summer (Summers, 1969; Serchuk and Rathjen, 1974; Tib- betts, 1977). Murawski (1993) defined L. pealeii as a member of a migratory, warm-water group of species, centered primarily in mid-Atlantic waters (par- ticularly in the spring), that make in- shore and northward migrations in the spring and offshore and southward mi- grations in late autumn. Geographic patterns in Northeast Fisheries Science Center (NEFSC ) sur- vey catches, from the Gulf of Maine to Cape Hatteras, show that L. pealeii are distributed over the entire conti- nental shelf (from inshore to offshore) in the autumn, are concentrated at the edge of the continental shelf and at the southern end of the survey area during winter and spring, and are con- centrated inshore in summer (Sum- mers, 1967; 1969; Serchuk and Rathjen, 1974; Vovk, 1978; Lange, 1980; Whita- ker, 1980; Lange and Waring, 1992). Analyses of sui-vey catches indicate that depth, time of day, and tempera- ture all influence cross-shelf distri- bution patterns (Summers, 1969; Ser- chuk and Rathjen, 1974; Lange and Warmg, 1992; Murawski, 1993; Brod- ziak and Hendrickson, 1999). Diel cor- rection factors have been applied to sui-vey indices in various studies to ad- just nighttime bottom trawl catches to daytime equivalents (daytime catches are higher when squid are concentrat- ed close to the bottom) (Lange and Sis- senwinel983; Lange and Sissenwine-). Research by Lange and Waring (1992) and Brodziak and Hendrickson (1999) demonstrated that the diel differences were size specific and that further con- sideration of these differences in cor- rection factors was warranted. Until recently, L. pealeii was thought to have a life span of up to three years, and the stock was assessed accord- ingly (Sissenwine and Tibbetts, 1977; Lange, 1981; Lange and Sissenwine, 1 NEFC (Northeast Fisheries Center I. 1986. Report of the second NEFC stock assess- ment workshop. NEFC Lab. Ref Doc. 88- 02, 114 p. [Available from NEFSC, 166 Water Street, Woods Hole. MA 02,543.1 - Lange, A. M. T, and M. P. Sissenwine. 1977. Lo/(go pea/ei stock status. North- east Fisheries Science Center Lab. Ref Doc. 77-28, 9 p. [Available from NEFSC, 166 Water Street, Woods Hole, MA 002543.] Hatfield and Cadrin: Geographic and temporal patterns in size and maturity of Loligo pcaleii 201 44"N 42" 40° 38° 1983; Lange''; Lange et al.''). Recent advances in the use of statoliths for age determination of squid (see reviews in Rodhouse and Hatfield, 1990; Jereb et al., 1991; Jackson, 1994) have enabled now esti- mates of life span to be derived for L. pcaleii ( Macy, 1995; Brodziak and Macy, 1996; Macy"*), which in- dicate that the life span of L. pealeii can be less than nine months. Back-calculations of hatching date from age data revealed that there is more than just a spring-summer spawning component of the population (Brodziak and Macy, 1996; Ma- cy''), with a small proportion of squid hatching dur- ing winter This winter spawning is presumed to occur offshore (Brodziak and Macy, 1996), in the vicinity of the submarine canyons along the edge of the northeastern U.S. continental shelf, from Hudson Canyon up to Georges Bank (Fig. 1). The possibility of winter spawning was raised initially by Summers ( 1969), based on length-frequency da- ta, but squid were not presumed to spawn until their second year because their growth was as- sumed to be too slow to allow spawning during their first summer Our study reports on two studies: 1) an analy- sis of survey data from spring and autumn NEFSC surveys from 1967 to 1998, and from winter NEF- SC surveys from 1992 to 1998, to describe gi-oss distribution patterns of L. pealeii over the north- west Atlantic continental shelf from Cape Hat- teras to the Gulf of Maine; 2) some results of a field study initiated in 1997 to investigate geographic and seasonal patterns of growth and maturity to determine if the winter spawning component off the northeastern United States can be defined by time and area. Materials and methods Survey analysis Length-frequency data for L. pealeii were analyzed from NEFSC bottom-trawl surveys conducted in the autumn (generally from mid-September to late October) from 1967 to 1997; in the spring (generally from March to early April) from 1968 to 1998; and in the winter (generally in Febru- ary) from 1992 to 1998. Data collection and processing and archiving methods are described by Azarovitz (1981). In 66'W 36 /y Hudson Canyon r/ MAB MAB - Mi(d-Atlantic Bight NE - New Englatitj GOM - Gulf of Maifie /Cape Hatteras Ja i—i , I , L_ J , I I I . L 3 Lange, A. M. T. 1984. An assessment of the long-finned squid resource ofFthe northeastern United States. Northeast Fisher- ies Science Center (NEFSC) Lab. Ref Doc. 84-.37. 24 p. [Avail- able from NEFSC, 166 Water St. Woods Hole, MA 02543.] '' Lange, A. M. T, M. P. Sissenwine, and E. D. Anderson. 1984. Yield analysis of long-finned squid, Loligo pealei (LeSueur). Northwest Atlantic Fisheries Organization (NAFO) SCR Doc. 84/1X/97, 29 p. ^ Macy W. K. 1995. Recruitment of long-finned squid in New- England (USA) waters. ICES CM 1995/K:35, 18 p. [Available from W. K. Macy, Graduate School of Oceanography, Univ. Rhode Island, South Ferry Road, Narragansett, RI 02882]. Figure 1 Map of the survey areas for longfin inshore squid off the northeastern coast of the United States 1 1967-98). the autumn and spring sui-veys the same trawl-sampling gear (Yankee-36 trawl) has been used since 1967, except during 1973-81, when a Yankee-41 (high rising) trawl was substituted in the spring surveys. In the winter sui-veys the trawl gear was larger and the Gulf of Maine was not sampled. The NEFSC sui-vey area was divided into two geograph- ic regions (the region north of Hudson Canyon to the Gulf of Maine [designated New England, NE] and the region south of Hudson Canyon to Cape Hatteras (designated Mid-Atlantic Bight [MAB]) and into four bottom depth zones (27-55 m, 56-110 m, 111-185 m, and 186-366 m [Fig. 1] ). The 1-26 m depth zone was not sampled in NEF- SC offshore surveys, but "inshore" strata were added to the survey in 1972. For the spring and winter surveys, the combined effects of annual abundance (numbers of squid per standardized trawl haul), survey stratum, and time of day (night, 20:00-03:59; dawn and dusk, 04:00-07:59, 16:00-19:59; and day, 08:00-15:59), as described by Brodziak and Hen- drickson (1999) for the autumn survey, were analyzed to determine adjustment factors for diel differences in log- transformed survey catches of prerecruit squid (<80 mm dorsal mantle length [ML], the minimum size in com- mercial catches) and recruits (>80 mm ML). The derived factors were then used to adjust all survey catches to 202 Fishery Bulletin 100(2) their daytime equivalent. The size groups (<80 mm ML, >80 mm ML) were chosen to allow comparisons with re- sults from previous studies (e.g. Lange, 1980: 1981; Lange and Sissenwine, 1983: Brodziak and Hendrickson. 1999: NEFC': Lange and Sissenwine-: Lange'^). Alternative anal- yses were performed for different size groups (<50 mm ML, >50 mm ML and <100 mm ML, >100 mm ML) to assess the sensitivity of the results to the choice of size groups. The combined effects of geographic region (NE and MAE), depth zone, and year on sui-vey catches were tested by using generalized linear models (GLM) to derive main effects and coefficients for each sui-v^ey. Paii^wise compari- sons were tested by using a Mest with Bonferroni adjust- ments (Sokal and Rohlf, 1995) to compare specific regions, seasons, and depth zones. All tests were analyzed at the 5'y( significance level. Differences between seasons and re- gions were tested between autumn and spring sui'veys for the years 1968 to 1997, between autumn and winter for 1992 to 1997, and between spring and winter surveys from 1992 to 1998. Proportion of catches <50 mm ML were ana- lyzed to evaluate the relative distribution of juvenile L. pealeii. Biological analysis Subsamples of 50-100 individuals were obtained from five different survey time series: NEFSC autumn (September- October 1997), winter (February 1998), and spring (March 1998), inshore Massachusetts (Howe^) (October 1997), and Connecticut (Johnson") (Long Island Sound, May 1998). The samples were analyzed from each of five depth zones (1-26 m, 27-55 m, 56-100 m, 111-185 m, 186-366 m), within each of three geographic regions (Gulf of Maine [GOM): Georges Bank-Southern New England, north of Hudson Canyon |SNE|: and Mid-Atlantic Bight, see above [MABl). A fourth region, south of Cape Hatteras (SOH) was added later. Each sample comprised a nonrandom selection of lengths to represent the size range present in a tow. In total, 2156 individuals were subsampled from 53 sun'ey tows. Sexes were determined and specimens were measured to enable the morphometric maturity analyses of Macy (1982): each individual squid was also weighed on a top-loading balance to 0.1 g. The morphometric method uses a suite of length measurements for female and male squid to determine maturity stage (measured on a scale of 1 to 4, where 1 is immature and 4 is fully mature). Oppor- tunistic commercial samples from early winter (December 1998 and January 1999) were also analyzed (118 individu- als) to bridge the temporal gap in sui-vey coverage. Data on dorsal mantle length (ML, mm) and total body mass (BM, g) for each maturity stage were used to esti- mate proportions for each maturity stage across the length ^ Howe, A. B. 1989. State of Massachusetts inshore bottom trawl survey. Atlantic States Marine Fisheries Conimi-~sion (ASMFC) Spec. Rep. 17:33-38. [Available from ASMFC, 1444 Eye Street, N.W., sixth floor, Washington. DC 2000.5.] " Johnson, M. 1994. State of Connecticut marine finfish trawl survey. Atlantic States Marine Fisheries Commission (ASMFC ) Spec Rep. 35:24-26. [Available from ASMFC, 1444 Eve Street, N.W., sixth floor, Washington, DC 20005.1 and weight range. These proportions were used to deter- mine the sizes at which squid of both sexes changed from one maturity stage to the next. Maturity-at-length data were weighted by diurnally ad- justed catch-at-length data for each depth zone and region to provide population-weighted maturity patterns, assuming that sui"vey length distributions accurately represent rel- ative proportions of population components. Catch-weight- ed data were analyzed to derive 1 ) the patterns of matu- rity for each sex at different times of the year: 2 ) estimates of proportions of each maturity stage sampled by survey: 3 ) mean length for each maturity stage of each sui-vey: 4 ) mean length for each region of each survey: and 5) mean length for each depth stratum of each survey. A small pro- portion (8.4^^ ) of survey catches <50 mm ML were not sub- sampled: these were assigned to the juvenile stage. Catches at larger sizes, which were not subsampled (6.8% of sui-vey catches), were removed from the analysis because sex or maturity stage could not be assigned with any degree of cer- tainty Individual squid, or size classes of squid, were not weighed during NEFSC surveys: therefore the maturity da- ta from the biological analysis could only be catch-weighted by length because length was measured on a random sub- sample of squid caught at each station in NEFSC surveys. Results Survey analysis Patterns of diurnal distribution were different among sea- sons surveyed (Table 1). In winter surveys, from 1992 to 1998, prerecruit (i.e. <80 mm ML) catch was lower at night and during dawn and dusk than during daylight hours (65'/J and 81% of daytime catch, respectively). However, for recruits (i.e. >80 mm ML), catch was higher both at night and at dawn and dusk than during the day ( 131% and 115% of daytime catch respectively) in winter sui-veys. In autumn and spring sui'veys, from 1968 to 1998. both prerecruits and recruits showed a lower catch at night and during dawn and dusk than by day: recruits showed a lesser diurnal variation than prerecruits. Results from analyses with dif- ferent size groups (<50 mm ML, >50 mm ML and <100 mm ML, >100 mm ML) were very similar, suggesting that the interaction of size and time of day is gradual. Catch rates varied significantly by season (Table 2). During winter and spring, survey catches were greater in the MAB by a factor of approximately four (Fig. 2). How- ever, there was no significant difference in sui'vey catches between geographic regions in autumn (Fig. 2). Pairwise comparisons showed that mean number-per-tow was sig- nificantly greater in autumn than in spring within both geographic regions and was greater in autumn than in winter in the NE. There were no significant differences be- tween autumn and winter means in the MAB nor between spring and winter means in either the MAB or the NE. Catch by depth, pooled over the MAB and NE, varied by season (Table 3). Pairwise comparisons of each depth for each season showed that winter and spring survey catch- es were lowest in the shallowest stratum (27-55 mi, in- Hatfield and Cadrin Geographic and temporal patterns in size and matunty of Loligo pealeii 203 Table 1 Ki'lativo catch rates for small i<80 mm dorsal mantle length (ML) 1 and large (>80 mm ML) L bottom-trawl surveys, 1967-98, by time of day (in relation to catch rates during daytime). ollfil l)C(ll< 11 in three seasonal NEFSC Winter Spring Autumn Time of day <80 mm >80 mm <80 mm >80 mm <80 mm >80 mm Night 0.65 1.30 0.51 0.72 Dawn and dusk 0.81 1.14 0.79 0.92 Day 1.00 1.00 1.00 1.00 0.09 0.46 1.00 0.34 0.83 1.00 Table 2 Results of generalized linear model (GLM) of !■ of freedom; SS=sum of squares; F=F-statistic ui-vey mean numbers-per-tow by P=probability ), for Loligo pealeii vear, depth zone, and geographic off the northeast LInited States region basec (df=degrees on NEFSC bottom-trawl sui-vey data 1967 -98. Season and effect df Type III SS Mean square F P Winter year 6 42.51 7.09 2.71 0.0133 depth 3 339.62 113.21 43.36 0.0001 region 1 233.73 233.73 89.53 0.0001 Spring vear 30 409.47 13.64 4.44 0.0001 depth 3 1696.53 565.51 183.79 0.0001 region 1 983.51 983.51 319.64 0.0001 Autumn year 30 557.85 18.59 5.17 0.0001 depth 3 1134.09 378.03 105.20 0.0001 region 1 9.12 9.12 2.54 0.1113 Table 3 Diurnally adjusted, mean numbers-per-tow of Loligo pea- leii from the three annual NEFSC bottom-trawl surveys, 1967-98, by season. Depth »ne (m) 27-55 56-110 111-185 186-366 Winter 42.9 103.1 215.5 30.3 Spring 42.6 90.7 342.5 91.9 Autumn 853.3 352.8 377.2 66.5 creased to peak values in deeper strata ( 111-185 m, great- er than 10 times the catches in the shallowest strata), and were low in the deepest stratum (>185 m). Converse- ly, autumn survey catches were highest in the shallowest stratum (27-55 m) and lowest in the deepest stratum (>185 m). These patterns were also generally significant within both geographic regions (Table 4). Table 4 Diurnally adjus tec , mean numbers-per tow of Loligo pea- \ leii from the th ree annual NEFSC bott om-trawl sui-veys. 1967-98, by area (NE=north of Hudson Canyon to the 1 Gulf of Maine; MAB=south of Hudson Canyon to Cape Hatteras. - Depth zone ( m I 27-55 56-110 111-185 186-366 Winter NE 6.0 38.9 160.4 28.7 Winter MAB 83.5 396.2 510.7 36.0 Spring NE 2.3 24.7 259.5 81.4 Spring MAB 86.8 392.3 787.3 129.7 Autumn NE 644.2 365.9 392.8 72.4 Autumn MAB 1082.5 293.1 293.7 45.5 In the pairwise comparisons of the winter and spring surveys, catches were significantly greater at 111-185 m than in other depth zones. In the 27-55 m depth zone, au- 204 Fishery Bulletin 100(2) tumn catches were considerably higher than in winter or spring. For all other depth zones and sui-vey comparisons, the differences were not significant. In the autumn survey, the proportion of small squid (<50 mm ML) was highest at 27-55 m depths (over 50% of the sampled squid in that depth zone in the NE and almost 75*^7^ of the squid in that depth zone in the MAB, Table 5). Proportions of small squid at greater depths were consid- erably lower. These patterns show higher relative recruit- ment into the population in the shallow waters of the con- tinental shelf in the autumn. Similarly in the MAB during winter and spring, small squid form a higher proportion of squid sampled in the two shallowest depth zones than at greater depths. A higher percentage of small squid was present in spring than in winter, with over GC/f of squid sampled in the MAB from '27-55 m being <50 mm ML. However, the highest propor- tion of small squid in the NE during winter and spring was at intermediate depths. Biological analysis The raw data of numbers sampled for each sex, length, and maturity stage are given in Table 6. For all seasons combined, the ML at 50'7( maturity during 1997-98 was approximately 200 mm ML for females and males (Table 7, Hatfield and Cadrin Geographic and temporal patterns in size and matunty of Loligo pealeii 205 Table 5 The percentage o( Loligo pealeii <50 mm ML in each depth zone, for each region and survey, and for the number these data were available (NE=north of Hudson Canyon to the Gulf of Maine; MAB=south of Hudson Canyon to of years for which Cape Hatteras). Depth zone (m) Winter Spring Aut umn Winter Spring Autumn NE (No. of years) NE (No. of years) NE (No. of years) MAB (No. of years I MAB (No. of years ) MAB (No. of years) 27-55 3 6 17 20 51 31 33 7 64 30 73 31 .56-110 11 7 40 31 19 31 27 7 48 31 27 31 111-185 27 7 24 31 31 31 14 7 22 31 9 31 186-366 7 3 2 31 14 30 10 3 14 31 2 31 Table 6 Numbers oi Loligo pealeii measured, for each sex and maturity stage, rity was based on a four-stage scale for sexual maturity for each sex length. from samples taken for biological analy where 1 was immature and 4 was fully sis in 1997- mature. ML 99. Matu- = mantle ML ( mm ) Females Males Stage 1 Stage 2 Stage 3 Stage 4 Total 2 Stage 1 Stage 2 Stage 3 Stage 4 Total S 30 1 1 40 5 5 2 2 50 31 1 32 18 18 60 59 3 1 63 30 5 1 36 70 73 17 1 91 32 22 2 1 57 80 57 30 1 1 89 27 37 12 1 77 90 51 45 5 101 10 45 16 4 75 100 27 51 2 3 83 2 50 17 2 71 110 10 62 5 6 83 2 35 30 8 75 120 8 86 14 4 112 1 32 40 4 77 130 3 79 6 4 92 24 37 11 72 140 3 71 10 8 92 14 39 10 63 1.50 2 38 16 14 70 14 43 10 67 160 25 6 7 38 14 28 8 50 170 13 2 7 22 12 31 7 50 180 7 2 3 12 6 18 7 31 190 9 1 2 12 3 19 9 31 200 1 5 6 5 9 14 210 1 1 1 3 2 8 11 220 1 3 4 4 7 12 230 5 6 240 1 5 7 250 1 3 5 260 1 1 270 1 1 3 3 280 1 1 290 1 1 Total 330 539 74 69 1012 124 318 345 126 913 206 Fishery Bulletin 100(2) 00 1 A 80- 60- 40- 20- 0- 200 300 nn-i E 80- A V ^ 60- / / / 40- male / / I — A 20- IL-^ — 3+4 —2+3+4 100 100 200 Mantle length (mm) 300 100 200 Body mass (g) Figure 3 Percent mature ofLoligopealeii for dorsal mantle length (ML mm) (Al and wet body mass IBM gi iB). The percentage at each maturity stage is shown for female ML (C) and BM (D) and for male ML lE) and BM iFi. Table 7 Size at maturity for Loligo pealcii in mantle length iML, mml and body mass IBM, g), by sex. i — denotes a missing value). Female Male Proportion mature ML BM ML BM 25% 166 111 50% 207 — 75% 2.38 — 184 113 196 146 241 184 Fig. 3, A and B). In terms of body mass, the size at 50% maturity for males was approximately 1.50 g, but, accord- ing to our samples, female maturity did not seem to be as closely associated with body mass (e.g. even squid in the heaviest size class were less than 50% mature). In females, maturity stage 2 was reached at a relatively small size (Fig. 3, C and D). To reach stage 3 requires a considerable increase in length or body mass, whereas fe- males in stage 4 are neither much longer, nor heavier than stage-3 females. Thus the transition from stage 3 to stage 4 (full maturity) takes place over a lesser period of somatic growth (and therefore possibly a shorter time period) than the transition from stage 2 to stage 3. In Macy's ( 1982) maturity stage notation, stage-3 females have no mature Hatfield and Cadrin Geographic and temporal patterns in size and maturity of Loligo pealeii 207 Table 8 DiLU'iially adjusti'd. t iTspond to unscxcd j (SOH: south of Cape atch-vveighted uvcnilcs and a Hatteras). mean numbers-per four-stage scale for tow o{ Li ill i^d H'xual maturi pealeii sampled in each maturity stage. Maturity stages cor- ty for each sex where 1 was immature and 4 was fully mature Juvenile Female Me le Totals Stage 1 Stage 2 Stage 3 Stage 4 Stage 1 Stage 2 Stage 3 Stage 4 Autumn 240.0 94.0 16.0 0.0 2.0 64.0 49.0 4.0 3.0 472.0 Winter 5.0 13.9 10.9 0.4 0.2 4.8 6.1 6.2 1.1 48.6 Spring plus SOH 26.3 12.7 10.0 2.3 1.4 5.4 5.0 5.0 2.4 70.5 Long Island Sound 436.0 0.0 52.0 378.0 176.0 161.0 0.0 196.0 144.0 1543.0 Commercial 0.0 5.0 155.0 0.0 0.0 8.0 198.0 331.0 15.0 712.0 Totals 707.3 125.6 243.9 380.7 179.6 243.2 258.1 542.2 165.5 2846.1 Table 9 1 Maturity patterns in Loligo pealeii (percentage of sample at each sex and stage derived from diurnally adjusted, catch-weighted mean numbers-per-tow. Maturity stages correspond to unsexed juveniles and a four stage scale for sexual maturity for each sex where 1 was immature and 4 was fully mature (SOH: south of Cape Hatteras). % juvenile Female Male % stage 1 % stage 2 % stage 3 % stage 4 % stage 1 % stage 2 7f stage 3 7f. stage 4 Autumn 50.7 20.0 3.4 0.5 13.6 10.3 0.9 0.6 Winter 10.3 28.7 22.4 0.7 0.4 9.8 12.5 12.7 2.2 Spring plus SOH 37.3 18.0 14.2 3.2 1.9 7.7 7.0 7.1 3.5 Long Island Sound 28.3 3.4 24.5 11.4 10.4 12.7 9.3 Commercial 0.7 21.8 1.1 27.8 46.5 2.1 oocytes (therefore they would not be considered to be close to maturity) and stage-4 females are fully mature. In comparison, the transition of male squid (Fig. 3, E-F) from stage 2, to stage 3, to stage 4, seems more evenly spaced, with a more gradual development seen over the course of the maturation process. If anything, there seems to be a greater transition from stage 3 to stage 4, than from stage 2 to stage 3. The difference between stage-2 and stage-3 males is measured only as elongation of the testis in conjunction with a reduction in the ratio of mantle cir- cumference to mantle length. In stage-4 males "elongate mature spermatophores are visible both in the Needham's sac and the penis." In subjective terms, stage 2 represents definitely immature, stage 3, maturing, and stage 4, fully mature males. Thus the rate at which full maturity is ap- proached is very different between the sexes. The raw data (numbers sampled in each size class, sex, and maturity stage ) were then catch-weighted to be repre- sentative of the squid sampled in the different surveys and from commercial data. Catch-weighted proportions of fe- male and male L. pealeii at each maturity stage are shown in Tables 8 and 9. In the autumn and spring surveys, and in the May Long Island Sound (LIS) samples, juvenile squid were the most abundant stage within the sampled popula- tion. In winter surveys the juveniles were one of the least abundant stages. Mature squid were never abundant in the NEFSC survey subsamples, nor in the inshore autumn (October) Massachusetts survey (combined with NEFSC data for autumn surveys). Most mature squid were seen in the LIS samples. In NEFSC surveys, the majority of both sexes were immature and stages 1 and 2. In LIS samples more squid were either stage 3 or 4 than immature. No ma- ture females were observed in the commercial samples, al- though a small proportion of males were mature. The catch-weighted mean ML for each maturity stage, by season, is given in Table 10. There was little difference between the size of juvenile squid between seasons. For both sexes, squid at stages 1, 2, and 3 were all longest in the autumn samples and shortest in the spring and LIS samples. Conversely, mature squid of both sexes (stage 4) were considerably larger in the spring than in autumn. In the LIS samples, mature feinale squid were the same size as in spring survey samples; mature male squid, on the other hand, were smaller than in the spring survey but larger than the autumn sui-vey samples. Commercial sam- ples showed a larger size for each maturity stage sampled. 208 Fishery Bulletin 100(2) Table 10 Mean dorsal mantle length (ML, mm) o{ Lotigo pealeii for each sex and maturity stage (con-esponding to four stage scale for sexual maturity for each sex where 1 was immature and 4 was fully mature), from d weighted mean numbers-per-tow data (SOH: south of Cape Hatteras). unsexed juveniles and a lurnally adjusted, catch- Juvenile Female Male Stage 1 Stage 2 Stage 3 Stage 4 Stage 1 Stage 2 Stage 3 Stage 4 Autumn 34 77 133 115 70 105 161 94 Winter 41 71 122 166 164 54 93 131 169 Spring plus SOH 37 58 90 122 139 61 83 114 143 Long Island Sound 47 90 79 138 70 83 109 Commercial 83 160 79 150 169 180 Table 11 Mean dorsal mantle length (ML mm) and perce ntage of each sex foi each sample of Loligo pcalcii by depth stratum per survey. Derived from diurnally adjusted. catch-wei ghted tow data. Depth zone Depth zone Depth zone Depth zont Depth zone <27m "'r 27-55 m % 56-110 m I'/ 111-185 m % 186-366 m % Autumn juvenile ML 32 73 29 T2 47 58 48 5 Autumn female ML 67 13 87 13 95 25 92 21 123 68 Autumn male ML 67 14 105 15 111 17 80 74 118 27 Winter juvenile ML 42 10 39 4 39 5 46 5 Winter female ML 100 68 84 63 97 47 111 47 Winter male ML 105 22 93 33 97 48 131 48 Spring juvenile ML 35 42 43 3 37 54 40 3 43 1 Spring female ML 60 32 113 61 78 26 100 47 123 48 Spring male ML 59 25 129 36 90 18 104 49 136 51 Long Island Sound juvenile ML 47 28 Long Island Sound female ML 97 39 Long Island Sound male ML 86 33 Commercial female ML 158 23 Commercial male ML 161 78 The distribution of the catch-weighted mean ML, by sex, by depth zone, for each season separately is shown in Ta- ble 11. For juvenile L. pealeii, in autumn, at 1-55 m depth, there was little difference in size at depth. In the 56-110 m depth zone, juveniles in autumn were considerably larg- er than at shallower depths. In winter, there was evidence for a slight increase in the size of juveniles with increasing depth. In spring survey samples, juveniles were similar in size across the depth range sampled. Female and male L. pealeii were generally smaller in the shallowest depth zone (1-26 m, only sampled in au- tumn and spring sui-veys), and much larger at depths greater than 185 m, for each survey. There was no clear pattern for intermediate depths. In autumn surveys, squid were generally smaller at 27-55 m depth than in deeper water. In winter and spring, however, squid at this depth were longer than at 56-185 m. The LIS samples showed larger mean sizes for each group at <27 m depth than in autumn and spring samples. In the autumn survey, squid of all maturity stages (ex- cept juveniles) were generally largest in the south (MAB) and smallest in the north (SNE and GOM, Table 12). Some of this distribution may have been an artifact of the sam- pling design because no squid were sampled in the MAB region in the 1-26 m depth zone, whereas this zone was sampled in the SNE, and was the only zone for which data were available for the GOM region. In the winter survey, the general pattern was the re- verse of that seen in the autumn. In this survey squid were generally smaller in the south (MAB) than in the Hatfield and Cadrin: Geographic and temporal patterns in size and maturity of Loligo pealeii 209 Table 12 Mean dorsal mantl e length (ML. mml by sex and maturitv stage (corresponding to unsexed juveniles and a four-stage scale for sexual maturitv for each sex, where 1 was immature and 4 was fully mai\\re)oi Loligo peateii by region per survey. Derived from | diurnallv adjusted. catch-weighted mean n umbers-per-tow data. MAB = Mid-Atlantic Bight; SNE = Georges Bank-Southern New | England, north of Hudson Canyon; GOM = Gulf of Mexico; SOH = South of Cape Hatteras. Juvenile Female Ma le Stage 1 Stage 2 Stage 3 Stage 4 Stage 1 Stage 2 Stage 3 Stage 4 Autumn MAB •23 98 132 162 69 120 172 170 Autumn SNE 38 71 130 85 69 102 147 96 Autumn GOM 33 68 65 69 71 Winter MAB 42 67 116 186 180 59 94 128 151 Winter SNE 50 80 127 148 159 52 94 134 171 Spring MAB 37 62 93 127 154 66 82 110 150 Spring SNE 41 51 95 132 151 54 88 125 183 Spring SOH 56 70 11 137 54 81 98 162 Long Island SNE 47 90 79 138 70 83 109 Commercial SNE 83 160 79 150 169 180 north (SNE only, no GOM samples were taken in the win- ter survey). The exception in winter were the few stage-3 and stage-4 females sampled. In the spring survey, the same pattern as in the winter survey was observed; squid in the south were smaller than in the north. In the spring survey, samples available from south of Cape Hatteras (the limit of the MAB samples) followed the same trend because the observed mean sizes were smaller than the IVLAB samples. The exception to this were maturity-stage- 1 squid. Discussion Survey analysis The high proportion of small squid (<50 mm ML) in the winter and spring surveys corroborated the occurrence of an early winter hatching event, documented from age data determined by squid statolith analysis (Brodziak and Macy, 1996; Macy^). Mean numbers-per-tow of juvenile squid in the MAB were considerably higher than in the NE in all seasons surveyed. Recruitment of squid into the population was highest in autumn, but juvenile squid were distributed more widely over the continental shelf in the spring. Per- haps the MAB component of the L. pealeii stock was larger because it is more stable — a result of the higher propor- tion of squid recruited into the area each winter, spring, and autumn. Murawski (1993) inferred a centering of the population in the MAB subject to the issue that portions of the stock are outside the area of the NEFSC surveys. Our data sug- gested that the area south of Cape Hatteras may play an important role in reproductive dynamics and recruitment to the population, suggesting that a considerable portion of the stock is south of the surveyed area, particularly dur- ing winter and spring. South of Cape Hatteras a second loliginid species, L. plei. is abundant (Roper et al., 1984). In our study, all Loligo specimens were examined carefully to ensure that only L. pealeii were measured and included in the biological analyses. Diel differences in catches of L. pealeii have been ob- sei-ved in a number of studies (Summers, 1969; Serchuk and Rathjen, 1974; Roper and Young, 1975; Sissenwine and Bowman, 1978; Lange and Sissenwine, 1983; Lange and Waring, 1992), where catches were consistently high- er in daytime than at night. To account for diel effects on minimum swept-area estimates of L. pealeii biomass and stock size, nighttime catches were adjusted to daytime equivalents by using the diel correction factors of Lange and Sissenwine (1983). However, these correction factors were not size specific. Brodziak and Hendrickson (1999) applied size-specific diel correction factors to squid from the autumn sui-vey (1967-94), splitting the data into pre- recruits (<80 mm ML) and recruits (>80 mm ML). In the autumn surveys the nighttime catch of prerecruits was only 8.79c of the daytime catch. The nighttime catch of recruits was 34'7( of the daytime catch. These differences were attributed to the different feeding behavior of juve- nile and adult squid, in that juvenile squid might need to undertake more vertical migrations at night to meet their higher metabolic requirements. In our present study, pre- recruit nighttime catch differed to a lesser degree from the daytime catch in the winter and spring (65% in the win- ter surveys, and 51% in the spring surveys) than in the autumn. In the winter surveys, nighttime catches of re- cruits exceeded daytime catches by 31%. In spring, night- time tows showed a lower catch of recruits, 72% of day- time values, than in winter. Patterns of diel differences reported in another study by Lange and Waring (1992) for spring catches were similar to those in our study. Their 210 Fisher/ Bulletin 100(2) reported autumn catches (Lange and Waring, 1992) were higher at night than in our study, but still lower than day- time catches. The behavior of squid at both prerecruit and recruit sizes therefore appears to be different in the winter and spring than ui the autumn. The prerecruit nighttime catch in winter and spring was half, or more than half of day- time catches, as opposed to 9% in autumn. For recruits, there was an even gi-eater difference among seasons. In winter, almost 1.5 times as many squid were caught at night, than by day. In spring the nighttime catch was T2''k of the daytime catch, twice the proportion of the autumn catch. Vovk (1978; 1985) and Maurer and Bowman ( 1985) have documented large changes in the diet of squid in dif- ferent seasons, relating these changes in feeding activity and dietary preference to changes in the size composition of the squid population, movements of squid in search of food concentrations, seasonal abundance of prey, and envi- ronmental conditions that affect both prey and predator Vovk (1985) noted that in autumn L. pealeii are daytime predators and do not feed extensively at night when they occur at shallow depths. In autumn, squid are more abun- dant near the seafloor by day (Brodziak and Hendrickson, 1999). Vovk (1985) also noted that feeding activity was generally low from December to April and related this to possible prey abundance. Perhaps the different diel behav- ior patterns of L. pealeii in winter and spring, when they appear to be more available to capture by bottom-trawls, are related to a lower prey abundance and a requirement for more time to be spent searching for prey. Some of these differences might be temperature related as well. In the autumn, with strong vertical stratification of the water column, there may be some physiological benefit for squid to move off the bottom at night into warmer waters. In winter and spring when there is no vertical stratification, no advantage is conferred by a strong diel migration as seen in the autumn. Comparisons in performance and catchability of the two trawl nets used in the spring and autumn sui-\'eys (Yankee 36 and Yankee 41) have been conducted (Sissenwine and Bowman, 1978), but any differences in catchability were confounded by diel and vessel differences. In our study, we corrected for diel differences, and vessel differences were not found to be significant (NEFSC/'*). Squid catches are more abundant in the autumn sur- veys than in winter and spring (Serchuk and Rathjen. 1974; Lange and Waring, 1992; Lange and Sissenwine-). This difference may be related to the recruitment of large numbers of small squid, present m shallow water, in au- tumn. Temperature, however, is a major factor limiting distribution. In winter and spring much of the continental shelf water is below the preferred temperature minimum for the species (ca. 8°C) (Summers, 1969; Serchuk and Rathjen, 1974; Murawski, 1993); therefore squid are ap- parently less abundant than in the autumn surveys when *• NEFSC (Northeast Fisheries Science Center). 1996. Report of the '21st Northeast Regional Stock Assessment Workshop (SAW 21). Center Reference Document 90-05, 200 p. jAvail- able from NEFSC, 166 Water St., Woods Hole, MA 02543.1 temperature is not a limiting factor The greatest differ- ences between autumn and spring or winter sui-veys are most apparent in the NE, whei-e a large portion of the stock may be outside the sui-veyed area owing to tem- perature limits (e.g. off the shelf or south of Cape Hat- teras). In the MAB there are few differences in catches be- tween spring and autumn at depths deeper than 27-55 m. Spring numbers are higher than autumn numbers from 111 to 185 m. The patterns between autumn and winter catches are similar to each other at that depth. Catches are always lower in winter than in autumn, but the differ- ences arc only significant from 27-110 m depth. In winter and spring sui-veys, catches were highest from 1 1 1 to 185 m, both in the NE and the MAB. Catches were higher, in both surveys, in the MAB than in the NE. These patterns also were obsei-ved for L. pealeii in survey anal- yses from 1967 to 1971 (Summers, 1969; Serchuk and Rathjen, 1974), from 1970 to 1977 (Lange, 1980), and from 1975 to 1986 (Lange and Waring, 1992). This finding may imply that geographic distribution is relatively stable for L. pealeii. during February and March, at least within the areas sui-veyed. In autumn sui-veys. Serchuk and Rathjen (1974). Lange (1980). and Lange and Waring (1992) also reported highest catches in the MAB. However, Serchuk and Rathjen ( 1974) showed a relatively higher abundance of squid taken from 56-110 m depth than that obsei-ved in our study, where mean numbers per tow in the depth zone 27-55 m were greater than three times higher than those in the zones 56-110 m and 111-185 m. The difference be- tween the 27-55 m zone and deeper strata was less notice- able in the NE, although mean numbers-per-tow were al- most twice as high in the 27-55 m zone. The differences may result from diurnal adjustments, or from the inclusion of more years of data in the sui-vey database. Lange ( 1980) found mean numbers-per-tow in the NE autumn to be high- est from 111 to 185 m depths. Biological analysis We found that squid mature at greater lengths than pre- viously reported for L. pealeii (NEFSC"*). Figure 3 (C-F) shows that using stage 3 or greater to indicate maturity may be an adequate proxy for females. For example, the size at which 50*^^ of females are mature is 198 mm with stages 3 and 4, and 207 mm with only stage 4 (Fig. 3C). Such a proxy may be valuable for samples with few obser- vations of stage-4 females (e.g. the body mass at 50% maturity is 120 g with stages 3 and 4 [Fig. 3D1 ). For males, however, the size difference between stage 3 and stage 4 is considerable, and substantial somatic growth is required to develop from stage 3 to stage 4. Com- bining the two maturity stages in males is therefore un- supported biologically, and the combined data would un- derestimate size at 50'^r maturity. That mature squid are largest in winter and smallest in autumn samples (and intermediate in size in the early INEFSCl and late |LIS| spring) has been noted previous- ly (Summers, 1971; Lange, 1980; Macy, 1980). Prior to the availability of age data for the species, the size dif- ferences at maturitv were ascribed to different year class- Hatfield and Cadrin Geographic and temporal patterns in size and maturity of Loligo pealeii 211 es (Summers, 1971; Lange, 1980; Macy 1980). The obser- vation that immature squid are larger in autumn than in winter and spring was documented by Lange (1980) but not interpreted. Some of this variabiHty might be ex- plained as a function of temperature. If L. pealeii have a life-span of 9-12 months (Brodziak and Macy, 1996), then females that are mature in September-October sam- ples would have hatched between November and January. Females that are mature in March would have hatched around May or June. Brodziak and Macy (1996) showed that squid hatched between November and April had a lower growth rate than those that hatched between May and October. Recent laboratory studies on small L. pealeii have indicated that squid grow significantly faster at high- er temperatures (Hatfield et al., 2001), in accord with results described for other cephalopod species (Octopus bimaculoideg — Forsythe and Hanlon. 1988; L. forbesi — Forsythe and Hanlon. 1989; Sepia officinalis — Forsythe et al., 1994). These laboratory studies also found that the effect of temperature on growth is most pronounced dur- ing the early life cycle of these cephalopods, nominally the first three months. If temperatures experienced by L. pealeii hatching in May and June are warmer than tem- peratures experienced by squid hatching from November to January, then the gi'owth potential will be lower for winter-hatching squid (seen as mature squid in autumn sui^veysi, resulting in the obser\'ed lower size at maturity in autumn versus winter and spring. The same phenom- enon would explain the size differences of immature squid among seasons. The large immature squid caught in the autumn survey are probably the same squid that become the large, mature squid in the winter, spring, and LIS sam- ples. The small immature squid in the spring are probably those squid which become the small mature squid seen in autumn survey samples. If the winter-hatching squid are from southern spawning events, then the temperature difference between winter-spawned and summer-spawned squid may not be very large. However, age data for L. pealeii show that growth rates are generally slower for winter hatched squid. Also, growth studies on L. forbesi have shown that a temperature difference of just 1"C can change growth rates of squid by 2% body weight/day and produce a threefold difference in weight at 90 days after hatching (Forsythe and Hanlon, 1989). The high numbers of juvenile squid in the autumn sur- vey were from protracted spawning in inshore waters some 4-5 months previously (documented since Verrill [1882] first reported inshore spawning). The high propor- tion in spring therefore reflects a period of spawning, pos- sibly also some 4-5 months earlier, around September or October of the previous year. Juvenile squid in winter and spring survey samples denote the presence of a hatching component other than the main inshore autumn compo- nent. The high proportion obser\'ed in LIS (May) samples probably reflects an extended winter-spring spawning pe- riod because squid of the size found in Long Island Sound in May could not have been the result of that seasons in- shore spawning. The inshore spawning season does not usually begin until late April and incubation time may require up to 4 weeks at the temperatures at that time of year (27 days at 12°C, McMahon and Summers, 1971). Brodziak and Macy ( 1996) showed a pattern of year-round hatching, which is consistent with the patterns suggested by our data. Summers (1969), Serchuk and Rathjen ( 1974), and Brod- ziak and Hendrickson (1999) all reported an increase in size of L. pealeii with increasing depth. We found the smallest squid in the shallowest water and the largest squid in the deepest water, confirming that nearshore wa- ters of the continental shelf are a preferred habitat for ju- venile L. pealeii during the autumn (as described by Brod- ziak and Hendrickson [1999]). The pattern of ontogenetic descent exhibited by other loliginid species (L. vulgaris — Worms, 1983; L. gahi—Hatfiek\ et al., 1990; L. vulgaris reynaudii — Augustyn et al., 1992) is consistent in L. pea- leii, but less marked at intermediate depths. In winter and spring, mean length is generally higher in the NE and lower in the MAB. In autumn, mean length is generally lower in the NE and higher in the MAB. Lange (1980) showed a similar pattern for immature females from the autumn survey. Males, however, showed the op- posite pattern. Langes (1980) winter and spring survey data showed the same pattern as our study for immature females and all males, except the fully mature squid. Commercial samples from early winter (December and January) contained no mature female squid that might produce the winter hatching component evident from age data and aggi-egated sui-vey length-frequency distribu- tions. Egg masses are only found consistently in one small offshore area (off Chesapeake Bay) by commercial fisher- men in the early winter (see Fig. 4). However, the com- mercial samples were from the southern edge of Georges Bank, and the survey data suggest that the winter recruit- ment originates from the southern part of the MAB. The scarcity of mature squid in NEFSC sui-vey samples sug- gests that sampling did not occur consistently in the right areas or seasons to identify major spawning peaks. Whita- ker 1 1978) documented that about 40% of males were fully mature in January and February in the region south of Cape Hatteras, off the coast of South Carolina. There was a large proportion of mature squid in samples from March and April, with 74% of females and 56% of males fully ma- ture. There is evidence for both spawning and hatching in the SNE from March to April; as in 1999, egg masses were caught incidentally from northeast of Hudson Can- yon and up towards the southern flank of Georges Bank, at depths of about 200 m, from mid-March to late April (StommelP). In the 1998 Massachusetts spring survey, in mid-May, a high abundance of small squid, about 30 mm ML, were caught south of Marthas Vineyard, at depths of <27 m (senior author, personal obs.). These observations suggest that spawning is probably protracted, from early to late winter, and early winter spawning is more dom- inant in the southern end of the U.S. continental shelf Thus, the available fishery-dependent samples may not be indicative of the population. The information on age and maturity in Brodziak and Macy (1996) was derived from ^ Stommell, M. 1999. Personal, comniun. ¥W Nobska, Woods Hole, MA 02543. 212 Fishery Bulletin 100(2) data collected from the winter fishery, most of which occurs north of the MAB region. A more structured design is required to address some of these issues. The entire size and maturity stage range needs to be sampled across the geographic range of each survey and these data should be augmented with opportunistic sampling outside the area or time frame in which the sui-veys are carried out. In summary, results from these two studies complement each other to reveal patterns of re- productive dynamics for L. pealeii that have been suggested previously in other studies but that have never thoroughly been investigated. The high frequency of small squid present in spring sui-vey catches indicates that winter spawning is indeed an important component of reproduction for the population, and biological analyses sug- gest that this winter spawning occurs primarily south of Cape Hatteras, rather than in the vicin- ity of the offshore submarine canyons along the edge of the northeastern U.S. continental shelf from Hudson Canyon up to Georges Bank. Acknowledgments 76" 44"N 42" 40" 36= i\. 74" TTTJ- 72° -J- 70° 68° 66 W ( V '^1 V \ / f May-Jul f Jun-Aug A-' Hudson Canyon ■s. Mar-Apr . May-Jun // 38"k..>-y We would like to extend our thanks to John Gal- braith of NEFSC. Ai-nie Howe of the Massachu- setts Division of Marine Fisheries, Dave Simpson of the State of Connecticut Marine Division, for coordinated field sampling from sui-veys; Glenn Goodwin and Gier Monsen of Seafreeze and Joe Mantineo of Ruggierio Seafoods for commercial samples. Lars Axelson, Glenn Goodwin, and Jim Ruble shared their obsei^vations on spawning grounds of Loligo pealeii with us. Chad Keith and Lynette Suslowicz provided technical help in the cutting room. Jon Brodziak provided the code for. and assistance with, the diel correction analyses. We would like to thank John Boreman. Steve Murawski, and Fred Serchuk for their thoughtful and instructive reviews of this manuscript. EMCH would like especially to acknowledge Steve Murawski's guidance and help throughout the course of her Research Associate- ship with NMFS. We also thank other scientific personnel on NEFSC, MA, and CT bottom-trawl surveys and last, but most definitely not least, the thousands of squid that have sacrificed their lives for the advancement of sciencel Fund- ing for EMCH was provided though the National Research Council Research Associateship program. Literature cited Augustyn. C. J.. M. R. Lipinski, and W. H. H. Sauer. 1992. Can the Loligo squid fishery be managed effectively? A synthesis of research on Loligo vulgaris reynaudii. In Benguela trophic functioning (A. I. L. Payne, K. H. Brink, K. H. Mann, and R. Hilborn. eds.i. p. 90.3-918. S. Afr. J. Mar. Sd. 12. f V >. i*, Dec-Jan -May Atlantic Ocean Figure 4 Persistent spawning areas and seasons for Loligo pealeii. as indicated from incidental catches of eggs in commercial squid trawls. Azarovitz, T. R. 1981. A brief historical review of the Woods Hole Laboratory trawl survey time series. Can. Spec. Publ. Fish. Aquat. Sci. 58:62-67. Brodziak, J. K. T, and L. C. Hendrickson. 1999. An analysis of environmental effects on survey catches of squids. Loligo pealeii and Ille.x illecehrotiua in the north- west Atlantic. Fish. Bull. 97:9-24. Brodziak. J. K. T. and W. K. Macy III. 1996. Growth of the long-finned squid. Loligo pealei. in the northwest Atlantic. Fi.-^h. Bull. 94:212-236, Cohen, A. C. 1976. The systematics and distribution of Loligo i Cephalop- oda, Myopsida) in the western North Atlantic, with descrip- tions of two new species. Malacologia 15:299-67. Forsythe, J. W., and R. T. Hanlon. 1988. Effect of temperature on laboratory growth, repro- duction and life span of Octopun himaculoides. Mar. Biol. 98:369-379. 1989. Growth of the Eastern Atlantic squid. Loligo forbesi Steenstrup (Mollusca: Cephalopoda). Aquaculture Fish. Manag. 20:1-4. Forsythe. J. W., R. H. DeRusha. and R. T. Hanlon. 1994. Growth, reproduction and life span of Sepia officina- lis (Cephalopoda: Mollusca) cultured through seven consec- utive generations. J. Zool. 233:175-192. Hatfield and Cadnn: Geographic and temporal patterns in size and maturity of Loligo pealeii 213 Garthwaite, R. L., C. J. Berg, and J. Harrigan. 1989. Population genetics of the common squid Loligo pealei LeSueur. 1821. from Cape Cod to Cape Hatteras. Biol. Bull. 177:287-294. Hatfield, E. M. C, R. T. Hanlon, J. W. Forsythe, and E. P. M. Grist. 2001. Laboratory testing of a growth hypothesis for juve- nile squid Loligo pealeii (Cephalopoda: Loliginidae). Can. J. Fish. Aquat. Sci. 58:845-857. Hatfield, E. M. C, P G. Rodhouse, and J. Porebski. 1990. Demography and distribution of the Patagonian squid {Loligo gahi. d'Orbigny) during the austral winter J. Cons. Int. Explor. Mer 46:306-312. Jackson, G. D. 1994. Application and future potential of statoHth incre- ment analysis in squids and sepioids. Can. J. Fish. Aquat. Sci. 51:2612-2625. Jereb, P. S., S. Ragonese. and S. von Boletzky. 1991. Squid age determination using statoliths: proceedings of the international workshop held in the Istituto di Tec- nologia della Pesca e del Pescato (ITPP-CNR), Mazara del Vallo, Italy 9-14 October 1989. N.T.R.-I.T.RP (Note Tec- niche e Reprints dellTnstituto di Tecnologia della Pesca e del Pescato) special publication 1, 128 p. Lange,A. M.T. 1980. The population dynamics of the squids {Loligo pealei and Illex illecebrosus) from the northwest Atlantic. M.S. thesis, Univ. Washington, Seattle, Wa, 178 p. 1981. Yield-per-recruit analyses for squid, Loligo pealei and Illex illecebrosus. fi-om the Northwest Atlantic. J. Shell- fish Res. 1:197-207. Lange, A. M. T, and M. P. Sissenwine. 1983. Squid resources of the northwest Atlantic. In Ad- vances in assessment of world cephalopod resources. FAO Fish. Tech. Pap. 231:21-54. Lange, A. M., and G. T Waring. 1992. Fishery interactions between long-finned squid (Lo//go pealei) and butterfish tPeprilus triacanthiis) off the north- east USA. J. Northwest Atl. Fish. Sci. 12:49-62. Macy W. K. 1980. The ecology of the common squid, Loligo pealei LeSueur, 1821, in Rhode Island waters. Ph.D. diss., Univ. Rhode Island, Narragansett, RI, 249 p. 1982. Development and application of an objective method for classifying long-finned squid, Loligo pealei. into sexual maturity stages. Fish. Bull. 80:449-459. 1995. The application of digital image processing to the aging of long-finned squid, Loligo pealei, using the stato- lith. In Recent developments in fish otolith research (D. H. Secor, J. M. Dean, and S. E. Campana, eds.), p. 283-302. Univ. South Carolina Press, Columbia, SC. Maurer, R.O., and R. E. Bowman. 1985. Food consumption of squids t Illex illecebrosus and Loligo pealei) off the northeastern United States. NAFO (Northwest Atlantic Fisheries Organization! Sci. Coun. Studies 9:117-124. McMahon. J., and W. C. Summers. 1971. Temperature effects on the developmental rate of squid (Loligo pealei) embryos. Biol. Bull. 141:561-567. Murawski, S. A. 1993. Climate change and marine fish distributions: fore- casting from historical analogy. Trans. Am. Fish. Soc. 122: 647-658. Rodhouse, P.G., and E.M.C, Hatfield. 1990. Age determination in squid using statolith growth increments. Fish. Res. 8:323-334. Roper, C. F. E., M. J. Sweeney, and C. K. Nauen. 1984. FAO species catalogue. Vol. 3. Cephalopods of the world. An annotated and illustrated catalogue of species to interest to fisheries. FAO Fish. Synop. ( 125) Vol. 3, 277 p. Roper, C. F E., and R. E. Young. 1975. Vertical distribution of pelagic cephalopods. Smith- sonian Contributions to Zoology 209, 51 p. Serchuk, F M., and W. F Rathjen. 1974. Aspects of the distribution and abundance of long- finned squid, Loligo pealei. between Cape Hatteras and Georges Bank. Mar Fish. Rev. 36:10-17. Sissenwine, M. P., and E. W. Bowman. 1978. An analysis of some factors affecting the catchability of fish by bottom trawls. ICNAF (Int. Comm. for the Northwest Atlantic Fisheries) Res. Bull. 13:81-87. Sissenwine, M. P., and A. M. Tibbetts. 1977. Simulating the effect of fishing on squid i Loligo and Illex) populations off the northeastern United States. ICNAF (Int. Comm. for the Northwest Atlantic Fisheries) Sel. Pap. 2:71-74. Sokal, R. R. and F J. Rohlf 1995. Biometry, the principles and practice of statistics in biological research. W. H. Freeman and Company, New York., NY, 887 p. Summers, W.C. 1967. Winter distribution of Loligo pealei determined by exploratory trawling. Biol. Bull. 133:489. 1969. Winter population of Loligo pealei in the mid-Atlantic bight. Biol. Bull. 137:202-216. 1971. Age and gi-owth of Loligo pealei. a population study of the common Atlantic coast squid. Biol. Bull. 141:189-201. Tibbetts, A. M. 1977. Squid fisheries (Loligo pealei and Illex illecebrosus) off the northeastern coast of the United States of America, 1963-1974. ICNAF (Int. Comm. for the Northwest Atlan- tic Fisheries) Sel. Paps. 2:85-109. Verrill, A. E. 1882. Report on the cephalopods of the northeastern coast of America. U.S. Comm. Fish and Fish., part 7, Rep. Comm. 1879:211-450. Vovk, A. N. 1978. Peculiarities of the seasonal distribution of the North American squid Loligo pealei (Lesueur 1821). Malac. Rev. 11:130. 1985. Feeding spectrum of Longfin squid (Loligo pealei) in the northwest Atlantic and its position in the ecosystem. NAFO (Northwest Atlantic Fisheries Organization) Sci. Coun. Studies 8:33-38. Whitaker, J. D. 1978. A contribution to the biology of Loligo pealei and Loligo plei (Cephalopoda, Myopsidal off the southeastern coast of the United States. M.S. thesis. College of Charles- ton, Charleston, SC, 165 p. 1980. Squid catches resulting from trawl surveys off the Southeastern United States. Mar Fish. Rev. 42:39-43. Worms, J. 1983. Loligo vulgaris. /;i Cephalopod life cycles, vol. I, spe- cies accounts. Academic Press, London, 475 p. 214 Abstract— Samples of the commer- cially and recreationally important West Australian dhufish iGlaucosoma Iwbra- icum) were obtained from the lower west coast of Australia by a variety of methods. Fish <300 mm TL were caught over flat, hard substrata and low-lying limestone reefs, whereas larger fish were caught over larger limestone and coral reef formations. Maximum total lengths, weights, and ages were 981 mm, 15.3 kg, and 39 years, respectively, for females and 1120 mm, 23.2 kg, and 41 years, respectively, for males. The von Bertalanffy growth curves for females and males were significantly different. The values for L^, k, and t„ in the von Bertalanffy growth equations were 929 mm, 0. Ill/year, and -0.141 years, respectively, for females, and 1025 mm, 0.111/year, and -0.052 years, respectively, for males. Preliminary estimates of total mortality indicated that G. hebraicum is now subjected to a level of fishing pressure that must be of concern to fishery man- agers. Glaucosoma hebraicum, which spawns between November and April and predominantly between December and March, breeds at a wide range of depths and is a multiple spawner The /.^(I's for females and males at first maturity, i.e. 301 and 320 mm, respec- tively, were attained by about the end of the third year of life and are well below the minimum legal length (MLL) of 500 mm. Because females and males did not reach the MLL until the end of their seventh and sixth years of life, respectively, they would have had. on average, the opportunity of spawning during four and three spawning sea- sons, respectively, before they reached the MLL. However, because G. hebra- icum caught in water depths >40 m typically die upon release, a MLL is of limited use for consei-\'ing this spe- cies. Alternative approaches, such as restricting fishing activity in highly fished areas, reducing daily bag limits for recreational fishermen, introducing quotas or revising specific details of cer- tain commercial hand-line licences (or doing both) are more likely to provide effective conservation measures. Age and size composition, growth rate, reproductive biology, and habitats of the West Australian dhufish (Glaucosoma hebraicum) and their relevance to the management of this species Sybrand A. Hesp Ian C. Potter Norman G. Hall Centre for Fish and Fisheries Research School ol Biological Sciences and Biotechnology Murdoch University South Street Murdoch, Western Australia 6150, Australia Email address (for I C Potter, contact author) i-potlenSpossum murdocti edu au Manuscript accepted 3 October 2001. Fish. Bull. 100:214-227 (2002). The West Australian dhufish (Glauco- soma hebraicum ), also known as the Westrahan jewfish (McKay, 1997), is one of the most commercially valuable and recreationally sought after finfish in Western Australia ( Sudmeyer et al. ' ). This species is confined to southwest- ern Australia, where its distribution ranges southwards from Shark Bay at 26°00'S, 113°00'E down the west coast and then eastwards along the south coast to the Recherche Aj-chipelago at 34°10'S, 122n5'E(HutchinsandThomp- son. 1995). Glaucosoma hebraicum is one of four members of the monogene- ric family Glaucosomatidae, which also includes the pearl perch (Glaucosoma scapulare) that is fished commercially and recreationally in eastern Australia (McKay. 1997). Despite the high quality of the flesh of Glaucosoma species and the commercial and recreational importance of G. hebra- icum in particular, there are no refer- eed papers on the biology of any species in this genus. Furthermore, the results of undergi-aduate studies on the biology of G. hebraicum, which were collated by Sudmeyer et al.,' included estimates of age that were based on the number of growth zones in whole otoliths, an ap- proach that would almost certainly have underestimated the age of many older fish (see "Results" section). Numerous commercial and recre- ational fishermen report that they now fish much farther offshore in order to obtain catches of dhufish comparable with those they had previously been able to obtain from nearer the coast. This consistent circumstantial evidence strongly suggests that the abundance of G hebraicum in more inshore waters has declined in recent years and that this is particularly the case in areas near the city of Perth where this spe- cies has been targeted by recreational fishing crews. The indications that the abundance of dhufish in nearshore wa- ters was declining led the Australian Fisheries Research and Development Corporation to fund the current study, with a view to producing biological da- ta for managing G hebraicum . During the present study, we first determined whether the otoliths of G hebraicum had to be sectioned to de- tect all of their opaque zones and we then validated that these opaque zones are formed annually and could thus be used to age this species. As accurate estimates of the age of a fish are de- pendent on a reliable birth date, the trends exhibited by reproductive vari- ables were used to estimate the dura- tion of the spawning period and, in par- ticular, when spawning activity peaked. The age of each fish was then deter- mined and the resulting length-at-age Sudmeyer, J. E.. D. A. Hancock, and R. C. J. Lenanton. 1992. Synopsis of Westralian jewfish iGlaucosoma hebraicum^ (Richard- son, 1845 1 (Pisces: Glaucosomatidae). Fish- eries Research Report 96, Western Austra- lian Marine Research Laboratory, Fisheries Western Australia, PO Box 20, North Beach 6020, Western Australia. Hesp et a\ Age and size composition, growth rate, reproductive biology, and habitats of Glaucosoma hebraicum 215 data employed to determine the age compositions and fiirowtli rales ot female and male G. hebraicum. The repro- ductive variables were also used to help ascertain where (1. hcbr'ciiciim spawns and whether this species is a mul- tiple spawner sciisu deVlaming (1983), i.e. whether indi- vidual females release eggs on more than one occasion in a spawning season. The lengths of both sexes at first ma- turity were calculated to determine whether they lay be- low the minimum legal length (MLL) of 500 mm and were thus appropriate for helping to conserve this species. The ages at which females and males mature were also deter- mined in order to elucidate whether fish might spawn in one or more spawning seasons before they attain the MLL. Attempts were also made to ascertain the types of habitat occupied by G. hebraicum at different stages during its life cycle and to obtain preliminary mortality estimates which could be used as an indicator of whether this species is being lightly or heavily fished. Finally, the data collected during this study were used to discuss ways in which the fishery for G. hebraicum might be managed most appro- priately in the future. Materials and methods Glaucosoma hebraicum, that were less than the MLL of 500 mm total length (TL), were collected between May 1996 and June 1999 by commercial trawls, hand-lines, and a recreational spear diver under a research collection permit issued by Fisheries Western Australia — the gov- ernment agency responsible for managing the fishery for this species. Filleted carcasses of G. hebraicum >500 mm TL, together with their gonads, were obtained monthly between May 1996 and April 1998 from commercial fish- processing plants and weigh-ins at local recreational fish- ing club competitions. These fish had been caught by commercial or recreational rod and hand-lines along the lower west coast of Australia between Mandurah (32°32'S) and the Houtman Abroholos <28°35'S), i.e. within that part of the distribution of G. hebjxiicum where this spe- cies is considered to be most abundant and is most heavily fished. The total length of each fish was measured to the near- est 1 mm and the weight of each fish <500 mm was weighed to the nearest 1 g. The weights of 334 females and 442 males >500 mm TL were weighed to the nearest 10 g prior to filleting. The relationship between total length (L) in mm and total wet weight ( W) in g of each sex was Females logW = logO.0000417 + 2.859 logL (/(=486, r'^=0.995) Males logW = logO.0000322 -f- 2.898 logL (w=572, r2=0.995). These relationships were then used to estimate the weights of the female and male fish that had been filleted but not weighed. Note that all of the logarithm values recorded in this paper are natural logarithms. On several occasions, a video camera, attached by cable to a television monitor and video recorder, was lowered over the substrata during commerical hand-line fishing for dhufish. Video footage of the substrate over which dhufish were caught was later examined to determine the types of habitat occupied by this species. Age determination The two sagittal otoliths of each fish were removed, cleaned, dried, and then stored in paper envelopes. All sagittal oto- liths were sectioned, except for those which, when placed in methyl salicylate and examined microscopically under reflected light against a black background, could clearly be seen to possess either no opaque zones or only a single opaque zone. However, because the opaque zones in the whole otoliths of large fish were so numerous and closely spaced that they were often difficult to distinguish from one another and because previous estimates of the age of dhufish were based on counts of opaque zones in whole otoliths (Sudmeyer et al.'), the number of opaque zones visible in 100 otoliths, obtained from a wide size range of fish, were compared prior to and after sectioning to ascer- tain whether sectioning increased one's ability to detect the opaque zones. For sectioning, the otoliths were mounted in clear ep- oxy resin and cut into 500 pm sections with a low-speed diamond saw (Buehler). The sections were cleaned and mounted on slides with DePX mounting medium and ex- amined under reflected light with a dissecting microscope attached to a video camera (Panasonic WV-CD20). The im- age was analyzed by using the computer imaging package Optimas 5 (Optimas, 1995). The number of opaque zones in each otolith was always counted twice and on different days and without knowledge of either the date of capture or the size of the fish from which the otolith came, and also, in those cases where the two counts differed, on a third occasion. Although the number of times that a third count did not agree with either of the two previous counts was negligible for otoliths with less than 15 opaque zones, such disagreement increased to ca. 10% for otoliths with 15-25 opaque zones and ca. 30% for those with more than 25 opaque zones. When a third count was necessary and was not the same as either of the two previous counts, fur- ther counts were made until successive counts did not dif- fer by more than two opaque zones. On such occasions, the final count was recorded. An independent reader counted the number of opaque zones on 110 otoliths from a wide size range offish. Eighty four percent of the counts of the number of opaque zones made by this independent reader were the same as those of the senior author for 50 sectioned otoliths that had been judged by the senior author to have up to 10 such zones and, in those cases where there were discrepancies, the dif- ferences were never more than one opaque zone. Eighty percent of the counts made by the independent reader of the number of opaque zones on 50 sectioned otoliths re- corded as possessing between 11 and 25 such zones by the senior author were the same or differed by only one from those of the senior author and, where there were discrepan- cies, these rarely exceeded three opaque zones. In the case of ten otoliths with >25 opaque zones, the maximum dis- 216 Fishery Bulletin 100(2) crepancy between the counts recorded by the independent reader and the senior author was five. After consultation, it was agreed that, in many of the cases where the counts differed by one opaque zone, the independent reader had failed to discern the outermost opaque zone at the periph- ery of the otohths. Moreover, the extent of any discrepancy between the counts of the independent reader and the se- nior autiior declined if the independent reader continued to recount the number of opaque zones on the otoliths. Validation that the opaque zones in the otoliths of G. he- braiciim are formed annually was carried out by analyzing the trends exhibited throughout the year by the marginal increments on whole otoliths, when only one opaque zone was present, and on sectioned otoliths when two or more opaque zones were present. For this purpose, the marginal increment on each otolith, i.e. the distance between the out- er edge of the single or outermost opaque zone and the edge of the otolith, was expressed either as a proportion of the distance between the primordium and the outer edge of the opaque zone, when only one opaque zone was pres- ent, or as a proportion of the distance between the outer edges of the two outermost opaque zones, when two or more opaque zones were present. Each of the above requisite dis- tances was measured perpendicular to the opaque zone(s) and without knowledge of the date of capture of the fish and was recorded to the nearest 0.01 mm by using Optimas 5. The values for the marginal increments were separated into groups according to the number of opaque zones on the otoliths, i.e. 1, 2-5, 6-8, 9-11 etc., after which the values for each of those groups in each corresponding month of the year between May 1996 and April 1998 were pooled. Von Bertalanffy growth equations The time when spawning peaked was estimated from the trends exhibited throughout the year by gonadosomatic indices, gonadal maturity stages, and pattern of oocyte development. Tliis time was considered to coiTespond to the birth date of G. hehraicum and could thus be used, in combi- nation with the number of opaque zones on the otolith and the time when the annulus becomes delineated on the oto- lith, to determine the age of individual fish on their date of capture. Because the sex offish <150 mm could not be deter- mined, the lengths-at-age of these small fish were randomly allocated in equal numbers to the data sets for female and male fish used for constructing the gi-owth curves. Assumptions are made concerning the distribution of er- rors when fitting von Bertalanffy growth cui-ves to length- at-age data. Kimura (1980) discussed the implications of the following three assumptions, namely that 1) the indi- vidual lengths-at-age have a constant variance, 2) the mean lengths-at-age have a constant variance and 3) the vari- ance of the lengths-at-age is dependent on age. The assump- tion most frequently adopted in growth studies is that the individual lengths-at-age have a constant variance. As dis- cussed by Kimura (1980), different assumptions regarding the error variance require modifications to the objective function to ensure that the parameters are estimated ac- curately and that any comparisons between gi-owth cui-ves, that are based on the likelihood ratio, are appropriate. A von Bertalanffy growth equation was fitted to the lengths-at-age of female and male fish with the traditional assumptions that the lengths-at-age are normally distrib- uted around the values predicted from the growth equa- tion and that the variance of this distribution is constant for each sex over all ages. However, visual examination of the residuals for each curve suggested that it was not ap- propriate to make the latter assumption. Further study showed that the variance of the residuals is approximately proportional to the age of the fish, as above in assump- tion 3 of Kimura ( 1980). Thus, the von Bertalanffy growth equation was fitted to the length-at-age data for each sex by using the assumption that the residuals were normally distributed, where the variance of this distribution was proportional to age but dependent on sex. That is, L, =L.{l-exp[-/?(^, -^„)]} and Li = L, 4- f,, where, for each sex, L^ = the observed length at age; L = the estimated length-at-age; t^ = the age; and fj= the error associated with the 7th fish. For the growth curves for females and males, L .^ is the mean asymptotic length predicted by the equation, k is the growth coefficient, and t^^ is the hypothetical age at which fish would have zero length if growth had followed that predicted by the equation. The errors are assumed to be normally distributed, such that f~NiO, cj^), where c,. is the constant of proportionality between the variance of the residuals and age for fish of sex .s. The growth equa- tions were fitted to the observed length-at-age data for both sexes by maximizing the log-likelihood of the data. The log-likelihood for the combination of male and female fish. A, may be written as A: \m-^ ^f' where A^ = the log-likelihood associated with females or males and may be calculated as A, = -^log(2ff)-- > log(cJ,) > \ — '- ^— k and n^ = the number offish of that sex in the length-at- age data. The maximum likelihood estimate off,, for each sex is given by The SOLVER routine in Microsoft EXCEL (Microsoft Corp., 2000) was used to estimate the parameters that Hesp et a\ . Age and size composition, growth rate, reproductive biology, and liabitats of Glaucosoma hebraiaim 217 would maximize the log-likelihood function, by fitting the equations to the combined set of length-at-age data for both females and males. The growth curves for the fish of each sex were com- pared following the Hkelihood ratio method described by Ivimura (1980) and Cerrato (1990). The test of the like- lihood ratio, A. that was applied was to reject the null hypothesis il (that there was no difference between the cui-v'es) at the « level of significance when fM+ff -F", ih,*U where A = . 2 \-"f '2 ' Fil ) and where n - /; ,, + nf, f = n - 3; and q = the number of linear constraints of the form 0J^^ = Op, where 6 is one of the parameters M and F = males and females, respectively. This test was developed by Gallant (1975), as described by Cerrato (1990). The gi'owth curves were fitted under all possible para- meter sets, and the best of both the 4- and 5-parameter, models, i.e. those that maximized the log-likelihood, were selected. The resulting 3-, 4- and 5-parameter models were compared with the 6-parameter model by using the above test to determine which of these three models, w. was of minimum complexity and not significantly different from the 6-parameter model i2. The model selected on the basis of these tests was the simplest model that, in the statisti- cal sense, provided the best description of the data. Reproductive biology The gonads of each fish that could be sexed macroscopi- cally were removed and weighed to the nearest 0.01 g. Each gonad was allocated to a maturity stage, based on the scheme of Laevastu (1965), but which, in the case of females, also took into account the histological character- istics of the ovaries (see "Results" section). The percentage contributions made by the different go- nadal stages in sequential 50-mm length intervals were calculated for both female and male G. hebraicum. The lengths at which 50'7f of female and male G. hebraicum reach sexual maturity (L50' were determined by fitting the logistic curve to the percentage of female and male fish which, during the spawning period, possessed gonads at stages III to VIII (see "Results" section for rationale for us- ing these six stages for this purpose). The logistic curve was fitted by employing a nonlinear technique (Saila et al., 1988) and by using a routine statistical method pro- vided in SPSS (SPSS Inc., 1988). The logistic equation is Pf^ = 1/(1 -H e'"*'''-'), where P^ is the proportion offish with mature gonads at the mid-point of the length class, L, and a and b are constants. The L-,, for each sex was derived from the equation Lr,,, = y . The ages at which 50% of fe- males and males reached maturity, i.e. the Ar,,,, were esti- mated, as follows, from the inverse von Bertalanffy growth equations for the two sexes (see Stergiou, 1999): ■^0 ~ 'i ■o-(i|log h. Gonadosomatic indices (GSIs) of females and males >Lr^Q at first maturity were determined from the equation W1/W2 X 100, where W\ = wet weight of the gonad; and W2 = wet weight of the whole fish. Mortality PreliminaiT analysis of catch curves demonstrated that the mortality estimates derived for commercially and rec- reationally caught fish that were greater than the MLL of 500 mm and eight years old and thus fully recruited (see "Results" section) were similar. Thus, the data from the commercial and recreational samples were pooled for esti- mating mortality. An estimate of the instantaneous coef- ficient of natural mortality, M, was determined from the von Bertalanffy gi'owth coefficient, k. with the regression equation developed by Ralston (1987), i.e. M = 0.0189 -i- 2.06/;. The instantaneous coefficient of total mortality, Z, was determined by maximizing the likelihood, when fitting the estimated age composition resulting from that mortal- ity to the observed age composition data for those dhufish that were gi'eater than the MLL of 500 mm and eight years old. In order to assess whether the observed age composi- tion data reflected decreasing levels of total mortality in earlier years, the catch cui-ve analysis was repeated with different initial ages, ranging from 10 to 30 years. Values of Z were also estimated by using the observed maximum age '^jiov* fo'' *'^^ sampled dhufish, employing both the regres- sion equation reported for fish by Hoenig ( 1983), i.e. log(Z)= 1.46- 1.01 log(C„,,,,^), and the equation for the expected value of the maximum age in a sample of size n, i.e. £(C) = 1 V- 1 Ir'^' where /, = the age at which fish become fully recruited to the fishery (Johnson and Kotz, 1970, p. 216, as reported by Hoenig, 1983). Results Habitats of Glaucosoma hebraicum Glaucosoma hebraicum <150 mm TL and <14 months old were caught regularly by trawlers offshore in water 218 Fishery Bulletin 100(2) depths of 27 to 33 m. The depth sounder indicated that these small G. hebratcum were most consistently caught over hard substrate that lay adjacent to reefs — a con- clusion later confirmed by video footage. Although a considerable amount of effort and a variety of techniques were employed in attempts to catch fish with lengths of 150-300 mm, only a small number offish of this size were collected. However, a few G. hebraiciim of this length class were caught by an experienced spearfisher while diving over low-lying reefs with rock ledges <30 cm high. Large numbers of dhufish >300 mm in length were obtained from rod and hand- line fishermen who were fishing in waters that were shown by video camera and com- mercial echo sounders to be located over limestone and coral reef formations and, in particular, where the "drop-offs"( reef edges) were two or more metres in height. Comparisons between number of opaque zones visible in whole and sectioned otoliths Sectioned otoliths Numbef ol otolil Number of opaque zones §5 hs 10 3 7 10 10 10 10 10 10 10 12 10 10 10 10 10 9 10 6 4 2 2 2 2 111 2 3 4 5 6 7 8 9 10 11 12 13141516 17 18 1920 2122 23 24 26 28 31 20 30 ao do 50 60 40 40 50 10 - 1 - • • • • • • • • • 10 20 40 30 20 10 40 50 2 ~ • • • 20 • • 10 • • 50 10 33 • 25 40 3 ~ • • 20 • • 30 30 • 33 • • - 4 " • • • • ?S 40 -5 ' 33 • • 20 50 6 ~ • • • 50 50 50 - 7 ~ • • • 100 -8 50 • 100 -9 " • 50 • 50 10 - • • u - 12 — 100 • The number of opaque zones observed in each sectioned otolith, in which up to six such zones could be seen, was the same as those visible on the same otolith prior to sectioning (Fig. 1). However, this fre- quently did not apply when a greater number of opaque zones were present. Fur- thermore, where such discrepancies occurred, the differ- ences between the number of opaque zones detected prior to and after sectioning rose as the number of opaque zones increased. In all cases where there were discrepan- cies, the number of opaque zones detected after section- ing was greater than prior to sectioning. Underestimates of the number of growth zones with whole otoliths, based on comparisons with those detected in sectioned otoliths, rose from one in whole otoliths with seven to nine opaque zones to between one and seven in those with 10-21 opaque zones (Fig. 1). In otoliths with a large number of opaque zones, the differences sometimes exceeded eight and for one such otolith was as high as twelve. These com- parisons demonstrated that, for validation that opaque zones are formed annually and that these zones can thus be used for aging G. hehraicum. experiments should be conducted on sectioned otoliths. Validation that opaque zones are formed annually The mean monthly marginal mcrements on sectioned oto- liths with 2 to 16 or more opaque zones rose from a low level in January to a maximum in September, before declining precipitously to a minimum in October and then rising slightly in December (Fig. 2). They thus reached high levels in early spring, before declining markedly in mid-spring, as the outermost opaque zone became delin- Figure 1 Comparisons between the number of opaque zones obsei-ved on the otoliths oi Glaucosoma hebraicum prior to and after the sectioning of those otoliths. The numbers above enclosed circles represent the percentage number of underestimates of the number of opaque zones when using whole rather than sectioned otoliths. eated through the formation of a new translucent zone, and then increased progressively in the ensuing inonths as the translucent region increased in width. Although fish possessing otoliths with one opaque zone were not caught in all months, the trends exhibited by the mean monthly marginal increments for those months when such fish were caught were consistent with those exhibited by otoliths with a larger number of opaque zones. Because the mean monthly marginal increment rose and declined only once during the year, irrespective of the number of opaque zones in the otolith, a single opaque zone IS laid down in the otoliths of G. hebraicum each year. The number of opaque zones in sectioned otoliths can thus be used, in conjunction with the birth date of G. hebrai- cum and the month when the opaque zone(s) become de- lineated, to age this species. Growth of Glaucosoma hebraicum Because the trends exhibited by the GSIs and stages in gonadal maturation and oocyte development demonstrated that the spawning of G. hebraicum peaked from late Janu- ary through early February, this species was assigned a birth date of 1 Februai'v. Age O-i- G. hebraicum were first caught by trawling over hard substrate in April and May, when their lengths ranged from 57 to 81 mm (Fig. 3). How- ever, substantial numbers of the O-i- age class were not Hesp et al : Age and size composition, growth rate, reproductive biology, and habitats of Glaucosoma hebraicum 219 0.4 r 02 25 55 08 0.4 2-5 opaque zones g 7 28 16 '' 20 14 6-8 opaque zones 04 9-11 opaque zones 0.8 0.4 12-15 opaque zones g .j. >- > 16 opaque zones 0.4 JFMAMJJASOND Month Figure 2 Mean monthly marginal increments ±1SE for sag- ittal otoliths of Glaucosoma hebraicum. Sample size is given for each month. In this Figure and Figure 5. the closed rectangles on the horizontal axis refer to summer and winter months and the open rectangles to autumn and spring months. 10 - 5 15 10 15 10 5 ■5 10 5 D 15 10 5 15 10 5 '- ' 5 - ■0 - ■0 15 10 ■0 5 ■5 10 [H] r"~^ April n=3 r\/1ay n=8 June n=13 July n=3 August n=11 October n=51 November n=63 December Q n=60 January n=27 February n=27 Marcti n=45 Total length (mm) Figure 3 Length-fi-equency distributions for Glaucosoma hebraicum caught by trawling along the lower west coast of Australia by using data for corresponding months in the period May 1996 to June 1999. "denotes mean lengths of 0-t- and early 1+ fish. His- tograms in gray refer to 0+ fish, and those in black and white refer to 1+ and 2-1- fish, respectively. caught until October, presumably reflecting the time typi- cally required for the 0-f age class to be recruited into these areas from those in which spawning occurs. The mean length of the O-i- age class had reached 95 mm by October, in which month the first opaque zone became delineated on the otoliths, and 108 mm by January, when fish were approaching the end of their first year of life. The mean length of the corresponding cohort, now early l-i-, was 127 mm in March, after which month the number of l-i- fish caught in trawl samples declined markedly (Fig. 3). The best of the 4- and 5-parameter growth curves were selected as the models with the largest log-hkelihood at that level of model complexity. The best 4-parameter model was the cui-ve that assumed different asymptotic lengths for the sexes, and the best 5-parameter model was that which assumed that the growth coefficients were equal. Comparisons between the curves demonstrated that the model that assumed common growth coefficients for fe- males and males was not significantly different {P>0.05) from the more complex model, which assumed that all 220 Fishery Bulletin 100(2) parameters of the gi-owth curves differed between the sexes. For this cui^ve, the parameters L., k, and t„ and their 959; confidence hmits were estimated to be 929 (908 to 949) mm, 0.111 (0.107 to 0.116)/year, and -0.141 (-0.183 to -0.100) years, respectively, for females, and to be 1025 (1003 to 1048) mm, 0.111 (0.107 to 0.116)/year. and -0.052 (-0.088 to -0.0161 years, respectively, for males. The growth curves for females and males were significant- ly different (P<0.001), with the asymptotic lengths hav- ing the most influence on the difference between the sex- es. The estimated constant of proportionality between the variance of the residuals and age were 363 for females and 320 for males. The von Bertalanffy growth cui-ves demonstrated that females gi'ow slightly slower than males. Thus, at ages 2 to 5. females had reached lengths of 196, 273, 342, and 404 mm, compared with 209, 294, 371, and 440 mm for males. By the time G. hebraicum had attained 10, 15, and 20 years, the females had reached ca. 628, 756, and 830 mm, respectively, and the males had reached ca. 689, 832, and 914 mm, respectively (Fig. 4). The maximum ages re- corded for females and males were 39 and 41 years, re- spectively, and the maximum total lengths of females and males were 981 mm (=ca. 15.3 kg) and 1120 mm (=ca. 23.2 kg), respectively. The ages at which female and male G. he- braicum reach the minimum legal length for capture (500 mm TLi were 7.0 and 6.0 years, respectively. Trends exhibited by reproductive variables The macroscopic characteristics of the different stages in gonadal development, and of the cytological characteris- tics of the ovaries at different stages based on an examina- tion of histological sections, are given in Table 1. Because stages I (virgin) and II (immature) in the de- velopment of both the ovaries and testes of G. hebraicum were difficult to separate macroscopically, data for these two stages were pooled in the case of both sexes. Further- more, it is also important to recognize that spawning stage (VI) ovaries are distinguished from prespawning stage (V) ovaries almost exclusively on the basis of their posses- sion of hydrated oocytes or postovulatory follicles (or both) when histological sections were employed to examine the ovary at a finer scale. However, because G. hebraicum is a multiple spawner, i.e. produces eggs in batches at in- tervals, any "prespawning" stage ovary may already have produced some hydrated oocytes, but been at an interme- diate phase in which the next batch of yolk granule oocytes had not yet become hydrated. The prevalence of females with prespawning ovaries that had already spawned on one or more occasions would be expected to increase dur- ing the spawning period. Likewise, the main difference be- tween prespawning and spawning testes, i.e. the ability of the testes to produce milt when subjected to physical pressure, may often represent different phases in the cy- clical changes undergone in the testis during the spawn- ing period. For the above reasons, the data on stage-V and stage-Vl ovaries and testes were pooled for describing the change in compositions of the gonadal maturity stages of each sex during the year. 1250 1000 750 ■0^^ 500 ^k'-' '■ Females 250 - /" o (UIUJ) 1 / c 1 1 H 1250 1000 >-^^— ^ ' ■'■'S^^^^^' 750 m 500 ap.' Males jJF n=799 250 f J 10 20 30 40 50 Age (years) Figure 4 Von Bertalanffy growth curves fitted to length-at-age data for females and males of Glaucosoma hebraicum caught on the lower west coast of Australia. ;; = number offish used to construct growth curves. Between May 1996 and April 1998, the mean monthly GSIs for females of G. hebraicum that were greater than the L-ii of 301 mm at first maturity were always low in winter (June to August) and early spring (September), i.e. <1.0 but then rose sharply to reach a peak of ca. 2.8 in mid-summer (January), before declining markedly during early to mid-autumn ( March and April i ( Fig. 5 ). The trends displayed by the mean monthly GSIs for males of G. he- braicum, that were greater than the L-^of 320 mm at first maturity, paralleled those just described for females. Because the trends exhibited by the mean monthly GSIs for females and males were the same during both 12 month periods, the percentage contributions of the different go- nadal stages of the females and males of G. hebraicum, that were longer than the L^,,, were pooled for each of the cor- responding calendar months. The gonads of female G. he- braicum in July were at stages I-II, i.e. virgin or immature (Fig. 6). Fish with ovaries at stage III (developing) were first found in August, albeit only a few fish, and those at stage IV Hesp et al : Age and size composition, growth rate, reproductive biology, and habitats of Glaucosoma hebmicum 221 Table 1 Characteristics of the macroscopic stages in the development ofthe gonads of Glai/cuKDiua lichmiciini and, in the case of the ova- ries, the histological characteristics of each corresponding ovarian stage. Terminology for oocyte stages follows Khoo ( 1979). Stage Macroscopic appearance Histological characteristics l-ll (virgin and immature) III (developing) rV (maturing) V (prespawning) VI (spawning) VII (spent) VIII (.recovering spent) Gonads very small. Ovaries transparent and oocytes not visible. Testes strandlike and gray-white. Gonads slightly larger than at stage I or II. Ovaries pinkish, blood capillaries visible in ovary walls. Testes white and more lobular. Gonads markedly larger. Ovaries reddish-orange, capillaries more conspicuous and some yolk granule oocytes visible through ovary wall. Milt is not extruded when pressure is applied to testes. Ovaries orange and occupy most of space in body cavity. Extensive capillaries in ovary walls. Milt appears when testes placed under firm pressure. Same as for stage V, but with hydrated oocytes visible through ovarian wall and only slight pressure required to release milt from testes. Gonads smaller than at stages V or VI. Ovaries flaccid. Some yolk granule oocytes still visible through ovary wall. Testes pinkish-red. Gonads greatly reduced in size and dark red. Testes strandlike. Ovigerous lamellae highly organized. Oogonia and chromatin nucleolar oocytes and, in more advanced ovaries, early perinucleolar oocytes are present. These oocyte stages are present in all subsequent ovarian stages. Early and late perinucleolar oocytes and yolk vesicle oocytes present. Yolk vesicle and yolk granule oocytes abundant. Yolk granule oocytes abundant and in tight groups. Hydrated oocytes or postovulatory follicles (or both) present. Remnant yolk granule oocytes present, typically under- going atresia. Lamellae not organized as in early stages of develop- ment and contain extensive scar tissue. Any remain- ing yolk gi'anule oocytes are atretic. (maturing) and stages V and VI (prespawning and spawn- ing) were first recorded in September and October, respec- tively. Stage-V and stage-VI ovaries collectively became the most prevalent group in females in November and formed the most dominant group by far in December to March. The samples in February and March contained a few fe- male fish with stage I-II ovaries, but none with ovaries at either stage III or FV (Fig. 6). These trends provide over- whelming circumstantial evidence that any female whose ovaries have developed to at least stage III by November will progress through to maturity during the following months of the spawning period. Thus, the L^,, for females at first maturity was calculated by using the percentage of ovaries with stages III and FV, as well as those with stages V-VIII. Although females with stage-VII (spent) and stage- VIII (recovering spent) ovaries were found between Janu- ary and May, the majority of ovaries were at stages I-II in the latter month and all were at stages I-II in June. The trends exhibited by the pattern of gonadal development in males were essentially the same as those just described for females and thus the L-q of males was likewise calculated with the percentage of testes at stages III- VIII (Fig. 7). The following account of the trends exhibited by the oo- cyte composition of ovaries is based on an histological ex- amination of the ovaries of large fish well above the L^g at first maturity. The oocytes in ovaries in July and Au- gust were almost exclusively at the chromatin nucleolar stage. Ovaries with yolk vesicles first appeared in Septem- ber, and those with yolk granules were first found in Oc- tober. Yolk granule oocytes became increasingly prevalent in ovaries in November and dominated the complement of their larger oocytes between December and March. Some of the residual yolk granule oocytes in April and all of those in May were undergoing atresia. No yolk vesicle or yolk granule oocj^tes were found in June. Hydrated oocytes were first found in ovaries in November and were present in many ovaries between December and March and in a few ovaries in April, but were found neither in May nor in the immediately ensuing months. Small numbers of post- ovulatory follicles were present in about a third of the ovaries of large females caught between December and March. The oocyte diameters of individual large G. hebra- iciim caught in each month of the spawning period pro- duced a series of modes (data not shown). 222 Fishery Bulletin 100(2) Length and age at maturity The sex oiGlaucosoina hebraicurn was not able to be deter- mined by macroscopic examination of the gonads until it had reached ca. 150 mm in length. During the main part of the spawning period, i.e. December to March, the gonads of all female and male G. hebraiciun <250 mm were at the earliest stages of development, i.e. I-II (Fig. 7). Gonads at stages III-VIII were first found in the 250-299 mm length class of females and in the 300-349 mm length class of males. The presence of such gonads demonstrated that the fish were maturing or that spawning was occurring or had been completed (see earlier). The prevalence of ova- ries at stages III-VIII increased to ca. SO'/r in the 300-349 mm length class and to 1009i^ in all females >450 mm. The gonads of all males >450 mm were at stages III-VIII (Fig. 7). The L^^'s for the lengths of female and male G. hebraicum at first maturity, derived from the logistic cui-ve 4 r Females o ^ 03 C5 02 0.1 Males fitted to the percentage contributions of fish with gonads at stages III-VIII in sequential 50-mm length classes, were 301 and 320 mm, respectively (Fig. 7). Individual G. hebraicum could first be sexed macroscopi- cally during their second year of life. Although relatively few two-, three- and four-year-old fish were caught, the trends exhibited by the proportion of gonads at stages III- VIII in both sexes during the spawning period were con- sistent. One female and no males at two years of age pos- sessed gonads at stage III or gi-eater (Fig. 8). However, 50'X of three-year-old female and male fish, and all five-year-old females and all six-year-old males possessed such gonads and were thus regarded as mature. The Ar^^s for the age at first maturity of females and males were 3.4 and 3.3 years, respectively. Mortality Using the regi'ession equation developed by Ralston (1987), in combination with the esti- mated value for the von Bertalanffy growth coefficient, k = 0. Ill/year, we estimated the instantaneous coefficient of natural mortality, M, to be 0.25/year The catch curve analysis of the combined age composition data, for the 620 dhufish older than 9 years and longer than the MLL (Fig. 9), produced an estimate of the instantaneous coefficient of total mortality, Z, of 0.21/year (95'^f confidence intei-val: 0.19 to 0.23/year). The estimate of Z remained at about this level as the initial age was increased to 15 years and then declined to ca. 0.15/year at 24 to 27 years (Fig. 9). It subsequently became less precise as the initial age increased. An esti- mate of Z of 0. 10/year was obtained when the obsei-ved maximum age of 41 years was substi- tuted into Hoenig's ( 1983) regression equation. However, when the sample size of 620 fish was taken into account, with the expression for the expected maximum age (Hoenig 1983, Appen- dix A), Z was estimated to be 0.22/year. M J J A S O N D J I I 1996 I FMAMJJASONDJFMA 1997 I 1998 ' Month Figure 5 Mean monthly gonadosomatic indices ±1SE for females and males of Glaucosoina hebraicum caught in offshore waters between May 1996 and April 1998. Data in this Figure and Figure 6 are restricted to females and males 2Lg„ at first maturity. Numbers offish used to calcu- late each mean GSI arc shown. Discussion Ontogenetic changes in habitat of Glaucosoma hebraicum Extensive sampling for G. hebraicum during the present study, allied with data obtained with an echo sounder and video footage, dem- onstrate that this species changes habitat as it increases in size. Thus, G. hebraicum <150 mm was found to live in areas near reefs where the substrate is firm and sponges often occur (Bergquist and Skinner, 1982). The reduction in the numbers of l-t- dhufish caught by trawl- ing in this type of habitat in late autumn, when their lengths were about 130 mm, probably reflects a movement by the members of this Hesp et al.: Age and size composition, growth rate, reproductive biology, and habitats of Glaucosoma hebmicum 223 cohort, as they increase in size, from a hab- itat that could be trawled to one where reefs occur and where it was not possible to trawl. This conclusion is supported by the fact that the few dhufish of 150-300 mm that were caught were collected from low- lying reefs, i.e. reefs that contained rock ledges up to 30 cm in height. In contrast, G. hebraiciim >300 mm typically occupy areas where there are substantial lime- stone and coral reef formations and their large size would make them less suscepti- ble to predation in a habitat where large predatory species, such as the Samson fish iScriola hippos) and the pink snapper {Pagrus auratus ) are found (Hesp, personal obs.). Aging The trends exhibited by the marginal increments on sectioned otoliths of G. heb?-aiciim show that an opaque zone is formed annually in the otoliths of this spe- cies. However, comparisons between the number of opaque zones on individual oto- liths prior to and after sectioning demon- strate that, after this species has reached six years in age, one or more of these zones often become visible only after the otolith has been sectioned. This demon- strates that earlier estimates of the age of older G. hebraiciim, which were based on the number of opaque zones visible in whole otoliths (Sudemeyer et al.M, were almost certainly often too low. An inability to detect all of the opaque zones in the whole otoliths of older fish is largely attributable to the fact that, as the otolith increases in width, it becomes in- creasingly difficult to distinguish between the zones at the periphery of the otolith. This problem parallels the situation re- corded for several other medium-size to large teleosts, such as Pacific hake (Mer- luccius productus) (Beamish. 1979), starry flounder iPlaty- ichthys stellatiis) (Campana, 1984) and blue-spotted flat- head (Platycephalus speculator) (Hyndes et al., 1992). Our results demonstrate that, although most G. hebraiciim are less than 25 years old, some females and males live for longer than 30 years and very occasionally for up to about 40 years. Other species that are typically caught by commercial and recreational rod and hand-line fishermen when fish- ing for dhufish include pink snapper iPagrus auratus), Sampson fish (Seriola hipposK silver trevally iPseudo- caranxdentexK breaksea cod iEpincphelides armatus). and occasionally also King George whiting (Sillaginndes punc- tata). The maximum total lengths and weights recorded for these five species are 1300 mm and 19.5 kg for pink ,oo.r-, Females Males "=32 [!□ July n=45 100 r |_ m n=83 .; [—1 August otl- n=120 100 r ^ J September Q 1 r -'1 , . n=93 ^DDa™ n=105 100 r ^ LLJc3=_ October n=85 :=□□= n=83 100 r November & □ nizjQ n=72 ;_nnn_ n=100 S 100 r 3 ~r December cr m ° =. nDl- n=25 □ n n=20 en 100 r B -, January c y J„„ n=99 = Da„ n=90 ^ 100 j- February n=78 Dn„ n=61 100 r = Li _o March n=63 := an^ n=65 100 r April D aunU n=83 .□ —Llczi n=90 100 |- J □ _□ May L U C=Z1 _ IZ=1 □ n=105 100 f— June n=46 1 ■-; n=55 '-' L Lil 1=1 HI 11 IV V-VI VII VIII HI III IV V-VI VII VIII Gonadal stage Figure 6 Percentage frequency histogi' ams for gonadal maturity stages of females and males of Glaucosoma hebraiciim. Data have been pooled for corresponding | months in the period May 1996 to April 1998. Sample sizes (n ) are given for each month. snapper, 1753 mm and 53.6 kg for Samson fish, 938 mm and 10.0 kg for silver trevally, 550 mm and 2.9 kg for breaksea cod, and 690 mm and 4.8 kg for King George whiting, compared with 1219 mm and 25.8 kg for dhufish (Hutchins and Thompson 1995). Although reliable data have been obtained for the age and growth of a number of commercial and recreational fish species that live in nearshore coastal or estuarine waters in southwestern Australia (e.g. Chubb et al., 1981; Hyndes et al., 1992, 1996; Hyndes and Potter, 1996, 1997; Lauren- son et al, 1994; Fairclough et al., 2000; Sarre and Potter, 2000), comparable data for those species that are found in and around reefs in deeper waters in southwestern Aus- tralia are restricted to those recorded for G. hebraicum in this paper and for the King George whiting Sillagitiodes 224 Fishen/ Bulletin 100(2) 100 75 50 25 Females 4 3 5 12 16 15 15 3236242228 33 ?1 ^ 3 5 CT S Males I 100 Q. 75 50 W 3 1 9 10 18 8 17 25 34 2522 24 22 2610 12 2 25 |y Ijir lllllllllllll 200 400 600 800 Total length (mm) 1000 1200 Figure 7 Percentage frequency of occurrence of gonads at stages I-II (D) and stages III-VIII (Dt in each sequential 50-mm class of female and male Glaucosoma hehrai- cum caught between December and March. The logistic curve has been fitted to the data for fish with gonads at stages III-VIII. The sample size is given for each length class. punctata by Hyndes et al. (1998). Using data from the low- er west coast of Australia, Hyndes et al. (1998) estimated the von Bertalanffy growth parameters, L, k, and f^, for King George whiting to be 538 mm TL, 0.47/year, and 0.13 years, respectively, for females, and 500 mm TL, 0.53/year and 0.16 years, respectively, for males. Although there are no published studies on the growth of pink snapper and sil- ver trevally in Western Australia, the growth of these two species has been investigated in New Zealand by Francis et al. ( 1992) and by James ( 1984) respectively. The param- eters L , k, and t,^ were estimated to be 720 mm FL (fork length), 0.106/year and -0.75 years, respectively, for pink snapper, and i-anged from 436 to 448 mm FL, from 0.27 to 0.43/year and from -1.6 to -0.6 years, respectively, for silver trevally Although the growth coefficient, k. for dhufish, i.e. 0. Ill/year, was similar to that for pink snapper, it was ap- preciably less that that for both Iving George whiting and Females 4 6 4 1010 7 9 12 7 58 75 50 25 =1 ^ CT (D m Males c o 36 10 96997 48 Q. 75 50 - 25 1 234 56789 ^10 Age (years) Figure 8 Percentage frequency of occurrence of gonads at stages I-II (D) and stages III-VIII (D) in each sequential age class of female and male Glaucosoma hebraicum caught between December and March. The sample size is given for each age class. silver trevally. Dhufish had an asymptotic length ca. 35% greater than that of pink snapper and approximately twice those of King George whiting and silver trevally. The lengths of females at maturity have been reported as ca. 350 mm FL for silver trevally (James, 1984) and 413 mm TL for King George whiting (Hyndes et al, 1998) and have been estimated as 237 mm FL for pink snapper (calculated from Crossland, 1977). Although maturity is first achieved by the females of snapper and dhufish at 30% of their respective asymptotic lengths, it is at- tained by the females of King George whiting and silver trevally at 75-80'7f of their asymptotic lengths. This find- ing implies that the last two species have a higher repro- ductive load sensu Gushing (1981). Thus, although these four species reach maturity and begin to occupy promi- Hesp et al : Age and size composition, growth rate, reproductive biology, and habitats of Claucosoma hebra/cum 225 nent reefs at similar lengths, the growth of silver trevally and King George whiting slows after mat- uration, whereas that of snapper and dhufish continues appreciably after they have reached maturity. Spawning location, period, and mode 125 r 100 •= 75 50 25 Glaucosoma hebraicum with gonads at stages VI (spawning) and VII (spent) were caught in waters rang- ing from 10 to 150 m in depth and at distances of 3 to 50 km from the shore and between latitudes 28°55' and 32°45'S. Thus, the spawning of dhufish is not apparently restricted to any particular water depth or region along the coast. However, because G. hebraicum greater than 300-320 mm in length (their size at first maturity) were almost invari- ably caught only around limestone or coral reef formations, this species apparently spawns in the vicinity of reefs. Because hydrated eggs and postovulatory follicles were found in at least some of the ovaries of large females in each month between November and April, it is evident that G. hebraicum spawn between the end of spring and middle of autumn. Although some fish commenced spawn- ing in November, the mean GSI of female fish in that month was still well below its maximum. This indication that only a small amount of spawning occurs in Novem- ber is consistent with the fact that many of the ovaries of large fish were still at stages III and IV. Although most of the ovaries of large females caught in May 1997 con- tained some vitellogenic oocytes, these oocytes were usu- ally undergoing atresia and the ovaries of other large fish in that month were either spent or resting. Furthermore, none of the ovaries of large G. hebraicum caught in May contained hydrated oocytes. This finding provides strong evidence that the spawning period does not extend into May. There is also strong evidence that spawning peaks in January and February. For example, by January, the ova- ries and testes of most large fish were at stages V or VI, i.e. prespawning or spawning, and, for the first time, some were spent or even recovering spent (stages VII and VIII). The maintenance of the GSIs of females at their maxima in both January and February is attributable to the fact that, because G. hebraicum is a multiple spawner, new batches of hydrated oocytes were continually being devel- oped in the ovary during these two months. However, the GSIs of females and males both declined precipitously in March, which demonstrates that, in the case of ovaries, the release of eggs during spawning was not being com- pensated for by a comparable production of new batches of mature eggs. As spawning activity peaked in January and February, it was appropriate to use 1 February as the birth date of G. hebraicum when assigning an age to each fish. 0.4 03 2 0.2 :i - 1 tB J n-m-;^^ n=^ 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 Age (years) Figure 9 Age composition of Glaucosoiiia hebraicum. that were caught at lengths > the mini- mum legal length of .500 mm, and maximum likelihood estimates (±95% confidence limits) for the instantaneous coefficients of total mortality/year, Z, determined from subsets of these data selected by using different initial ages. The fact that, during the spawning season, mature ova- ries of G. hebraicum often contained yolk vesicle, yolk gran- ule, and hydrated oocytes and, in some cases, also post- ovulatory follicles, implies that this species is a multiple spawner sensu deVlaming (1983), i.e. individual females release eggs on more than one occasion in a spawning sea- son. The oocytes of individual female G. hebraicum during the spawning period ranged widely in size and, in many cases, their diameters formed relatively discrete modes in oocyte diameter-frequency distributions. The ovaries of G. hebraicum thus contain hatches of oocytes that are pre- sumably released at different times. Multiple-batch spawn- ing over a protracted period enables a greater total number of eggs to be produced and released during a spawning pe- riod and results in eggs becoming discharged at different times (McEvoy and McEvoy, 1992), which would increase the overall chance of recruitment success. Implications of the biology of Glaucosoma hebraicum for fisheries management The age composition data for dhufish older than 9 years and larger than the MLL reflect an average level of the instan- taneous coefficient of total mortality. Z. of 0.21/vear This value is consistent with the estimate obtained from the observed maximum age, taking into account the sample size of 620 fish (Hoenig. 1983). The much lower value obtained for Z, with Hoenigs ( 1983) regression equation for fish, i.e. 0.10/year, does not take into account sample size. The estimate of the instantaneous coefficient of natural mortality, M, of 0.25/year, that was calculated with Ralston's (1987) equation, exceeds the average value of 0.2L/year for the instantaneous coefficient of total mortality, which was estimated from the catch curve. Examination of the residu- als from the regression line fitted by Ralston (Fig. 8.1 in 226 Fishery Bulletin 100(2) Ralston, 1987) suggests that the precision of this estimate of the instantaneous coefficient of natural mortality is like- ly to be relatively low. It was therefore concluded that the value for M derived from Ralston's equation represented an overestimate. A more detailed examination of the catch cun'e data suggested, but was unable to demonstrate con- clusively, that the level of total mortality experienced by the older fish when they were young was less than that which is now being experienced by the population. Indeed, if the decline in the estimated value for Z, displayed in Fig. 9, was extrapolated to an age of 40 years, the total mortal- ity exhibited by the oldest age classes (when, as young fish, they first became fully vulnerable to the fishery) would be ca. 0.1/year. Such a value, which might be only slightly greater than the natural mortality, matches the estimate of Z calculated from the obsei-ved maximum age with Hoe- nig's ( 1983) regi-ession equation. However, such agi'eement may be fortuitous because the latter estimate should rep- resent the total mortality experienced by the fish within the sample, i.e. from age 9, rather than just the mortality of the older fish. Nevertheless, if the level of natural mor- tality, M, is ca. 0. l/year and the average level of instanta- neous coefficient of total mortality, Z, from age 9 years is ca. 0.21/year, the current level of fishing mortality, F. would exceed 0.1 l/year. In the nearshore waters along the lower west coast of Australia where this species is most heavily fished, the abundance of G. hebraicum has declined to a level that is of concern to fishermen. Numerous anecdotal reports indicate that commercial and dedicated recreational fish- ermen, such as those who provided the samples for this study, now tend to move further offshore in order to obtain catches of G. hebraicum comparable with those they used to obtain in waters closer to the coast. However, many rec- reational fishermen still continue to fish for G. hebraicum (and other species) in the traditional areas where dhufish were fished in the past. The expansion of the fishery for dhufish to include waters farther offshore, allied with the increasing use of global positioning systems (GPSs) to im- prove fishing efficiency, is increasing the level of exploita- tion of the stock as a whole. The fact that there are indications that the fishing-in- duced mortality of dhufish may now exceed natural mor- tality and that ongoing expansion in the extent to which fishermen are moving offshore (and also, in the case of rec- reational fishermen, in a northwards and southwards di- rection from the main metropolitan region of Perth) will further increase fishing pressure, is of concern to the man- agers responsible for the fishery for G. hebraicum. How- ever, because our sampling regime was not designed spe- cifically at determining the levels of fishing mortality to which G. hebraicum is being subjected, there is clearly a need to undertake a study in which the main aim is to achieve this objective. If such research were to confirm our preliminary findings that fishing mortality is reaching an unacceptable level, there will be an urgent need to use the biological data produced during the current study to refine the management plans designed to consei-ve this species. Female and male G. hebraicum first roach sexual matu- rity at the end of their third year of life when they are just over 300 mm in length and they reach 500 mm, the MLL for capture, when they are about 7 and 6 years old, respec- tively. Thus, on average, the female and male dhufish that live until they reach the MLL will have had the opportu- nity to have spawned for four and three years, respectively, before they can legally be retained following capture. On- going research at the state fisheries laboratory in Western Australia has indicated that ca. 50'^'i of fish caught in wa- ters of 20-30 m depth die on being released back into the water and that this percentage increases to ca. 95'r for fish brought to the surface from depths greater than 40 m (Moran-). Thus, the use of a MLL is likely to be of only lim- ited value for consei-ving this species as fishing effort con- tinues to increase. It is therefore important to introduce measures that will conserve G. hebraicum by maintaining the catches of this species at a level consistent with the re- quirements for ecological sustainability. Examples of such management controls might include closing areas to com- mercial and recreational fishing ( particularly those around reefs that are especially heavily fished) introducing quo- tas for commercial fish catches, making adjustments to the number of commercial licenses, further restricting the bag limit for recreational fishermen, and limiting the number of recreational fishermen that can fish in a given area. Furthermore, because the Fremantle Maritime Centre has successfully cultured G. hebraicum (Cleary et al.-^), there is also now the potential for restocking this species in areas in which it has become severely depleted. Acknowledgments Gratitude is expressed to those recreational and commer- cial fishermen and fish processors and many friends and colleagues for their help in the collection of fish and to colleagues at the Centre for Fish and Fisheries Research at Murdoch University for their help and advice. Grati- tude is also expressed to Gavin Sarre for providing inde- pendent counts of the number of translucent zones on otoliths and to two anonymous referees for their construc- tive comments on the original text. Funding was provided by the Fisheries Research and Development Corporation (FRDC). Literature cited Beamish, R. J. 1979. Differences in the age of Pacific hake iMcrluccius pro- ductus) using whole and sections of otoliths, J. Fish. Res. Board Can. 36:141-151. - Moran, M. 2001. Personal commun. Western Australian Ma- rine Research Laboratory, Fisheries Western Australia. PO Box 20, North Beach 6020, Western Australia. ^ Cleary, J. J., G. I. Jenkins, and G. Partridge. 1999. Prelimi- nary manual for the hatchery production of WA dhufish tGlauco- soina hebraicum ). Interim report to FRDC ( Fisheries Research and Development Corp.), 30 June 1999. (Project 96/308), 36 p. Fremantle Maritime Centre. 1 Fleet St.. Fremantle. Western Australia. 6160. Hesp et a\ Age and size composition, growth rate, reproductive biology, and habitats of Glaucosoma hebraicum 227 Bcrgquist, P. R., and I. G. Skinner. 1982. Sponges (phylum Porifera). ChapU'r 3 lit Marino invertebrates of southern Austraha, part I (S. A. Shepherd and I. M. Thomas, eds. ), p. 38-72. D. J. Woolmnn. Onvern- ment Printer, South Australia. Campana, S. E. 1984. Comparison of age determination methods for the starry flounder. Trans. Am. Fish. Soc. 1 13:365-369. Cerrato, R. M. 1990. Interpretable statistical tests for gi-owth comparisons using parameters in the von Bertalanffy equation. Can. J. Fish. Aquat. Sci. 47:1416-26. Chubb, C. F, I. C. Potter. C. J. Grant, R. C. J. Lenanton, and J. Wallace. 1981. Age structure, growth rates and movements of .sea mullet, Miigil ccphahis L., and yellow-eye mullet, Aldri- chetta forsteri (Valenciennes), in the Swan-Avon river sys- tem. Western Australia. Aust. J. Mar. Freshwater Res. 32; 605-628. Crossland, J. 1977. Seasonal reproductive cycle of snapper Chrysophrys aiirariis (Forster) in the Hauraki Gulf N.Z. J. Mar. Fresh- water Res. 11:37-60. Gushing, D. H. 1981. Fisheries biology: a study in population dynamics, 2"'' ed. Univ. Wisconsin Press, Madison, WI, 295 p. deVHaming, V. L. 1983. Oocyte development patterns and hormonal involve- ments among teleosts. In Control processes in fish physi- ology (J. C. Rankin, T. J. Pitcher, and R. T. Duggan, eds.), p. 176-199. Croom Helm, Beckenham. Fairclough, D. F, W. F Dimmich, and I. C. Potter 2000. Length and age compositions and growth rates of the Australian herring Arripis georgiana in different regions. Mar Freshwater Res. 51:631-640. Francis, R. I. C. C, L. J. Paul, and K. P Mulligan. 1992. Ageing of adult snapper (Pagrus aurattis) from otolith annual ring counts: validation by tagging and oxytetracy- cline injection. Aust. J. Mar. Freshwater Res. 43:1069-89. Gallant, A. R. 1975. The power of the likelihood ratio test of local nonlin- ear models. J. Am. Stat. 29:73-81. Hoenig, J. M. 1983. Empirical use of longevity data to estimate mortality rates. Fish. Bull. 82:898-903. Hutchins, B. H., and M. Thompson. 1995. The marine and estuarine fishes of south-western Aus- tralia: a field guide for anglers and divers. Western Aus- tralian Museum, Perth, 103 p. Hyndes, G. A., N. R. Loneragan, and I. C. Potter 1992. Influence of .sectioning otoliths on marginal incre- ment trends and age and growth estimates for the flathead Platycephalus speculator. Fish. Bull. 90:276-284. Hyndes, G. A., M. E. Platell, I. C. Potter and R. C. J. Lenanton. 1998. Age composition, growth, reproductive biology, and recruitment of King George •whiting.Sillagmodes punctata . in coastal waters of southwestern Australia. Fish. Bull. 96: 258-270. Hyndes, G. A., and I. C. Potter 1996. Comparisons between the age structures, growth and reproductive biology of two co-occurring sillaginids, Sillagu rohusta and S. bassenNis, in temperate coastal waters of Australia. J. Fish Biol. 49:14-32. 1997. Age, growth and reproduction ofSillago schomburgkii in south-western Australian, nearshore waters and com- parisons of life history styles of a suite oi Sillago species. Environ. Biol. Fish. 49:435-447. Hyndes, G. A., I. C. Potter and S. A. Hesp. 1996. Relationships between the movements, growth, age structures and reproductive biology of the teleosts Sillago burrus and S. vittata in temperate marine waters. Mar Biol. 126:549-558. James, G. D. 1984. Trevally, Caranx georgianus Cuvier: age determina- tion, population biology and the fishery. Fish. Res. Bull. NZ. 25, 51p. Johnson, N. L. and S. Kotz 1970. Continuous univariate distributions-1. John Wiley, New York. NY. 300 p. Khoo. K. H. 1979. The histochemistry and endocrine control of vitello- genesis in goldfish ovaries. Can. J. Zool. 57:617-626. Kimura. D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull. 77:765-776. Laevastu, T. 1965. Manual of methods in fisheries biology. FAO, Rome, 51 p. Laurenson, L. J. B., I. C. Potter and N. G. Hall. 1994. Comparisons between generalised growth curves for two estuarine populations of the eel tailed catfish Cnido- glanis macrocephalus. Fish. Bull. 92:880-889. McEvoy, L. A., and J. McEvoy 1992. Multiple spawning in several commercial fish species and its consequences for fisheries management, cultivation and experimentation. J. Fish Biol, (suppl. B) 41:125-136. McKay, R. J. 1997. Pearl perches of the world. FAO species catalogue, vol. 17. FAO, Rome, 24 p. Microsoft Corporation. 2000. Excel. Microsoft Corporation, Seattle, WA. Optimas Corporation. 1995. Optimas 5, user's quide and technical reference, vol.1, 7'h ed. Bothell, WA, 524 p. Ralston, S. 1987. Mortality rates of snappers and groupers. //? Tropical snappers and groupers. Biology and fisheries management (J. J. Polovina and S. Ralston, eds.), p. 375-404. Westview Press, Boulder, CO. Saila, S. B., C. W Recksieck, and M. H. Prager 1988. Basic fisheries science programs. Elsevier, New York, NY, 230 p. Sarre, G. A., and I. C. Potter 2000. Variation in age compositions and growth rates of Acanthopagrus hulcheri iSparidael among estuaries: some possible contributing factors. Fish. Bull. 98:785-799. SPSS Inc. 1988. SPSS-X™ users guide. SPSS Inc., Chicago, IL, 806 p. Stergiou, K. 1999. Intraspecific variation in size- and age-at-maturity for red bandfish, Cepola macrnphthalma. Environ. Biol. Fish. 54:151-160. 228 Abstract— Longitudinal surveys of ang- lers or boat owners are widely used in recreational fishery management to estimate total catch over a fishing season. Survey designs with repeated measures of the same random sample over time are effective if the goal is to show statistically significant differ- ences among point estimates for succes- sive time intervals. However, estimators for total catch over the season that are based on longitudinal sampling will be less precise than stratified estimators based on successive independent sam- ples. Conventional stratified variance estimators would be negatively biased if applied to such data because the samples for different time strata are not independent. We formulated new general estimators for catch rate, total catch, and respective variances that sum across time strata but also account for correlation stratum samples. A case study of the Japanese recreational fishery for avu tPlecoglossus altivelis) showed that the conventional stratified variance estimate of total catch was about 10^:i of the variance estimated by our new method. Combining the catch data for each angler or boat owners throughout the season reduced the vari- ance of the total catch estimate by about 75%. For successive independent surveys based on random independent samples, catch, and variance estima- tors derived from combined data would be the same as conventional stratified estimators when sample allocation is proportional to strata size. We are the first to report annual catch estimates for ayu in a Japanese river by formu- lating modified estimators for day-per- mit anglers. Longitudinal logbook survey designs for estimating recreational fishery catch, with application to ayu iPlecoglossus altivelis) Shuichi Kitada Tokyo University of Fisheries 4-5-7 Konan, Minato Tokyo 108-8477, Japan Email address kiladaia'tokyo-u-fish.ac ip Kiyoshi Tezuka Nakagawa Branch Tochigi Prefectural Fishenes Expenmental Station Ogawa, Nasu Tochigi 324-0501, Japan Manuscript accepted 14 September 2001 Fish. Bull. 100:228-243 (2002). Angler surveys are widely used in fish- ery management to estimate recreation- al catch, and there is an extensive body of literature on this subject (see Guthrie et al., 1991). Pollock et al. (1994) pub- hshed a manual on angler sui-vey meth- ods and their applications in fishery management. The first purpose of our study is to make two very important points for the designers of recreational fishery surveys: 1 ) longitudinal surveys taking repeated measures on the same random sample of anglers or boats over time are better than successive indepen- dent surveys if the goal is to determine significant trends in catch and fishing effort, and 2) stratified surveys, or suc- cessive independent surveys based on random independent samples of ang- lers or boats, are better than longitu- dinal surveys if the goal is to obtain pre- cise estimates of annual total catch and fishing effort. If longitudinal fisheries data sets are used to estimate annual catches, then correlations between monthly sample observations must be taken into account when evaluating the precision of catch estimates. This problem is not addressed in the litera- ture, and the variance estimation pro- cedures for this situation are unclear. The second purpose of our study is to estimate the annual catch of ayu iPleco- glossus altivelis) in a river because no estimates have been reported in Japan. In our study, we formulated a new method for accurate variance estima- tion with longitudinal fishery data, ex- emplified by a case study of the rec- reational fishery for ayu in Nakagawa River in Tochigi Prefecture, Japan (Fig. 1). Annual catch estimates based on sums of monthly estimates were com- pared with those based on combined data for each angler or boat through- out the fishing season. We demonstrate how use of a design with repeated mea- sures facilitates determination of sig- nificant seasonal trends in catches, and show the usefulness of combined (non- stratified) data analysis. We also esti- mate the total annual catch of the ayu fishery by formulating modified esti- mators for day-permit anglers (anglers who are granted permits to take fish for one day). Materials and methods Case study of ayu We used longitudinal data collected for the Japanese ayu fishery to compare estimators of effort and catch and their associated variance estimators. Ayu is the most popular target species of rec- reational anglers in rivers in Japan. In the Nakagawa River (Fig. 1), the upstream run of wild juvenile ayu from the coast begins in late March to early April, and is completed by early July Ayu mature and spawn from September to November, and then die after spawn- ing. Cooperatives release both hatch- ery-produced and wild juveniles caught Kiiada and Tezuka: Survey designs for estimating recreational fishery catch 229 Figure 1 Location of the angler survey. Bold lines in the lower figure indicate the area of the Nakagawa River where the sui"vey was conducted. in Biwako Lake, Shiga Prefecture (Fig. 1) from early April to the end of May. Thus, recreational anglers catch both naturally recruited wild juveniles and transplanted wild juveniles from Biwako Lake and hatchery-produced ayu released by the cooperatives. Both hatchery and wild stocks consist of a single year class that recruits in the spring. The river fishing season for ayu begins on June 1st and closes at the end of October To estimate the annual ayu catch by recreational anglers in the Nakagawa River, we conducted a longitudinal log book survey in 1993. Sampling procedure There are four cooperatives that set fishing rights on the Nakagawa River in Tochigi Prefecture. Fishing permits for ayu are sold at the cooperatives and fishing tackle shops, and these permits are valid over the entire Nak- agawa River in the prefecture. The cooperatives record the total number of season- and day-permits sold, and a complete list of season-permit anglers is available. An a priori sample size of 120 anglers (an expected sampling fraction of about 0.5% of the total number of season-per- mit anglers) was allocated to the four cooperatives in pro- portion to the number of season-permits sold (Table 1). Anglers who possess a permit (season or day! can fish for ayu over the whole Nakagawa area, regardless of where the permit was purchased. Hence, we treated the samples as if they were drawn from the population by simple random sampling, even though they were drawn by strati- fied random sampling of cooperatives. We asked the cooperatives to select samples randomly, but the samples were drawn arbitrarily. The selection, however, was not a purposive sampling; therefore we treat- ed them as random samples. The sampled season-permit anglers were asked to record catch data throughout the fishing season, including each fishing date, the number of ayu caught, and the fishing site, on a printed form, which was returned after the fishing season was over. To estimate the total catch in weight, we also surveyed the body weights of ayu in recreational catches by month. The 230 Fishery Bulletin 100(2) The number of tions along the permits Nakaga sold waP Table 1 in 1993. sample size, and the number of logbooks i iver. ;; = number of anglers sampled. etu •ned by th c four fishermen's cooperative associa- Fishermen's cooperative associations Number of season permits sold Number of day permits n Number of logbooks returned Hokubu Nanbu Chuo Motegi Total 11,314 6.391 1911 2346 21,962 6520 1946 231 369 9066 70 30 10 10 120 64 21 10 9 104 primary sampling unit in a population of season-permit anglers or boat owners (i.e. party-boat owner.s or per.son- al-boat owners) was an angler or a boat owner, and the secondary sampling unit was a fishing day. We selected anglers or boat owners by simple random sampling with- out replacement from a list of anglers or boat owners, and asked a sample of anglers or boat owners to record catch data on all fishing days throughout the survey. Because all the secondary sampling units were surveyed, we regard- ed this as a single-stage cluster sampling procedure (Co- chran, 1977). Estimation of total catch by month V{R,J- N N-n Mr (2) This variance estimator is obtained by dividing Equation 6.9 in Cochran ( 1977, p. 155 ) by the total number of fishing days in the /i'th month M^,. In Equation 2, M,. is unknown; hence we approximated the variance estimate by using the estimator of M/, as follows (Thompson, 1992, p. 621: ViR,^ = —± N-u NMpiUi - 1) t" ^iC,,-Mj,f, (3) Estimation for season-permit anglers or boat owners The principal notations for estimation of the total catch for season-permit anglers or party-boat owners are as follows (Cochran, 1977; Pollock et al., 1994): N = total number of sampling units (season-permit angler or party-boat or personal-boat owner; known number); n = sample size drawn from the population A'^; M). = total number of fishing days in the population in /(•th month (to be estimated); M,;, = total number of fishing days in kth month of selected /th sample; M^_ = mean number of fishing days per sampling unit in kih month (to be estimated); C^f. = number offish caught by ;th sample in /;th month; R/^ = catch rate of /?th month (to be estimated); Cj,"' = total catch by season-permit anglers or by party- boat owners or personal-boat owners in A'th month (to be estimated). The catch rate is the number offish caught per sampling unit each day. For logbook surveys, the ratio of the mean is the preferable estimator of catch rate (Jones et al.,1995; Pollock etal., 1997). The ratio and the variance for the catch rate is estimated by R, X—' " I,.l^-'- (1) where M^, = NMj. The variance of M^. is estimated by ViM/. ) = N'ViM,. ). where M,, is the mean number of fish- ing days per sampling unit in /;th month. The estimator and the variance are as follows (Cochran, 1977, p. 249, fromEq. 9A.2): V(M,) = Nnin-1)'^ ''■ '• (4) (5) The total number offish caught in A'th month is estimat- ed by CI"' = Mi,Ri, = NM,M,, T.J'- TJ- N :7V- Ic.. (6) This results in an unbiased estimator. Wlien the total number of fishing days M,. is unknown, the ratio estima- tor coincides with the unbiased estimator. The variance is evaluated by (Cochran. 1977, p. 249, from Eq. 9A.2): Kitada and Tezuka Survey designs for estimating recreational fishery catch 231 iHn-l) ^ where Q = V C,^, / n. W'jl'" = Q'"wj(. In general, season-permit and day-permit (7) anglers fish in the same location; therefore we assumed the same mean body weight for both kinds of anglers. The samples for estimating C/. and iv f, are independent; therefore the variances are estimated by using Goodman's (1960) formula: Estimation for day-permit anglers We estimated the catch of day-permit anglers separately. The notations for day- permit anglers catch estimation are as follows: D = total number ofday-permits sold through the season (known); Dj^ = total number of day-permits sold in k th month (known); dj. = total number of day-permit anglers who returned logbooks in kth month; Rl''' = catch rate of A'th month for day-permit anglers; Cjf' = total catch by day-permit anglers in ^th month (to be estimated). The estimator of total number of fish caught by day-per- mit anglers in A'th month is v(w,j = v(w;;') + v(wi" (14) where ViWl"') = W^V(C'^-") + cf V(JZ7,, ) + V(C'^")V{UJ^ ) and V(Wl'") = ZZvfV(Ci'" ) + C','"''V{W, ) + V(C['' ' )V{Ui, ). The mean body weight offish in Ath month is estimated from a sui-vey of individual body weights (u\^, of /,, fish caught on the fishing grounds). The estimator for Ath month and its variance estimator are and ViUJ^ ' = Z !li ' "' * ' '"^^ '" / (/* (4 - !>)■ Cl'"=DX''\ (8) where the catch rate for day-permit anglers, estimated by the sample mean, is The total fishing days in Ath month is estimated as the sum of the fishing days estimates of the season-permit an- glers and the day-permit anglers by K'" ^ d,: (9) The variance estimator of Cj/' is (Cochran, 1977, p. 26, from Eq. 2.20): nCl'"}= D-iViRi'"} ^ ^^'A-c^^) |-(C,, - R,'^)\ d,id, (10) Total monthly catch of season- and day-permit anglers The total catch by season- and day-permit anglers in ^th month is Q, which is estimated by C,=Cl'' + C['". (11) Mj.f. =M^+Di^ (15) and the variance is estimated by V(M.,.i.) = N'-ViMi^) be- cause Di is known. Estimation of annual catch Method 1 (based on monthly estimates) The annual catch is estimated by summing monthly catch estimates over the entire fishing season (A' months). The point estimator c=£Q=£cr+|;ci'"=c-+c"^'. (16) These two total estimates are obtained from independent samples (season- and day-permit anglers); therefore the variance is estimated by adding the variances: V(C.) = V(C,") + V{CY (12) When the same sample of anglers reports catches through- out the season, the sampling is not independent in each month, and monthly catch estimates are auto-correlated. Taking this correlation into account, the variance estima- tor is The total catches in weight in kth month are estimated by using the mean body weight of the species in ^th month ( w f^) estimated from the survey of individual body weights by W,=W-' + <'=C,i^, (13) where VV^~ and W^'^'are the total catches in weight for sea- son- and day-permit anglers, given by W"^" = C^'W^ and V(C) = V(C'") + V(C'") = XV(Cr') + 2;^c7v(Cr,Ci^') K K (17) +^v(c,'")+2^cov(c;'",c;^'). 232 Fishery Bulletin 100(2) The covariance between two total estimates of season-per- mit anglers is estimated by (see Appendix 1, from Cochran, 1977, p. 25) CovCC;' N(N-n ^C-, = i^^^lZii!^,C„-C,)(C,,-C,). (18) nin-1) ^ The fourth term of Equation 17 is equal to if a different sample of day-permit anglers is drawn in each month. The total catch in weight is estimated by K K K w = ^Wi = ^w;;' + '^Wi';" = w"-' + w"^' (i9) and the variance estimator is similar to Equation 17 but has a slightly different covariance which is Cov(w;-^',w;'') - w^m^.Cov{c;f\ci- (20) This covariance estimator was derived by the delta method (Seber, 1982, p. 7, see Appendix 2). which coincides with a covariance when Wj and w,. are constant. The mean annual catch rate is estimated for season-per- mit anglers or boats by The estimator of covariance between C"" and M"' is simi- lar to Equation 18 (see Appendix 1): C^{C"",M"")= ^'^~"' y (C ,-C)(M ,-M). (25) where C, and M, are the number of fish caught and the number of fishing days of (th season-permit angler or boat owner throughout the season, respectively, and ^" = l.lf-"^"- and M y" M,l n. The mean number of fishing days per sampling unit (season-permit angler or boat) is estimated by — M" M = N and the variance estimator is V(M) = -^ViM'"}, R C" M'" (21) where M'" = ^M^ = iV^M* Here M /, is given by Equation 4. The total effort is esti- mated by M'^'+ D. The approximate variance of ¥(/?'"") is given by the delta method (Seber, 1982, p. 7; Appendix 3), that is ViR"") M' V(C"") + c \2 M'- ViM'") M' (Cov(C'",M'-') (22) where V(C' ' ) is given by Equation 17, and the variance of the total number of fishing days is estimated by V(M"') = iV-'K^V(M;,)-^2^Cov(M^,M;..) . (23 Here ViM^.) is given by Equation 5 and the covariance of two sample means is estimated by (Cochran, 1977, p. 25) Cov(M,. M.. ) = ^ " T^M,, - M.)(M,,. - My ). (24 where M'"'and V(M'"') are already derived from Equation 21 and Equation 23, respectively. The catch rate of day-permit anglers over the season is estimated by R r'' D and the variance is V{k•") = A^V^C■-"\, D- where C'"'' and V(C"'') are given in Equation 8 and Equa- tion 10, respectively. Method 2 (based on total annual catch of each angler or boat owner) Another procedure for estimating annual catch is to use annual data rather than monthly catch for individual anglers or boats. The advantages of this proce- dure are that covariances between months do not have to be considered and estimators are much less complicated than those obtained using method 1. Equations derived for monthly estimation can be used without modification for this procedure. Modified estimators for day-permit anglers We could not conduct a sui-vey of day-permit anglers, so we substituted i?^ for R^" in Equation 8. In addition, the total number of day permits sold in /;th month ^D^) was Kilada and Tezuka: Survey designs for estimating recreational fishery catch 233 unknown. Hence, we slightly modified the procedure for estimating D,. by Dpi., where p^. is the proportion of day permits sold in ^th month to the annual total number of day permits sold (D). Day permits issued by the cooperatives are sold mainly in fishing tackle shops. We selected four tackle shops and surveyed the total number of day permits sold at the se- lected jth fishing tackle shop (D^), and the total number of day permits sold at the selected /th fishing tackle shop in /.■th month (D^i.). The proportion of day permits sold in ^th month was estimated by Pk = " n i;.^. where h is the number of fishing tackle shops selected from a total of// shops. The evaluation of the variance of P/. was similar to Equation 3: Vlftl = "'""'tI'"..-".'.''- D'hih- Some day permits, however, were sold at the fishing sites, and the above variance estimator was not appropriate for this situation. Assuming S'^'^j/), was selected by simple random sampling from D, we evaluated the variance by (Cochran, 1977, p. 52, Eq. 3.8) ^E>.-i)' The modified estimator for the total number of fish caught by day-permit anglers in /;th month is k DpA (26) Here p^ and Rf, are independent because these are esti- mated from different survey data. The variance of revised d''' was estimated by using Goodman's (1960) method; ViCi"'): : d-'[r'^V(p,) + pIv(R, ) + V(ft mR, )}. The total fishing days in /jth month was estimated by Equation 15 but in this case /)^ was unknown. The modi- fied estimator was Mn = M, + Dp, and the variance was slightly revised as Vmn) = N''Vm,) + E)'V(p,). The annual catch was estimated by Equation 16, sub- stituting Equation 8 by Equation 26. In this case, we estimated C[ and C,- ' from the same sample of season- permit anglers. Hence, the fourth term of the covariance in Equation 17 must be considered. The approximate covari- ance is estimated by (see Appendix 4) c'o^(c;;",c;/') = D'~p,,p,,c^v{R,^,Rf^.), ai) where the covariance between /?^, and /?;,. is Cov(R,,Rk)- N' M,M,, ^^Cov(C;-'',M^.) M,Ml M'kM^. Cov( Mi,C;?' + ± Z Cov(M^,M^.) Here Cov(Ci'",Ci?') and Cov(M,..Mi,.] are already given by Equation 18 and Equation 24. The other covariance com- ponents are c^v(&;\M,. ) = ^-'\ f (C,, - ^, )(M„. - M,.) n~{n - 1) ■^ C'^(M„C;?') = ^~'\ j^(C„. - 4 "M,, - M„ ). n (n -1)~ The annual catch in weight was estimated by Equation 19, substituting Equation 8 with Equation 26. The covariance in the fourth term of V(W) in Equation 17 was estimated with Equation 20 by c^( W;!'", vv;.'" ) = (t,(tj..(5rv(c;'",c;?' ) where Cov(C/ ,C^' ) is given in Equation 27. Results In 1993, 21,962 season permits and 9066 day permits were sold (Table 1). The total number of day permits sold at the four fishing tackle shops was 4776, and the number (pro- portion, p^) sold was 2732 (0.572) in June, 1,189 (0.249) in July, 716 (0.150) in August, 124 (0.026) in September, and 15 (0.003) in October. We received 104 logbooks from the 120 season-permit anglers sampled, a return rate of 86.7%. In addition, two anglers voluntarily submitted logbooks, but we did not in- clude these unsolicited returns because they were not ran- domly selected. The modes of the catch rates by the sam- pled anglers were from five to ten fish per month (Fig. 2). The histograms show a large variation in the catch rate among season-permit anglers. The peak fishing season was from June to July. In September, the number of anglers decreased, and the fishing season ended in October The 234 Fishery Bulletin 100(2) 1- 10 20 30 40 50 60- July 3n 1- Mil □_ 10 20 30 40 50 60- October n n 10 20 30 40 50 60- August Ha ^ r-l _ 4 2 10 20 30 40 50 60- Total nnn 10 20 30 40 50 60- 10 20 30 40 50 60- Catch (number of fish) Figure 2 Distributions of the catch rate of ayu by month for 104 sampled season-permit anglers in the Nakagawa River. modes of the number of fishing days per season-permit an- gler were five for all months. The variation in the niunber of fishing days among anglers was also large (Fig. 3). Monthly plots of the total number of fish caught versus the number of fishing days showed linear relationships (Fig. 4). The variation in Figure 4 indicates differences in the skill of the anglers. The monthly number of anglers decreased over the fishing season. Figure 5 shows the monthly changes in the total number of fishing days, the total number of fish caught, and the catch rate for the 104 sampled anglers. The decline in number offish caught was largely due to the decrease in fishing days. The change in catch rate indicated a decline in the abundance of the stock. The mean body weight of ayu was greatest in June (Fig. 6) and was affected by a method of fishing for ayu called "Tomo-zuri" angling, which takes advantage of the attack- ing behavior of ayu when another fish enters its territory. Anglers attach a "call" fish (a live ayu) above a treble hook that snares the territorial wild fish, as it attacks the "call" fish. Because larger individuals establish territories ear- lier than smaller ayu, fish caught in June were predomi- nantly the larger individuals. Reflecting the monthly trend in the number of fishing days, 89'7,_ of the total annual catches of season-permit an- glers and 98'>; of those of day-permit anglers were taken from June to August (Table 2). The catch by day-permit anglers was substantially smaller than anticipated, esti- mated at about 29^ of season-permit anglers' catch in both numbers and total weight. CVs ranged from 7% to 12% in June and July for all parameters; however, they were higher in August and September, ranging from 10% to 209; . In October, CVs exceeded 43% for total catch in num- ber and weight. The decreasing precision of the monthly catch rate estimates was caused by the decrease in anglers («,,) (Figs. 4 and 5). _ The CVs of annual estimates of M and Mj, by method 1 were about T'r. but that of R'"' was about 20':i (Table 2i because we evaluated the covariance terms for the number of catches and fishing days between months; those were Kitada and Tezuka: Survey designs for estimating recreational fishery catch 235 w 8- o 4- n E 8- June September £2. 5 10 15 20 25 July 8- 4- II 5 10 15 20 25 October 5 10 15 20 25 5 10 15 20 25 4n August 2- Total n^r. 5 10 15 20 25 10 30 50 70 90 Number of fistiing days Figure 3 Distributions of the ayu fistiing days per angler by niontii for 104 sampled season-permit anglers in the Nakagawa River Cov{C'if\Cl-[) in V(C'")and Cov(M;,,,M,. ) in Equation 22. The CVs of C and W were also evaluated at about 21% and were strongly affected by the covariances between months in Equation 17. The variance of the total number estimate V[C) was 1.2604 x 10'-. and variance by neglecting the covariance term in Equation 17 was 1.2230 x 10". The CV of C without the covariance term was 6.53%. If we ne- glect the covariance, the variance is substantially under- estimated. The variance was 10.31 times larger when the covariance term was included. We obtained similar point estimates of anijual catch by method 2 iTotal'"'^'^. 2 ;„ -pable 2 ). The CVs of M , Mj, C, and W for day-permit anglers were about 7%, but that of i?'" was reduced from 19.7% to 6.6% by not considering the co- variance terms. The CVs of C and W dropped about 10% from 21% without the covariance. Similar point estimates and smaller variance estimates were obtained. The vari- ance estimate of the annual catch obtained by method 1 with covariances ( 1.2604 xlO'-'l was 4.11 times larger than that by method 2 13.0667 xlO"). The relationship between the sample size and the pre- cision of the annual catch estimate for season-permit an- glers was examined. We calculated the values ViC) for var- ious values of n by using Equation 7. To obtain precision over the season for CVs of C''"( = Vviti/r) below 10%, a sam- ple size of 120 or more is required (Table 3). A high positive correlation in catches between adjacent months was detected (Table 4). We mapped anglers (ob- jects) and fishing days (categories) into a two-dimension- al graph by correspondence analysis (Hayashi, 1950; Ben- zecri, 1992) using the function "pqS.prcomp" in S version 4 (Chambers and Hastie, 1992). Correspondence analysis showed the relations between rows and columns in a fre- quency table gi'aphically as points in a common low-di- mensional space (Clausen, 1998). Both objects (rows) and categories (columns) of variables are represented as points in such a way that an object is relatively close to its catego- ry and relatively far from other categories (Leeuw and van Rijckevorsel 1988). For example, the 72nd angler fished 10 days in June, five days in July, seven days in August, three days in September, one day in October, and this angler was mapped closed to June, reflecting the month of his high- est fishing effort (Fig. 7). The results suggest several fish- ing patterns with high catch seasons in June-July, July- 236 Fishery Bulletin 100(2) 6-1 June n=95 September n=52 4- o ° 1 , . 5 10 15 20 25 5 10 15 20 25 o 6-| o 6-1 July o n=100 October n=16 CD O 4- 4- SI c/) O OJ E -=> c 2- 0- C „ oo oO o° o O °° 8 2" 5 10 15 20 25 5 10 15 20 25 6- tir' Total Q n=102 4- 12- 2- 0- 8- ooigo o ° 4. D 5 10 15 20 25 o O o J^^°o D 10 30 50 1 1 70 90 Number of fishing days Figure 4 Relat ionship between the number of fishing days and the number of fi; h cau ?ht by 104 sampled season-permit anglers in the Nakagawa River August, August-September, and September-October, re- sulting high correlations between adjoining months, and large covariances between distant months (Table 4). The prime advantage of a longitudinal study is its ef- fectiveness for studying change, and a repeated measures analysis of variance can be applied to a complete data set with a constant correlation (Diggle et al., 1994). However, our data set was incomplete because the number of anglers who fish in each month changed (see n in Fig. 4) and had a different correlation structure among month (Table 4). We tested the differences between successive monthly catch estimates of season-permit anglers by using a parametric bootstrapping method. In the central limit theorem, the sample distribution of a monthly total catch estimate can be regarded as a normal with the mean C^'-^'and the variance ViCj."). Based on the two point estimates, vari- ance estimates and the correlation coefficient between suc- cessive two months, we generated 10,000 bivariate normal random variables (Gentle, 1998). The means and 95% con- fidence intervals of the differences between two monthly total catch estimates were -226,561 1-650,404-203,080) for June and July, 870,720 1455,091-1,277,402] for July and August, 470,594 1161,013-783,1681 for August and September, and 537,488 1290,905-782, 727| for September and October. Significance levels were corrected for multi- ple testing by using the Bonferroni ajustment factor (So- kal and Rohlf, 1995). The confidence interval for June and July straddled 0, showing no significant difference. On the other hand, three other confidence intervals did not in- clude 0; therefore the monthly differences were statisti- cally significant (Fig. 8). Discussion Bias and source of variation The estimate of the total annual catch of ayu by the rec- reational fishery was the first obtained in Japan and was much larger than expected. The total number of day-per- mits sold was 9066, and was quite small (1.9%) compared with the estimated total number of anglers (477,520). Kitadd and Tezuka Survey designs for estimating recreational fishery catch 237 Although the difference in the catch rate between day- and season-permit anglers was unknown, the influence of this bias on the total catch estimates would be minor In order to check the bias, however, one could conduct a logbook survey of day-permit anglers. Sixteen anglers of the total sample ( 13'/r I in our study did not return logbooks and therefore may have caused a bias in our estimates; however no attempt was made to evaluate the difference between nonrespondents and respondents. The angler sample was drawn arbitrarily by the cooperatives but was not a random sample in the strictest sense. If cooperative anglers tended to be selected, this could have been a source of bias. The source of variation in total catch is the variation in the catch of the sampling unit, including differences in fishing days, skill of the anglers, and the number of an- glers that a party boat could accommodate. A stratified sampling scheme based on categories of anglers or boats is effective for this situation. The weakest point in the use of logbook surveys, perhaps, is that the catch data are report- ed by those who catch the fish and by boat owners with monetary interests. To what extent the anglers might have exaggerated or under- reported their catch is not known. Party boat owners may record lower than actual catches to reduce taxes. To examine this possible source of bias, on- site sui-veys should be conducted. For the ayu fishery in the Nakagawa River, an access point survey may be prac- tical (Pollock et al., 1994). When comparatively complete lists of boat owners and anglers are available, logbook sur- veys based on these lists, combined with on-site surveys, are appropriate. Longitudinal and stratifled survey designs Longitudinal surveys taking repeated measures on the same random sample over time are better than successive independent surveys if the designer's goal is to show sta- tistically significant differences in the estimates between time intervals. Monthly estimates showing seasonal trends 238 Fishery Bulletin 100(2) 30- 20- 10- 30- 20- _g 10- E 30- 20- 10- June n=47 Mean=56 Ma 20 40 60 80 100 July n=74 IVIean=46 n n n n II n 20 40 60 80 100 August n=68 Mean=52 n nn 30-1 20- 10- September n=63 Mean=45 n n 80-1 60- 40- 20- 20 40 60 80 100 Total n=252 IVlean=49 _IZL M JTL 20 40 60 80 100 20 40 60 80 100 Body weight (g) Figure 6 Distributions of body weight by month foi' ayu caught in the Nakagawa River- can be obtained by the equations derived in our study. In such repeated measure designs, the most precise esti- mates of annual catch are obtained by method 2. On the other hand, stratified surveys, or successive independent sui-veys between time intervals based on random samples of anglers or boats, are better than longitudinal surveys if the designer's goal is to obtain precise estimates of total effort, or catch (or total effort and catch), over the entire season. Stratification by month would improve the preci- sion of annual estimates even more if estimates varied greatly across months. Stratified sampling allows inde- pendent monthly estimates, and monthly estimates can be summed to produce precise estimates over time. In the absence of correlations between monthly sample obser- vations, the estimated variance of annual estimates can be obtained simply by adding the estimated variances of the monthly estimates. The estimated variances of annual estimates stratified by month would be considerably less than those of annual estimates based on repeated monthly observations of a one-time annual sample. If method 2 is used to analyze data obtained by such in- dependent surveys, how would the precision of the estima- tor compare with the precision of a stratified estimator? For simplicity, we consider a population that is divided into two subpopulations of A^,, N., units, respectively. The stratified estimator of the population total and the respec- tive variance are V{Y} = N^^V{y,) + N'^V{y,), where Vj and y-, are the sample means for sample sizes of «, and /!2. On the other hand, those obtained by method 2 are - TV, + N.-, _ r, = — ' =- 1 " i.v, + n.,y.:, ), ^,^., J ..(7V..N,, ]-y, j ,MiV, + 7V,, -^,-^, «] +n., n, + n.. Subtracting i' from Y. we have Kitada and Tezuka: Survey designs for estimating recreational fisliery catch 239 Table 2 Estima ed pai ameters and coefficients of variation lin pare nthescs). Total"""' ' = total estimated by summing iionthly estimates ( method 1 ). Total""""' ^ = total estimated by combining data throughout the season (method 2) fli- ' = catch rate for season-permit anglers M = mean number of fishing days per season-permit anglers. Mj. = total number of fishing days (s eason-i-day-permit | anglers ). C"^' = total catch in lumber for season-permit angl ers. C''' = total catch in number for day-permit angl srs. C = total catch in number (season+day-perm it anglers). W" = total catch in weight for season permit anglers. W-' = total catch in weight for day- | permit inglen . W = total catch in weight (season+day-permit anglers). Parameter June July August September October Total""-"' ' Total""^"' '^ fl'"' 11 12 12.37 10.29 9.29 5.35 11.04 11.04 (0.087) (0.075) (0.115) (0.187) (1.495) (0.197) (0.066) M 6.92 7.05 4.62 2.80 0.28 21.66 21.65 (0.073) (0.075) (0.101) (0.153) (0.291) (0.073) (0.073) M^ 157,231 157,047 102,722 61,687 6,152 484,839 484.628 (0.071) (0.0741 (0.100) (0.152) (0.290) (0.072) (0.071) QS) 1,690,440 1.914,495 1,042,772 570,800 32,732 5,251,241 5.251,241 (0.116) (0.115) (0.154) (0.179) (0.437) (0.214) (0.105) Qd) 57,658 27,915 13,982 2,186 152 101,894 100,108 (0.087) (0.077) (0.118) (0.197) (1.528) (0.056) (0.066) c 1,748,099 1.942,410 1,0.56,755 572,987 32,884 5,353,135 5,351,349 (0.112) (0.114) (0.152) (0.178) (0.435) (0.210) 10.103) w""(n 94.64 88.30 53.26 25.85 1.48 263.. 53 253.90 (0.125) (0.122) (0.159) (0.184) (0.439) (0.213) (0.107) W"''(n 3.23 1.29 0.71 0.10 0.01 5.34 4.84 (0.099) (0.086) (0.124) (0.202) (1.530) (0.065) (0.069) WU) 97.86 89.59 53,98 25.95 1.49 268.86 258.74 (0.122) (0.120) (0.157) (0.183) (0.437) (0.209) (0.105) Table 3 Coeffic ent of variations of the total catch es timate C'-" (method 2) for various sample sizes (number of anglers). ;; CV n CV n CV 10 0.3401 110 0.1025 230 0.0709 20 0.2405 120 0.0982 250 0.0680 30 0.1963 130 0.0943 300 0.0621 40 0.1700 140 0.0909 400 0.0.538 50 0.1521 1.50 0.0878 500 0.0481 60 0.1388 160 0.0850 600 0.04.39 70 0.1285 170 0.0824 700 0.0406 80 0.1202 180 0.0802 800 0.0380 90 0.1134 190 0.0780 900 0.0358 100 0.1075 200 0.0760 1000 0.0340 Y-Y.. n.>Ni-niN2 (>'i-y2'- showing that the two methods provide different estimates with the extent of the difference, depending upon the Table 4 Estimated variance-covariance matrix ( xlO"') for the monthly estimates of catch (number) by season-permit 1 anglers in number by Equation 17 (lower diagonal) and the correlation coefficient r(C,(., C,.) (upper diagonal, m bold font). The diagonal component refers to V'(Ct and the lower half refers to Cov(Cj",Cj'' ). Month Jun Jul Aug Sep Oct Jun 3.837 0.67 0.43 0.30 -0.03 Jul 6.011 4.862 0.65 0.49 0.08 Aug 9.422 4.434 2.578 0.65 0.28 Sep 10.681 6.258 1.456 1.038 0.44 Oct 10.725 6.336 1.522 0.069 0.020 sizes of the strata, the sample sizes, and the estimates of the sample means. According to Cochran (1977), method 2 works well enough if the sample allocation is propor- tional because a simple random sample distributes itself approximately proportionally among strata. With propor- tional allocation N^/n^ = NJn.^; therefore the difference of the two estimators is 0. In our case study, the annual 240 Fishery Bulletin 100(2) 100 97 O - d 20 99 ,7 12 2|4 10 ,^95 ^1 ^1, ond axis 0.0 1 ^J430 2Au^^g '' ^393 4 1 80 Juty ags ^' 77 93 31 32 34J, 90 Sept. ^.sse. ^^^^ -P'j^ ,6 71 f»6 35 57 ^^^02 3B2 June72 ^ ^^6 « in o 9 47 51 ^3 " 18 ^' 55 o 5 " 70 Oct. 1 1 1 1 1 1 1 -008 -006 -004 -002 00 02 0.04 First axis Figure 7 Plots of the determined quantities for sample (anglers) and category (months) from the correspondence analysis for the fishing days of 104 sam- pled anglers. The numbers in the figure refer to the number of anglers that were sampled catch estimates for the season-permit anglers C"' were the same value for metho(i 1 and method 2. The population size N and the sample size /; were in proportion constant throughout the season; therefore NJn^ = NIn was same for all strata. The ratio of the variances is V(Y^ ) {N,+ N., fU'fViy, ) + nlV(y, )| ' which also shows that the precision of both methods depends on the sizes of the strata, the sample sizes, and the variance of the sample means. If the allocation is proportional, the variance ratio becomes 1 and the two variances coincide. The objective of stratified surveys is to obtain precise total effort or catch estimates (or both) over the entire season. A proportional sample allocation is recommended, which allows a simple calculation with method 2 and improves precision of estimates at the same time. Acknowledgements We extend our gratitude to the anonymous referees and to John V. Merriner, Katherine Myers, and Sharyn Matriotti for their critical readings that greatly improved the man- uscript. We also thank John M. Hoenig, John B. Pearce, Yashushi Taga, and Geoff Gordon for their comments on earlier versions of the manuscript. A portion of this study was funded by the Fisheries Agency of Japan (Japan Sea- Farming Association ). Kilada and Tezuka: Survey designs for estimating recreational fishery catcfi 241 >. -1,000,000 it ^„„ Jul vs. Aug 500 ^ 300 500,000 100 _jl 500 300 100 200,000 600,000 1,000.000 Sep vs. Oct 500,000 1,500,000 200,000 600,000 1,000,000 Difference between two total estimates Figure 8 Bootstrap distributions of diffences between successive monthly ayu catch estimates. Literature cited Benzecri, J, P. 1992. Correspondence analysis handbook. Dekker. New York, r^, 665 p. Chambers J. M., and, T. J. Hastie (eds.) 1992. Statistical models in S. Chapman and Hall, New York, NY, 608 p. Clausen, S.-E. 1998. Applied correspondence analysis: an introduction. SAGE Publications, Thousand Oaks, CA, 69 p. Cochran, W. G. 1977. Sampling techniques, third ed. John Wiley and Sons, New York, NY, 41.3 p. Diggle, P. J, K-Y. Liang, and S. L. Zeger. 1994. Analysis of longitudinal data. Oxford Univ. Press, Oxford, 253 p. Gentle, J. E. 1998. Random number generation and monte carlo meth- ods. Springer, New York, NY, 247 p. Goodman, L. A. 1960. On the exact variance of products. J. Am. Stat. Assoc. .55:708-713. Guthrie, D.. J. M. Hoenig, M. Holliday, C. M. Jones, M. J. Mills, S. A. Moberiy, K. H. Pollock, and D. R. Talhelm (eds). 1991. Creel and angler sui-veys in fisheries management. Am. Fish. Soc. Symp. 12, Betliesda, MD. 528 p. Hayashi, C. 1950. On the prediction of phenomena from mathematical statistic point of view. Annals Inst. Stat. Math. 3:69-98. Jones, C. M., D. S. Robson, H, D. Lakkis, and J. Kressel. 1995. Properties of catch rates used in analysis of angler surveys. Trans. Am. Fish. Soc. 124: 911-928. Leeuw, J. de, and J. L. A. van Rijckevorsel. 1988. Beyond homogeneity analysis. In Component and corre- spondence analysis (J. L. A. van Rijckevorsel and J. de Leeuw, eds.), p. 56-57. John Wiley and Sons, New York, NY. Pollock, K. H., J. M. Hoenig, C. M. Jones, D. S. Robson, and C, J. Greene. 1997. Catch rate estimation for roving and access point sur- veys. N.Am. J. Fish. Manag. 17:11-19. 242 Fishery Bulletin 100(2) Pollock, K. H., C. M. Jones, and T. L. Brown. 1994. Angler survey methods and their applications in fish- eries management. Am. Fish. Soc. Special Publication 2.5, Bethesda, MD, 371 p. Sober, G. A. F. 1982. The estimation of animal abundance and related parameters, second ed. Griffin, London, 6.54 p. Sokal, R. R.,andF.J. Rohlf 1995. Biometry, third ed. Freeman and Company, New York, NY, 887 p. Thompson. S. K. 1992. Sampling. John Wiley and Sons, New York, NY', .343 p. Appendix 1 : The covariance of two estimators from sample means First we consider the covariance of two total estimates. Let X, and Yj be simple random samples (i=l n) from a population of size N with mean //^ and ;/^, and A' and Y be two sample means. Cochran (1977. p .2.5) derived the covariance of two- sample mean, that is Cov( X. y ) = ^^^!— ^ - Cov( X, y ) N n N-n N tAY^X -^ )(Y -u ). This is estimated by Cov(x, y ) = ^^-^ - Cov( X, y N n > (A, -X){Y, -Y Nnin- W, = C,tv, + w, (C,-C,) + C,iw,-IU,). From Taylor's series (mentioned above), the approximate covariance is obtained. Cov( w' " , w;.' ' ) = £( w; ■" - w; -^ ' ) ( w'. ' -w:?') -£[^,(Cl'^'-Cr') + C-'([Z7,. A^ Here Q''|' and iZij , are independent, and both Wk and u'a; are estimated from different samples. Therefore Cov (QVilij. ) = Cov(w,,Cj:^ = Covi w,,w,. ) = 0, then we get the covariance as only the first term. Appendix 3: Approximate variance of ^ Taylor's series of R with respect to C and M is obtained by" ^ = — -F— (C-C)--^(M-M). MM M- Then the approximate variance is obtained by V(i?) = £;(^-i?)-"- J-V(C) + -£-V'(M) M' M' -'iS-CoviCM). The covariance between two population total estimators is defined by Cov(X,y ) = CmtNX.NY) = EiNX - N^, HNY - N/u^. ) ., -^ NiN-n) — = Af-Cov(X,y)= cov(X.y). n This is estimated by Appendix 4: Covariance between C'lf' and C, By expanding C,.'' and C",''' we get CI'' ' = Dp,R, + DR, ( p, - p, ) + Dp, ( i?, - «, ). Q'^' = Dp,R, + DR,Ap,. -p,) + Dp.AR,. - R, (d) k c7viX,Y) = ^^^^^^^^±iX,-XnY,-Y). For the monthly total catches, we get c^(cr',c-' ) = ^'^";'' X ic,, - ^, )(c„. - 1 nin - 1) ■'— ' Appendix 2: Approximate covariance between Wl"and Wl" Taylor's series of W^ with respect to the random variables is obtained by then the approximate covariance is given by Cov(c;,'",c;'" ) = £(c;.'" - Dp,,/?,, kc;;" - Dp,yR,, = D- R,M,Covi p, , p, I + /?,p, Cov( p, .R,) +p,M,XoviR,.p,, ) + p,p, Cov(i?,.i?,. I If the first three covariance components are ecjual to because of independent sampling, then we have Cov(c;;",c;;" i = £»-p,,p,,Cov( /?,.,/?, ). Here we can write R/. and Rf. as the ratio of two random variables from Equation 1 by Kitada and Tezuka Siirvey designs for estimating recreational fishery catch 243 R,= C(s) By using a method similar to that given in Appendix 3, we get Cov(R,.,R.,)- N' 1 M,M,. -cov(c;-'",c;f') M,A/^. -CoviCl",M,.) J^'^ Cov(M„C;r') Cov{M,,,Mi,) Cov(Q'",Af^,.) and Cov(M,,,Cj.?'). These are the covariances between a total estimate and a sample mean. By a method similar to that in Appendix 1, we have Cov(i-,y ) = EiNX - NiJ^)(Y - fi^.) = NCov(X,y ) = Cov(X.y). The covariance is estimated by Cov(i'T)= ^ " y{X~X)(Y-Y). n(n - 1) f^ For our case, the two covariances are as follows; n (« - 1) ■"' Here Cov(C'i'",C'i'')and CovlM^M,. I are already given by Equation 18 and Equation 24. Hence we can estimate Cov(M, , CI? ' ) = !y " X ' C,*- - Q- ) (^,* - M, nin- ll •^ 244 Abstract— Juvenile chinook salmon, Oncorbyncluis tshawytscha. from natal streams in California's Central Valley demonstrated little estuarine depen- dency but grew rapidly once in coastal waters. We collected juvenile chinook salmon at locations spanning the San Francisco Estuary from the western side of the freshwater delta — at the con- fluence of the Sacramento and San Joa- quin Rivers — to the estuary exit at the Golden Gate and in the coastal waters of the Gulf of the Farallones. Juveniles spent about 40 d migrating through the estuary at an estimated rate of 1.6 km/d or faster during their migration season (May and June 1997) toward the ocean. Mean growth in length (0.18 mm/d) and weight (0.02 g/d> was insignificant in young chinook salmon while in the estuary, but estimated daily growth of 0.6 mm/d and 0.5 g/d in the ocean was rapid (PsO.OOll. Condition (A' factor) declined in the estuary, but improved markedly in ocean fish. Total body pro- tein, total lipid, triacylglycerols (TAG), polar lipids, cholesterol, and nonesteri- fied fatty acids concentrations did not change in juveniles in the estuary, but total lipid and TAG were depleted in ocean juveniles. As young chinook migrated from freshwater to the ocean, their prey changed progressively in importance from invertebrates to fish larvae. Once in coastal waters, juve- nile salmon appear to employ a strat- egy of rapid growth at the expense of energy reserves to increase survival potential. In 1997, environmental con- ditions did not impede development: freshwater discharge was above aver- age and water temperatures were only slightly elevated, within the species' tolerance. Data suggest that chinook salmon from California's Central Valley have evolved a strong ecological pro- pensity for a ocean-type life history. But unlike populations in the Pacific Northwest, they show little estuarine dependency and proceed to the ocean to benefit from the upwelling-driven, bio- logically productive coastal waters. Physiological ecology of juvenile chinook salmon iOncorhynchus tshawytscha) at the southern end of their distribution, the San Francisco Estuary and Gulf of the Farallones, California* R. Bruce MacFarlane Elizabeth C. Norton Santa Cruz Laboratory Southwest Fisheries Science Center National Marine Fisheries Ser^/lce, NOAA 110 Shaffer Road Santa Cruz, California 95060 E-niail address (for R B MacFarlane) Bruce MacFarlane a noaa gov Manuscript accepted 23 August 2001. Fish. Bull. 100:244-2.57 (2002). Estuaries are considered important in the development of juvenile salmon. In the Pacific Northwest, estuaries have been shown to provide nursery and rearing conditions for juveniles emigrat- ing from streams of birth to the ocean (Reimers, 1973; Healey, 1982; Levy and Northcote, 1982; Myers and Horton. 1982; Simenstad et al, 1982; McCabe et al, 1986). The San Francisco Estuary is the largest estuary on the West Coast and is a segment in the migration corri- dor for chinook salmon (Onc-orhyru'luis tshawytscha ) from natal streams in the watersheds of the Sacramento and San Joaquin Rivers, known as California's Central "Valley The Central Valley is unique by having four runs of chinook salmon which constitute a significant socioeconomic resource. Ocean harvest south of Pt. Arena (estimated as 85-95% from Central Valley stocks) and spawn- ing escapement range between 0.5 and 1.3 X 10'' chinook salmon per year ( 1970-98) and represent about $60 mil- lion (U.S.) in personal income annually (PFMCM. Beyond the direct value of Central Valley chinook salmon, their demography and welfare significantly affect the financial and societal aspects of water rights decisions. Chinook salmon populations migrat- ing through the San Francisco Estuary are at the southern limit of the species' geographical range and are subject to the impacts of a highly urbanized, in- dustrialized, and agricultural freshwa- ter and estuarine system (Nichols et al., 1986).A11 chinook salmon runs originat- ing in the Central Valley are in jeopar- dy. Before 1900, spawning runs were es- timated at 2 X 10« adults (Fisher, 1994), but in 1998 only an estimated 0.25 X 10'' returned of which about 30% were of hatchery origin (PFMCM. The Sacramento River winter-run chinook was the first Pacific salmonid species listed under the U.S. Endangered Spe- cies Act of 1973 (ESA). Originally cat- egorized as threatened in 1989, its sta- tus was changed to endangered in 1994. Chinook salmon of the Central Valley spring run, once forming the dominant chinook race in California (Clark, 1929). were listed as threatened in 1999. Even the fall run, by far the dominant run to- day ( 92% of all Central Valley spawn- ers, 1990-98 IPFMC'l), has uncertain status and is an ESA candidate. Hatch- ery production supports the natural fall run, and the other runs to a much lesser degree. Annually, about 35 million chi- nook salmon are produced by state and federal hatcheries in the Central Val- ley; the fall run