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 Homa