St ( 1 1 a a rs3 r i s (4 U.S. Department of Commerce Volume 96 Number 3 July 1 998 Fishery Bulletin £ U.S. Department of Commerce William M. Daley Secretary National Oceanic and Atmospheric Administration D. James Baker Under Secretary for Oceans and Atmosphere National Marine Fisheries Service Rolland Schmitten Assistant Administrator for Fisheries The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fish- eries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115- 0070. Periodicals postage is paid at Se- attle, WA, and at additional mailing of- fices. POSTMASTER: Send address changes for subscriptions to Fishery Bul- letin, Superintendent of Documents, Attn. : Chief, Mail List Branch, Mail Stop SSOM, Washington, DC 20402-9373. Although the contents of this publica- tion have not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this peri- odical is necessary according to law for the transaction of public business of this De- partment. Use of funds for printing of this periodical has been approved by the Di- rector 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: $34.00 domestic and $42.50 foreign. Cost per single issue: $13.00 doemestic and $16.25 foreign. See back for order form. Scientific Editor Dr. John B. Pearce Editorial Assistant Laura Garner Northeast Fisheries Science Center National Marine Fisheries Service, NOAA 1 66 Water Street Woods Hole, Massachusetts 02543-1097 Editorial Committee Dr. Andrew E. Dizon National Marine Fisheries Service Dr. Harlyn O. Halvorson University of Massachusetts, Boston Dr. Ronald W. Hardy University of Idaho, Hagerman Dr. Richard D. Methot National Marine Fisheries Service Dr. Theodore W. Pietsch University of Washington, Seattle Dr. Joseph E. Powers National Marine Fisheries Service Dr. Harald Rosenthal Universitat Kiel, Germany Dr. Fredric M. Serchuk National Marine Fisheries Service Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C 1 5700 Seattle, Washington 98115-0070 The Fishery Bulletin carries original research reports and technical notes on investiga- tions in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Begin- ning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 96 Number 3 July 1998 Fishery Bulletin The National Marine Fisheries Service (NMFS) does not approve, recommend, or endorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, recommends, or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the adver- tised product to be used or purchased because of this NMFS publication. Contents Articles 395-414 415-427 428-437 438-450 Ault, Jerald S., James A. Bohnsack, and Geoffrey A. Meester A retrospective ( 1 979-1996) multispecies assessment of coral reef fish stocks in the Florida Keys Collins, L. Alan, Allyn G. Johnson, Christopher C. Koenig, and M. Scott Baker Jr. Reproductive patterns, sex ratio, and fecundity in gag, Mycteroperca microlepis (Serranidae), a protogynous grouper from the northeastern Gulf of Mexico Gannon, Damon P., James E. Craddock, and Andrew J. Read Autumn food habits of harbor porpoises, Phocoena phocoena, in the Gulf of Maine Gorfine, Harry K., David A. Forbes, and Anne S. Gason A comparison of two underwater census methods for estimating the abundance of the commercially important blacklip abalone, Haliotis rubra 451-461 462-481 482-491 492-501 502-515 Jones, Cynthia ML, and Brian Wells Age, growth, and mortality of black drum, Pogonias cromis, in the Chesapeake Bay region Kimura, Daniel K., Allen M. Shimada, and Franklin R. Shaw Stock structure and movement of tagged sablefish, Anoplopoma fimbria, in offshore northeast Pacific waters and the effects of El Nino-Southern Oscillation on migration and growth Liu, Kwang-Ming, Po-Jen Chiang, and Che-Tsung Chen Age and growth estimates of the bigeye thresher shark, Alopias supercihosus, in northeastern Taiwan waters Love, Mlilton S., Jennifer E. Case He, and Kevin Herbinson Declines in nearshore rockfish recruitment and populations in the southern California Bight as measured by impingement rates in coastal electrical power generating stations Lowe, Sandra A., Donald M. Van Doornik, and Gary A.Winans Geographic variation in genetic and growth patterns of Atka mackerel Pleurogrammus monopterygius (Hexagrammidae), in the Aleutian archipelago ii Fishery Bulletin 96(3), 1998 516-524 Martini, Frederic H., Michael R Lesser, and John B. Heiser A population profile for hagfish, Myxine glutinosa, in the Gulf of Maine. Part 2: Morphological variation in populations of Myxine in the North Atlantic Ocean 525-537 Merkouris, Susan E., Lisa W. Seeb, and Margaret C. Murphy Low levels of genetic diversity in highly exploited populations of Alaskan Tanner crabs, Chionoecetes bairdi, and Alaskan and Atlantic snow crabs, C. opilio 538-546 IWIIunro, Peter T. A decision rule based on the mean sguare error for correcting relative fishing power differences in trawl survey data 547-561 Nichol, Daniel G. Annual and between-sex variability of yellowfin sole, Pleuronectes asper, spring-summer distributions in the eastern Bering Sea 562-574 Rocha-Olivares, Axayacatl Age, growth, mortality, and population characteristics of the Pacific red snapper, Lutjanus peru, off the southeast coast of Baja California, Mexico 575-588 Scharf, Frederick S., Richard M. Yetter, Adam P. Summers, and Francis Juanes Enhancing diet analyses of piscivorous fishes in the Northwest Atlantic by identification and reconstruction of original prey sizes from ingested remains 589-602 Schmid, Jeffrey R. Marine turtle populations on the west-central coast of Florida: results of tagging studies at Cedar Keys, Florida, 1986-1995 603-613 Secor, David H., and Troy E. Gunderson Effects of hypoxia and temperature on survival, growth, and respiration of juvenile Atlantic sturgeon, Acipenser oxyrinchus 614-620 Waldman, John R., Reese E. Bender, and Isaac 1. Wirgin Multiple population bottlenecks and DNA diversity in populations of wild striped bass, Morone saxatilis 621-623 Notes Bej da, Allen J., and Beth A. Phelan Can scales be used to sex winter flounder, Pleuronectes americanus ? 624-627 Gelsleichter, James, Enric Cortes, Charles A. Manire, Robert E. Hueter, and John A. Musick Evaluation of toxicity of oxytetracycline on growth of captive nurse sharks, Ginglymostoma cirratum 628-632 Koeller, Peter, and Gregory Crowell Electrotaxis in American lobsters, Homarus americanus, and its potential use in sampling early benthic-phase animals 633-640 Pepin, Pierre, John F. Dower, and William C. Leggett Changes in the probability density function of larval fish body length following preservation 641-646 Tanabe, Toshiyuki, and Kodo Niu Sampling juvenile skipjack tuna, Katsuwonus pelamis, and other tunas, Thunnus spp., using midwater trawls in the tropical western Pacific 647-650 Wells, Randall S., Suzanne Hofmann, and Tristen L. Moors Entanglement and mortality of bottlenose dolphins, Tursiops runcatus, in recreational fishing gear in Florida 651 Awards 652 Subscription form 395 Abstract .—A baseline assessment of 35 economically and ecologically im- portant Florida Keys reef fish stocks is provided by using a systems approach that integrates sampling, statistics, and mathematical modeling. Quantita- tive fishery-independent data from reef fish visual surveys conducted by SCUBA divers from 1979 to 1996 were used to develop estimates of population abundance, assemblage composition, and stock structures in relation to key physical and habitat factors. Exploita- tion effects were assessed with a new length-based algorithm that calculates total mortality rates from estimates of “average length of fish in the exploit- able phase of the stock.” These esti- mates were highly correlated for two statistically independent data sources on reef fish: fishery-independent diver observations and fishery-dependent head boat catches. We developed a reef fish equilibrium exploitation fishery simulation (REEFS) model and used es- timates of fishing mortality to assess yield-per-recruit in relation to fishing intensity and gear selectivity and to assess spawning potential ratio (SPR) in relation to U.S. federal “overfishing” standards. Our analyses show that 13 of 16 groupers (Epinephelinae), 7 of 13 snappers (Lutjanidae), one wrasse (Labridae), and 2 of 5 grunts (Haemuli- dae) are below the 30% SPR overfish- ing minimum. Some stocks appear to have been chronically overfished since the late 1970s. The Florida Keys reef fishery exhibits classic “serial overfish- ing” in which the largest, most desir- able, and vulnerable species are de- pleted by fishing. Rapid growth of the barracuda population (Sphyraenidae) during the same period suggests that fishing has contributed to substantial changes in community structure and dynamics. Manuscript accepted 16 December 1997 Fishery Bulletin 96(3):395-414 (1998). A retrospective (1979-1 996) multispecies assessment of coral reef fish stocks in the Florida Keys Jerald S. Ault Rosenstiel School of Marine and Atmospheric Science University of Miami 4600 Rickenbacker Causeway, Miami, Florida 33 1 49 E-mail address: ault@shark.rsmas.miami.edu James A. Bohnsack Southeast Fisheries Science Center National Marine Fisheries Service, NOAA 75 Virginia Beach Drive, Miami, Florida 33149 Geoffrey A. Meester Rosenstiel School of Marine and Atmospheric Science University of Miami 4600 Rickenbacker Causeway, Miami, Florida 33149 The Florida Keys support a rich tropical marine ecosystem, a pro- ductive multispecies coral reef fish- ery, and a billion dollar tourist economy. The Florida Keys are also considered an “ecosystem-at-risk” as one of the nation’s most signifi- cant yet most stressed marine re- sources under management of the National Oceanic and Atmospheric Administration (NMFS, 1996). Con- cern about habitat degradation and escalating resource uses from a rap- idly growing human population in southern Florida resulted in the es- tablishment of the Florida Keys Na- tional Marine Sanctuary ( FKNMS ) in 1990. Because they are one of the most complex ecosystems on earth, coral reefs are a particular concern. The diverse fish community of these coral reefs is influenced by compli- cated biological and physical inter- actions (Sale, 1991; Lee et al., 1992; Polunin and Roberts, 1996). Reef fisheries can target a number of eco- nomically and ecologically impor- tant species (e.g. groupers, snap- pers, lobsters, conch, sponges, and corals). Over the past several de- cades, public use of and conflicts over fishery resources have increased sharply, while some fishery catches from historically productive snapper and grouper stocks have declined (Bohnsack et al., 1994). Concomi- tantly, the status and biological dy- namics of these reef fishery re- sources are not well understood, and important stock assessment data are not available. Another concern regarding reef fishery resources is the restoration of the Everglades north of the Florida Keys. Hydrological projects of historic proportions are expected to make a substantial change in the timing, volume, and location of fresh- water outflows into the coastal ma- rine environment (Harwell et al., 1996). These changes could affect the survivorship of juvenile reef fishes in critical shallow nursery areas of Florida Bay and Biscayne Bay and ultimately affect the productivity of the entire coral reef ecosystem. 396 Fishery Bulletin 96(3), 1998 The condition of most reef fish stocks is unknown because of the large number of species in the fishery, a lack of fishery-effort and landings data, and the quantity of population dynamics data needed to do traditional stock assessments. The goal of this pa- per is to develop a technically sound quantitative method for multispecies management assessments. For this purpose, we present an integrated baseline assessment to reference the status of the multispecies fishery so that the effects of management changes in the FKNMS may be accurately evaluated in the future (U.S. Dep. Commerce, 1996). Using fishery- independent data, we conducted an 18-year retro- spective, analytical yield assessment of economically important Florida Keys reef fish stocks to elucidate the effects of fishing and to help define an effective fishery management strategy. Hypothesis A key to our ability to assess reef fish stocks was the use of “average size” (in length) of fish in the exploit- able phase of the population ( L ) as an indicator of stock status. Average size of fish was derived from visual survey data or headboat landings data. Headboats are party boats that carry more than 15 anglers per fishing trip (Dixon and Huntsman1). The use of L in stock assessment has deep roots in tra- ditional fisheries management (Beverton and Holt, 1956, 1957; Ricker, 1975). The statistic provides a population level metric that integrates individual metabolic variables such as interdependent_growth, mortality, and reproductive processes. The L statis- tic also is an important index of fishing effects be- cause persistent heavy fishing reduces the average size of the population over time, making the stock younger through a process known as “juvenescence” which successively eliminates older, more fecund size classes (Ricker, 1963). This is extremely important in the context of stock and recruitment because the fecundity potential of individuals increases exponen- tially with size. In general, the average length of fish in the exploitable phase (i.e. between the size at first capture, L', and the maximum size, Lx ) is highly corre- lated with average population size and thus reflects the rate of fishing mortality operating in the fishery. Theoretically, the average size of fish landed for any given species should be equal to the average size in the exploited phase of the remaining population 1 Dixon, R. L., and G. R. Huntsman. 1992. Estimating catches and Fishing effort of the southeast United States headboat fleet, 1972-1982. Beaufort Laboratory, Southeast Fisheries Science Center, Natl. Mar. Fish. Serv., NOAA, Beaufort, NC 28516. Draft report. just after fishing. In other words, we hypothesize that fishery-independent survey estimates of average length derived from visual data reported by divers should equal fishery-dependent estimates derived from catch data reported by headboat anglers. The greater the correlation between the two independent estimates of L , the more robust “average length” ( should be as an indicator of stock status subject to exploitation. Methods and materials Study area The Florida Keys coral reef ecosystem is a unique tropical coastal marine environment stretching about 370 km from Key Biscayne southwest to the Dry Tortugas (Fig. 1). Situated parallel to the Florida current and Florida Bay, the coastal ecosystem en- compasses many varied habitats comprising fresh- to saltwater marshes, estuaries, lagoons, mangrove stands, coral islands, sea grass beds, and coral reefs. Florida Bay and adjacent coastal estuaries serve as nursery areas for spiny lobster and many juvenile fishes that occupy reefs as adults. The clear water and high diversity of reef fish in the Florida Keys coral reef tract provide a unique environment to as- sess multispecies fisheries. Here we use a “systems approach” to facilitate effective decision making and to improve fishery management performance (Ault and Fox, 1989; Rothschild et al., 1996). Reef fish surveys Fishery-independent visual estimates of the abun- dance and size distributions of multispecies reef fish populations were taken along the Florida Keys reef track continuously from 1979 to 1996 (Table 1) by 12 highly trained and experienced divers using the sta- tionary visual survey method of Bohnsack and Bannerot (1986). This nondestructive method pro- vides reliable quantitative estimates of species abun- dance, frequency-of-occurrence, and size structure for the reef fish community. Divers recorded the abun- dance as well as the minimum, mean, and maximum lengths of each species seen during 5 minutes within randomly selected 7.5-m radius circular quadrats. Underwater visual estimates of reef fish size and abundance have frequently been made (Bellwood and Alcala, 1988; Harvey and Shortis, 1996); however, accurate and precise visual estimates of fish length require well-trained and experienced observers be- cause objects in water appear magnified and closer than their actual range (Bell et al., 1985; Harvey and Ault et al.: A multispecies assessment of coral reef fish stocks 397 Figure 1 Map of the Florida Keys coastal marine ecosystem showing the coral reef tract (light gray) running offshore from Key Biscayne southwest to the Dry Tortugas, and the spatial relationships of Florida Bay, Key West, and the Miami urban center. Numbered darkened circles show the 89 reefs where 4,571 visual survey samples of the reef fish community were taken from 1979 to 1996. Open circles around numbers indicate sanctuary preservation areas (SPAs). Table 1 Visual survey sampling effort (number of point samples) by habitat areas and depth intervals conducted from 1979 to 1996 in the Florida Keys reef tract. The offshore zone is exposed to the Florida Current. Reef zone habitat type Artificial Coral Hard Sand Depth reef reef bottom bottom Totals Feet Meters Inshore Inshore Offshore Inshore Offshore Inshore Total Percent 0-10 0-3.05 0 848 171 26 8 13 1,066 23.32 10-20 3.05-6.10 5 726 816 14 207 38 1,806 39.51 20-30 6.10-9.14 85 403 561 4 146 0 1,199 26.23 30-40 9.14-12.19 28 31 92 0 40 0 191 4.18 40-50 12.19-15.24 9 9 81 0 48 0 147 3.22 50-60 15.24-18.29 0 15 65 0 47 0 127 2.78 60-70 18.29-21.34 0 1 34 0 0 0 35 0.77 Total 127 2,033 1,820 44 496 51 4,571 Percent 2.78 44.48 39.82 0.96 10.85 1.12 100.00 Shortis, 1996). To improve accuracy, divers continu- ously calibrated their length estimates with a 30-cm ruler attached perpendicular to the far end of a meterstick. Divers without calibration sticks have been shown to obtain a mean accuracy of 86% for length estimates (St. John et al., 1990). 398 Fishery Bulletin 96(3), 1 998 Maximum use of the visual survey data required statistical intercalibration of the sampling efficiency of each diver. We used multiple regression analysis (Neter et al., 1996) to estimate relative sampling ef- ficiency by adapting the “fishing power” model of Robson (1966) C(d,t) = F(d,t)N(t)i;(d,t) = q(d)f(d,t)N(t)i;(d,t) C(d,t) ■» . = q(d)N (d,t), f(d,t) (1) where C(d,t) = the fish count of diver d at reef date t; Nit) = the average population size at reef date t; q(d) = the coefficient of sampling efficiency for diver d; f(d,t) = the nominal survey effort of diver d at reef date t, and %(d,t) = a log normally distributed error variable. To account for any sampling bias that may have been introduced by differences among divers, a simple log- linear transformation of Equation 1 makes it pos- sible to obtain minimum variance estimates of rela- tive sampling efficiency, given fish counts by species, by diver, and by reef date. Data used for model de- velopment were derived from a series of controlled experiments conducted during a 9-day sampling ex- pedition to the Dry Tortugas during June 1994. A matrix of estimated efficiency coefficients for divers by species was used to adjust an individual diver’s results in relation to a standard-normal diver, here the most experienced diver in the group. After stan- dardization, all the individual visual “catch-per-unit- of-effort” measurements were comparable over time and space (Ault et al.2). Spatial and temporal pat- terns in abundance and correlative linkages to habi- tat types were qualitatively analyzed with 3-D visu- alization software (IDL, 1995) by reef site through- out the Florida Keys for various survey years. Mul- tivariate statistical analysis (Johnson and Wichern, 1992; Venables and Ripley, 1994) was used to assess variance-covariance and correlation structures be- tween reef fish density and selected environmental and fish community auxiliary covariates. We also used the 1981-95 NMFS headboat catch- and-effort data (Bohnsack et al., 1994; Dixon and Huntsman, 1992) to provide fishery-dependent popu- lation estimates comparable to those from the visual 2 Ault, J. S., J. A. Bohnsack, and G. Meester. 1998. The rela- tive fishing power of divers in tropical reef fish visual surveys. Unpubl. manuscript. survey. Headboat data provide total numbers of in- dividuals in the catch as well as total weight in the catch by species by year. Stock assessment indicator variable A stock assessment indicator variable is a quantita- tive measure that reflects the status of a population subjected to fishing or other environmental changes. Because reef fishes integrate aspects of the coastal ocean environment over their lifetime, a robust mea- sure of population “health” or status can provide a sensitive indicator of direct and indirect stress on the stock, and perhaps on the regional marine eco- system (Fausch et al., 1990). Population health for reef fish communities can best be described with the metabolic-based pool variable “average length in the exploitable phase of the stock.” Therefore, to assess the health of each of the s stocks in the reef fish com- munity over the past two decades, the statistic “av- erage length in the exploitable phase of the stock,” L ( t ), was found for each stock by integrating between the population age limits from t ' ( minimum age at first capture) to tx (oldest age in the stock), written as tk F(t)J N(a,t)L{a,t)da F(t)J N(a,t)da t ' where N(a,t) = abundance for age class a at time t; L(a,t) = length for class a at time t; and Fit) = the instantaneous fishing mortality rate at time t. Estimates of the mean, variance, and 95% confidence interval of the mean were computed by the methods of Sokal and Rohlf ( 1969). To estimate the annual total instantaneous mor- tality rate Zit) for each fish stock in each year t from population size structure and abundance statistics, we used a length-based method (Ault and Ehrhardt, 1991; Ehrhardt and Ault, 1992) particularly appli- cable to reef fish population dynamics Z(t)(L' - L(t)) + K(L„ - Lit)) Z{t)(Lx - Lit)) + K(L„ - Lit)) ’ <3) where, Lx _V Lit) maximum size; the length at first capture; the average size in the exploitable phase in year t; and Ault et al.: A multispecies assessment of coral reef fish stocks 399 K and - parameters of the von Bertalanffy equation. Thus, the only unknown variable in Equation 3 is the total mortality rate in year t, Z(t), which can be esti- mated fairly easily with an iterative algorithm called LBAR (Ault et al., 1996). Finally, by assuming that M, the annual instantaneous rate of natural mortality, is known and constant for the interval At, we can esti- mate the fishing mortality rate as F(t)= Z it)-M. Reef fish exploited-populafion simulation model To achieve a better understanding of the dynamics of multispecies tropical coral reef fish stocks, their response to exploitation, and the accuracy and pre- cision of statistical estimates from the sampling sur- veys, we developed an object-oriented computer simu- lation model for exploited reef fish populations in the Florida Keys fishery (REEFS [reef fish equilibrium exploitation fishery simulator (Ault3)]). The funda- mental population-dynamic processes of growth, mortality, and recruitment are relatively similar for fishes of the temperate, boreal, and tropical seas; however, some distinct differences in rates exist for tropical marine fishes as reflected by quasicontinuous growth, protracted spawning and recruitment, and competition-based population dynamics (Ault and Fox, 1990; Sparre and Venema, 1992; DeMartini, 1993). To represent the continuous time dynamics of a tropical coral-reef fish population in the numerical model, following Ault and Olson (1996), we formalized the conservation law for population abundance as dN(a,t ) dN(a,t ) dN(a,t) = — — : — da + — dt = da dt - Zia,t)N(a,t)dt . (4) This partial differential equation expresses popula- tion age structure in terms of the average number of fish by age over time. The term dN/da is the contri- bution to the change in N(a,t) resulting from indi- viduals getting older. Because the variable a is tied stepwise to chronological age, for each time step t, a gets one unit older, so that da/dt = 1. This condition holds for t > 0 and a > 0. Equation 4 requires two conditions on N(a,t ): one initial condition for N(0,t); and a boundary condition in age 0 tied to reproduc- tion for N(0,t). Integration of Equation 4 with a growth function allows efficient estimation of popu- lation biomass and average size in the stock over 3 Ault, J. S. 1998. Tropical coral reef fishery resource decision dynamics. Unpubl. manuscript. time. In the numerical model REEFS, we modified Equation 4 to a stochastic age-independent length- based population dynamics model to simulate effi- ciently the average or ensemble number at a given length for the entire population age structure as N(L) = J*R(t - a)Sia)Q(a)P(L\a}da , (5) tT where R( z-a) = recruitment lagged back to cohort birth date; S(a) = survivorship to age a; Q(a) - sex ratio at age a; and P(L | a) = the conditional probability of a fish being length L given that it is age a. The ensemble average length L at age a is repre- sented by the von Bertalanffy growth function. The conditional probability distribution for length and age was assumed to be bivariate normal. The reported maximum age of fish in the stock tf (equal to a generation), usually obtained from age and growth studies by using either scales or otoliths, allows application of a convenient and consistent method to normalize the annual instantaneous natu- ral mortality rate M to life span. First, we assume that Sit,), the fraction of the initial cohort numbers surviving from recruitment t to the maximum age, can be expressed as Nit,) N(tr) ) = e (6) Then, assuming an unexploited equilibrium, by set- ting the probability of survivorship of recruits to the maximum age to be 5% (i.e. S(tA)=0.05), and letting t be equal to 0, we rearranged Equation 6 in order to provide an estimate of the natural mortality rate (7) Mortality and growth estimation in tropical fishery populations are normally approached from a size-based perspective because of difficulties in ageing fish. Aver- age size can be converted to mean age by making two assumptions: 1) that age a maps directly into, or is a function of, size L(a)\ and 2) that mean length-at-age from the von Bertalanffy equation can be inverted as -In L „ - Lia) K + tn (8) 400 Fishery Bulletin 96(3), 1 998 Numbers-at-length were converted to numbers-at- weight for each species by means of simple allomet- ric relationships. sustainable basis. SPR is a fraction expressed as the ratio of exploited spawning stock biomass in rela- tion to the equilibrium unexploited SSB Assessment of exploitation effects The REEFS model was configured to assess two fish- ery management decision-making endpoints, yield- per-recruit (YPR) and spawning potential ratio (SPR). Fishery management endpoints are relatively robust measures of potential yields and recruitment. As such, they help to focus on biological (size) and fishing (intensity) controls for managing current and future fishery production. Because biomass B(a,t) is the product of numbers-at-age multiplied by weight- at-age, yield in weight Yw from a given species s was calculated as Yw(F,L',t) = F(f)J B(L\a,t)dL = v Lr (9) F(m N(L\a,t)W (Lja,f)dL . L' Yield-per-recruit (YPR) is then calculated by scaling yield to average recruitment from the right-hand side of the above equation. Spawning stock biomass (SSB, in metric tons [t] ) is a measure of the stock’s repro- ductive potential or capacity to produce newborn, ultimately realized at the population level as suc- cessful cohorts or year classes. Spawning stock bio- mass is obtained by integrating over individuals be- tween the minimum size of first maturity, L , and maximum reproductive size (here assumed to be the maximum size L, ) in the stock Li SSB{t)= j B(L\a,t)dL. (10) L The size of first capture, L \ is that regulated by re- gional fishery management. The modeled fishing mortality rate of headboats (and “viewing power” of divers) was assumed to remove (and sight) fish with a “knife-edged selectivity pattern” (see Gulland, 1983) over the range of exploitable sizes JO if L\a < L' { F(t) if L\a>L' ' (11) SPR = BSB exploited B B B u n f. xpl oit e d (12) Resultant estimated SPRs are then compared to the U.S. Federal standards which define 30% SPR as the “overfishing” threshold (Rosenberg et al., 1996). Lin- ear regressions of estimated SPRs for snapper and grouper were made on 1996 average exvessel prices obtained from voluntary Monroe County dealer re- ports (NMFS4). Because sale of jewfish was prohib- ited, a theoretical 1996 jewfish price was estimated as 0.438 of the price of gag grouper on the basis of historical average annual price ratios (1987-90). Management analyses This assessment focuses on 35 reef fish species in 5 families: groupers, Epinephelinae; snappers, Lutjanidae; grunts, Haemulidae; the hogfish, Lachnolaimus maximus , Labridae; and the great barracuda, Sphyraena barracuda, Sphyraenidae. These are the primary targets of the recreational, commercial, and headboat fleets. Population dynam- ics parameters for each of these fish used in the analyses were gleaned from summaries in Claro (1994) and taxa-specific literature (Tables 2 and 3). The hogfish was grouped with snappers for analyti- cal purposes. The assessment used the following 7 steps (Fig. 2): Step 1 : Conduct visual survey for the reef fish com- munity in year t and intercalibrate diver sampling efficiency by species, site, and year. Step 2: Begin management analysis for species s using intercalibrated visual survey data to compute within-year estimates of L and associated 95% confidence intervals from the size and abundance data, by species, integrated over the range of exploitable sizes. Step 3: Compute a statistically independent set of annual mean estimates of L using fishery- dependent headboat data and compare these with visual survey estimates. Spawning potential ratio (SPR) is a contemporane- ous management endpoint that measures the stock’s potential capacity to produce optimum yields on a 4 NMFS. 1996. Fishery Statistics Div., Southeast Fisheries Sci. Center, Natl. Mar. Fish. Serv., NOAA, 75 Virginia Beach Dr., Miami FL 33149-1003. Ault et al.: A multispecies assessment of coral reef fish stocks 401 Figure 2 Flow chart showing the steps used in the multispecies reef fish assessment. See text for details. Step 4: Use the population dynamics parameters of Table 3 to determine parameters of the LBAR (Ault et al., 1996) and the REEFS computer algorithms. Step 5: Use L ( t ) estimate in LBAR to estimate by year fishing mortality rates for the two data sources; i.e. time series of visual and headboat data. Step 6: Use the REEFS model to estimate: 1) ex- pected Lit ) given the reported population dynamics and F parameter values; 2) YPR and assess growth overfishing; and 3) spawning stock biomass (SSB) for the fishery in both exploited and unexploited states (i.e. F= F , and F= 0, respectively) and evaluate SPR to assess for recruitment over- fishing. Step 7: From these results, make specific fishery management recommendations on control strategies of F and L' consistent with eu- metric fishing principles that minimize the potential for overfishing. 402 Fishery Bulletin 96(3), 1998 Results Fishing effort and sampling intensity Trends in nominal fishing effort, measured as the numbers of licensed recreational, commercial, and headboats vessels in Monroe County, show that rec- reational fishing effort has increased sharply since 1965 (Fig. 3). Since 1981, the largest increase has clearly come from the recreational sector and con- tinues to increase whereas commercial and headboat sectors have been relatively stable. In the 18-year (1979-96) visual survey, 4,571 point samples were collected over a variety of bottom types from 72 reefs located throughout the Florida Keys at depths to 21 m (Table 1; Fig. 1). The complete da- tabase contains information on 226 reef fish species in 55 families with 42 biological, habitat, and physi- cal covariates. Table 2 Parameters, definitions, and units for population dynamics variables common to the LBAR and REEFS numerical mod- els used in simulation analysis of Florida Keys reef fish popu- lation dynamics. See Table 3 for parameter values. Parameter Definition Units s Reef fish species (s=l, . . . ,n) a Cohort age class (a=l, . . . ,tx) t r Age of recruitment months Lr Size at recruitment mm L Minimum age of maturity months K Minimum size of maturity mm t' Minimum age of first capture months L' Minimum size of first capture mm h Oldest (largest) age years W Largest (oldest) size mm Ultimate weight kg Ultimate length mm K Brody growth coefficient per year 1 0 Age at which size equals 0 years aWL Scalar coefficient of weight-length function dimensionless P\VL Power coefficient of weight-length function dimensionless W(a,t ) Weight at age a at time t g L(a,t ) Length at age a at time t mm N(a,t) Numbers at age a at time t number of fish M(a,t) Natural mortality rate at age a at time t per year F(a,t) Fishing mortality rate at age a at time t per year S(a ) Survivorship to age a dimensionless Z(t) Total mortality rate in year t dimensionless 0(a) Sex ratio at age a dimensionless B(a,t) Biomass at age a in year t kg YJt) Yield in weight in year t metric tons SSB(t) Spawning stock biomass in year t metric tons SPR(t) Spawning potential ratio in year t percent Average size and mortality Average annual length was estimated for headboat catch statistics (1981-95) and for visual survey data (1979-96). Headboat data were used in the compara- tive analysis with the visual survey data because they provide consistent catch statistics and effort data. Typical comparisons of average length in the exploit- able phase of the stock for the two data sets are shown for eight representative, economically important reef fishes: black grouper, red grouper, gray snapper, yel- lowtail snapper, white grunt, bluestriped grunt, hog- fish, and great barracuda (Fig. 4). The 95% confi- dence intervals were computed for the visual esti- mates but could not be determined for the headboat data at this time owing to the survey estimation pro- cedures used to calculate total numbers and total weight for the entire Florida Keys. In a few instances (e.g. 1985 and 1986), the computed confidence bounds were large owing to low sample sizes, but these mean estimates still correlated well with the rest of the data. The estimated average lengths in the exploitable phase from the two independent data sources were highly correlated for groupers, snappers, and grunts (Fig. 4). The trend in average size also was relatively flat over the last 18 years and close_to L' (Fig. 4). Although the relation between L for visual and headboat data was similar for all groupers, snappers, and grunts, it differed somewhat for hogfish and barracuda (Fig. 4, G and H). Average length ( L ) for hogfish was consistently smaller in visual samples than in headboat landings. In both data sets, how- ever, L declined in the early 1980s but has steadily increased since the late-1980s (F=3.96, df=9, PcO.OOOl) (Fig. 4G). Average length for barracuda increased significantly (F= 2.2, df=10, P<0.018) in visual surveys but declined in headboat landings beginning in the early 1980s (Fig. 4H). Increased mean barracuda size in visual samples indicates that there has been a corx^sponding increase in abun- dance because larger L requires increased survival. In visual samples, barracuda are now the top ranked species in biomass among all Florida Keys reef fishes. Ault et al.: A multispecies assessment of coral reef fish stocks 403 Tabie 3 Florida Keys reef fish population dynamics parameters for 46 species used in mortality estimations and fishery simulations. Population dynamics parameter definitions and units are given in Table 2. The symbol * indicates that the species is present in recreational catch but not headboat catches or the visual survey. The dash ( — ) indicates that insufficient population dynamic data were available to conduct a management analysis. Complete parameter sets were available for 35 species. Population parameters Species groups M tX L„ K ^0 tm L' t' awi Ki L, Groupers (n = 18) Black Grouper Mycteroperca bonaci 0.150 20 1200.0 31.6 0.160 -0.300 48 508.0 39 4.27E-06 3.2051 1153.1 Coney Epinephelus fulvus 0.180 17 698.9 1.5 0.145 -1.080 13 203.2 19 7.29E-05 2.5700 332.5 Gag Grouper Mycteroperca microlepis 0.200 13 1187.2 25.1 0.149 -0.802 60 508.0 36 1.21E-05 3.0305 1034.4 Graysby Epinephelus cruentatus 0.200 15 415.0 1.1 0.130 -0.940 36 203.2 52 1.22E-05 3.0439 362.5 Jewfish Epinephelus itajara 0.081 37 2394.0 244.9 0.054 -3.616 72 508.0 68 2.09E-05 2.9797 2328.0 Marbled Grouper * Epinephelus inermis — — — — — — — — — — — — Misty Grouper * Epinephelus mystacinus — — — — — — — — — — — — Nassau Epinephelus striatus 0.180 17 698.9 5.9 0.145 -1.080 83 508.0 95 3.83E-06 3.2292 648.2 Red Grouper Epinephelus morio 0.180 17 938.0 11.9 0.153 -0.099 48 508.0 61 1.13E-05 3.0350 869.0 Red Hind Epinephelus guttatus 0.180 17 392.7 1.1 0.207 -0.831 49 203.2 33 1.80E-04 2.6140 382.9 Rock Hind Epinephelus adscensionis 0.250 12 486.1 2.3 0.191 -2.160 48 203.2 9 6.00E-06 3.1930 453.3 Scamp Mycteroperca phenax 0.143 21 999.7 19.3 0.126 -1.357 48 508.0 52 2.02E-05 2.9932 932.2 Snowy Grouper Epinephelus niveatus 0.130 15 1091.3 19.5 0.113 -0.915 48 508.0 57 2.45E-05 2.9300 909.0 Speckled Hind Epinephelus drummondhayi 0.200 15 967.0 16.6 0.130 -1.010 48 508.0 58 1.11E-05 3.0730 861.0 Warsaw Grouper Epinephelus nigritus 0.080 41 2394.0 244.9 0.054 -3.616 48 508.0 68 2.09E-05 2.9797 2328.0 Yellowedge Grouper Epinephelus flavolimbatus 0.180 15 860.0 15.7 0.170 0.000 67 508.0 64 2.82E-05 2.9800 960.0 Yellowfin Grouper Mycteroperca venenosa 0.180 15 860.0 15.7 0.170 0.000 67 508.0 64 2.82E-05 2.9800 960.0 Yellowmouth Grouper Mycteroperca interstitialis Snappers (n = 13) and hogfish (n = l) 0.180 17 881.8 8.6 0.063 -9.030 36 508.0 56 2.58E-05 2.8937 710.7 Black Snapper Apsilus dentatus 0.300 10 618.3 3.2 0.097 -1.728 29 203.2 30 4.52E-05 2.8146 418.4 Blackfm Snapper Lutjanus buccanella 0.230 9 729.7 2.4 0.084 -2.896 20 304.8 43 7.40E-06 2.9735 458.8 Cubera Snapper Lutjanus cyanopterus 0.150 20 1200.0 34.9 0.160 -0.300 28 304.8 19 1.32E-05 3.0601 910.0 Dog Snapper Lutjanus jocu 0.333 9 854.0 10.2 0.100 -2.000 28 304.8 30 4.28E-05 2.8574 790.0 Gray Snapper Lutjanus griseus 0.300 10 722.3 5.2 0.136 -0.863 24 254.0 29 3.05E-05 2.8809 556.2 Lane Snapper Lutjanis synagris 0.300 10 618.3 3.2 0.097 -1.728 29 203.2 30 4.52E-05 2.8146 418.4 continued 404 Fishery Bulletin 96(3), 1998 Table 3 (continued) Population parameters Species groups M h K *0 L V t' awi Pwl L, Mahogony Snapper Lutjanus mahogoni 0.300 10 618.3 3.2 0.097 -1.728 29 304.8 64 8.18E-05 2.7190 418.4 Mutton Snapper Lutjanus analis 0.214 14 938.7 14.1 0.129 -0.738 24 304.8 29 1.57E-05 3.0112 797.8 Red Snapper Lutjanus campechanus 0.190 16 975.0 13.7 0.162 -0.010 28 508.0 55 2.04E-05 2.953 955.0 Schoolmaster Lutjanus apodus 0.250 12 570.0 3.3 0.180 0.000 20 254.0 40 2.04E-05 2.9779 503.8 Silk Snapper Lutjanus vivanus 0.230 9 781.1 9.3 0.092 -2.309 37 304.8 38 1.00E-05 3.1000 512.0 Vermillion Snapper Rhomboplites aurorubens 0.230 10 613.6 2.8 0.206 0.111 43 254.0 33 1.72E-05 2.9456 541.6 Yellowtail Snapper Lutjanus chrysurus 0.214 14 454.7 1.3 0.209 -0.712 24 304.8 56 7.75E-05 2.7180 433.4 Hogfish Lachnolaimus maximus Grunts (n=13) and barracuda (n=l) 0.250 12 566.0 3.8 0.190 -0.776 18 203.2 20 2.55E-05 2.9700 439.0 Black Margate Anisotremus surinamensis — — — — — — 33 203.2 — 2.39E-06 3.3916 — Bluestriped Grunt Haemulon sciurus 0.500 6 289.6 0.47 0.484 -0.011 12 203.2 31 1.94E-05 2.9996 273.5 Caesar Grunt Haemulon carbonarium — — — — — — 27 203.2 — 1.29E-05 3.0559 — Cottonwick Haemulon melanurum — — — — — — 27 203.2 — 2.52E-05 2.9527 — French Grunt Haemulon flavolineatum — — — — — — 18 203.2 — 9.06E-06 3.1581 — Margate Haemulon album 0.374 8 752.6 8.57 0.174 -0.450 34 203.2 17 1.52E-05 3.0423 578.4 Porkfish Anisotremus virginicus — — — — — — 25 203.2 — 1.01E-05 3.1674 — Sailors Choice Haemulon parrai 0.428 7 400.2 1.24 0.220 -0.355 12 203.2 35 2.02E-05 2.9932 320.1 Smallmouth Grunt Haemulon chrysargyreum — — — — — — 24 203.2 — 2.77E-03 2.1567 — Spanish Grunt Haemulon macrostomum — — — — — — 39 203.2 — 2.28E-05 3.0295 — Striped Grunt Haemulon striatum — — — — — — 21 203.2 — 1.39E-05 3.0988 — Tomtate Haemulon aurolineatum 0.333 9 441.6 1.89 0.091 -2.095 24 203.2 57 6.19E-06 3.2077 279.9 White Grunt Haemulon plumieri 0.375 8 511.9 3.06 0.186 -0.776 18 203.2 24 8.35E-06 3.1612 410.3 Great Barracuda Sphyraena barracuda 0.200 15 1238.3 14.03 0.172 -0.461 36 619.2 44 4.11E-06 3.0825 1151.5 Management analyses We also conducted a multispecies stock assessment and management analysis with the estimates of fish- ing mortality to examine if current exploitation lev- els are commensurate with sustainable fisheries. Although 46 exploited reef fish species had been seen or captured in the visual and headboat surveys, only 35 species had population dynamics parameter sets sufficient to conduct a management analysis (Table 3). We noted striking similarities in key relations within taxa as shown by the somewhat discrete clusters of taxa when maximum size dependent on maximum age for a variety of species was plotted (Fig. 5). Mean F estimates for visual survey and headboat data were used to encompass conservatively the range of fea- sible fishing mortality rates experienced in the fish- ery over the last two decades. A comparison of meth- ods and data sources also allowed us to consider risks associated with the overall uncertainty bounds for each stock assessment. Results of an example assess- ment analysis is shown for gray snapper in which Ault et al.: A multispecies assessment of coral reef fish stocks 405 Year Figure 3 Time series of three types of nominal fishing effort directed at Florida Keys reef fish from 1965 to 1993 based on numbers of recreational ( ) and com- mercial (sfc) vessels registered in Monroe County, Florida. Headboats ( + ) are considered to be a subset of the commercial fishing fleet. management endpoints like yield-per-recruit (YPR) and spawning potential ratio (SPR) are provided (Fig. 6). These results are typical of the management in- formation provided for each species. Our analyses indicated that the average size in the exploitable phase for many economically impor- tant reef fish populations was marginally above the minimum size of capture regulated by fishery man- agement agencies (i.e. South Atlantic Fishery Man- agement Council and the Florida Marine Fisheries Commission). To assess the impact of these mortal- ity rates on the stock production, we performed an analytical yield analysis to estimate the YPR for each stock, and evaluated the current YPR status in rela- tion to “eumetric” (cf. Beverton and Holt, 1957) or well-balanced fishing in terms of the minimum size of fish captured and level of fishing effort (Fig. 6A). An issue of paramount concern in assessing analyti- cal yield models are the settings of the two dynamic control variables, F and L' (Fig. 6A). From the per- spective of optimal decision making, any transition of the fishery over the yield surface should minimize negative risks, while maximizing the economic, eco- logical, social, and aesthetic aspects. The YPR analy- sis for gray snapper suggests that the current size (or age) of first capture, L' at 254 mm ( = 10 in), likely results in “growth overfishing.” To maximize the YPR for the range of F operating in the fishery for the last two decades, L' should be increased to greater than 350 mm (Fig. 6B). Figure 6C shows progressive reductions of gray snapper stock biomass when F is increased to the estimate’s lower bound shown in Figure 6B. The like- lihood range of F (i.e. F=[0. 5,1.1] ) is likely also con- tributing to a very low SPR for gray snapper. In fact, it is below the Federal minimum of 30% SPR (Fig. 6D). The minimum estimated F still reduces SPR to about 29% of the unexploited state, whereas the up- per bound estimate results in 15% SPR. The major- ity of the range of estimated F suggest that the gray snapper stock is also “recruitment overfished” as re- flected by reduced spawning potential. The summary of the SPRs for Florida Keys reef fish (Fig. 7) shows that a total of 13 of 16 groupers, 7 of 13 snappers, and 2 of 5 grunts, for which there are data, are below the SPR that constitutes overfishing by Federal definitions. Overall, 63% of the 35 stocks that could be analyzed were overfished. Linear re- gressions of SPR on exvessel price showed a signifi- cantly negative (F1>14=7.55, P=0.0157) slope for grou- per and a marginally significant (F1 n=4.77, P= 0.0514) negative slope for snapper (Fig. 8). Discussion Fishery dynamics Our results indicate that Florida Keys reef fish popu- lations have been heavily fished for at least the last 406 Fishery Bulletin 96(3), 1998 c Yellowtail snapper — i — • — i — i — 1 — i — i — i — i — 1 — “ t 1 1 1 1 1 1 1 1 1 1 1 h— u t — i— i — i — i — i — i — i — i — i— i — i — i — i i i — i — i — i — i — i — i — i — i— 1975 1980 1985 1990 1995 2000 Figure 4 Average length (cm) in the exploitable phase of the stock estimated from headboat (•) and visual survey (0) data for several reef fish stocks during the period 1979 to 1996: (A) red grouper, (B) black grouper, (C) yellowtail snapper, (D) gray snapper, (E) white grunt, (F) bluestriped grunt, (G) hogfish, and (H) great barracuda. Bars around visual survey estimates are 95% confidence intervals of the mean. Horizontal dotted lines show the minimum size at first capture L' regulated by regional fishery management. two decades. Total fishing effort has increased sub- stantially because of greater average fishing power per vessel and a much larger recreational fishery. Mace (1997) estimated that the average “fishing power” per vessel (i.e. the average proportion of the stock removed per unit of fishing effort) has increased 4-fold over the previous 25 years mainly because of improved technology involving better vessel designs, hydroacoustics, hydraulics, and navigation (GPS, Loran C, charts). The arithmetic increase of recre- ational fishing vessels is an important factor also, although its absolute effect on reef fish stocks is un- known because the recreational fleet is distributed diffusely and heterogeneously and has not been well sampled to date. Stock assessment indicator variable The estimated average lengths (size) of reef fish in the exploitable phase, determined from statistically independent visual and headboat data, were highly comparable for groupers, snappers, and grunts, sup- porting their use in the multispecies assessment. Average sizes of hogfish and barracuda, however, differed between the two data sets. The larger aver- age hogfish size in headboat samples appears to be the result of life history patterns and different re- sponses to fishing gears with depth. Hogfish tend to move from shallow to deeper water with age (Davis, 1976) and are more vulnerable to spearfishing than hook-and-line gear. Divers, however, are effectively Ault et a I.: A multispecies assessment of coral reef fish stocks 407 CD .C X 0> c QJ CD CD CB ffl > < Year Year Figure 4 (continued} restricted to shallower depths for safety reasons. Thus, large hogfish that inhabit depths below safe diving limits are available to only the headboat fleet. In shallow water, divers are more likely to see small hogfish because they are more abundant and because large hogfish are more likely to be selectively depleted by spearfishing. The increased average size in both data sets since the mid-1980s (Fig. 4G) is most likely the result of increased spearfishing regulation, im- posed recreational bag limits, and initiation of mini- mum size limits. The average size of barracuda diverged in visual samples and headboat landings since the early 1980s (Fig. 4H). This pattern is likely the result of the pro- motion and expansion of catch-and-release fishing for barracuda during that time period. Decreased fishing mortality resulting from more released bar- racuda would increase the average size of fish in the exploitable phase and be detected in visual samples. Headboat landings, however, could trend toward smaller fish because more large barracuda increase the frequency of angler “breakoffs” (i.e. fish biting through lines or leaders) and the proportion of re- leases because only small barracuda are normally retained for human consumption. Large barracuda are avoided because they carry greater risk of ciguatera poisoning, which can result in convulsions and death for humans (de Sylva, 1994). The trend in average size for grouper, snapper, and grunt stocks was relatively flat over the past 18 years and close to the minimum exploitable length (Fig. 4). The flatness is explained by considering expected L from a modeled range of F in an analytical model, given knowledge about current values of F. The slope of L on F was very shallow in the range of the ana- lytical model (Fig. 9), corroborating the empirical estimates in Figure 4. Some stocks appear to have been chronically overfished since the late 1970s. We 408 Fishery Bulletin 96(3), 1998 0 10 20 30 40 Maximum age (years) Figure 5 Relationship between the population dynamic parameters maximum size Lx and maximum age tk for three economically important Florida Keys reef fish taxa: groupers, Epinephelinae (■); snappers, Lutjanidae (A); and grunts, Haemulidae (•). also noted similarities in key relations within vari- ous taxa that separated out into somewhat discrete clusters when maximum size versus maximum age by species is plotted (Fig. 5). This pattern of species clusters suggests that species within the various taxa groupings will likely respond to exploitation in a simi- lar manner. The sensitivity to exploitation is high- est for groupers, followed by snappers, and then grunts. Overfishing and community shifts Despite conservative assumptions, the estimated fishery exploitation rates suggest that many Florida Keys reef fish stocks are overfished according to defi- nitions for U.S. fisheries (Rosenberg et al., 1996) (Figs. 7 and 9). Many desirable grouper and snapper stocks have low spawning potential ratios (SPRs). Inverse relationships between increased fishing ef- fort (particularly by the recreational sector) (Fig. 3) and the long-term decreased average size and stock biomass (e.g. Fig. 60 of the most desirable species (e.g. groupers and snappers) are particular concerns. The Florida Keys reef fishery shows the classic pattern of serial overfishing, in which the more vul- nerable species are progressively depleted (Munro and Williams, 1985; Russ and Alcala, 1989). The long- est-lived, latest-maturing, and lowest mortality (M) stocks [i.e. groupers] are those first to experience sig- nificant declines in population biomass, followed in sequence by intermediate-lived stocks [snappers], and finally by short-lived stocks [grunts] [Fig. 7]). Within families, the inverse relations between the spawning potential ratio and exvessel market price (Fig. 8) are consistent with serial overfishing. As ex- pected, the most valuable snapper and grouper also tend to have the lowest spawning potentials. During the time frame of this study, numerous measures were taken to reduce fishing mortality in state and federal waters. Fish traps were progressively elimi- nated between 1980 and 1992, and numerous bag limits and minimum size limits were imposed. Fish- eries were closed for queen conch ( Strombus gigas), jewfish ( Epinephelus itajara), and Nassau grouper ( E . striatus). These actions are evidence of trends re- ported in this study. Our data suggest that there may have been sub- stantial changes in the composition of the biomass and abundance of the reef fish community over the past several decades. Although many groupers and snappers have declined, apparently in response to growing fishing effort, some grunts have increased in relative abundance. Claro (1991) noted a similar process in the Golfo de Batabano, Cuba, and hypoth- esized that chronic overharvesting of snappers re- sulted in shifts in community composition in favor Ault et al.: A multispecies assessment of coral reef fish stocks 409 A C Figure 6 Example calculation of an analytical yield model management analysis for gray snapper: (A) 3-dimensional yield-per-recruit YPR surface (g) as a function of the fishing mortality rate F (per yr) and size at first capture L' (mm); (B) YPR isopleths in lifetime yield-per-recruit; (C) stock biomass dependent on size for several exploitation levels; and (D) spawning potential ratio (SPR) as a function of F and L'. REEFS model input parameter values are from Table 3. Shaded rectangles in panels A, B, and D show estimated range of F and L'. Darkened rectangle in panel D indicates the range of F estimates that exceed the Federal 30% SPR overfishing definition following Rosenberg et al. (1996). of grunts. Another indication of significant change was the explosive growth of barracuda (Fig. 4H) which may be explained by several factors. First, there is little directed commercial or recreational fish- ing for barracuda as food because of health concerns. Second, growth of catch-and-release fishing by sport anglers and reduced emphasis on spearfishing may have substantially lowered barracuda mortality. Third, other top predators, such as groupers, snap- pers, and sharks, have been intensively fished, there- fore probably lowering competition for food, while, at the same time, barracuda still retain a large and possibly increasing prey base of grunts and other small fishes. Increased abundance and biomass of a top predator like barracuda could be a management concern if barracuda substantially impact reef fish community dynamics. For example, excessive pre- dation on popular sport fishes like snappers could counteract potential reductions in fishing mortality sought by traditional management. An adjustment of minimum sizes of first capture (L1) and fishing mortality rates (F) may mitigate the apparent growth and recruitment overfishing condi- tions in the fishery. This adjustment should be done in a multispecies context to optimize the biotic and fishery potential of the reef fish assemblage. How- ever, traditional management actions alone are un- likely to be sufficient because they can be circum- 410 Fishery Bulletin 96(3), 1998 Figure 7 Estimates of percent spawning potential ratio (SPR%) for 35 species of Florida Keys reef fish comprise groupers, snappers, grunts, hogfish, and great barracuda. Darkened bars indicate stock “overfishing,” and open bars indicate the stock is above the 30% SPR U.S. Federal standard. vented and habitually fail to control fishing effort effectively, particularly in an open access fishery (Waters, 1991; Bohnsack and Ault, 1996). For ex- ample, bycatch mortality and high fishing effort from the expanding fleets can make size limits ineffective. In theory, every fish can be caught once it reaches minimum legal size with the result that insufficient mature adults survive to reproduce. The tradition of open-access management systems coupled with risk- prone management decisions remains a principal obstacle to achieving renewable resource sustain- ability (Rosenberg et al., 1993). Reversing adverse trends in the Keys reef fishery are likely to require other innovative approaches for controlling exploitation rates. Rothschild et al. (1996) recommended that fishery management maintain a systems view of the resources, emphasizing strategy over tactics. With this in mind, we recommend cou- pling traditional management measures with a spa- tial network of areal closures called “no take” ma- rine reserves. Marine reserves provide an ecosystem management strategy for achieving long-term goals of protecting biodiversity while maintaining sustain- able fisheries. The establishment of a network of small (16 to 3,000 ha) no-take reserves in the FKNMS on 1 July 1997 (U.S. Dep. Commerce, 1996) is a first step. A key to the success of this effort is a conscien- tious, continuous assessment program for integrating fishery-independent and fishery-dependent data to evaluate the effectiveness of these reserves (Bohnsack and Ault, 1996). With adaptive management (Walters, 1986), improvements can be implemented over time. Multispecies assessment Our overall goal was to improve the scientific basis for managing tropical multispecies fisheries by pro- viding an efficient, quantitative framework to assess multispecies fisheries. New assessment methods are particularly needed for complex fisheries that have been poorly documented, are not well understood, and face increased exploitation. Traditional single-spe- cies stock assessment methods are at times inappro- priate or inadequate to deal with the dynamics and structure of multispecies assemblages with large numbers of exploited species (Caddy, 1981; Ault and Fox, 1990; Appledorn, 1996). Owing to a lack of data and basic biological information, only a few reef spe- cies in the entire southeastern U.S. have had com- prehensive stock assessments. We emphasized a multispecies ecosystem approach because traditional fishery models have been inef- fectual in creating sustainable fisheries (Ludwig et al., 1993; Sharp, 1995; Caddy, 1996; Russ, 1996). Ault et a I.: A multispecies assessment of coral reef fish stocks 41 I 80 w c m o CL Figure 8 Linear regressions of estimated species spawning potential ratio (y) on average 1996 Monroe County exvessel price (x) for: (A) snapper: y = 119.59 - 41.38.x; and, ( B) grouper: y = 47.35 - 15.03*. Single-species fishery management models built to maximize fishery yield and economic rent ignore critical biological and physical interactions and cumulative stresses on habitats. Reef fish stocks are likely to be regulated by trophic interactions at the in- dividual, population, and community levels. Also, application of “traditional” fishery man- agement models developed for temperate spe- cies to tropical coral reef assemblages is tenu- ous. In response to these problems, the Na- tional Research Council Committee on Fish- eries (1994) recommended developing multi- species ecosystem management programs for building sustainable fisheries. Successful implementation of such programs will require innovative research, new management strat- egies, less destructive and wasteful fishing methods, protection of critical and sensitive habitats, and more effective education. Our retrospective analysis emphasized fish- ery-independent data. Although fishery-inde- pendent assessments can provide reliable measures of fish abundance, population dy- namics, and community composition ( Gunder- son, 1993), their application in multispecies fishery assessments have been limited. We predict an increasing need to rely on such data for assessment purposes because fishery-de- pendent data will become less available and less useful as new regulations are imposed that will establish larger size limits, closed seasons, closed fishing areas, and species pro- hibited from being fished. Also, the shifting emphasis from commercial to recreational fishing makes collecting fishery-dependent data much more difficult and expensive. Results from this 18-year retrospective assess- ment are encouraging in providing estimates of stock status. However, several assumptions that simplify the population dynamics of the various species make it prudent to consider population es- timates first-order approximations. The error intrin- sic in population-rate estimates derived from sur- veys depends on the accuracy and precision of the basic survey. Errors ultimately propagate upwards during a series of calculations used to determine the average size, total mortality rates, fishing mor- tality rates, yield-per-recruit, and finally the spawn- ing potential ratio, which is the current focus of management decisions. Also, although the Florida Keys fishery represents a major fishing area, it does not necessarily represent the entire stock range for an individual species. It is possible that mature stock components exist outside the fishing area. Six actions could improve future assessments. First, develop suitably structured spatial models for linkages between habitat use and fish ontogeny to “fill-in” the map of population estimates for areas not sampled. Second, calibrate the relative statisti- cal power of diver surveys and headboat fishing gear. Ideally, diver observations should relate to what fish- ermen catch. Because the fishing mortality rate of headboats (and viewing power of divers) are consid- ered strictly proportional to average population abun- dance, we must understand the fraction of the stock assessed per unit of effort and the interrelationship of the efficiency of the two “gear” types. Third, in- crease temporal and spatial sampling coverage to increase survey precision and resolution. Fourth, tune fishery-independent surveys with other indices of stock abundance. In this retrospective analyses, no attempt had been made to have the two survey 412 Fishery Bulletin 96(3), 1998 types coincide with respect to sampling effort or lo- cations within the Florida Keys. Fifth, employ new sampling technologies, such as hydroacoustics, green band lasers, and stereo video cameras to improve the accuracy and cost effectiveness of biomass and abun- dance estimates. Sixth, improve basic biological in- formation on growth, reproduction, mortality, feed- ing, and recruitment, which are fundamental ele- ments of stock assessment. The models and conclu- sions presented here are strongly influenced by the accuracy of the parameter estimates and the source for these estimates is not always reliable. Conclusions We used a new approach involving fishery-indepen- dent data to conduct a quantitative retrospective multispecies assessment of changes in the Florida Keys multispecies reef fish community. Our results show that fishing effort and mortality levels are very intense, that many stocks are “overfished,” and that exploitation has likely altered the structure and dy- namics of the reef fish community. Inevitable in- creases in fishing effort, particularly by recreational anglers, combined with habitat degradation by rapid growth of human populations in the region, if un- abated, will increase the potential for overfishing and ecosystem changes. Without effective intervention by regional fishery management to bring fishing effort under control, reef fish stocks will likely continue to decline. A spatial network of “no take” marine re- serves, combined with traditional management mea- sures, have the potential to reverse these trends for many species and to allow the long-term goals of building sustainable fisheries and protecting biodi- versity to be achieved. Acknowledgments We thank Guillermo Diaz, Doug Harper, Ken Linde- man, Jiangang Luo, Elizabeth Maddox, and Dave Mc- Clellan for technical assistance, and John Caddy, Ault et al.: A multispecies assessment of coral reef fish stocks 413 Mark Hixon, Daniel Pauly, Joseph E. Powers, Will- iam J. Richards, Joe Serafy, and Steven G. Smith for their critical review of the manuscript. Logistical support for field sampling was provided by the FKNMS, the National Undersea Research Center (NURC), and the National Park Service. This re- search was sponsored by the United States Man and the Biosphere Marine and Coastal Ecosystem Direc- torate Grant 4710-142-L3-B, and the NOAA Coastal Ocean Program Grant NA37RJ0200. Literature cited Appledorn, R. S. 1996. Model and method in reef fishery assessment. In N. V. C. Polunin and C. M. Roberts (eds.). Reef fisheries, p. 219-248. Fish and Fisheries Series 20, Chapman & Hall, London. Ault, J. S., and N. M. Ehrhardt. 1991. Correction to the Beverton and Holt Z-estimator for truncated catch length frequency distributions. ICLARM Fishbyte 9( l):37-39. Ault, J. S., and W. W. Fox Jr. 1989. FINMAN: simulated decision analysis with multiple objectives. In E. F. Edwards and B. A. Megrey (eds.), Mathematical analysis of fish stock dynamics, p. 166- 179. Am. Fish. Soc. Symp. 6. 1990. Simulation of the effects of spawning and recruitment patterns in tropical and subtropical fish stocks on tradi- tional management assessments. Proc. Gulf Carib. Fish. Inst. 39:361-388. Ault, J. S., R. N. McGarvey, B. J. Rothschild, and J. Chavarria. 1996. Stock assessment computer algorithms. In V. F. Gallucci, S. Saila, D. Gustafson, and B. J. Rothschild (eds.), Stock assessment: quantitative methods and applications for small scale fisheries, p. 501-515. Lewis Publishers (Division of CRC Press), Chelsea, MI. Ault, J. S., and D. B. Olson. 1996. A multicohort stock production model. Trans. Am. Fish. Soc. 125(3):343-363. Bell, J. D., G. J. S. Craik, D. A. Pollard, and B. C. Russell. 1985. Estimating length frequency distributions of large reef fish underwater. Coral Reefs 4:41-44. Bellwood, D. R., and A. C. Alcala. 1988. The effect of minimum length specification on visual census estimates of density and biomass of coral reef fishes. Coral Reefs 7:23-27. Beverton, R. J. H., and S. J. Holt. 1956. A review of methods for estimating mortality rates in exploited fish populations, with special reference to sources of bias in catch sampling. Rapp. P.-V. Reun., Cons. Int. Explor. Mer 140:67-83. 1957. On the dynamics of exploited fish populations. Ministry of Agriculture, Fisheries and Food, Fishery In- vestigations (UK), series II, vol. XIX, Lowestoft, England, 533 p. Bohnsack, J. A., and J. S. Ault. 1996. Management strategies to conserve marine biodiver- sity. Oceanography 9( 1 ):73— 82. Bohnsack, J. A., and S. P. Bannerot. 1986. A stationary visual census technique for quantita- tively assessing community structure of coral reef fishes. U.S. Dep. Commer., NOAA Tech. Report NMFS 41, 15 p. Bohnsack, J. A., D. E. Harper, and D. B. McClellan. 1994. Fisheries trends from Monroe County, Florida. Bull. Mar. Sci. 54(3):982-1018. Caddy, J. F. 1981. Use of Beverton and Holt yield tables for prelimi- nary assessment of effects of changes in size at first cap- ture and fishing effort in a mixed species fishery. Annex M in Second technical consultation on stock assessment in the balearic and gulf of lions statistical divisions, p. 141- 149. FAO Fisheries Report 263, FAO, Rome. 1996. Regime shifts and paradigm changes: Is there still a place for equilibrium thinking? Fish. Res. 25:219-230. Claro, R. 1991. Changes in fish assemblages structure by the effect of intense fisheries activity. Trop. Ecol. 32( 1 ):36— 46. 1994. Ecologia de los peces marinos de Cuba. Centro de Investigaciones de Quintana Roo, Mexico, 547 p. Davis, J. C. 1976. Biology of the hogfish, Lachnolaimus maximus (Walbaum), in the Florida Keys. M.S. thesis, Univ. Mi- ami, Coral Gables, FL, 86 p. DeMartini, E. E. 1993. Modeling the potential of fishery reserves for man- aging Pacific coral reef fishes. Fish. Bull. 91:414-427. de Sylva, D. P. 1994. Distribution and ecology of ciguatera fish poisoning in Florida, with emphasis on the Florida Keys. Bull. Mar. Sci. 54(3):944-954 Dixon, R. L., and G. R. Huntsman. 1992. Estimating catches and fishing effort of the south- east United States headboat fleet, 1972-1982. U.S. Dep. Commer., NMFS Tech. Rep., 23 p. and append. Ehrhardt, N. M., and J. S. Ault. 1992. Analysis of two length-based mortality models ap- plied to bounded catch length frequencies. Trans. Am. Fish. Soc. 121( 1 ): 1 15—122. Fausch, K. D., J. Lyons, J. R. Karr, and P. L. Angermeier. 1990. Fish communities as indicators of environmental degradation. Am. Fish. Soc. Symp. 8:123-144. Gulland, J. A. 1983. Fish stock assessment: a manual of basic methods. FAOAViley Series on Food and Agriculture, vol. 1, New York, NY, 223 p. Gunderson, D. R. 1993. Surveys of fisheries resources. John Wiley & Sons, New York, NY, 248 p. Harvey, E., and M. Shortis. 1996. A system for stereo-video measurement of sub-tidal organisms. MTS Journal 29:10-22. Harwell, M. A., J. F. Long, A. M. Bartuska, J. Gentile, C. C. Harwell, V. Myers, and J. C. Ogden. 1996. Ecosystem management to achieve ecological sustain- ability: the case of South Florida. Environ. Manage. 20(41:497-521. IDL (Interactive Data Language). 1995. User’s guide: Interactive Data Language 4.0. Re- search Systems, Inc., Boulder, CO, var. pag. Johnson, R. A., and D. W. Wichern. 1992. Applied multivariate statistical analysis. Prentice Hall, Englewood Cliffs, NJ, 642 p. Lee, T. M., C. Rooth, E. Williams, M. McGowan, A. Szmant, and M. E. Clarke. 1992. Influence of Florida Current, gyres and wind-driven 414 Fishery Bulletin 96(3), 1998 circulation on transport of larvae and recruitment in the Florida Keys coral reefs. Cont. Shelf Res. 12:971-1002. Ludwig, D., R. Hilborn, and C. J. Walters. 1993. Uncertainty, resource exploitation, and conservation: lessons from history. Science (Wash., D.C.) 260:17 and 36. Mace, P. 1997. Developing and sustaining world fishery resources: state of science and management. In D. A. Hancock, D. C. Smith, A. Grant, and J. P. Beumer (eds.), Developing and sustaining world fisheries resources: the state of sci- ence and management, p. 1-20. Proc. Second World Fish- ery Congress, Brisbane, Australia. Munro, J. L., and D. M. Williams. 1985. Assessment and management of coral reef fisheries: biological, environmental, and socio-economic aspects. Proc. Fifth Int. Coral Reef Congress, Tahiti 4:544-578. National Research Council (U.S.) Committee on Fisheries. 1994. Improving the management of U.S. marine fisheries. National Academy Press, Washington, DC, 62 p. Neter, J., M. H. Kutner, C. J. Nachtsheim, and W. Wasserman. 1996. Applied linear statistical models, 3rd ed. Richard D. Irwin, Homewood, IL, 1408 p. NMFS. 1996. Our living oceans: report of the status of U.S. Living marine resources, 1995. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-F/SPO-19, 160 p. Polunin, N. V. C., and C. M. Roberts (eds). 1996. Reef fisheries. Fish and Fisheries Series 20. Chap- man & Hall, New York, NY, 477 p. Ricker, W. E. 1963. Big effects from small causes: two examples from fish population dynamics. J. Fish. Res. Board Can. 20:257-264. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 191, 382 p. Robson, D. S. 1966. Estimation of relative fishing power of individual ships. International Commission for the Northwest At- lantic Fisheries (ICNAF) Res. Bull. 3:5-14. Rosenberg, A. A., M. J. Fogarty, M. P. Sissenwine, J. R. Beddington, and J. G. Shepherd. 1993. Achieving sustainable use of renewable resources. Science (Wash., D.C.) 262:828-829. Rosenberg, A (convener), P. Mace, G. Thompson, G. Darcy, W. Clark, J. Collie, W. Gabriel, P. Goodyear, A. MacCall, R. Methot, J. Powers, V. Restrepo, T. Wainwright, L. Botsford, J. Hoenig, and K. Stokes. 1996. Scientific review of definitions of overfishing in U.S. fishery management plans. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-F/SPO-17, 205 p. Rothschild, B. J., J. S. Ault, and S. G. Smith. 1996. A systems science approach to fisheries stock assess- ment and management. In V. F. Gallucci, S. Saila, D. Gustafson, and B. J. Rothschild (eds.), Stock assessment: quantitative methods and applications for small scale fish- eries, p. 473-492. Lewis Publishers (Division of CRC Press), Chelsea, MI. Russ, G. R. 1996. Fisheries management: What chance on coral reefs? ICLARM 19(3): 5-9. Russ, G. R., and A. C. Alcala. 1989. Effects of intense fishing pressure on an assemblage of coral reef fishes. Mar. Ecol. Prog. Ser. 56:13-27. Sale, P. F. (ed.). 1991. The ecology of fishes on coral reefs. Academic Press, San Diego, CA, 754 p. Sharp, G. D. 1995. It’s about time: new beginnings and old ideas in fish- eries science. Fisheries Oceanography 4(4):324-341. Sokal, R. R., and F. J. Rohlf. 1969. Biometry: the principles and practice of statistics in bio- logical research. W. H. Freeman, San Francisco, CA, 776 p. Sparre, P., and S. C. Venema. 1992. Introduction to tropical fish stock assessment. FAO Fish. Tech. Paper 306/1 (rev. 1), FAO, Rome, 376 p. St. John, J., G. R. Russ, and W. Gladstone. 1990. Accuracy and bias of visual estimates of numbers, size structure, and biomass of coral reef fish. Mar. Ecol. Prog. Ser. 64:253-262. U.S. Department of Commerce. 1996. Florida Keys National Marine Sanctuary final manage- ment plan/environmental impact statement (FMP/EIS), vol. 1: the management plan. NOAA, Washington, D.C., 319 p. Venables, W. N., and B. D. Ripley. 1994. Modern applied statistics with s-plus. Springer- Verlag, New York, NY, 462 p. Walters, C. J. 1986. Adaptive management of renewable resources. Inter- national Institute for Applied Systems Analysis. Macmil- lan, New York, NY, 374 p. Waters, J. R. 1991. Restricted access vs. open access methods of man- agement: toward more effective regulation of fishing effort. Mar. Fish. Rev. 53(3):1-10. 415 Abstract .—We determined sex ra- tio, spawning season, batch fecundity, spawning frequency, and annual fecun- dity for gag (Serranidae: Mycteroperca microlepis), a protogynous hermaphro- dite. Gag were randomly sampled from 1991 through 1993 (n=1398) and selec- tively sampled with a bias towards heavily pigmented fish in 1994 (n=648). All samples were taken from commer- cial and recreational fisheries in the northeastern Gulf of Mexico. Sex ratio was 49 females: 1 male for the 1991-93 samples. Heavily pigmented gag (n= 62), commonly thought to be males, were col- lected from waters >41.0 m during all months except January and March. Of these, 60 were histologically sexed and found to be 5.0% female, 3.3% possibly early transitional male, and 91.7% male. Indeterminate spawning gener- ally occurred from February through April in water depths >30.5 m. Hy- drated oocytes were homogeneously distributed in hydrated ovaries. Short- est length and youngest age at spawn- ing were 577 mm total length (TL) and 3 years for females, and 981 mm TL and 8 years for males. The highest batch fecundity (865,295 hydrated oocytes) occurred in an 8-year-old, 1038-mm-TL gag. Batch fecundity had a significant (a=0.01) positive correlation with TL, gutted body weight, and age but was most strongly correlated with TL. Aver- age annual spawning frequency ranged from 8 to 27 for 2- to 9-year-old fish, vary- ing significantly by year (P<0.05) and among years and age classes (P< 0.02). Annual fecundity estimates ranged from 0.065 to 61.4 million and corre- lated well with age. Manuscript accepted 25 November 1997. Fishery Bulletin 96:415-427 (1998). Reproductive patterns, sex ratio, and fecundity in gag, Mycteroperca microlepis (Serranidae), a protogynous grouper from the northeastern Gulf of Mexico L. Alan Collins All yn G. Johnson Christopher C. Koenig M. Scott Baker Jr. Panama City Laboratory National Marine Fisheries Service, NOAA 3500 Delwood Beach Road, Panama City, Florida 32408 E-mail address (for L A. Collins): acollins@nmfspc.ssp.nmfs.gov The spawning potential ratio (SPR), which is used to estimate the adult stock size as a percentage of the unfished stock, is an important as- sessment tool in the management of fish stocks (Goodyear, 1993). An accurate determination of SPR re- quires a basic understanding of re- production (Huntsman and Schaaf, 1994), including information on sex ratio, spawning season duration, and size (and age) at first spawn- ing, as well as estimates of batch fecundity, spawning frequency, and annual fecundity by fish size. Gag (Serranidae: Mycteroperca microlepis), a large, economically important, shallow-water grouper found from Massachusetts to Rio de Janeiro, Brazil, (Briggs, 1958) is heavily fished in the Gulf of Mexico on Florida’s western shelf, with 1986-92 annual commercial land- ings of about 6.80 x 105 kg (1.5 mil- lion lb) and recreational landings of 0.2-0. 6 million fish (Gulf of Mexico Fishery Management Council1 ; Schirripa and Goodyear2 ). The gag is a winter-spring spawning, mon- andric protogynous hermaphrodite that matures sexually around 550 mm total length and 3 to 6 yr (McErlean and Smith, 1964; Collins et ah, 1987; Bullock and Smith, 1991; Hood and Schlieder, 1992; Coleman et al., 1996; Koenig et al., 1996). Gilmore and Jones (1992), who described the reproductive be- havior and color variation of gag off the east coast of Florida, assumed that large gag with heavy pigmen- tation (“black-belly” and “black- back”) were males. Similarly, Bul- lock and Smith ( 1991) assumed that large gag “with dark pigmentation ventrally” from the Gulf of Mexico were male. More recently, Koenig et ah (1996) and Coleman et ah (1996) have described reproductive styles of gag and the apparent effects of fish- ing on gag spawning aggregations. Although these studies (above) have examined some aspects of gag reproduction, few have addressed 1 Gulf of Mexico Fishery Management Council. 1989. Amendment number 1 to the reef fish fishery management plan. GMFMC, Tampa, Florida, 356 p. 2 Schirripa, M. J., and C. P. Goodyear. 1994. Status of the gag stocks of the Gulf of Mexico: assessment 1.0. Miami Labo- ratory, Natl. Mar. Fish. Serv., NOAA, 75 Virginia Beach Drive, Miami, FL 33149. 416 Fishery Bulletin 96(3), 1998 either fecundity or spawning frequency, and none have used the methods developed by Hunter et al. (1985) for multiple-spawning fishes. Hunter’s meth- ods have been used successfully to estimate batch fecundity, spawning frequency, and annual fecundity in other Gulf of Mexico fishes similar to gag in size and longevity (e.g. Fitzhugh et al., 1988; Taylor and Murphy, 1992; Render and Wilson, 1992; Fitzhugh et al., 1993; Nieland and Wilson, 1993; Wilson and Nieland, 1994; and Collins et al., 1996). Such esti- mates obtained with these newer methods are far more accurate than the previously used “ovarian egg number” methods for multiple spawning fishes (as described in Hunter et al., 1985). In this study, we estimate sex ratio, spawning sea- son duration, length and age at first spawning, depth of spawning, batch fecundity, spawning frequency by year, spawning frequency by size and age, and an- nual fecundity in gag from the Gulf of Mexico. We also test the assumption that heavily pigmented gag represent only males using histological examination of gonads of heavily pigmented fish. Methods Gag were randomly sampled from commercial and recreational landings from Panama City to St. Pe- tersburg, Florida, in 1991, 1992, and 1993. In 1994, we asked commercial fishermen to leave all heavily pigmented gag ungutted so that they could be exam- ined for sex. Gag were considered heavily pigmented only if they retained the heavy-black pigment on the ventral portion of the abdomen after death (e.g. plate XVI, Fig. D., page 237, in Bullock and Smith, 1991). Total length (TL, in mm), fork length (FL, in mm), and total (ungutted) weight (TW, in g) were recorded before removing gonads, which were kept on ice prior to examination. Gonads of reproductively active gag are so large that they require subsampling for determination of sex. To ensure that a single tissue sample per female would be adequate to estimate maximum oocyte diameter and hydrated oocyte counts, we first tested for tissue ho- mogeneity. To test for homogeneity of oocyte diameter, three bilobed ovaries with hydrated oocytes were di- vided into six equal sections (three from each lobe) and subsampled to determine the frequency distributions of oocyte diameters; they were tested for homogeneity with a Kolmogorov-Smimov two-sample test (Sokal and Rohlf, 1981). To test for homogeneity of density of hy- drated oocytes, we used six hydrated ovaries divided in the same manner and determined the number of hydrated oocytes per gram; samples were tested for ho- mogeneity with a two-way AN OVA (SAS, 1988). We made a preliminary determination of sex in gag using fresh, unstained samples. A small sample was removed from each gonad, teased apart, and viewed microscopically (250x). Sex was assigned according to the following criteria: females had oocytes and no tissues that could be mistaken for spermatogenic tis- sue; males had no oocytes (or few small oocyte rem- nants); and transitional males had high numbers of oocytes and possible sperm (Shapiro et al., 1993). Female gonads were then microscopically staged. Maximum oocyte diameter (MOD) was recorded; the most advanced oocyte type present was noted and used to assign individuals to one of five ovarian-matu- ration stages: 1) mature, resting (MOD<0. 12 mm and clear); 2) early developing (MOD<0.20 mm and slightly opaque); 3) late developing (MOD=0.20 to 0.59 mm and opaque); 4) ripe (MOD>0.59 mm and mostly transparent); and 5) spent (much loose par- ticulate matter in flaccid ovary) (West, 1990). After fat and mesentery were removed, gonads were blot- ted dry, weighed to the nearest 0.1 g., and placed in 10% buffered formaldehyde solution. We then selected and prepared some of the gonad samples (above) to verify sex and stage using stan- dard histological techniques. Five-pm-thick sections were prepared from tissues embedded in paraffin and stained with Harris’s hematoxylin and eosin. Ova- rian stages were assigned on the basis of the most advanced oocyte or follicle stage present: 1) primary growth; 2) cortical alveolar; 3) vitellogenic; 4) hy- drated; and 5) spent (presence of postovulatory fol- licles [POFs]). Stages 1-4 followed Wallace and Selman (1981). Histological stages of possible early transitional males and functional males were as fol- lows: 6) possible early transitional males had active female tissue with possible crypts containing primary spermatocytes (Moe, 1969; Johnson, 1995); 7) ma- ture inactive males had a few secondary spermato- cytes and remnant oocytes (primary growth and atretic only); 8) mature active males had many sec- ondary spermatocytes and remnant oocytes; 9) rip- ening males had spermatids in ducts; and 10) ripe males had tailed spermatozoa in ducts. Sex and stage were compared by gonad-region for all heavily pig- mented gag. Physical condition of the gonad (e.g. in- tactness, presence of parasites, possible preservation problems) was also noted. We estimated duration of the spawning season by using two techniques: 1) plotting gonadosomatic in- dex (GSI=gonad weight (1Q0)/TW) by month for all gag with gonads in good condition (>90% intact) and 2) determining when hydrated ovaries with non- degenerated POFs appeared. Proof of actual spawn- ing (and not just the presence of “ripe” ovaries, as discussed in Sadovy et al., 1994) was provided by Collins et al.: Reproductive patterns, sex ratio, and fecundity of Mycteroperca microlepis 417 POFs. Size and age at first spawning, and water depth of spawning, were determined with only his- tological stage-4 females and stage- 10 males. Depth of spawning was indicated by depth of catch, as de- termined from interviews with fishermen. Batch fecundity estimates involved counting whole, hydrated oocytes taken from cross sections of gonads and weighed (to the nearest 0.001 g) fol- lowing Hunter et al. (1985). Samples were exam- ined histologically for the presence of POFs and hydrated oocytes. The POFs in samples were noted as 0 (absent), 1 (degenerating), or 2 (nondegen- erating) (Fitzhugh and Hettler, 1995). Ovaries with nondegenerating POFs were omitted from batch fecundity estimates to avoid biases in determin- ing the number of hydrated oocytes per gram of tissue (Hunter et al., 1985). Batch fecundity esti- mates were calculated by dividing the product of gonad weight and number of hydrated oocytes by the sample weight. Any fish collected late in the spawning season were omitted from batch fecun- dity analyses because of apparent decreases in batch size (e.g. Fitzhugh et al., 1988). Spawning frequency (the number of spawnings per year by a female) was estimated by dividing the duration of the spawning season by the aver- age number of days between spawning for all fe- males (Hunter and Macewicz, 1985; Hunter et al., 1986). Duration of the spawning season was the number of days between the first and last occur- rence of hydrated oocytes or POFs each year. The average number of days between spawning for all females (> or = TL of the smallest hydrated fish) was 100% divided by the percentage frequency of hydrated females. For example, if the spawning season was 100 days long and females spawned every two days (with 50% hydrated), then spawn- ing frequency would be 50. We obtained annual fecundity estimates by multi- plying batch fecundity by spawning frequency. Age- dependent spawning frequency was estimated with stage-4 females within each age group. Because spawning frequency varied by age and year, these variables were considered separately in calculating annual fecundity. We analyzed the following relation- ships using regression: batch fecundity (BF) x TL, BF x age, and BF x gutted body weight ( GBW=total weight — gonad weight = TW - GW); spawning fre- quency x age; annual fecundity (AF) x TL, AF x age, and AFxGBW. Most fish were aged by using whole or sectioned sagittal otoliths (Johnson et al., 1993). Whole otoliths were adequate for ageing smaller fish (<900 mm TL). We sectioned otoliths for larger gag (>899 mm TL) to facilitate reading outer rings. Figure 1 Area sampled for gag from commercial and recreational catches, 1991-94. Total area sampled, along the 40-80 m contour, is encompassed by brackets; most samples were col- lected from gag caught in the area south of Panama City to the Florida Middle Ground (the latter is indicated by the box). Results Of all the gag collected from February 1991 through December 1994 (rc=2,046), most (61.7%) were taken from February through May between Panama City and the Florida Middle Ground (Fig. 1). A few samples were collected from off Pensacola, St. Pe- tersburg, and Ft. Myers, FL. The majority of com- mercial and recreational landings were <24 h old when sampled. Depth data recorded with samples were predominately from commercial catches (57.2%). Commercial catches were usually from greater water depths (mean=57.2 m, n- 697 ) than those for rec- reational catches (mean=32.6 m, n= 73). Most 1991 samples (n= 463) were from the commercial fishery (78.1%). Most 1992 samples (n=318) were from the rec- reational fishery (85.5%). Samples in 1993 (z? =617 ) came from each of the fisheries in approximately equal numbers (commercial =49%). Most 1994 samples (n=648) were from the commercial fishery (68.6%). Because neither oocyte diameter-frequency distribu- tions for oocytes >0.08 mm in diameter (a=0.05; Kolmogorov-Smirnov two-sample test; dQ 05=0.1103; 4 ! 8 Fishery Bulletin 96(3), 1998 Table 1 Effect of location of gag tissue samples on hydrated oocyte counts per unit of weight (g). Locations are anterior (1), mid (2), and posterior (3) of ovarian lobes. Analysis of variance indicates significance of location within a lobe for number of hydrated oocytes per gram of tissue. SS = sum of squares; MS = mean square. Positions of sample in ovary Lobe 1 Lobe 2 Both lobes X SD n X SD n X SD n 1 918 92 6 951 107 6 934 97 12 2 914 81 6 932 180 6 923 133 12 3 875 140 6 931 144 6 903 138 12 Total 902 103 18 938 138 18 920 121 (36) Source df SS MS F P>F Lobe 1 11,460.23 11,460.23 0.74 0.40 Region 1 6018.21 6018.21 0.39 0.54 Interaction 1 803.81 803.81 0.05 0.82 Error 32 496,556.28 15,517.38 Total 35 514,838.52 n= 5,450 oocytes; range=0.04 to 1.20 mm) nor hydrated oocyte counts per gram (Table 1) differed significantly among ovarian regions, we randomly sampled one re- gion on each ovary to determine maturation stage, maximum oocyte diameter, and fecundity. Gag hydrated ovaries contained a mean of 920 hydrated oocytes per gram of ovarian tissue (range=875 to 951, SD=121, n= 6 fish) (Table 1). The distinct clusters of oocytes of vari- ous diameters, including stage-1 and stage-2 oocytes, in gag ovaries throughout the spawning season indi- cated that gag is a multiple and indeterminate spawner. Gag ovaries were usually much larger than testes (mean ovary weight=46.5 g, mean testes weight= 26.7 g). The largest ovary and testes weighed 1.60 kg and 120.0 g, respectively. Wormlike parasites were common in ovaries. Histological appearance of gag ovaries (non- transitional) was Typical for marine teleosts (Wallace and Selman, 1981), except that some functional females appeared to contain crypts of spermatogonia. An unusual condition observed in three females (stage 1 or 2) was the presence of a 5-mm thick lin- ing (around the inside perimeter of the abdominal cavity) of compressed, hydrated oocytes apparently caused by rupture of the ovarian wall. These three fish were caught after the peak spawning season (May 1994, June 1993, and July 1994). We found no such lining in any of the other specimens examined. We found that sex determinations and maturation- stage determinations were made with variable accu- racy in fresh and histologically prepared samples. The two examinations gave the same results 89% of the time for sex determination but only 70% of the time for stage determination. The main reason for this poor comparison of the two staging techniques was that we could not accurately identify spent fe- males (with POFs) or transitional males from fresh samples, because POFs and sperm could only be posi- tively identified histologically. Sex ratio Sex ratio from all random samples (pooling data for 1991-93) was approximately 49 females: 1 male. Our emphasis on collecting gag with dark ventral pig- mentation in 1994 clearly increased the overall per- centage of males: 52.5% of all males for 1991-94 (n= 59) were collected in 1994. Percentage of males was 1.9% in 1991-93 and 4.7% in 1994. Females were collected year round and males were collected in all months but January. Possible early transitional male gag (n= 3) were found only during February, April, and May. Ventral pigmentation Gag with heavy ventral pigmentation (n-6 2) were collected in all months but January and March in waters >41 m (Table 2). Sex determination (h=60) showed that 5% were females (Fig. 2A), 3.3% were pos- sible transitional males (Fig. 2, B and C), and 91.7% were males (Fig. 2D). Sex and stage were homogeneous among six gonad regions in all heavily pigmented fish. Spawning period Winter-spring spawning, indicated by the presence of hydrated oocytes, was corroborated by GSI (Fig. Collins et al.: Reproductive patterns, sex ratio, and fecundity of Mycteroperca microlepis 419 Table 2 Results of histology on gonads of gag observed with heavy ventral pigmentation, 1991-94 (M = male, F = female, T = possible early transitional male). Stage (see “Methods” section): 1 = primary growth; 3 = vitellogenic; 4 = hydrated; 6 = possible sperm crypts; 7 = inactive; 8 = active; 9 = ripening; and 10 = ripe. C = commercial and R = recreational. Numbers in parentheses indicate number aged. Year and month n Total length (mm) Age (yr) Gonad wt.(g) Catch depth (m) Fishery Sex Stage 1991 Feb i 985 — 31.2 - C T 6 i 886 — 15.4 45.7-54.9 C M 8 May 2 1140,1150 12,14 47.2,69.4 — C M 10 Sep 1 1140 — 33.1 — R M 10 Dec 1 1170 — 28.9 — C M 10 Total 6 886-1170 12,14 (2) 15.4-69.4 45.7-54.9 C,R T,M 6-10 1992 Jun 2 1100,1162 8,11 19.2,22.4 — R M 8-10 Oct 1 1276 20 8.9 39.6-J R M 9 Dec 6 965-1143 11(1) 24.7-58.6 57.9-68.6 C M 7-10 Total 9 965-1276 8-20 8.9-58.6 57.9-68.6 C,R M 7-10 1993 Feb 1 1170 17 84.5 54.9-76.2 C M 10 1 858 7 54.6 — R F 3 Jun 1 1030 9 55.2 — R F 1 Jul 6 1070-1290 10-20 12.8-25.7 57.9-59.4 C,R M 7-10 Aug 1 1105 11 13.9 54.9-67.1 C M 9 Oct 2 1065,1095 13,- 13.6,17.9 115.8 C M 9 Total 12 858-1290 7-20 12.8-84.5 54.9-115.8 C,R M,F,T 1-10 1994 Apr 8 980-1201 7-12 35.7-49.9 67.1-109.7 C M 10 1 1120 12 267.7 67.1 C T 6 1 1090 9 152.6 67.1 C F 4 May 11 1055-1240 8-22 6.4-120.0 53.3-106.7 C M 8-10 Jul 4 1130-1210 11-20 12.8-22.7 54.9-99.1 C M 9-10 Aug 4 1085-1235 — 18.2-27.8 79.2 C M 9-10 Sep 1 1175 — 40.9 50.3 c M 10 Oct 1 1242 — 30.9 79.2 R M 10 Nov 1 1230 — 21.8 — C M 10 Dec 1 981 — 5.4 41.1-44.2 c M 10 Total 33 980-1240 7-22 5.4-267.7 41.1-109.7 c M,F,T 4-10 Grand total (1991-94) 60 858-1290 7-22 5.4-267.7 41.1-115.8 C,R M,F,T 1-10 1 Questionable depth from spearfishing tournament. 3, A and B). Female GSIs (rc-1,695) ranged from 0.01 to 8.29, with greatest mean GSI (0.21-2.60) occur- ring from February through March in all years (Fig. 3A). Male GSIs (n=59) ranged from 0.02 to 0.82, with greatest GSI occurring during February (n- 1) in 1991 (Fig. 3B). Possible early transitional male GSI (n=l) was 1.65 in April 1994 (only one possible transitional male with an intact gonad and an accurate total weight was collected). Histological staging of ovaries collected in 1993 (n=209) and 1994 (r? =437) also confirmed winter-spring spawning (Fig. 4). All ovaries were stage 1 from June through September and stage 2 from October through January. Stage-3 ovaries first appeared in December and stage-4 in February. Stage-4 and stage-5 ovaries indicated that spawning occurred from February through April in 1993-94, corroborating GSI analysis. Stage-5 ovaries (containing nondegenerating POFs) were extremely rare (n- 7), occurring only in January, March, and July 1993, and April 1994. Small numbers of hydrated oocytes appeared as remnants in the cen- ter of stage- 1 ovaries after the spawning season. 420 Fishery Bulletin 96(3), 1998 Figure 2 Photomicrographs of histological sections of gag with heavy ventral pigmenta- tion: (A) ovarian tissue from an 858-mm-TL female caught in February 1993; (B) tissue from a 985-mm-TL possible early transitional male caught in February 1991; (C) tissue from a 1120-mm-TL possible early transitional male caught in April 1994; (D) testicular tissue with remnant oocytes from a 1143-mm-TL male caught in December 1992; PG = primary growth oocyte; CA = cortical alveolar oocyte; V = early vitellogenic oocyte, EHO = early hydrated oocyte (with coalesc- ing yolk globules); PPS = possible primary spermatocytes; PS = primary sperma- tocytes; SS = secondary spermatocytes; MS = mature spermatids; and RO = rem- nant oocyte. Magnification was 132x for A and 264x for B-D. Spawning mainly occurred from February through April 1991-94, but some evidence of spawning also occurred in December, January, May, June, and July. A total of 173 stage-4 and stage-5 ovaries were found during February, March, and April {n- 895). No hy- drated ovaries and only one ovary with POFs were found during December {n= 58) and during January {n- 77). Only one hydrated ovary, one ovary with hy- Collins et al.: Reproductive patterns, sex ratio, and fecundity of Mycteroperca microlepis 421 Figure 2 (continued) C drated oocyte remnants, and no ovaries with POFs were found during May (n=306). Hydrated oocyte remnants (in two ovaries) and POFs (in one ovary) were found in June (n = 180) and July (rc=159). Length and age at first spawning, and depth of spawning Length at first spawning (see “Methods” section for definition) was 710 mm TL in 1991, 570 mm TL in 1993, and 670 mm TL in 1994 for females, and 980 mm TL for males in 1994. Age at first spawning was 3 years for females and 7 years for males. Stage-4 females were caught at depths of 30 m and greater. All males were caught at depths greater than 41m. Batch fecundity Batch fecundity estimates (n=39) ranged from 10,864 to 865,295 hydrated oocytes (Fig. 5). These estimates came from gag 690-1065 mm TL, 4.5-16.5 kg TW, 23.4-871.9 g ovary weight, and 3-9 years of age. Regression analysis showed significant positive linear correlations between batch fecundity estimates 422 Fishery Bulletin 96(3), 1998 (BFE) and TL, gutted body weight (GBW), and age. Total length was the best predictor of batch fecun- dity (Fig. 5; BFE = 1.773 x 103(TL) - 1.119 x 106; r2=0.600, P<0.0001, n=39); followed by GBW ( BFE - 5.942 x 104(G£W) - 9.517 x 104; r2= 0.576, P<0.0001, n= 35); then age ( BFE = 9.476 x 104(age) - 1.213 x 105; r2=0.474, P<0.0001, n=38). Spawning frequency: by year, and age x year Spawning frequency estimates were determined for females >709 mm TL in 1991, >576 mm TL in 1993, and >669 mm TL in 1994. Variations in spawning frequency were significantly different among years tested (/-test, P<0.02). Spawning frequency for fe- males in 1991 was 27 times, with 29.3% (93/317) of individual females spawning at an average of every 3.4 d. Spawning frequency was not determined in 1992 owing to the absence of samples from January and February. Spawning frequency in 1993 was 14 with 16.1% (36/224) of individual females spawning at an average of every 6.2 d. Spawning frequency for 1994 was 8 with 14.0% ( 18/129) of individual females spawning at an average of every 7.2 d. Spawning frequency also varied significantly among ages (/-test, P<0.01) and by age x year (ANOVA, P<0.05) for gag (n >5 per age x year) (Table 3; Fig. 6). Age-2 females in 1993 had a spawning frequency = 0, although one spawn- ing female of this age occurred in 1994. Age-3 females in 1991 also had a spawning frequency = 0; 1993 and 1994 spawning frequencies were also low (6 and 7, respectively). Spawning fre- quency varied the most (6-71) among years for 4- to 8-yr-old gag. Nine year old gag spawned 41 times in 1991. All sample sizes for 10- to 12-yr- old, 16-yr-old, and 26-yr-old gag were inadequate (n<6) for estimation of spawning frequency. Annual fecundity Annual fecundity estimated with spawning fre- quency estimates by age and year ranged from 0.065 to 61.4 million (Table 4; Fig. 7). Only spawning frequencies for aged fish (n>5 per age and year) were used. Age was an effective pre- dictor of annual fecundity (AFE= 9.276 x 103(age)3-94; r2= 0.76, P<0.0001; n= 33). Discussion We found both the peak and maximum length of the spawning period for gag to be slightly longer than previously determined (e.g. McErlean, 1963; Hood and Schlieder, 1992; Coleman et ah, 1996; Koenig et ah, 1996) in the Gulf of Mexico. Both GSI and percent frequency of hydrated oocytes and POFs were greatest in February through April, but a few POFs were present in December-January and May-July. Our sampling may not have included peak spawning from 1992 through 1994, as it had in 1991. McErlean (1963), through macroscopic examination of gonads, estimated that spawn- ing occurred January through March. Hood and Schlieder ( 1992), through histological examina- tion, showed that peak spawning occurred in Collins et a I.: Reproductive patterns, sex ratio, and fecundity of Mycteroperca microlepis 423 February and March and females were spent from December to May. Coleman et al. (1996) and Koenig et al. (1996) used GSIs, fresh “squashes,” histological examinations, and ju- venile otolith daily increments and found peak spawning to occur in February-March, but some gag were reproductively active from January through mid-May in 1991. McErlean’s (1963) estimates of (annual) fe- cundity in 7-8 yr-old gag (n=3, 930-946 mm TL) were somewhat similar to our batch fecundity estimates for the same age fish (his and our range of estimates were 0.526-1.457 million and 0.249-0.865 million, respectively) although the exact oocyte size and stage counted by McErlean are unknown. We can infer, however, from the “bright-yellow eggs” he counted (McErlean3) that they were vitellogenic (nonhydrated) oocytes because fresh hydrated oocytes are clear. We counted only hydrated oocytes (n= 12 fish, 860- 1065 mm TL). McErlean’s ovary weights averaged 231.1 g, whereas ours averaged 656.7 g. This com- parison of ovary weight of similar size and age fish strongly suggests that McErlean’s gag were not ready to spawn. Gag population changes be- tween 1961 (McErlean, 1963) and 1991-94 (present study) may also have affected length, age, ovary weight, and fecundity (Johnson et al., 1993). Our estimates of gag batch fecundity (0.011- 0.865 million), spawning frequency by age and year (6-71; see Tables 3 and 4), and annual fe- cundity (0.065-61.4 million) are similar to those of several multiple spawning, demersal marine species similar to gag in size and longevity. Fitzhugh et al. (1993) estimated mean batch fecundity (1.6 million) and spawning frequency (46, n = 25) for black drum, Pogonias cromis, whereas Nieland and Wilson (1993) estimated ranges of batch fecundity (0.510-2.42 million), spawning frequency (25-28), and annual fecun- dity (13.0-67.0 million, n=41) for the same spe- cies. Wilson and Nieland (1994) estimated batch fecundity (0.160-3.27 million), and their data indicated a spawning frequency of 1-30 (calcu- lated by the senior author (LAC) of this paper) for red drum, Sciaenops ocellatus (n=51). Re- cently, Collins et al. (1996) estimated batch fe- cundity (0.001-1.70 million), spawning frequency (21-35), and annual fecundity (0.012-59.7 mil- lion) for red snapper, Lutjanus campechanus (n= 66). Spawning frequencies and possibly GSIs may be useful in forecasting trends in fishery production. For instance, our estimates of GSI and spawning fre- quency were greatest during 1991 and 1993, suggest- ing that population reproductive efforts may have been greater and gag juveniles may have been more 3 McErlean, A. J. [retired]. 1997. 4364 Hickory Shores Blvd., Gulf Breeze, FL 32561. Personal commun. 424 Fishery Bulletin 96(3), 1998 abundant in these years. This finding is corroborated by studies of Johnson and Koenig (in press), wherein abundance of age-0 gag in north Florida seagrass beds was higher in 1991 and 1993 than it was in 1992 and 1994. Johnson and Koenig (in press) also found that an alternating pattern of abundance was evident in the age structure of the gag fishery from 1984 through 1989, with odd years producing domi- nant year classes, if the same number of females are assumed to spawn each year. Table 3 Spawning frequency estimates (SFEs) by age and year for female gag (>total length (TL) of smallest female with hy- drated oocytes (HOs) or postovulatory follicles (POFs) dur- ing the spawning season). Age (yr) Year n n with HOs/POFs % with HOs/POFs SFE 2 91 0 — — — 93 6 0 0 0 94 1 1 100.0 57 3 91 9 0 0 0 93 119 8 6.7 6 94 8 1 12.5 7 4 91 20 6 30.0 28 93 10 2 20.0 17 94 78 8 10.3 6 5 91 105 28 26.7 25 93 34 6 17.6 15 94 8 1 12.5 7 6 91 8 2 25.0 23 93 13 4 30.8 26 94 12 3 25.0 14 7 91 13 1 7.7 7 93 28 13 46.4 40 94 2 0 0 0 8 91 6 1 16.7 16 93 6 5 83.3 71 94 8 2 25.0 14 9 91 9 4 44.4 41 93 1 1 100.0 85 94 2 0 0 0 10-12 91 2 1 50.0 47 93 0 — — — 94 3 1 33.0 19 16 91 1 0 0 0 93 0 — — — 94 0 — — — 26 91 1 0 0 0 93 0 — — — 94 0 — — — 3-26 91 174 43 24.7 23 2-9 93 217 39 18.0 15 2-12 94 122 17 13.9 8 The scarcity of nondegenerating POFs and the rela- tively low r2 values of regression equations suggest that some of our annual fecundity estimates are low, a problem prevalent in some past studies on other species (Nieland and Wilson, 1993; Taylor and Murphy, 1992). The age of POFs is an important fac- tor in judging whether POFs in a sample are from the same spawn as that for hydrated oocytes (Fitzhugh and Hettler, 1995). Delayed preservation in formalin may have caused some incompletely spawned ovaries to appear fully hydrated (if the frag- ile POFs were lost [Hunter et al., 1985], thus caus- ing low estimates of batch fecundity, spawning fre- Table 4 Data used for gag annual fecundity estimation, 1991-94. SFE=spawning frequency estimate by age and year. Catch date Total length (mm) Age (yr) Batch fecundity estimate SFE Annual fecundity estimate 1991 21 Mar 820 5 265,294 25 6,632,350 5 Apr 732 4 82,183 28 2,301,124 5 Apr 755 5 176,552 25 4,413,800 5 Apr 810 5 208,438 25 5,210,950 8 Apr 830 5 243,231 25 6,080,775 8 Apr 870 5 276,416 25 6,910,400 8 Apr 710 5 144,265 25 3,606,625 1993 25 Feb 870 7 335,275 40 13,411,000 25 Feb 1065 7 657,268 40 26,290,720 25 Feb 1038 8 865,295 71 61,435,945 25 Feb 950 8 696,251 71 49,433,821 20 Mar 860 7 772,158 40 30,886,320 21 Mar 878 8 368,974 71 26,197,154 21 Mar 914 8 510,637 71 36,255,227 22 Mar 983 7 661,993 40 26,479,720 23 Mar 1000 8 834,425 71 59,244,175 23 Mar 930 6 625,159 26 16,254,134 28 Mar 762 3 191,767 6 1,150,602 29 Mar 920 5 302,604 15 4,539,060 29 Mar 715 3 208,878 6 1,253,268 29 Mar 740 3 10,864 6 65,184 29 Mar 820 5 484,514 15 7,267,710 29 Mar 900 5 458,392 15 6,875,880 29 Mar 930 6 633,481 26 16,470,506 29 Mar 995 7 635,329 40 25,413,160 29 Mar 940 7 717,445 40 28,697,800 3 Apr 700 3 304,997 6 1,829,982 3 Apr 735 3 224,684 6 1,348,104 1994 15 Mar 846 4 428,371 6 2,570,226 15 Mar 690 4 187,216 6 1,123,296 15 Mar 827 5 310,357 7 2,172,499 15 Mar 910 6 487,719 14 6,828,066 6 Apr 1008 8 249,162 14 3,488,268 Collins et a I.: Reproductive patterns, sex ratio, and fecundity of Mycteroperca microlepis 425 quency, and annual fecundity). Varying amounts of “missing” hydrated oocytes due to partial spawning would have then caused the low r2s. An incomplete hydration process in some of the gag is another possible explanation for the low r2 values; samples were taken from gag caught at different times of the day, causing variable hydration (Hunter et al., 1985). Relatively low r2 values for the relation between batch fecun- dity and length and weight were also found for black drum (r2<0.47, in Nieland and Wilson, 1993) and swordfish, Ziphias gladius (r2<0.41, in Taylor and Murphy, 1992). Because dockside sampling was also used in these studies, de- layed preservation may have caused some black drum and swordfish POFs to be missed, thereby causing low r2 values. In this study, we used histological sections to identify sex in gag with heavy abdominal pig- mentation. Because male behaviors are associ- ated with fish demonstrating this pigment pat- tern (Gilmore and Jones, 1992), the assump- tion is widespread that all these fish are males. Yet we found that 5% of such pigmented fish were females. It is possible that these fish were actually in the early stages of sex change, a phase undetectable even with histology. Our observation of a lining of hydrated oo- cytes around the outside of the abdominal cav- ity in gag suggests that the ovary or the ovi- duct, or both, had ruptured at the time of hy- dration. This rupture probably did not occur during the fish’s capture and handling because the lining of hydrated oocytes was dry and firm when we observed this phenomenon on the boat. A similar rupture may have occurred in a cap- tive population of laboratory-matured female gag where two individuals spontaneously de- veloped hydrated oocytes but did not ovulate; no oocytes were shed although no ovarian rup- ture was noted, possibly due to the absence of a male.4 Although the proportion of males re- quired for adequate fertilization is unknown for gag, the sharp decline of males observed by Coleman et al. (1996), Koenig et al. (1996), McGovern et al. (in press), and in this study suggests a restricted availability of males dur- ing the spawning period. Ovarian hydration in females in the absence of males could lead to ovarian rupture. Although recent studies have increased our knowl- edge of gag reproduction, several significant ques- LU Li. 40 cn * ¥ 3 4 5 6 7 Age (years) 1991 ▼ 1993 * 1994 Figure 6 Gag spawning frequency estimates (SFE) by age and year, for ages with n> 5 by year. 4 Carr, W. 1992. Whitney Laboratory, University of Florida, St. Augustine, FL. Personal commun. tions remain. Gag appear to spawn at a discrete depth; therefore should only those females from a known spawning area be used for spawning fre- quency and fecundity estimates? Do fewer males in proportion to females prevent some mature female gag from spawning every year (Coleman et al., 1996; 426 Fishery Bulletin 96(3), 1 998 Koenig et al., 1996)? Would oocytes hydrate and then be resorbed if there was no male in close proximity? Can gag spawn as females early in the spawning sea- son, change sex, and then spawn as males late in the same season, as suggested for red grouper, Epin- ephelus morio, by Moe (1969)? Would measurement of hormones be a better indicator of the state of tran- sition than our histological examination? Does the possibility that sex change in serranids may occur quickly (captive Anthias squamipinnis changed from female to male in two weeks [Fishelson, 1970]) ex- plain why our number of possible early transitional males was low? Questions such as these can only be answered with specifically designed laboratory and field studies. Acknowledgments We are indebted to many persons and organizations. Several commercial fishermen in the Panama City, Florida, area allowed us to sample their catches, in- cluding Captains Ray Ward, Sigurd Smeby, Mark Raffield, Mike Raffield, Dannie Lee, and Steve Smeby. Jerry and Carl Anderson let us routinely sample their seafood markets at Panama City Beach. Charterboat captains Bill Archer and Charles Paprocki also participated in sampling. Special thanks are extended to NMFS Fishery Reporting Specialist Debbie Fable for her help in the field. Bruce Thompson, Jeff Render (deceased), and Cheryl Crowder (Louisiana State University) advised us on methods and ovarian histology. Lewis Bullock, Ron Taylor, and Ruth Reece (Florida Marine Research Institute, St. Petersburg) effectively helped with sam- pling, methods, and histology. David Wyanski (South Carolina Dep. Natural Resources, Charleston) as- sisted with identification of transitional males. Re- views by Felicia Coleman, Gary Fitzhugh, Yvonne Sadovy, and Lewis Bullock greatly improved the manuscript. Technical support was provided by An- drew David, John Brusher, Chris Keim, Marc Remy, Sean Murray, Mike Strohmenger, John Dahl, Sandra Neidetcher, Guy Pizzuti, and Sara Heath (NMFS, Panama City). Carol Parker and Betsy Black typed many drafts of this ms. This research was funded by the MARFIN (Marine Fisheries Initiative) Program of the NMFS, Southeast Region (Grant #94 MFIH09). Literature cited Briggs, J. C. 1958. A list of Florida fishes and their distribution. Bull. FL State Mus. Biol. Sci. 2:223-318. Bullock, L. H., and G. B. Smith. 1991. Seabasses (Pisces: Serranidae). Florida Mar. Res. Inst, (part II) vol. VIII, 243 p. Coleman, F. C., C. C. Koenig, and L. A. Collins. 1996. Reproductive styles of shallow water grouper (Pisces: Serranidae) in the eastern Gulf of Mexico and the conse- quences of fishing spawning aggregations. Environ. Biol. Fishes 47:129-141. Collins, L. A., A. G. Johnson, and C. P. Keim. 1996. Spawning and annual fecundity of red snapper (Lutjanus campechanus) from the northeastern Gulf of Mexico. In F. Arreguin-Sanchez, J. L. Munro, M. C. Balgos, and D. Pauly (eds.), Biology, fisheries and culture of tropical groupers and snappers, p. 174-188. ICLARM Conf. Proc. 48. Collins, M. R., C. W. Waltz, W. A. Roumillat, and D. L. Stubbs. 1987. Contribution to the life history and reproductive bi- ology of gag, Mycteroperca microlepis (Serranidae), in the South Atlantic Bight. Fish. Bull. 85:648-653. Fishelson, L. 1970. Protogynous sex reversal in the fish Anthias squami- pinnis (Teleostei, Anthiidae) regulated by the presence or absence of a male fish. Nature (Lond.) 227:90-91. Fitzhugh, G. R., T. G. Snider III, and B. A. Thompson. 1988. Measurement of ovarian development in red drum ( Sciaenops ocellatus ) from offshore stocks. Contrib. Mar. Sci., suppl. to vol. 30:79-83. Fitzhugh, G. R., B. A. Thompson, and T. G. Snider III. 1993. Ovarian development, fecundity, and spawning fre- quency of black drum Pogonias cromis in Louisiana. Fish. Bull. 91:244-253. Fitzhugh, G. R., and W. F. Hettler. 1995. Temperature influence on postovulatory follicle de- generation in Atlantic menhaden, Brevoortia tyrannus. Fish. Bull. 93:568-572. Gilmore, R. G., and R. S. Jones. 1992. Color variation and associated behavior in the epinepheline groupers Mycteroperca microlepis (Goode and Bean) and M. phenax Jordan and Swain. Bull. Mar. Sci. 51(1):83-103. Goodyear, C. P. 1993. Spawning stock biomass per recruit in fisheries man- agement: foundation and current use. In S. J. Smith, J. J. Hunt, and D. Rivard (eds.). Risk evaluation and biologi- cal reference points for fisheries management, p. 67- 81. Can. J. Fish. Aquat. Sci. 120. Hood, P. B., and R. A. Schlieder. 1992. Age, growth and reproduction of gag, Mycteroperca microlepis , (Pisces: Serranidae) in the eastern Gulf of Mexico. Bull. Mar. Sci. 51(3):337-352. Hunter, J. R., N. C. H. Lo, and R. J. H. Leong. 1985. Batch fecundity in multiple spawning fishes. In R. Lasker (ed.l, An egg production method for estimating spawning biomass of pelagic fish: application to the north- ern anchovy, Engraulis mordax, p. 67-77. U.S. Dep. Commer. NOAA Tech. Rep. NMFS 36. Hunter, J. R., and B. J. Macewicz. 1985. Measurement of spawning frequency in multiple spawning fishes. In R. Lasker (ed. ), An egg production method for estimating spawning biomass of pelagic fish: ap- plication to the northern anchovy, Engraulis mordax , p. 79- 94. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 36. Hunter, J. R., B. J. Macewicz, and J. R. Sibert. 1986. The spawning frequency of skipjack tuna, Katsuwonus pelamis , from the South Pacific. Fish. Bull. 84:895-903. Collins et a\.: Reproductive patterns, sex ratio, and fecundity of Mycteroperca microlepis 427 Huntsman, G. R., and W. E. Schaaf. 1994. Simulation of the impact of fishing on reproduction of a protogynous grouper, the graysby. N. Am. J. Fish. Manage. 14:41-52. Johnson, A. G., L. A. Collins, and J. J. Isely. 1993. Age-size structure of gag, Mycteroperca microlepis , from the northeastern Gulf of Mexico. Northeast Gulf Sci . 13:59-63. Johnson, A. G., and C. C. Koenig. In press. Age and size structure of the fishery and juve- nile abundance of gag Mycteroperca microlepis from the north- eastern Gulf of Mexico. Proc. Gulf Caribb. Fish. Inst. Johnson, A. K. 1995. Comparison of gonadal histology and sex steroid lev- els over the seasonal reproductive cycle of red grouper, Epinephelus morio. M.S. thesis, Univ. South Florida, Tampa, FL, 100 p. Koenig, C. C., F. C. Coleman, L. A. Collins, Y. Sadovy, and P. L. Colin. 1996. Reproduction in gag (Mycteroperca microlepis ) in the eastern Gulf of Mexico and the consequences of fishing spawning aggregations. In F. Arreguin-Sanchez, J. L. Munro, M. C. Balgos, and D. Pauly (eds.), Biology, fisher- ies and culture of tropical groupers and snappers, p. 307- 323. ICLARM Conf. Proc. 48. McErlean, A. J. 1963. A study of the age and growth of the gag, Mycteroperca microlepis Goode and Bean (Pisces: Serranidae) on the west coast of Florida. Florida Board Conserv. Mar. Lab., Tech. Ser. 41, 29 p. McErlean, A. J., and C. L. Smith. 1964. The age of sexual succession in the protogynous her- maphrodite Mycteroperca microlepis. Trans. Am. Fish. Soc. 93:301-302. McGovern, J. C., D. M. Wyanski, O. Pashuk, C. S. Manooch III, and G. R. Sedberry. In press. Changes in the sex ratio and size at maturity of gag, Mycteroperca microlepis , from the Atlantic coast of the southeastern United States during 1976-1995. Fish. Bull. Moe, M. A., Jr. 1969. Biology of the red grouper Epinephelus morio (Valenciennes) from the eastern Gulf of Mexico. Florida Dep. Nat. Resour. Mar. Res. Lab. Prof. Pap. Ser. 10, 95 p. Nieland, D. L., and C. A. Wilson. 1993. Reproductive biology and annual variation of repro- ductive variables of black drum in the northern Gulf of Mexico. Trans. Am. Fish. Soc. 122:318-327. Render, J. H., and C. A. Wilson. 1992. Reproductive biology of sheepshead in the northern Gulf of Mexico. Trans. Am. Fish. Soc. 121:757-764. Sadovy, Y., A. Rosario, and A. Roman. 1994. Reproduction in an aggregating grouper, the red hind, Epinephelus guttatus. Env. Biol. Fishes 41:269-286. SAS Institute, Inc. 1988. User’s guide, release 6.03. SAS Institute, Inc., Cary, NC. Shapiro, D. Y., Y. Sadovy, and M. A. McGehee. 1993. Periodicity of sex change and reproduction in the red hind .Epinephelus guttatus, a protogynous grouper. Bull. Mar. Sci. 53(3): 1 151-1 162. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. W. H. Freeman and Co., NY, 859 p. Taylor, R. G., and M. D. Murphy. 1992. Reproductive biology of the swordfish Xiphias gladius in the straits of Florida and adjacent waters. Fish. Bull. 90:809-816. Wallace, R. A., and K. Selman. 1981. Cellular and dynamic aspects of oocyte growth in teleosts. Am. Zool. 21:325-343. West, G. 1990. Methods of assessing ovarian development in fishes: a review. Aust. J. Mar. Freshwater Res. 41:199-222. Wilson, C. A., and D. L. Nieland. 1994. Reproductive biology of red drum, Sciaenops ocellatus, from the neritic waters of the northern Gulf of Mexico. Fish. Bull. 92:841-850. 428 Abstract .—This study describes the stomach contents of 95 harbor por- poises (Phocoena phocoena) killed in groundfish gill nets in the Gulf of Maine between September and December, 1989-94. The importance of prey was assessed by frequency of occurrence, numerical proportion, and proportion of ingested mass. Atlantic herring ( Clupea harengus) was the most impor- tant prey, occurring in 78% of noncalf porpoise stomachs and contributing 44% of ingested mass. Pearlsides ( Maurolicus weitzmani ), silver hake (Merluccius bilinearis ), and red and white hake (Urophycis spp. ) were com- mon prey items. There were no signifi- cant differences among diets of sex and maturity groups, but the calf diet dif- fered significantly from adults in num- ber of Atlantic herring eaten and the total mass of food consumed. At four to seven months of age, calves were eat- ing pearlsides, small silver hake, and eu- phausiids ( Meganyctiphanes norvegica) while still nursing. Manuscript accepted 10 December 1997 Fishery Bulletin 96:428-437 (1998). Autumn food habits of harbor porpoises, Phocoena phocoena, in the Gulf of Maine Damon R Gannon Duke University Marine Laboratory, Nicholas School of the Environment 1 35 Duke Marine Lab Road, Beaufort, North Carolina 285 1 6 E-mail address: dpg3@acpub.duke.edu James E. Craddock Woods Hole Oceanographic Institution Woods Hole, Massachusetts 02543 Andrew J. Read Duke University Marine Laboratory, Nicholas School of the Environment 1 35 Duke Marine Lab Road, Beaufort, North Carolina 285 1 6 Harbor porpoises ( Phocoena pho- coena) from the Bay of Fundy and Gulf of Maine are believed to com- prise a single population, hereafter referred to as the Gulf of Maine population (Palka et al., 1996; Wang et al., 1996). To date, studies of the food habits of this population have been restricted to samples collected in the Bay of Fundy during summer, where porpoises feed primarily on Atlantic herring ( Clupea harengus ; Smith and Gaskin, 1974; Recchia and Read, 1989; Smith and Read, 1992). Many porpoises leave the Bay of Fundy in fall, moving south- ward into the Gulf of Maine (Gas- kin, 1977; Gaskin, 1984; Read and Westgate, 1997). During winter, a portion of the population disperses over the continental shelf from New England to North Carolina (Pola- check et al., 1995; Read et al., 1996). Because of their small size and limited energy stores, harbor por- poises must remain close to food resources to avoid starvation (Koop- man, 1994). Moreover, their un- usual life history incurs high ener- getic costs; most females attain sexual maturity at three years of age and give birth to a calf each year (Read and Hohn, 1995). Lactation lasts for at least eight months; thus mature females spend most of their lives simultaneously pregnant and lactating. This intensive reproduc- tive schedule requires calves to be- come nutritionally independent at a relatively early age, usually before the end of their first year (Smith and Read, 1992). Large numbers of harbor por- poises are killed each year in gill nets in the Bay of Fundy, Gulf of Maine, and Mid-Atlantic Bight (Read and Gaskin, 1988; Read et al., 1993; Bravington and Bisack, 1996). For the Gulf of Maine, the estimated average annual harbor porpoise bycatch for 1990 to 1995 was 1800 (Bisack1 ). Little is known about the process by which porpoises become entangled in gill nets, and thus ef- forts are hampered in mitigating this conservation problem. Porpoises may become entangled because they feed on fish species targeted by the fish- 1 Bisack, K. 1996. Harbor porpoise bycatch estimates in the U.S. Gulf of Maine sink gillnet fishery: 1994 and 1995. Paper pre- sented to the International Whaling Com- mission Scientific Committee Meeting in Aberdeen, Scotland, June 1996 (in review). Gannon et al.: Food habits of Phocoena phocoena 429 Figure 1 Capture locations of harbor porpoises taken during the autumn ( 1989-94) in the Gulf of Maine sink gillnet fishery and used in this analysis of food habits. The isobath shown is 91.4 m (50 fathoms). ery or because they feed on the same prey as the target species. In this paper, we examine the stom- ach contents of harbor porpoises in the Gulf of Maine during autumn and in- vestigate dietary differences amongst various sex and maturity categories. Our main objectives were to elucidate seasonal changes in the harbor por- poise diet and expand our knowledge of the dynamics between porpoises and their prey that may be responsible for entanglement of porpoises in gill nets. Methods Sample collection The sample consisted of 95 porpoises killed in gill nets during autumn (1 September- 31 December) of 1989 and 1991-94. All porpoises were captured in bottom tending gill nets set for groundfish, principally cod ( Gadus morhua), pollock (Pollachius virens), goosefish ( Lophius americanus ), and several species of flatfish. Most por- poises were taken in the vicinity of Jeffreys Ledge in the west central Gulf of Maine, at water depths between 35 and 185 m (Fig. 1). All samples were obtained by fisheries observers work- ing onboard gillnet vessels. Observers were instructed to retain whole por- poise carcasses whenever possible, but when sea conditions or other factors prevented re- tention of carcasses, observers excised stomachs in the field. Carcasses and excised stomachs were fro- zen after the vessels returned to shore (usually 12- 48 hours post mortem) for later examination. On the basis of age (determined from dentinal growth layers and body length; see Read and Hohn, 1995) and reproductive condition (determined by examination of gonads and mammary glands; see Read and Hohn, 1995), porpoises were classified to the following sex, maturity, and reproductive catego- ries: porpoises were considered calves (less than one year of age, not fully weaned), juveniles (older than one year but sexually immature), or sexually mature. The sex and maturity composition of the sample was as follows: (males and females combined) calves = 13; female juveniles = 12; male juveniles = 18; fe- male mature adults = 10; male mature adults = 34; and unknown sex or maturity = 8. Because sample sizes were small, pregnant (n= 4), simultaneously pregnant and lactating (n=5), and resting adult fe- males (rc=l) were pooled in the “mature female” group for statistical analyses. However, to facilitate com- parisons with the findings of Recchia and Read (1989), data for lactating and nonlactating mature females are also presented separately. Prey identification The contents of all three stomach chambers were examined in the laboratory. Intact prey were removed first, then loose flesh was decanted. The remaining stomach contents were poured through a 1-mm metal sieve to separate hard parts from liquefied digesta. Solid prey remains used for identification were sepa- rated from other skeletal remains by hand. Struc- tures used to identify partially digested food items included sagittal otoliths, dentary bones, and skulls 430 Fishery Bulletin 96(3), 1998 of teleosts; lower mandibles (“beaks”) from cephalo- pods; tooth cusp plates (“combs”) from agnathans; and exoskeletons and eyes from crustaceans. Prey items were identified with the aid of a laboratory reference collection and published guides, including those of Bigelow and Schroeder (1953), Clarke ( 1986), Harkonen (1986), and Scott and Scott (1988). Prey importance Relative food importance in the autumn diet of the harbor porpoise was determined by 1) frequency of occurrence, 2) proportion of numerical abundance, and 3) proportion of total ingested mass. Frequency of occurrence is the percentage of porpoise stomachs containing a particular food type. Proportion of nu- merical abundance is the number of individuals of a prey species recovered from all stomachs, divided by the total number of all prey from all stomachs. The number of individuals from each fish species in each stomach was determined by summing the number of intact fish and half the number of free otoliths. The number of either upper or lower beaks (whichever were more abundant) from each species was used to determine the number of squid present. Proportion of prey mass is the percentage of total prey mass in the stomach at the time of death that was represented by a particular species. Reconsti- tuted mass, or the mass of prey prior to ingestion, rather than the existing mass of partially digested prey, was used in this calculation. Reconstituted prey masses were estimated from body lengths of intact prey and the lengths of otoliths or cephalopod beaks (Table 1). If a stomach contained more than 25 otoliths from the same species, all otoliths from that species were counted, and a subsample of 25 was randomly selected and measured. Otoliths were scored on a scale from 0 (undamaged otoliths re- Table 1 Equations used to estimate length and mass of harbor porpoise prey. ML = mantle length; H = length; OL = otolith length; LRL = lower rostral length; and SL = standard length. Length is in hood length; M = mass; FL = fork millimeters and mass is in grams. Prey species Equations Source Bathypolypus arcticus ML = 15.4 + 12.28 H Clarke, 1986 (North Atlantic octopus) In M= 1.06 + 2.55 In H Clarke, 1986 Clupea harengus FL = 69.23 OL - 27.48 Recchia and Read, 1989 (Atlantic herring) log M = 3.12 logFL - 5.41 Recchia and Read, 1989 Gadus morhua ln(FL/10) = 3.3138 + 1.6235 ln(OL/10) Hunt, 1992 (Atlantic cod) M= 0.0124 (FL/10)2 93 Bowen and Harrison, 1994 Illex illecebrosus (Northern short-fin squid) lnM= 1.773 + 2.4 1 nLRL Clarke, 1962 Loligo pealei log ML = 1.767 + 1.4 log LRL Gannon et al., 1997b (Long-fin inshore squid) M = 0.25662 (ML/10)2 1582 Lange and Johnson, 1981 Maurolicus weitzmani1 FL = 9.82 + 28.75 OL Harkonen, 1986 (Weitzman’s pearlsides) M = 0.3737 OL2 503 Harkonen, 1986 Merluccius bilinearis FL = 20.9 L- 0.41 Recchia and Read, 1989 (Silver hake) logM = -2.26 + 3.08 log(FL/10) Kohler et al., 1970 Peprilus triacanthus2 SL= -9.15919 + 25.01871 OL Present study (r2=0.983) (Butterfish) log M = -0.67576 + 3.222 logOL Present study (r2=0.924) Pollachius virens ln(FL/10) = 3.251 + 1.6251 ln(OL/10) Harkonen, 1986 (Pollock) M = 0.0134 (FL/10)294 Bowen and Harrison, 1994 Scomber scombrus FL/ 10 = 7.33 OL + 0.37 Recchia and Read, 1989 (Atlantic mackerel) M = 0.00756 (FL/10)3 082 Kulka and Stobo, 1981 Sebastes spp.3 FL = 16.165 L1 224 Harkonen, 1986 (Rockfish) M = 0.0741 OL3 295 Harkonen, 1986 Urophycis spp.4 FL/10 = 1.525 OL1 1456 Clay and Clay, 1991 (Red and white hake) M = 0.003998 (FL/10)3 1718 Clay and Clay, 1991 1 Taxonomy of the genus Maurolicus has been revised recently (Parin and Kobylansky, 1996). The equations used to estimate M. weitzmani size are those given by Harkonen (1986) for M. muelleri. 2 Standard length range: 49-153 mm; weight range: 3-104 g; n = 44. 3 Equations given by Harkonen (1986) for S. marinus. 4 Equations given by Clay and Clay (1991) for U. tenuis. Gannon et al.: Food habits of Phocoena phocoena 431 trieved from skulls) to 5 (severely degraded, free otoliths) following the methods of Recchia and Read (1989). Otoliths categorized as 3 or higher were not used in size estimations, unless no undamaged otoliths were present. When only damaged otoliths from a particular prey species were present in a por- poise stomach, the available skeletal structures were measured; consequently the reconstituted prey mass for that stomach may have been underestimated (see Jobling and Breiby, 1986; Sekiguchi and Best, 1997) These three measures of prey importance were applied to data from the 82 noncalf porpoises as a group and to each sex and maturity class. Food habit studies in which different methods are used can yield widely disparate results, making it difficult to draw comparisons between studies (Gannon et al., 1997a, 1997b). Because one of the primary objectives of this research was to obtain information on seasonal changes in the diet, it was important for these data to be treated in a manner similar to those of Recchia and Read (1989) and Smith and Read (1992). Results Overall sample Table 2 lists the numbers and mean sizes of 15 prey taxa recovered from the 95 porpoise stomachs. At- lantic herring ( 78%), silver hake ( Merluccius bilinearis, 68%), pearlsides ( Maurolicus weitzmani, 38%), and red and white hake ( Urophycis spp., 29%) occurred most frequently in the stomachs of the 74 noncalf porpoises (Table 3). Atlantic herring represented only 7% of the food by proportion of numerical abundance but accounted for 44% of ingested mass. Pearlsides ac- counted for 67% of food by proportion of numerical abundance but only 3% by ingested mass, owing to their small size. The unknown fish present in por- poise stomachs may have been alewives ( Alosa pseudoharengus ) but this could not be determined with certainty. Both red and white hake ( Urophycis chuss and U. tenuis ) were present; however it is dif- ficult to differentiate between small, eroded otoliths from red and white hake, therefore all Urophycis otoliths were grouped together. Atlantic hagfish (Myxine glutinosa) and euphausiids ( Meganycti - phanes norvegica ) were included in analyses of fre- quency of occurrence only because the numerical abundance and mass of these two species were diffi- cult to estimate. To allow comparisons to be drawn with the summer diet, data from Recchia and Read (1989) are also given in Table 3. Figure 2 shows length-frequency distributions for the three most abundant prey: pearlsides, silver hake, and Atlantic herring. On average, Atlantic herring was the largest prey consumed by length (254 mm ±36 SD ) with a range from 159 to 339 mm. The average fork length Table 2 Number and mean sizes of food items present in the stomachs of harbor porpoises sampled in the Gulf of Maine during autumn. ML = mantle length, FL = fork length, and SL = standard length. Present = present in porpoise stomach contents but numerical abundance not determined. Food item n Length measurement Mean length ± SD (mm) Mean mass ± SD 300 dives) but neither hav- Gorfine et al.: Two methods for estimating abundance of Haliotis rubra 441 Table 1 Analysis of variance in density estimates, made from radial transect collections, among fifteen sites nested within three locations and among four divers for each of three size classes of H. rubra (type IV SS). ns = nonsignificant; * = P<0.10; ** = P<0.05; *** = PcO.Ol). co2 = relative magnitude of variance estimate (%). MS = mean square. Class Source df MS F P c o 2 Significance Juvenile Location 2 0.14 2.42 0.1312 14 ns Diver 3 0.05 1.92 0.1593 14 ns Location x diver 4 0.04 1.36 0.2822 3 ns Site (location) 12 0.06 1.69 0.0904 41 * Diver x site (location) 20 0.03 0.75 0.7631 25 ns Residual 63 0.04 Prerecruit Location 2 2.36 19.24 0.0002 44 *** Diver 3 0.06 1.10 0.3723 0 ns Location x diver 4 0.08 1.32 0.2936 2 ns Site (location) 12 0.12 3.56 0.0003 37 *** Diver x site (location) 21 0.06 1.70 0.0475 18 ** Residual 81 0.03 Postrecruit Location 2 2.60 9.98 0.0028 43 *** Diver 3 0.33 3.16 0.0461 16 ** Location x diver 4 0.14 1.30 0.3027 3 ns Site (location) 12 0.26 1.95 0.0393 27 * Diver x site (location) 21 0.10 0.78 0.7390 11 ns Residual 90 0.13 ing previously used the other method, performed both techniques to census abalone populations at each study site. Abalone counts in four radial transects and four timed collections of 10-min duration were made by each diver at each site so that the effects of divers and methods were fixed and the design bal- anced. Timed searches were performed after transect counts in each instance to avoid disturbance due to removal of abalone during timed searches. Separate analyses for each method were made with general linear models to determine the significance of the effects of diver and site on abundance esti- mates. A combined analysis was not possible because of the different base sampling units for each method. Where Cochran’s test indicated significant hetero- geneity of variance, data were log-transformed prior to analysis to achieve homoscedasticity (Winer, 1971). Diver x site interaction was omitted from the analy- ses presented in Table 1 because preliminary analy- ses showed this effect to be nonsignificant. Trial stock surveys The effects of divers on the detection of differences in abundance were further investigated during trial surveys of substocks at Gabo Island and Sandpatch Point near Mallacoota and Sticks Reef near Altona in Port Phillip Bay (Fig. 1). These surveys were aimed at determining how well variation among several locations was detected with this method when it was applied over a spatial scale comparable to that used in routine stock surveys. Five sampling sites were randomly selected at each location, and for each site, three out of five divers were selected to undertake the sampling depending on their availability and fitness to dive. Two to three sites at each location were sampled during each day of the surveys. At each site, each diver made three replicate radial transect collections of emergent prerecruit and postrecruited abalone. Each diver also collected juvenile abalone (<80 mm) in one of their three assigned transects by searching cryptic habi- tat and turning over small boulders. All abalone col- lected were brought on board the research vessel, and the maximum shell length of each was measured to the nearest mm. Analysis of variance of the mixed linear model was used to test the effects of locations, sites nested within locations, and divers on abundance estimates. The SAS® general linear model (GLM) procedure was used (SAS Institute, Inc., 1989). Separate analyses were performed for three size classes of abalone (ju- venile <80 mm, 80-mm < prerecruit < LML, and postrecruit > LML). Data were converted to number per m2. The relative treatment magnitude, omega squared (to2), for each effect in the GLM was calculated (Keppel, 1991). 442 Fishery Bulletin 96(3), 1 998 Design of stock monitoring program Abundance estimates of abalone stocks in Victoria were made annually at 60 sites, using radial transects during 1992 to 1994. These estimates were made during routine stock surveys for the Victorian program for abalone stock assessment. At each site, nine replicate abalone collections were made. Sites were selected from abalone habitat with commercial quantities of abalone. During 1992, sites for radial transect surveys were initially chosen at random from between the 5-m and 18-m isobaths to avoid the wave-break zone and to minimize hyperbaric exposure. Accurate navigational fixes were made at each site with a global positioning system (GPS). During successive years each site was resurveyed after having been located with GPS. This mixed sam- pling design has been shown to have greater power than nested designs for benthic monitoring and avoids the potential bias of fixed designs (Van der Meer, 1997). Application of methods to stock monitoring An analysis of variance was applied to results from stock monitoring surveys to determine the inter- annual variation in abundance for each zone. This was in the form of a mixed linear model in the SAS® general linear model procedure. The relative treatment magnitude for each effect in the GLM was calculated with the same procedure as that used in the trial stock surveys. The effect of differences between research divers on abalone abundance estimates from radial transects was further investigated (with two- and three-way interactions omitted) across all zones from an analysis of 1992-94 stock surveys. During each field survey of a group of adjacent sampling sites, the dive team was composed of three or four out of a possible twelve divers. Differences in the composi- tion of dive teams between surveys occurred for lo- gistical reasons. At each site three divers performed three radial transects each where the allocation of directions for these transects among divers was ran- dom. Differences among divers were compared by using Ryan’s test (Day and Quinn, 1989), and the effect of diver experience on abundance estimates was determined by regressing divers’ mean collections of abalone per transect against the number of transects each diver performed. Power to detect changes in abundance Monte Carlo simulations were made to estimate the probability of detecting cumulative annual changes in abundance. The simulations were based on data from stock surveys conducted during 1992 and 1993 that employed radial transects as well as data from 1989 and 1991 surveys based on timed searches. Sce- narios involving different combinations of annual change increments (increases) and number of years sampled were each simulated 200 times for each management zone (i.e. central, eastern and western; Fig. 1). Divers were randomly selected from those who participated in the surveys, and the abundance estimates for each radial transect were adjusted for the diver factor and variation within each site, whereas the abundance estimates for timed searches were standardized by using a linear regression be- tween pairs of divers (McShane, 1994). The respec- tive analyses of variance used to analyze the stock survey data for each method were applied to each simulation, and the proportion of tests that showed a significant (a=0.05) annual change in abundance (year effect) was calculated. The proportion of sig- nificant results for each management zone was then plotted against the number of years simulated for each increment level. A smoothing function was applied to each curve to eliminate fluctuations due to changes in confidence levels among years in the series. Results Operator effects on methods Diver effects on radial transects During the first experiment at Gabo Island, the precision (SE/ic) of the divers’ abalone counts for each transect ranged from 0.04 to 0.27. Cochran’s test showed the divers’ variances of precision to be homoscedastic (a=0.01). Differences among divers’ mean precision values were not significantly different (F=1.43; df =3, 8; P>0.10). Collectively, the four groups of twelve estimates of abundance represent four identical surveys of the same site which yielded similar mean abundance estimates on each occasion (Fig. 2). Estimates of mean abalone abundance for the site during the sec- ond experiment were also similar (Fig. 3) and no sig- nificant differences were detected among the four divers (F=0.01; df=3, 33; P>0.10). Comparison of diver effects on radial transect and timed search survey methods There were no sig- nificant differences between the two divers when using radial transects (F=0.19; df=l, 2; P>0.10) nor were there significant differences when using timed searches (F=7.31; df=l, 2; P>0.10). Notably, the dif- ferences between the divers’ mean abundance esti- mates (Fig. 4) were much smaller for radial transects Gorfine et al.: Two methods for estimating abundance of Haliotis rubra 443 Figure 2 Comparison of mean abundance estimates of H. rubra at Gabo Island among surveys using radial transects. Error bars are 95% confidence limits. (effect size=9%) than for timed searches (ef- fect size=44%); thus power in the analysis was likely to be low. Trial stock surveys with radial transects There were significant differences in den- sity between all locations for the prerecruit and postrecruit size classes, but locations were not significantly different for juvenile abalone (Table 1). Locations and sites within locations accounted for most of the variation in density of pre- and postrecruit abalone. Significant differences between divers occurred for postrecruits, but the magnitude of the variance effect for divers was only about one third of the variance effect for sites. Sites varied significantly within locations for each size class; for ju- veniles and post recruits this was at the 10% confidence level, and for prerecruits at the 1% confi- dence level. Only one interaction effect, diver x site (location) for prerecruit density, was significant. Application of methods to stock monitoring Variation in abalone abundance estimates from both methods showed significant diver effects in two out of three instances (Tables 2 and 3). However, the only significant year effect detected was for estimates from radial transect surveys in the central zone of the fish- ery. The lack of a significant diver effect in this in- stance may have increased the power to detect a year effect. Consistent with the high spatial variability that characterizes abalone populations, site effects were significant in all instances. Year x site interac- tions were significant in two instances for each method, indicating that interannual changes in abun- dance varied among sites within the affected zones. Only one year x diver interaction was significant and because some divers were not represented in each year, the interpretation of this effect in this instance is problematic. For each method only a relatively small amount of the variation in abundance could be attributed to the effect of year. Sites generally accounted for most of Diver 1 Diver 2 Diver 3 Diver 4 Figure 3 Comparison of mean abundance estimates of H. rubra at Gabo Island among divers using radial transects. Error bars are 95% confidence limits. 70 60 50 40 30 20 10 Radial transect Timed search o -I . 1 , i Diver 1 Diver 2 Diver 1 Diver 2 Figure 4 Diver effects on mean abundance estimates of H. rubra for each survey method. Error bars are 95% confidence limits; base sam- pling units were 30 m2 for radial transects and 10 min for timed searches. 444 Fishery Bulletin 96(3), 1998 Table 2 Analysis of variance for interannual differences in abundance estimates of H. rubra for each of Victoria’s fishery management zones, made from timed collection surveys, during 1989-91 (type IV SS). (ns = nonsignificant; * = PcO.lO; ** = P<0.05; *** = P<0.01). ( o 2 = relative magnitude of variance estimate (%). Management zone Source df MS F P a? Significance Central Year 2 136.60 0.69 0.5141 i ns Site 28 286.04 2.42 0.0008 45 *** Diver 16 205.08 1.73 0.0525 13 * Diver x site 83 92.85 0.79 0.8714 20 ns Year x site 21 198.87 1.68 0.0468 16 * Year x diver 4 25.70 0.19 0.9404 4 ns Year x site x diver 10 138.32 1.17 0.3207 2 ns Residual 98 118.28 Eastern Year 2 203.64 1.30 0.2914 1 ns Site 13 181.97 1.99 0.0230 18 ** Diver 8 54.01 0.59 0.7863 5 ns Diver x site 71 77.91 0.85 0.7872 15 ns Year x site 24 156.81 1.71 0.0241 24 ** Year x diver 4 60.83 1.59 0.1961 3 ns Year x site x diver 41 38.34 0.42 0.9993 34 ns Residual 224 91.66 Western Year 2 160.84 1.40 0.2814 1 ns Site 15 836.77 5.32 0.0001 67 *** Diver 5 779.21 4.95 0.0008 20 *** Diver x(site 43 186.48 1.18 0.2755 8 ns Year x site 13 114.87 0.73 0.7262 4 ns Year x diver 0 — — — — — Year x site x diver 0 — — — — — Residual 54 157.38 the variation in the ANOVA model. For timed searches the variance estimate for diver effects was about one third that for sites; however for transects, the relativity between diver and site variance effects was more variable. The values for the relative mag- nitude of variance estimates attributable to divers were similar between the two methods. Mean abundance estimates from radial transects during stock monitoring surveys varied significantly among divers; divers 4 and 5 in particular collected relatively fewer abalone per transect and diver 8 col- lecting more than the other divers (Table 4). Ryan’s test did not reveal a significant difference between the remaining nine divers’ mean relative abundance estimates (Table 4). A regression of the divers’ mean relative abundance estimates against the respective numbers of radial transects performed (r2=0.01, n = 1339) showed that the slope of the relationship was not significantly different from zero (P>0.10). Power to detect changes in abundance The number of years of sampling required for an 80% or greater predicted probability of detecting a sig- nificant cumulative change in abundance with ra- dial transects ranged from 1-3 years for all zones. The western zone required 3 years for 2.5% and 5% increments to become significant on 80% of occasions and 2 years for the 10% increment. The other two zones required only one year for all increments to become significant (Fig. 5A). In contrast, timed searches (Fig. 5B) required three years to detect a cumulative annual increase of 10% in the central zone with 80% probability, and five years for the same size change in both the eastern and western zones. In the eastern zone the probabil- ity of detecting increases with timed searches was simi- lar to that for the central zone with transects. In this instance only one year was required to have an 80% probability of detecting each increment of change. Discussion Radial transects provided precise estimates of aba- lone abundance and did not vary significantly be- tween different divers in three of the investigations made during this study. However, the importance of selecting scientific divers with aptitude for these types of surveys is underscored by the fact that three Gorfine et a I,. Two methods for estimating abundance of Haliotis rubra 445 Table 3 Analysis of variance for interannual differences in abundance estimates of H. rubra for each of Victoria’s fishery management zones, made from radial transect surveys, during 1992-94 (type TV SS). ns = nonsignificant; * = PcO.lO; ** = P<0.05; *** = P<0.01. ( o 2 = relative magnitude of variance estimate (%). Management zone Source df MS F P of2 Significance Central Year 2 515.43 4.08 0.0222 3 ** Site 29 378.00 5.04 0.0001 54 *** Diver 10 93.24 1.24 0.2611 1 ns Diver x site 123 47.16 0.63 0.9989 21 ns Year x site 56 126.36 1.69 0.0022 18 *** Year x diver 5 31.23 0.49 0.7842 1 ns Year x site x diver 35 64.17 0.86 0.7066 2 ns Residual 479 75.00 Eastern Year 2 130.23 0.87 0.4312 0 ns Site 14 801.99 5.33 0.0001 38 *** Diver 5 1170.54 7.78 0.0001 21 *** Diver x site 46 274.05 1.82 0.0021 23 *** Year x site 28 150.21 1.00 0.4728 0 ns Year x diver 4 301.05 5.22 0.0057 11 *** Year x site x diver 18 57.69 0.38 0.9899 7 ns Residual 242 150.48 Western Year 2 36.54 2.30 0.1214 1 ns Site 12 261.00 7.14 0.0001 41 *** Diver 7 100.08 2.74 0.0096 7 *** Diver x site 40 47.61 1.30 0.1201 7 ns Year x site 24 147.51 4.03 0.0001 40 *** Year x diver 6 65.70 1.31 0.3016 1 ns Year x site x diver 18 50.04 1.37 0.1201 4 ns Residual 225 36.54 divers’ mean transect collections differed significantly from the remaining nine divers during three annual abalone stock surveys. During routine monitoring, two divers (4 and 5) consistently collected less aba- lone than the others and frequently aborted dives after having failed to complete all their allocated transects. Reasons given for terminating their dives included problems with middle ear equalization, dif- ficulty coping with the surge and backwash created by waves, and fouling of lines and air hoses due to entanglement. These observations concur with Shep- herd (1985) who found that surge and kelp density were two significant factors affecting research diver efficiency in underwater censuses of abalone popu- lations. In contrast to divers 4 and 5, diver 8 consis- tently collected more abalone than the other research divers. However, diver 8 participated only in those parts of the study conducted in the eastern zone, where abalone stocks are generally more abundant. Also, as an ex-abalone diver, diver 8 could reason- ably be expected to possess a superior ability in per- forming abalone searches and collections that may have biased his estimates. These observations suggest that some divers may not possess the aptitude required to conduct under- TabJe 4 Mean number (least squares) of H. rubra collected per transect by each diver during 1992-94 Victorian abalone stock surveys. Vertical bars in “ns" column indicate groups of means that were not detected as significantly different by Ryan’s test (a=0.05). Is = least squares; SE = standard error of the least squares means. Diver n Is mean ± SE ns 8 36 19.48 1.92 9 147 16.03 0.95 1 467 12.62 0.53 12 38 12.00 1.87 3 408 11.26 0.57 11 66 10.74 1.42 7 6 10.50 4.71 6 17 10.03 2.80 2 119 10.02 1.06 10 15 9.73 2.98 4 107 7.83 1.12 5 12 3.58 3.33 water censuses and highlight the need for under- standing the nature of research being conducted. The ability to work at sea in often arduous conditions, discipline in adherence to sampling protocols, and 446 Fishery Bulletin 96(3), 1 998 the importance of avoiding a competitive atmosphere in which bagging the most abalone becomes the ob- jective are essential characteristics of an effective abalone survey team. The problem of competition implies that nondestructive sampling, such as count- ing, may reduce potential bias associated with col- A Radial transects B Timed searches Simulated year Figure 5 Predicted probabilities for detecting cumulative annual changes in H. rubra abundance from (A) radial transect surveys and (B) timed searches. Gorfine et al.: Two methods for estimating abundance of Haliotis rubra 447 lecting. Indeed, the diver who consistently collected more abalone from transects during routine stock surveys was not significantly different from other members of the dive team when only counts of aba- lone were required. Similarly, McShane ( 1995) found that by reducing the handling time involved with the collection of large numbers of abalone from dense aggregations he reduced the operator bias in the es- timation of patch frequency. Counting also has the advantage of considerably reducing the time needed to complete each transect. However, collecting be- comes necessary when length-frequency data are required because in-situ measurements are impracti- cal under the surgy conditions commonly encountered. Experience in the radial transect method did not have a significant influence on the number of aba- lone estimated despite the fact that the number of transects completed per diver varyied over a wide range (6-438 transects). Differences among divers are expected when those who sampled at only a few locations are compared with divers who sampled over a larger range of abalone populations. This expecta- tion is supported by the lower standard deviations associated with the mean collections per transect of the less experienced divers. The standard deviations of the divers’ mean rela- tive abundance estimates from radial transects were relatively large, especially for those divers who had completed many such transects. This large range reflects the high spatial variability in abalone den- sities both within and between sampling sites which is evident from the relatively large proportion of vari- ance contributed by location and site effects. During the trial stock surveys, variation among sites within locations was mostly similar to the variation among locations (except for juveniles). Although higher pre- cision in abundance estimates may be obtained by sampling on a small scale, our results support the notion that abalone distributions are variable over the range of spatial scales commonly used to census abalone populations. Under these circumstances stratification of sites may provide little increase in the power of population surveys to detect interannual changes in abundance. The higher proportion of vari- ance between sites in relation to locations with sam- pling juvenile abalone was not unexpected because of the smaller number of replicates and the greater degree of difficulty in observing this cryptic size class (Nash et al., 1994). The importance of adopting a sampling design that minimizes spatial effects is highlighted by the low proportion of the total varia- tion attributable to the year effect. The mixed de- sign we used provided a compromise between the advantages offered by fixed designs in reducing within-site variability and those offered by random designs in improving the precision of estimates of site means. The consequent increase in power over a random design has allowed us to economize on the number of sites sampled annually. The use of transects to estimate the abundance of abalone stocks has been criticized (McShane, 1994) as a time-consuming approach that does not allow for the patchy nature of abalone distributions over relatively small spatial scales. However, strip tran- sects provide an objective and reproducible approach to population surveys that appear to be somewhat less affected by research diver differences than are observations made against time. Our results dem- onstrate that variability among most scientific divers in the use of both radial transects and timed searches generally has a relatively small effect on variation in abalone abundance estimates. This finding sug- gests that standardization of abundance data from underwater censuses of abalone populations may be unnecessary. Because adaptive approaches to the setting of TACs require a time series of abundance indices, it is unlikely that all scientific dive team members will be retained for the required period. Thus effects of variation between diver and year will confound abundance estimates and interannual variation in abundance. The general linear model we used to analyze abalone abundance data includes the effect of variation among divers. For both methods, diver effects were smaller than site effects. Previous analyses of timed search estimates involved stan- dardization to eliminate diver effects, with the as- sumption that all differences between within-site replicates were due to diver variation (McShane and Smith, 1990). With this approach there is a risk of underestimating the variances of mean abalone abundances because variation between divers is con- founded with intrasite variation. Another problem with standardization of data to eliminate diver ef- fects prior to analysis is the resulting confidence lev- els associated with performing multiple regressions between different diver pairs. The greater the num- ber of divers the lower the level of confidence. This finding contrasts with the higher confidence level provided by the general linear models procedure, which takes account of diver x site interaction, thereby requiring only one comparison among divers. The radial transect method is a credible alterna- tive to timed searches for the conduct of underwater census of abalone. The notion that transect surveys are inefficient and time consuming ignores the im- portance of how the transects are applied. The num- ber of radial transects that can be completed in a typical day of sampling compares favorably with the number of timed searches that can be completed within the same period. When timed searches were 448 Fishery Bulletin 96(3), 1998 used to survey Victorian abalone stocks, about 20- 28 samples were completed daily; the radial transect method provides 27 samples during a typical day and requires only one-third the number of ascents per diver, reducing the risk of decompression illness (Marks and Fallowfield, 1994; Oxer, 1994). As with timed searches, the radial transect method has been successfully applied under rough sea conditions typi- cal of the exposed coastal reefs colonized by abalone. McShane and Smith ( 1990) identified the similar- ity between the timed search method and the tech- niques employed by commercial abalone divers to target aggregations of abalone. From this similarity one can reasonably infer that in some instances the abalone collection rate of research divers may exhibit (albeit over a smaller spatial scale) the hyperstability that characterizes the catch per unit of effort of com- mercial abalone divers. Indeed, in a study involving intensive experimental fishing McShane and Smith (1989) could not detect a significant difference be- tween pre- and postfishing abundance estimates of postrecruited abalone from timed searches despite a 50% decrease in CPUE. In his discussion of a patch- frequency estimation method, McShane (1995) noted the bias in timed searches when a large proportion of the search time is spent collecting abalone from dense aggregations. Radial transects target a spe- cific area (30 m2); consequently handling time does not affect the numbers of abalone counted or collected and divers are unable to target aggregations to the exclusion of more sparsely distributed abalone. Therefore it is reasonable to expect that surveys based on radial transects more accurately reflect true abalone abundance than do those based on timed searches. Although there is no doubt that dense patches of abalone are important with respect to fishing and the maintenance of profitable catch rates (McShane, 1995), recent work on H. rubra in Victoria (Officer1) has demonstrated the importance of postfishing movement and reaggregation in maintaining these patches. The Victorian studies also showed that dense patches contributed most of the variation in abun- dance estimates and that the standard deviations of samples of sparsely distributed abalone were rela- tively small. It is from these sparsely distributed abalone, and those occupying cryptic habitat, that reaggregation appears to have occurred. Further re- search is required to determine the relative impor- tance of sparsely distributed abalone in assessing the impact of fishing. Abundance estimates from transect 1 Officer, R. 1997. Marine and Freshwater Resources Institute, PO Box 114, Queenscliff, Victoria 3225, Australia. Unpubl. manuscript. sampling may prove to be more effective indicators of the size of stocks if dense aggregations are excluded from surveys. Because the application of radial transects avoids targeting some emergent abalone to the exclusion of others, there is less potential for divers to bias their sample towards larger abalone as may occur with timed searches. By measuring the length of each abalone collected, separate estimates of abundance for prerecruit (shorter than the legal minimum length) and postrecruit (recruited to the stock) aba- lone can be made. Timed searches do not necessarily permit this separation of prerecruits from post- recruits because of the potential for divers to collect larger, more accessible abalone at the expense of smaller abalone. Timed search estimates of abun- dance reported in McShane and Smith (1989) show an almost threefold increase in prerecruits after fish- ing, suggesting that abundances of prerecruits may have been underestimated prior to fishing because of the prevalence of postrecruits. Our trial stock surveys provided evidence that the transect method has the power to detect differences in abalone abundance (among locations and sites) and is robust in respect to differences in divers’ abilities to perform the method. Moreover, annual stock sur- veys based on radial transects were able to detect a significant change in abundance between consecu- tive years and had a high power of detecting smaller effect sizes over 2-3 years, thus showing their use- fulness as an effective stock monitoring tool. The radial transect method as it is currently applied for monitoring Victorian abalone stocks should have suf- ficient power to detect changes in abundance of about 10% after 3 years sampling. Monte Carlo simulations showed that surveys based on timed searches gener- ally would require several years of sampling and con- sequently would have lower power than radial transect surveys in detecting the same rate of change. The simulation results for the eastern zone, where it was predicted that only one year would be required for a 2.5% change, proved to be the only exception. However, unlike the timed search sampling of the central and western zones, all sites in the eastern zone were sampled twice annually during 1989-91. This additional sampling would be expected to in- crease substantially the statistical power to detect changes in abundance. Surveys based on radial transects have been used to monitor H. rubra stocks in the Victorian abalone fishery since 1992 with the objective of establishing a temporal series of abundance data to assist man- agers in determining sustainable levels of fishing. In monitoring Victoria’s abalone stocks, the aim is to detect an overall change in abundance across a man- Gorfine et al.: Two methods for estimating abundance of Haliotis rubra 449 agement zone against a background of increases and decreases at different sites within that zone. Al- though the Monte Carlo simulations predict that, with the use of transects, such overall changes should become detectable within several years, it is obvious that a much longer time series than three years is required to determine if significant interannual changes form part of a trend in abundance. In addi- tion to the temporal scale required for effective moni- toring of abalone stocks, there is also the issue of selecting an appropriate spatial scale over which to sample. Many abalone fisheries are managed in spa- tial units involving several hundreds of kilometres of coastline, and the relatively high cost of fishery- independent surveys ensures that compromises must be made to maximize the benefits from stock assess- ment programs. Consequently, the biologist investi- gating abalone is faced with selecting between a lim- ited number of intensive surveys of substocks over small spatial scales and extensive surveys that are less detailed but at the scale over which the fishery is managed (McShane et al., 1994). In their discus- sion of the design of surveys for abundance indices, Hilborn and Walters (1992) favored extensive sur- veys with low sampling intensities that cover the entire fishing grounds. Their preference for many sampling stations and minimal effort per station has recently been supported by Van der Meer’s ( 1997 ) study. This is the approach we are currently using to survey abalone stocks in Victoria, although we maintain rela- tively high within-site replication to allow greater pre- cision for site means. Whether power is traded off against precision depends on the extent to which infer- ences are to be drawn regarding individual sites. Acknowledgments We would like to thank staff at MAFRI who provided assistance to the Abalone Stock Assessment Program Team in the development and implementation of abalone stock assessment surveys. In particular, we thank Mike Callan, Cameron Dixon, Mark Ferrier, Steve Frlan, and Bruce Waters who performed much of the scientific diving for this study. Nik Dow as- sisted with experimental design and statistical analy- sis. Miriana Sporcic provided statistical advice on the final version of the manuscript. We also thank Dave Allen and Murray Smith, who provided their expertise and experience in conducting abalone dive surveys, and the Abalone Fishermen’s Co-operative Ltd. (Mallacoota), who provided vessel support for investigations in the eastern zone. Cameron Dixon assisted with data manipulation, Monte Carlo simu- lations, and the production of figures. Rob Day, An- thony Hart, and David Smith provided many valu- able suggestions for improving this manuscript. Fi- nally, three anonymous reviewers provided construc- tive comments on a previous draft. Literature cited Breen, P. A. 1992. A review of models used for stock assessment in aba- lone fisheries. In S. A. Shepherd, M. J. Tegner, and S. A. Guzman del Proo (eds.), Abalone of the world: biology, fish- eries and culture, p. 253-275. Blackwells, Oxford. Day, R. W., and G. P. Quinn. 1989. Comparisons of treatments after an analysis of vari- ance in ecology. Ecol. Mono. 59:433-463. Hart, A. M., and H. K. Gorfine. 1997. Abundance estimation of blacklip abalone (Haliotis rubra ) II. A comparative evaluation of catch-effort, change- in-ratio, mark-recapture and diver-survey methods. Fish. Res. 29:171-183. Hilborn, R., and C. J. Walters. 1992. Quantitative fisheries stock assessment: choice, dy- namics and uncertainty. Chapman and Hall, New York, NY, 570 p. Keppel, G. 1991. Design and analysis: a researcher’s handbook, 3rd ed. Prentice-Hall, Englewood Cliffs, NJ, 594 p. Marks, A. D., and T. L. Fallowfield. 1994. A retrospective study of decompression illness in rec- reational SCUBA divers and SCUBA instructors in Queensland. In Queensland diving industry workplace health and safety committee, proceedings of safe limits: an international dive symposium, Cairns, October 1994, p. 52-59. Worksafe, Australia. McShane, P. E. 1994. Estimating the abundance of abalone ( Haliotis spp. ) stocks — examples from Victoria and southern New Zea- land. Fish. Res. 19: 379-94. 1995. Estimating the abundance of abalone: the importance of patch size. Mar. Freshwater Res. 46:657-62. McShane, P. E,, K. H. H. Beimssen, and S. Foley. 1986. Abalone reefs in Victoria: a resource atlas. Victorian Department of Conservation, Forests and Lands, Marine Science Laboratories Technical Report Series 47, 50 p. McShane, P. E., S. F. Mercer, and J. R. Naylor. 1994. Spatial variation and commercial fishing of New Zealand abalone (Haliotis iris and H. australis). N. Z. J. Mar. Freshwater Res. 28:345-55. McShane, P. E., and M. G. Smith. 1989. Direct measurement of fishing mortality in abalone (Haliotis rubra Leach) off southeastern Australia. Fish. Res. 8:93-102. 1990. Victorian abalone monitoring: first review August 1990. Victorian Department of Conservation, Forests and Lands, Marine Science Laboratories, Program Review Se- ries 100, 47 p. 1992. Shell growth checks are unreliable indicators of age of the abalone Haliotis rubra (Mollusca: Gastropoda). Aust. J. Mar. Freshwater Res. 43:1215-19. Nash, W. J., T. L. Sellers, S. R. Talbot, A. J. Cawthorn, and W. B. Ford. 1994. The population biology of abalone ( Haliotis species) in Tasmania. I. Blacklip abalone (H. rubra) from the north 450 Fishery Bulletin 96(3), 1998 coast and islands of Bass Strait. Sea Fisheries Division, Tasmania, Tech. Rep. 48, 69 p. Oxer, H. F. 1994. Safe limits — assessing the risks. In Queensland diving industry workplace health and safety committee, Proceedings of safe limits: an international dive symposium, Cairns, October 1994, p. 86-92. Worksafe, Australia. Prince, J. D., and S. A. Guzman del Proo. 1993. A stock reduction analysis of the Mexican abalone (haliotid) fishery. Fish. Res. 16:25-49. SAS Institute, Inc. 1989. SAS/STAT user’s guide, version 6, 4th ed, vol. 2. SAS Institute, Inc., Cary, NC, 846 p. Shepherd, S. A. 1985. Power and efficiency of a research diver, with a de- scription of a rapid underwater measuring gauge: their use in measuring recruitment and density of an abalone population. In C. T. Mitchell (ed.). Diving for science, p. 263-272. Am. Acad. Underwater Science, La Jolla, CA. Van der Meer, J. 1997. Sampling design of monitoring programmes for ma- rine benthos: a comparison between the use of fixed ver- sus randomly selected stations. J. Sea Res. 37: 67-79. Winer, B. L. 1971. Statistical principles in experimental design, 2nd ed. McGraw-Hill, New York, NY, 907 p. 451 Abstract .—We used otolith ageing to describe the population dynamics of black drum, Pogonias cromis, collected over a three-year period from the Chesapeake Bay region’s commercial and recreational fisheries. Black drum average age, total length, and weight were 26 years, 109.5 cm, and 22.1 kg respectively. The oldest fish was 59 years and fish older than 50 years were present in the catch from 1990 to 1992. Growth in length slowed by age 20, whereas growth in weight did not slow until age 45. A von Bertalanffy growth function was fitted to our data (Lj= 117.3 cm, K=0.105, t0=- 2.3 yr) and was similar to that for northeast Florida, but dissimilar to that for the Gulf of Mexico. Fish grow slower but reach larger sizes in the Atlantic than in the Gulf. Estimates of instantaneous total mortality, Z, from maximum age and catch-curve analyses were low, 0.08- 0.13, indicating that fishing mortality is also low in the Chesapeake Bay re- gion. Studies to date lend support to the hypothesis that black drum from the east coast of the United States are from a common stock. The fishery of the Chesapeake Bay region is made up of old, large migrants from that larger popula- tion and should be managed accordingly. Manuscript accepted 28 October 1997. Fishery Bulletin 96:451-461 (1998). Age, growth, and mortality of black drum, Pogonias cromis, in the Chesapeake Bay region Cynthia M. Jones Applied Marine Research Laboratory and Department of Biological Sciences Old Dominion University, Norfolk, Virginia 23529-0456 E-mail address: jones@estuary.amrl.odu.edu Brian Wells Department of Biological Sciences Old Dominion University, Norfolk, Virginia 23529 Black drum, Pogonias cromis, is the largest member of the family Sciaeni- dae in the western North Atlantic Ocean. Black drum range in U.S. waters from New England south through Florida and across the northern Gulf of Mexico, with Chesa- peake Bay being near the northern end of the breeding range (Welsh and Breder, 1923; Hildebrand and Schroe- der, 1928). Black drum support im- portant recreational and commercial fisheries throughout their range in the United States. Their population abundance has been historically greater on the Florida coast than northward (Welsh and Breder, 1923), but the degree of stock unity along the east coast of the United States has not yet been determined. Black drum is migratory in the Chesapeake Bay region. Frisbie (1961) speculated that juveniles move offshore and southward in the fall. Richards (1973) reported that black drum were absent from ma- rine waters off Virginia during win- ter. Although occasionally caught inshore during winter, black drum generally move inshore to spawn in spring and offshore to overwinter in the fall. The migratory behavior of this fish complicates interpretation of the biological characteristics of the Atlantic coast fishery. Proper management of the black drum population depends on knowl- edge of their basic biology through- out their range, particularly their resilience to harvesting. Yet much is unknown about their adult life history and biology in the Chesa- peake Bay region where studies have concentrated on early life his- tory. Initial studies of eggs, larvae, and juveniles (Frisbie, 1961; Joseph et al., 1964; Richards and Castagna, 1970) failed to clarify the geographic extent of the spawning and nursery regions. A recent study by Daniel and Graves (1994) concluded that egg production of black drum had been overestimated because of misindentification and that previ- ously reported egg distributions (Jo- seph et ah, 1964) may be incorrect. Little work has been directed at adult black drum in the Chesapeake Bay region, aside from general fau- nal studies like that of Hildebrand and Schroeder (1928), and only one study is recent. Studies of early life history by Frisbie (1961) and Joseph et al. (1964) provide little informa- tion that can be used in yield mod- eling to evaluate resilience to har- vest. The only studies that provide information specifically useful for modeling include Richards (1973) and Desfosse (1987), both on age and growth. Desfosse (1987) re- ported ages of 4-15 years with 10- year-olds predominant in the catch, whereas Richards (1973) estimated 452 Fishery Bulletin 96(3), 1998 maximum age at 35 years. Unfortunately, these studies relied on scales to age black drum. Fur- thermore, Beamish and McFarlane (1983) re- ported that scales were not a reliable hard part to age older fish of many species. Hence, the use of scales for ageing black drum in the Chesa- peake Bay region may give unreliable results. Only one recent study of black drum life his- tory has focused on the Chesapeake Bay region; more work has been done in Florida and Gulf of Mexico waters. Pearson ( 1929) first described the early life stages for black drum in Texas waters. Egg and larval distributions have been reported (Jannke, 1971; Holt et al., 1985; Ditty, 1986), as well as adult distributions (Cody et al., 1978; Ross et al., 1983). Recent studies, based on otolith ageing, report maximum ages of 43 years in the northern Gulf of Mexico (Beck- man et al., 1990) and 58 years off the northeast coast of Florida (Murphy and Taylor, 1989). Al- though Pearson ( 1929) described spawning mi- grations of fish over 80 cm, most young fish show little movement between embayments (Osburn and Matlock, 1984). This paper describes fundamental biological characteristics of black drum in the Chesapeake Bay region that support stock unity of east coast fish. These data can be used as a basis for yield modeling and evaluation of black drum’s resil- ience to harvest. We present the first otolith- based age determination for Chesapeake Bay black drum, which includes characteristics of catch, growth, and mortality. We compare these life history parameters with those derived from other geographic regions. Methods Black drum (n=853) were collected March through June, 1990-92, from commercial and recreational fisheries on the eastern shore of Virginia where more than 90% of the catch is landed (Jones et al., 1990). Commercial landing sites were located at Willis Wharf, Oyster, and Bayford; recreational sites were at Cape Charles and Cherrystone Point (Fig. 1). Fish- ermen were asked for the location of their catches. Collection sites were visited daily once the first land- ings were made. Additionally, in the fall of 1990 and 1992, we obtained juveniles (n=10) from special sam- pling of pound nets near the bay mouth. Fish were sexed and measured for total length (TL), standard length (SL), total weight (TW), gonad weight (GW), girth at the preopercle (Gl), and maxi- mum girth (G2). Sagittal otoliths, dorsal spines, and Figure 1 Map of Chesapeake Bay showing Chesapeake Bay region sampling sites. fin rays were taken from each specimen. One otolith, chosen randomly from each pair, was transversely sectioned through the core on a Beuhler low-speed Isomet saw. Three sections of about 300-m thickness were mounted with Flo-texx mounting medium on a slide and read under a dissecting microscope ( lOx) with transmitted light and bright field. Dorsal spines and fin rays were processed similarly (10-40x) but sectioned perpendicular to the long axis of the growth plane, close to the base. To compare hard parts, we read random sections without knowledge of length or collection date of specimen. Ages were assigned on the basis of counts of an- nuli. We call them presumptive annuli in this paper because we have not completed validation of ages 44-59. However, otolith annuli have been validated to age 43 in the Gulf of Mexico through marginal increment analysis (Beckman et al., 1990; Fitzhugh and Beckman1), and we have recently shown corre- 1 Fitzhugh, G. R., and D. W. Beckman. 1987. Age, growth and reproductive biology of black drum in Louisiana waters. Coastal Fisheries Institute, Center for Wetland Resources, Louisiana State University, Final Report of Funded projects FY 1986—1987, 89 p. Jones et al.: Population dynamics of Pogonias cromis 453 spondences between bomb radiocarbon chronologies from the atmosphere and those from otolith cores of black drum (Campana and Jones, 1998). Average birth date was arbitrarily taken to be 1 January (Jearld, 1983). To assess ageing precision, all hard parts (n =30) were read twice by each of two readers, and agreement between and within readers was evaluated by percent agreement methods (Beamish and Fournier, 1981; Chang, 1982). Disagreements were resolved by a third reading. To evaluate changes in otolith size in relation to fish total length and age, otoliths from 300 fish ( 1990 collections; ages 0-57; 22.9-130.0 cm TL) were mea- sured for maximum length (otolith length [OL] +0.01 mm), radius along the sulcal grove (otolith radius [OR] ±0.001 mm), maximum thickness (otolith width [OWID] ±0.01 mm), and weight (otolith weight [OWT] ±0.001 g). Relation between otolith measurements and fish TL and age were evaluated by simple linear regression analysis. To evaluate growth, observed individual lengths- at-age were fitted to the von Bertalanffy growth func- tion, VBGF (Ricker, 1975), by using nonlinear regres- sion, SAS NLIN procedure DUD method (SAS, 1988). Likewise, individual weights-at-age were fitted to the VBGF. Model parameters were the following: Lx , the mean asymptotic length; Wx, the mean asymptotic weight; K and K', respectively; the Brody growth co- efficient on length and weight; and t0 and t’Q, the theoretical age at which the fish would have zero length on length and weight (Ricker, 1975). Growth curve parameters were compared between years and sexes with maximum likelihood ratio tests (Kimura, 1980). Linear regression was used to determine length- weight relationships for fish ranging from 22.9 to 130.0 cm TL and 0.6 to 49.4 kg TW. Differences be- tween sexes were tested with Rawlings’ ( 1988) tests of homogeneity of slopes and intercepts by using PROC REG in SAS (Littell et al., 1991). The hypoth- esis of isometric growth (Ricker, 1975) was tested with a Utest. Instantaneous total annual mortality rates, Z, were estimated from maximum age with Hoenig’s pooled regression equation (Hoenig, 1983), by calculating a theoretical total mortality for the entire life span fol- lowing the reasoning of Royce (1972), and with the regression method, i.e. with a catch curve combin- ing loge-transformed recreational and commercial abundance data. In the latter method, mortality es- timates were based on data from ages 21-43 and 21- 59. Younger ages were truncated because the age group at the apex of the catch curve (age 20 ) may not have been fully recruited to the fishery ( Everhart and Youngs, 1981). Older ages were truncated at the first age class (age 44) with fewer than five fish following Chapman and Robson ( 1960). Data from 1990 to 1992 were combined to minimize effects of variation in year-class strength (Robson and Chapman, 1961). The right limb of the catch curve was tested for devia- tion from linearity by analysis of variance (AN OVA). Estimates of Z were converted to total annual mor- tality rates (A=l-e 2; Ricker, 1975). All statistical analyses were performed with SAS (SAS, 1988). Rejection of the null hypothesis was based on a = 0.05, F-tests in ANCOVA were based on type-III sum of squares (Freund et al., 1986), and assumptions of linearity were checked with residual plots (Draper and Smith, 1981). Data were log10- transformed to correct for nonlinearity and hetero- geneity of variance when necessary. Log-transformed data are presented in graphs and tables in original units, unless otherwise stated. Variables that could not be normalized were compared with Wilcoxon’s two-sample test or a Kruskal-Wallis test for more than two samples, and large-sample approximate 2- scores or were reported. Results Hard part comparisons All hardparts showed regular, concentric marks that could be interpreted as annuli. However, marks were not equally clear or consistent between all hard parts. Otoliths were the clearest and most precise of the hard parts to interpret. One hundred percent of otoliths, 36.7% of dorsal spines, and 63.7% of fin rays had marks clear enough to read. Between-reader precision was 100% for otoliths, 27.3% for dorsal spines, and 47.4% for fin rays. Compared with otoliths, dorsal spines and fin rays underestimated age; this underestimation worsened with increasing age (Kruskal-Wallis distribution-free multiple com- parison test, MSD=15.81,P<0.05). Underageing was especially marked with dorsal spines. On the basis of these results and otolith growth patterns (see next section), we deemed otoliths the clearest, most reli- able hard part, and used them for all ageing. Otolith size relationships to fish size and age Black drum otoliths continue to increase in size with fish length and age, apparently throughout life. All measures of otolith size — OL, OWT, OR, OWID — were significantly and positively related to fish length and age. Although black drum otoliths continue to increase in size, the relations of various otolith sizes to fish length and age were not consistent. Relations 454 Fishery Bulletin 96(3), 1998 Age (years) Figure 3 Observed weights-at-age and fitted von Bertalanffy regres- sion line for black drum from the Chesapeake Bay region, 1990-92. between fish total length and otolith maximum length (OL= 2.69 + 0.20 TL), and otolith maximum width (OWID= 2.69 + 0.14TL), were isometric, rea- sonably linear, and therefore were useful for back- calculation of fish lengths. Other relations between total length and all relations on age were exponen- tial functions (OWT=1.72 x 10-5 TL2-66; OR- 6.02 x 10 ~3TLlA6; OL= 10.78 Age °-256; OWID=8.7l Age °-231; OWT=0.231 Age0 025) OR=0. 964 Age0 541 ). Annuli on black drum otoliths continue to be depos- ited with increasing fish size. Annuli counts were sig- nificantly and positively related to fish length (Fig. 2) and weight (Fig. 3). Fitted regression lines and data plots indicate counts continue to increase most clearly with weight. However, they also increase with length even though there is a leveling off at greater numbers of annuli. Although usually used merely to describe growth patterns, Figures 2 and 3 provide evidence — usually not stated — that otolith age is valid. Age and size compositions The Chesapeake Bay fishery generally captures old black drum. Mean age was 26 years (Fig. 4). Ages ranged from 6 to 59 years in the regularly sampled catch, but several juveniles were obtained from sam- pling pound nets. Median age in the catch was con- sistent from year to year (1990=25.0, 1991=23.0, 1992=24.0; Kruskal-Wallis %2=4.53, P>0.05) and be- Age (years) Figure 4 Overall age distribution of black drum in the Chesapeake Bay fishery, 1990-92. Juveniles were taken in the fall of 1990 and 1992 in special sampling of the pound nets. tween sexes ( S = $ =24.0; Wilcoxon 2=1.01, P>0.05). Age at the 95th percentile was 48 years, indicating that many older fish were landed. The youngest fish Jones et al.: Population dynamics of Pogonias cromis 455 Year of birth Figure 5 Year class distribution of black drum in the Chesapeake Bay fishery, 1990-92; juveniles excluded. 100 - 50 - 0 -h I I Male 75 100 125 Total length (cm) 150 Figure 6 Distribution of total lengths of black drum in the Chesa- peake Bay fishery, 1990-92. caught, apart from young-of-the-year, was age 6, and age at the 5th percentile was 16 years. No fish be- tween 1 and 5 years were found. Recruitment to the gear appears to be complete by age 20 or 21. Black drum recruitment in Chesapeake Bay is characterized by occasional, dominant year classes (Fig. 5). Exceptionally large year classes occurred in 1934 and 1942, demonstrated in an abundance that fell above the 95% confidence band of expected year class strength around the catch curve. Abundant, but not exceptional, year classes occurred in 1933, 1943, and 1968. Poor year classes, those that fell below the lower 95% confidence interval, occurred in 1939, 1946, 1951, and 1958. We lack information on re- cruitment after 1972 because black drum are not fully recruited to the bay fishery until age 21. Total length of adult black drum in Chesapeake Bay averaged 109.5 cm, ranging from 78.7 to 130.2 cm (Fig. 6). Median length (cm) in the catch was not significantly different from year to year (1990=109.2, 1991=108.0, 1992=110.5; Kruskal-Wallis %2=4.52, P>0.05), although females were slightly longer than males ( S =109.2, $ =109.5, Wilcoxonz=2.06, P< 0.05). Length at the 95th percentile was 121.9 cm, indicat- ing that many large fish were landed. Mean total weight of adult black drum in Chesa- peake Bay differed slightly between sexes and among years. Total weight of adults averaged 22.1 kg over the period 1990-92 and ranged from 11.3 to 49.4 kg (Fig. 7). Females were slightly heavier (ANOVA, F=8.23, P<0.05), probably due to their reproductive product. The difference between sexes, 1.1 kg, amounts to only 5% of average total weight. Fish in 456 Fishery Bulletin 96(3), 1 998 1990 (23.0 kg) were slightly heavier ( ANOVA, F= 4.67, P<0.05) than those landed in 1991 (21.4) and 1992 (22.3 kg). Again, the difference among years is only 7% of average total weight. Comparisons between areas and gears Black drum collected in Chesapeake Bay and coastal waters did not differ in simple biological attributes. Catches from the two areas showed no significant differences in age (bay=25.8 yr, coastal=26.8 yr, Z=- 1.21, P>0.05), total weight (bay=21.7 kg, coastal=22.4 kg, Z=-1.32, P>0.05), or total length (bay=109.5 cm, coastal=109.5 cm, Z=0.09, P>0.05). Hence, data from both areas were pooled in all other analyses. Recreational and commercial catches showed sta- tistically significant differences in total length (Z=2.13, P<0.05), but not in total weight (commer- cial=22.1 kg, recreational=22.2 kg, Z=0.76, P>0.05), or age (commercial=26.3 yr, recreational=26.9 yr, Z=1.60, P>0.05). Mean TL of the commercial catch was 109.0 cm (n=698, SE=8.7 cm), and recreational mean TL was 110.4 cm (n=166, SE=8.6 cm). Mean, median, ranges, and quantile measures of TL are almost identical for these two fisheries. Although the differences in TL are statistically significant because of large sample size, they are not biologically mean- ingful. Hence, data from these fisheries were pooled to analyze growth and mortality. Growth Observed lengths varied greatly within age (Fig. 2). Growth was rapid before 15 years of age but slowed by age 20. Lengths thereafter varied asymptotically about the mean. Black drum have achieved 58% of Lto, by age 6, when fish are first caught in the bay, and have achieved 90% by age 20, after which they are fully recruited to the gears. Apparently growth was very rapid in the first 5 years, ages absent from our collections. The VBGF equation for data pooled over the period 1990-92 is Lt = 117.3(l- e“0105u+2-3)). No differences were found in growth curve param- eters in length between the sexes (P>0.05) or years (P>0.05 ). We observed large numbers of fish at older age, permitting a good estimate for Lx ( zz =87 1 ; in- cludes juveniles, r2=0.998). However, because we observed no fish between 1 and 5 years, our estimate of K is not optimum. Parameters estimated and asymptotic standard errors are given in Table 1. Observed weights of Chesapeake Bay black drum varied greatly within age (Fig. 3). As with age-length Table 1 Summary of parameter estimates for the von Bertalanffy growth equation on total length (cm) and total weight (kg) of Chesa- peake Bay region black drum, Pogonias cromis (199092). 95% confidence intervals Parameter Estimate SE Lower Upper 117.3 0.4 116.5 118.1 K 0.105 0.003 0.099 0.111 t0 -2.3 0.2 -2.7 -1.9 37.4 1.7 34.0 40.8 K' 0.033 0.003 0.027 0.039 t 0 -1.5 0.9 -3.3 0.3 data, growth was rapid for the first 6 years. Although it slowed thereafter, fish still grew appreciably in weight until growth slowed substantially at 45 yr. Black drum have reached 22% of by age 6 when they first appear in the bay as adults, 51% of W ^ by age 20, and 78% by age 45. Hence, they grew more slowly in weight than in length. The VBGF equation for data pooled over the period 1990-92 is Wt =37.4(l- e-°-03(i+1-5)). We observed large numbers of older fish, permit- ting a good estimate for ( n= 586, r2=0.977). How- ever, because we observed no fish between ages 1 and 5, our estimate of K' is not optimum. Para- meters estimated, asymptotic standard errors, and 95% confidence intervals are given in Table 1. No differences were found in weight-growth curve parameters between the sexes (P>0.05). However, pairwise comparisons showed parameters differed between the years 1990 and 1991 (1990: 17^=57.2 kg, fG=0.0 18/yr, P0=-2.24 yr; 1991: Wm=29.8 kg, -K^O-052/yr, P0=0.06 yr, likelihood ratio test: %2= 10.54, P< 0.05). Fish captured in 1991 weighed less at older ages than in 1990 and 1992. We have no explanation for this; causes could be minor, i.e. sam- pling error, a slightly greater proportion of older fish that had completed spawning in 1992, or perhaps fish that were in worse condition in 1991. A pooled length-weight regression was developed (Fig. 8) with the equation TW = 1.01 X 10"2 TL311 (r2= 0.97; n=599; PcO.Ol). The slope of the regression line (6=3.11; SE=0.03) was significantly different from 3.00 (Gtest; t= 3.75; P<0.05), indicating allometric growth. Jones et a I.: Population dynamics of Pogonias cromis 457 Mortality Mean instantaneous total mortality rates, Z, ranged from 0.08 to 0.13. Estimates obtained from a maxi- mum observed age of 59 years, and for age truncated at the 95th percentile — 48 years, were 0.08 (A= 8%) and 0.09 (A=10%) with Hoenigs (1983) method, and 0.08 (A=8%) and 0.10 (A=10%) with Royce’s (1972) method. A regression estimate obtained from the slope of a catch curve truncated at older ages (Fig. 9) was 0.12 (A=13%) with 95% confidence intervals of 0.11 (A=12%) and 0.13 (A= 14%). This regression line did not deviate significantly from linearity ( ANOVA; F=1.18; P> 0.05). A regression estimate obtained from the slope of the full catch curve, i.e. with all older cohorts even when n< 5, was 0.09 (A=9%) with 95% confidence intervals of 0.08 (A=8%) and 0.09 (A= 10%). This regression line, too, did not deviate sig- nificantly from linearity (ANOVA; F=1.29; P> 0.05). Discussion Age determination methods We believe otoliths are the preferred, most reliable hard part to use for ageing black drum. Reasons for this include high precision and readability of otoliths, their continued growth with increase in fish size and age, the increase in the number of annuli with size, and validation over most of the life span. Otolith annuli are extremely clear and easy to read, even out to 59 annuli, and agreement between readings was absolute, 100%. In contrast, fin rays and spines often produced unreadable sections, and fewer bands were counted than on otoliths, especially at older ages. We have not yet been able to validate black drum otolith ages completely in the Chesapeake Bay re- gion with marginal increments or other analyses. However, evidence from other regions indicate that black drum otoliths are valid throughout much of their life. For example, Fitzhugh and Beckman,1 and Beckman et al. (1990) used marginal increment analysis to validate otolith annuli formation in black drum to age 43 from Louisiana. However, because most of their fish were age 5 to 27, they had to group the few fish at older ages. Our putative ages extend an additional 15 years beyond the range these au- thors described. Other evidence indicates members of the family (Sciaenidae) consistently produce an- nuli throughout life. Beckman et al. (1989) used marginal increment analysis to confirm annulus for- mation to age 37 in red drum, Sciaenops ocellatus. Ross et al. ( 1995) confirmed annuli formation in two red drum aged 38 and 40 through oxytetracycline marking of otoliths. Although we have not yet fully 458 Fishery Bulletin 96(3), 1998 validated ages from 44 to 59 years, we have found (see above) that black drum otoliths satisfy the cri- teria of Van Oosten (1929) for annuli: the number of rings increased with mean size, rings were consis- tently located on otoliths of different putative ages, and otolith radii correlated highly with putative age. Although we did not evaluate scales because of their problematic use in ageing, we believe there is direct evidence that they underestimate black drum age in the Chesapeake Bay region. Beamish and McFarlane ( 1983) documented the tendency of scales to underestimate age, especially at older ages. Richards (1973) and Desfosse (1987) estimated maxi- mum ages for Chesapeake Bay region black drum of only 35 and 10 yr, respectively, using scales. Consid- ering that size composition has not changed over the intervening years (Desfosse, 1987; Hutchinson and Rogers2 *), these ages are much younger than we ob- served. Richards should have seen maximum ages of at least 41, Desfosse at least 57. We therefore ar- gue against using scales for ageing black drum. Implications of age structure Although we had only three years of data, the long lives of black drum allowed our collections to repre- sent a history of recruitment of over 50 years — as was the case with Pereira et al. (1995) for freshwa- ter drum ,Aplodinotus grunniens. Our data show that recruitment of black drum from the Chesapeake Bay region generally appears to be low, with only occa- sional strong year classes that persist for many years, for example the 1934 and 1942 cohorts. Moreover, low average recruitment is anticipated for a species with a long reproductive lifespan (20 years at the age of capture), high batch fecundity (1-14 million eggs), and several batches in a spawning season (Wells, 1994), especially when the population remains at low abundance throughout the years. Our recruitment history of black drum also showed an absence of fish ages 1 to 5, which is consistent since at least the 1960s. There are several possible causes: 1) low abundance of black drum young that is hard to measure, 2) recent complete recruitment failure, 3) gear specificity, and 4) migration south- ward during this life stage and later northward mi- gration. We review the evidence in support of these alternatives briefly. Given its demography, this stock should have a low survival rate during the early life stages that is dif- ficult to distinguish from zero. Black drum’s poten- tial lifetime production of 60-840 million eggs requires 2 Hutchinson R., and C. Rogers. 1969. Salt water fishing in Virginia. Dep. Conserv. and Econ. Devel., Richmond, VA, 41 p. mortalities of at least 106 or 107 during larval and ju- venile stages to maintain stable populations. Hence, the high mortality seen in the field (Cowan et al., 1992) is predictable and is difficult, if not impossible, to dif- ferentiate from 100% in the field during early life. The absence of several year classes in the catch of a fishery could also signify complete recruitment fail- ure. Yet indirect evidence does not support this throughout the east coast range. Frisbie ( 1961 ) noted the virtual absence of young black drum in the bay and Richards ( 1973) stated that “black drum of more than 220 to less than 800 mm in length were not readily available . . .”. These observations correspond to cohorts from the late 1950s and 1960s which, seen retrospec- tively in modern catches, showed normal recruitment levels. Even though fish of the expected size of 1-5 year-olds are not typically seen in the bay, these young fish are not missing from the entire geographic range. Fish ages of l—k years are found in bycatch from north- east Florida (Murphy and Taylor, 1989). Hence, exami- nation of the catch argues against complete recruit- ment failure throughout the stock’s range. Fishing gear and practices used for black drum in Chesapeake Bay target large fish and may exclude small fish. The commercial fishery uses anchored and drifted gill nets with 33-cm stretch mesh, which al- low smaller fish to escape. Likewise, recreational anglers use hooks that target large fish. Hence, we can explain some of the absence of smaller fish by gear selectivity in the directed fishery. However, if these fish were present in the bay, we would expect to see them in other fisheries, but fishermen have told us that they have never seen these fish in their gear — gear such as pound nets and gill nets of 7.6- 15.2 cm (3-6 inch) stretch mesh that would retain these smaller sizes. Perhaps the strongest alternative explanation for missing 1-5 year-olds lies in the migratory patterns seen in many sciaenids. Specifically, black drum un- dergo long-range migration along the coasts of the southeast states. Although black drum have been noted as far north as Canada (Welsh and Breder, 1923; Silverman, 1979), they occur more commonly from Delaware south to Florida. Even in the Chesa- peake Bay, however, black drum are not resident year round. Frisbie (1961) suggested a southward migra- tion of young fish from Chesapeake Bay in the fall, and the same pattern of fall emigration of juveniles has been shown for Delaware Bay (Thomas and Smith, 1973). Thereafter, only larger and older fish migrate into the bay in the spring — with few younger than six years. In contrast with the Chesapeake Bay pattern, Murphy and Taylor (1989) found that only 20% of their sample from Florida included fish older than age four. Our adult catch data could be ex- Jones et al.: Population dynamics of Pogonias cromis 459 plained by the differential seasonal migration north- ward of older, larger fish from a population centered farther south, as was first postulated by Welsh and Breder (1923). Finally, two fish tagged in northeast Florida were captured about four months later at the mouth of Chesapeake Bay; thus long-range migra- tions do occur (Murphy3). In summary, migration and gear selectivity are likely explanations of the age structure of the Chesa- peake Bay fishery and the apparent absence of age 1-5 fish in this region. However, given our data, we cannot rule out local recruitment failure. Movement and exchange is supported by similar sizes-at-age in fish from Florida and Chesapeake Bay (Table 2): mean maximum length is 117.2 cm TL for Florida, 117.3 cm TL for Virginia; maximum ages along the east coast are 58 for Florida (Murphy and Taylor, 1989), 46 for Georgia (Music and Pafford, 1984), and 59 for Virginia (this study). Stock unity Several lines of evidence suggest that black drum on the U.S. east coast are from a common stock. Fish throughout the area appear to have similar growth. Von Bertalanffy growth function parameters that we estimated for the Chesapeake Bay region (1/^= 117.3 cm; 7f=0. 105/'yr; tQ=- 2.3 yr) were similar to those that Murphy and Taylor (1989) found in northeast Florida (Lm=ll7 .2 cm; if=0.124/yr; t0=- 1.29 yr). In contrast, black drum from the Gulf of Mexico grow more quickly, are smaller at age, and have a smaller maxi- mum size (Table 2). Mitochondrial DNA evidence also suggests a common stock in the western North At- lantic Ocean. No significant differences in frequency 3 Murphy, M. D. 1995. Florida Marine Research Institute, Department of Environmental Protection, 100 Eighth Ave. S.E., St. Petersburg, FL 33701. Personal commun. of mtDNA haplotypes were found in fish taken from Virginia and the east coast of Florida (Gold4). How- ever, Atlantic east coast fish differed from those sampled in the northern Gulf of Mexico (Gold et al., 1995). Finally, limited tagging data directly suggest black drum move between Chesapeake Bay and Florida (as noted previously). Implications of mortality estimates The long life we found in black drum indicates a low mortality rate for larger fish and a stock that cannot support heavy fishing pressure. Our greatest esti- mate of instantaneous total mortality, Z, converts to an annual total mortality (A) of less than 13%. As Z = F + M, natural mortality must also be less than 13%. Because black drum do not completely recruit to the fishery until age 21 in the Chesapeake Bay region, our estimates of total mortality apply to the period of 21 years ago and earlier. For our estimates to be valid today, fishing mortality on young fish must still be low throughout the stock’s range. Values of Z have important implications for management. Stocks with high M generally can withstand the highest fish- ing mortality because fishing simply takes fish that would otherwise die from natural causes. In contrast, stocks with low M (like black drum) do not have a potential for such “excess” natural mortality that can be diverted into fishing mortality (Gulland, 1983; Murphy and Taylor, 1989). Life history strategy Black drum have an unusual life history for a long- lived fish. They achieve a large size quickly — 84% of 4 Gold, J. R. 1995. Center for Biosystematics and Biodiversity, Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843. Personal commun. Table 2 Estimates of von Bertalanffy growth function parameters from various studies of black drum. Standard errors in parentheses (when available). Growth parameters Sample size Total length range (cm) Area and study L^lcm) K *0 Atlantic coast Murphy and Taylor (1989) 117.2 (0.9) 0.124(0.003) -1.29 (0.08) 397 20.2-127.5 Northeast Florida Present study 117.3 (0.4) 0.105(0.003) -2.3 (0.2) 871 22.9-130.2 Gulf of Mexico Doerzbacher et al. (1988), Texas Beckman et al. (1990), Louisiana 79.8(4.2) 110.0 0.219 (0.027) 0.038 -16.42 383 1072 20.3-99.1 460 Fishery Bulletin 96(3), 1998 their total potential growth is accomplished in only 20% of their life span. Moreover, they become sexu- ally mature at age 5-6 years (Murphy and Taylor, 1989) and appear reproductively active over a po- tential lifespan of some 60 years. Life history theory indicates that species that have an early age at first reproduction and fast growth tend to be short lived (Begon et al., 1990; Chamov, 1993). Typically, long-lived fishes grow slowly and mature late, like sturgeons (Jenkins and Burkhead, 1993) and redfishes, Sebastes (Scott and Scott, 1988; Beverton, 1992). Black drum are as long-lived as these fishes but have faster early growth and a relatively early age of first reproduction. This strategy may give black drum a capacity to main- tain population stability greater than that seen in simi- larly long-lived fishes in the presence of heavy fishing. Acknowledgments We would like to thank the eastern shore fishermen, and especially the late Clyde Bradford, for helping us to obtain samples. Collections for 1990 and part of 199 1 were made by Stephen J. Bobko, and the 1990 data were the basis of his M.S. thesis. The following people assisted in collections: Stephen Nixon, Robert Skinner, Sean Priest, and Bill Sharp. Qing Yang and Tung Quash assisted in otolith processing, and Hassan Lakkis and Qing Yang assisted with statistical analyses. This re- search was funded by a Wallop/Breaux Program Grant for Sport Fish Restoration from the U.S. Fish and Wildlife Service through the Virginia Marine Re- source Commission, Project F-88-R3-7. Literature cited Beamish, R. J., and G. A. McFarlane. 1983. The forgotten requirement for age validation in fish- eries biology. Trans. Am. Fish. Soc. 112: 35-743. Beamish, R. J., and D. A. Fournier. 1981. A method for comparing the precision of a set of age determination. Can. J. Fish. Aquat. Sci. 38:982-983. Beckman, D. W., C. A. Wilson, and A. L. Stanley. 1989. Age and growth of red drum, Sciaenops ocellatus, from offshore waters of the northern Gulf of Mexico. Fish. Bull 87:17-28. 1990. Age and growth of black drum in Louisiana waters of the Gulf of Mexico. Trans. Am. Fish. Soc. 19:537-544. Begon, M., J. L. Harper, and C. R. Townsend. 1990. Ecology: individuals, populations and communities, 2nd ed. Blackwell Scientific Publications, Boston, MA, 945 p. Beverton, R. J. H. 1992. Patterns of reproductive strategy parameters in some marine teleost fishes. J. Fish. Biol. 41 (suppl. B):137-160. Campana, S. E., and C. M. Jones. 1998. Radiocarbon from nuclear fallout as applied to the age validation of black drum, Pogonias cromis. Fish. Bull. 96(2)185-192. Chang, W. B. 1982. A statistical method for evaluating the reproducibil- ity of age determinations. Can. J. Fish. Aquat. Sci. 39:1208-1210. Chapman, D. G., and D. S. Robson. 1960. The analysis of a catch curve. Biometrics 16:354- 368. Charnov, E.L. 1993. Life history invariants. Oxford Univ. Press, New York, NY, 167 p. Cody, T. J., K. W. Rice, and C. E. Bryan. 1978. Commercial shrimp and panaeid shrimp studies, northwestern Gulf of Mexico. Texas Parks and Wildlife Dep. Coast. Fish. Proj. Report 2-276-R, 49 p. Cowan, J. H. Jr., R. S. Birdsong, E. D. Houde, J. S. Priest, W. C. Sharp, and G. B. Mateja. 1992. Enclosure experiments on survival and growth of black drum eggs and larvae in lower Chesapeake Bay. Estuaries 15:392-402. Daniel, L. B., and J. E. Graves. 1994. Morphometric and genetic identification of eggs of spring-spawning sciaenids in lower Chesapeake Bay. Fish. Bull. 92:254-261. Desfosse, J.C. 1987. Preliminary analysis of Virginia’s black drum (Po- gonias cromis) recreational and commercial fisheries. Virginia Marine Resources Commission Report 87-7, New- port News, VA, 27 p. Ditty, J.G. 1986. Ichthyoplankton in neritic waters of the northern Gulf of Mexico off Louisiana: composition, relative abundance, and seasonality. Fish. Bull. 84:935-946. Doerzbacher, J. F., A. W. Green, and G. C. Matlock. 1988. A temperature compensated von Bertalanffy growth model for tagged red drum and black drum in Texas bays. Fish. Res. 6:135-152. Draper, N. R., and H. Smith. 1981. Applied regression analysis, 2nd ed. John Wiley & Sons, New York, NY, 709 p. Everhart, W. H., and W. D. Youngs. 1981. Principles of fishery science, 2nd ed. Cornell Univ. Press, Ithaca, NY, 349 p. Freund, R. J., R. C. Littell, and P. C. Spector. 1986. SAS system for linear models, 1986 ed. SAS Insti- tute Incorporated, Cary, NC, 211 p. Frisbie, C. M. 1961. Young black drum, Pogonias cromis, in tidal fresh and brackish waters, especially in the Chesapeake and Delaware Bay area. Chesapeake Sci. 2:94-100. Gold, J. R., L. R. Richardson, C. Furman, and F. Sun. 1995. Mitochondrial DNA diversity and population struc- ture in marine fish species from the Gulf of Mexico. Can. J. Fish. Aquat. Sci. 51 (suppl. 1):205-214. Gulland, J. A. 1983. Fish stock assessment: a manual of basic methods. FAO/Wiley series on food and agriculture, vol 1. John Wiley and Sons, New York, NY, 223 p. Hildebrand, S. F., and W. C. Schroeder. 1928. The fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43:1-388. Hoenig, J. M. 1983. Empirical use of longevity data to estimate mortal- ity rates. Fish. Bull. 82( l):898-903. Holt, G. J., S. A. Holt, and C. R. Arnold. 1985. Diel periodicity of spawning in sciaenids. Mar. Ecol. Prog. Ser. 27:1-7. Jones et a I.: Population dynamics of Pogonias cromis 461 Jannke, T. E. 1971. Abundance of young sciaenid fishes in Everglades Na- tional Park, Florida, in relation to season and other variables. Univ. Miami Sea Grant Program. Sea Grant Fish Bull. 11, 128 p. Jearld, A., Jr. 1983. Age determination. In L. A. Nielsen and D. L. Johnson (ed.). Fisheries techniques, p 301-324. Am. Fish. Soc., Bethesda, MD. Jenkins, R. E., and N. M. Burkhead. 1993. Freshwater fishes of Virginia. Am. Fish. Soc., Bethesda, MD, 1079 p. Jones, C. M., K. H. Pollock, A. Ehtisham, and W. Hinkle. 1990. Assessment of the black drum recreational fishery, 1989, in Virginia. Old Dominion Univ. Res. Found. Tech. Rep. 90-2. Norfolk, VA, 100 p. Joseph, E. B., W. H. Massmann, and J. J. Norcross. 1964. The pelagic eggs and early larval stages of the black drum from Chesapeake Bay. Copeia 2:425-434. Kimura, D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull. 77:765-776. Littell, R. C., R. J. Freund, and P. C. Spector. 1991. SAS system for linear models, 3rd ed. SAS Insti- tute, Inc., Cary, NC, 329 p. Murphy, M. D., and R. G. Taylor. 1989. Reproduction and growth of black drum, Pogonias cromis, in northeast Florida. Northeast Gulf Sci. 10:127- 137. Music, J. F., and J. M. Pafford. 1984. Population dynamics and life history aspects of ma- jor marine sportfishes in Georgia’s coastal waters. Geor- gia Dep. Natural Resources, Coast. Div. Contrib. Ser. 38: 1-382. Osburn, H. R., and G. C. Matlock. 1984. Black drum movement in Texas Bays. N. Am. J. Fish. Manage. 4:523-530. Pearson, J. C. 1929. Natural history and conservation of redfish and other commercial sciaenids on the Texas coast. Bull. U.S. Bur. Fish. 44:129-214. Pereira, D. L., C. Bingham, G. R. Spangler, D. J. Conner, and P. K. Cunningham. 1995. Construction of a 110-year biochronology from sagittae of freshwater drum ( Aplodinotus grunniens ). In D. H. Secor, J. M. Dean, and S. E. Campana (eds), Recent developments on fish otolith research, p 177-196. Univ. South Carolina Press, Columbia, SC. Rawlings, J. O. 1988. Applied regression analysis: a research tool. Wads- worth, Inc., Belmont, CA, 553 p. Richards, W. E. 1973. Age, growth and distribution of black drum ( Pogonias cromis) in Virginia. Trans. Am. Fish. Soc. 102:584-590. Richards, C. E., and M. Castagna. 1970. Marine fishes of Virginia’s eastern shore (inlet and marsh, seaside waters). Chesapeake Sci. 11:235-248. Ricker, W. E. 1975. Computations and interpretation of biological statis- tics of fish populations. Bull. Fish. Res. Board Canada 191, 382 p. Robson, D. S., and D. G. Chapman. 1961. Catch curves and mortality rates. Trans. Am. Fish. Soc. 90:181-189. Ross, J. L., J. S. Pavela, and M. E. Chittenden Jr. 1983. Seasonal occurrence of black drum, Pogonias cromis, and red drum, Sciaenops ocellatus, off Texas. Northeast Gulf Sci. 6:67-70. Ross, J. L., T. M. Stevens, and D. S. Vaughan. 1995. Age, growth, mortality, and reproductive biology of red drums in North Carolina waters. Trans. Am. Fish. Soc. 124:37-54. Royce, W. F. 1972. Introduction to the fisheries sciences. Academic Press. New York, NY, 351 p. SAS (Statistical Analysis System). 1988. SAS/STAT user’s guide, 6.02 ed. SAS Institute Inc., Cary, NC, 1028 p. Scott, W. B., and M. G. Scott. 1988. Atlantic fishes of Canada. Can. Bull. Fish. Aquat. Sci. 219, 731 p. Silverman, M. J. 1979. Biological and fisheries data on black drum, Pogonias cromis (Linnaeus). NOAA, Natl. Mar. Fish. Serv. Tech. Rep. 22. Sandy Hook Laboratory, Northeast Fisheries Cen- ter. Highlands, NJ, 35 p. Thomas, D. L., and B. A. Smith. 1973. Studies of young of the black drum, Pogonias cromis, in low salinity waters of the Delaware Estuary. Chesa- peake Sci. 14:124-130. Van Oosten, J. 1929. Life history of the lake herring ( Leucichthys artedi Le Sueur) of Lake Huron as revealed by its scales with a cri- tique of the scale methods. Bull. U.S. Bur. Fish. 44:265-428. Wells, B. K. 1994. Reproductive biology of Chesapeake Bay black drum, Pogonias cromis, with an assessment of fixatives and stains for histological examination of teleost ovaries. M.S. the- sis, Old Dominion Univ., Norfolk, VA, 67 p. Welsh, W. W„ and C. M. Breder Jr. 1923. Contributions to the life histories of Sciaenidae of the eastern United States coast. Bull. U.S. Bur. Fish. 39:141-201. 462 Stock structure and movement of tagged sablefish, Anoplopoma fimbria, in offshore northeast Pacific waters and the effects of El Nino-Southern Oscillation on migration and growth Daniel K. Kimura Allen M. Shimada Franklin R. Shaw Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 7600 Sand Point Way NE, Seattle, WA 98 1 1 5-0070 E-mail address (for D.K. Kimura): Dan.Kimura@noaa.gov Abstract .-Sablefish in the north- east Pacific are found in commercial quantities from the Bering Sea, the Aleutian Islands, throughout the Gulf of Alaska, and south along the west coast of Canada and the U.S. to Baja California. Tag-recovery data support a two-population hypothesis through- out the North American range: an Alaska population ranging from the Bering Sea, including the Aleutian Is- lands and extending down through the Gulf of Alaska to northwest Vancouver Island, Canada; and a west coast popu- lation extending from southwest Van- couver Island to Baja California. Tag recoveries indicate that these two popu- lations mix off southwest Vancouver Is- land and northwest Washington, and to a lesser extent off southern Washing- ton and Oregon. Alaska sablefish, which commonly migrate over 500 n mi, are more mo- bile than west coast sablefish. Tag re- coveries for sablefish tagged in Alaska have shown strong mutual exchanges between nearly all areas. In contrast, west coast sablefish have shown far less migratory behavior. Tagging data with respect to bathymetry are difficult to interpret in both regions owing to the fact that tagging and recovery effort do not cover the full bathymetric range of adults. Results of analysis of tag-recapture growth data were consistent with pat- terns observed for several other pelagic and demersal species. That is, El Nino- Southern Ocean Oscillation events ap- peared to retard the growth of sable- fish along the west coast and to enhance growth of Alaska sablefish. The timing of recoveries from sablefish tagged off Alaska and recovered off southwest Vancouver Island and Washington- Oregon suggests that movement south correlates positively with strong up- welling in this southern area. Although sablefish trap-index surveys show a north to south cline in the percentage of large sablefish ( >60 cm, and possi- bly of Alaska origin) sampled in length frequencies along the west coast, we were unable to correlate annual fluc- tuations in these percentages with up- welling strength. Manuscript accepted 23 September 1997. Fishery Bulletin 96:462-481 (1998). Adult sablefish have a proclivity for great depths (200-1500 m) where the ocean environment is relatively constant over extensive geographic distances and consequently can be found from central Baja California along the Pacific coast through the Gulf of Alaska and Aleutian Islands, along the Bering Sea slope to the Russian coast, and down the Kam- chatka Peninsula all the way to southern Japan (OCSEAP, 1986; Allen and Smith, 1988). The marked absence of sablefish eggs and larvae from the Bering Sea above 55°N is usually attributed to an intolerance to low (<2°C) tem- peratures (OCSEAP, 1986). Sable- fish occurrence in the Bering Sea and Aleutian Islands is thought to be dependent on egg and larval drift from the northeast Pacific, princi- pally through westward transport from the Gulf of Alaska by the Alas- kan Stream (OCSEAP, 1986). Al- though sablefish occur along the coast of Asia, only adult fish (>2 yr) are found (Kodolov, 1968) and they are believed to have been recruited from the northeast Pacific stock. Ocean currents from the western Bering Sea favor dispersal of sable- fish along the Asian coast by means of the East Kamchatka and Oyashio Currents. Commercial catches of sablefish appear to be absent off Japan (Chikuni, 1985). We do not know if commercial catches of sablefish oc- cur off Russia, although Kodolov (1968) describes the adult distribu- tion of sablefish as extending from the Bering Sea all along the Kam- chatka Peninsula. In contrast, sablefish is an important commer- cial species throughout the north- east Pacific with average annual landings (1984-93) of 2150 t for the eastern Bering Sea, 2405 t for the Aleutian Islands, 22,590 t for the Gulf of Alaska, 4835 t for British Columbia, Canada, and 11,129 t for the U.S. West Coast.1 Sablefish fecundity is determi- nate and eggs are spawned in three or four batches (Hunter et al., 1989; Macewicz and Hunter, 1994); eggs are semipelagic (McFarlane and Beamish, 1992) and have been found at depths ranging from 200 to 800 m (Thompson, 1941; Kodolov, 1968; Moser et al. 1994). Age-0 (yr) larvae initially inhabit offshore sur- face waters (McFarlane and Beam- ish, 1992), move inshore during the 1 Aleutian Islands and eastern Bering Sea catches (Lowe, 1995), Gulf of Alaska catches (Fujioka, 1995), Canadian west coast catches (Saunders et al., 1995), U.S. west coast catches (Methot et al., 1994). Kimura et a I.: Stock structure and movement of Anoplopoma fimbria 463 summer of their first year (still 0-yr-old) (Rutecki and Varosi, 1997a), and usually become demersal in outer coastal waters during their second year of life at age 1 yr (Rutecki and Varosi, 1997a, 1997b). From there, it is believed that they undergo a protracted ontoge- netic migration to greater depths (Saunders et al., 1997). Adult sablefish recruit to offshore demersal fish- eries at about age 3-4 yr (Sasaki, 1985; McFarlane and Saunders, 1997). Routine trawl, longline, and trap surveys by the National Marine Fisheries Ser- vice principally cover depths from 200 to 1000 m (Parks and Shaw, 1988; Lauth et al., 1998). These surveys in Alaska, and along the west coast typically indicate sablefish are abundant at the maximum depths fished. Spawning off California occurs at depths beyond 800 m (Hunter et al., 1989). In Monterey Bay, Parrish2 measured dissolved oxygen and noted that peak longline catches occurred in the oxygen-minimum zone at around 730 m. A special longline survey targeting deeper waters (Wilkins3) and deep trap sets (Parks and Shaw, 1988) has shown that sablefish along the west coast occur to depths of up to -1500 m. Beamish et al. (1979) reported ex- ploratory catches of sablefish to depths of 2740 m. Pearcy et al. ( 1982) noted the occurrence of sablefish in the Astoria and Cascadia abyssal plains down to a depth of 2560 m. Deepwater trawl and photographic work by Wakefield (1990) showed that sablefish oc- cur as far down as -1500 m but appear to become scarce beyond this depth. Former naval dumping sites off San Francisco, at depths from 2000 to 3200 m, showed an absence of sablefish (Cailliet et al.4). A reasonable interpretation might be that although sablefish can occur at depths beyond 1500 m, sable- fish abundance can be expected to decline consider- ably beyond this depth. Because of the many interesting questions raised by their broad geographic and bathymetric distribu- tions, sablefish have become one of the most tagged demersal fish species in the northeast Pacific. On the basis of tagging studies, authors have empha- sized both the resident nature of some sablefish 2 Parrish, R. H. 1975. The relationships of oxygen concentra- tion and depth of capture with size and abundance of sablefish ( Anoplopoma fimbria) in Monterey Bay, California. Current address: Pacific Fisheries Environmental Lab., 1352 Lighthouse Ave., Pacific Grove, CA 93940. Unpubl. manuscript. 3 Wilkins, M. E. 1997. RACE Division, Alaska Fisheries Sci- ence Center, Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA 98115. Personal commun. 4 Cailliet, G. M., W. W. Wakefield, G. Moreno, and K. Rhodes. 1992. The deep-sea fauna from the proposed navy ocean dis- posal site, using trap, otter and beam trawl, and camera sled samples. Navy CLEAN Contract No. N62474-88-D-5086, pre- pared for PRC Environmental Management, San Francisco, CA, and Honolulu, HI, 69 p. (Wespestad, 1983; Beamish and McFarlane, 1983, 1988; Maloney and Heifetz, 1997) and the proclivity of other sablefish to be highly migratory (Bracken, 1983; Dark, 1983; Heifetz and Fujioka, 1991; McFar- lane and Saunders, 1997). Generally fish tagged in- shore, off the west coast of Vancouver Island, or off the U.S. west coast were more likely to be character- ized as resident and nonmigratory. In this paper we examine several hypotheses con- cerning the movement of sablefish, using recapture data gathered from offshore tagging of sablefish from throughout the northeast Pacific Ocean. The first hypothesis is that sablefish off Alaska and Canada, i.e. north of 50°N latitude, are a separate population from the sablefish south of 50°N latitude found along southern Canada and the U.S. west coast. The sec- ond hypothesis is that Alaska sablefish tend to mi- grate much farther than west coast sablefish. An- other question is whether sablefish migrate to greater depths as they become older. Such hypotheses are of practical importance because they can affect the way resource assessment survey results are interpreted. A study in which growth curves were estimated from a sablefish tag data set has been completed (Kimura et al., 1993). In that study differences in the growth curves of Alaska and west coast popula- tions were reported. In this paper, we examine the effects of El Nino-Southern Oscillation (ENSO) events on the growth of sablefish in the northeast Pacific Ocean. Finally, we examine the migration of fish tagged in Alaska to areas along the west coast. Upwelling along the west coast (42-48°N latitude) appears to be positively correlated with these migrations, and possible explanations are given for why a correla- tion exists. Materials and methods In this paper we analyze tag recoveries of the sable- fish tagging program conducted by the National Marine Fisheries Service (NMFS). Tagging occurred during surveys designed to measure relative abun- dance of sablefish throughout its range in offshore U.S. waters. From 1971 to 1993, approximately 218,255 fish were captured with trawl, trap, and longline gears and later released after having been tagged with anchor tags (Table 1). With the excep- tion of Canadian waters, sablefish were tagged from southern California to the Bering Sea. Analyses of sablefish tagged in Canadian waters have been pre- viously reported (e.g. Beamish et al. 1979; Beamish and McFarlane, 1983, 1988; McFarlane and Saunders, 1997). 464 Fishery Bulletin 96(3), 1 998 From 1971 to 1976, sablefish were tagged and re- leased through a cooperative program involving the NMFS, California Department of Fish and Game, Oregon Department of Fish and Wildlife, and re- search vessels from Russia and the Republic of Ko- rea. Trawls, traps, and longlines were the predomi- nate gears used to capture sablefish, accounting for approximately 99% of the tag releases. In Alaska from 1978 to 1993, the U.S. -Japan Co- operative Longline Survey was the primary source of sablefish tag releases; a few releases were made from NMFS trap and trawl surveys. Along the U.S. west coast from 1979 to 1993, most tagged sablefish were released through NMFS sablefish trap index surveys and the remaining fish were released dur- ing NMFS trawl surveys. Tagging from 1971 to 1976 was carried out mainly on the west coast and in southeastern Alaska (Tables 2 and 3). In 1978, tagging operations expanded into the Gulf of Alaska; in 1979, into waters off the Aleu- Table 1 Releases and recoveries by gear and region of release. Also shown are the proportion of recoveries for each combina- tion of gear and region of release. Bottom table shows tag recoveries by gear and region of recovery. Tag types were all anchor tags. Gear of West release Alaska coast Total Tag releases by gear and region of release Trawl 10,920 28,899 39,819 Trap 7248 30,722 37,970 Longline 136,077 0 136,077 Unknown 4206 183 4389 Total 158,451 59,804 218,255 Tag recoveries by gear and region of release Trawl 717 1320 2037 Trap 977 3717 4694 Longline 7987 0 7987 Unknown 166 19 185 Total 9847 5056 14,903 Recovery proportions by gear and region of release Trawl 0.0657 0.0457 0.0512 Trap 0.1348 0.1210 0.1236 Longline 0.0587 — 0.0587 Unknown 0.0395 0.1038 0.0422 Total 0.0621 0.0845 0.0683 Tag recoveries by gear and region of recovery Trawl 552 1863 2415 Trap 1331 1856 3187 Longline 6342 826 7168 Unknown 1209 603 1812 Total 9434 5148 14,582 tian Islands; and by 1982, into the eastern Bering Sea. Tag releases from 1982 onward has been fairly consistent in all areas. Tagging All sablefish were tagged with anchor tags (Floy FD- 68). The vinyl tubing (yellow, orange, or blue) on each tag was 60 mm long, 2 mm in diameter, and bore a unique number and a legend of where to return the tag. Captured sablefish were routinely put into “live” tanks supplied with fresh running sea water imme- diately after the catch was brought on board. Anes- thetics were not used. Usually within 15 minutes of completion of each haul, sablefish were dipped from Table 2 We divided the northeast Pacific into 27 areas based on the criteria described below (Fig. 1). The historic regions: eastern Bering Sea (EBS), Aleutian Islands (AI), Gulf of Alaska (GOA), and west coast (WC) essentially maintain their integrity. Nominally, Alaska includes EBS, AI, and the GOA; whereas the west coast includes only WC. Areas are defined in decimal degrees of longitude and latitude. Area number and region Area definition Longitude Latitude 1. EBS 170°E-175°E Above 55°N 2. EBS 175°E-180°E Above 55°N 3. EBS 175°W-180°W Above 55°N 4. EBS 170°W-175°W Above 55°N 5. EBS 165°W-170°W Above AK Pen. 6. EBS 160°W-165°W Above AK Pen. 7. AI 170°E-175°E 50-55°N 8. AI 175°E-180°E 50-55°N 9. AI 175°W-180°W 50-55°N 10. AI 170°W-175°W 50-55°N 11. GOA 165°W-170°W Below AK Pen. 12. GOA 160°W-165°W Below AK Pen. 13. GOA 155°W-160°W Below AK Pen. 14. GOA 150°W-155°W Below AK Pen. 15. GOA 145°W-150°W 16. GOA 140°W-145°W 17. GOA 135°W-140°W 18. GOA 52.5°N -60°N 19. GOA 50°N-52.5°N 20. WC 47.5°N-50°N 21. WC 45°N-47.5°N 22. WC 42.5°N-45°N 23. WC 40°N-42.5°N 24. WC 37.5°N-40°N 25. WC 35°N-37.5°N 26. WC 32.5°N-35°N 27. WC 30°N-32.5°N Kimura et al.: Stock structure and movement of Anoptopoma fimbria 465 466 Fishery Bulletin 96(3), 1998 the tank, placed in a padded tagging cradle, mea- sured, tagged, and released. Each anchor tag was inserted between and engaged behind the ptery- giophores of the dorsal fin. Fork length, tag number, capture depth, geographical position, and date of release were recorded for each fish. Only fish judged in viable condition were tagged. Sablefish of all sizes (20-110 cm) taken by the capture gear were tagged. Recovery A recovery program was promoted through the use of posters, news releases, and letters explaining the research and enlisting the cooperation of those who might encounter tagged sablefish, particularly fish- ermen, fish processors, and members of state fisher- ies agencies. Individuals finding a tagged fish were requested to return the tag with information about the date, location, fishing gear, depth of capture, and fish length. A reward and the release history of the tagged fish were provided to those who returned tags. By the end of 1993, 14,903 recoveries had been made (6.82%). The information garnered from these recover- ies constitutes the core of this paper. Owing to the wide- spread geographic availability of sablefish and the broad coverage of the surveys and fisheries, tagged sablefish were released and recaptured from locations throughout its range in the northeast Pacific Ocean. Data analysis Our philosophy in this paper is to present the data as they are and not to adjust the number of tag re- coveries for exploitation rates. Such adjustments require knowing the catch and population biomass, or standardized effort measures, by area. If these are incorrectly specified, the data may be further dis- torted rather than corrected. When the recovery data are being adjusted over an extremely broad geo- graphic range, where dominant gears are different and where even the same gears fish differently, such adjustments become particularly difficult. Because of the high numbers and long time series of tag returns, basic descriptive tools appear to re- veal large-scale migration and stock-structure pat- terns. Accordingly, we used basic descriptive tools such as mapping locations of tag recovery. Although such visual representations are useful, we also needed to aggregate the data so that general pat- terns could be more easily discerned. Therefore, we divided the range of sablefish (encompassing the eastern Bering Sea, the Aleutian Islands, Gulf of Alaska, and the west coast of the U.S. and Canada) into 27 areas of moderately small scale. We used large nominal regions (the eastern Bering Sea, the Aleu- tian Islands, and the Gulf of Alaska) to assist in these demarcations and also a straight line drawn down the Alaska Peninsula passing through 150°W, 60°N and 168. 5°W, 53°N and continuing on to 170°W lon- gitude. All areas west of 130°W longitude were di- vided into areas 5° wide in longitude, and areas east of 130°W longitude were divided into areas 2.5° wide in latitude (Fig. 1; Table 2). The habitat of adult sable- fish includes the 400-m depth contour shown in Fig. 1. Once these areas were described we were able to tabulate the tag recoveries in a table whose rows are the areas of release and whose columns are the ar- eas of recovery. We feel this is the single most useful piece of information that can be garnered from any study of migration based on a tagging experiment. We also characterized net movements by mapping average positions of release and recovery. The aver- age position of tagging and recovery was calculated for all fish tagged in area i, for each area i. An arrow was then drawn from the area of tagging to the area of recovery. The back end of the arrow was in area i, and the position of the arrow point provided some indication of the average movement of the tagged sablefish (i.e. the arrow points to where the tagged fish in area i go). Similarly, the average position of tagged and recovered fish was calculated for all fish recovered in area i, for each area i. The pointed end of these arrows were in area i, and the back end of the arrow gave some indication of where these fish, on average, originated (i.e. the arrow indicated where the tagged fish recovered in area i came from). Distance traveled was examined by calculating a histogram and empirical cumulative distribution function (cdf) of the distance traveled between tag- ging and recovery. Movement of tagged fish in rela- tion to depth was examined by simply tabulating the depth of tagging and the depth of recovery. Modeling ENSO growth effects Modeling approaches to analyzing ENSO growth ef- fects was difficult because the best model for this apparently simple task was unclear. Assuming an initial size at tagging of Sp a size at recovery of S2, and a time at liberty of At, the most direct model would seem to be something like S2 = S1 + Afi. How- ever, this model ignores the common sense notions that the growth increment per unit time would be expected to decrease as both Sj and At increase and suggests the model S2 = Sx + At exp (fienso + PlS1 + /324?), where Penso represents an ENSO effect, and ^ and (50 can be presumed negative. A tagged fish was assumed exposed to ENSO if it was at liberty any- time during a recognized ENSO event: 1972-73, 1976-77, 1982-83, 1986-87, and 1991-93. Kimura et al.: Stock structure and movement of Anoplopoma fimbria 467 Figure 1 Map of 27 areas (described in Table 2) used in this paper to describe analysis of sablefish tag-recovery data. Principal regions are also described in Table 2. Also shown are the 400-m depth contour, which is the approximate habitat of adult sablefish, and the oceanographic circulation that are hypothesized to affect movements of sablefish. Migration from Alaska to the west coast A well-defined sablefish migration seems to occur from Alaska to the northern portion of the west coast. To explore this phenomenon, we plotted the number of tag recoveries from Alaska, recovered along the west coast, against upwelling strength in the north- ern portion of the west coast. Length frequencies from Alaska and west coast sablefish trap index surveys (Parks and Shaw, 1983) were also used to examine the hypothesis that a substantial mixing of large ( >60 cm) sablefish of Alaska origin occurs in west coast sablefish stocks from Vancouver Island and Wash- ington State to California’s San Francisco Bay. Results Descriptive results Tag recovery rates by fishing gear type used for ini- tial capture show that the tag recovery rate for trap gear (12.4%) is twice that for longline (5.87%) and trawl (5.12%) gears (Table 1). Because these results hold for fish tagged in Alaska and off the west coast, it is likely that the much higher tag recovery rate for trap gear is due to the superior condition of and pre- sumably lower mortality rate for trap-caught fish. Tag recovery rates were remarkably consistent for all areas of tagging (Table 4). This finding would suggest that the exploitation rate for sablefish stocks are similar throughout the northeast Pacific. Recov- ery rates were higher along the west coast (8.4%) than from Alaska waters (6.0%). This is just about the magnitude that would be expected when taking into account the mixtures of gears used for tagging (i.e. if we apply both the recovery rates for trawl and trap caught fish observed from the west coast and the longline recovery rate observed in Alaska to the num- bers released in Alaska, we would arrive at the 6% re- covery rate that was actually observed in Alaska). Sablefish are long-lived; ages over 40 yr are regu- larly documented (Kimura et al., 1993) and a maxi- mum recorded age of 94 yr has been recorded at AFSC ( Anderl5). The instantaneous natural mortality rate 5 Anderl, D. M. 1997. Age and Growth Task, Alaska Fisheries Science Center, Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA 98115. Personal commun. 468 Fishery Bulletin 96(3), 1 998 used for stock assessments is M = 0.10 or less (Methot et al., 1994). Thus fish tagged in the early 1970s were still contributing to recoveries as of 1993 (Table 5). It is also notable that sablefish tagged offshore are usually not recaptured immediately; recoveries are typically protracted over many years. Results related to population structure Mapping the location of the 14,903 sablefish tag recov- eries shows a geographic continuity in sablefish distri- bution and commercial fishing effort (Fig. 2). The few offshore recoveries in midocean were located on sea- mounts. If we plot the source of these seamount recov- eries (Fig. 3), we see that they came from virtually all areas of release (see Shaw and Parks, 1997, for details). Table 6 lists the area of release against the area of recovery for the 27 areas in Figure 1. The data show that sablefish tagged off Alaska north of 50°N lati- tude have a 96.5% chance of being recaptured north of 50°N. Similarly, sablefish tagged off the west coast south of 50°N latitude have a 95.6% chance of being recovered south of 50°N. Table 6 suggests that Alaska sablefish mix in ar- eas 1-19. It is noteworthy that all Alaska areas show reciprocal migrations to all other Alaska areas. Mix- ing is also suggested from the fact that the recovery rate for fish tagged in all these areas is similar (Table 4). In contrast, west coast sablefish (areas 20-27) largely confined their movements within the area of release (64.4% of west coast tag recoveries came from the same areas as their release compared with 35.8% for Alaska). Tag recoveries indicate that Alaska sablefish on average move considerably further than west coast sablefish (Fig. 4). More than 30% of tagged Alaska Table 4 Table showing the percentage of tags recovered from releases in each area. Also shown by size category are the percentage of tagged fish moving north or west, staying put, and moving east or south. Areas are described in Table 2 and Figure 1. Size less than 57 cm Size greater than 57 cm All sizes Percent Percent Percent Percent Percent Percent Percent north or in area of east or north or in area of east or Area recovered west tagging south west tagging south 1 2 3 5.4 0.0 11.8 88.2 0.0 2.2 97.8 4 6.3 0.0 24.4 75.6 0.4 13.0 86.6 5 n 6.2 2.4 32.7 64.9 1.8 20.9 77.3 D 7 3.8 0.0 0.0 100.0 0.0 29.2 70.8 8 5.3 0.0 28.6 71.4 2.9 28.3 68.8 9 5.5 9.1 30.5 60.4 4.1 26.8 69.1 10 5.8 20.0 16.5 63.5 10.6 18.5 70.9 11 6.1 26.6 21.3 52.1 4.8 15.6 79.6 12 6.8 17.4 31.3 51.4 10.5 18.8 70.7 13 5.8 19.3 27.4 53.3 10.9 17.9 71.2 14 5.4 20.5 36.6 42.9 9.8 29.2 61.1 15 4.9 34.9 29.5 35.7 15.2 43.4 41.5 16 4.9 46.4 34.0 19.6 25.9 42.8 31.3 17 6.3 41.9 36.7 21.3 24.2 56.8 19.0 18 19 20 7.9 40.9 42.7 16.4 29.2 59.3 11.6 8.0 10.7 74.5 14.8 6.4 76.0 17.6 21 7.6 22.5 66.6 10.9 19.5 71.4 9.1 22 6.0 16.1 73.6 10.4 18.3 70.7 11.0 23 14.8 17.6 64.9 17.5 18.9 67.1 14.0 24 8.4 31.5 61.9 6.6 38.7 55.5 5.9 25 9.1 41.3 54.3 4.5 32.9 64.4 2.7 26 8.8 44.6 54.4 0.9 53.6 46.4 0.0 27 6.7 36.3 63.7 0.0 57.7 42.3 0.0 Kimura et al.: Stock structure and movement o f Anoplopoma fimbria 469 Latitude Latitude 470 Fishery Bulletin 96(3), 1998 Figure 2 Position of recovery for 14,903 sablefish recovered from 1971 to 1993. Offshore recov- eries, separated by line segments, were from seamounts. 62°00’N 54°00'N 46°00'N 38°00'N 30°00'N Longitude Figure 3 Source of seamount tag recoveries can be seen to be from almost all areas. Figure adapted from Shaw and Parks (1997). Kimura et al.: Stock structure and movement of Anoplopoma fimbria 471 472 Fishery Bulletin 96(3), 1998 sablefish were observed to move over 500 n mi from tagging to recovery, and more than 10% moved over 1,000 n mi. In contrast fewer than 10% of west coast sablefish moved over 500 n mi. For the Gulf of Alaska, there is a strong proclivity for fish tagged in the southeast (areas 16-18) to move north and west and for fish tagged in the west (ar- eas 3-14) to move east and south (Tables 4 and 6). It is convenient to define movement north or west as a recovery from an area number less than the area of tagging, and movement east or south as a recovery from an area number greater than the area of tag- ging (Fig. 1). Several authors (Fujioka et al., 1988; Heifetz and Fujioka, 1991) point out that these directional movements are most pronounced for fish tagged at less then 57 cm in southeast Alaska, and for fish tagged at greater than 57 cm in the western Gulf of Alaska. The greater intensity of these size-se- lective migrations are supported by our data (Table 4). The patterns of migration can be illustrated visu- ally by considering maps of average movement where tag recoveries are aggregated by the area of tagging or the area of recovery. Figure 5 shows the average position of tagging (tail of arrow) for each of the 27 areas for which sablefish were tagged (i.e. 23 areas) and the average position where subsequent recoveries occurred. Similarly, Figure 6 shows the average posi- tion of recoveries (head of arrow) in each area (i.e. 24 areas) where more than 2 recoveries occurred and the average position of tagging (tail of arrow) for each ag- gregation of tags. For the Alaska population, Fig- ures 5 and 6 seem to show an ex- pansion and return about the cen- ter of sablefish abundance in the central Gulf of Alaska. By this we mean that fish tagged away from the center of the Gulf of Alaska tend to be recovered toward the center of the Gulf, and fish recov- ered away from the center of the Gulf tend to have originated to- ward the center of the Gulf. Such a pattern could represent random movement throughout the Alaska range, with concentration of tag- ging and recovery effort (Tables 3 and 6) occurring in the central Gulf of Alaska. For the west coast popu- lation, the movement is more of a slight but steady northward shift whether the aggregation is by area of tagging or area of recovery. Such movement may be the result of mi- grations or may represent a biolog- ically static situation, one in which recoveries are influenced by increas- ing fishing pressure towards the north. As pointed out in the introduc- tion, sablefish inhabit a very broad bathymetry. In order to try to learn something of sablefish depth mi- grations, we tabulated depth at release for all tagged sablefish and depth at release versus depth at recovery for sablefish that were tagged and recovered (Table 7). Fish that were tagged in Alaska and recovered off the west coast, Alaska West coast Nautical miles traveled Figure 4 Histogram and cumulative distribution function (cdf) of the distance traveled (n mi) for sablefish that were tagged and recovered. Distances are straight line distances that are minimums; actual distance traveled must have been greater. Histograms are binned in 50 n mi intervals, with the 0-50 n mi bin not plotted so that more detail could be seen. For Alaska the 0-50 n mi bin count was 2744, for the west coast the 0-50 n mi bin count was 3075. Kimura et al. : Stock structure and movement of Anoplopoma fimbria 473 Figure 5 Average movement of tagged and recovered sablefish aggregated by area of tagging. Arrows point to the average location where sablefish tagged in a specified area were recovered (i.e. where fish were going). Tail of arrow originates in one of the 27 areas in which tagging occurred. Tagging did not occur in all areas. or vice versa, were excluded from Table 7. For Alaska sablefish, the depth of tag recovery generally had the same distribution regardless of the depth of tagging. For west coast sablefish, recoveries tended to be in the same depth zone as the depth of tagging. How- ever, few recoveries were made from the relatively large number of tags released in the 400-1000 m depth zone. Results related to ENSO events and migration from Alaska to the west coast Global-scale El Nino-Southern Oscillation (ENSO) events are the strongest climatic feature occurring in the Pacific Ocean. Along the west coast it weak- ens upwelling and is known to lower food availabil- ity and to stunt growth of pelagic species (Bakun, 1996). To the north, off Alaska, ENSO may have the opposite effect, enhancing growth and biological pro- ductivity (Beamish and Bouillon, 1993; Bakun, 1996). A significant negative effect on sablefish growth due to ENSO events was detected (P[ 1 1 | >3.973] < 7.2xl0-5, df=3666) off the west coast, and a significant posi- tive effect on growth was detected (P[ \ t | >2.604] < 0.01, df=6030) for Alaska sablefish (Table 8). In this analysis growth increment data were used only if tagging and recovery occurred in the same region. Residuals histograms suggested a normal pattern of errors for the model fits (Fig. 7). We found that sablefish that migrate from Alaska waters to the west coast appear to concentrate in areas off Vancouver Island, Washington, and Oregon. Area 20, between 47°30' and 50°N latitude (Fig. 1), is of particular interest because it is an area of mix- ing for Alaska and west coast sablefish (Table 6, col- umn 20). The question arises whether we can detect factors that might influence the migration from Alaska to areas 20 and 21 off the west coast. One factor that seems to have a positive correlation with this migration south is the strength of upwelling in these southern areas (Fig. 8, two-tailed, sign. a= 0.01). Length frequencies from sablefish trap index sur- veys (1978-91), aggregated over all years, strongly distinguish between fish from Alaska (areas 17 and 18) and the west coast (areas 20-27) (Fig. 9A). How- ever, the length frequencies for fish greater than 60 cm (Fig. 9B) suggest three groups: Alaska (areas 17 and 18), west coast north (areas 20-24), and west coast south (areas 25-27). The west coast north (ar- 474 Fishery Bulletin 96(3), 1 998 Figure 6 Average movement of tagged and recovered sablefish aggregated by area of recov- ery. Head of arrow lies in one of the 27 areas in which more than two tag recover- ies occurred. Rear of arrows lies in the average location where sablefish were tagged (where fish were coming from). eas 20-24) group appears to have an excess of large (>60 cm) sablefish, but we cannot be sure of their origin. We were unsuccessful in our attempt to cor- relate annual fluctuations in the abundance of large sablefish with upwelling, but this may have been due to limitations in the samples that were available on a biennial basis from the west coast north area. Discussion Comparison and interaction of Alaska and west coast populations Our tag-recovery data alone provide compelling evi- dence (Table 6) that Alaska and west coast sablefish constitute separate populations that, for practical purposes, remain largely independent. In the long- term, approximately 3.5% of Alaska fish migrate to the west coast, and approximately 4.4% of west coast fish migrate to Alaska. Short-term migration rates will be small and justify the separation of these popu- lations for fishery management purposes. However, biologically, these exchange rates are probably suffi- cient to consider sablefish a single biological popula- tion throughout its range, provided these populations are not reproductively isolated. Notice that these migration rates say nothing of net migration which is dependent on the absolute magnitude of the Alaska and west coast populations. Stock assessment scientists have long felt that sablefish constitute two distinct stocks between Alaska and the U.S. west coast, largely because their von Bertalanffy growth parameters and size-at-ma- turity differ so dramatically (Table 9; McDevitt, 1990; Kimura et al., 1993). However, early genetic studies (Tsuyuki and Roberts, 1969; Wishard and Aeber- sold, 1979; Gharrett et al., 1983) indicate that al- though the northeast Pacific may support many “somewhat discrete” populations, a mechanism for gene transfer was expected to explain the observed polymorphism. In light of our strong evidence for migratory behavior in tagged sablefish, it is not sur- prising that significant gene flow occurs throughout the range. The recapture rate for tagged sablefish appears to be remarkably similar for all areas of tagging (Table 4), especially if the differential survival rate from the fishing gears used for initial capture is taken into account. These recapture rates cannot tell us much Kimura et al. Stock structure and movement of Anoplopoma fimbria 475 Table 7 Number of sablefish tag releases and recoveries by depth category (in meters). Proportions are proportion of total releases that were recovered and the proportion of recoveries that were made at each depth category. Alaska and the west coast was divided at 50°N latitude, with tagging and recovery occurring within the respective regions. Alaska Number recovered at depth Total recoveries Depth released Total released 0-200 200-400 400-600 600-800 800-1000 0-200 11240 75 98 264 237 49 723 200-400 19334 73 365 704 568 116 1826 400-600 108131 51 268 866 856 208 2249 600-800 17572 11 41 206 312 64 634 800-1000 683 2 0 7 31 14 54 Alaska Number recovered at depth Depth Proportion released recovered 0-200 200-400 400-600 600-800 800-1000 0-200 0.064 0.104 0.136 0.365 0.328 0.068 200-400 0.094 0.040 0.200 0.386 0.311 0.064 400-600 0.021 0.023 0.119 0.385 0.381 0.092 600-800 0.036 0.017 0.065 0.325 0.492 0.101 800-1000 0.079 0.037 0.000 0.130 0.574 0.259 West coast Number recovered at depth Depth Total Total released released 0-200 200-400 400-600 600-800 800-1000 recoveries 0-200 9208 376 386 34 3 0 799 200-400 15283 309 2054 339 12 0 2714 400-600 23624 6 108 118 7 0 239 600-800 6208 1 5 7 2 0 15 800-1000 4327 0 0 0 0 0 0 West coast Number recovered at depth TT 4 * uepui released recovered 0-200 200-400 400-600 600-800 800-1000 0-200 0.086 0.471 0.483 0.043 0.004 0.000 200-400 0.178 0.114 0.757 0.125 0.004 0.000 400—600 0.010 0.025 0.452 0.494 0.029 0.000 600-800 0.002 0.067 0.333 0.467 0.133 0.000 800-1000 0.000 0.000 0.000 0.000 0.000 0.000 about the absolute exploitation rate on sablefish, because tag loss, survival of tagged fish, and report- ing rates are unknown. However, the data seem to indicate that the relative exploitation rate of substocks is roughly of similar magnitude through- out their range. Adult sablefish make two types of migrations that are quite striking in nature: migrations from the con- tinental slope across abyssal plains to seamounts, and long-range migrations along the continental slope that can extend all the way from the Bering Sea to southern California in either direction. The mechanism for accomplishing these long-range mi- grations has been discussed very little in the litera- ture, although Moser et al. (1994) suggest that mi- gration to seamounts may be midwater over the abyss. Because adult sablefish are demersal, one possible mechanism for accomplishing these long-range mi- grations is that sablefish follow the continental slope, or sea floor. Because these deep areas are in the low- oxygen zone (Bakun, 1996), this route would appear 476 Fishery Bulletin 96(3), 1998 o o 00 o — o C CD D O o O o o o C\J ...ill ii„ -20 0 20 40 Alaska residuals o o 00 o o CD O o H o o C\J ..ll [III,.. -40 -20 0 20 40 West coast residuals Figure 7 Histograms of residuals from fits to the growth model used to detect ENSO growth ef- fects suggest that residuals from the model fits are approximately normally distributed. Figure 8 Plot showing correlation of the number of sablefish tagged off Alaska and recovered off the west coast (dash) with the strength of upwelling off the west coast (solid). One year was subtracted from the year of tag recovery, allowing time for the tag recovery to take place. The Bakun upwelling indices were mean coastal summer upwellings (April-August) from 42-48°N latitude (Mason and Bakun, 1986). to be a difficult and ineffi- cient for long migrations. Also, it would seem that such migrations would take time, but several authors have noted that for tagged sablefish, there appears to be little relation between time at liberty and distance traveled (Bracken, 1983; Dark, 1983; Beamish and McFarlane, 1988). A second possibility is that sablefish are using the dominant ocean circulation patterns to redistribute themselves. The dominant circulation patterns in the northeast Pacific — the counter-clock- wise Alaska Gyre (and Alas- ka Current) and the clock- wise Central Pacific Gyre (and California Current) — suggest both a physical basis for separation and a mecha- nism for partial exchanges between Alaska and west coast stocks (Fig. 1). North of 50°N latitude, all life stages of sablefish occur in the counterclockwise rotating gyre. In con- trast, south of 50°N latitude, adults live in a northward flowing undercurrent, and pe- lagic juveniles and benthic subadults live in the southward flowing surface current (i.e. the California Current). Tag recoveries from midocean seamounts come from all areas of tagging and appear to demonstrate that sablefish routinely mi- grate on (and mix on) open ocean currents from release areas along the continental slope (Fig. 3). The Alaska Current (and Gyre) could provide a passive means for sablefish to make long migrations through- out the Gulf of Alaska. Similarly, the Califor- nia Current (flowing offshore and southerly on the surface) and the California Undercur- rent (flowing nearshore and northerly at depth) could provide the physical conveyance for sablefish to migrate passively up and down the North American west coast. Our findings concerning exchanges among areas in Alaska stocks corroborate earlier findings (Sasaki, 1985; Fujioka et al., 1988; Heifetz and Fujioka, 1991). Perhaps the relative strength and general timing of prevailing currents available to Alaska sablefish are what allow them to migrate farther and with greater frequency than west coast sablefish (Fig. 4). The lati- tudinal range of the Alaska population is only 50- 60°N, half that of the west coast population which is 30-50°N. Perhaps this narrower latitudinal range also facilitates migrations. Earlier, we described how migrations from Alaska to the west coast are reciprocated by migrations from the west coast to Alaska. Although this is true, sablefish Kimura et a!.: Stock structure and movement of Anopiopoma fimbria 477 Length (cm) Figure 9 (A) Length frequencies from trap index surveys, plotted by areas, showing the dichotomy between Alaska (areas 17 and 18) and west coast (areas 20-27) sablefish. The two distribu- tions shifted to the right are from Alaska. (B) The same length frequencies but only for sable- fish with length greater than 60 cm. Top distributions are from Alaska (areas 17 and 18), the middle distributions are from the west coast north (areas 20-24), and the lower distributions are from the west coast south (areas 25-27). migrating from Alaska to the U.S. west coast appear to concentrate in the northern area of the west coast (areas 20-22, Table 6). Methot6 recognized from fishery data that west coast sablefish varied greatly in size at age. He sug- gested that this was due to a commingling of south- ern California and Alaska “morphs.” Extending this hypothesis, tag recoveries (Table 6) suggest that area 20 contains a substantial number of fish from Alaska, but that areas farther south contain substantially fewer migrants. Similarly, length frequencies of large fish from the NMFS trap index survey (Fig. 9B) sug- gest three populations: Alaska fish from areas 17 and 18; a mixture of Alaska and smaller southern Cali- fornia fish from areas 20-24; and small southern California fish from areas 25-27. On the basis of size and age-at-depth data, adult sablefish are believed to inhabit greater depths as they grow older owing to a protracted, ontogenetic migration to greater depths (Fujioka et al., 1988; Saunders et al., 1997). Norris (1997) disagreeing with this view, conjectured that the broad bathymetric range was due to radiative evolutionary adaptation of enzyme systems to greater depths. Norris (1997) regarded the age-at-depth data as an artifact caused by varying size-selective fishing mortality with depth. He felt that reproducing populations inhabit vari- ous depth zones. Sigler et al. ( 1997 ) hypothesized that 6 Methot, R. D. 1993. Latitudinal and bathymetric patterns in sablefish growth and maturity off the U.S. West coast. Paper presented at the international symposium on the biology and management of sablefish, Alaska Fisheries Science Center, April 13-15, 1993. Table 8 Parameter estimates and their statistical significance for a tag-recovery growth model intended to detect a growth effect due to ENSO events. Data were divided between Alaska and the west coast and the ENSO parameter was included when a particular tagged fish was at liberty dur- ing any ENSO event. The model used was S2 = S j + At exp < A?nso + Pi^i + A>A), where S1 was the size of fish at the time of tagging, S2 was the size of fish at time of recovery, A, was the time elapse between tagging and recovery, and Penso represents an effect from any exposure to ENSO events. Both release and recovery were in the same region for the data used. Parameter Estimate SE f-value Alaska modeling results. Degrees of freedom for t-statistic=6,030. P 0.05781 2.2200 x 10"2 /3j -0.08307 4.8603 x 10"4 P2 -0.00031 9.9375 x 10“6 West coast modeling results. Degrees of freedom for t-statistic=3,666. P -0.12918 3.2513 x 10"2 Px -0.09636 7.4322 x 10"4 P2 -0.00036 1.4059 x 10“5 2.604 -170.910 -31.093 -3.973 -129.651 -25.853 age-at-depth data could be explained by considering the movement of sablefish to greater depths as be- ing a random walk (i.e. resulting from a sequence of random movements). From this point of view, greater age-at-depth could be viewed as an ontogenetic phe- nomenon resulting from statistical behavior. 478 Fishery Bulletin 96(3), 1998 Table 9 Von Bertalanffy growth parameters and lengths at maturity for Alaska (AK) and west coast sablefish. Growth parameters are based on age data (Kimura et al., 1993). Lengths at maturity for Alaska are from Sasaki (1985), and for the west coast are from Parks and Shaw (1983). Region Location Sex L_ (cm) K t0 (yr) Length at maturity (cm) Alaska male 70.2 0.120 -8.06 female 86.7 0.106 -6.15 Bering Sea male 65 female 67 Aleutian Islands male 61 female 65 Gulf of AK male 57 female 65 West coast male 54.7 0.472 -1.82 female 61.0 0.499 -0.81 Bodega Canyon male 52.7 female 55.3 Patton Escarpment male 54.8 female 56.3 Depth-related tagging data for sablefish are very difficult to interpret for several reasons. First of all, surveys, and therefore tag releases, did not cover the full bathymetric distribution of sablefish. Second, commercial fishing, and therefore recovery effort, did not cover the full bathymetric distribution of sable- fish. And finally, the bathymetric distribution of a species abundant across such a broad latitudinal range as sablefish, might be expected to vary by lati- tude.7 Most of our evidence supports the hypothesis that sablefish seem to have a deeper, lower limit to their distribution off the west coast, compared with their distribution off Alaska. The principal depth-related result for Alaska sable- fish is that the depth distribution of recoveries is simi- lar regardless of the depth of tagging (Table 7). Al- though the depth distribution of recoveries naturally follows recovery effort, it also suggests a certain amount of random movement in relation to depth over time. The depth-related data for west coast sablefish are even more difficult to decipher. These data show an extremely low recovery rate from fairly heavy tagging of fish from the 400-1000 m depth zone. The reason for this low recovery rate is un- known. Perhaps, the strongest conclusion to emerge 7 One reviewer noted the term “latitudinal emergence,” where deeper dwelling species are thought to have shallower upper limits on their bathymetric distributions in the more northern limits of their range. However, we did not find this term to be widely used in the literature. from our study is that we still have more to learn about sablefish depth distributions. Effects of ENSO on migration and growth It seems natural that the growth of pelagic species off the west coast, such as Pacific whiting ( Merluc - cius productus) and Pacific salmon ( Onchorynchus spp. ), would be adversely affected by ENSO events (Beamish and Bouillon, 1993; MacLellan and Saunders, 1995). It is perhaps more interesting that along the west coast, the growth and somatic condi- tion of adult rockflsh have been negatively affected by ENSO (Lenarz et al., 1995). That the observed growth of a deepwater species, such as sablefish, can be negatively affected suggests that the feeding of many demersal species along the west coast may suffer some ill effects from ENSO. In the Gulf of Alaska, ENSO growth effects are thought to be op- posite of those experienced off the west coast. Growth of salmon and groundfish species are thought to be significantly enhanced by ENSO events (Beamish and Bouillon, 1993; Bakun, 1996). Thus our findings that ENSO events retard growth of sablefish off the west coast but enhance growth of sablefish off Alaska are consistent with this literature. However, this may be the first documentation of this effect for a demer- sal species. The question arises as to why there is a positive correlation between Bakun’s upwelling index and west coast tag recoveries of Alaska fish (Fig. 8). Are Kimura et al.: Stock structure and movement of Anoplopoma fimbria 479 migrating sablefish taking advantage of enhanced feeding opportunities presented by strong upwelling or are they displaced sablefish subject to irregular but recurring episodes of ocean-forcing events and passively entrained in strengthened currents southward? Dorn (1992) noted two oceanographic phenomena that make west coast waters a highly productive re- gion: coastal upwelling and the advection of cool, zoo- plankton-rich water from the Alaska Subarctic Gyre. These phenomena could be distinguished (Roesler and Chelton, 1987) and Hollowed and Wooster (1992) have pointed out that they are positively correlated. Hollowed and Wooster (1992) noted that climate and ocean circulation in the northeast Pacific oscil- lates between a type-A pattern (i.e. weak Aleutian Low and Gulf of Alaska circulation, intensified Cali- fornia Current, and strong upwelling off the west coast) and type-B ENSO-associated conditions (i.e. intensified winter Aleutian Low and Gulf of Alaska circulation, weak California Current, and weak coastal upwelling). It is the more prevalent type-A scenario, with a strong California Current, that pos- sibly facilitates the migration of sablefish south. Thus, the strength of migration from Alaska to the west coast would be negatively correlated with ENSO events. Dorn (1995) concluded that increased north- ward movements of Pacific whiting along the west coast during El Nino years were due to intensified northward currents. Whether or not a subset of Alaska sablefish follow enhanced feeding gradients, a strong California Current could deliver them to the west coast. Because the California Current is a surface cur- rent and because there exists a natural boundary between the California Current and the Alaska Gyre, it is somewhat speculative to claim that a strong California Current plays an important role in the migration of Alaska sablefish to the west coast. Nev- ertheless, Alaska sablefish do make this migration, and the migrations coincide with an enhanced Cali- fornia Current and increased upwelling. It would seem difficult to conclude that these phenomena are totally unrelated. One difficulty in discussing sablefish migration is that we cannot be sure that the larger fish found along the northern portion of the west coast are of Alaskan origin. A partial explanation could be that the strong upwelling zone from Cape Mendocino to Point Conception in central California, as measured by Ekman transport (Parrish et al. 1981), represents a physical barrier to larger (Alaska?) sablefish. This barrier seems plausible because large sablefish be- come relatively rare in the length frequencies for the areas where upwelling intensifies. The second part of this theory would tell us whether Alaska sablefish are able to reproduce in waters off Washington and Oregon. It is possible that the gross numbers of Alaska sablefish migrating to the west coast are not substantial enough to affect length frequencies di- rectly but that the numbers may be substantial enough to affect the population genetically over a much broader geographic range. Bakun (1996) ar- gued that the Southern California Bight and the Columbia River Plume offer two locations where fish larvae can survive the disruptions of a strong up- welling environment. These locations may corre- spond to those for northerly and southerly west coast sablefish populations. However, Moser et al. ( 1994) believed that although sexually mature and spawning sablefish are encoun- tered as far south as Baja California, virtually none of their reproductive potential survives below the Southern California Bight and they do not contrib- ute to the standing stock in any meaningful way. They concluded that it is the advection of eggs and larvae from the north that ensures the stability of the popu- lation south of Pt. Conception. This hypothesis, that sablefish abundance off the west coast (and particu- larly California) is dependent on a steady “leakage” of eggs, larvae, or juvenile fish from more northern center) s) of abundance, parallels the hypothesis that Asian stocks are dependent on the migration of adult fish from northern and eastern areas of the Pacific. Although the appearance of northern fish in the mixed-zone off the Washington-Oregon coast is well established, we cannot yet quantify the net contri- bution of immigrants. Also, we found little evidence of tagged adult fish migrating from areas north of central California to southern and Baja California. However, this does not preclude Moser et al.’s ( 1994) conjecture that the southern California population is dependent on eggs and larvae advected from the north. The dependence of sablefish stocks on spawn- ing populations appears not to be well understood. Acknowledgments We thank Heather A. Parker, LTJG/NOAA, Pacific Fisheries Environmental Group, for providing data on the Bakun upwelling indices, and Sandra A. Lowe for providing catch data and a careful review. Jef- frey T. Fujioka, Anne B. Hollowed, and William T. Peterson (Northwest Fisheries Science Center, New- port Laboratory) provided extremely helpful inter- nal reviews. We would like to give special thanks to three referees who pointed out numerous errors in our manuscript. One referee in particular shared a great deal of his or her knowledge of sablefish biol- ogy and bathymetry. Because of divergent views ex- 480 Fishery Bulletin 96(3), 1998 pressed concerning sablefish biology and movement, it is important for us to state that the conclusions described here are those of the authors. Literature cited Alien, M. J., and G. B. Smith. 1988. Atlas and zoogeography of common fishes in the Bering Sea and northeast Pacific. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 66, 151p. Bakun, A. 1996. Patterns in the ocean: ocean processes and marine population dynamics. California Sea Grant College Sys- tem [La Jolla, California] in cooperation with the National Oceanic and Atmospheric Administration and Centro de Investigaciones Biologicas del Noroeste, 323 p. Beamish, R. J., and D. R. Bouillon. 1993. Pacific salmon production trends in relation to climate. Can. J. Fish. Aquat. Sci. 50:1002-1016. Beamish, R. J., C. Houle, and R. Scarsbrook. 1979. A summary of sablefish tagging and exploratory trap- ping studies conducted during 1978 by the Pacific Biologi- cal Station. Can. Data Rep. Fish. Aquat. Sci. 162, 113 p. Beamish, R. J., and G. A. McFarlane. 1983. Summary of results of the Canadian sablefish tag- ging program. In Proceedings of the international sable- fish symposium, Lowell Wakefield Fisheries Symposia Se- ries, p. 147-183. Univ. of Alaska, Alaska Sea Grant Re- port 83-8. 1988. Resident and dispersal behavior of adult sablefish (Anoplopoma fimbria) in the slope waters off Canada’s west coast. Can. J. Fish. Aquat. Sci. 45:152-164. Bracken, B. E. 1983. Sablefish migration in the Gulf of Alaska based on tag recoveries. In Proceedings of the international sablefish symposium, Lowell Wakefield Fisheries Symposia Series, p. 185-190. Univ. of Alaska, Alaska Sea Grant Report 83-8. Chikuni, S. 1985. The fish resources of the northwest Pacific. FAO Fish. Tech. Pap. 266, 190 p. Dark, T. A. 1983. Movement of tagged sablefish released at abundance index sites off southeastern Alaska, Washington, Oregon, and California during 1978-1981. In Proceedings of the international sablefish symposium, Lowell Wakefield Fish- eries Symposia Series, p. 191-207. Univ. of Alaska, Alaska Sea Grant Report 83-8. Dorn, M. W. 1992. Detecting environmental covariates of Pacific whit- ing Merluccius productus growth using a growth-increment regression model. Fish. Bull. 90:260-275. 1995. The effects of age composition and oceanographic conditions on the annual migration of Pacific whiting, Merluccius productus. Calif. Coop. Oceanic Fish. Invest. Rep. 36:97-105. Fujioka, J. T. 1995. Sablefish. In Stock assessment and fishery evalua- tion report for the groundfish resources of the Gulf of Alaska as projected for 1996, p. 4-3 to 4-17. North Pacific Fish- ery Management Council, Anchorage, AK. Fujioka, J. T., F. R. Shaw, G. A. McFarlane, T. Sasaki, and B. E. Bracken. 1988. Description and summary of the Canadian, Japanese, and U.S. joint database of sablefish tag releases and re- coveries during 1977-83. U.S. Dep. Commer. NOAA Tech. Memo. NMFS F/NWC-137, 34 p. Gharrett, A. J., M. A. Thomason, and L. N. Wishard. 1983. Progress in the study of biochemical genetics of sablefish. In Proceedings of the international sablefish symposium, Lowell Wakefield Fisheries Symposia Series, p. 143-146. Univ. of Alaska, Alaska Sea Grant Report 83-8. Heifetz, J., and J. T. Fujioka. 1991. Movement dynamics of tagged sablefish in the north- eastern Pacific. Fish. Res. 11:355-374. Hollowed, A. B., and W. S. Wooster. 1 992. Variability of winter ocean conditions and strong year classes of Northeast Pacific groundfish. ICES Mar. Sci. Symp. 195:433-444. Hunter, J. R., B. J. Macewicz, and C. A. Kimbrell. 1989. Fecundity and other aspects of the reproduction of sablefish, Anoplopoma fimbria, in central California waters. Calif. Coop. Oceanic Fish. Invest. Rep. 30:61-72. Kimura, D. K., A. M. Shimada, and S.A. Lowe. 1993. Estimating von Bertalanffy growth parameters of sable- fish Anoplopoma fimbria and Pacific cod Gadus macroceph- alus using tag-recapture data. Fish. Bull. 91:271-280. Kodolov, L. S. 1968. Reproduction of sablefish. ( Anoplopoma fimbria (PALL.)). Vop. Ikhtiol., 8(4):662-668 [In Russian, transl. in Prob. Ichthy., 8(4):531— 555.] . Lauth, R. R., M. E. Wilkins, P. A. Raymore Jr. 1998. Results of trawl surveys of groundfish resources off the West Coast upper continental slope from 1989 to 1993. U.S. Dep. Commer., NOAA Tech. Memo. NMFS- AFSC-79, 342 p. Lenarz, W. H., D.A. VenTresca, W.M. Graham, F. B. Schwing, and F. Chavez. 1995. Exploration of El Nino events and associated biologi- cal population dynamics off central California. Calif. Coop. Oceanic Fish. Invest. Rep. 36:106-119. Lowe, S. A. 1995. Sablefish. In Stock assessment and fishery evalua- tion report for the groundfish resources of the Bering Sea/ Aleutian Islands regions as projected for 1996, pp. 8-1 to 8-23. North Pacific Fishery Management Council, Anchor- age, AK. Macewicz, B. J., and J. R. Hunter. 1994. Fecundity of sablefish, Anoplopoma fimbria , from Oregon coastal waters. Calif. Coop. Oceanic Fish. Invest. Rep. 35:160-174. MacLellan, S. E., and M. W. Saunders. 1995. A natural tag on the otoliths of Pacific hake (Mer- luccius productus ) with implications for age validation and migration. In Recent developments in fish otolith re- search, p. 567-580. The Belle W. Baruch Library in Ma- rine Science Number 19. Maloney, N. E., and J. Heifetz. 1997. Movements of tagged sablefish, Anoplopoma fimbria, released in the eastern Gulf of Alaska. In Biology and management of sablefish, Anoplopoma fimbria, p. 115- 121. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 130. Mason, J. E., and A. Bakun. 1986. Upwelling index update, U.S. West, 33N-48N latitude. U.S. Dep. Commer., NOAA Tech. Memo. NMFS- SWFC-67, Southwest Fish. Sci. Cent., La Jolla, CA, 81 p. McDevitt, S. A. 1990. Growth analysis of sablefish (Anoplopoma fimbria) from mark-recapture data from the northeast Pacific. M S. the- sis, Univ. Washington, Seattle, WA, 87 p. Kimura et al : Stock structure and movement of Anoplopoma fimbria 481 McFarlane, G. A., and R. J. Beamish. 1992. Climatic influence linking copepod production with strong year classes in sablefish, Anoplopoma fimbria. Can. J. Fish. Aquat. Sci. 49:743-753. McFarlane, G. A., and M. W. Saunders. 1997. Dispersion ofjuvenile sablefish, Anoplopoma fimbria , as indicated by tagging in Canadian waters. In Biology and management of sablefish, Anoplopoma fimbria , p. 137- 150. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 130. Methot, R., R. Lauth, F. Shaw, and M. Wilkins. 1994. Assessment of the west coast sablefish stock in 1994. Pacific Fishery Management Council, 2130 SW Fifth Avenue, Suite 224, Portland, OR 97201. Moser, H. G., R. L. Charter, P. E. Smith, N. C. H. Lo, D. A. Ambrose, C. A. Meyer, E. M. Sandknop, and W. Watson. 1994. Early life history of sablefish, Anoplopoma fimbria , off Washington, Oregon, and California, with application to biomass estimation. Calif. Coop. Oceanic Fish. Invest. Rep. 35:144-159. Norris, J. G. 1997. Adaptive radiation and sablefish, Anoplopoma fim- bria. In Biology and management of sablefish, Anoplo- poma fimbria , p. 99-114. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 130. OCSEAP (Outer Continental Shelf Environmental Assessment Program). 1986. Marine fisheries: resources and environments. In D.W. Hood and S.T. Zimmerman (eds.), The Gulf of Alaska: physical environment and biological resources, p. 417- 458. Minerals Manage. Serv. MMS 86-0095, Anchorage, AK. Parks, N. B., and F. R. Shaw. 1983. Changes in relative composition of sablefish (Ano- plopoma fimbria) in coastal waters of California, 1980- 82. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F / NWC-51, 16 p. 1988. Abundance and size composition of sablefish (Anoplo- poma fimbria ) in coastal waters of California and south- ern Oregon, 1984-86, U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-125, 29 p. Parrish, R. H., C. S. Nelson, and A. Bakun. 1981. Transport mechanisms and reproductive success of fishes in the California Current. Biol. Oceanogr. 1( 2): 175— 203. Pearcy, W. G., D. L. Stein, and R. S. Carney. 1982. The deep-sea benthic fish fauna of the Northeastern Pacific Ocean on Cascadia and Tufts abyssal plains and adjoining continental slopes. Biol. Oceanogr. 1(4):375- 428. Roesler, C. S., and D. B. Chelton. 1987. Zooplankton variability in the California Current, 1951-1982. Calif. Coop. Oceanic Fish. Invest. Rep. 28:59- 96. Rutecki, T. L., and E. R. Varosi. 1997a. Migrations of juvenile sablefish, Anoplopoma fim- bria, in southeast Alaska. In Biology and management of sablefish, Anoplopoma fimbria , p. 123—130. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 130. 1997b. Distribution, age, and growth of juvende sablefish, Anoplopoma fimbria, in southeast Alaska. In Biology and management of sablefish, Anoplopoma fimbria, p. 45- 54. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 130. Sasaki, T. 1985. Studies on sablefish resources in the North Pacific Ocean. Bull. Far Seas Fish. Res. Lab. 22:1-108. Saunders, M. W., B. M. Leaman, and G. A. McFarlane. 1997. Influence of ontogeny and fishing mortality on the interpretation of sablefish, Anoplopoma fimbria, life history. In Biology and management of sablefish, Anoplopoma fimbria, p. 81-92. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 130. Saunders, M.W., G. A. McFarlane, M. Stocker, and B.M. Leaman. 1995. Sablefish. In M. Stocker and J. Fargo [eds.], Ground- fish stock assessments for the west coast of Canada in 1994 and recommended yield options for 1995, p. 223-278. Can. Tech. Rep. Fish Aquat. Sci. 2069. Shaw, F. R., and N. B. Parks, 1 997. Movement patterns of tagged sablefish, Anoplopoma fimbria, recovered on seamounts in the northeast Pacific Ocean and Gulf of Alaska. In Biology and Management of Sablefish, Anoplopoma fimbria, p. 151-158. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 130. Sigler, M. F., S. A. Lowe, and C. R. Kastelle. 1997. Area and depth differences in the age-length rela- tionship of sablefish, Anoplopoma fimbria, in the Gulf of Alaska. In Biology and Management of Sablefish, Anoplo- poma fimbria, p. 55-63. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 130. Thompson, W. F. 1941. A note on the spawning of the black cod, Anoplopoma fimbria. Copeia, No. 4:270. Tsuyuki, H., and E. Roberts. 1969. Muscle protein polymorphism of sablefish from the eastern Pacific Ocean. J. Fish. Res. Board Canada 26:2633-2641. Wakefield, W. W. 1990. Patterns in the distribution of demersal fishes on the upper continental slope off central California with studies on the role of ontogenetic vertical migration in particle flux. Ph.D. diss., Scripps Inst. Oceanography, La Jolla, CA, 281 p. Wespestad, V. G. 1983. Movement of sablefish, Anoplopoma fimbria, in the northeastern Pacific Ocean as determined by tagging ex- periments (1971-1980). Fish. Bull. 81:415-420. Wishard, L., and P. Aebersold. 1979. A biochemical genetic study of northeastern Pacific sablefish (Anoplopoma fimbria). Northwest and Alaska Fisheries Center, National Marine Fisheries Service, 2725 Montlake Boulevard East, Seattle, WA, 98112, 33 p. 482 Age and growth estimates of the bigeye thresher shark, Alopias superciliosus, in northeastern Taiwan waters Kwang-Ming Liu Po-Jen Chiang Che-Tsung Chen Department of Fishery Science National Taiwan Ocean University Keelung 20224, Taiwan E-mail address (for K.-M. Liu): kmiiu@ntou66.ntou.edu.tw Abstract.-Age and growth of Alo- pias superciliosus in waters off north- eastern Taiwan were determined from vertebral band counts on 321 specimens (214 females and 107 males) and veri- fied with a length-frequency analysis of 821 specimens (491 females and 330 males). Growth bands formed once a year according to marginal increment analysis and numbered up to 21 and 20 bands for females and males, respec- tively. The parameters of von Berta- lanffy growth equations estimated from vertebral readings were the following: asymptotic precaudal length (L^) = 224.6 cm, growth coefficient ( K) = 0.092/ yr, age at zero length (f0) = -4.21 yr for females; and = 218.8 cm, K = 0.088/ yr, t0 = -4.24 yr for males. The ages at maturity were estimated to be 12.3- 13.4 yr for females, 9-10 yr for males. The largest female aged from vertebrae was 20 yr old, the largest male 19 yr old. Length-frequency analysis supported our vertebral ageing estimates. Manuscript accepted 13 November 1997. Fishery Bulletin 96:482-491 (1998). The bigeye thresher shark, Alopias superciliosus (Lowe), is circumtrop- ical and subtropical in distribution (Compagno, 1984). It is found over the continental slope around Taiwan and is abundant at depths of 40- 100 m in waters off northeastern Tai- wan (Fig. 1). Based on daily catch data (1989-1994) from Nan Fan Ao fish market, near Suao, estimates of annual landings of bigeye threshers in number of fish are about 3300, and 220 metric tons (t) in weight, 13% of the total annual shark catch. Biological information, particu- larly on reproduction, for bigeye thresher is abundant (Nakamura, 1935; Cadenat, 1956; Bass et ah, 1975; Stillwell and Casey, 1976; Gruber and Compagno, 1981; Gil- more, 1983, 1993; Moreno and Mo- ron, 1992; Chen et al., 1997) but little is known about age and growth. Other than Gruber and Compagno’s (1981) preliminary es- timation on the age and growth of bigeye threshers, no studies have been published. Accurate age struc- ture of stocks is essential for stock assessment and fishery manage- ment. This study provides the first information concerning age and growth of bigeye threshers from waters off northeastern Taiwan. Age and growth, from vertebral band counts, were verified with length- frequency analysis. Materials and methods Thresher sharks caught in north- eastern Taiwan waters by commer- cial longlines were sampled at Nan Fan Ao fish market, from Septem- ber 1993 to October 1994. Precaudal length (PCL), fork length (FL), and total length (TL) (in cm) were measured and weighed at the fish market. Because caudal fins of bigeye thresher sharks are of- ten damaged, PCL is used through- out this paper, unless otherwise noted. After specimens were gutted, a total of 371 (245 females and 126 males) precaudal vertebrae were removed for age determination. Ver- tebrae located under the first dor- sal fin are often used for age deter- mination (Casey et al., 1985; Bran- stetter and Stiles, 1987; Chen et al., 1990); samples from two specimens ( 166 cm and 171 cm PCL) were used to compare variations in banding patterns from centra at different locations along the vertebral col- umn. Precaudal vertebrae had the same band counts as those under Liu et ai.: Age and growth of Alopias superciliosus 483 the dorsal fin and were used for age analysis in this study because these are the only vertebrae readily available at market. X-ray radiography (Cailliet et al., 1983; Cailliet, 1990; Joung, 1993) and staining of vertebrae (Stevens, 1975; Casey et al., 1985; Tanaka, 1991; Joung, 1993) are two methods commonly used to enhance the clarity of bands on vertebrae. The sil- ver nitrate technique of Stevens (1975) was tried but no increase in clarity of bands was observed; hence, we did not use the technique. The x-ray method, which is simpler and yields better clarity of bands with a Java video image analysis system, was used. A Rigaku industrial x-ray apparatus (model: radioflex 90 GSB) with Fuji industrial x-ray film (fine grain and high contrast) was used to take x-radio- graphs of vertebral centra. Banding patterns in ev- ery x-radiograph were counted three times by one of the authors during a three-month period; only those centra whose band counts were the same for at least two of the three readings were accepted for further analysis. Growth bands included fast-growth and slow-growth zones. The radii of each band and ver- tebrae were measured on a line from the nucleus (notochord) to the outer edge of each band and to the outer margin of each vertebrae with x-radiography and the Java video image analysis system (Fig. 2). The time of band formation was estimated from monthly change of marginal increment (MI) with the following equation: M/ = (R-rn)/(rn-rn_1), where R - the vertebral radius; and rn and r , = radii of the ultimate and penultimate bands, respectively. 121°E 122°E 123°E 124°E Figure 1 Sampling area for Alopias superciliosus in northeast- ern Taiwan waters. The asymptotic weight was obtained by substitut- ing into the weight-length relationship; the VBGE on weight can be expressed as Wt=W00[l-e~K{t~to))b , where Wf = the weight at age t; Wco = the asymptotic weight; and b - the exponent of the weight-length relationship. In addition, Tanaka and Mizue’s (1979) method was used to identify the ultimate band on the basis of the following stages: opaque zone, narrow translu- cent zone (the width of translucent zone is smaller than half width of opaque zone), and wide translu- cent zone (the width of translucent zone is larger than half width of opaque zone). The growth of bigeye thresher sharks was described with the von Bertalanffy growth equation (VBGE) as Lt =LM(l-e“*u-‘o)), where L( = Lm = K = the length at age t\ the asymptotic length; the growth coefficient; and the theoretical age at zero length. The parameters of VBGE were estimated from a Walford plot (Walford, 1946). Length-frequency distributions of 821 individuals (330 males, 491 females), grouped by season and sex, were analyzed with the computer package MULTIFAN (Fournier et al., 1990) to estimate the parameters of von Bertalanffy growth equations. Initial values of the L ^ and K were adapted from those obtained in the previous section. The relation between precaudal length and verte- bral radius was tested with an F-test, and the differ- ence in regression lines between sexes was tested with analysis of covariance. Results The relation between precaudal length and vertebral radius showed a slightly curved trend for both sexes (Fig. 3) and was statistically significant (P<0.05). 484 Fishery Bulletin 96(3), 1998 Figure 2 X-ray radiograph of Alopias superciliosus used for age determination. B = birth mark, R = vertebral centrum radius. Radiating bands are indicated in even years: 2, 4, 6, 8, and 10. Analysis of covariance indicated that there is signifi- cant difference in PCL-R relationships between fe- male and male (P<0.05). Therefore, the PCL-R rela- tionships were expressed by sex as follows: female: PCL = 50.84 R° 494 (r2=0.75, n= 214), male: PCL = 48.91 R° 492 (r2=0.75, n=107). Growth bands included two types of growth zones, as shown on the x-radiographs. These varied with season; usually a more calcified zone (translucent zones in Fig. 2) represented growth in summer (Casey et al., 1985; Cailliet et al., 1986; Branstetter and McEachran, 1986). The vertebral bands numbered up to 21 and 20 for females and males, respectively. Owing to a lack of smaller specimens, the mean vertebral radii of bands I— III for females and I-V for males were interpolated from the relationship between vertebral radius and number of bands. The back-calculated length at band formation was able to be obtained by substituting the mean band radii (r.) for R in the PCL-R relation- ship (Table 1). The monthly change in vertebrae (MI ) indicated that MI reached its peak in February, 0.60 for males, de- creasing thereafter to its lowest value of 0.08 in June, and increasing thereafter (Fig. 4). This trend sug- gested that a vertebral band was formed once a year Centrum radius (mm) Figure 3 Relationship between precaudal length and centrum radius of Alopias superciliosus. = female; • = male. in the period April to July. A similar trend was found for females. Following Tanaka and Mizue’s (1979) ageing criteria, we found that translucent zones (fast- growth zones) were formed between late spring and early summer, slow growth zones in winter, compa- Liu et al.: Age and growth of Alopias superciliosus 485 486 Fishery Bulletin 96(3), 1 998 rable to the results from MI. This finding indicates that one band is formed per year by bigeye thresher sharks. Chen et al. (1997) estimated PCL at birth for the bigeye thresher shark to be 73.7 cm, similar to our estimated mean length at the birth mark formation (69.6 cm). Because no growth band was found for em- bryos, the first band was assumed to be a birth mark. For comparison with other literature that reported total length (TL), PCL and FL can be converted to TL with the following equations: female: TL = 15.3 + 1.81PCL TL = 13.3 + 1.69FL male: TL = 15.1 + 1.76PCL TL = 26.3 + 1.56FL (r2=0.90, n=177); (r2=0.89, n = 177); (r2=0.88, n=6 8); (r2=0.81, ;j=6 8). The relations between body weight and total length, and body weight and precaudal length are described as follows: female: W = 6.87 x 10"5 PCL2 769 (r2=0.88, n=421); W = 1.02 x 10~5 TL2 78 (r2= 0.90, n=175); Figure 4 Monthly change of marginal increment of male Alopias superciliosus. Numbers indicate sample size and vertical bars indicate +SE. male: W = 9.93 x 10~5 PCL2 685 (r2=0.83, n=187); W = 3.73 x 10-5 TL2-57 (r2= 0.80, «=65). The parameters of VBGE estimated from band counts for females and males are given in Table 2 and the VBGE in PCL (cm) are female: Lt = 224.6 ( 1 - e-oo92«+4.2p). male: Lt = 218.8 ( 1 - e-o.o88S. auriculatus; bocaccio, S. paucispinis; blue rockfish, S. mystinus\ treefish, S. serriceps; and grass rockfish, S. rastrel- liger ) accounted for 99% of all rockfish caught. Most of these fishes were be- tween 0 and 2 years old. Catch rates for all six of these species have dropped substantially since the inception of the survey in 1977. Catch rates peaked in the early 1980s, dropped by a factor of over 100 to a low in 1984, and have gen- erally remained low through 1993. One species, blue rockfish, has not been taken since 1984. We compared our rockfish impingement data from one power station in King Harbor, Redondo Beach, with data from scuba transects conducted during the same period within King Harbor. The results of the two surveys strongly suggest that the catch rates of rockfishes by power plants reflect the abundance of these fishes surrounding the plants. We sug- gest that the reduction in the abun- dance of nearshore rockfishes in the southern California Bight is due to both decreased recruitment success, reflect- ing long-term adverse oceanographic conditions, and to overfishing. Manuscript accepted 26 January 1998. Fishery Bulletin 96:492-501 (1998). The oceanographic regime off south- ern California is extremely dy- namic, undergoing decadal changes in temperature, upwelling, and off- shore transport (Roemmich and McGowan, 1995; MacCall, 1996). Beginning in 1976, the waters of the southern California Bight (SCB) began to warm and temperatures have generally remained warmer than those of the previous four de- cades (MacCall, 1996), characteriz- ing part of a process of fluctuating water temperatures that has oc- curred for thousands of years (Soutar, 1967; Soutar and Isaacs, 1974; Baumgartner et al., 1992). This warming trend is associated with a decline in upwelling and sub- sequent decreased zooplankton bio- mass (Roemmich and McGowan, 1995; Hayward et al., 1996). In ad- dition, several studies in the SCB indicate that this change in the physical regime has led to changes in reef fish population size, commu- nity structure, and recruitment (Stephens et al., 1986; Holbrook and Schmitt, 1996). An ongoing 20-year survey at King Harbor, Redondo Beach, as well as one of 13 years du- ration at Santa Cruz Island, north- ern Channel Islands, has docu- mented population declines in many fish species (Stephens et al., 1994; Holbrook and Schmitt, 1996). In particular, the King Harbor study shows a severe decline in the abun- dances of rockfishes (genus Sebas- tes). Some species, such as the blue rockfish, S. mystinus, that were very common in the mid-1970s, vir- tually disappeared by the mid- 1980s and have remained absent (Stephens et al., 1986, 1994). Despite widespread recognition that long-term data are essential for understanding population fluctua- tions and for correlating changes in population size with environmental events, most studies of reef fishes are limited in both temporal and spatial scales. The few cases where populations have been tracked for many years are limited in spatial scale. Both the Santa Cruz Island and the King Harbor studies, al- though of relatively long duration, have tracked fish populations at only a single site. A more accurate portrayal of fish abundances would come from multisite surveys, con- ducted over a number of years. Data from the fish-impingement studies Love et al.: Declines in rockfish recruitment and populations 493 Figure 1 Location of the eight Southern California Edison coastal electric generating stations. For this sur- vey, we used fish-impingement data from the four stations indicated in bold. of the Southern California Edison Company (SCE) provide a unique long-term look at fish populations in the SCB. The impingement data are relatively long-term (17 years), are collected at fine temporal resolution (at minimum several surveys per month), and encompass much of the broad spatial distribu- tion of the SCB. This paper is the first in a series investigating changes in fish abundances over the past 17 years in the SCB using the extensive data set collected by SCE. Here, we report on patterns of rockfish (genus Sebastes ) impingement from 1977 to 1993. Although rockfish made up slightly less than 1% of all species impinged in the SCE stations, as a group they are extremely important in both recreational and com- mercial fisheries in southern California (Wine1; Ally et al.2; Barsky3). In addition to documenting changes in abundance of common species, we discuss patterns 1 Wine, V. 1979. Southern California independent sport fish- ing survey. Annual report. 3. Calif. Dep. Fish Game, Mar. Res., Admin. Rep. 79-3. 2 Ally, J. R. R., D. S. Ono, R. B. Read, and M. Wallace. 1991. Status of major southern California marine sport fish species with management , based on analyses of catch and size compo- sition data collected on board commercial passenger fishing ves- sels from 1985 through 1987. Calif. Dep. Fish and Game, Mar. Res. Div., Admin. Rep. 90-2, 376 p. 3 Barsky, K. 1996. California Department of Fish and Game, 530 E. Montecito St., Santa Barbara, CA, 93103. Personal commun. of change in abundance in relation to changes in sea surface temperature over the same time period. Materials and methods Data on fish impingement were obtained from the biological monitoring program conducted by South- ern California Edison at coastal electric generating stations throughout much of the SCB. Although fish impingement was monitored at all eight stations, we chose to analyze data from the four stations that had the most continuous sampling effort over the period 1977-93 (Fig. 1). Three of the plants, Ormond Beach, Redondo Beach, and Huntington Beach, are fossil fuel fired; San Onofre is a nuclear generating station. Al- though the timing of sampling among the stations was haphazard, samples were taken at least once and up to 11 days per month. The number of samples taken per year at each station are shown in Table 1. All power plant intakes are situated on sandy bot- tom and all are surrounded by varying amounts of rock rubble. Intake openings are at approximately equal depths (8-10 m) with the exception of Redondo Beach station where intakes are situated in slightly deeper water (14 m). The intake openings vary in their distance to the shoreline from 285 m at Redondo Beach to 965 m at San Onofre. 494 Fishery Bulletin 96(3), 1998 Table 1 Number of fish-impingement samples taken during normal operations at four coastal electrical power generating stations from 1977 to 93. Year Huntington Beach Ormond Redondo (units 7-8) San Onofre 1 San Onofre 2 San Onofre 3 1977 0 0 87 76 0 0 1978 13 13 67 90 0 0 1979 55 68 73 93 0 0 1980 98 100 103 34 0 0 1981 60 58 52 42 0 0 1982 47 43 35 19 27 0 1983 57 55 48 23 61 8 1984 57 50 58 28 65 70 1985 56 54 55 63 84 67 1986 53 57 49 28 69 101 1987 45 57 43 51 57 60 1988 41 47 25 42 57 47 1989 38 46 47 40 46 58 1990 26 37 30 36 55 43 1991 16 43 50 48 46 57 1992 37 44 38 58 78 68 1993 67 40 31 24 65 69 Total 753 799 737 629 710 648 All data presented were obtained during “normal” power plant operations. During normal operation, fishes are entrained in the cooling water and flow through conduits to the station where they are im- pinged on traveling screens in the forebay of the power plant. “Normal” fish-impingement surveys consist of counting all fishes impinged during a 24-h period of routine plant operations (e.g. on full flow days). Impinged fishes are separated from inciden- tal debris, sorted by species, identified, counted, and measured to standard length (mm). Because flow rates differed slightly among stations and years, all abundance data are presented as the number of fishes per million gallons of water pass- ing through the intake. Flow rates also varied within each station from month to month but remained con- stant during any given month. Thus, it was neces- sary to calculate monthly mean catch rates for each site. Annual mean catch rates are presented in this paper. For most analyses, annual catch is the grand mean of all monthly means across all sites for that year. For one comparison, mean annual catch rates are calculated for the Redondo station only. Results During this survey, 27,546 rockfishes, representing 16 identifiable species, were caught (Table 2). Olive rockfish ( Sebastes serranoides) were the most com- Table 2 Number of rockfish impinged during normal operations at four coastal electrical power generating stations from 1977 to 93. Species Common name Number Sebastes serranoides Olive rockfish 14,571 S. auriculatus Brown rockfish 5586 S. paucispinis Bocaccio 3262 S. mystinus Blue rockfish 1640 S. serriceps Treefish 1124 S. rastrelliger Grass rockfish 1092 Sebastes spp. Unidentified rockfish 129 S. carnatus Gopher rockfish 65 S. dalli Calico rockfish 44 S. caurinus Copper rockfish 9 S. miniatus Vermilion rockfish 8 S. atrovirens Kelp rockfish 8 S. goodei Chilipepper 5 S. melanops Black rockfish 2 S. oval is Speckled rockfish 1 S. maliger Quillback rockfish 1 S. flavidus Yellowtail rockfish 1 Total rockfish 27,546 monly taken, followed by brown rockfish {S. auricu- latus ), bocaccio ( S . paucispinis), blue rockfish {S. mystinus), treefish (S. serriceps ), and grass rockfish {S. rastrelliger). These six most abundant species represented 99% of all rockfish caught. From the lengths of the fishes sampled (Fig. 2), we determined Love et al.: Declines in rockfish recruitment and populations 495 O-Hi i t T"T"“T™T"T"T“T“T“T“i i m i r 10 30 50 70 90 110 130 150 170 190 210 230 250 270 290 310 330 350 X> 6 10 30 50 70 90 110 130 150 170 190 210 230 250 270 290 310 330 350 10 30 50 70 90 no 130 150 170 190 210 230 250 270 290 310 330 350 200 P Blue rockfish 150- 100- 50- o-*-i 1 1 1 — T"""i 1 1 1 1 1 r 10 30 50 70 90 no 130 150 170 190 210 230 250 270 290 310 330 350 Length (mm) Figure 2 Length frequencies (standard length mm ) of the four most commonly observed rockfishes in the Southern California Edison impingement studies, 1977-93. Short lines above bars indicate mean length. that most of the rockfishes impinged were between 0 and 2 years old (Love and Westphal, 1981; Love et al., 1990; Love and Johnson, in press). The life histories of these six spe- cies encompass a range of habitat preferences and behaviors. Olive rockfish and blue rockfish are near- shore, midwater fishes. Grass rock- fish and treefish are shallow-water benthic species, usually inhabiting high relief. Brown rockfish are found primarily along sand-rock interfaces or over low relief. Bocaccio are mid- water fish; the young are found in shallow water and they migrate into deeper water as subadults (Feder et al., 1974). We analyzed abundance patterns over time for all rockfishes combined and separately for the six most com- mon species. Since the inception of the survey in 1977, mean catch rates for all rockfishes combined have dropped substantially (Fig. 3). Catch rates peaked in the early 1980s, dropped precipitously (by a factor of over 100) to a low in 1984 and have generally remained low through 1993. The exception was a one-year, nearly tenfold rise between 1987 and 1988 which was due primarily to a large influx of young bocaccio, al- though olive, brown, and grass rock- fish catches also increased slightly during this period (Fig. 4). Despite very different life histo- ries, the six most abundant species showed generally similar impinge- ment patterns over time (Fig. 4). In all six species, peak impingement occurred in the late 1970s or early 1980s. Maximum catches occurred in 1977 (bocaccio), 1980 (grass, blue, and brown rockfish), or 1981 (olive rockfish and treefish). Between 1983 and 1993, with only one exception, impingement of these six species was either extremely low or nonexistent. The excep- tion occurred in 1988, when relatively large num- bers of bocaccio were impinged. The 1988 peak in bocac- cio abundance was the result of catches at only two stations, Redondo and San Onoff e, in May of that year. Between the late 1970s and early 1990s, there was a sharp drop in the amount of rockfish impingement in the SCE coastal generating stations. We believe that this pattern reflects the abundance of these fishes in nearshore waters at the time. To address this issue, we compared our impingement data for olive rockfish, blue rockfish, and bocaccio with that from visual diving surveys conducted on transects in King Harbor, Redondo Beach (Stephens4). The vi- sual surveys primarily record juvenile abundance 4 Stephens, J. 1997. Department of Biology, Occidental Col- lege, 1600 Campus Rd., Los Angeles, CA, 90041. Unpubl. data. 496 Fishery Bulletin 96(3), 1 998 Year Figure 3 Annual number of rockfishes impinged per million gallons of water pumped, for all species of rockfishes combined, 1977-93. Note log scale. See text for description of the calculation of mean. Error bars are ±1 SE. because rockfish of the three target spe- cies tend not to remain in King Harbor as adults. For this comparison, we used only the yearly catch rates from the Re- dondo Beach electrical power generating station that is closest to the King Har- bor transects. In the dive survey, all three species were common during the mid- to late-1970s (Fig. 5). However, by 1980, individuals of these species were rarely seen. Blue rockfish and bocaccio have not been observed since that time. Olive rock- fish have remained scarce, although small population increases were noted in the late 1980s and early 1990s. The pat- terns of impingement for these three spe- cies at Redondo show similar trends to that from the diver surveys. During the late 1970s, olive rockfish were commonly im- pinged but catches declined markedly by 1980 (Fig. 5). Two of the slight increases noted in the diver survey in 1985 and 1991 were reflected in the impingement study, although the magnitude of the increase is greater in the SCE data in 1985 and in the diver surveys in 1991. In the impingement survey, blue rockfish catches were relatively high in the late 1970s and declined to zero in 1981 (Fig. 5). This pattern matches precisely the pattern observed in the diver surveys, with one exception. The exception was a one- day pulse of small blue rockfish caught in a single day at the Redondo station in 1982. Regarding blue rockfish, the two data sets are in general agreement; no blue rockfish have either been observed or cap- tured since 1983. Changes in abundance of bocaccio are also similar between the impingement data and the diver surveys (Fig. 5), with the exception of a large pulse of young bocaccio impinged during a single collection in 1988. Discussion From at least the 1950s through the late 1970s, black- and-yellow, blue, gopher, and olive rockfishes, as well as young bocaccio, were important components of the inshore rocky reef community of the SCB (Limbaugh, 1955; Ebeling et al., 1980; Larson, 1980; Stephens et al., 1984, 1986; Patton et al., 1985). In particular, blue rockfish and olive rockfish were among the domi- nant species over many reefs (Carlisle et al., 1964, Ebeling et al., 1980). However, since the early 1980s, most species of rockfishes have nearly disappeared from the nearshore waters of the SCB (Stephens et al., 1994; Larson5; Schroeder6). On many of the reefs that once held substantial numbers of these species, very few rockfishes remain. Results of the fish-im- pingement surveys conducted since 1977 support the observation that several species of rockfish are less abundant now than in the late 70s and early 80s. We find it particularly compelling that for at least two species (blue rockfish and grass rockfish), not a single individual was collected in the past ten (blue rock- fish) or three (grass rockfish) years. We feel that the pattern of changing abundance in the impingement study is an accurate reflection of the pattern of change in the nearshore environment. Despite two very different survey methods, the pat- terns of rockfish abundance derived from the im- pingement data and the visual survey data from King Harbor were similar. On the gross level, the patterns for three species show amazing similarity; all have declined drastically since the late 70s and have re- mained low in the 80s and early 90s. On a finer scale, there were several large peaks in the impingement data that were not evident in the visual survey data (Fig. 5: blue rockfish in 1982, bocaccio in 1988). These differences may be due to differences in the ages of some of the fish recorded in the two surveys. Although the visual surveys mainly record juvenile blue rock- fish, olive rockfish, and bocaccio, they may occasion- ally include older individuals, whose population lev- els may be buffered from the potentially large varia- tions in recruitment by mortality. The impingement collections comprised mainly 0-2 year olds. Both of the large pulses seen in the impingement data were collections taken on a single day. 5 Larson, R. 1997. Department of Biological Sciences, San Francisco State University, San Francisco, CA 94132. Personal commun. 6 Schroeder, D. 1997. Marine Science Institute, University of California, Santa Barbara, CA 93106. Personal commun. Love et al.: Declines in rockfish recruitment and populations 497 Olive rockfish Brown rockfish Figure 4 Annual number of rockfishes impinged per million gallons of water pumped for the six most commonly impinged species, 1977-93. Note log scale. See text for description of the calculation of mean. Error bars are ±1 SE. It is likely that the decline in rock- fish abundance in the southern Califor- nia Bight was well underway by 1977 when the current impingement surveys started. The King Harbor survey began in 1974 and results showed that the abundance of blue rockfish and bocac- cio were at much higher levels from 1974 to 1977 than in later years. For example, blue rockfish were extremely abundant in King Harbor in 1974 but since 1983, not a single blue rockfish has been observed or impinged. There were also occasional impingement col- lections dating back to 1975 and these also indicate that pre-1977 rockfish catches were as high or higher than 1977 levels. These pre-1977 impinge- ment surveys were too temporally sparse to be included in the complete data set, but they are suggestive. Thus, it appears that, despite larger variation in year to year rockfish impingement during the late 70s and early 80s, there has been an overall decrease since that time which probably reflects a decrease in nearshore populations of these spe- cies throughout the southern California Bight (Stephens4; Love et al., in press). Although none of the previously men- tioned nearshore studies conducted dur- ing the 1950s-1970s were designed to focus on rockfish year-class strength, it appears from these surveys that the widespread abundance of nearshore rockfishes before the early 1980s was generated by a series of strong year classes. During this same period, South- ern California trawl studies of two other deeper-dwelling rockfish species im- plied a similar phenomenon (Mearns et al., 1980). It is likely that recent de- clines in abundance of these nearshore species were due to a decade-long se- ries of very poor year classes. What might be responsible for this poor rockfish recruitment? The succes- sion of poor year classes off southern California is likely linked to decade-long changes in oceanographic conditions. Most rockfishes are primatively vivipa- rous (Boehlert et al., 1987) and rockfish larvae are found primarily in the upper mixed layer (Ralston and Howard, 1995). After ap- proximately one month, rockfish larvae metamor- phose into pelagic juveniles that spend 3-6 months in the water column as plankton and micronekton 498 Fishery Bulletin 96(3), 1998 (Love et al., 1991). It is during the larval and pelagic juvenile stages that rockfish year-class strength is determined (Ralston and Howard, 1995) and up- welling appears to be a particularly important fac- tor affecting survival during these stages. Years with intermediate levels of upwelling seem to correspond with strong year classes (Ainley et al., 1993). Since the late 1970s, waters in the southern Cali- fornia Bight have warmed approximately 1.5°C and upwelling has declined to approximately one half of the levels observed in the late 70s and early 80s (Norton and Crooke, 1994). In turn, this has led to reduced zooplankton production (Roemmich and McGowan, 1995) and an apparent reduction in the larval and juvenile survivorship of many marine fish species (Holbrook and Schmitt, 1996). The current low-upwelling conditions are probably part of a long- term alternation of warm- and cold-water regimes that extend along much of the northeast Pacific (MacCall, 1996). Poor pelagic juvenile rockfish sur- vivorship of many species also extends at least into central and northern California (Ralston7). Juvenile 7 Ralston, S. 1996. National Marine Fisheries Service, Tiburon Laboratory, 3150 Paradise Dr., Tiburon, CA, 94920. Personal commun. bocaccio recruitment indices for central- northern California (Ralston et al., 1996) show a steady decline similar to the decline we observed for bocaccio in the impingement data. There has also been a sharp decline in the numbers of both subadult and adult inshore rockfishes in southern California. These reductions are due both to natural mortality and heavy fishery exploitation. Unlike many deeper- water and more northerly species, the inshore rockfishes of California have relatively short life spans, often less than perhaps 25 years (Miller and Geibel, 1973; Love and Westphal, 1981; Love and Johnson, in press). Thus, even without fishing pressure, the number of adults of many of these species would tend to decrease relatively rapidly dur- ing extended periods of poor reproduc- tion. In addition, these are also very heavily fished species. Until the early 1990s, most of the catch was made by recreational fishermen (Wine1; Ally et al.2). Beginning in the early 1990s, a live-fish commercial fishery developed that targets shallow-water species (Barsky3). The sharp decreases in inshore rockfish popula- tions are also mirrored by similar declines in the populations of deeper-water species. Between 1980 and 1996, it appears that there was also a substan- tial decrease in the numbers of deeper-water rock- fishes off southern California. An analysis of the rec- reational rockfish catch in southern California since 1980 shows steep declines in catches of most of the important species (Love et al., in press). As an ex- ample, catches of chilipepper rockfish (S. goodei ) have declined to 0.5%, bocaccio to 1%, and widow rockfish (S. entomelas ) to 1.25% of 1980 levels. These declines are almost certainly reflective of lowered abundances of many species rather than a shift in fishing em- phasis by recreational vessels. It is generally held that although reef fish popula- tions exhibit large temporal variations, even under- going local extinctions, external sources of new young will eventually provide new recruits. Referring spe- cifically to the California coast, MacCall (1996) specu- lated on the establishment of southern marine spe- cies, such as the Pacific seahorse ( Hippocampus ingens), at the northern ends of their ranges off Cali- fornia. He noted that these animals probably estab- lish themselves during warm-water periods and, in the absence of fishing pressure, may survive colder Love et al.: Declines in rockfish recruitment and populations 499 Blue rocktish Figure 5 A comparison of impingement and diver observations of olive rockfish, blue rock- fish and bocaccio from King Harbor, Redondo Beach (1974—94, Stephens4) and the SCE Redondo Beach station (1977-94, this study). periods if the adults suffer low mortality. A similar phenomenon has been described for California sheephead [Semicossyphus pulcher) which exhibit episodic recruitment related to anomalous events in cur- rent flow (Cowen, 1985). The reverse of this phenomena has occurred for at least some of the inshore rockfish species, par- ticularly blue rockfish and olive rockfish. In the SCB, both species are near the southern end of their usual geographic ranges and large-scale recruitment may occur only during cold-water cycles, as occurred during the 1960s and early 1970s. During this current warm-water period, recruitment waned and the adult population was expected to decline slowly. However, the continuing fishing pressure on the populations accel- erated this process. On the basis of current flow in the SCB, it is likely that even during periods of successful recruitment, many of the rockfishes in southern California are generated from southern California adults (Reid et al., 1958; Schwartzlose, 1963; Browne, 1993). If true, the sharp drop in the adult populations of many rockfishes is particularly troubling and raises the issue of recruitment overfishing. This is a particularly strong possibility be- cause there is little incentive for rec- reational anglers to decrease fish- ing activities on shallow reefs. These rockfishes are caught as part of a species assemblage that in- cludes not only various rockfishes, but also such species as kelp bass (Paralabrax clathratus ), Pacific bar- racuda (Sphyraena argentea), Cali- fornia sheephead ( Semicossyphus pulcher ), and ocean whitefish ( Caulolatilus princeps). As long as even moderate numbers of any recreational reef species are taken, recreational vessels will con- tinue to fish on rockfish-depleted reefs and continue to reduce already low numbers of rockfish. Because vir- tually all inshore reefs in southern California are heavily fished, successful recruitment will likely continute to be hazardous. Acknowledgments This work was conducted through a cooperative agreement with the Biological Resources Division, U.S. Geological Survey, contract number 1445-CA- 0995-0386. As usual Lyman Thorsteinson was very supportive of our work. We would like to thank John Stephens for providing the King Harbor survey data. 500 Fishery Bulletin 96(3), 1998 We would like to thank Donna Schroeder and two anonymous reviewers for helpful comments on ear- lier versions of the manuscript. Literature cited Ainley, D. G., W. J. Sydeman, R. H. Parrish, and W. H. Lenarz. 1993. Oceanic factors influencing distribution of young rock- fish ( Sebastes ) in central California: a predator’s perspective. Calif. Coop. Oceanic Fish. Invest. Rep. 34:133-139. Baumgartner, T. R., A. Soutar, and V. Ferreira-Bartrina. 1992. Reconstruction of the history of Pacific sardine and northern anchovy populations over the past two millennia from sediments of the Santa Barbara Basin, California. CalCOFI Rep. 33:24-40. Boehlert, G. W., M. Kusakari, and J. Yamada. 1987. Reproductive mode and energy costs of reproduction in the genus Sebastes. In Proceedings of the international rock- fish symposium; Oct. 1986, Anchorage, Alaska, p. 143-152. Alaska Sea Grant Rep. 87-2, Univ. Alaska, Fairbanks, AK. Browne, D. R. 1993. Understanding the oceanic circulation in and around the Santa Barbara Channel. Proc. 8th Ann. Information Trans. Meeting, MMS 93-0058, p. 95-108. U.S. Dep. of Interior, Washington, D. C. Carlisle, J. G., Jr., C. H. Turner, and E. E. Ebert. 1964. Artificial habitat in the marine environment. Calif. Dep. Fish Game, Fish. Bull. 124. Cowen, R. K. 1985. Large scale pattern of recruitment by the labrid, Semicossyphus pulcher: causes and implications. J. Mar. Res. 43:719-742. Ebeling, A. W., R. J. Larson, W. S. Alevizon, and R. N. Bray. 1980. Annual variability of reef-fish assemblages in kelp for- ests off Santa Barbara, California. Fish. Bull. 78:361-377. Feder, H. M., C. H. Turner, and C. Limbaugh. 1974. Observations on fishes associated with kelp beds in southern California. Calif. Dep. Fish Game, Fish Bull. 160. Hayward, T. L., S. L. Cummings, D. R. Cayan, F. P. Chavez, R. J. Lynn, A. W. Mantyla, P. P. Niiler, F. B. Schwing, R. R. Veit, and E. L. Venrick. 1996. The state of the California Current in 1986-1996: Continuing declines in macrozooplankton biomass during a period of nearly normal circulation. Calif. Coop. Oce- anic Fish. Invest. Rep. 37:22-37. Holbrook, S. J., and R. J. Schmitt. 1996. On the structure and dynamics of temperate reef fish assemblages — are resources tracked? In M. L. Cody and J. A. Smallwood (eds.). Long-term studies of vertebrate communities, p. 19-48. Academic Press, San Diego, CA. Larson, R. J. 1980. Competition, habitat selection, and the bathymetric segregation of two rockfish ( Sebastes ) species. Ecol. Monogr. 50:221-239. Limbaugh, C. 1955. Fish life in the kelp beds and the effects of kelp har- vesting. Univ. Calif., Inst. Mar. Res., IMR Ref. 55-9, 158 p. Love, M. S., M. H. Carr, and L. J. Haldorson. 1991. The ecology of substrate-associated juveniles of the genus Sebastes. Environ. Biol. Fishes 30:225-243. Love, M. S., J. Caselle, and W. Van Buskirk. In press. A severe decline in the commercial passenger fish- ing vessel rockfish ( Sebastes spp.) catch in the southern California Bight, 1980-1996. CalCOFI Rep. Love, M. S., and K. Johnson. In press. Aspects of the life histories of grass rockfish, Sebastes rastrelliger and brown rockfish, S. auriculatus, from southern California. Fish. Bull. Love, M. S., P. Morris, M. McCrae, and R. Collins. 1990. Life history aspects of 19 rockfish species (Scor- paenide:Seftastes) from the southern California Bight. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 87. Love, M. S., and W. V. Westphal. 1981. Growth, reproduction, and food habits of olive rock- fish, Sebastes serranoides , off central California. Fish. Bull. 79:533-545. MacCall, A. D. 1996. Patterns of low-frequency variability in fish popula- tions of the California Current. Calif. Coop. Oceanic Fish. Invest. Rep. 37:100-110. Mearns, A. J., M. J. Allen, M. D. Moore, and M. J. Sherwood. 1980. Distribution, abundance, and recruitment of soft-bot- tom rockfishes (Scorpaenidae: Sebastes) on the Southern California Mainland Shelf. Calif. Coop. Oceanic Fish. Invest. Rep. 21:180-190. Miller, D. J. and J. J. Geibel. 1973. Summary of blue rockfish and lingcod life histories; a reef ecology study; and gian kelp, Macrocystis pyrifera experiments in Monterey Bay, California. Calif. Dep. Fish and Game, Fish Bull. 158. Norton, J. G. and S. J. Crooke. 1994. Occasional availability of dolphin, Coryphaena hip- purus , to Southern California commercial passenger fish- ing vessel anglers: observations and hypotheses. Calif. Coop. Oceanic Fish. Invest. Rep. 35:230-239. Patton, M. L., R. S. Grove, and R. F. Harman. 1985. What do natural reefs tell us about designing artificial reefs in southern California? Bull. Mar. Sci. 37:279-298. Ralston, S., and D. F. Howard. 1995. On the development of year-class strength and co- hort variability in two northern California rockfishes. Fish. Bull. 93:710-720. Ralston, S., J. N. Ianelli, R. A. Miller, D. E. Pearson, D. Thomas, and M. E. Wilkins. 1996. Status of bocaccio in the Conception/Monterey/Eu- reka INPFC areas in 1996 and recommendations for man- agement in 1997. Appendix B, in Status of the Pacific Coast groundfish fishery through 1996 and recommended acceptable biological catches for 1997, stock assessment and fishery evaluation, p. 1-48. Pacific Fisheries Man- agement Council, 2130 SW 5th Ave, Suite 224, Portland, OR 97201. Reid, J. L., Jr., G. I. Roden, and J. G. Wyllie. 1958. Studies of the California Current system. CalCOFI Rep. 5:27-56. Roemmich, D., and J. McGowan. 1995. Climatic warming and the decline of zooplankton in the California Current. Science (Wash., D.C.) 267:1324- 1323. Schwartzlose, R. A. 1963. Nearshore currents of the western United States and Baja California as measured by drift bottoles. CalCOFI Rep. 9:15-22. Soutar, A. 1967. The accumulation of fish debris in certain California coastal sediments. CalCOFI Rep. 11:136-139. Love et a!.: Declines in rockfish recruitment and populations 501 Soutar, A., and J. D. Isaacs. 1974. Abundance of pelagic fish duringthe 19th and 20th centuries as recorded in anaerobic sediment off the Californias. Fish. Bull. 72:257-274. Stephens, J. S., Jr., G. A. Jordan, P. A. Morris, M. M. Singer, and G. E. McGowen. 1986. Can we relate larval fish abundance to recruitment or population stability? A preliminary analysis of recruit- ment to a temperate rocky reef. Calif. Coop. Oceanic Fish. Invest. Rep. 27:65-83. Stephens, J. S., Jr., P. A. Morris, D. J. Pondella, T. A Loonee, and G. A. Jordan. 1994. Overview of the dynamics of an urban artificial reef fish assemblage at King Harbor, USA, 1974-1991: a re- cruitment driven system. Bull. Mar. Sci. 55:1224-1239. Stephens, J. S., Jr., P. A. Morris, K. Zerba, and M. Love. 1984. Factors affecting fish diversity on a temperate reef: the fish assemblage of Palos Verdes Point, 1974- 1981. Env. Biol. Fish. 11:259-275. 502 Abstract Atka mackerel, Pleuro- grammus monopterygius, growth data differed significantly by area for length and weight characteristics, suggesting that local aggregations of this species develop differential growth patterns. We analyzed length-at-age and weight- length growth patterns from over 500 fish and 37 protein-coding gene loci to determine the relation among Atka mackerel stocks in the Aleutian Is- lands. However, the potential stock de- lineations based on the growth patterns were not supported by the genetic data. Atka mackerel showed a high degree of genetic variability. Variation was detected at 30 of 37 loci; 14 of the 30 loci were variable at the P0 95 level. Av- erage heterozygosity for 329 specimens was 0.137, an unusually high value for a marine fish. Between-sample varia- tion among samples was extremely low (Fst=0.004), suggesting considerable gene flow throughout the range repre- sented by the samples. On the basis of the genetic data, we cannot reject the null hypothesis that our samples came from a single genetically homogenous population of Atka mackerel. We pre- sume that gene flow occurs throughout the population through the dispersal of pelagic larvae and juveniles. We con- clude that despite genetic homogeniza- tion, phenotypic variation in Atka mack- erel adult life history stages warrants consideration by fishery managers. Manuscript accepted 17 September 1997. Fishery Bulletin 96:502-515 (1998). Geographic variation in genetic and growth patterns of Atka mackerel, Pleurogrammus monopterygius (Hexagrammidae), in the Aleutian archipelago Sandra A. Lowe Alaska Fisheries Science Center 7600 Sand Point Way NE, BIN C 1 5700 Seattle, Washington 981 I 5-007 0 E-mail address: sandra.lowe@noaa.gov Donald M. Van Doornik Gary A. \X/i nans Northwest Fisheries Science Center 2725 Montlake Blvd. E. Seattle, Washington 98112 Atka mackerel, Pleurogrammus mon- opterygius, a member of the greenling family Hexagrammidae, is distrib- uted throughout the north Pacific Ocean, the southern Bering Sea, and the Gulf of Alaska. The center of abundance of this semipelagic species is the Aleutian Islands.1 2 Greenling larvae and fingerlings (25-30 mm) undergo a pelagic stage and assume an oceanic mode, dur- ing which time they reside in the surface layers of the open waters and migrate for considerable dis- tances out to sea (Gorbunova, 1962). Adult (3+ years) Atka mackerel are pelagic during much of the year (in waters <200 m depth) but migrate annually to moderately shallow wa- ters where they become demersal and spawn in areas of strong currents (Gorbunova, 1962). They have spe- cific spawning site preferences and spawning has been observed in island passes in the Aleutian, Shumagin, and Commander Island archipela- goes (Turner, 1886; Rutenberg, 1962). Historically, large inshore concen- trations of spawning Atka mackerel have been the target of subsistence fisheries by native Aleuts (Turner, 1886 ). The first large-scale commer- cial catches in the Aleutian Islands and western Gulf of Alaska were taken by Russian fleets beginning in the early 1970s, followed by the Re- public of Korea, and to a lesser ex- tent, Japan in the early 1980s (Murai etal., 1981; Berger etal., 1986). Pres- ently, a U.S. trawl fishery harvests Atka mackerel from the eastern Bering Sea and Aleutian Islands re- gions, with minimal catches from the Gulf of Alaska. From 1992 to 1994, catches averaged about 70,000 met- ric tons (t) (valued at 17 million U.S. dollars [exvessel] in 1994) and in- creased to 104,000 t in 1996. 1,2 1 Lowe, S. A., and L. W. Fritz. 1996a. Atka mackerel. In Stock assessment and fishery evaluation report for the ground- fish resources of the Bering Sea/ Aleutian Islands regions as projected for 1997, p. 369-420. North Pacific Fishery Manage- ment Council, 605 W. 4th Avenue, Suite 306, Anchorage, AK 99501-2252. 2 Lowe, S. A., and L. W. Fritz. 1996. Atka mackerel . In Stock Assessment and Fish- ery Evaluation Report for the Groundfish Resources of the Gulf of Alaska as pro- jected for 1997, p. 331-361. North Pacific Fisheries Management Council, 605 W. 4th Avenue, Suite, Anchorage, AK 99501-2252. Lowe et at: Geographic variation in genetic and growth patterns of Pleurogrammus monopterygius 503 An analysis of morphological and meristic data by a Russian scientist indicated separate populations in the Gulf of Alaska and the Aleutian Islands.3 The meristic study compared the number of dorsal, anal, and pectoral fin rays, total number of vertebrae, and number of gill rakers on the first gill arch from a sample of 100 fish collected off Kodiak Island in the Gulf of Alaska and the Rat Islands in the Aleutian Islands. The morphological study consisted of a sta- tistical comparison of various partial fish body lengths as a percentage of fork length, by area. Lee (1985) also conducted a morphological study analyz- ing the covariance between four partial fish lengths and fork length by area and sex from samples taken from the Bering Sea, Aleutian Islands, and the Gulf of Alaska. The data showed some differences (al- though not consistent by area for each characteris- tic), suggesting a certain degree of reproductive iso- lation. On the basis of an analysis of variance of Aleu- tian Islands growth data with year, area, and sex as factors, Kimura and Ronholt ( 1988) found significant differences in weight and length-at-age of Atka mack- erel from six different areas in the Aleutian Islands, indicating potential stock differentiation. Kimura and Ronholt (1988) suggested therefore that Atka mackerel appear to be distributed in localized groups or assemblages once they assume the more demer- sal phase of their life history. Recent analysis of Aleu- tian Islands Atka mackerel growth data by current fishery management areas shows a distinctive size cline, with length at age smallest in the western Aleu- tians and largest in the eastern Aleutians1,2 Differential growth by area can be an indication of stock delineations. The presumption for Atka mackerel, based on the morphological and growth analyses, was of at least a discontinuous distribu- tion throughout the Aleutian Archipelago. Genetic tests are necessary to confirm separate stocks and help to further our understanding of the life history and distribution patterns of Atka mackerel. Infor- mation about the extent and nature of stock differ- ences is also critical to improve stock assessments for long-term management of the Atka mackerel fisheries in the Gulf of Alaska and Aleutian Islands. To date, there have been no population genetics surveys of this species to determine the existence of discrete stocks. A preliminary survey of allozymes from 40 indi- viduals indicated a sufficient number of polymorphic loci to conduct a full study of the genetic population structure of this species (Winans et al., 1995). We report here 1) length-at-age and length-weight rela- 3 Levada, T.P. 1979. Comparative morphological study of Atka mackerel. Pac. Sci. Res. Inst. Fish. Oceanogr.5(TINRO ), Vladivostok, U.S.S.R. Unpubl. manuscript, 7 p. tions derived from samples taken throughout the Aleutian Archipelago and 2) a genetic survey of Atka mackerel from samples taken from four locations in the Aleutian Islands ranging from 169°W to 172°E (Fig. 1). Our null hypothesis is that there are no dif- ferences between samples taken along the Aleutian Archipelago, from Umnak Island in the east to Attu Island in the west. Materials and methods Age , weight, and length data Biological samples and length and weight informa- tion were collected during National Marine Fisher- ies Service (NMFS) triennial trawl surveys, June to August of 1993 and 1994, from the Gulf of Alaska and Aleutian Islands region, respectively. The sam- pling coincided with the summer spawning period of Atka mackerel (July to October; McDermott and Lowe, 1997). Random samples of Atka mackerel from the trawl survey catches were sorted by sex, and in- dividual weight and fork-length data were collected. Length was estimated to the nearest one centime- ter, and weight was estimated to the nearest gram with a platform scale when weather conditions al- lowed (Martin and Clausen, 1995). Age structures (sagittal otoliths) were collected by using a length- stratified sampling scheme of five fish per sex, per centimeter length category. An attempt was made to distribute the otolith collections over the entire sur- vey area. Otoliths were placed in vials with 50% etha- nol, and age was determined by personnel in the NMFS Age Determination Laboratory. A total of 510 otoliths were collected and aged (Table 1). Otoliths were prepared by snapping each along the dorsal-ventral plane and passing the bro- ken surface over a flame. The burnt cross-section was examined under a dissecting microscope and illumi- nated by reflected light (Anderl et. al, 1996). Growth parameters were estimated by fitting the age-length data to the widely used von Bertalanffy (1938) growth equation: /t = L00fl-e~KF Age 11 14,576.45 1325.13 275.00 <0.0001 Area 3 5380.33 1793.44 372.19 <0.0001 Age x area 20 260.09 13.00 2.70 <0.0001 Error 477 2298.51 4.82 Table 5 Two-way factorial analysis of variance of Atka mackerel length-weight data collected from the 1993 and 1994 trawl surveys of the Gulf of Alaska and Aleutian Islands, respec- tively. The model contains the factors length and area, and an interaction term of length x area. SS = sum of squares. Mean Source df SS square F-value P>F Length 28 59.73 2.13 278.13 <0.0001 Area 3 1.21 0.40 52.77 <0.0001 Length x area 56 1.65 0.10 3.85 <0.0001 Error 425 3.26 0.01 TafaSe 6 Allele frequencies for 22 loci in Atka mackerel. “No.” represents the number of successfully screened fish. Rare variation is described for 7 alleles1 and 8 additional loci.2 Seven loci were monomorphic: AK *, CK*, bGALA*, bGLUA *, mMDH-2*, sMDH-2*, and PEPB-2*. Sample location Sample location Locus, Locus, sample size and alleles Umnak Island Seguaxn Pass Kiska Island Attu Island sample size and alleles Umnak Island Seguam Pass Kiska Island Attu Island AAT-1* AH-2*3 No. 39 100 73 74 No. 56 96 80 71 *-100 1.000 0.995 0.993 0.993 *100 0.536 0.521 0.613 0.514 *-63 0.000 0.005 0.007 0.007 *115a 0.384 0.354 0.300 0.366 AAT-2* *120 0.054 0.063 0.044 0.070 No. 62 100 80 73 *117 0.009 0.000 0.019 0.014 *100 0.815 0.860 0.875 0.829 *94 0.009 0.005 0.000 0.014 *60 0.185 0.135 0.112 0.164 *84 0.009 0.057 0.025 0.021 *115 0.000 0.000 0.006 0.007 ALAT* *82 0.000 0.005 0.006 0.000 No. 63 92 78 73 AAT-3* *100 0.563 0.533 0.564 0.555 No. 70 89 76 73 *70 0.413 0.457 0.436 0.425 *100 0.900 0.916 0.888 0.925 *120 0.024 0.011 0.000 0.021 *110 0.100 0.084 0.112 0.075 EST* ADA* No. 62 74 73 61 No. 59 96 78 72 *100 0.645 0.608 0.541 0.648 *100 0.534 0.417 0.449 0.444 *90 0.290 0.345 0.390 0.303 *75 0.280 0.292 0.269 0.285 *110 0.040 0.041 0.062 0.033 *89 0.025 0.036 0.026 0.021 *85 0.024 0.007 0.007 0.016 *110 0.008 0.031 0.064 0.049 G3PDH-1* *115 0.000 0.010 0.026 0.000 No. 74 94 80 74 *120 0.110 0.172 0.135 0.174 *-100 0.993 0.995 0.981 1.000 *130 0.034 0.016 0.032 0.014 *-70 0.007 0.005 0.019 0.000 *140 0.000 0.010 0.000 0.000 G3PDH-2*1 *155 0.008 0.016 0.000 0.014 No. 72 98 67 74 ADH* *100 0.993 0.964 0.985 1.000 No. 69 100 76 67 *125 0.007 0.031 0.015 0.000 *-100 0.870 0.875 0.908 0.836 GPI-1* *-183 0.094 0.100 0.086 0.149 No. 73 100 80 74 *-250 0.022 0.025 0.007 0.015 *100 0.904 0.880 0.881 0.926 *-25 0.014 0.000 0.000 0.000 *25 0.096 0.120 0.119 0.074 continued 51 0 Fishery Bulletin 96(3), 1998 Table 6 (continued) Sample location Locus, sample size Umnak Seguam Kiska Attu and alleles Island Pass Island Island IDDH*1 Sample location Locus, sample size Umnak Seguam Kiska Attu and alleles Island Pass Island Island No. 63 *100 0.984 *165 0.016 *25 0.000 IDHP-1* No. 100 *100 1.000 *65 0.000 *140 0.000 MPI* *100 0.971 *110 0.014 *116 0.014 *90 0.000 PEPB-1*1 2 3 n 74 *100 0.993 *90 0.007 No. 73 *100 0.562 *110 0.397 *118 0.007 *90 0.000 *113 0.034 PEPD* No. 63 86 76 0.948 0.954 0.006 0.007 0.041 0.039 80 74 1.000 0.988 0.000 0.000 0.000 0.000 0.969 0.994 0.020 0.006 0.000 0.000 0.010 0.000 90 80 0.994 1.000 0.000 0.000 99 80 0.576 0.544 0.394 0.387 0.005 0.019 0.005 0.000 0.020 0.050 93 80 *100 74 *110 0.973 *120 0.007 *92 0.020 PGDH*1 No. *100 0.980 *80 0.007 *115 0.014 PGK-1*1 No. 0.993 *100 0.007 *71 0.000 *45 0.000 *15 PGM*1 74 No. 0.993 *100 0.007 *80 74 SOD * 0.500 No. 0.439 *100 0.034 *190 0.007 TPI* 0.020 No. *-100 74 *-500 0.738 0.742 0.222 0.226 0.024 0.011 0.016 0.022 69 99 0.819 0.818 0.152 0.152 0.022 0.030 72 100 0.785 0.765 0.097 0.115 0.097 0.090 0.014 0.030 74 99 0.926 0.960 0.074 0.040 75 96 0.993 0.964 0.007 0.036 74 100 1.000 1.000 0.000 0.000 0.794 0.804 0.175 0.182 0.025 0.014 0.006 0.000 79 72 0.816 0.771 0.152 0.153 0.032 0.076 79 74 0.829 0.770 0.082 0.122 0.082 0.101 0.006 0.007 80 74 0.944 0.953 0.050 0.047 80 74 0.981 0.993 0.019 0.007 80 74 0.994 0.986 0.006 0.014 1 Variation at alleles not listed above where the frequency of the allele is < 0.010: G3PDH-2*90 and IDDH*40 in Seguam Pass; IDHP-1*90 and *50 in Kiska Island; PEPB1*115 in Seguam Pass; PGDH*70 and PGK1*55 in Gulf of Alaska; and PGM*115 in Kiska Island. 2 Variation at 8 loci not listed above where the frequency of the alternate allele is < 0.010: AH-1*118 in Attu Island; ENO*-180 in Seguam Pass; FH*77, and GPI-2*93 , and *105 in Kiska Island; IDHP-2*73 in Seguam Pass; LDH*155 in Attu Island; MDH-1*150 (n=100) and PK*-350 in Seguam Pass. 3 A seventh allele (*110) could not be reliably distinguished from the *115 allele and therefore was pooled with the latter allele. used von Bertalanffy curve, and the length-weight data were fitted to an allometric length-weight rela- tionship. The F-tests on the growth curves deter- mined that there was significant variability in the data among areas that was explained by separate growth curves rather than a single growth curve for all areas combined. Although the Gulf of Alaska and eastern Aleutian Islands data sets were small, data fits to the separate von Bertalanffy models were good, and an F-test indicated that separate growth curves were appropriate. There were low numbers of fish greater than five years in the Gulf of Alaska; how- ever, the data set still allowed for fairly precise esti- mation of the asymptotic length (CV=0.02), the pa- rameter most likely to be affected by the lack of older ages. Morphological and meristic evidence and growth patterns suggest that Atka mackerel may have some Table 7 Within-sample genetic variability over 37 loci. Location Average heterozygosity Percent of loci P0 95 Average number of alleles/locus Umnak Island 0.135 35.1 2.2 Seguam Pass 0.142 32.4 2.4 Kiska Island 0.135 35.1 2.3 Attu Island 0.137 32.4 2.2 degree of stock separation, at least by broad geo- graphic areas (the eastern Bering Sea, Aleutian Is- lands, and Gulf of Alaska12,3; Lee, 1985; Kimura and Ronholt, 1988). The present study is the first genetic Lowe et a I.: Geographic variation in genetic and growth patterns of Pleurogrammus monopterygius 51 1 survey of Atka mackerel intended to help understand the level and pattern of reproductive isolation among areas thought to be reflected in the areal growth dif- ferences. Samples of Atka mackerel were collected dur- ing the spawning season when groups of this species are locally aggregated, presumably providing the best separation of potentially reproductively isolated groups. Our results show that Atka mackerel have above- average levels of genetic variation for a marine fish. The average heterozygosity per locus for Atka mack- erel is 0. 137, well above the average reported by Ward et al. (1994) for 57 species of marine fish (0.064 ± 0.004). Assuming that the majority of the allozyme variation is selectively neutral (Kimura, 1968), we believe that large levels of genetic variation are most parsimoniously explained by large, historically stable populations. An alternative explanation is that large levels of variability reflect a response to inhabiting a heterogeneous habitat (Avise, 1994). Nonetheless, 30 polymorphic loci provide substantial statistical power to assess the level of between-sample differentiation. Between-sample variation was extremely low among the four samples of Atka mackerel. In a study where electrophoretic data for marine, freshwater, and anadromous fishes were compared. Ward et al. (1994) calculated an average GgT ( roughly equiva- lent to FgT) of 0.062 for 57 species of marine fish (com- pared with 0.22 for freshwater species of fish). The FgT for Atka mackerel (0.004) is far smaller than this average, providing evidence that a large amount of gene flow is occurring throughout the range repre- sented by these samples. Very low genetic distance values between samples were observed; the largest distance value between two samples was 0.00017, indicating little or no stock differentiation. Further- more, the nonsignificant Hardy- Weinberg test of the four samples pooled together did not indicate any between-sample heterogeneity. In light of these two genetic results, we can not reject the null hypothesis that our samples came from a single genetically ho- mogenous population of Atka mackerel. There was also no apparent gradual differentiation throughout the Aleutian Archipelago corresponding to the clinal geographic variation seen in the growth data. Stock delineations based on the age-length and weight- length relationships were not supported by the allozyme data. Concordance of electrophoretic and growth data sets A lack of congruence between genetic and life his- tory characteristics was also found for Pacific ocean perch by Seeb and Gunderson (1988). They analyzed data from the Washington coast to the Bering Sea. Stock delineations based on age structure, age-length relationships, and ages at maturity were not sup- ported by the allozyme data. Although they did not find clear genetic stock differentiation, they did find a cline of gene frequencies within the Gulf of Alaska and significant allele frequency differences between the extremes of the geographic range for some loci. There are at least three possible explanations for the lack of genetic stock differentiation for Atka mackerel in the Aleutian Islands. First, the genetic technique surveyed invariant gene loci when in fact genetic differences may exist among the sampled stocks. As in any genetic study, the absence of ge- netic differentiation does not preclude the possibil- ity that true genetic differences exist, and other ge- netic techniques may be employed to further exam- ine this possibility (Avise, 1994). Second, the species could perform a major spawning migration encom- passing the entire population, with mixing and spawning in one area. Third, separate spawning loca- tions are used but gene flow occurs among locations through the dispersal of pelagic larvae and juveniles. We consider the third explanation most likely given the growth differences seen in adult Atka mackerel. The Aleutian Archipelago encompasses several wide and deep straits that could form barriers to sig- nificant movement across the island chain for adult demersal fish residing over the continental shelf. The pelagic behavior of larval and juvenile fish and their presumed distribution in the open ocean habitat, however, would make them susceptible to wide dis- persal by currents, thus accounting for the genetic homogeneity among samples. A schematic diagram of the basic surface currents in the north Pacific is shown in Figure 6 (McAlister and Favorite, 1977). The Alaska stream flows westward throughout the Aleutian Archipelago, with significant northward flow to the Bering Sea through Amukta, Amchitka, and Buldir passes. The North Aleutian current flows eastward to Umnak Island providing a thorough mixing mechanism across the Aleutian Island chain. The currents could thus provide the mechanism for sufficient gene flow through the Aleutian chain to actually prevent genetic differentiation. It is pre- sumed that adults do not undertake large-scale mi- grations and assume a fairly localized existence, which might thus account for the differences in length, age, and weight comparisons of the adults. Observed spawning areas in the Aleutian, Shumagin, and Commander Islands, which have been histori- cally referenced in the literature (Turner, 1886; Rutenberg, 1962), are thus presumed to attract nearby resident schools that migrate inshore. Although there are no major currents flowing east- ward from the Aleutian Islands to the Gulf of Alaska, 512 Fishery Bulletin 96(3), 1998 c o> 03 J- C b £ 3 6 5 S . 03 pQ 03 ►, -D t £ •: 03 -*-> £ o £ □ cn > cd • -g i> p 33 t> fc 2 05 £o J rH n <1 a> ~j c J < a Q1 -M P"0 43 £ £ +3 jl3 a3 ^ ^ Jr cd ^ 03 § II to 03 lO ■-£ H • - b * t) 5 c o ^ 3 +j d 5 c C/3 ^ 03 _ C/3 C0 C Cd 03 Sm -3 J-H 03 < a «? II cd 0) > • - cd 03 O 4> X V ^ -*4> cj cd sp « -f, £ ^ 03 .£ pO £ £ C rn . „ £ o W M43 II & CO Sm o CO £ 3 "u ^ £ " -M O Q) « ■£ > Jd ^ C/3 cd w £ ■s « § 3 >> t- CO O H the western Gulf of Alaska portion of the population may be the result of juve- nile or adult migration (or both) or habitat expan- sion. Kimura and Ronholt (1988) postulated that Gulf of Alaska Atka mack- erel are at the extreme limit of their geographic range (the extreme limit of which is populated only during periods of favorable environmental conditions). Zolotov ( 1984) found that Atka mackerel spawning habitat in Russian waters extended from the Kam- chatka Peninisula through the Kurile Islands almost without breaks. Continuous distribution of Atka mack- erel spawning habitat and dispersal of larvae into the open ocean pointed to an absence of mechanisms that would result in repro- ductively isolated stocks in Russian waters. Zolotov (1984) thus concluded that Atka mackerel did not form ecological groupings along the Kamchatka Pen- insula and Kurile Islands but made up a single popu- lation. Although we cannot rule out that there are un- known mechanisms or processes which might re- sult in reproductive isola- tion of groupings of Aleu- tian Island Atka mackerel, life history information suggests that a process for extensive mixing occurs during the early life stages. We currently do not have any observations to sup- port the necessary pro- cesses (e.g. larval reten- tion, natal homing, genetic imprinting), that would result in reproductive iso- lated populations in the Aleutian Islands. Lowe et a I.: Geographic variation in genetic and growth patterns of Pleurogrammus monopterygius 513 We interpret the areal growth differences exhib- ited by Atka mackerel as indications of localized as- semblages made up of groupings from a single mix- ing stock that aggregates during the adult portion of its life history. We suggest that the local specific growth variability seen in adults is a reflection of environmental effects. It is unknown which environ- mental characteristics might be most influential on growth of Atka mackerel; however, temperature and salinity are noted to be the most important hydro- logical factors affecting the distribution of hexagram- mids, and food, predators, and parasites the most important biological factors (Rutenberg, 1962). Seeb and Gunderson (1988) noted that local growth and age-at-maturity differences could also reflect histori- cal influences and fishing pressure. A large and sus- tained Atka mackerel fishery has been conducted throughout the Aleutian Islands since the early 1970s. Catches have fluctuated with the demise of the foreign fishery and the development of the do- mestic fishery, and in recent years the fishery has been concentrated in the eastern Aleutian Islands where the largest fish reside. The fish in the west- ern Aleutian Islands were not heavily exploited from 1980 to 1994, but they have historically been the smallest fish. The geographic size cline noted in the growth data seems to run counter to what we might expect given the differential fishing pressure. Stock assessment and management implications The stock structure of Atka mackerel has stock assess- ment and fishery management implications. From a stock assessment perspective, we are interested in elu- cidating the underlying population processes that would result in stock separation, or lack of, and in evaluating the impacts of the fishery on the genetic stock(s), par- ticularly on the spawning concentrations. From a fish- ery management perspective, the practical recognition of fishery-targeted assemblages (not necessarily geneti- cally distinct) and the spatial and temporal affects of harvesting these assemblages are of interest. Historically, small-scale Alaskan subsistence fish- eries targeted spawning concentrations of Atka mack- erel, but the very shallow and presumably rough habitat of the spawning grounds are not accessible to the current large-scale commercial fisheries. Analysis of commercial fishery data indicates that bottom trawling is probably not disturbing the nest- ing sites.6 Thus, the unique reproductive life history features of Atka mackerel may provide for some pro- tection of the spawning stock. Booke ( 1981) distinguished between inherited (ge- netic) versus “acquired” or “environmentally induced” characters; the latter including morphological and phenotypic characteristics. Because acquired pheno- typic markers by definition are appropriated through contact with the environment, they may not reflect the population genetic characteristics (Avise, 1994). However, acquired markers can serve an important role in population analysis because they can reveal where individuals have spent various portions of their lives (Avise, 1994). From both a stock assess- ment and fishery management perspective, the lo- calized aggregations of adult Atka mackerel, al- though not genetically distinct, are important to our understanding of the population dynamics of this species and the impact of the fishery. Indications of localized populations of Atka mack- erel raise the issue of potential localized depletion by the fishery. Another fundamental fishery manage- ment and assessment question has been the relation of Gulf of Alaska Atka mackerel to Aleutian Islands Atka mackerel. Atka mackerel are currently man- aged by two major areas, the Gulf of Alaska and the Aleutian Islands, and by three subareas within the Aleutian Islands. For management purposes, there are at least two catch quotas that must be set given the management area boundary at 170°W which di- vides the Gulf of Alaska and the Aleutian Islands. Any further subdivisions of the quota are generally based on stock separation rationale or the desire to spread out fishing effort over large geographic areas (or both). The Gulf of Alaska Atka mackerel are as- sessed separately from Aleutian Islands Atka mack- erel, mainly because of the different sources of data (two different survey time series), and to a lesser extent on the presumption that Gulf of Alaska Atka mackerel showed some level of separation from Aleu- tian Islands Atka mackerel. The results of this study show no evidence of Atka mackerel genetic stock separation between the western Gulf of Alaska and throughout the Aleutian Islands chain. However, because adults show evidence of localized aggrega- tions, it seems appropriate to set at least two sepa- rate quotas in Alaskan waters (Gulf of Alaska and Aleutian Islands), with further subdivisions to dis- tribute fishing effort. The North Pacific Fishery Management Council (NPFMC) implemented Amendment 28 to the “Fish- ery Management Plan for the Groundfish of the Bering Sea and Aleutian Islands,” creating three subareas within the large Aleutian Islands manage- ment area (Fig. 1). The impetus for the implementa- 6 Fritz, L. W., and S. A. Lowe. 1997. Seasonal distributions of Atka mackerel (Pleurogrammus monopterygius) in commer- cially-fished areas of the Aleutian Islands and Gulf of Alaska. NMFS. Alaska Fisheries Science Center, 7600 Sand Point Way NE, Seattle, WA 98115-0070. Unpubl. manuscript. 514 Fishery Bulletin 96(3), 1998 tion of additional management areas was the possi- bility of localized depletion of Atka mackerel by the fishery given the presumed localized distribution of this species and the localized, highly concentrated fishery effort. This study shows that genetically dis- tinct populations of Atka mackerel do not exist in this area, and therefore a significant amount of gene flow may occur among locales within this species. However, gene flow is presumably occurring during the early life stages (larval, juvenile) and is depen- dent on dispersal by currents. The potential for lo- calized depletion by the fishery at the adult stage is still a concern. Until we have a better understand- ing of the dispersal and distribution mechanisms affecting Atka mackerel early life stages, manage- ment strategies should strive to spread out fishing effort to help reduce the potential for localized deple- tion. We conclude that despite the genetic homog- enization, the phenotypic variation in adult Atka mack- erel warrants consideration by fishery managers. Acknowledgments We acknowledge the field biologists on surveys who accommodated our sampling plan and painstaking collected the biological samples, in particular, Bill Flerx, Michael Martin, and Robin Harrison. This work would not be possible without their efforts. We also thank David Kuligowski for his laboratory as- sistance in processing the genetic samples and Lowell Fritz, Dan Kimura, Anne Hollowed, and three anony- mous reviewers, whose reviews of our manuscript improved it considerably. Literature cited Aebersold, P. B., G. A. Winans, D. J. Teel, G. B., Milner, and F. M. Utter. 1987. Manual for starch gel electrophoresis: a method for the detection of genetic variation. U.S. Dep. Conimer., NOAA Tech. Rep. NMFS 61, 19 p. Anderl, D. M., A. Nishimura, and S. A. Lowe. 1996. Is the first annulus on the otolith of the Atka mack- erel, Pleurogrammus monop terygius, missing? Fish. Bull. 94:163-169. Avise, J. C. 1994. Molecular markers, natural history and evolution. Chapman and Hall, New York, NY, 511 p. Berger, J. D., J. E. Smoker, and K. A. King. 1986. Foreign and joint venture catches and allocations in the Pacific Northwest and Alaska under the Magnuson Fishery Conservation and Management Act, 1977- 84. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F- NWC-99, 53 p. Booke, H.E. 1981 . The conundrum of the stock concept — are nature and nurture definable in fishery science? Can. J. Fish. Aquat. Sci. 38:1479-1480. Chambers, J. M., and T. J. Hastie (eds). 1992. Statistical models in S. Wadsworth & Brooks/Cole Advanced Books & Software, Pacific Grove, CA, 698 p. Gorbunova, N.N. 1962. Razmnozhenie i razvite ryb semeistva terpugovykh (Hexagrammidae) (Spawning and development of green- lings (family Hexagrammidae)). Tr. Inst. Okeanol., Akad. Nauk SSSR 59:118-182. In Russian. [Trans, by Isr. Program Sci. Trans., 1970, p. 121-185. In T.S. Rass ( ed. ), Greenlings: taxonomy, biology, interoceanic transplantation. Available Natl. Tech. Inf. Serv., Springfield, VA., as TT 69-55097.] IUBMBNC (International Union of Biochemistry and Molecular Biology, Nomenclature Committee). 1992. Enzyme nomenclature 1992: recommendations of the Nomenclature Committee of the International Union of Biochemistry and Molecular Biology on the nomenclature and classification of enzymes. Academic Press, San Di- ego, CA, 862 p. Kimura D .K., and L. L. Ronholt. 1988. Atka mackerel. In R. G. Bakkala (ed.), Condition of groundfish resources of the eastern Bering Sea and Aleu- tian Islands region in 1987, p. 147-171. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-139. Kimura, M. 1968. Evolutionary rate at the molecular level. Nature (Lond.) 217:624-626. Lee, J. U. 1985. Studies on the fishery biology of the Atka mackerel Pleurogrammus monopterygius (Pallas) in the north Pa- cific Ocean. Bull. Fish. Res. Dev. Agency 34, p.65-125. Martin, M. H., and D. M. Clausen. 1995. Data report: 1993 Gulf of Alaska bottom trawl survey. U.S. Dep. Commer., NOAA Tech. Memo. NMFS- AFSC-59, 217 p. McAlister, B. W., and F. Favorite. 1977. Oceanography. In M. L. Merritt and R. G. Fuller (eds.), The environment of Amchitka Island, p. 331- 353. Prepared for Division of Military Application, En- ergy Research and Development Administration. [Avail- able Natl. Tech. Infor. Ser., Springfield, VA as TID-26712.] McDermott, S. F., and S. A. Lowe. 1997. The reproductive cycle of Atka mackerel (Pleurogram- mus monopterygius) in Alaskan waters. Fish. Bull. 95: 321-333. Murai, S., H. A. Gangmark, and R. R. French. 1981. All-nation removals of groundfish, herring, and shrimp from the Eastern Bering Sea and northeast Pacific Ocean, 1964-80. U.S. Dep. Commer., NOAA Tech. Memo. NMFS/F/NWC-14, 40 p. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89:583-590. Neter, J., W. Wasserman, and M. H. Kutner. 1985. Applied linear statistical models. Richard D. Dar- win, Inc., Homewood, IL, 1127 p. Rutenberg, E. P. 1962. Survey of the fishes Hexagrammidae. In T. S. Rass (ed.), Tr. Inst. Okeanol., Akad. Nauk SSSR 59, p. 1-103. In Russian. [Trans, by Isr. Program Sci. Trans., 1970, T.S. Rass (ed.), Greenlings: taxonomy, biology, interoceanic transplantation. Available Natl. Tech. Infor. Serv., Spring- field VA, as TT 69-55097.) Lowe et a!.: Geographic variation in genetic and growth patterns of Pieurogrammus monopterygius 515 Seeb, L. W., and D. R. Gunderson. 1988. Genetic variation and population structure of Pacific ocean perch (Sebastes alutus). Can. J. Fish. Aquat. Sci. 45:78-88. Shaklee, J. B., F. W. Allendorf, D. C. Morizot, and G. S. Whitt. 1990. Gene nomenclature for protein-coding loci in fish. Trans. Am. Fish. Soc. 119:2-15. Sokal, R. R., and F. J. Rob If. 1981. Biometry. W. H. Freeman and Company, San Fransisco, CA, 859 p. Swofford, D. L., and R. B. Selander. 1981. BIGSYS-1: a FORTRAN program for the comprehen- sive analysis of electrophoretic data in population genet- ics and systematics. J. Hered. 72:281-283. Turner, L. M. 1886. Results of investigation made chiefly in the Yukon District and the Aleutian Islands; conducted under the auspices of the Signal Service, United States Army, extend- ing from May, 1874, to August, 1881. Contribution to the Natural History of Alaska. No. II. Arctic Series of Publica- tions issued in connection with the Signal Service, U.S. Army. Government Printing Office, Washington, D.C. Von Bertalanffy, L. 1938. A quantitative theory of organic growth (inquiries on growth law II). Hum. Biol. 10(2):181-213. Ward, R. D., M. Woodwark, and D. O. F. Skibinski. 1994. A comparison of genetic diversity levels in marine, fresh- water, and anadromous fishes. J. Fish Biol. 44:213-232. Winans, G. A., P. B. Aebersold, and S. A. Lowe. 1995. Allozyme variation in Atka mackerel, Pieurogrammus monopterygius. Isozyme Bull. 28, p. 52. Zolotov, O. G. 1984. Biology of the northern Pieurogrammus monop- terygius (Pallas) in the waters of Kamchatka and the Kurile Islands. Ph.D diss., Kamchatka Pacific Scientific Institute, 107140, Moscow, ul. V. Krasnoseljskaya, d. 17, 22 p. 516 A population profile for hagfish, Myxine glutinosa, in the Gulf of Maine. Part 2: Morphological variation in populations of Myxine in the North Atlantic Ocean Frederic H. Martini Michael R Lesser John B. Heiser Shoals Marine Laboratory, G-14 Stimson Hall Cornell University, Ithaca, New York 1 4853 Current address (for F H. Martini): 507 1 Hana Hwy. Haiku, Hawaii 96708 E-mail address (for F H Martini) martini@maui.net Abstract .—The species Myxine glutinosa has long been recognized as encompassing both eastern and west- ern North Atlantic populations. Wisner and McMillan (1995) have proposed splitting the species into M. limosa (Girard, 1858; western North Atlantic) and M. glutinosa (Linnaeus, 1758; east- ern North Atlantic). We examined a variety of morphological characteris- tics, including cusp counts, slime pore counts (total, prebranchial, trunk, and tail), total length, and body proportions (prebranchial, trunk, and tail length, maximum width and depth, and depth at cloaca). Several western Atlantic populations of varying size were com- pared with one large sample from the waters between Sweden and Denmark. The results indicate that although specimens from the Gulf of Maine dif- fer significantly from those collected in the eastern North Atlantic, specimens collected from the mid-Atlantic coastal region of the United States and from northern Canadian waters are less dis- tinctive. In addition, there are signifi- cant morphological differences among the populations sampled in the west- ern North Atlantic. It is therefore sug- gested that until and unless molecular data indicate otherwise, the species name M. glutinosa be retained as en- compassing both eastern and western North Atlantic populations. Manuscript accepted 13 January 1998. Fishery Bulletin 96:516-524 (1998). Only one hagfish, Myxine glutinosa (Linnaeus, 1758), has been found on both sides of the North Atlantic Ocean; this is the only hagfish re- ported from within the Gulf of Maine. Myxine glutinosa are an important species in the Gulf of Maine because 1) they are present in substantial numbers (densities may reach 500,000/km2), and they may have a considerable impact on the benthic community (Lesser et al., 1996; Martini et al., 1997b), 2) they have direct and indirect effects on commercial fisheries in the Gulf of Maine, through predation on groundfish in fixed gear fisheries and through competition for shrimp, and 3) they have commercial value as the basis for a leather fishery that provided Gulf of Maine fisher- men with $1.2 million in 1996. The first part of this study (Mar- tini et ah, 1997a) presents morpho- logical data and a population pro- file for hagfish in the inner Gulf of Maine. In this report we compare the morphological characteristics of this population with those of M. glutinosa populations in other geo- graphical regions. Wisner and McMillan ( 1995) proposed splitting M. glutinosa into M. glutinosa (Linnaeus, 1758) for the eastern North Atlantic (ENA) and Myxine limosa (Girard, 1858) for the west- ern North Atlantic ( WNA). The pro- posed separation was based on dif- ferences in size at sexual maturity, maximum size, and the color of pre- served specimens. We undertook this analysis initially to see if there were significant morphometric dif- ferences that would support the separation of species. This ulti- mately led us to perform statistical comparisons among our morpho- logical data, collected in the inner portion of the Gulf of Maine, and the data of other researchers working with eastern and western North At- lantic populations of M. glutinosa. Materials and methods Wisner and McMillan included mor- phometric data for Myxine collected over a very broad area. The latitu- dinal limits were 33°46'N to 66°39'N, and the longitudes ranged from 79°42'W to 52°13'W. Of the 138 specimens cited in their report, only 28 (20%) were collected within the boundaries of the Gulf of Maine. Their study provided detailed data Martini et a!.: A population profile for Myxine glutinosa 517 on the numbers of slime pores (prebranchial, trunk, tail, and total) and total cusp counts for Myxine in the eastern and western North Atlantic. Through correspondence with Robert L. Wisner,1 we obtained data and collection information for 73 western North Atlantic (WNA) and 179 eastern North Atlantic (ENA) specimens. These data sets permitted exten- sive comparisons with our morphological data for 1478 specimens from the inner Gulf of Maine (WNA). Methods of measuring and counting followed those of Fernholm and Hubbs (1981) and McMillan and Wisner (1984). The body axis was divided into prebranchial, trunk, and tail regions. The pre- branchial measurement extends from the tip of the snout to the anterior margin of the pharyngocu- taneous duct (pcd), the trunk continues to the ante- rior margin of the cloaca, and the tail region extends from that point to the tip of the tail. The sum of these measurements is equal to the total length (TL). The proportional measurements were recorded in milli- meters and converted to percentages of total length. Additional data included body depth, body width, cloacal depth, and tail depth (in mm and %TL); total slime pore count (TP) with prebranchial, trunk, and tail counts (as values and %TP), and total cusp counts. Morphological data were collected from 306 hag- fish from the inner Gulf of Maine during five years of trap surveys at depths of 120-150 m by staff and students of the Shoals Marine Laboratory. Samples were collected from late June to early September. The 1 Wisner, R. L. 1996. Marine Biology Research Division, Univ. California, San Diego, La Jolla, CA 92093-0202. Unpubl. data. traps used in this study were comparable to those set by the commercial hagfishing fleet. The trap used on most trips consisted of a weighted garbage can with 5-7.5 cm holes in the side and an inner, baited trap of wire screening. After collection, the animals were transferred to tanks of chilled seawater (2-6°C) at 32-35 ppt salinity for transport to the marine labo- ratory. All measurements were taken on fresh speci- mens. Total length data from these samples were combined with length data provided by the New England Fisheries Development Association for 1172 hagfish collected by fishermen within 80 km of the shoreline in the southern Gulf of Maine over the pe- riod of June-July 1995. All specimens were collected at depths of 120-180 m. Samples from elsewhere in the North Atlantic were collected by either trawling or trapping at depths of 30-600 m between 1965 and 1987 (Wisner1). Collec- tions in the eastern North Atlantic were from mid- summer (August); those from the western North At- lantic were from all seasons. The morphological data from these collections were recorded from preserved specimens by Robert L. Wisner and Charmion McMillan. An analysis of variance was performed on the data with the fixed effect of collection site by us- ing Statview 4.5 (Abacus Concepts Inc., 1995). This analysis consisted of post-hoc multiple comparison tests (Scheffe F-test) at the 5% level of significance. Results When plotted out on regional charts (Fig. 1 ), the data Figure 1 Sampled populations of M. glutinosa in the eastern and western North Atlantic. The shaded areas indicate the approximate range of collection sites for each data set. (Left) Collection sites within the Gulf of Maine. The 100-m, 200-m, and 300-m depth contour lines are indicated. OGM = outer Gulf of Maine; IGM = inner Gulf of Maine; BB = Brown’s Bank; GB = George’s Bank (Right) Other collection sites in the North Atlantic. ENA = eastern North Atlantic; NWA = Northwest Atlantic; G = Gulf of Maine (see [left]); and MAC = mid-Atlantic coast. 518 Fishery Bulletin 96(3), 1998 Table 1 Morphological measurements for Myxine glutinosa in the eastern and western North Atlantic. Asterisks indicate values signifi- cantly different (PcO.OOOl) from the Eastern Atlantic population. Character Mean SD Range %TL; SD Range n Eastern North Atlantic group (ENA) Total length (mm) 299 31 227-387 100 179 Prebranchial (snout:pcd) 82 8 61-101 27.4 1.1 24-31 179 Trunk 174 20 133-227 58.2 1.5 54-62 179 Tail 44 5 33-59 14.9 0.9 12-18 179 Width 14 2 10-21 4.9 0.5 3-6 179 Depth (trunk) 18 3 13-26 6.1 0.6 5-8 179 Depth (cloaca) 14 2 10-20 4.7 0.5 4-7 179 Depth (tail) 15 2 12-20 5.1 0.4 4-6 179 Inner Gulf of Maine group (IGM) Total length (mm)2 524* 102 170-950 100 1478 Prebranchial (snout:pcd) 135 27 54-200 26.7* 1.6 24-37 143 Trunk 311 72 28-459 61.4 5.9 42-83 143 Tail 64 14 25-106 12.6* 1.2 9-17 143 Width 14 7 4-35 2.7* 1.1 2-6 91 Depth (trunk) 22 7 8-35 4.2* 0.7 2-7 87 Depth (cloaca) 19 5 6-28 3.7* 0.5 2-5 198 Depth (tail) 20 5 8-30 4.0* 0.5 2-5 97 1 %TL = Percentage of total length. Values were log-transformed before comparison. 2 Combined data, from Martini et al., 1997a, and Kuenstner, 1996. collection sites fell into four groups: 1 ) mid-Atlantic coast (MAC), from the latitude of Charleston, South Carolina, to the southern coast of New England (n=51, with data on total length, slime pore counts, and cusp counts), 2) the outer Gulf of Maine (OGM), from the area of Brown’s Bank (>z = 13, with data on total length, slime pore counts, and cusp counts), 3) the northwestern Atlantic (NWA) off Labrador, in- cluding Davis Strait (n= 9, with data on total length, slime pore counts, and cusp counts), and 4) the east- ern North Atlantic (ENA), from the Skaggerrack, between Sweden and Denmark (n = 179, data on to- tal length, proportional lengths, slime pore counts, and cusp counts). Our data set comprised a separate group, 5) the inner Gulf of Maine (IGM), represented by samples from within 80 km of the shoreline (n=1478 for total lengths, n=143 for proportional measurements, n =94-97 for slime pore counts and cusp counts). Comparisons of morphometric data for the ENA and IGM populations, the two groups for which the greatest number of measurements were available, showed differences in total length and proportional measurements (Table 1). Table 2 presents data on cusp counts and slime pore counts for all groups (NWA, IGM, OGM, MAC, and ENA). The propor- tional measurements expressed as %TL and %TP were log-transformed before analysis and retrans- formed for presentation in the tables. Figure 2 illus- trates the results of the post-hoc multiple compari- son tests (Scheffe F-test) at the 5% level of signifi- cance. Table 3 details these results. Table 4 compares the data on total length for the sampled populations. These results can be briefly summarized as follows: 1 The NWA sample, closest geographically to the ENA population, can be distinguished from the ENA only in terms of the trunk slime pores as a percentage of the total slime pore count. The to- tal length data for the NWA and ENA groups are not significantly different. 2 The NWA sample is distinct from the OGM sample in terms of the total slime pore count and differs from the IGM sample with regard to prebranchial, trunk, and total slime pore counts as well as the total cusp count. The differences in regional slime pore count between the NWA and IGM samples were not significant when compared as percent- ages of the total slime pore count. 3 The OGM sample differs from the ENA sample in terms of the total slime pore count, and from the IGM sample in terms of the trunk slime pore count (as a value, not as a percentage of total slime pores) and in the total slime pore count. The to- tal length data for the OGM and ENA groups were not significantly different, but the OGM animals were significantly smaller than those of the IGM. 4 The mid-Atlantic coastal group (MAC) has charac- ters that overlap those of other groups. The MAC Martini et ai.: A population profile for Myxine glutinosa 519 Table 2 Tooth cusp counts and slime pore counts for populations of Myxine glutinosa L. in the eastern and western North Atlantic. Parenthetical values refer to the data expressed as a percentage of the total number of slime pores. Character Population Mean SD Range n Total cusp count NWA 32 1 31-34 9 OGM 35 2 33-39 7 IGM 35 2 28-40 97 MAC 35 2 30-38 51 ENA 34 1 29-38 101 Total slime pore count NWA 95 2 91-101 9 OGM 106 4 100-113 13 IGM 114 7 91-128 94 MAC 100 4 91-108 51 ENA 96 5 85-108 143 Prebranchial slime pores (%) NWA 26 (27) 3 (2) 23-32 (25-29) 9 OGM 31 (29) 1 (1) 28-32 (28-30) 13 IGM 33 (29) 4(3) 20-45 (21-40) 94 MAC 28 (28) 2 (2) 25-33 (25-33) 51 ENA 27 (28) 3 (2) 20-36 (24-33) 143 Trunk slime pores (%) NWA 58 (61) 3 (1) 56-66(59-62) 9 OGM 63 (59) 2 (1) 59-67 (57-61) 13 IGM 67 (59) 4(3) 51-77 (51-69) 94 MAC 59 (59) 2 (2) 54-65 (55-64) 51 ENA 56 (58) 3 ( <0.5 ) 50-63 (57-60) 143 Tail slime pores (%) NWA 11(12) 1 (2) 8-13 (8-14) 9 OGM 13 (12) 1 (1) 11-15(10-14) 13 IGM 13(11) 2 (1) 8-19 (9-15) 94 MAC 12 (12) 1 (1) 10-14 (10-14) 51 ENA 12 (13) 1 (1) 8-15 (9-14) 143 data and the NWA sample differ significantly in the total cusp count. The MAC sample differs from the OGM sample in terms of the total slime pore count; it differs from the IGM sample in terms of pre- branchial, trunk, tail, and total slime pore counts; the regional differences are not significant when compared as percentages of the total slime pore count. The MAC sample differs from the ENA sample in the trunk slime pore count (as a value, not as a percentage of total slime pores) and total cusp count. The total length data for the MAC, OGM, and ENA groups are not significantly different; all are sig- nificantly smaller than the IGM animals. 5 The IGM data is distinct from the ENA group for trunk, tail, and total slime pore and cusp count data. The IGM and ENA groups also differ sig- nificantly in prebranchial, trunk, and tail lengths, body width, body depth, tail depth, and cloacal depth (as percentages of total body length). 6 Hagfish from the inner Gulf of Maine were sig- nificantly larger than hagfish collected in any other location (P<0.0001). The average lengths of the OGM and MAC samples (315 mm and 280 mm respectively) were not significantly different from that of the ENA population (290 mm; see Tables 4 and 5). Further, we are unaware of any records for M. glutinosa larger than 450 mm out- side of the Gulf of Maine or the adjacent conti- nental slopes. Wisner and McMillan (1995) re- ported total lengths of 117-501 mm for their WNA sample (n= 78). With deletion of the 13 animals known to be from the outer Gulf of Maine, the size range becomes 117-446 mm. This range, which still includes 15 animals from the outer Gulf of Maine,2 is within the size range reported for eastern North Atlantic M. glutinosa (maxi- mum size of 450 mm. It may also be significant that one of the OGM specimens, 350 mm in total length, contained fully mature eggs (SI075-689, from 42°40.5' N, 66°37'W; Wisner1). This is above the size of sexual maturity for eastern North At- lantic M. glutinosa (200 mm) but below the ap- parent length at maturity for specimens in the IGM sample (400 mm; Martini et al., 1997a). The large maximum size and large size at maturity 2 Data forms did not permit determination of individual sizes, only ranges for these collection sites. 520 Fishery Bulletin 96(3), 1998 Relationships among the sampled populations of M. glutinosa for cusp counts and slime pore counts. A solid line connecting two populations indicates no significant difference exists between the absolute values. A dashed line indicates no significant difference exists between the values expressed as percentages of the total slime pore counts. (A) Total cusp counts; (B) prebranchial slime pore counts; (C) trunk slime pore counts; (D) tail slime pore counts; (E) total slime pore counts. NWA = Northwest Atlantic; IGM = inner Gulf of Maine; OGM = outer Gulf of Maine; ENA = eastern North Atlantic; and MAC = mid-Atlantic coast. that has been widely attributed to M. glutinosa in the western North Atlantic thus appears to be valid only for the inner Gulf of Maine. In summary, the IGM group is significantly larger in total length than any other population sampled. Proportional measurements could be compared only between the IGM and ENA samples. In addition to having much greater total length, hagfish from the IGM sample are more slender (both in width and depth) and have shorter prebranchial segments and shorter, narrower tails than ENA specimens. Allo- metric effects are probably not responsible for these differences, which remain significant even when the IGM data set is restricted to animals of the same size range as that in the ENA sample (<400 mm). Total length, slime pore counts, and total cusp counts could be compared among all groups. The highest variability observed was in the total slime pore count. Specimens from the Gulf of Maine (IGM and OGM) differed from one another and from all other sample groups. In terms of the regional distribution of slime pores, there were significant differences between the IGM and ENA samples when compared as absolute values or as percentages of the total slime pore count. Differences in the regional distribution of slime pores Martini et at: A population profile for Myxine glutinosa 521 Table 3 Results of post-hoc multiple comparison tests (Scheffe F-test) at the 5% level of significance. For counts, means, and standard deviations, refer to Table 2. The asterisks (*) indicate values significant at the 5% level (Scheffe F-test). Groups compared Mean diff. P-value Groups compared Mean diff. P-value Prebranchial slime pore counts, as value and percentage of IGM vs. OGM -0.276 (-0.010) 0.9987 (0.0694) total (in parentheses) MAC vs. NWA 0.660 (0.001) 0.9538 (0.9995) IGM vs. MAC 5.216 (0.020) <0.0001* (0.4127) MAC vs. OGM -1.271 (-0.004) 0.2460 (0.9376) IGM vs. NWA 7.517 (0.033) <0.0001* (0.1905) MAC vs. ENA -0.201 (-0.006) 0.9940 (0.1040) IGM vs. OGM 2.756 (0.005) 0.2380 (0.9979) NWA vs. ENA -0.861 (-0.007) 0.8116 (0.6125) IGM vs. ENA 6.045 (0.014) <0.0001* (0.0915) NWA vs. OGM -1.932 (-0.006) 0.1156 (0.9204) MAC vs. NWA 2.301 (0.019) 0.7105 (0.8340) OGM vs. ENA 1.075 (-0.001) 0.3745 (>0.9999) MAC vs. OGM -2.46 (0.009) 0.4471 (0.9913) MAC vs. ENA .828 (-0.010) 0.8798 00.9999) Total slime pore counts NWA vs. ENA -1.472 (-0.019) 0.9448 (0.8091) IGM vs. MAC 14.227 <0.0001* OGM vs. NWA 4.761 (0.028) 0.0865 (0.6534) IGM vs. NWA 18.429 <0.0001* OGM vs. ENA 3.288 (0.009) 0.0661 (0.9849) IGM vs. OGM 7.566 0.0039* IGM vs. ENA 17.38 <0.0001* Trunk slime pore counts as value and percentage of total MAC vs. NWA 4.203 0.6730 (in parentheses) MAC vs. OGM -6.661 0.0353* IGM vs. MAC 8.550 (0.002) <0.0001* (0.9985) MAC vs. ENA 3.154 0.0922 IGM vs. NWA 9.197 (-0.017) <0.0001* (0.2559) NWA vs. OGM -10.863 0.0054* IGM vs. OGM 4.838 (0.004) 0.0022* (0.9985) NWA vs. ENA -1.076 0.9995 IGM vs. ENA 11.170 (0.010) <0.0001* (0.0212)* OGM vs. ENA 9.787 <0.0001* MAC vs. NWA 0.647 (-0.019) 0.9997 (0.2076) MAC vs. OGM -3.712 (0.002) 0.0861 (>0.9999) Total cusp counts MAC vs. ENA 2.619 (0.008) 0.0030* (0.3125) IGM vs. MAC 0.432 0.9625 NWA vs. ENA 1.972 (0.027) 0.8591 (0.0054)* IGM vs. NWA 3.409 0.0006* NWA vs. OGM -4.359 (0.022) 0.2478 (0.2986) IGM vs. OGM 0.235 >0.9999 OGM vs. ENA 6.331 (0.006) <0.0001* (0.9760) IGM vs. ENA 1.883 <0.0001* MAC vs. NWA 2.978 0.0109* Tail slime pore counts as value and percentage of total MAC vs. ENA 1.667 0.0003* (in parentheses) MAC vs. OGM -0.197 >0.9999 IGM vs. MAC 0.995 (-0.006) 0.0166* (0.0694) NWA vs. ENA -1.310 0.7119 IGM vs. NWA 1.655 (-0.005) 0.1012 (0.9683) NWA vs. OGM -3.175 0.1263 IGM vs. ENA 0.794 (-0.012) 0.0088* (<.0001)* OGM vs. ENA 1.864 0.4313 between the groups IGM-MAC, IGM-NWA, IGM- OGM, and MAC-ENA were not significant when com- pared as percentages of the total slime pore count. Total cusp counts differed significantly between NWA-IGM, NWA-MAC, IGM-ENA, and MAC-ENA. Discussion The primary goal of this analysis was to assess the validity of the proposed splitting of M. glutinosa L. into two species. With the exception of the IGM data, the proposal by Wisner and McMillan ( 1995) was based on a general morphological comparison of animals col- lected from the eastern and western North Atlantic. This study used their data, assigned to the NWA, OGM, MAC, and ENA groups. As is often the case when work- ing with fishes whose lifestyles, habits, and population dynamics are poorly understood, there are a number of potential complicating factors that could bias these data. For example, both the Wisner and McMillan study and our own have compared specimens collected by 1) different methods, 2) at different times, and 3) at dif- ferent depths. This is not unusual; the majority of the 59 currently recognized species of hagfishes are known from small numbers of animals (often just one) caught accidentally in mobile fisheries gear. The limitations of our data sets are therefore shared not just with Wisner and McMillan ( 1995 ) but with many other stud- ies of hagfish systematics. Our goal was to determine if — given the limitations of the available data — the pro- posed split could be justified on the basis of the avail- able morphological data. We will now discuss each of these complicating factors as they influence hagfish collections in gen- eral, and the data in this study in particular. The collection method might affect the size range of animals captured Comparative data on trawl versus trap collection of Myxine are unavailable, but it is known that trawl- 522 Fishery Bulletin 96(3), 1 998 Table 4 Size distribution in populations of Myxine glutinosa in the eastern and western North Atlantic. Total length measure- ments (in mm ) reported for eastern and western Atlantic popu- lations (all groups). See text for explanation of abbreviations. NWA OGM IGM MAC ENA 315 509 280 299 Average 190-405 215-510 195-724 220-380 227-392 range (rc=ll)2 (n= 13)7 (n=306)3 (n=13)J (n = 179)1 529 312 103-405 170-950 220-420 253-362 (n= 8)2 ( /i = 1 1 72 ) 4 (n=37)2 (n= 8)5 <4506 ‘ From Wisner and McMillan (1995). 2 From raw data provided by R. Wisner (see Footnote 1 in main text). 3 From Martini et al. (1997a). 4 From Kuenstner ( 1996). 5 From Fernholm (1981). 6 From Adams and Strahan (1963). ing estimates of hagfish population density produce gross underestimates (Wakefield, 1990). Given the maximum swimming rate of adult Myxine (<1 m/sec, Foss, 1968), hagfish of any size are probably unable to outrun a trawl. We would predict that trapping would collect smaller individuals that might slip through the trawl mesh, whereas trawling could col- lect mature animals that might not be attracted to traps (breeding hagfish may not feed; Walvig, 1963). However, because we have no indication of allomet- ric effects of body size on any morphometric charac- ter for M. glutinosa, the statistical comparisons of slime pore counts, cusp counts, or proportional mea- surements should be unaffected by variations in to- tal length among the sample populations. Neither traps with large-bore entrances or trawl nets, the collection methods used for these data, should bias the maximum recorded size. Collections made at different times of the year may produce biased samples owing to seasonal migrations or breeding activities Although no large-scale tagging studies have been performed, field observations and their relatively inefficient and slow swimming speed suggest that M. glutinosa are relatively sedentary animals with small home ranges. Among hagfishes, only Eptatretus burgeri is known to have seasonal migrations (Fernholm, 1974), and their migration is related to a specific breeding cycle that is unique among hagfishes. Myxine glutinosa has no particular breed- ing period, and adults at all stages of gonadal devel- opment are present throughout the year (Walvig, Table 5 Results of post-hoc multiple comparison tests of total lengths; asterisks (*) indicate values significant at the 5% level (Scheffe /'’-test). See text for explanation of abbreviations. Groups compared Mean difference P-value IGM vs. MAC 231.490 <0.0001* IGM vs. NWA 213.798 <0.0001* IGM vs. OGM 196.381 <0.0001* IGM vs. ENA 213.099 <0.0001* MAC vs. NWA -17.692 0.9996 MAC vs. OGM 35.109 0.9698 MAC vs. ENA -18.391 0.9943 NWA vs. ENA -1.217 >0.9999 NWA vs. OGM -17.417 0.9996 OGM vs. ENA 16.718 0.9904 1963; Martini et al., 1997b). There were no signifi- cant differences in the population profiles for ani- mals collected at our primary study site in June- September from one year to the next, and although weather and sea-state conditions prohibited collec- tion in midwinter, no population differences, in terms of size range or abundance, were apparent in ROV surveys performed in December-January as com- pared to July-August (Martini, personal obs.). Collections made at different depths may yield different sex ratios and population profiles The reported depth range of M. glutinosa is exten- sive (50-1100 m). There are no reports of depth strati- fication by size or sex for any species of Myxine, and only suggestions of unequal depth distribution (by size and sex) for two species of Eptatretus (E. stouti and E. deani) (Johnson, 1994; Wakefield, 1990). All known M. glutinosa populations have sex ratios that are highly skewed in favor of females, but above the size at sexual maturity there are males and females of all sizes. Any variations in the sex ratios would not affect the mor- phological parameters we compared because no sexual dimorphism has been reported for these characters in M. glutinosa or any other hagfish species. The IGM data were collected from fresh, rather than preserved, specimens Fixation shrinkage of 10-15% may occur in preserved specimens (Wisner1). This shrinkage would not af- fect parameters such as cusp counts or slime pore counts, but it would potentially affect total length. Shrinkage alone, however, could not account for the magnitude of the observed differences in maximum total length (950 mm for IGM, versus a maximum of Martini et a!.: A population profile for Myxine glutinosa 523 510 mm elsewhere in the WNA, and 450 mm in the ENA) or the minimum size at sexual maturity (400 mm for the IGM vs. 200 mm for the ENA). Shrink- age may have affected some of the proportional com- parisons between the IGM and ENA samples, but the effect is not straightforward. The IGM hagfish (fresh measurement) have proportionately shorter prebranchial segments and tails than the ENA speci- mens (preserved measurement), and the IGM speci- mens are more slender (both in width and depth). The morphological measurements were made by different groups We do not believe that the differences observed among these populations reflect variation in proto- col between the research groups because differences in slime pore counts and cusp counts among the sampled populations remain if the IGM data are set aside, and a single team (Wisner and McMillan) col- lected the NWA, OGM, ENA, and MAC data sets. Conclusions Our analyses of the available data do not support a clean division between eastern and western North Atlantic populations of M. glutinosa. We therefore support retention of the species name M. glutinosa for both eastern and western populations pending the results of mtBNA analysis or other molecular comparisons. Specific arguments against species or subspecies division include the following: 1) Considerable vari- ability exists in the total slime pore count among the WNA groups examined; 2) Although there were sig- nificant regional differences in slime pore count be- tween the IGM and ENA samples, when compared as absolute values or as percentages of the total slime pore count, this was not the case for the other WNA groups; 3) Total cusp counts were also variable, with significant differences noted between NWA-IGM, NWA- MAC, IGM-ENA, and MAC-ENA; and 4) The degree of differentiation versus the ENA group, in increasing order, would be NWA (trunk slime pore percentage) — » OGM (total slime pore count) — > MAC (trunk slime pore count and cusp count) — » IGM (to- tal, trunk, and tail slime pore counts and cusp count). This pattern suggests, but does not prove, the ex- istence of clinal variations that may reflect the de- gree of relative isolation of the populations. Although on a map the Gulf of Maine appears continuous with the western North Atlantic, in fact it is almost com- pletely isolated from the offshore waters by exten- sive banks and shoals (Fig. 1A). There is only one deepwater connection (260-270 m) between the Gulf of Maine and the North Atlantic. This narrow con- nection, the Northeast Channel, extends between Brown’s Bank (BB), where the majority of the OGM samples were collected, and a shallow ridge that ex- tends northeast from George’s Bank (GB). Oceano- graphically the Gulf of Maine resembles a landlocked sea, like the Mediterranean, rather than a contigu- ous portion of the North Atlantic. For example, the salinity, temperature, tides, and current dynamics of the Gulf of Maine are distinct from those of the North Atlantic. This combination of factors could iso- late hagfish populations within the Gulf of Maine from those elsewhere in the western North Atlantic. The distinct characteristics of the Gulf of Maine en- vironment may also be linked to physiological differ- ences between M. glutinosa in the Gulf of Maine ver- sus the eastern North Atlantic. For example, M. glutinosa in the eastern North Atlantic have been maintained at water temperatures as high as 15°C (Palmgren, 1927; Gustafson, 1935), whereas speci- mens at the Shoals Marine Laboratory (Gulf of Maine) quickly become moribund as temperatures approach 10°C (Martini, personal obs.). The size difference between the IGM and other sampled populations could be a function of age, with members of the IGM population having longer lifespans. However, this is impossible to evaluate without growth rate or longevity data — which does not presently exist for this or any other species of hagfish. Alternatively, the large total size and large size at maturity for hagfish within the Gulf of Maine may reflect the relative availability of food. It is in- teresting to note that Kendall described winter floun- der (Pleuronectes americanus [formerly Pseudo- pleuronectes americanus ] ) of the George’s Bank re- gion as a separate species from those found elsewhere in the western North Atlantic, on the basis of its unusually large size; this proposal was ultimately rejected because there were no other morphological differences. It is not clear, however, that the ecosys- tem in the inner Gulf of Maine is significantly more productive than that of the Skaggerack in the east- ern North Atlantic; both areas support substantial commercial fisheries. Wisner and McMillan (1995) proposed species sepa- ration based on maximum size, size at maturity, and differences in the color of preserved specimens. As indicated above, the large maximum size and large size at maturity appear to be characteristic of the IGM population only, rather than a general charac- teristic of WNA populations of Myxine. We remain unable to evaluate the significance of the color dif- ferences in preserved specimens noted by Wisner and McMillan (1995) for eastern versus western North 524 Fishery Bulletin 96(3), 1 998 Atlantic Myxine. The patterns they described as typi- cal for the WNA are not found in our preserved speci- mens from the IGM. Further, our field data, which included ROV and submersible observations and trap collections, indicate that the described color patterns are not characteristic of living members of the inner Gulf of Maine populations. Whether they are char- acteristic of other WNA populations, or typical of only a few subpopulations, remains to be determined. Acknolwedgments This study was funded in part by a grant from the National Oceanic and Atmospheric Administration’s National Undersea Research Center at the Univer- sity of Connecticutt, Avery Pt., CT. The crew of the research vessel Seward Johnson and the Johnson SeaLink (Harbor Branch Oceanographic Institution) made possible direct underwater observations of hagfish at study sites within the Gulf of Maine. Their professionalism and their tolerance of the logistical complications introduced by large quantities of hag- fish slime are greatly appreciated. Time aboard the research vessel John M. Kingsbury and both person- nel and logistical support were generously provided by the Shoals Marine Laboratory, Cornell Univer- sity, Ithaca, NY. Robert L. Wisner (Scripps Institute of Oceanography, University of California, San Di- ego, CA) provided the raw data from Wisner and McMillan ( 1995) and additional information on hag- fish collections. Data extracted from these sources are included in the Tables above. Alexander Gryska and Susan Kuenstner of the New England Fisheries Development Association provided catch statistics and length data from their work with the Gulf of Maine hagfish fishery. Literature cited Abacus Concepts Inc. 1995. Statview 4.51. Abacus Concepts Inc., 1918 Bonita Ave., Berkeley, CA. Adam, H., and R. Strahan. 1963. Systematics and geographical distribution of myxi- noids. In A. Brodal and R. Fange (eds.), The biology of Myxine , p. 1-8. Universitetsforlaget, Oslo. Fernholm, B. 1974. Diurnal variations in behavior of the hagfish, Eptatretus burgeri. Mar. Biol. 27: 351-356. 1981. A new species of hagfish of the genus Myxine, with notes on other eastern Atlantic myxinids. J. Fish Biol. 19:73-82 Fernholm, B., and C. L. Hubbs. 1981. Western Atlantic hagfishes of the genus Eptatretus (Myxinidae) with description of two new species. Fish. Bull. 79:6-83. Foss, G. 1968. Behavior of Myxine glutinosa in natural habitat; in- vestigation of the mud biotype by a suction technique. Sarsia 31:1-13. Girard, C. 1858. Ichthyological notes. Proc. Acad. Nat. Sci. of Phila- delphia, p. 223-224. Gustafson, G. 1935. On the biology of Myxine glutinosa L. Ark. f. Zool. 28(2): 1-8. Johnson, E. W. 1994. Aspects of the biology of Pacific ( Eptatretus stouti) and black ( Eptatretus deani) hagfishes from Monterey Bay, California. M.S. thesis, School of Natural Sciences, Cali- fornia State Univ., Fresno, CA, 130 p. Kendall, W. C. 1912. Notes on a new species of flatfish from off the coast of New England. Bull. U. S. Bur. Fish. 30:391-394. Kuenstner, S. E. 1996. Harvesting the value-added potential of Atlantic hag- fish. New England Fisheries Development Assoc., Bos- ton, MA, 46 p. Lesser, M., F. H. Martini, and J. B. Heiser. 1996. Ecology of hagfish, Myxine glutinosa L., in the Gulf of Maine. I. Metabolic rates and energetics. J. Exp. Mar. Biol. Ecol., 208:215-225. Linnaeus, C. 1758. Systema naturae, per regna tria secundum classes, ordines, genera, species cum characteribus, differentiis, synonymis, locis, editio decima, reformata, tomus I. Laurentii Salvii, Holmiae, 824 p. Martini, F. H., J. B. Heiser, and M. P. Lesser. 1997a. A population profile for hagfish, Myxine glutinosa L., in the Gulf of Maine. Part 1: Morphometries and repro- ductive state. Fish. Bull. 95:311-320. Martini, F., M. Lesser, and J. B. Heiser. 1997b. Ecology of the hagfish, Myxine glutinosa, L., in the Gulf of Maine: II. Potential impact on benthic communi- ties and commercial fisheries. J. Exp. Mar. Biol. Ecol. 214:97-106. McMillan, C., and Wisner, R. L. 1984. Three new species of seven-gilled hagfishes (Myxini- dae, Eptatretus) from the Pacific Ocean. Proc. Cal. Acad. Sci. 43(6):249-267 Palmgren, A. 1927. Aquarium experiments with the hag-fish (Myxine glutinosa L .). Acta Zool. VIII: 1-16. Wakefield, W. W. 1990. Patterns in the distribution of demersal fishes on the upper-continental slope off central California, with stud- ies on the role of ontogenetic vertical migration on particle flux. Ph.D. diss., Scripps Institute of Oceanography, Univ. Califonia at San Diego, La Jolla, CA, 178 p. Walvig, F. 1963. Gonads and the formation of sexual cells. In A. Brodal and R. Fange (eds.), The biology of Myxine, p. 530- 580. Universitatsforlaget, Oslo. Wisner, R. L., and C. B. McMillan. 1995. Review of the new world hagfishes of the genus Myxine (Agnatha, Myxinidae) with descriptions of nine new species. Fish. Bull. 93:530-550. 525 Abstract .—We used allozyme analyses to investigate genetic varia- tion among commercially exploited populations of Chionoecetes bairdi (Tanner) and C. opilio (snow) crabs in Alaskan waters. Data were collected from 34 presumptive loci in 1002 C. bairdi and 539 C. opilio sampled throughout the commercially important range of each species in Alaska. Aver- age observed heterozygosities were 0.027 for C. bairdi and 0.013 for C. opilio. Low levels of geographic differ- entiation were detected among popula- tions of C. bairdi and C. opilio , and our data suggest that subpopulations of C. bairdi exist within the Bering Sea. Fur- ther, evidence of gene introgression was found between C. bairdi and C. opilio in the Bering Sea. We also included a geographic isolate, North Atlantic C. opilio, in the analyses. Little differen- tiation was found, and no private alle- les were detected in North Atlantic C. opilio despite significant geographic separation from Alaskan C. opilio. Manuscript accepted 24 October 1997. Fishery Bulletin 96:525-537 ( 1998). Low levels of genetic diversity in highly exploited populations of Alaskan Tanner crabs, Chionoecetes bairdi, and Alaskan and Atlantic snow crabs, C. opilio* Susan E. Merkouris Lisa W Seeb Genetics Laboratory Commercial Fisheries Management and Development Division Alaska Department of Fish and Game 333 Raspberry Road, Anchorage, Alaska 995 1 8-1 599 E-mail address (for S E Merkouris): SueMOfishgame. state. ak. us Margaret C. Murphy PO. Box 25526 Juneau, Alaska 99802-5526 Five species of the genus Chio- noecetes, Majidae, are described from the North Pacific region (Rathbun, 1925; Garth, 1958). The nearly circumpolar range of C. opilio (snow crab) includes the Bering Sea, the Arctic Ocean, the western North Pacific coast of Asia, and the northern Atlantic Ocean. Chionoecetes bairdi , C. angulatus, and C. tanneri are widespread in the eastern North Pacific (Garth, 1958). Chionoecetes japonicus is found only in the western North Pa- cific along the coast of Asia. In the Alaskan waters of the Bering Sea, the distribution of C. bairdi is strongly associated with the conti- nental slope areas along the coast of the Alaska Peninsula, and the Pribilof Islands (Otto, 1982), and there is considerable overlap in the distribution of C. bairdi and C. opilio (Karinen, 1974). Commercial fisheries for male C. bairdi and C. opilio in Alaska, along with king crabs ( Paralithodes and Lithodes) have long been the world’s most abundant sources of crabs and have considerable current and his- torical commercial importance in Alaska (Otto, 1990). Chionoecetes bairdi are faster growing, larger, and more valuable than C. opilio. Commercial harvests of C. bairdi in the Bering Sea and Gulf of Alaska fisheries peaked in the late 1970s, declined throughout the 1980s, and although some populations of C. bairdi have recently rebounded, many fisheries remain closed owing to low abundance (Kruse, 1993). Rapid development of the Bering Sea C. opilio fishery coincided with the decline of C. bairdi fisheries, and although C. opilio have domi- nated landings of Bering Sea crabs since the mid-1980s, catches have also declined dramatically in recent years (ADF&G, 1994). Declining abundances of Bering Sea and Gulf of Alaska crab popu- lations have intensified competition for the remaining resources and have lead to re-evaluation of crab fishery management practices. Gen- * Contribution number PP-123 of the Alaska Department of Fish and Game, Commer- cial Fisheries Management and Develop- ment Division, Juneau, Alaska. 526 Fishery Bulletin 96(3), 1998 erally when a population or stock is modeled, the assumption is made that the individuals within the population are uniform, but it is important that this assumption be verified (Cobb and Caddy, 1989). In practice, a stock is often defined as 1) a production or management unit about which conclusions can be made without regard for differences within the group and exchanges with other groups, or as 2) biologi- cally, a genetically discrete population (reviewed in Cobb and Caddy, 1989). However, stock identity of decapod populations has been largely ignored or prob- lematic (Cobb and Caddy, 1989; Kruse, 1993). Existing shellfish management units in Alaska were originally established according to historical fishing grounds of red king crab. Current area lines for C. bairdi and C. opilio were based on mark-and- recapture data, natural geographic barriers, and ar- eas of stock abundance grouped by major fishing grounds. Although genetic surveys were not con- ducted on unexploited Chionoecetes populations in Alaska, the establishment of a genetic baseline is critical for verifying current fishery management units and for monitoring potential fishery-induced genetic changes. Additionally, the genetics of C. bairdi and C. opilio populations in the Bering Sea is complicated by the presence of hybrids of these two species. Chionoecetes bairdi and C. opilio appear to have many similar morphological, physiological, and reproductive features (Watson, 1970; Slizkin, 1990) that allow them to hybridize in areas of range over- lap (Karinen and Hoopes, 1971; Somerton, 1981; Hoopes et al.1 ). Many studies suggest that commercial fishing ac- tivities may have significant genetic effects on fish stocks without reducing them to near extinction (for example, see reviews in Allendorf et al., 1987;Thorpe, 1993), and genetic selection against fast growth may result from intense fishing pressure (Kruse, 1993; Stevens et al., 1993); however, these effects can be assessed only if there are comparative baseline data. In 1990 we began genetic investigations of C. bairdi and C. opilio populations in Alaska using allozyme electrophoresis. Our objectives were 1) to assess ge- netic variation in exploited populations of C. bairdi and C. opilio in Alaska and 2) to determine if signifi- cant differentiation exists to warrant re-examination of current management units. Materials and methods Population sample collections Samples of C. bairdi and C. opilio were obtained be- tween 1989 and 1993 from population assessment Table 1 Collection information for Chionoecetes specimens. Location number Location2 Date n Chionoecetes bairdi l2 Bristol Bay Jun 90 50 Bristol Bay Jun 91 50 22 Bering Sea and Pribilof Islands Jul 90 50 Bering Sea and Pribilof Islands Apr 91 75 Bering Sea and Pribilof Islands Jul 92 30 Bering Sea and Pribilof Islands Mar 93 50 Bering Sea and Pribilof Islands Jul 93 80 3 Port Moller Jun 90 42 4 Sand Point and Pavlof Bay Aug 90 50 5 Kodiak N. Mainland Feb 90 50 6 Kodiak S. Sitkalidak Strait Feb 90 50 7 Kamishak Bay Jun 90 50 8 Montague Strait Jul 90 50 9 Kachemak Bay Jun 90 50 10 Prince William Sound Aug 90 50 11 Sullivan Island Jul 93 100 12 Seymour Canal Sep 89 50 Seymour Canal Jul 93 75 Chionoecetes opilio I2 Bering Sea Apr 91 75 Bering Sea Mar 93 50 II2 St. Matthew Island Jul 90-Aug 90 100 St. Matthew Island Jul 92-Aug 92 100 III2 Pribilof Island Jul 90 44 Pribilof Island Aug 90 50 Pribilof Island Jul 92 40 Pribilof Island Jun 93-Jul 93 80 Nova Scotia, North Atlantic Sep 91 97 1 Latitude and longitude locations available from authors. - Collections within geographic location pooled for analyses; C. bairdi collection sites in the Bering Sea and Pribilof Islands were overlapping. trawl and pot survey catches, test fishery pot catches, and dockside commercial pot catches. Crabs caught in pots set in adjacent areas were pooled into a single collection (Table 1). Chionoecetes bairdi were col- lected from sites ranging from Seymour Canal in Southeast Alaska to northwest of the Pribilof Islands in the Bering Sea (Table 1; Fig. 1). Chionoecetes opilio were collected from sites in the Bering Sea. In addi- tion, we obtained a collection of C. opilio from the North Atlantic for comparison with Bering Sea samples. Eighteen collections of C. bairdi and nine collections of C. opilio were analyzed in this study. 1 Hoopes, D. T., J. F. Karinen, and M. J. Pelto. 1970. King and Tanner crab research. International North Pacific Fisheries Commission Annual Rep. 1970:110-120. Merkouris et al.: Genetic diversity in Chionoecetes bairdi and C opilio 527 Tissues (muscle, gill, hepatopancreas, and heart) were dissected from each individual, placed in a la- belled tube, which was chilled on wet ice, capped, and frozen at -15°C (1989-90 collections) or in liq- uid nitrogen (1991-93 collections). Freezing gener- ally occurred within 20 minutes after dissection but in some cases was as long as 1 hour. Tissues were transported to the laboratory on dry ice or liquid ni- trogen and stored at -80°C until analysis. Several locations were sampled in multiple years. Allozyme electrophoresis Procedures for horizontal starch gel electrophoresis followed those of Harris and Hopkinson (1976) and Aebersold et al. (1987). Activity reflecting 34 pre- sumed loci were resolved with the following buffers (Table 2): 1) N-(3-aminopropyl)-morpholine, citrate (AC, pH 6.1, 6.9) (Clayton and Tretiak, 1972); 2) Tris, borate, citrate, lithium hydroxide (TBCL, pH 8.7) (Ridgway et al., 1970); 3) Tris, citrate (TC, pH 7.0) (Shaw and Prasad, 1970); 4) Tris, citrate (TC, pH 8.0) (Selander et al., 1971); and 5) Tris, borate, EDTA (TBE, pH 8.5) (Boyer et al., 1963). Gene nomencla- ture followed Shaklee et al. (1990). Allelic standards were used to compare relative mobilities between species. Genetic differentiation We estimated allele frequencies, calculated average observed heterozygosities, and tested conformation of genotype frequencies to Hardy-Weinberg expected frequencies with log-likelihood ratios (modified from Weir, 1990) for each collection (a=0.05, adjusted for multiple tests [Rice, 1989] ). Samples ofC. bairdi col- lected from the Bering Sea and Pribilof Island areas were pooled because sample sites were overlapping. Interannual heterogeneity of multiple-year collec- tions in Bristol Bay, Bering Sea, St. Matthew Island, Pribilof Islands, and Seymour Canal was tested by using log-likelihood statistics (Sokal and Rohlf, 1995). Multiple-year collections within a site were pooled for further analyses if no significant heterogeneity existed (P<0.01). We estimated degree of population subdivision by using Wright’s ( 1978) nonhierarchical F statistics and a hierarchical log-likelihood analy- sis. We used FSTAT (version 1.2)(Goudet, 1995) to 528 Fishery Bulletin 96(3), 1998 Table 2 Allozyme protocols to resolve enzyme coding loci in Chionoecetes samples. Enzyme nomenclature follows Shaklee et al. (1990), and locus abbreviations are given. Tissue abbreviations are: (M) muscle, (H) heart, (G) gill, and (P) hepatopancreas. Buffers are described in the text. Enzyme or protein Enzyme number Locus Tissue Buffer Aspartate aminotransferase 2.6.1. 1 AAT-1*, AAT-2* M TBE Aconitate hydratase 4.2. 1.3 AH-2*2 M TC7.0 AH-3* M TC7.0 Adenosine deaminase 3. 5. 4.4 ADA-1*, ADAS* G,M,H TBCL ADAS* P TBCL Alanine aminotransferase 2.6. 1.2 ALAT* M TBCL Cytochrome-b52 1.8. 1.4 CBYR* M TBCL p- N - Acetylgalactosaminidase 3.2.1.53 PGALA* G,P TBCL, TBE Glyceraldehyde-3-phosphate dehydrogenase 1.2.1.12 GAPDH* M AC6.1 N-Aeetyl-/)-glucosaminidase 3.2.1.30 PGLUA* G,P TBCL Glycerol-3-phosphate dehydrogenase 1.1. 1.8 G3PDH-1* M TC7.0 G3PDH-2* H,M TC7.0 Glucose-6-phosphate isomerase 5.3. 1.9 GPI-A1* M TBCL Isocitrate dehydrogenase (NADP+) 1.1.1.42 IDHP-1* H TC7.0 Malate dehydrogenase 1.1.1.37 MDH-A1*, MDH-A2* M AC6.1 Malic enzyme (NADP+) 1.1.1.40 MEP-1* H TC7.0 Mannose-6-phosphate isomerase 5.3. 1.8 MPI* M TBE Dipeptidase 3.4.-.- PEPA* H,P TBE Proline dipeptiase 3.4.13.9 PEPD-2* G,P TBCL Phosphogluconate dehydrogenase 1.1.1.44 PGDH* H TC7.0 Phosphoglucomutase 5.4.2. 2 PGM-1* M,H TC7.0 General (unidentified) protein PROT-1*, PROT-2*, PROT-3* M TBCL Superoxide dismutase 1.15.1.1 SOD-1*2 M TBCL Triose-phosphate isomerase 5.3. 1.1 TPI-1* M TBE 1 This locus expressed with both CBYR* and GR*; preferential stain is CBYR* 2 Used in C. opilio statistical analysis only. test significance of FgT values (2000 permutations). Allele frequencies were used to generate distance matrices of Cavalli-Sforza and Edwards chord dis- tances (Cavalli-Sforza and Edwards, 1967). Compu- tations were made with S-Plus analytical software (MathSoft Inc., 1997). Results Chionoecetes bairdi Twenty-seven loci were scored consistently in C. bairdi populations and were used in the data analyses. Tem- poral variation was examined for multiple-year collec- tions of C. bairdi from Bristol Bay, the Bering Sea and Pribilof Islands area, and Seymour Canal (Table 1). No significant interannual differences (P<0.01 ) were found within each geographic location; therefore these col- lections were pooled for subsequent analyses. Fifteen loci, AAT-1*, AAT-2*, AH-3*, ALAT *, CBYR*, G3PDH-1*, G3PDH-2*, GPI-A 1 *, IDHP-1*, MDH-A1*, PEPA*, PGDH*, PGM-1 *, PROT-3* , and TPI-1* were polymorphic (Table 3). At four loci, G3PDH-1*, GPI-A1 *, IDHP-1 *, and PEPA*, the most common allele had a frequency of <0.95 in at least one population. Twelve monomorphic loci, ADA-1*, ADA-2*, ADA-3*, pGALA*, GAPDH*, pGLUA*, MDH-A2*, MEP-1 *, MPI*, PEPD-2*, PROT-1 * , and PROT-2* were included in the 27-locus analyses. Seven monomorphic loci, AH-2*, mMDH-1* , PEPD- 1*, SOD-1*, SOD-2*, TPI-2*, and XO* were not scorable in all populations and were not included in our analyses. We could not interpret the genetic ba- sis for the variation we observed in two consistently resolved zones of esterase activity; therefore those data were excluded. Three collections of C. bairdi. Sand Point and Pavlof, Port Moller, and Prince William Sound, with sample sizes of <25 at informative loci, AH-3* and IDHP-1*, were not included in the population data analyses; how- ever, we report allele frequencies in Table 3. Genotype frequencies at all loci conformed to Hardy- Weinberg expectations; therefore we assumed all sam- Merkouris et al.: Genetic diversity in Chionoecetes bairdi and C. opilio 529 Table 3 Allele frequency estimates for Chionoecetes bairdi and C. opilio collections, quency of 0.000. ND indicates no data. Dashed lines indicate fre- Population2 AAT-l * AAT-2* AH-2* n *100 *64 *210 *120 n *100 *131 *69 n *100 *110 *83 Chionoecetes bairdi Bristol Bay pooled 100 0.995 — — 0.005 100 1.000 — - ND Bering Sea pooled 283 1.000 - - — 274 0.998 — 0.002 280 1.000 — — Port Moller 1990 42 1.000 — — — 42 1.000 — — 42 1.000 — — Sand Pt/Pavlof 1990 37 1.000 — — — 47 1.000 — — ND Kodiak N. 1990 48 1.000 — — — 48 0.990 0.010 — 38 1.000 — — Kodiak S. 1990 43 0.988 — — 0.012 43 0.988 0.012 — 36 1.000 — — Kamishak 1990 50 0.990 0.010 — — 50 1.000 — — 50 1.000 — — Montague St. 1990 42 1.000 — — — 49 1.000 — — 50 1.000 — — Kachemak Bay 1990 50 1.000 — — — 50 0.990 0.010 — 50 1.000 — — Prince Wm. Sd. 1990 50 1.000 — — — 50 1.000 — — 50 1.000 — — Sullivan Is. 1993 100 1.000 — — — 100 0.995 0.005 — 100 1.000 — — Seymour C. pooled 125 1.000 - — — 123 1.000 - — 125 1.000 - - Chionoecetes opilio Bering Sea pooled 124 0.992 — 0.004 0.004 123 1.000 — — 115 0.991 0.009 — St. Matt. Is. pooled 192 0.992 — — 0.008 193 0.987 0.005 0.008 200 1.000 — — Pribilof Is. pooled 211 0.991 0.005 — 0.005 193 0.995 0.003 0.003 214 0.993 0.002 0.005 Atlantic 0. 1991 97 1.000 - — - 97 0.995 0.005 - 97 1.000 - — AH-3* ALAT* CBYR* G3PDH-1 * G3PDH-2*2 Population' n *100 *94 *91 *106 n *100 *87 n *100 *117 n *100 *117 n *100 *86 *111 Chionoecetes bairdi Bristol Bay pooled 100 0.005 0.995 — — 100 1.000 — 100 1.000 — 96 0.974 0.026 100 1.000 — — Bering Sea pooled 279 0.014 0.982 0.004 — 283 0.998 0.002 285 0.998 0.002 265 0.958 0.042 196 0.949 0.008 0.043 Port Moller 1990 42 — 1.000 — — 42 1.000 — 42 1.000 — 41 0.976 0.024 20 0.950 0.025 0.025 Sand Pt/Pavlof 1990 3 — 1.000 — — ND 50 1.000 — 1 1.000 — ND Kodiak N. 1990 38 0.013 0.987 — — 50 1.000 — 50 1.000 — 46 0.967 0.033 50 0,990 — 0.010 Kodiak S. 1990 33 0.015 0.985 — — 50 1.000 — 50 1.000 — 37 0.946 0.054 42 1.000 — — Kamishak 1990 50 — 1.000 — — 50 1.000 — 50 1.000 — 49 0.990 0.010 50 1.000 — — Montague St. 1990 49 — 1.000 — — 49 1.000 — 50 1.000 — 50 0.970 0.030 31 1.000 — — Kachemak Bay 1990 50 — 1.000 — — 50 1.000 — 50 1.000 — 50 0.950 0.050 50 0.970 — 0.030 Prince Wm. Sd. 1990 50 — 0.990 0.010 — 50 1.000 — 50 1.000 — 50 0.970 0.030 47 1.000 — — Sullivan Is. 1993 95 0.011 0.979 0.011 — 97 1.000 — 50 1.000 — 89 0.933 0.067 49 0.980 — 0.020 Seymour C. pooled 122 0.004 0.992 0.004 — 125 1.000 - 100 1.000 - 122 0.959 0.041 109 1.000 - - Chionoecetes opilio Bering Sea pooled 124 0.923 0.077 — — 123 1.000 — 125 0.996 0.004 125 0.988 0.012 96 0.995 — 0.005 St. Matt. Is. pooled 197 0.939 0.058 — 0.003 195 1.000 — 200 1.000 — 198 0.990 0.010 174 0.994 — 0.006 Pribilof Is. pooled 213 0.937 0.063 — — 214 1.000 — 214 1.000 — 214 0.998 0.002 183 0.992 0.003 0.005 Atlantic 0. 1991 95 0.937 0.063 - - 97 1.000 - 97 1.000 - 97 1.000 - 94 1.000 - - continued pies were from single panmictic populations. Average observed heterozygosities ranged from 0.016 to 0.035. Asnong-population variation Various measures indicated a low level of population differentiation. FST values were low, however two loci, G3PDH-2* and PEPA* with respective FgT values of 0.0137 (P=0.0005) and 0.0154 (P=0.0020), and the overall FgT value of 0.0046 for 15 polymorphic loci (P-0.0295) differed significantly from zero. A com- parison of all C. bairdi populations with a hierarchi- cal log-likelihood analysis, indicated significant het- erogeneity at four loci, G3PDH-2*, IDHP-1*, PEPA*, and PROT-3*; however the overall statistic was not significant (Table 4). In contrast, when populations 530 Fishery Bulletin 96(3), 1 998 Table 3 (continued) Population' GPl-Al* IDHP-1 * MDH-A1* n *100 *142 *60 *158 *9 *55 *19 n *100 *81 *119 *90 n *100 *79 *128 Chionoecetes bairdi Bristol Bay pooled 100 0.970 — 0.020 — 0.005 — 0.005 79 0.633 0.278 0.082 0.006 100 1.000 — — Bering Sea pooled 283 0.971 0.005 0.018 — 0.002 0.002 0.002 275 0.562 0.358 0.074 0.006 283 0.995 0.005 — Port Moller 1990 42 0.988 — 0.012 — — — — 9 0.611 0.389 — — 42 1.000 — — Sand Pt/Pavlof 1990 50 0.980 — 0.010 — 0.010 — — 34 0.691 0.162 0.147 — 8 1.000 — — Kodiak N. 1990 49 0.959 0.010 0.031 — — — — 49 0.592 0.357 0.051 — 49 1.000 — — Kodiak S. 1990 50 0.930 0.020 0.030 — 0.010 — 0.010 36 0.722 0.264 0.014 — 39 1.000 — — Kamishak 1990 50 0.990 — — — 0.010 — — 30 0.683 0.267 0.050 — 50 1.000 — — Montague St. 1990 50 1.000 — — — — — — 50 0.510 0.430 0.060 — 50 1.000 — — Kachemak Bay 1990 50 0.980 — 0.010 — 0.010 — — 28 0.571 0.286 0.107 0.036 50 1.000 — — Prince Wm. Sd. 1990 49 0.990 — — — 0.010 — — 7 0.643 0.214 0.143 — 50 1.000 — — Sullivan Is. 1993 100 0.985 - 0.015 — — — — 82 0.634 0.341 0.024 — 99 1.000 — — Seymour C. pooled 124 0.992 - 0.004 - 0.004 - - 122 0.574 0.348 0.078 - 125 0.996 - 0.004 Chionoecetes opilio Bering Sea pooled 125 0.984 0.004 0.004 0.004 — 0.004 — 125 1.000 — — — 125 0.992 0.008 — St. Matt. Is. pooled 200 0.988 0.013 — — — — — 196 1.000 — — — 200 0.973 0.027 — Pribilof Is. pooled 214 0.986 0.012 0.002 — — — — 207 0.990 0.005 0.005 — 213 0.986 0.014 — Atlantic O. 1991 97 0.974 0.026 - - - - - 97 1.000 - - - 97 0.990 0.010 - MDH-A2* PEPA* PGDH * PGM-1 Population' n *100 *65 n *100 *138 *64 n *100 *112 *124 *91 n *100 *82 *114 *103 *88 *75 Chionoecetes bairdi Bristol Bay pooled 100 1.000 — 92 0.940 0.022 0.038 92 0.995 0.005 — — 100 0.995 — — 0.005 — — Bering Sea pooled 283 1.000 — 279 0.991 0.004 0.005 269 0.996 — 0.004 — 268 0.996 0.004 — — — — Port Moller 1990 42 1.000 — 39 0.987 — 0.013 18 1.000 — — — 42 1.000 — — — — — Sand Pt/Pavlof 1990 8 1.000 — 40 0.975 — 0.025 47 1.000 — — — 13 1.000 — — — — — Kodiak N. 1990 49 1.000 - 48 0.990 0.010 — 40 1.000 — — — 47 0.989 — — — — 0.011 Kodiak S. 1990 40 1.000 — 43 0.977 0.012 0.012 34 1.000 — — — 38 0.987 0.013 — — — — Kamishak 1990 50 1.000 — 49 1.000 — — 45 1.000 — — — 50 0.980 0.010 — — — 0.010 Montague St. 1990 50 1.000 — 50 1.000 — — 49 1.000 — — — 42 1.000 — — — — — Kachemak Bay 1990 50 1.000 — 50 1.000 — — 46 0.989 — — 0.011 50 1.000 — — — — — Prince Wm. Sd. 1990 50 1.000 — 47 0.989 — 0.011 49 1.000 — — — 50 1.000 — — — — — Sullivan Is. 1993 99 1.000 — 92 0.995 — 0.005 84 0.994 — 0.006 — 99 0.995 0.005 — — — — Seymour C. pooled 125 1.000 - 118 0.983 0.004 0.013 119 0.992 — 0.008 — 125 0.992 — — — 0.008 — C. opilio Bering Sea pooled 125 1.000 — 125 0.992 0.008 — 123 0.980 — 0.020 — 125 0.992 0.008 — — — — St. Matt. Is. pooled 200 0.998 0.002 187 1.000 — - 183 0.970 — 0.030 — 198 0.982 0.005 0.003 0.003 0.008 — Pribilof Is. pooled 214 1.000 — 199 0.997 0.003 — 196 0.967 0.008 0.025 — 214 0.993 0.002 — — 0.005 — Atlantic O. 1991 97 1.000 — 97 0.995 0.005 — 97 0.985 0.005 0.010 - 97 0.964 - - - 0.036 - PROT-3* SOD-1* TPI-1 * Population' n *100 *88 n *100 *109 n *-100 *100 *-41 Chionoecetes bairdi Bristol Bay pooled 100 0.010 0.990 125 1.000 — 100 1.000 — — Bering Sea pooled 283 0.016 0.984 273 1.000 - 284 0.996 - 0.004 Port Moller 1990 42 — 1.000 39 1.000 — 42 1.000 — — Sand Pt/Pavlof 1990 50 — 1.000 50 1.000 — 47 0.979 — 0.021 Kodiak N. 1990 50 — 1.000 50 1.000 — 50 0.990 — 0.010 Kodiak S. 1990 50 — 1.000 50 1.000 - 47 1.000 — — Kamishak 1990 50 — 1.000 50 1.000 — 49 1.000 — — Montague St. 1990 50 - 1.000 50 1.000 - 50 1.000 - - continued Merkouris et at: Genetic diversity in Chionoecetes bairdi and C opilio 531 Table 3 (continued) PROT-3* SOD-1 * TPI-1* Population1 2 n noo *88 n *100 *109 n *-100 *100 *-41 Chionoecetes bairdi, continued Kachemak Bay 1990 50 — 1.000 50 1.000 — 50 1.000 — — Prince Wm. Sd. 1990 50 — 1.000 50 1.000 — 50 1.000 — — Sullivan Is. 1993 100 — 1.000 ND 100 1.000 — — Seymour C. pooled 125 — 1.000 125 1.000 — 125 1.000 — — Chionoecetes opilio Bering Sea pooled 124 0.992 0.008 122 1.000 - 125 0.992 0.004 0.004 St. Matt. Is. pooled 196 0.997 0.003 198 0.997 0.003 198 0.995 0.002 0.002 Pribilof Is. pooled 214 0.986 0.014 211 0.998 0.002 214 1.000 — — Atlantic 0. 1991 97 1.000 — 97 1.000 — 97 0.990 0.010 — 1 Port Moller, Sand Point/Pavlof Bay, and Prince William Sound C. bairdi populations not included in analyses. 2 G3PDH-1*87 was pooled with *100. were grouped by major geographic regions (Bering Sea, Gulf of Alaska, and Southeast Alaska) for com- parison, the overall log-likelihood statistic was highly significant (P<0.01, Table 4). G3PDH-2* and PROT- 3* contributed most to the observed among-region heterogeneity. Within the Bering Sea, pooled samples collected from Bristol Bay (east of approximately 162°45’W long.) were significantly different from pooled samples from the Bering Sea and Pribilof Is- land areas (west of approximately 167°00'W long.). In this comparison, G3PDH-2* , IDHP-1*, and PEPA * allelic frequencies differed significantly. The following were the low-frequency alleles detected according to their respective major geographic regions: 1) Bering Sea, AAT-2*69, ALAT*87, CBYR*117 , G3PDH-2*86 , GPI-A1*55, MDH-A1*79, PGDH*112, PGM- 1*103, and PROT-3* 100\ 2) Gulf of Alaska, AAT- 1*64, PGDH*91, and PGM-1*75\ and 3) Southeast Alaska, MDH-A1 *128 and PGM- 1 *88. The PROT-3* 100 allele, likely an introgressed C. opilio allele, contrib- uted significantly to the overall heterogeneity observed among the major geographic regions of Bering Sea, Gulf of Alaska, and Southeast Alaska. Genetic distance measurements (Cavalli-Sforza and Edwards, 1967) within C. bairdi populations ranged from 0.031 to 0.059. Chionoecetes opilio Twenty-nine loci were scored consistently in all col- lections and were used in the data analyses. No sig- nificant temporal variation (P<0.01) was found within multiple-year collections in the Bering Sea, St. Matthew Island, and the Pribilof Islands. We pooled these collections for further analyses. Seventeen loci, AAT-1*, AAT-2* , AH-2* , AH-3*, CBYR*, G3PDH-1*, G3PDH-2*, GPI-A1*, IDHP-1*, MDH-A1*, MDH-A2*, PEPA*, PGDH*, PGM-1*, PROT-3*, SOD- 1 *, and TPI- 1 *, were polymorphic in at least one population (Table 3). Twelve monomorphic loci were ADA-1*, ADA-2*, ADA-3*, ALAT*, pGALA*, GAPDH*, PGLUA*, MEP-1*, MP1 *, PEPD-2*, PROT- 1 *, and PROT-2*. One polymorphic locus, AH -3*, was vari- able at a frequency of >0.05 in at least one population. Genotype frequencies at all loci conformed to Hardy- Weinberg expectations; therefore we assumed all samples were from single panmictic populations. Average observed heterozygosities were lower than those for C. bairdi and ranged from 0.012 to 0.013 in Alaskan populations; the average heterozygosity was 0.010 in Atlantic Ocean C. opilio. Among-population variation Fst of 17 polymorphic loci was not significantly dif- ferent from zero for Alaskan populations (P=0.4270) or when the Atlantic Ocean collection was included (P=0.7130). However, the FgT value for PGM-1* was highly significant (P=0.0050) in comparisons of At- lantic Ocean and Alaskan C. opilio. A hierarchical log-likelihood analysis of these loci revealed very low levels of heterogeneity among all C. opilio (Table 5). In this analysis, PGM-1* was also highly significant (P=0.0071) in comparisons of Atlantic Ocean and Alaskan C. opilio. Among all C. opilio, PGM-1* was significant (P=0.0361). The overall log-likelihood sta- tistic for all loci among all C. opilio was significant (P=0.0382)(Table 5). Low-frequency alleles detected in Alaskan collections were AAT-1 *64, *210, and *120, AAT-2* 69, AH-2* 110 and *83,AH-3*106, CBYR*117, G3PDH-1*117, G3PDH-2*86 and *111, GPI-A1*60, 532 Fishery Bulletin 96(3), 1998 *158, and *55, IDHP-1*81 and *119, MDH-A2*65, PGM-1*82, *114, and *103, PROT-3*88, SOD -1*109, and TPI-1*-41 . Differing PGM-1* alleles and allele frequencies contributed most to the overall levels of differentiation observed between Alaskan and Atlan- tic Ocean C. opilio populations. Multilocus between- population chord distance measures ranged from 0.029 to 0.033 in Alaskan collections and increased to 0.043 when we added the Atlantic Ocean population. Comparison of C. bairdi and C. opilio from Alaska A total of 27 loci, AAT-1 *,AAT-2*,AH-3*, ADA-1 *, ADA- Table 4 Hierarchical log-likelihood analysis for Chionoecetes bairdi. df AAT-1 * df AAT-2* df AH -3* df ALAT* Total 16 13.10 16 12.70 16 14.72 8 2.21 Among 2 1.23 2 6.35* 2 1.90 1 1.60 Within 14 11.87 14 6.35 14 12.82 7 0.61 Bering Sea 2 2.69 2 0.62 2 2.54 1 0.61 Gulf of Alaska 12 9.18 12 5.73 12 10.28 6 0.00 Among 2 1.49 2 0.10 2 1.54 1 0.00 Within 10 7.69 10 5.63 10 8.74 5 0.00 Northern Gulf 8 7.69 8 4.83 8 5.38 4 0.00 Southeast 2 0.00 2 0.80 2 3.36 1 0.00 df CBYR* df G3PDH-1 * df G3PDH-2 * df GPI-A1* Total 8 2.21 8 8.25 16 43.07** 40 37.66 Among 1 1.61 1 0.31 2 13.82** 5 3.89 Within 7 0.60 7 7.94 14 29.25** 35 33.77 Bering Sea 1 0.60 1 1.01 2 16.84** 5 3.50 Gulf of Alaska 6 0.00 6 6.93 12 12.41 30 30.27 Among 1 0.00 1 1.16 2 0.49 5 6.25 Within 5 0.00 5 5.77 10 11.92 25 24.02 Northern Gulf 4 0.00 4 3.85 8 9.94 20 21.57 Southeast 1 0.00 1 1.92 2 1.98 5 2.45 df IDHP-1 * df MDH-A1* df PEP A* df PGDH* Total 24 38.63* 16 10.40 16 30.30* 24 15.60 Among 3 3.79 2 5.96 2 4.55 3 2.87 Within 21 34.84* 14 4.44 14 25.75* 21 12.73 Bering Sea 3 3.56 2 1.82 2 14.04 3 3.91 Gulf of Alaska 18 31.28 12 2.62 12 11.71 18 8.82 Among 3 5.51 2 0.78 2 2.62 3 0.77 Within 15 25.77 10 1.84 10 9.09 15 8.05 Northern Gulf 12 20.83 8 1.84 8 8.22 12 7.13 Southeast 3 4.94 2 0.00 2 0.87 3 0.92 df PGM-1* df PROT-3 * df TPI-1* df Overall Total 32 27.40 8 18.22* 8 6.29 256 280.75 Among 4 6.43 1 17.83** 1 0.58 32 72.71** Within 28 20.97 7 0.39 7 5.71 224 208.04 Bering Sea 4 3.87 1 0.39 1 1.21 32 57.22 Gulf of Alaska 24 17.10 6 0.00 6 4.50 192 150.82 Among 4 3.01 1 0.00 1 0.77 32 24.49 Within 20 14.09 5 0.00 5 3.73 160 126.33 Northern Gulf 16 13.38 4 0.00 4 3.73 128 108.40 Southeast 4 0.71 1 0.00 1 0.00 32 17.93 * Test is significant at a = 0.05. ** Test is significant at a = 0.01. Merkouris et al.: Genetic diversity in Chionoecetes bairdi and C. opilio 533 2*, ADA-3*, ALAT*, CBYR*, (5GALA*, GAPDH*, pGLUA*, G3PDH-1*, G3PDH-2*, GPI-A1*, IDHP-1*, MDH-A1 *, MDH-A2*, MEP-1*, MPI*, PEPA*, PEPD- 2*, PGDH*, PGM-1*, PROT-1*, PROT-2*, PROT-3* , and TPI-1*, were scored in all C. bairdi and C. opilio populations analyzed; 16 loci, AAT-1*, AAT-2*, AH - 3*, ALAT*, CBYR*, G3PDH-1*, G3PDH-2*, GPI-A1*, IDHP-1*, MDH-A1* , MDH-A2*, PEPA*, PGDH*, PGM-1*, PROT-3*, and TPI-1*, were polymorphic in at least one population of either species. Eleven loci, ADA-1*, ADA-2*, ADA-3*, pGALA*, GAPDH*, pGLUA*, MEP-1*, MPI*, PEPD-2*, PROT-1*, and PROT-2*, were monomorphic for the identical allele. Significant differences (P<0. 01) between the two spe- cies were detected at eight loci, AH-3*, G3PDH-1*, GPI-A1*, IDHP-1*, MDH-A1 *, PEPA*, PGDH*, and PROT-3*. Two loci, AH-3* and PROT-3*, were par- ticularly informative with nearly fixed differences for alternate alleles (Table 3). In addition, with the ex- ception of C. opilio from the Pribilof Island area col- lections, C. opilio expressed only the IDHP-1* 100 allele, whereas C. bairdi were highly variable. Sev- eral low-frequency variants (P<0.05) also contributed to the overall differences between the species. Numer- ous low-frequency alleles detected in only Bering Sea collections of C. bairdi were also detected in Alaskan C. opilio collections, including the following: AAT-2 *69, CBYR* 117, G3PDH-2*86, GPI-A1*55, MDH-A1*79, PGM-1*103, and PROT-3*88. Also, a rare allele, PGDH* 112, was detected only in Bristol Bay C. bairdi and in Pribilof Island and Atlantic Ocean C. opilio. Multilocus variation between species was estimated by genetic-distance measures. Between-species ge- netic-distance measures ranged from 0.222 to 0.251. The largest genetic distance was between Atlantic C. opilio and two northern Gulf of Alaska C. bairdi popu- lations, those of Kachemak Bay and Montague Strait. The smallest between-species genetic distance was that for Pribilof Island and Bering Sea C. opilio compared with that for Bering Sea and Pribilof Island C. bairdi. Table 5 Hierarchical log-likelihood analysis for Chionoecetes opilio. Source df Overall Total 96 121.93* Among 32 41.63 Within 64 80.30 Bering Sea 64 80.30 Atlantic Ocean 0 0.00 * Significant value (P< 0.05 ) Chiocoetes bairdi x C. opilio hybrids Hybrids between C. bairdi and C. opilio crabs in the Bering Sea have been identified morphologically and genetically (Karinen and Hoopes, 1971; Johnson, 1976; Grant et al., 1978; Hoopes et al.1). Several col- lections in our study had allele frequencies suggest- ing either low levels of introgression between the species or the inclusion of hybrid or backcross indi- viduals in the collection. For example, low frequen- cies of PROT-3*100 were observed in C. bairdi crabs from Bristol Bay, the Bering Sea, and the Pribilof Islands. PROT-3* 100 was not observed in C. bairdi from non-Bering Sea collections. Similarly, PROT- 3*88 was observed in C. opilio collections from the Bering Sea, St. Matthew Island, and the Pribilof Is- lands, and low frequencies of IDHP-1*81 and *119 were observed in Pribilof Island C. opilio collections; these alleles were not observed in C. opilio collected from the Atlantic Ocean. Discussion Allozyme electrophoresis techniques have been used extensively to describe evolutionary relationships within genera of decapod crustaceans (Bert, 1986; Bert and Harrison, 1988; Busack, 1989; Abdullah and Shukor, 1993); however, these techniques have gen- erally revealed very low levels of intraspecific genetic variation (Nelson and Hedgecock, 1980; Smith et al., 1980; Busack, 1988; Seeb et al., 1990b). Exceptions to this generalization are found in species that occur over broad geographic areas or in widely different environments (Nelson and Hedgecock, 1980; Mulley and Latter, 1981; Kartavtsev et al., 1991). Signifi- cant population heterogeneity has been found in spe- cies that exhibit highly specialized life history at- tributes (Stevens, 1991) and in some freshwater spe- cies (Macaranas et al., 1995; Fetzner et al., 1997). Within a marine species, Seeb et al. ( 1990a) discrimi- nated populations of red king crab from major geo- graphic areas of the Gulf of Alaska and Bering Sea. Allozyme techniques have also been used to examine seasonal variability of decapod larval, megalopal, and adult allelic frequencies (Kordos and Burton, 1993). Davidson et al. (1985) examined population struc- ture in North Atlantic C. opilio using allozymes and interpreted their observed esterase polymorphisms as phenotypic expressions of probable genotypic dif- ferences. However, we feel caution should be used in interpreting their data until genetic transmission of these markers can be shown by inheritance studies because we were unable to interpret the genetic ba- sis of observed esterase activity. 534 Fishery Bulletin 96(3), 1998 The importance of tissue quality for allozyme stud- ies is well documented (e.g. Utter et al., 1987) but is particularly critical for crustacean tissue, which de- grades more rapidly than finfish tissue. Laboratory analysis revealed obvious differences in enzyme reso- lution for several loci that resulted in conservative allele pooling and loss of some population data. Op- timal sample handling may have increased detect- able variation. Population diversity of C. bairdi We detected differences among C. bairdi populations of the three major geographic regions: Bering Sea, Gulf of Alaska, and Southeast Alaska. Most notably, we detected population subdivision within the Bering Sea. Samples collected in Bristol Bay (east of approxi- mately 162°45'W long.) were statistically different from samples collected near the Pribilof Islands (west of approximately 167°00'W long.). Other biological data support our inference that subpopulations of C. bairdi occur within the Bering Sea. Using trawl survey data from 1974 onward, Otto (1982) reported size-frequency distributions of C. bairdi in the Pribilof Islands that were different from those in the area north of the Alaska Peninsula, in- cluding Bristol Bay. Further, as National Marine Fisheries Service (NMFS) trawl assessment surveys expanded westward, it became evident that crabs in the Pribilof area were larger in size compared with those along the continental slope north and west of the Pribilof Islands. Somerton (1981) found size dif- ferences among large females from the eastern Bering Sea seemed to partition into two subareas along 167°15'W long. The mean size of adult females was quite constant in the eastern subarea (east of 167°15'W long.) but decreased steadily in the west- ern subarea. Although both investigators speculated upon the possibility of genetic or environmental fac- tors causing these differences, the species has been considered a single Bering Sea stock for management purposes (Otto, 1982). In our study, we lacked C. bairdi specimens from west of approximately 173°W long, and thus were unable to test whether crabs collected from near the Pribilof Islands differed from crabs collected near the continental slope. Although there does not appear to be a precise correlation be- tween our findings and the observations of Otto (1982) and Somerton (1981), all data seem to sug- gest that C. bairdi in the Bering Sea may not be com- posed of a single panmictic population. However, any conclusions we may draw from bio- chemical data should be confirmed by other types of data. The American lobster ( Homarus americanus ) is a case in point. Tracey et al. (1975) noted genetic and morphological differences between inshore and offshore populations. In tagging studies, Fogarty et al. ( 1980) demonstrated limited movement of lobsters released at inshore stations, in contrast with exten- sive seasonal migrations by lobsters tagged at off- shore locations. These studies indicated that despite the potential for genetic exchange during seasonal mixing, the inshore and offshore populations retained their genetic identity. The planktotrophic larvae of Chionoecetes may spend up to two months in surface waters (Slizkin, 1990), raising the possibility that larvae released in one area may become recruits in another. Further, most of what is known or hypothesized about migration of Bering Sea C. bairdi is based upon abundance and dis- tribution estimates from annual NMFS trawl surveys and subsequent fishery captures.2 Migration studies of Chionoecetes in Alaska have been hindered by tag losses after molting. Development of a permanent tag for Chionoecetes would provide a valuable tool for ex- amining population migration and further for evaluat- ing the null hypothesis of panmixia. Our findings of regional and within-Bering Sea heterogeneity among C. bairdi populations merit further investigation. Population diversity of C. opilio We detected only small differences between Bering Sea and North Atlantic C. opilio, primarily in PGM-1* al- lele frequencies. The low-frequency alleles of other loci observed in Alaskan C. opilio did not contribute sig- nificantly to overall differences. Although sampling error may have been a factor (Gregorius, 1980) in de- tection of rare alleles, no private alleles were found in the North Atlantic. The minimal genetic differentia- tion among this circumpolar species contrasts with find- ings in halibut (Hippoglossus), herring ( Clupea ) and cod (Gadus) (Grant, 1987) where significant differen- tiation has been described across a similar geographic range. The close genetic affinity of the Bering Sea C. opilio collections with the collection from the North Atlantic and the detection of low-frequency Bering Sea Chionoecetes alleles in the North Atlantic suggest re- cent or ongoing gene flow. This possibility is supported by Garth ( 1958) whose range description includes a dis- tribution of C. opilio northward (from the Bering Strait) and eastward through the Beaufort Sea and Davis Strait to the North Atlantic. This species tolerates very low temperatures; the optimum temperature for im- mature individuals has negative values even in the summer (Slizkin, 1990). Thus, it is plausible to hypoth- 2 Morrison, R. 1997. Alaska Dep. Fish and Game, RO. Box 920587, Dutch Harbor, AK 99692. Personal commun. Merkouris et at: Genetic diversity in Chionoecetes bairdi and C opilio 535 esize gene flow through these northern seas even in the presence of sea ice pack. Alternatively, it is pos- sible that gene flow is restricted, but that the differen- tiating forces of drift, selection, and mutation have not yet produced significant detectable genetic divergence. Species differentiation, hybrids, and gene introgressiom The morphological criteria for differentiating C. bairdi and C. opilio have been well described (Rathbun, 1925; Garth, 1958; Karinen and Hoopes, 1971). However, a few crabs in our study that met one or more morphological criteria were probably hybrid or recombinant individuals. Their composite genotypes were atypical of the parental species and were instead indicative of Fj hybrids or backcross matings. Species differentiation in allelic frequen- cies was observed between C. bairdi and C. opilio , most notably for AH-3*, IDHP-1* , and PROT-3 *. The unusual composite genotypes were most clearly dif- ferentiated by the PROT-3* 100 allele in C. bairdi collections, and conversely, by the PROT-3*88 allele in the C. opilio collections from the Bering Sea. The PROT-3* 100 allele contributed significantly to the among-regional differences in C. bairdi. This marker is also very useful for investigation of hybridization between these species. Gene flow and geographic barriers Our findings are consistent with previous electro- phoretic studies that infer that decapod crustaceans, especially large, mobile species, exhibit low levels of variation at allozyme markers (Nelson and Hedge- cock, 1980). The high dispersal potential of marine species is often used to explain the low levels of varia- tion detected among populations (Avise, 1994). Such distribution appears to be the case with Chionoecetes as well as with other marine organisms with similar geographic distributions in the North Pacific, Gulf of Alaska, and Bering Sea, such as Pacific halibut (Grant et al., 1984), Pacific herring (Grant and Ut- ter, 1984), Pacific cod (Gadus macrocephalus) (Grant et al., 1987), Pacific ocean perch ( Sebastes alutus) (Seeb and Gunderson, 1988), and red king crab (Paralithodes camtschaticus ) (Seeb et al., 1990a). Limits to the actual dispersal of marine species, despite high dispersal potential, may periodically or continuously limit gene flow in some directions (Avise, 1994 and references therein). The Alaska Peninsula-Aleutian chain appears to be a significant barrier to gene flow in some species, such as Pacific ocean perch (Seeb and Gunderson, 1988), Pacific herring (Grant and Utter, 1984), rock sole (Pleuro- nectes bilineatus) (Mulligan et al., 1995) and red king crab (Seeb et al., 1990a). The Alaska Peninsula-Aleutian Chain appears to limit gene flow among C. bairdi populations as evi- denced by the differentiation detected among regions above and below this geographic barrier. The lack of introgressed C. opilio alleles in the Gulf of Alaska and Southeast Alaska populations studied also sup- ports this view. Our findings suggest that the nu- merous rare alleles observed in Bering Sea, and not Atlantic, populations of Chionoecetes may be ancient alleles that have been maintained over time in the large populations that have inhabited these waters; however, larger sample sizes of Atlantic Ocean C. opilio are needed to confirm this hypothesis. Detection of low- frequency Bering Sea Chionoecetes alleles in North Atlantic C. opilio and the lower heterozygosity of North Atlantic C. opilio, coupled with the lack of congeners in the North Atlantic, allow speculation that the genus Chionoecetes may have arisen in the North Pacific. Acknowledgments The authors thank the following Alaska Department of Fish and Game personnel for population sample collections: Dean Beers, Cathy Bothelo, Bill Donaldson, Wayne Donaldson, Lee Hammarstrom, Ken Imamura, Dave Jackson, Al Kimker, Tim Koeneman, Al Spalinger, Dan Urban, and especially Donn Tracy, Ken Griffin, and Ranee Morrison. We thank the National Marine Fisheries Service person- nel for collections: Pete Cummiskey, Jan Haaga, John Karinen, Rich Macintosh, Eric Munk, and Brad Stevens, and we thank Bob Otto for his efforts in coordinating the Atlantic C. opilio collection. We also thank Dave Armstrong of the University of Wash- ington, and M. John Trembley of the Canadian De- partment of Fisheries and Oceans for their assistance. Penny Crane and three anonymous reviewers provided critical reviews. This research was funded in part by a cooperative agreement with the National Oceanic and Atmospheric Administration (award NA37FL0333). Literature cited Abdullah, R., and N. A. A. Shukor. 1993. Isozyme variation between two closely related spe- cies Crangon crangon (L.) and Crangon allmanni kinahan (Decapoda, Caridea). Crustaceana 64:114—120. Aebersold, P. B., G. A. Winans, D. J. Teel, G. B. Milner, and F. M. Utter. 1987. Manual for starch gel electrophoresis: a method for the detection of genetic variation. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 61, 19 p. 536 Fishery Bulletin 96(3), 1 998 ADF&G (Alaska Department of Fish and Game). 1994. Annual management report for the shellfish fisher- ies of the westward region, 1993. Division of Commercial Fisheries Management and Development, Kodiak, AK. Reg. Info. Rep. 4K94-29, 262 p. Allendorf, F. W., N. Ryman, and F. M. Utter. 1987. Genetics and fishery management: past, present, and future. In N. Ryman and F. Utter (eds.), Population ge- netics and fishery management, p. 1-19. Washington Sea Grant Program, Univ. Washington, Seattle, WA. Avise, J. C. 1994. Molecular markers, natural history and evolution. Chapman & Hall, Inc., New York, NY, 511 p. Bert, T. M. 1986. Speciation in western Atlantic stone crabs (genus Menippe 1: the role of geological processes and climatic events in the formation and dist ribution of species. Mar. Biol. 93:157-170. Bert, T., and R. G. Harrison. 1988. Hybridization in western Atlantic stone crabs (genus Menippe 1: evolutionary history and ecological context in- fluence species interactions. Evolution 42:528-544. Boyer, S. H., D. C. Fainer, and M. A. Naughton. 1963. Myoglobin: inherited structural variation in man. Science (Wash., D.C.) 140:1228-1231. Busack, C. A. 1988. Electrophoretic variation in the Red Swamp ( Procam - barus elarkii) and White River crayfish ( P . acutus ) (Decapoda: Cambaridae). Aquaculture 69:211-226. 1989. Biochemical systematics of crayfishes of the genus Procambarus, subgenus Seapulicambrus (Decapoda: Cambaridae). J. N. Am. Benthol. Soc. 8(2):180-186. Cavalli-Sforza, L. L., and A. W. F. Edwards. 1967. Phylogenetic analysis: models and estimation procedures. Evolution 21:550-570. Clayton, J. W., and D. N. Tretiak. 1972. Amine-citrate buffers for pH control in starch gel electrophoresis. J. Fish. Res. Board Can. 29:1169-1172. Cobb, J. S., and J. F. Caddy. 1989. The population biology of decapods. In J. F. Caddy (ed.), Marine invertebrate fisheries: their assessment and manage- ment, p. 327-374. John Wiley & Sons. New York, NY. Davidson, K., J. C. Roff, and R. W. Elner. 1985. Morphological, electrophoretic, and fecundity char- acteristics of Atlantic snow crab, Chionoecetes opilio , and implications for fisheries management. Can. J. Fish. Aquat. Sci. 42:474-482. Fetzner, J. W., Jr., R. J. Sheehan, and L. W. Seeb. 1997. Genetic implications of broodstock selection for cray- fish aquaculture in the midwestern United States. Aqua- culture 154(19971:39-55. Fogarty, M. J., D. V. D. Borden, and H. J. Russell. 1980. Movements of tagged American lobster, Homarus americanus, off Rhode Island. Fish. Bull. 78(31:771-780. Garth, J. S. 1958. Brachyura of the Pacific coast of America Oxyrhyn- cha. Univ. S. Cal. Press, Los Angeles, CA, 499 p. Goudet, J. 1995. FSTAT (version 1.2): a computer program to calcu- late F-statistics. J. Heredity 86(61:485-486. Grant, W. S., L. Bartlett, and F. M. Utter. 1978. Biochemical genetic identification of species and hy- brids of the Bering Sea Tanner crab, Chionoecetes bairdi and C. opilio. Proc. Natl. Shellfish Assoc. 68:127. Grant, W. S. 1987. Genetic divergence between congeneric Atlantic and Pacific Ocean fishes. In N. Ryman and F. Utter (eds.), Population genetics and fishery management, p. 225- 246. Washington Sea Grant Program, Univ. Washington, Seattle, WA. Grant, W. S., D. J. Teel, T. Kobayashi, and C. Schmitt. 1984. Biochemical population genetics of Pacific halibut ( Hippoglossus stenolepis 1 and comparison with Atlantic halibut ( H . hippoglossus). Can. J. Fish. Aquat. Sci. 41:1083-1088. Grant, W. S. and F. M. Utter. 1984. Biochemical population genetics of Pacific herring ( Clupea pallasi). Can. J. Fish. Aquat. Sci. 41:856-864. Grant, W. S., C. I. Zhang, T. Kobayashi, and G. Stahl. 1987. Lack of genetic stock discretion in Pacific cod ( Ga - dus macrocephalus 1. Can. J. Fish. Aquat. Sci. 44:490-498. Gregorius, H. R. 1980. The probability of losing an allele when diploid geno- types are sampled. Biometrics 36:643-652. Harris, H., and D. A. Hopkinson. 1976. Handbook of enzyme electrophoresis in human genetics. American Elsevier, New York, NY, 361 p. Johnson, A. G. 1976. Electrophoretic evidence of hybrid snow crab, Chionoecetes bairdi x opilio. Fish. Bull. 74:693—694. Karinen, J. F. 1974. King and Tanner crab research. 1971. INPFCAnnu. Rep 1972:102-111. Karinen, J. F., and D. T. Hoopes. 1971. Occurrence of Tanner crabs (Chionoecetes sp.) in the eastern Bering Sea with characteristics intermediate be- tween C. bairdi and C. opilio. Proc. Natl. Shellfish Assoc. 61:8-9. Kartavtsev, Y. P., B. I. Berenboim, and K. I. Zgurovsky. 1991. Population genetic differentiation of the pink shrimp, Pandalus borealis Kroyer, 1838, from the Barents and Bering Seas. J. Shellfish. Res. 10:333-339. Kordos, L. M., and R. S. Burton. 1993. Genetic differentiation of Texas Gulf Coast popula- tions of the blue crab Callinectes sapidus. Mar. Biol. 117:227-233. Kruse, G. H. 1993. Biological perspectives on crab management in Alaska. In Management strategies for exploited fish popu- lations, p. 355-384. Alaska Sea Grant Report 93-02, Univ. Alaska, Fairbanks, AK. Macaranas, J. M., P. B. Mather, P. Hoeben, and M. F. Capra. 1995. Assessment of genetic variation in wild populations of the red claw crayfish ( Cherax quadricarinatus, von Mar- tens 1868) by means of allozyme and RAPD-PCR markers. Mar. Freshwater Res. 46:1217-1228. MathSoft, Inc. 1997. S-PLUS programmer’s guide. Data Analysis Prod- ucts Div., MathSoft, Seattle, WA. Mulley, J. C., and B. D. H. Latter. 1981. Geographic differentiation of eastern Australian prawn populations. Aust. J. Mar. Freshwater Res. 32(61:889-906. Mulligan, H., A. W. Kendall, and A. C. Matarese. 1995. The significance of morphological variation in adults and larvae of the rock sole ( Pleuronectes bilineatus ) from the Bering Sea and Northeastern Pacific Ocean. In Pro- ceedings of the international symposium on North Pacific flatfish, p. 133-151. Alaska Sea Grant Report 95-04, Univ. Alaska, Fairbanks, AK. Nelson, K., and D. Hedgecock. 1980. Enzyme polymorphism and adaptive strategy in the decapod Crustacea. Am. Nat. 116:238-280. Merkouris et al.: Genetic diversity in Chionoecetes bairdi and C. opilio 537 Otto, R. S. 1982. An overview of eastern Bering Sea Tanner crab fisheries. In Proceedings of the international symposium on the genus Chionoecetes, p. 83-115. Alaska Sea Grant Report 82-10, Univ. Alaska, Fairbanks, AK. 1990. An overview of eastern Bering Sea King and Tanner crab fisheries. In Proceedings of the international sym- posium on king and Tanner Crabs, p. 9-26. Alaska Sea Grant Report 90-04, Univ. Alaska, Fairbanks, AK. Rathbun, M. J. 1925 The spider crabs of America. Smithsonian Institu- tion, United States National Museum, Government Print- ing Office, Washington, Bulletin 129, 613 p. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43: 223-225. Ridgway, G. J., S. W. Sherburne, and R. D. Lewis. 1970. Polymorphism in the esterases of Atlantic herring. Trans. Am. Fish. Soc. 99:147-151. Seeb, L. W., and D. R. Gunderson. 1 988. Genetic variation and population structure of Pacific ocean perch ( Sebastes alutus). Can. J. Fish. Aquat. Sci. 45:78-88. Seeb, J. E., G. H. Kruse, L. W. Seeb, and R. G. Week. 1990a. Genetic structure of red king crab populations in Alaska facilitates enforcement of fishing regulations. In Proceedings of the international symposium on king and Tanner Crabs, p. 491-502. Alaska Sea Grant Report 90- 04, Univ. Alaska, Fairbanks, AK. Seeb, L. W., J. E. Seeb, and J. J. Polovina. 1990b. Genetic variation in highly exploited spiny lobster Panulirus marginatus populations from the Hawaiian Archipelago. Fish. Bull. 88:713-718. Selander, R. K., M. H. Smith, S. Y. Yang, W. E. Johnson, and J. B. Gentry. 1971. Biochemical polymorphism and systematics in the genus Peromyscus. I. Variation in the old field mouse ( Peromyscus polionotus). Studies in Genetics VI. Univ. Texas Publication 7103:49-90. Shaklee, J. B., F. W. Allendorf, D. C. Morizot, and G. S. Whitt. 1990. Gene nomenclature for protein-coding loci in fish. Trans. Am. Fish. Soc. 119:2-15. Shaw, C. R., and R. Prasad. 1970. Starch gel electrophoresis of enzymes: a compilation of recipes. Biochem. Genet. 4:297-320. Slizkin, A. G. 1990. Tanner crabs (Chionoecetes opilio, C. bairdi) of the Northwest Pacific: distribution, biological peculiarities, and population structure. In Proceedings of the international symposium on king and Tanner crabs, p. 27-33. Alaska Sea Grant Report 90-04. Univ. Alaska, Fairbanks, AK. Smith, P. J., J. L. McKoy, and P. J. Machin. 1980. Genetic variation in the rock lobsters Jasus edwardsii and Jasus novaehollandiae. N.Z. J. Mar. Freshwater Res. 14:55-63. Sokal, R. R., and F. J. Rohlf. 1995. Biometry, 3rd ed. Freeman and Co., New York, NY, 887 p. Somerton, D. 1981. Regional variation in the size of maturity of two spe- cies of Tanner crab (Chionoecetes bairdi and C. opilio ) in the eastern Bering Sea, and its use in defining manage- ment subareas. Can. J. Fish. Aquat. Sci. 38:163-174. Stevens, B. G., W. E. Donaldson, J. A. Haaga, and J. E. Munk. 1993. Morphometry and maturity of paired Tanner crabs, Chionoecetes bairdi, from shallow and deep water environments. Can. J. Fish. Aquat. Sci. 50:1504-1516. Stevens, P. M. 1991. A genetic analysis of the pea crabs (Decapoda: Pinnotheridae) of New Zealand. II. Patterns and intensity of spatial population structure in Pinnotheres atrinicola. Mar. Biol. 108:403-410. Thorpe, J. E. 1993. Impacts of fishing on genetic structure of salmonid populations. In J. G. Cloud and G. H. Thorgaard (eds.), Genetic conservation of salmonid fishes, p. 67-80. Ple- num Press, New York, NY. Tracey, M. L., K. Nelson, D. Hedgecock, R. A. Shleser, and M. L. Pressick. 1975. Biochemical genetics of lobsters: genetic variation and the structure of American lobster (Homarus americanus) populations. J. Fish. Res. Board Can. 32(11 ):209 1— 2 10 1 . Utter, F. M., P. Aebersold, and G. Wimans. 1987. Interpreting genetic variation detected by electro- phoresis. In N. Ryman and F. Utter (eds.), Population genetics and fishery management, p. 21-45. Washington Sea Grant Program, Univ. Washington, Seattle, WA. Watson, J. 1970. Maturity, mating, and egg laying in the spider crab, Chionoecetes opilio. J. Fish. Res. Board Can. 27:1607-1616. Weir, B. S. 1990. Genetic data analysis. Sinaur Associates, Inc., Sun- derland, MA, 377 p. Wright, S. 1978. Evolution and the genetics of populations. Vol. IV: Variation within and among natural populations. Univ. Chicago Press, Chicago, IL, 580 p. Abstract .-Some combinations of trawler, trawl, and crew catch fish bet- ter than others. Systematic error en- ters catch per unit of effort (CPUE) data from trawl surveys through such rela- tive differences in efficiency among sampling instruments. Correcting rela- tive fishing power differences can re- move bias due to systematic error but may also increase the variance of the mean CPUE estimate. As a result, the overall error of the estimate may actu- ally become worse, even when the fish- ing power difference is statistically sig- nificant. A decision rule specified by the mean square error (MSE) of the mean CPUE estimate avoids this mistake: namely a correction that reduces the error in the mean CPUE estimate would be applied, a correction that in- creases the error would not be applied. I describe and demonstrate an algo- rithm, based on minimizing the MSE, for deciding to correct a fishing power difference. The strategy requires that a probability density function exist that models the CPUE data reasonably well. Manuscript accepted 3 March 1998. Fishery Bulletin 96:538-546 (1998). A decision rule based on the mean square error for correcting relative fishing power differences in trawl survey data Peter T. IWHunro Resource Assessment and Conservation Engineering Division Alaska Fisheries Science Center National Oceanic and Atmospheric Administration, NOAA 7600 Sand Point Way NE, BIN C- 1 5700 Seattle, Washington 981 15-0070 E-mail address: peter.munro@noaa.gov Trawl surveys are often used to col- lect data on catch observations stan- dardized by fishing effort (the catch per unit of effort [CPUE]), and the mean CPUE is often interpreted as an index of abundance. Ideally, fish- ing power, or fish catching effi- ciency, must be held constant in trawl surveys lest altered catch rates be confounded with changes in abundance of fish or inverte- brates. Unfortunately, the sampling instrument is a complex system that includes the vessel, vessel op- erators, and fishing gear, all of which may vary from survey to sur- vey, introducing changes in fishing power (Gulland, 1956). Correcting fishing power differences seems necessary for proper interpretation of mean CPUE. However, methods have not been established for deter- mining if an improved estimate of mean CPUE actually results from such correction. Relative differences in fishing power make standardization in trawl surveys difficult. Technologi- cal changes in fishing gear, as well as replacement of older research vessels, may affect fishing power (Azarowitz, 1981; Byrne et ah, 1991). Fishing power differences are an inherent part of multiple vessel surveys. Examples of this type of survey are annual and triennial surveys in several regions of the north Pacific Ocean and Bering Sea conducted by the Alaska Fisheries Science Center (AFSC) of the Na- tional Marine Fisheries Service (NMFS) (Harrison, 1992; Weinberg et al., 1994; Munro and Hoff, 1995; Goddard and Zimmermann1) and the International Young Fish Sur- vey in the North Sea (Anonymous2). Koeller and Smith (1983) reported change in a single vessel’s ability to measure speed over a 3-year period and hypothesized that this may have altered its fishing power from year to year. Operator effects have also been shown to account for fish- ing power differences among vessels (Munro and Hoff, 1995) and thus may be inferred for change in op- 1 Goddard, P., and M. Zimmermann. 1993. Distribution, abundance, and biological characteristics of groundfish in the east- ern Bering Sea based on results of the U.S. bottom trawl survey during June-Septem- ber 1991. U.S. Dep. Commer., NOAA, Natl. Mar. Fish. Serv., Alaska Fish. Sci. Cent., 7600 Sand Point Way NE, Seattle, WA 98115-0070, Proc. Rep. 93-15, 324 p. 2 Anonymous. 1986. Manual for the In- ternational Young Fish Surveys in the North Sea, Skagerrak, and Kattegat. In- ternational Council for the Exploration of the Sea, Palaegade 2-4, DK-1261, Copen- hagen K, Denmark. ICES Council Meet- ing 1986/H:2. Munro: Correcting relative fishing power differences in trawl survey data 539 follows, “vessel” refers to the entire system that com- prises the sampling instrument, from the bow of the fishing vessel to the codend.) A fishing power difference forces a choice between two estimators of mean CPUE, one that incorporates a fishing power correction and one that does not. Correcting a fishing power difference amounts to changing an observation to an estimate. The follow- ing model is used to estimate what a standard ves- sel would have caught had it executed exactly the same tow as was done by a nonstandard vessel: xt = yt FPC, where x, = estimated CPUE at station i by the stan- dard vessel; y . = observed CPUE at station i by the non- standard vessel; and FPC = estimated fishing power correction factor. Each original observation, yL, has no error beyond measurement error. Every x, has error due to the variance of the estimate of FPC. Consequently the mean CPUE estimated from the xt has at least two components of variation, one stemming from the usual sampling variance in the observations (y(), the other due to uncertainty in the estimate of FPC. This added component of estimation error has been rec- ognized as the cost of correcting systematic error in the observations, but only in passing, (Sissenwine and Bowman, 1978; Byrne et al., 1981; Koeller and Smith, 1983; Fanning, 1984). Estimators are often chosen on the basis of their relative error, yet most researchers have ignored this in deciding whether or not to correct a fishing power difference. Investigations have been focused on esti- mating the difference but have not evaluated it in terms of estimating mean CPUE. Very few decision rules have been explicitly stated, most having been implied by testing the statistical significance of the fishing power difference itself. Early CPUE calibra- tions (Gulland, 1956; Robson, 1966) were based on multiplicative models of different sources of variabil- ity in CPUE data, including vessel effects. Log trans- formation of the data produced linear models with coefficients that could be estimated with regressions and for which classical hypotheses could be formu- lated and tested. Sissenwine and Bowman (1978), Kimura (1981), and Gavaris (1980) have followed this strategy in their decisions to apply FPCs. Gavaris (1980) estimated the FPC using the method of Bradu and Mundlak (1970) and reported approximate con- fidence intervals, without explicitly stating a deci- sion rule. Fanning (1984) proposed an explicit deci- sion rule using a beta-distributed index for the fish- ing power difference in paired observations. If the confidence interval included the value that repre- sented identical fishing power, he recommended that the estimated FPC not be applied. Byrne and Fogarty ( 1985) tested the significance of fishing power differ- ences using Hotelling’s Usquared test when several species were considered simultaneously, or the non- parametric Friedman’s test for a single species. They offered no interpretation of a significant fishing power difference, in particular, whether or not it should be corrected. The Utest was used by Byrne et al. (1991) to determine the significance of a fishing power difference. In response to a significant differ- ence, they estimated an FPC using the method of Bradu and Mundlak ( 1970). They then produced con- fidence intervals for that estimate using a bootstrap approximation. However, they did not state an ex- plicit decision rule based on those intervals. Correcting a fishing power difference would be worthwhile only when it reduces the error in the es- timate of mean CPUE. Statistical significance of a fishing power difference is not a compelling justifi- cation because the cost of the added uncertainty may out weigh the benefit of removing bias that entered through systematic error in CPUE data. If the esti- mate of a correction factor has a lot of uncertainty, then the error of the estimate of mean CPUE could actually become worse by correcting data, even for a statistically significant fishing power difference. A decision rule for correcting a fishing power differ- ence must avoid this mistake by accounting for the cost of correcting as well as the benefit. Such a rule would permit choosing the estimate of mean CPUE that yields the lower total error. Methods The notion of the mean square error (MSE) lends a useful structure for defining such a decision rule. The MSE is a widely recognized measure of error between an estimator and its parameter (Mood et al., 1974). The MSE is defined as MSE [C] = E ( c-cr which is the expectation of the squared difference between the estimator of mean CPUE, C, and the parameter being estimated, true CPUE or, C. The MSE can be rewritten as MSE [C] = Var [C] + b2 [C] , or the sum of the variance and the squared bias of the estimator. By defining the following estimators, 540 Fishery Bulletin 96(3), 1 998 C(w) - estimator as a function of uncorrected data C(wFpC) = estimator as a function of data cor- rected with an estimated FPC, where w = (xv x2, . . . , , yv y2, . . . yn) = a mixed vector of CPUE observations from two vessels and wFpc = {xvx2, . . . , xn ,y1FPC,yt)FPC, . . . ynJPO - a mixed vector of CPUE observations and estimates, a model for the decision rule can be stated as: Apply the FPC if MSE C(w) > MSE C{w — FPC ' ference for which correcting reduces the MSE of the estimated mean CPUE can then be determined. This general strategy is illustrated by construct- ing an algorithm to apply it to a real problem. Any algorithm for implementing this decision rule will depend on specific circumstances. In this case the particulars are defined by a fishing power problem in an annual survey of the eastern Bering Sea, con- ducted by the AFSC (Wakabayashi et al., 1985; Goddard and Zimmermann1). Two vessels system- atically sample all strata, following interleaved sta- tion patterns that produce approximately equal num- bers of CPUE observations. From these two sets of unpaired data a fishing power difference between two vessels is estimated for each of a number of species. The question is “Should this estimated FPC be used to correct CPUEs of one vessel to the fishing effi- ciency of the other?” This general MSE decision rule takes the following form (the specifics of the Bering Sea survey being addressed within this framework): and do not apply the FPC if MSE\C{w) < MSE C (iv ppp ) (Note, this explication of the MSE of mean CPUE has been framed in terms of a single survey with two vessels. But the notion of MSE lends itself equally well to any situation in which data must be “corrected,” including the common case of a single, nonstandard vessel conducting a standard survey.) This decision rule is unattainable because it re- quires that the true value of the fishing power dif- ference and CPUE sampling distributions be known. However, simulations can be used to estimate the mean square error. One such simulation strategy takes the following form: Assume a probability dis- tribution for CPUE. Generate realizations of this dis- tribution to represent a survey in which a fishing power difference is suspected. Impose an assumed fishing power difference on the simulated survey. Estimate the fishing power difference and apply the estimate to correct the assumed fishing power dif- ference. ( Estimating the fishing power difference may require further distributional assumptions and simu- lations, depending on the estimator and the kind of data it requires, especially if fishing power differ- ence is to be estimated from experiments conducted independently of the survey.) The MSEs of the esti- mated mean CPUE are then calculated from the re- alizations with and with out the fishing power cor- rections. This procedure is repeated for a range of selected fishing power differences. Whether the ob- served fishing power difference (estimated from real data) falls within the range of the fishing power dif- 1 Simulate surveys from an appropriate sampling distribution for data collected by a “standard” vessel. 2 Impose a known fishing power difference on the CPUE data in each simulated survey. (In these examples half the data were altered to emulate a two-vessel survey.) 3 For each simulated survey, estimate an FPC to correct the fishing power difference that was im- posed in the previous step. (FPCs may be esti- mated from simulation from independent experi- mental data or, as in these examples, estimated from the simulated survey itself. The important aspect is that the error structure of the FPC esti- mator be incorporated in the simulation process.) 4 Estimate the mean CPUE for each simulated sur- vey with and without correcting for the fishing power difference. 5 Repeat steps 2 through 4 for a range of fishing power differences. 6 Compute MSEs for estimated mean CPUE for each level of fishing power difference. 7 Plot the estimated MSEs against the fishing power differences imposed in Step 2 (Fig. 1). 8 Determine the range of fishing power difference where the MSE for corrected data is lower than the MSE for uncorrected data (Fig. 1). The region of increased error is centered around the value 1.0, which represents equal fishing powers, and is sandwiched between regions of reduced error (Fig. 1). The smaller the true fishing power difference the more likely that correcting it will lead to increased error in mean CPUE, and the greater the fishing Munro: Correcting relative fishing power differences in trawl survey data 541 power difference the more likely that correct- ing it will improve error in mean CPUE. For the range of the relative fishing power differ- ence for which correction increases (becomes worse), the MSE will be called the “noncor- rection region.” The two ranges of the relative fishing power difference for which correcting reduces (improves) the MSE will be referred to collectively as the “correction region.” This procedure has four critical elements: simulating the CPUE data, the estimator of FPC, the estimator of mean CPUE, and the sample size in each simulation. The CPUE data (kilograms per hectare) were simulated with the A-distribution. The A-distribution has been pro- posed as an appropriate distribution for data that include the value 0.0 and that are heavily skewed to the right (Pennington, 1983; McCon- naughey and Conquest, 1992). The probability density function for the A-distribution has pa- rameters p, the probability of an observation with the value 0.0, and p and a, the conven- tional defining parameters of the lognormal dis- tribution, which are the population mean and standard deviation of the log-transformed ele- ments, respectively (Aitchison and Brown, 1957). The parameters for this distribution were calculated with CPUE data for flathead sole ( Hippoglossoides elassodon ) and walleye pollock ( Theragra chalcogramma ) collected in the 1992 eastern Bering Sea survey (Table 1; Fig. 2). These two species were chosen to illustrate cases of moderate and extreme skewness to the right. The mechanism for imposing a fishing power difference is also part of simulating each survey. In these examples the relative differ- ence was applied by multiplying each CPUE from the nonstandard vessel by a ratio that rep- resented the true mean nonstandard CPUE over the true mean standard CPUE. Two-hun- dred surveys were simulated at each of twenty preselected fishing power differences (Table 2). In each simulated survey, half of the data were selected to represent the standard vessel and the fishing power difference was imposed on the other half of the data, representing the non- standard vessel. The sample sizes in two of the simulations were based on the number of ob- servations used to calculate the parameters of the A-distribution: 149 per vessel for pollock and 144 per vessel for flathead sole. The third simulation was based on 50 observations per vessel for flathead sole. The smaller sample size is similar to the number of tows made in the larger strata of the annual Bering Sea survey (Goddard and Zimmermann1). -1 LU Z> Q_ O c a 15,000- 10,000- 5,000 - uncorrected data corrected data Variance 0.5 1.0 1.5 2.0 B 20,000-i "O 3 million metric tons) during 1982-84 were attributed in part to increased availability of yel- lowfin sole within the standard survey area. Manuscript accepted 10 December 1997. Fishery Bulletin 96:547-561 (1998). Annual and between-sex variability of yellowfin sole, Pleuronectes asper, spring-summer distributions in the eastern Bering Sea Daniel G. Nichoi Resource Assessment and Conservation Engineering Division Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 7600 Sand Point Way NE, BIN C I 5700 Seattle, Washington 981 I 5-0070 Yellowfin sol e, Pleuronectes asper, of North America has a Pacific coast distribution from British Columbia (49°N) north to the Chukchi Sea (70°N) and a distribution along the Asian coast that ranges from the Sea of Japan (35°N) north to the Gulf of Anadyr (Fadeev, 1970; Hart, 1973; Bakkala, 1981; Wilderbuer et al., 1992). The eastern Bering Sea shelf has supported the largest con- centration of yellowfin sole (Bak- kala, 1993), with survey biomass estimates exceeding 2 million met- ric tons (t) since the early 1980s (Wilderbuer et al., 1992). Spring-summer (June-August) distributions of yellowfin sole in the eastern Bering Sea are dependent upon the timing of their annual cross- shelf migration from overwintering grounds near the shelf-slope break (about 200 m bottom depth) to near- shore waters (<50 m bottom depth) of Bristol Bay north to Nunivak Is- land (Bakkala, 1981; Wakabayashi, 1989). This migration is thought to follow the ice edge as it recedes dur- ing spring (Bakkala, 1981). Many ju- venile yellowfin sole are likely nonmi- gratory and remain nearshore under ice-cover during winter (Fadeev, 1970). Juvenile and adult distribu- tions, therefore, overlap during spring-summer when adults enter nearshore waters to spawn (Nichoi, 1995, 1997). Past yellowfin sole abundance in the eastern Bering Sea has varied with the magnitude of fishery exploi- tation. A predominantly foreign-led commercial trawl effort (Japan and U.S.S.R.) accounted for annual land- ings in excess of 400,000 t from 1959 to 1962. This trawl effort led to a de- cline of the stock and to significantly lower annual catches (<100,000 t) through the 1970s (Wilderbuer et al., 1992). Lower exploitation rates dur- ing the late 1970s coincided with a major increase in stock abundance that appeared to peak in the early 1980s. Considering the high abun- dance, annual yellowfin sole catches have been moderate since 1982, av- eraging 147,433 1 (Table 1). Commer- cial trawl catches of yellowfin in re- cent years have been limited by fish- ery management time and area clo- sures owing to by catch of species such as Pacific halibut ( Hippoglossus stenolepis). Pacific herring ( Clupea pallasi ), Tanner crab ( Chionoecetes bairdi), and red king crab ( Para - lithodes camtschaticus ) (Witherell, 1995). Despite moderate exploitation, annual survey biomass estimates have fluctuated widely since the early 1980s (Wilderbuer1). These fluctua- 1 Wilderbuer, T. K. 1996. Yellowfin sole. Chapter 3 in Plan team for groundfish fish- eries of the Bering Sea/Aleutian Islands (ed.), Stock assessment and fishery evaluation continued on next page 548 Fishery Bulletin 96(3), 1 998 Table 1 Standard AFSC survey yellowfin sole ( Pleuronectes asper) biomass and population estimates, percentage of male and female CPUE (kg/hectare) in nearshore ( <30 m bottom depth) waters, sex-proportions, and mean CPUE-weighted bottom depths by sex, for years 1982-96. Commercial catch estimates for yellowfin sole in the eastern Bering Sea are also included. Year Survey biomass (metric tons) % CPUE; < 30 m Population numbers (billions) Proportion male (m/m+f)2 Mean depth (m)3 Commercial catch (metric tons)4 Males Females Males Females 1982 3,377,838 20.56 9.81 20.67 0.492 45.0 52.9 95,712 1983 3,535,269 22.04 5.41 16.94 0.486 46.3 58.5 108,385 1984 3,141,188 16.96 5.62 14.26 0.434 47.3 58.7 159,526 1985 2,443,666 29.63 11.39 10.60 0.437 43.3 55.2 227,107 1986 1,909,866 31.05 13.38 7.96 0.436 42.8 54.7 208,597 1987 2,613,067 27.47 13.18 10.35 0.438 44.6 53.6 181,428 1988 2,402,369 23.94 9.86 10.09 0.426 45.0 55.6 223,156 1989 2,316,249 28.83 20.05 9.66 0.454 42.3 48.3 153,165 1990 2,183,708 24.30 12.46 9.06 0.435 44.6 53.0 83,970 1991 2,393,268 19.22 5.94 9.54 0.453 44.5 53.8 115,842 1992 2,172,900 34.89 19.01 8.21 0.452 42.3 53.4 149,569 1993 2,465,443 31.64 11.59 10.03 0.442 43.8 54.2 106,101 1994 2,610,474 32.34 13.73 10.70 0.442 40.1 51.0 144,544 1995 2,009,671 33.02 13.66 8.24 0.445 40.2 51.3 124,740 1996 2,298,560 31.28 9.83 9.62 0.460 43.9 56.8 129,659 1 %CPUE=(ICPUE, (,<30)/ICPUE, i;=d/;) x 100 where CPUE,^ = station CPUE of the ith sex within the dth depth range. Note that this estimate is derived from the AFSC survey which did not cover the entire nearshore area. 2 Proportion male = population number of males divided by population number of males and females. 3 Mean depths are weighted by CPUE (kg/hectare) values at each tow location. 4 Total retained and discarded catch estimates for yellowfin sole from foreign ( 1982-87), joint-venture (1982-90), and domestic (1987-present) fisheries in the eastern Bering Sea ( Wilderbuer1). tions are inconsistent with age-structured stock syn- thesis models (Methot, 1990) that predict relatively stable abundance levels after 1981 (Wilderbuer1). Large scale spring-summer Alaska Fisheries Sci- ence Center (AFSC) resource assessment surveys of the eastern Bering Sea shelf began with a baseline survey in 1975 (Pereyra et al.2) and have continued annually since 1979 with the primary purpose of es- timating the abundance of important groundfish and crab species, including yellowfin sole (Gunderson, 1993). In 1982, a change in the standard survey trawl from a 400-mesh “eastern” trawl to a larger 83-112 “eastern,” increased the catch efficiency for most flat- fish species including yellowfin sole (Bakkala, 1993). Survey gear has remained constant since 1982. Al- 1 (continued, from previous page) report for the groundfish re- sources of the Bering Sea/Aleutian Islands region as projected for 1997. North Pacific Fishery Management Council, 605 W. 4th Avenue Suite 306, Anchorage, AK 99501. 2 Pereyra, W. T., J. E. Reeves, and R. G. Bakkala. 1976. Dem- ersal fish and shellfish resources of the eastern Bering Sea in the baseline year 1975. Northwest and Alaska Fisheries Cen- ter Proc. Rep. NOAA-NMFS, 619 p. Alaska Fisheries Science Center, Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA 98115-0070. though these surveys cover most of the shelf area north of the Alaska Peninsula and south of latitude 60°N (Fig. 1), nearshore waters have been excluded owing to shallow bottom depths and variable bottom substrate. Areas such as Togiak Bay (Fig. 1), where commercial trawlers have successfully targeted yel- lowfin sole in waters as shallow as 5-6 m (Low and Narita, 1990), have been excluded from resource as- sessment. Recognition that yellowfin sole spawn during the survey period (June-August) and that nearshore spawning grounds extend into these nonsurveyed areas (Nichol, 1995) has prompted speculation as to whether fluctuations in survey bio- mass may be due to annual variation in yellowfin sole distributions that overlap surveyed and nonsurveyed areas. This study investigates two potential sources of variation in yellowfin sole distribution during spring- summer, which may partially account for fluctuations in survey biomass estimates. In this paper, I describe variations in yellowfin sole distribution patterns be- tween sexes and among years. Data from exploratory nearshore samples are included to demonstrate prob- lems associated with the exclusion of nearshore ar- eas from resource assessments. Nichol: Annual and between-sex variability of Pleuronectes asper 549 Figure I Standard Alaska Fisheries Science Center eastern Bering Sea survey stations. Large numer- als (1-6) indicate strata areas: strata 1-2, <50 m; strata 3-4, 50-100 m; strata 5-6, 100-200 m. Materials and methods Surveys and gear Standard AFSC Owing to differences in standard trawl gear used prior to 1982, only the 1982-96 stan- dard surveys are considered here. About 350 stan- dard stations, ranging in depth from 16 m to 230 m, were sampled annually for each of the 1982-96 AFSC surveys (Fig. 1; Table 2). Survey stations were spaced 37 km (20 nautical miles) apart, following a 37 x 37 km grid pattern. Surveys generally began in inner Bristol Bay, followed transects directed north and south, and progressed east to west with each finished transect. Trawls were towed for about 30 min at a speed of 5.6 km/h (3 knots) at each station during daylight hours. The standard AFSC sampling gear, an 83-112 “eastern” trawl, was characterized by av- erage vertical and horizontal wing openings of 2.6 m and 16.5 m, respectively. Effective net spread and height measurements were obtained by means of a commercial trawl net-mensuration system (Scanmar) that was used during most tows. Footrope and headrope lengths were 25.3 m and 34.1 m, respec- tively. Paired 55-m dandylines (bridles) were at- tached to each wing and to 1.8 m x 2.7 m steel V- doors. Chain extensions measuring 61 cm in length were attached from each end of the footrope to the lower dandyline to improve bottom contact of the footrope. Two chartered commercial trawl vessels were employed each year to complete the standard survey. A total of 9 vessels ranging from 30.0 to 39.6 m in length have been used since 1982. Total yellowfin sole catch weights (kg) as well as random sex-specific length-frequency samples (total length in cm) were collected at each station (haul) where the species was captured. Haul positions, bot- tom depths, and surface and bottom temperatures were also recorded at each station. AFSC exploratory tows Additional 30-minute tows in waters nearer shore than the standard survey area were made during the course of the AFSC surveys with the same trawl gear and sampling procedures. Sampling depths ranged from 9.0 m nearshore to 29.3 m offshore. A total of twenty-five nearshore tows, 13 in the Togiak Bay area and 12 in the Kuskokwim Bay area, were made during the 1988-91 standard 550 Fishery Bulletin 96(3), I 998 Table 2 Summary of AFSC groundfish surveys on the eastern Bering Sea shelf. Tows in which yellowfin sole (YFS, Pleuronectes asper ) were captured are noted parenthetically. Survey Area2 Gear2 Date Bottom depth (m) Number of tows3 Number of YFS measured Standard EBS 83-112 June- Aug 1982 18-137 329 (251) 37,023 June-Aug 1983 20-134 354 (270) 33,924 June-Aug 1984 18-134 355 (279) 33,894 June-Aug 1985 18-154 353 (270) 33,824 June-Aug 1986 18-148 354(251) 30,470 June-Aug 1987 18-112 342 (235) 31,241 June-Aug 1988 18-130 353 (249) 27,121 June-Aug 1989 18-110 354 (236) 29,510 June-Aug 1990 16-126 352 (247) 30,491 June-Aug 1991 18-119 351 (248) 27,985 June-Aug 1992 16-152 336 (231) 23,626 June-Aug 1993 18-163 355 (243) 26,647 June-Aug 1994 16-113 355 (257) 24,420 June-Aug 1995 16-154 356 (249) 22,111 June-Aug 1996 20-114 355 (247) 27,190 Exploratory Kuskokwim Bay 83-112 June 1988 13-27 4(4) 786 area June 1989 9-20 4(4) 734 June 1990 13-29 2(2) 408 June 1991 13-29 2(2) 293 Togiak Bay area 83-112 June 1988 13-29 5(5) 1264 June 1989 16-27 5 (5) 1438 June 1990 11-26 2(2) 515 June 1991 15 1 (1) 107 Nearshore beam trawl Togiak Bay area PSB May, 1995 2-30 28 (28) 4134 1 EBS = eastern Bering Sea shelf area from the Alaska Peninsula north to latitude 62° N. 2 83-112 = Eastern otter trawl with 25.3 m (83 ft) headrope and 34.2 m (112 ft) footrope PSB = plumb-staff beam trawl with a 3.1 m (10 ft) aluminum beam 3 Number in parentheses indicates number of tows in which yellowfin sole were captured. surveys (Fig. 2; Table 2). These tows were conducted to collect preliminary data on spawning yellowfin sole, which were known from commercial trawl re- ports to occur in these nearshore areas. Extensive survey coverage of this area has been limited owing to the shallow bottom depths. Nearshore beam-trawl survey A 3. 1-m plumb staff beam trawl with specifications detailed in Gunderson and Ellis (1986) was used aboard a chartered 9.7-m (32-ft) purse-seine or gillnet vessel. This survey suc- cessfully sampled a total of 28 stations in the shal- low waters of the Togiak Bay area, from 18 May to 28 May 1995 (Fig. 2; Table 2). Sampling depths ranged from 1.8 m nearshore to 30.2 m offshore. Tows were made during daylight hours and were less than 10 min in duration. All yellowfin sole were sexed, measured, and assigned a maturity code. For the purpose of this examination, maturity was classified as either mature or immature. Females were con- sidered mature if ovaries were yolked or recently spent. Males were considered mature if testes were swollen, opaque colored, running with sperm, or re- cently spent. Analysis Yellowfin sole distributions were described in terms of catch per unit of effort (CPUE; kg/hectare) for each of the 1982-96 AFSC standard surveys. Calculation of CPUE followed area-swept methods described by Alverson and Pereyra (1969), whereby catch is the total weight (kg) of yellowfin sole within a station tow and effort is the area (hectare) swept by the path of the trawl. Area swept was computed as the prod- uct of the distance fished and the effective net-spread Nichol: Annual and between-sex variability of Pleuronectes asper 551 + Standard station a Exploratory tow ♦ Beam-trawl tow Figure 2 Geographic location of exploratory (1988-91) and beam-trawl tows (1995) made inshore of the standard AFSC survey stations. for each tow. All fish within the path of the trawl were assumed captured. Male and female catch rates at each station were calculated as follows: CPUE, = CPUE ■ ( Weight Proportion of ith sex ) = CPUE n (l) where CPUE = CPUE/ = n = ni = L = L = l a = b = catch per unit of effort (kg/hectare) of all yellowfin sole at a station; catch per unit of effort (kg/hectare) of the ith sex; total number of fish; number of fish of the ith sex; total length (cm) of the jth fish, males and females included; total length (cm) of the jth fish and ith sex; 0.007441; and 3.130. Constants a and b were calculated with nonlinear regression (nlin procedure; SAS Institute, 1989) with the equation: W = aLb (n=796, r2=0.99), where W=individual yellowfin sole weights (g) measured during the 1987 standard AFSC survey. Prior to es- timating a and 6, a test comparing the length- weight relations between males and females was performed. A test of the null hypothesis of no difference between male and female linear log( length )-log( weight) re- vealed no significant difference in either slopes (analysis of covariance; F- 2.46; df=l,792; P- 0.12) or intercepts (F=2.05; df=l,793; P=0.15). Sexes were there- fore combined to estimate constants a and b above. Annual yellowfin sole “survey biomass” was esti- mated as: n Biomass = ^^[CPUE k ■ Ak^j , (2) k=i where CPUEk = the mean of station CPUE values (Eq. 1) within the &th stratum (Fig. 1); and Ak - the area (hectares ) of the Mh stratum. Proportions of males (no. males/no. both sexes) at each station were averaged across stations by depth 552 Fishery Bulletin 96(3), 1998 group (10-19 m, 20-29 m, etc.), weighted by the sta- tion CPUE (no. fish/hectare). For nearshore beam- trawl samples, variance surrounding mean male-pro- portions was estimated following Cochran ( 1977) for estimation of proportions in cluster sampling. For standard and exploratory surveys, the error struc- ture of proportion means was approximated by the boot-strap method (Efron and Tibshirani, 1993). The station-specific data were randomly resampled and the mean estimated 1000 times. Twenty-fifth and 975th quantiles were chosen for the 95% confidence bounds. Results Between-sex variation Male yellowfin sole within the standard survey area were distributed closer to shore than were females in all years, 1982-96 (Fig. 3; Table 1). In addition to having an overall distribution in deeper water, fe- males also appeared to have a more extended north- westerly distribution than males (Fig. 3). Males out- numbered females near shore, whereas farther off shore ( >60 m bottom depth), females outnumbered males (Fig. 4). The proportion of males (no. males/ no. both sexes) and CPUE increased with decreas- ing depth (Fig. 4). Togiak and Kuskokwim Bay area samples from exploratory trawls were also charac- terized by a higher proportion of males as well as greater overall yellowfin sole concentrations. The proportion of males from Togiak and Kuskowim Bay areas averaged 0.60 and 0.67, respectively (Fig. 4). Sex proportions also varied between mature and immature yellowfin sole. Among beam-trawl samples from the Togiak Bay area (Fig. 2), the overall mean proportion of males was 0.54 (SE=0.0033). Among mature fish, however, the proportion of males aver- aged 0.65 (SE=0.0067) and among immature fish, mostly small <20 cm fish, the proportion of males averaged 0.50 (SE=0.0024). Within the spawning area (<30 m), the proportion of males among mature fish averaged 0.68 (SE=0.0056) (Fig. 5). Amomg-year variation Yellowfin sole spring-summer distributions varied from year to year, although the most notable differ- ence was for years 1982-84 when distributions of males in particular within the standard AFSC sur- vey area were shifted off shore (deeper waters) rela- tive to subsequent years ( 1985-96) (Fig. 3). A greater mean CPUE-weighted bottom depth in conjunction with a decreased percentage of nearshore ( <30 m) CPUE (kg/hectare) during 1982-84 confirmed a bathymetric shift to deeper waters for both male and female distributions in relation to subsequent years (Table 1). Interestingly, yellowfin sole biomass lev- els, as well as the overall proportion of males within the standard AFSC survey, were considerably higher during 1982-83 than in subsequent years (Table 1). Discussion Between-sex variation Distributional differences between male and female yellowfin sole exist primarily because males mature earlier than females. Standard AFSC surveys, ex- ploratory nearshore samples, and Togiak area beam- trawl samples have shown conclusively that male yellowfin sole outnumber females in the nearshore areas ( <30 m bottom depth) of the eastern Bering Sea during spring-summer. Given that males ma- ture approximately 4 years earlier than females (Wilderbuer et al., 1992), spawning males should considerably outnumber spawning females. Because yellowfin sole spawn primarily in nearshore areas ( <30 m) during spring-summer (Nichol, 1995), ma- ture males outnumber mature females by nearly 2:1 in this region. In deeper waters, females outnumber males because of the abundance of larger (25-32 cm TL) immature females that generally do not move into the spawning area (Nichol, 1997). Factors such as differential life-spans and differ- ential catchability between sexes have been shown to cause similar sex-ratio patterns in other species (Beverton, 1964). Such factors were assumed negli- gible in this case. Since 1982, maximum ages for yel- lowfin sole males and females averaged 25.7 and 26.1 years,3 respectively, and have not differed signifi- cantly (paired t-test; P-value=0.7758, df=28), suggest- ing similar life-spans for males and females. Beverton (1964) discussed the potential effects of differential natural mortality, fishing mortality, and catchability on sex-ratios of plaice (Pleuronectes platessa L.) in the North Sea. The possibility that male yellowfin sole may be more readily captured by trawl gear, at least among immature fish, is unlikely given that the sex proportion among immature fish in the spawning area was near 0.5 (Fig. 5). The possibility that spawning males are more catchable than fe- males owing to differential spawning behavior, as Beverton ( 1964) reported for plaice, is not known for 3 Ages were determined by the Age and Growth Unit of the Alaska Fisheries Science Center (AFSC) from annual otolith collections made during the standard AFSC groundfish trawl surveys in the eastern Bering Sea. Nichol: Annual and between-sex variability of Pleuronectes asper 553 180 177 174 171 168 165 162 159 156 180 177 174 171 168 165 162 159 156 180 177 174 171 168 165 162 159 156 180 177 174 171 168 165 162 159 156 CPUE (kg/hectare): - No Catch • < 50 • 50-99.9 * 100-199.9 • > 199.9 Figure 3 Geographical distribution of male and female yellowfin sole (Pleuronectes asper) in terms of catch per unit of effort (CPUE; kg/hectare) for each year, 1982-96. 554 Hshery Bulletin 96(3), 1998 180 177 174 171 168 165 162 159 156 CPUE (kg/hectare): * No Catch ■ < 50 • 50-99.9 • 100-199.9 • > 199.9 Figure 3 (continued) Nichol: Annual and between-sex variability of Rleuronectes asper 555 62 60 58 56 54 62 60 58 56 54 62 60 58 56 54 62 60 58 56 54 62 60 58 56 54 180 177 174 171 168 165 162 159 156 180 177 174 171 168 165 162 159 156 180 177 174 171 168 165 162 159 156 CPUE (kg/hectare): * No Catch • < 50 • 50-99.9 • 100-199.9 ♦ > 199.9 Figure 3 (continued} 556 Fishery Bulletin 96(3), 1998 yellowfin sole. However, because yellowfin sole males were dominant inshore and females were dominant offshore, the more likely cause for the sex-propor- tion patterns observed here was the differential dis- tribution patterns between sexes. Among-year variation Spring-summer distributions during 1982-84 may have been deeper than those in subsequent years, in part, because the population was younger and less mature. Wilderbuer et al. ( 1992) indicated that older ( >17 years) yellowfin sole were an insignificant part of the population prior to the mid-1980s compared with later years and indeed, length distributions were skewed toward smaller length groups during the 1982-84 surveys compared with later years (Fig. 6). Given estimated lengths at 50% maturity of 20.3 and 28.8 cm TL for male and female yellowfin sole, re- spectively (Wilderbuer et al., 1992), many of the fish constituting the strong modes from 1982 to 1984 were sexually immature. Certainly there was a progres- sion from immaturity to maturity for many of these fish. Nichol (1997) showed that larger (25-32 cm) im- mature females nearing maturity maintain a deeper bathymetric distribution than mature spawning fish during spring-summer. With increasing maturity of the population, therefore, we would expect to see a distribution shift to shallower spawning waters in years after 1984. Smaller males (22-27 cm TL) and females (25-32 cm TL) constituted a large pro- portion of the yellowfin sole biomass from 1982 to 1984 (Fig. 7). High concentrations of 22-27 cm males that resided at 40-49 m bottom depth dur- ing 1982-84 were not apparent from 1985 to 1996. Similarly, prominent modes of 25-32 cm females at 30-39 m and 70-79 m depths during 1982-84 were not nearly as apparent during 1985-96 (Fig. 7). Why then was the proportion of males within the standard area higher during 1982-83 than other years? The most plausible explanation is that the deeper overall yellowfin sole distribu- tions during these years rendered the population more available to the survey. Nearshore areas not Bottom depth (m) Figure 5 Mean male proportions (no. males/no. males and females) of mature and immature yellowfin sole ( Pleuronectes asper) by bottom depth during the 1995 beam-trawl survey of Togiak Bay. Mean pro- portions are weighted by the CPUE (no. fish/hect- are) at each station within each depth grouping. Bars are 95% confidence intervals. Population (billions) Nichol: Annual and between-sex variability of Pleuronectes asper 557 Figure 6 Length composition of male (solid line) and female (dashed line) yellowfm sole ( Pleuronectes asper) in terms of population numbers for years 1982-96. The vertical lines provide a reference for fish length at 30 cm TL. 558 Fishery Bulletin 96(3), 1 998 1985-96 Figure 7 Length composition of yellowfin sole (Pleuronectes asper ) in terms of bottom depth and bio- mass (metric tons) for years 1982-84 combined and 1985-96 combined. Shaded areas indi- cate 22-27 cm length groupings for males, and 25-32 cm length groupings for females. Each depth label represents a 10-m range (i.e. 40=40-49 m). sampled during the standard AFSC surveys likely contained more males than females. Hence, a shift of the population to deeper waters possibly exposed a higher proportion of males to the survey gear than during subsequent years. Effects of distribution on survey biomass Given that the standard AFSC survey area does not encompass the entire yellowfin sole distribution, annual distributional shifts between surveyed and nonsurveyed areas may explain the annual biomass fluctuations observed within the standard AFSC sur- vey area. Yellowfin sole biomass estimates within the standard AFSC survey area have fluctuated from a high of 3.5 million t in 1983 down to 1.9 million t in 1986 (Table 1). These fluctuations are not possible according to yellowfin sole life history and the rela- tively light fishing exploitation of this species since the late 1970s (Wilderbuer et al., 1992). Stock syn- thesis models based on catch-at-age data (Methot, 1990) predict much more stable changes in yellow- fin sole biomass (Fig. 8). Even though these models incorporate survey biomass estimates as auxiliary information (Wilderbuer1), predicted biomass esti- mates from 1982 to 1984 were much lower than what survey estimates indicated. I propose that the deeper overall spring-summer distributions from 1982 to 1984 contributed to an inflation of survey biomass estimates in contrast with subsequent years, owing to increased availability of yellowfin sole to the AFSC survey. Strong year classes that were present in the survey area during 1982-84 may have become less accessible to the survey gear during later years ow- ing to their sexual maturation and subsequent mi- gration into shallower nearshore spawning waters. Nearshore areas from the standard survey bound- ary to the coastline (excluding river systems and in- Nichol: Annual and between-sex variability of Pleuronectes asper 559 ner bays) totaled approximately 40,184 km2 (Fig. 1). In contrast, the two nearest shore strata within the AFSC survey area total 77,871 km2 (stratum 1) and 41,027 km2 (stratum 2), respec- tively. The coastline areas, thus, offered yellow- fin sole a considerable refuge from the AFSC bottom trawl survey. According to past samples from Togiak Bay and Kuskokwim Bay areas (Table 2), yellowfin sole are more concentrated in the coastline areas than within the standard AFSC survey area (Fig. 4). Considering the vast unsampled areas near shore, as well as the large yellowfin sole concentrations that inhabit these areas, annual fluctuations in biomass within the standard AFSC area should be expected. Biomass estimates generated from standard AFSC eastern Bering Sea trawl surveys are considered a relative measure of population abundance. However, because nonsurveyed nearshore areas contain higher concentrations of yellowfin sole and because males outnumber females there, standard area biomass estimates underestimate male abundance. Distribution plots suggest a more extended northwesterly distribution of females compared with males. Within the northwest area (subareas 2, 4, and 6) the average proportion of males from 1982 to 1996 was 0.33 ± 0.016 (95% confidence) compared with 0.50 ± 0.010 in the southeast area (subareas 1, 3, and 5). The lack of males in the northwest area may be an indication that a large percentage of males missed by the survey inhabit nearshore waters of Kuskokwim Bay and waters east of Nunivak Island (Fig. 1). Bottom temperature may be one factor that has influenced the availability of yellowfin sole to the standard AFSC survey. Except for the years 1982- 84, when fish availability was affected more by popu- lation structure, biomass estimates were generally lower during colder years (Fig. 9). Midshelf bottom temperatures were chosen to represent annual tem- perature regimes because water exchange is mini- mal when compared with the inner shelf (< 50 m) and outer shelf (>100 m) waters (Coachman, 1986; Wilderbuer et al., 1992). Distribution patterns, as well as timing of the annual yellowfin sole cross-shelf migration, may in part be dependent upon the an- nual temperature regime. Fadeev (1965) indicated that spring yellowfin sole migrations occurred one month earlier than usual during years when warm water developed earlier on the eastern Bering Sea shelf. Pola et al. (1985) also demonstrated, using simulation models, that summer yellowfin and rock sole distributions could be established two months earlier in a warm year than in a cold year. If ice-edge Figure 9 Increase in estimates of yellowfin sole ( Pleuronectes asper) biomass as related to mean midshelf (50-100 m) bottom temperatures for years 1985-1995. Line represents the lin- ear regression of biomass (dependent variable) against bottom temperature (independent variable); r2 = 0.32, n = 12, slope = 0.15 (two-tailed P=0. 054), intercept = 2.08. Note that 1982-84 data were excluded because the population structure was unique during those years. 560 Fishery Bulletin 96(3), 1998 retreat determines the timing of inshore yellowfm sole migration (Bakkala, 1981), and subsequently the timing of spawning, distribution patterns at the time of the survey could be affected. Increasing percent- ages of spent yellowfm sole during the June 1993 AFSC survey (Nichol, 1995) indicated that spawn- ing activity progresses within the duration of the survey. Spring-summer surveys conducted in warmer years may intercept higher proportions of spent individuals that no longer inhabit the unavail- able nearshore spawning areas. Interannual variation of yellowfin sole biomass estimates within the eastern Bering Sea may be re- duced under alternative survey designs (McAllister, 1995). I recommend the incorporation of additional standard stations in nearshore areas, such as the exploratory tows made in Togiak and Kuskokwim Bays (Fig. 2), and east of Nunivak Island. The addi- tion of such stations would likely reduce interannual variation of survey biomass estimates as well as pro- vide a more accurate representation of the popula- tion sex-composition. Acknowledgments I thank Terry Sample, Gary Walters, Claire Armi- stead, Pam Goddard, Bob McConnaughey, and nu- merous other AFSC personnel who contributed to the AFSC surveys from 1982 to 1996. I also appreciate the efforts of Tom Crawford, skipper of the FV Crawdad, during the Togiak Bay beam-trawl survey. Peter Munro and Steve Syrjala helped with the sta- tistical analysis. Dave Somerton, Jay Orr, Gary Walters, Gary Stauffer, James Lee, and Gary Duker provided valuable comments on early versions of this manuscript. David Sampson and two anonymous reviewers helped strengthen later versions of the manuscript. Literature cited Alverson, D. L., and W. T. Pereyra. 1969. Demersal fish explorations in the northeastern Pa- cific Ocean: an evaluation of exploratory fishing methods and analytical approaches to stock size and yield forecasts. J. Fish. Res. Board Can. 26:1985-2001. Bakkala, R. G. 1981. Population characteristics and ecology of yellowfin sole. In D. W. Hood and J. A. Calder (eds.). The eastern Bering Sea shelf: oceanography and resources, vol. 1, p. 553-574. U.S. Dep. Commer., NOAA, Off. Mar. Poll. As- sess., U.S. Gov. Print Off., Wash., D.C. 1993. Structure and historical changes in the groundfish complex of the eastern Bering Sea. U.S. Dep Commer., NOAA Tech. Rep. NMFS 114, 91 p. Beverton, R. J. H. 1964. Differential catchability of male and female plaice in the North Sea and its effect on estimates of stock abun- dance. Rapp. P.V. Reun. Cons. Int. Explor. Mer 155:103- 112. Coachman, L. K. 1986. Circulation, water masses, and fluxes on the south- eastern Bering Sea shelf. Continental Shelf Res. 5(1/ 2):23-108. Cochran, W. G. 1977. Sampling techniques, third ed. John Wiley and Sons, Inc., New York, NY, 428 p. Efron, B., and R. J. Tibshirani 1993. An introduction to the bootstrap. Chapman and Hall, New York, NY, 436 p. Fadeev, N. S. 1965. Comparative outline of the biology of flatfishes in the southeastern part of the Bering Sea and condition of their resources. Translated by Isr. Prog. Sci. Transl., 1968. In P. A. Moiseev (ed.), Soviet fisheries investigations in the northeastern Pacific, part 4, p 112-129. [Available from U.S. Dep. Commer., Natl. Tech. Inf. Serv., Springfield, VA, as TT 67-51206.] 1970. Fisheries and biological characteristics of the east- ern Bering Sea yellowfin sole. Translated by Isr. Prog. Sci. Transl., 1972. In P. A. Moiseev (ed.), Soviet fisheries in- vestigations in the northeastern Pacific, part 5, p 332- 396. [Available from U.S. Dep. Commer., Natl. Tech. Inf. Serv., Springfield, VA, as TT 71-50127.) Gunderson, D. R. 1993. Surveys of fisheries resources. John Wiley & Sons, Inc., NY, 248 p. Gunderson, D. R., and I. E. Ellis. 1986. Development of a plumb staff beam trawl for sam- pling demersal fauna. Fish. Res. 4:35-41. Hart, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can. Bull. 180, 740 p. Low, L., and R. E. Narita. 1990. Condition of groundfish resources in the Bering Sea- Aleutian Islands region as assessed in 1988. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-F/NWC-178, 224 p. McAllister, M. K. 1995. Using decision analysis to choose a design for sur- veying fisheries resources. Ph.D. diss. Univ. Washington, Seattle, WA, 293 p. Methot, R. D. 1990. Synthesis model: an adaptable framework for analy- sis of diverse stock assessment data. Symposium on appli- cation of stock assessment techniques to gadoids. Inter- national North Pacific Fisheries Commission (INPFC) Bull. 50:259-277. Nichol, D. G. 1995. Spawning and maturation of female yellowfin sole in the eastern Bering Sea. In Proceedings of the international flatfish symposium, October 1994, Anchorage, Alaska, p. 35- 50. Univ. Alaska, Alaska Sea Grant Rep. 95-04. 1997. Effects of geography and bathymetry on growth and maturity of yellowfin sole, Pleuronectes asper, in the east- ern Bering Sea. Fish Bull. 95(3):494-503. Pola, N. B., E. J. Rykiel Jr., and W. E. Grant. 1985. Numerical simulation of fish migrations in the east- ern Bering Sea. Ecol. Model. 29:327-351. SAS Institute Inc. 1989. SAS/STAT user’s guide, version 6, fourth ed., vol. 2. SAS Institute Inc., Cary, NC, 846 p. Nichol: Annual and between-sex variability of Pleuronectes asper 561 Wakabayashi, K. 1989. Studies on the fishery biology of yellowfin sole in the eastern Bering Sea. [In Jpn., Engl, summ.] Bull. Far Seas Fish. Res. Lab. 26:21-152. Wilderbuer, T. K., G. E. Walters, and R. G. Bakkala. 1992. Yellowfin sole, Pleuronectes asper , of the eastern Bering Sea: biological characteristics, history of exploita- tion, and management. Mar Fish. Rev. 54(4): 1-18. Witherell, D. B. 1995. Management of flatfish fisheries in the North Pacific. In Proceedings of the international flatfish sym- posium, October 1994, Anchorage, Alaska, p. 35-50. Univ. Alaska, Alaska Sea Grant Rep. 95-04. 562 Age, growth, mortality, and population characteristics of the Pacific red snapper, Lutjanus pern, off the southeast coast of Baja California, Mexico Axayacatl Rocha-Olivares Universidad Autonoma de Baja California Sur Departamento de Biologia Marina Apartado Postal 1 9-B, La Paz Baja California Sur, 23080 Mexico Present address: Department of Biological Sciences, 508 Life Sciences Building Louisiana State University, Baton Rouge, Louisiana 70803-1715 E-mail address: arocha@lsu.edu Abstract .—Pacific red snapper, Lutjanus peru, collected from commer- cial hook-and-line gear and shrimp trawls off the southeast coast of Baja California Sur, Mexico, were aged by using scales and otoliths (whole-otolith and sectioned-otolith readings). Com- parison of ages determined from these structures revealed that scales tend to underestimate ages beyond 5 years, of- ten by more than 1 year, and that they are the least precise structures for age- ing. Sectioned otoliths gave greater es- timates of age than whole-otolith counts (mean differences <1 yr) and were the most precise structure. The von Bertalanffy growth function de- scribed L. peru growth satisfactorily with length-at-age determined from whole otoliths and sectioned otoliths. Parameters for the entire population were Lx = 97.32 cm, K = 0.1111/yr, f0 = -0.316 yr ( rz = 1 180 ). No significant dif- ferences in length-at-age were found be- tween sexes. The largest individual was a 31-year-old 99.2-cm-TL male, consti- tuting a maximum age, and length record for this species. A multiple re- gression model of whole-otolith age as a function of otolith and fish measure- ments provided satisfactory results. Total mortality rates were significantly lower for females (Z=0.282/yr) than for males (Z=0.366/yr). Manuscript accepted 24 November 1997. Fishery Bulletin 96:562-574 (1998). The Pacific red snapper, Lutjanus peru (Nichols and Murphy), known as “huachinango” in Mexico, is dis- tributed throughout the lower Gulf of California to Peru and found in offshore schools over rocky bottoms to depths exceeding 100 m. It is a commercially important species in Mexico as well as in Central and South America (Thomson et ah, 1987; Gutierrez Vargas, 1990; Ra- mirez Rodriguez and Rodriguez Medrano, 1990). Lutjanus peru is fished in most Mexican states along the Pacific coast. The highly valued product is generally marketed whole and transported to inland cities, or oc- casionally exported. Nationwide, it shares a market with its congener, the Gulf of Mexico red snapper Lutjanus campechanus Poey. From 1980 to 1988, the reported catches of both species averaged 6556 met- ric tons (t) per year (SD 1290 t/yr), of which 56% originated in the Pa- cific.1 In Baja California Sur, Pacific red snapper is fished on a small artisanal scale; near Cerralvo Is- land, L. peru ranks within the six most important exploited finfishes (Fig. 1) (Ramirez Rodriguez and Rodriguez Medrano, 1990). As is the case with other snappers, juveniles of this species aggregate over soft bottoms where they are caught as bycatch during shrimp trawling ac- tivities (Van Der Heiden, 1985; author’s personal observations). Information on the biology of L. peru is limited despite its ecologi- cal and commercial importance. Gorelova (1979) performed feeding experiments on juvenile L. peru caught off the coast of Peru, whereas Ruiz Santos (1983) studied the re- productive biology of the species off the southwest coast of Mexico and Ruiz and Madrid ( 1992 ) studied the biology of a parasitic isopod and its effects on L. peru hosts off Michoacan. With the exception of Rocha Olivares and Gomez Munoz (1993), age and growth of the Pacific red snapper have been estimated either from scales in Mexico (Castro, 1981; Ruiz Luna et al., 1985; Aguilar Salazar, 1986) or from length-frequency dis- tributions in Costa Rica (Gutierrez Vargas, 1990). The use of scales, however, may result in less than accurate age estimates for some 1 Anuarios Estadisticos de Pesca, Secretaria de Medio Ambiente Recursos Naturales y Pesca. Av. Progreso Num. 5, Colonia Del Carmen, Coyoacan. Delegacion Coyoacan 04110, Mexico, D.F. Rocha-Olivares: Age, growth, and mortality of Lutjanus peru 563 species, including systematic underestimations (e.g. Bilton, 1973; O’Gorman et al., 1987; see review by Beamish and McFarlane, 1987, and papers therein). Because of the lack of reliable information on biological parameters for this commercially important tropical species, this work represents the first comprehensive study of growth and mortality of a L. peru population based on otolith age determinations. A comparison of age estimates from whole and sectioned otoliths and from scales is made to assess their relative use- fulness. Growth and mortality rates are esti- mated, and other aspects of the biology are pre- sented for an exploited population of the Pa- cific red snapper off the southeast coast of Baja California Sur. Materials and methods Study area and sampling scheme This study took place in Bahia de La Paz, one of two major bays in the Gulf of California {Walker, 1960). The most important fishing grounds for L. peru lie within and southeast of the Bay, around Espiritu Santo-Partida and Cerralvo Islands (Fig. 1). Peak catches occur during the summer. Pacific red snapper were collected from March 1989 to March 1991. Most samples came from La Paz City fish market, which was sampled at least weekly, March 1989 to February 1991. Snappers were also collected monthly with baited hook and line, May 1989 to February 1991. Juvenile red snappers were collected as shrimp trawler bycatch during Febru- ary and March 1991. Biological samples were taken from all fish collected (n=2605). In the market, sam- pling time was limited to the period before fish were stored, therefore biological sampling was restricted to the first 20 randomly selected individuals from each 4-cm length interval. Biological samples con- sisted of both sagittae, at least 10 scales from under- neath the left pectoral fin, and testes and ovaries, which were fixed in 10% buffered formalin and used for sex determination and reproduction studies. To- tal and standard lengths (TL and SL, respectively) were measured to the nearest mm; total and gutted weights (TW and GW, respectively) were recorded to the nearest 2 or 5 g, depending on fish size. Since fish were gutted on landing, some data and samples could not be obtained from the market. However, many gutted fishes could be sexed from gonadal re- mains. In 1989, when samples were most abundant, Figure 1 Map of study area showing approximate positions (•) of fishing grounds and sampling sites of Pacific red snapper Lutjanus peru. Depth contours are in fathoms. 3085 additional fishes were measured to construct sex-specific length-frequency distributions. Sample processing and age determination Because the left and right sagittae did not differ morphometrically2 or in weight (paired f-tests, P>0.5) in a random subsample of 50 fishes representative of the length range available and since no differences were found in their marking pattern or number of rings, the right otolith was used for age determina- tion when available. Otoliths were submerged in 90% glycerol for 24 hours before they were viewed under a dissecting scope with reflected light over a dark- ened background. Otoliths embedded in thermoplas- tic cement were sectioned with a low-speed saw. Three sections (0.3 mm thick), including the primor- dium and the flanking regions, were made orthogo- nal to the anteroposterior axis of the structure. Sec- tions were glued to glass slides with cyanoacrylate- based cement and prepared for observation as de- 2 Measurements compared are those described in Rocha Olivares and Gomez Munoz (1993). 564 Fishery Bulletin 96(3), 1998 scribed for whole otoliths. Scales were mounted intact between two glass slides and observed under a dissecting microscope at different magni- fications with transmitted light. Organisms were aged by counting the number of growth marks found in scales (scale age), whole otoliths (whole- otolith age), and sectioned otoliths (sectioned-otolith age). Because a large number of fishes occurred within a re- stricted length range (Fig. 2), age estimates for whole-otolith ring counts were made on 50 randomly subsampled otoliths from each 2-cm fish-length in- terval (72=1356) (FAO, 1982). After whole-otolith age deter- minations were made, a strat- ified random subsample of these otoliths was sectioned (n= 151) and the corresponding scales were used for ageing ( 10 per fish). Only nonregenerated scales were used for age determinations . Two independent whole-otolith age determinations were made at different times by two readers. A third estimate was then made three months later for those otoliths with different ages. Lack of consensus resulted in the otolith being discarded as “noninterpretable,” and excluded from the growth analysis. An effort was made to note the cause for rejection. Sectioned otoliths and scales were subject to two reading rounds by one reader separated by a three-month interval. Estimates of the precision (variation between dif- ferent reading rounds for each structure) and of the corroboration (variation among structures) were as- sessed by comparing the percentage of discordant determinations between two data sets (%D) and the index of average percent error of Beamish and Fournier (%E) (1981). Following the method proposed by Boehlert ( 1985) of using objective criteria and multiple regression models to determine fish age from otolith measur- able parameters, an effort was made to predict Pa- cific red snapper whole-otolith age by using otoliths and fish morphometries. This method, originally con- ceived to save time and costs as well as to reduce the subjectivity involved in otolith reading, was imple- mented on L. peru data and the resulting model used to determine the age of fishes whose otoliths could Total length (cm) Figure 2 Length-frequency distribution of Lutjanus peru collected by hook-and-line and shrimp trawling activities in the southeast coast of Baja California Sur, Mexico. Only fishes from trawls, hook and line, and the market were used for the ageing study. not be read. Independent variables included otolith weight (to the nearest mg), width, length, ventral radius, anterior radius, fish TL, GW, their logarith- mic transformations and squared terms, and a modi- fied condition factor K (gutted weight/ TL3). Data from a stratified random subsample (FAO, 1982) from the whole-otolith age determinations were used for this analysis (n=285). A stepwise procedure with an inclusion level ofP=0.05 was used to select variables for the model. Homoscedasticity and normality were evaluated by residual analysis (Zar, 1984). Growth Individual length-at-age (whole-otolith and sec- tioned-otolith) data were used to fit the von Bertalanffy growth function (VBGF): Lt^L„(l-e~KU~to))j where Lf = length at age t\ asymptotic length; growth coefficient; and age at which fish length would be zero if it grows according to the model. L K tn Hotelling’s T2 test (Bernard, 1981) was used to com- pare growth parameters. Rocha-Olivares; Age, growth, and mortality of Lutjanus peru 565 Month Figure 3 Percentage of translucent margins of the whole otoliths of Lutjanus peru. Sample sizes are shown next to symbols. Observed length-at-age data were obtained from whole-otolith readings for most fishes (n=1170) and from sectioned-otolith read- ings for the largest speci- mens (rc=10). To correct for within-year growth, ages were assigned as follows. Since translucent margin deposition on the otolith peaks in July (Rocha Oli- vares and Gomez Munoz, 1993; Fig. 3), only fishes caught during this month were assigned integer ages. For the rest, subsequent growth was accounted for by assigning fractional ages in proportion to the elapsed time (e.g. a fish with two annuli caught in January was assigned 2.5 years). Length-weight relation- ships were fitted to the data and used to calculate the von Bertalanffy asymptotic weight (Wm). Growth perfor- mance <|) = log10 ( K) + 2/3 logjoOFJ was computed from the growth parameters (Manooch, 1987). All coefficients of determination reported for non- linear models were computed by using the following expression (Draper and Smith, 1981): were used to obtain approximate estimates (Pauly, 1980; Ralston, 1987). A mean bottom temperature of 14°C was assumed for calculating M. This tempera- ture prevails throughout most of the Gulf at depths between 100 and 300 m (Maluf, 1983). R2 = 1- Yj{y~y)2 ^(y-y)2 Results where y = observed values; y = predicted values; and y = mean values. Mortality An age-length key was constructed from whole- otolith ages for the sexes combined and applied to the length-frequency distributions to construct popu- lation age distributions by sex. Total mortality rate (Z) was determined from the descending limb of the resulting catch curves for males, females, and pooled sexes. Hoenig’s ( 1983) combined regression equation was used to obtain another estimate of Z. Owing to the lack of data concerning fishing effort and the unavailability of unfished areas, no direct estima- tion of the natural mortality rate (M) was possible. Instead, empirical relationships between M and K Population parameters and morphometric relations During the sampling period L. peru abundance was variable, and in some months (e.g. March and De- cember 1989 and July to August 1990), sample size was small because of limited supply. Sex determina- tion was possible for only 13.1% of the specimens used for length-frequency distributions (males:females= 1:0.84; H0= 1:1 ratio, Xc=2.85, P=0.091). In the bio- logical sampling, the sex ratio was 1:0.85 (H0= 1:1 ratio, x2=5.36, P=0.021 ), indicating that no bias was introduced in the length-frequency sampling with only gonad fragments. Macroscopic gonad differentiation occurs in L. peru between 30 and 35 cm TL (Table 1 ). A significant frac- tion of the catch included individuals smaller than 50 cm TL (38% of a total of 2171 kg of gutted fish). 566 Fishery Bulletin 96(3), 1998 Table 1 Total length composition (cm) of the biological sample of Pacific red snapper by sex. The column labeled “unsexed” groups fishes with undifferentiated gonads (almost all of those <30 cm) and those gutted and with no traces of go- nadal tissue (most of those between 30 and 40 cm and all of those > 40 cm). Size range Males Females Unsexed Total 10.1-15.0 0 0 24 24 15.1-20.0 0 0 259 259 20.1-25.0 4 5 366 375 25.1-30.0 23 23 404 450 30.1-35.0 67 59 305 431 35.1-40.0 109 76 117 302 40.1-45.0 66 46 59 171 45.1-50.0 40 28 56 124 50.1-55.0 32 24 52 108 55.1-60.0 37 17 38 92 60.1-65.0 26 25 26 77 65.1-70.0 26 23 18 67 70.1-75.0 21 29 11 61 75.1-80.0 9 30 7 46 80.1-85.0 5 9 0 14 85.1-90.0 0 3 0 3 90.1-95.0 0 0 0 0 95.1-100.0 1 0 0 1 Total 466 397 1742 2605 Male and female length compositions were signifi- cantly different (Kolmogorov-Smirnov two-sample test, D- 0.0836, 0.010.05); therefore data were pooled: TL= 1.246SL + 0.104 TW=1.207GW - 0.020 GW= 1.763 x 10-5 TL2 877 TW= 1.816 x 10“5 TL2 905 (rc=2603, r2= 0.997, P<0.001) (n=841, r2=0.988, P<0.001) (n=2350, P2=0.990, P<0.001) (n=844, P2=0.980, P<0.001). Exponents of the length-weight relationships were significantly different from three (f-test, P<0.05). Age determination Concentric annuli and circuli were observed in most otoliths and scales. Most scales were found to be in- terpretable, although circuli were never as clearly defined as otolith annuli. Most samples in-92) pre- Table 2 Values of percent disagreement (%D) and percent average error (%E) for whole-otolith, sectioned-otolith, and scale age determinations for individual structures (precision) and between the structure and whole-otolith age (between structures) of Pacific red snapper. No. = number of determinations. Rejec- tions are otoliths not included in the analysis for being too large (size) or for lacking a clear marking pattern (other). Index Rejections Age %D %E No. Size Other Precision Whole-otoligh 16.90 3.89 143 7 1 Sectioned-otolith 10.42 2.13 144 1 6 Scale 34.01 4.19 147 2 2 Between structures Sectioned-otolith 26.36 2.82 129 Scale 58.65 8.86 133 sented less than 40% regenerated scales. A correla- tion was found between the percentage of regener- ated scales and fish length (r=0.310, PcO.001). Of the 1356 otoliths used for whole-otolith age de- terminations, 186 were discarded for the following reasons: large size prevented the enumeration of all the growth rings (2.2%), breakage or loss (1.4%), de- formities or abnormal calcification (2.3%), and non interpretable otoliths as defined in the “Methods” section (7.8%). Whole-otolith ages were available for 1170 individuals ranging from 10.2 to 83.5 cm TL. A number of scales and otoliths were also rejected in the comparison of whole-otolith, sectioned-otolith, and scale ages (Table 2). Noninterpretable otoliths were included in the computation of precision indi- ces, but not in the between-structures indices. Most of the rejections of whole otoliths were due to size, but a similar proportion of sectioned otoliths lacked a clear marking pattern. Sectioned otoliths and scales were the most and least precise structures, respec- tively, for age determination (Table 2). The %D indi- cates a factor of three difference in the precision of sectioned otoliths and scales. However, the %E shows that such differences are less important (less than a factor of two) because this index incorporates the dif- ference in age estimations (Beamish and Fournier, 1981). The difference between whole-otolith ages and sectioned-otolith or scale ages was much larger (Fig. 4). Sectioned-otolith ages were found to be at least twice as similar to whole-otolith ages as scale ages according to %D, and more than three times as similar according to %E (Table 2). The magnitude of the discrepancies in age determi- nations among structures increased with both age and length of the fish (Fig. 4). Differences between scale Rocha-Olivares: Age, growth, and mortality of Lutjanus pern 567 < / > I >. 05 Q. => O O) ® CT> < 16 r 14 12 10 8 6 4 2 0 16 r 14 12 10 8 6 4 2 0 4s L_ 0 6 12 1 3 5 1 2 2 15 1 S 2 2 21 7 2 2 1 8 1 _L 1 i 2 1 4 1 1 1 1 1 CO 0 5L CD 1 S zr o 8 10 12 14 16 10 20 30 40 50 60 70 80 90 Whole-otolith age (years) Total length (cm) Figure 4 (A) Scatter plots of sectioned-otolith and scale ages against whole-otolith ages. Numbers represent the frequency of occurrence. (B) Mean age difference (± standard error of the mean) of scale and sectioned-otolith ages minus whole-otolith age against length of Lutjanus peru. and whole-otolith ages were as large as six years, where- as for sectioned otoliths, differences did not exceed three years (Fig. 4A). Scale ages were consistently lower than whole-otolith ages in fish larger than 50 cm TL, whereas the largest sectioned-otolith and whole-otolith age de- viations were found in fish above 40 cm TL (Fig. 4B). The mean sectioned-otolith and whole-otolith age de- viations did not exceed one year, although mean scale and whole-otolith age deviations often exceeded one year. Six independent variables with highly significant coefficients were included in the multiple regression model (Table 3). The model explained 96% of the vari- ability in the data with a standard error of less than one year. A linear regression of whole-otolith age on otolith weight alone accounted for 93% of the vari- ance. The residuals revealed a trend indicating a certain degree of heteroscedasticity but were nor- mally distributed (x2=10.41, P= 0.005). Whole-otolith ages determined with the multiple regression model for 28 fish with unreadable otoliths were very close to the VBGF (see “Discussion” section). Growth Mean observed length-at-age were not significantly different between sexes, ft-tests, 0.95>P>0.09), however Table 3 Regression coefficients for the multiple regression model of whole-otolith age as a function of otolith and fish mea- surements of Pacific red snapper. SE = coefficient stan- dard error, SEE = standard error of the estimate. R = mul- tiple correlation coefficient; No. = number of observations. Variable Coefficients SE P Intercept (c) -6.183 1.0940 0.000 Total length squared 1.306E-03 1.4910E-04 0.000 ln( Gutted weight) 1.353E-05 3.1700E-06 0.000 Otolith weight squared 5.915 0.7313 0.000 Otolith width squared -7.391E-01 2.4400E-01 0.000 In (otolith ventral radius) 2.153E-02 7.6050E-03 0.005 Otolith weight -1.014E-02 3.3360E-03 0.003 R = 0.9797 SEE = 0.6469 No. = 283 male and female VBGF parameters computed sepa- rately from individual data were significantly differ- ent (P2=609.5, P«0.001) (Fig. 5). The 99% Roy-Bose simultaneous confidence intervals indicated that the three VBGF parameters significantly contributed to the observed difference in predicted growth (6.091. = 10.84 kg. The growth performance of L. peru was (J> = 1.736. The age composition of the sampled males differed significantly from that of the females (Kolmogorov- Smirnov two-sample test, D=0.0843, 0.01 90 1 1 Total 16 166 133 221 172 112 86 66 61 42 41 22 13 13 6 1 3 1 1 1 2 1 1180 Table 5 Comparison of growth parameters of Lutjanus peru obtained in studies from Mexico and Costa Rica (F-W=Ford-Walford plot, NL=nonlinear regression, n=sample size, LFA=length-frequency analysis, obs=observed lengths, back=back-calculated lengths, NA=not applicable). Reference A B C D E Study site Michoacan Michoacan Guerrero Oaxaca Baja California Sur NW Costa Rica (Los Cabos region) Baja California Sur (Bahia de La Paz region) Structure used scales scales scales NA otoliths Method F-W F-W F-W LFA NL Data obs. back obs. back NA obs L r (cm) 81.5 79.5 82.64 66.71 83.34 92.85 K (per yr) 0.196 0.191 0.11 0.23 1.46 0.12 t0(yr) 0.725 0.786 1.48 -0.54 0.04 0.14 n 1068 412 175 208 5902 1180 Age range (yr) 1-7 1-7 0-2 0-8 1-12 0-31 Size range (cm) 23-62 24-64 21-37 20-59 24-82 10-99 A = Ruiz et al. ( 1985); B = Aguilar Salazar ( 1986); C = Castro ( 1981 ); D = Gutierrez Vargas ( 1990); and E = this study. Rocha-Olivares: Age, growth, and mortality of Lutjanus peru 573 between sexes. The higher mortality rate of males may account for the biased sex ratio. In no other lutjanid species has such a large difference in total mortality been observed between males and females. Whether this difference results from different avail- ability to the fishing gear, spatial segregation, or for- aging behaviors (F- associated difference), or from dif- ferences in predator vulnerability or natural longevity (M-associated difference) requires further research. In this paper I have compared the merit of scales, whole otoliths, and sectioned otoliths to age the Pa- cific red snapper, L. peru. As is the case for other tropi- cal and semitropical fishes, the usefulness of whole- otolith age determination is limited up to a certain age, after which, sectioned otoliths are the only reli- able method for ageing and scales should be avoided. Age determinations from whole and sectioned otoliths of fishes covering most of the known size range of L. peru were used to determine the VBGF parameters. A 31-year-old fish measuring 99.2 cm is reported and constitutes a record size and age. Total mortality rates were higher for males resulting in a prepon- derance of females among the older fishes. Acknowledgments I am very grateful to Jon Elorduy Garay and Victor M. Gomez Munoz for their support and encourage- ment. Juan G. Diaz helped in the laborious task of reading otoliths. Thanks to Silvia Ramirez and Roberto Carmona for their help and enthusiasm. Initial drafts of the paper benefited from comments by Larry Jacobson, Nancy Lo, Richard Rosenblatt, and John Butler, and were prepared while the au- thor held a predoctoral fellowship from the Mexican Consejo Nacional de Ciencia y Tecnologia (CONACyT). Four anonymous reviewers provided insightful com- ments and kindly pointed to additional references. Financial support was obtained from DGICSA-SEP grants C89-01-0191 and C90-01-0406 to Jon Elorduy Garay. Literature cited Aguilar Salazar, F. A. 1986. Determinacion de la edad y estimacion de la tasa de crecimiento del huachinango del Pacifico mexicano Lutjanus peru (Nichols y Murphy, 1922), por el metodo de lectura de escamas. Tesis profesional, Facultad de Ciencias, Universidad Nacional Autonoma de Mexico, 74 p. Beamish, R. J., and D. A. Fournier. 1981. A method for comparing the precision of a set of age determinations. Can. J. Fish. Aquat. Sci. 38:982-983. Beamish, R. J., and G. A. McFarlane. 1987. Current trends in age determination methodology. In R. C. Summerfelt and G. E. Hall (eds.), Age and growth of fish, p. 15-42. Iowa State Univ. Press, Ames, IA. Bernard, D. R. 1981 Multivariate analysis as a means of comparing growth in fish. Can. J. Fish. Aquat. Sci. 38:233-236. Bilton, H. T. 1973. Effects of starvation and feeding on circulus forma- tion on scales of young sockeye salmon of four racial ori- gins and of one race of young kokanee, coho and chinook salmon. In T. B. Bagenal (ed.), The ageing of fish, p. 40- 70. Unwin Brothers Ltd., Surrey, England. Boehlert, G. W. 1985. Using objective criteria and multiple regression models for age determination in fishes. Fish. Bull. 83( 2 ): 103—1 17. Castro, F. C. 1981. Determinacion de la edad y crecimiento ( Lutjanus peru). Ciencias del Mar 1:4-8. Coggan, R., K. Skora, A. Murray, and M. White. 1990. A comparison between age determinations of the Antarctic fish Notothenia gibberifrons Lonnberg using scales and otoliths. Cybium 14(11:43-55. Draper, N. R., and H. Smith. 1981 Applied regression analysis. John Wiley, New York, NY, 709 p. FAO. 1982. Metodos de recoleccion y analisis de datos de talla y edad para la evaluacion de poblaciones de peces. FAO Fisheries Circular 736, FAO, Rome, 101 p. Ferreira, B. P., and G. R. Russ. 1994. Age validation and estimation of growth rate of the coral trout, Pleetropomus leopardus , (Lacepede 1802) from Lizard Island, Northern Great Barrier Reef. Fish. Bull. 92(0:46-57. Gorelova, T. A. 1979. Experimental data on the digestion rate in finger- lings of the snapper, Lutjanus peru. J. Ichthyol. 19(6): 161-164. Grimes, C. B. 1987. Reproductive biology of the Lutjanidae: a review. In J. J. Polovina and S. Ralston (eds.), Tropical snappers and groupers: biology and fisheries management, p. 239- 294. Westview Press, Inc., Boulder, CO. Gutierrez Vargas, R. 1990 Tasas de crecimiento, mortalidad, reclutamiento, rendimiento y biomasa relativos por recluta de Lutjanus peru (Perciformes: Lutjanidae) en el Pacifico Noroeste de Costa Rica. Rev. Biol. Trop. 38(2B):441-447. Hoenig, J. M. 1983. Empirical use of longevity data to estimate mortal- ity rates. Fish. Bull. 82(11:898-903. Libby, D. A. 1985. A comparison of scale and otolith aging methods for the alewife, Alosa pseudoharengus. Fish. Bull. 83(41:696-701. Lou, D. C. 1992. Validation of annual growth bands in the otolith of tropical parrotfishes Scarus schlegeli (Bleeker). J. Fish Biol. 41(5):775-790. Loubens, G. 1980. Biologie de quelques especes de poissons du lagon Neo-Caledonian III. Croissance. Cahiers de l’lndo-Pacific- que 2:101-153. Lowerre-Barbieri, S. K., M. E. Chittenden Jr., and C. M. Jones. 1994. A comparison of a validated otolith method to age 574 Fishery Bulletin 96(3), 1 998 weakfish, Cynoscion regalis, with the traditional scale method. Fish. Bull. 92(31:555-568. Maluf, L. Y. 1983. Physical oceanography. In T. J. Case and M. L. Cody (eds.), Island biogeography of the Sea of Cortez, p. 26- 45. Univ. California Press, Berkeley, CA. Manooeh, C. S., III. 1987. Age and growth of snappers and groupers. In J. J. Polovina and S. Ralston (eds.), Tropical snappers and grou- pers: biology and fisheries management, p. 329-373. Westview Press, Inc., Boulder, CO. Mathews, C. P., and M. Samuels. 1985. Stock assessment and management of newaiby, hamoor and hamra in Kuwait. In C. P. Mathews (ed.), Final report: the proceedings of the 1984 shrimp and fin fisheries, management workshop, p. 67-115. Mariculture and Fisheries Dep., Food Resources Division, Kuwait In- stitute for Scientific Research, Kuwait, 200 p. Munro, J. L. (ed.). 1983. Caribbean coral reef fishery resources. ICLARM Stud. Rev. 7, 276 p. Nelson, R. S., and C. S. Manooeh. 1982. Growth and mortality of red snappers in the west- central Atlantic Ocean and the northern Gulf of Mexico. Trans. Am. Fish. Soc. 111:465—475. Newman, S. J., D. M. Williams, and G. R. Russ. 1996. Variability in the population structure of Lutjanus adetii (Castelnau, 1873) and L. quinquelineatus (Bloch, 1790) among reefs in the Central Great Barrier Reef, Australia. Fish. Bull. 94( 2 ):3 13-329. O’Gorman, R., D., H. Barwick, and C. A. Bowen. 1987. Discrepancies between ages determined from scales and otoliths from alewives from the Great Lakes. In R. C. Summerfelt and G. E. Hall (eds.), Age and growth of fish, p. 203-210. Iowa State Univ. Press, Ames, IA. Pauly, D. 1980. On the interrelationships between natural mortal- ity, growth parameters, and mean environmental tempera- ture in 175 fish stocks. J. Cons. Int. Explor. Mer 39(21:175-192. Pearson, D. E., J. E. Hightower, and J. T. H. Chan. 1991. Age, growth, and potential yield for shortbelly rock- fish, Sebastes jordani. Fish. Bull. 89(3):403-410. Ralston, S. 1987. Mortality rates of snappers and groupers. In J. J. Polovina and S. Ralston (eds ), Tropical snappers and grou- pers: biology and fisheries management, p. 375-404. Westview Press, Inc., Boulder, CO. Ramirez Rodriguez, E. M., and C. Rodriguez Medrano. 1990 Composicion especifica de la captura artesanal de peces en la Isla Cerralvo, B.C.S., Mexico. Invest. Mar. CICIMAR (Cent. Interdiscip. Cienc. Mar) 5(21:137-141. Rocha Olivares, A. 1991. Edad y crecimiento del huachinango del Pacifico Lutjanus peru (Nichols y Murphy, 1922) (Perciformes: Lutjanidae) en la Bahia de La Pazy zonas adyacentes, B.C.S., Mexico. Tesis profesional, Departamento de Biologia Ma- rina, Universidad Autonoma de Baja California Sur, 121 p. Rocha Olivares, A., and V. Gomez Munoz. 1993. Validacion del uso de otolitos para determinar la edad del huachinango del Pacifico Lutjanus peru (Perciformes: Lutjanidae), en la Bahia de La Paz y aguas adyacentes, B.C.S., Mexico. Ciencias Marinas 19(3):32 1—331. Ruiz, A., and J. Madrid. 1992. Estudio de la biologia del isopodo parasito Cymothoa exigua Schioedte y Meinert, 1884 y su relacion con el huachinango Lutjanus peru (Pisces: Lutjanidae) Nichols y Murphy, 1922, a partir de capturas comerciales en Michoacan. Ciencias Marinas 18(1): 19-34. Ruiz Luna, A., E. Giron, J. Madrid, and A. Gonzalez. 1985. Determinacion de edad, crecimiento y algunas constantes biologicas del huachinango del Pacifico, Lutjanus peru (Nichols y Murphy, 1922). Memorias VIII Congreso Nacional de Zoologia, Saltillo, Coahuila Agosto 1985:181-201. Ruiz Santos, H. 1983. Reproduccion del huachinango Lutjanus peru (Pisces: Lutjanidae) del Pacifico Sur — Mexico. Tesis profesional, Universidad Nacional Federico Villareal, 72 p. Smith, D. C., G. E. Fenton, S. G. Robertson, and S. A. Short. 1 995. Age determination and growth of orange roughy ( Hoplo - stethus atlanticus): a comparison of annulus counts with radio- metric ageing. Can. J. Fish. Aquat. Sci. 52(2):391-401. Thomson, D. A., L. T. Findley, and A. N. Kerstitch. 1 987 . Reef fishes of the Sea of Cortez. Univ. Arizona Press, Tucson, AZ, 302 p. Van Der Heiden, A. M. 1985. Taxonomia, biologia y evaluacion de la ictiofauna dem- ersal del Golfo de California. In A. Yanez-Arancibia (ed.), Recursos pesqueros potenciales de Mexico; la pesca acompanantedelcamaron.p. 149-200. Programa Universi- tario de Alimentos, Instituto de Ciencias del Mar y Lim- nologia, Instituto Nacional de la Pesca, UNAM, Mexico, D.F. Walker, B. W. 1960. The distribution and affinities of the marine fish fauna of the Gulf of California. Syst. Zool. 9( 1-4):123-133. Wilson, C. D., and G. W. Boehlert. 1990. The effects of different otolith ageing techniques on estimates of growth and mortality for the splitnose rock- fish, Sebastes diploproa , and canary rockfish, Sebastes pinniger. Calif. Fish Game 76(3 ): 146—16. Zar, J. H. 1984. Biostatistical analysis. Prentice Hall, Englewood Cliffs, NJ, 718 p. 575 Enhancing diet analyses of piscivorous fishes in the Northwest Atlantic through identification and reconstruction of original prey sizes from ingested remains Frederick S. Scharf Department of Forestry and Wildlife Management University of Massachusetts, Amherst, Massachusetts 0 1 003 Present address: Texas Parks and Wildlife Department, Coastal Fisheries Division Seabrook Marine Laboratory RO. Box 8, Seabrook, Texas 77586 E-mail address: fred.scharf@tpwd.state.tx, us Richard M. Yetter Department of Forestry and Wildlife Management University of Massachusetts, Amherst, Massachusetts 0 1 003 Adam R Summers Organlsmic and Evolutionary Biology University of Massachusetts, Amherst, Massachusetts 0 1 003 Francis Juanes Department of Forestry and Wildlife Management and Graduate Program in Organismic and Evolutionary Biology University of Massachusetts, Amherst, Massachusetts 0 1 003 Abstract.— Biological interactions among species can play a dominant role in structuring marine fish communi- ties. Specifically, predation may repre- sent a significant source of mortality for larval and juvenile fishes. Analysis of predator diet requires accurate infor- mation on the identity as well as the sizes of prey consumed. In examina- tions of stomach contents of piscivorous fishes, the condition of recovered prey items varies substantially not only in the large range of digestive states en- countered but also in the occurrence of partially consumed fishes. To estimate the original sizes of well-digested and partially consumed prey fishes we con- structed a series of predictive equations relating total length, fork length, and weight of fish to seven morphometric measurements including dorsoventral body depth, eye diameter, caudal pe- duncle depth, pectoral-fin length, opercle length, cleithrum length, and dentary length for ten common prey fishes in the Northwest Atlantic. All relationships were highly significant, with coefficients of determination typi- cally exceeding 0.90 and mean percent prediction errors less than 10%, indi- cating that reliable original size esti- mates are obtainable from incomplete fish remains. To aid in field-based iden- tification of prey fishes, we extracted and examined opercles, cleithra, and dentaries from each fish. Careful ex- amination of bones revealed prominent diagnostic characteristics with clear differences among family taxa, demon- strating their potential utility as iden- tification tools. Used collectively, the predictive equations and the diagnos- tic features of the bones should allow for inclusion in diet analyses of prey items previously designated as uniden- tifiable or unmeasurable, and thus in- crease the amount of dietary information obtainable from stomach contents analy- ses of Northwest Atlantic piscivores. Manuscript accepted 22 October 1997. Fishery Bulletin 96:575-588 (1998). Biological interactions among spe- cies can significantly affect the dy- namics of marine fish populations (Overholtz et al., 1991; Rothschild, 1991). Specifically, predation by pi- scivorous fishes is recognized as an important mechanism in structur- ing fish communities (Hackney, 1979; Lyons and Magnuson, 1987; Tonn et al., 1992), particularly in the role it plays to potentially regu- late natural mortality rates and re- cruitment of larval and juvenile fishes (Sissenwine, 1984; Houde, 1987; Bailey and Houde, 1989). In order to estimate consumption rates of key predators, details are needed on the types and sizes of prey fishes consumed across spatial and tem- poral scales. However, because of the substantial variability in the condition of recovered prey items, the collection of stomach contents data in the field is often limited to only those items that are readily identified and measured, resulting in the loss of potentially important diet information. Identification of piscine prey in- gested by fish, avian, and mamma- lian predators has frequently in- volved the use of diagnostic bones including vertebrae (Pikhu and Pikhu, 1970; Feltham and Marquiss, 1989; Carss and Elston, 1996), pha- ryngeal arches (McIntyre and Ward, 1986; Raven, 1986), opercular bones (Newsome, 1977), the axial skeleton 576 Fishery Bulletin 96(3), 1998 and hypurals (Trippel and Beamish, 1987), and cleithra and dentaries (Hansel et al., 1988). Measure- ments of the dimensions of diagnostic bones have often then been used to estimate original prey size (Trippel and Beamish, 1987; Hansel et al., 1988; Carss and Elston, 1996). Previous studies on the food habits of piscivores have used fish otoliths as an aid in prey identification and original prey size estima- tion ( Jobling and Breiby, 1986). However, acidic pre- servatives, such as formaldehyde, can dissolve otoliths, resulting in unreliable otolith length and fish length relationships (McMahon and Tash, 1979). In addition to bones, external morphological mea- surements, such as eye diameter and caudal peduncle depth, have been used successfully to reconstruct prey body size for recently consumed prey fishes (Crane et al., 1987; Scharf et al., 1997). However, the majority of fish feeding studies employing such techniques have been directed at freshwater piscivores, whereas researchers examining the diets of marine fishes have utilized these techniques much less frequently (Crane et al., 1987; Scharf et al., 1997). Moreover, no such techniques are currently available to enhance diet analyses and more com- pletely define the role of piscivorous fishes in struc- turing fish communities in the Northwest Atlantic.1 Here, we generate a series of predictive regression equations to estimate prey fish total length, fork length, and weight for ten fish species. Original fish size is estimated from four external morphometric measurements including maximum body depth, eye diameter, caudal peduncle depth, and pectoral-fin length, as well as from length measurements of three diagnostic bones: the opercle, the cleithrum, and the 1 Rountree, R. A. 1997. Food Chain Dynamics Investigation (FCDI). Northeast Fisheries Science Center (NEFSC), National Marine Fisheries Service (NMFS), 166 Water St., Woods. Hole, MA 02543. Personal commun. dentary. The prey species used here include several of commercial and recreational importance and com- monly occur in the diets of marine piscivores in the Northwest Atlantic, representing approximately 52% of the total fish prey consumed by piscivorous fishes during 1973-90 (Grosslein et al., 1980; Langton and Bowman, 1981). 2 We also identify and describe the unique characteristics of the three diagnostic bones for each fish species and assess their potential value as tools for field identification of prey fishes recov- ered from predator stomachs. Materials and methods Over 700 fish representing ten species in five fami- lies ranging in size from 52 to 340 millimeters (mm) total length (TL) were measured and dissected (Table 1). Fish were collected as part of Food Chain Dynam- ics Investigation (FCDI) (NEFSC, NMFS) research cruises conducted on Georges Bank during June and August of 1995 and 1996. FCDI research cruises were sponsored through the National Oceanic and Atmo- spheric Administration Coastal Ocean Program- Georges Bank Predation Study. Fish were also col- lected from waters off the Northeast U.S. coast dur- ing September and October of 1996 as part of the NEFSC Fall bottom trawl surveys. Additional fish were collected by personnel from the Massachusetts Division of Marine Fisheries as part of a coastal trawling survey conducted in May and September of 1996. All fish were immediately frozen until they could be returned to the laboratory. In the labora- tory, fish were thawed and weighed wet to the near- 2 Food Chain Dynamics Investigation (FCDI). Northeast Fisher- ies Science Center, NMFS, 166 Water St., Woods Hole, MA 02543. Unpublished data. Family, species, number ( n ), and size range (total length Table I in millimeters) of prey fishes used to construct predictive equations. Family Species Common name n Size range Clupeidae Alosa pseudoharengus Alewife 137 73-282 Alosa aestivalis Blueback herring 38 83-134 Clupea harengus Atlantic herring 84 97-269 Scombridae Scomber scombrus Atlantic mackerel 56 158-330 Stromateidae Peprilus triacanthus Butterfish 108 54-193 Ammodytidae Ammodytes dubius Sand lance 75 109-209 Gadidae Urophycis chuss Red hake 45 108-340 Merluccius bilinearis Silver hake 95 75-300 Melanogrammus aeglefinus Haddock 47 79-202 Gadus morhua Atlantic cod 31 52-140 Scharf et a L Diet analysis of piscivorous fishes 577 est 0.01 gram (g) before external measurements were completed to the nearest 0.01 mm with digital cali- pers. To remove diagnostic bones, fish were placed in boiling water for 30-90 seconds, depending on fish size. Bones were then extracted from the soft tissue and measured to the nearest 0.025 mm with an ocu- lar micrometer (2.75x). Least squares regression equations (StataCorp., 1995) were generated to predict original total length, fork length, and weight from measurements of body depth, eye diameter, caudal peduncle depth, pecto- ral-fin length, opercle length, cleithrum length, and dentary length. The left eye, left pectoral fin, and diagnostic bones from the left side of each fish were used consistently unless damaged. Body depth was measured as the maximum linear dorsoventral dis- tance with fins depressed. Eye diameter was mea- sured horizontally along the anteroposterior axis. Caudal peduncle depth was measured dorsoventrally. Pectoral-fin length was measured from the anterior most point of fin insertion to the tip of the longest fin ray. Opercle length was measured as the maxi- mum linear dorsoventral distance, usually from the dorsal most point to the tip of the primary ray (Fig. 1 ). Cleithrum length was measured from the tip of the dorsal process to the tip of the ventral process (Fig. 2). Dentary length was defined as the maximum linear anteroposterior distance from the symphyseal mar- gin located anteriorly between the left and right dentaries to the posterior tip of the dorsal or ventral process, whichever was longer (Fig. 3). A series of least squares regression equations to predict fish weight from total or fork length for each species was also generated. To further assess the strength of in- dividual bivariate relationships, mean percent pre- diction errors (Smith, 1980) were determined for each regression by averaging the percent prediction error calculated for each observation as [(Observed - Predicted) / Predicted ] x 100. Forward and backward stepwise linear regressions (StataCorp., 1995) were performed in an attempt to identify the best set of predictor variables. To ensure that all variables in the stepwise model were indi- vidually significant, the value of the ^-statistic used to determine variable inclusion was set at four (Draper and Smith, 1981). Opercles, cleithra, and dentaries were also care- fully examined to identify distinguishing character- istics that may be potentially useful for identifica- tion of each fish species. Several features of each bone were examined for differences among species, includ- ing the general shape of each bone; the numbers and orientation of ridges on the opercle and the curva- ture of opercle margins; the numbers, sizes, and shapes of cleithrum processes; and the presence or absence of teeth on the dentary, with attention given to overall tooth size, shape, and orientation, as well as the shapes and relative lengths of dentary processes. Resuits Predictive equations Regressions relating external morphological mea- surements to total and fork length were all highly significant (P<0.0001), with r2 values ranging from 0.63 to 0.99 and mean percent prediction errors rang- ing from 2.68 to 10.83 (Table 2). Regressions from measurements of eye diameter to predict original fish length were typically more variable than those from other external measurements. This was likely due to measurement error associated with damage of the soft adipose tissue surrounding the eye incurred dur- ing the freezing and thawing processes. Regressions relating diagnostic bone measurements to total and fork length were also highly significant (P<0.0001), with r2 values ranging from 0.81 to 0.99 and mean percent prediction errors ranging from 1.26 to 8.48 (Table 3). Compared with regressions from external morphological measurements, variation in diagnos- tic bone measurements typically explained more of the variation in original fish length (94% ofr2 values >0.90) and bones were generally more precise in pre- dicting fish length (87% of mean %PEs < 5.00). Stepwise linear regressions indicated that cleithrum length was the most consistent predictor of original fish length and it was included in the best set of pre- dictor variables for 8 of 10 species (Table 4). Pecto- ral-fin length was included in the best set of predic- tors for 6 species, whereas dentary length was in- cluded for 5 species. The remaining independent vari- ables were included in the best set of predictors for either 3 or 4 species, respectively. Regressions relating external morphological mea- surements and diagnostic bone measurements to fish weight were also each highly significant (P<0.0001 ), with r2 values ranging from 0.71 to 0.99 (Table 5). Mean percent prediction errors for regressions pre- dicting fish weight ranged from 5.97 to 39.17. These were typically higher compared with prediction er- rors for regressions predicting fish length, indicat- ing that estimates of original fish length were more precise than estimates of original prey weight. Simi- lar to length regressions, diagnostic bone measure- ments yielded higher r2 values and lower mean per- cent prediction errors when regressed against fish weight compared with external morphological mea- 578 Fishery Bulletin 96(3), 1 998 surements. Stepwise linear regressions indicated that body depth, caudal peduncle depth, and cleithrum length were the most consistent predic- tors of original fish weight and were included in the best set of predictor variables for 8, 9, and 8 species, respectively (Table 6). Pectoral-fin length was in- cluded in the best set of predictor variables for 5 spe- cies, whereas opercle length was included for 3 spe- cies. Dentary length and eye diameter appeared to be the least consistent predictors of fish weight and were included in the best set of predictor variables for 1 and 0 of the species, respectively. Lastly, mea- surements of total and fork length were significantly related to total weight for each species (P<0.0001) with Scharf et a I.: Diet analysis of piscivorous fishes 579 r2 values typically greater than 0.95 and mean percent prediction errors generally less than 10% (Table 7). Diagnostic bones For each of the bones examined here, clear differences in diagnostic features exist among family taxa. Distin- guishing bone characteristics were also evident among genera within the families Gadidae and Clupeidae. Differences between two species from the same genus (Alosa ) were, however, difficult to discern from the three bones used here. Therefore, bone illustrations are pre- sented for only nine species, with alewife, rather than blueback herring, representing the genus Alosa. Opercles The opercle is the largest and most dorsal bone in the opercular series, which consists of the opercle, the subopercle, and the interopercle. Together, these three bones pro- vide the skeletal support for the muscular operculum, or gill cover in fishes. The opercle articulates with the opercular process of the hyomandibula. The articulation site is located in the anterodorsal region of the opercle, and repre- sents a consistent morphological feature for orientation of the bone during examination. The opercles of the fishes examined here can be differentiated by two major diag- nostic characteristics. First, the general shape of the opercle is clearly unique to several of the families examined. The opercles of some families share a general tri- angular shape with three well-de- fined margins, whereas others are not triangular and possess four definable margins. A second dis- tinctive feature is the presence or absence of ridges originating at the site of hyomandibular articulation and extending ventrally or posteri- orly along the margins of the opercle. Of the taxa examined here, sil- ver hake ( Merluccius bilinearis ) and red hake ( Urophycis chuss ) have opercles with the most pro- nounced triangular shape and three clearly defined margins (Fig. 1, A and B ). The hake opercles are further distinguished by the pres- ence of two prominent ridges origi- nating at the site of hyomandibu- lar articulation and extending ven- trally and posteriorly along the anterior and dorsal margins, re- spectively. Hake opercles differ 580 Fishery Bulletin 96(3), 1 998 slightly because the posterior region of the red hake opercle is longer in relation to the body of the opercle than that of silver hake and is also narrower, ending in a sharper point in contrast with the opercle of sil- ver hake. The opercles of Atlantic cod ( Gadus morhua ) and haddock (. Melanogrammus aeglefinus) share a general triangular shape with the hakes, although it is less pronounced in cod and haddock (Fig. 1, C and D). Opercles of Atlantic cod and had- dock can be further separated from hake opercles by the presence of only one prominent ridge, which ex- tends posteriorly from the site of hyomandibular at- tachment, rather than the two ridges found in hake. There appear to be no consistent, observable differ- ences between the opercles of Atlantic cod and had- dock. The opercles of alewife ( Alosa pseudoharengus) and Atlantic herring ( Clupea harengus ) are not tri- angular and have a distinct notch that is dorsal to the site of hyomandibular articulation along the dor- sal margin, with a pronounced curvature along the posterior margin (Fig. 1, E and F). The opercles of alewife and Atlantic herring each possess one ridge, extending ventrally from the site of hyomandibular articulation along the anterior margin, and each also has four definable margins. Alewife and Atlantic herring opercles differ mainly in the shape of the dorsal notch — the notch being cup-shaped in alewife, compared with v-shaped in Atlantic herring. In At- lantic mackerel (Scomber scombrus), the opercle has a pronounced dorsal curvature with a notch located centrally along the posterior margin (Fig. 1G). The ridge extending from the hyomandibular articulation site is unique to Atlantic mackerel in that it branches out broadly as it extends posteriorly from the articu- lation site. The opercles of butterfish ( Peprilus triacanthus) are unique in general shape, resembling a butterfly wing, with four clearly defined margins and a ridge extending ventrally from the hyomandib- ular articulation site along the anterior margin (Fig. 1H). Sand lance (Ammodytes dubius ) represent the only species examined here with opercles not possess- ing ridges along any margins. Sand lance opercles also Scharf et al.: Diet analysis of piscivorous fishes 581 Table 2 Least squares regression equations relating measurements of body depth (BD), eye diameter (ED), caudal peduncle depth (CP), and pectoral-fin length (PF) to total length (TL) and fork length (FL) for ten prey species in the Northwest Atlantic. All measure- ments are in millimeters. sb = standard error of the regression coefficient; r2 = coefficient of determination; %PE = mean percent prediction error; n = number of fish measured; Range = size range in total length. Fork lengths were not measured for red hake, silver hake, or Atlantic cod. Total length Fork length Species Equation sb r2 %PE Equation S6 r2 %PE n Range Alewife TL = 3MIBD + 11.012 0.073 0.95 4.38 FL = 3.469BD + 10.856 0.067 0.98 4.04 137 73-282 TL = 24.906 ED - 61.344 0.781 0.88 7.30 FL = 23.335 ED - 57.728 0.562 0.96 6.13 137 73-282 TL = 12.876 CP- 1.784 0.182 0.97 3.53 FL = 11.348CP + 1.313 0.163 0.99 3.31 137 73-282 TL = 6.840 PF + 0.532 0.100 0.99 3.15 FL = 5.982 PF + 1.785 0.093 0.98 3.34 66 73-282 Blueback herring TL = 4.97 OBD + 1.439 0.268 0.91 3.59 FL = 4.620BB - 1.573 0.246 0.91 3.66 38 83-134 TL = 25.318-ED - 35.605 2.117 0.80 5.33 FL = 23.792 ED - 37.352 1.865 0.82 5.42 38 83-134 TL= 13.684CP + 0.194 0.752 0.90 3.53 FL = 12.585CP- 1.782 0.754 0.89 4.06 38 83-134 TL = 7.760 PF + 2.935 0.383 0.92 3.58 FL = 7.198PF- 0.005 0.358 0.92 3.79 38 83-134 Atlantic herring TL = 4.561BD + 27.341 0.089 0.97 4.22 FL = 4.036 BD + 26.880 0.081 0.97 4.17 84 97-295 TL = 26.717BD - 51.235 0.594 0.96 5.08 FL = 23.633 ED - 42.586 0.537 0.96 5.00 84 97-295 TL = 16.303CP - 17.288 0.296 0.97 4.55 FL = 14.441CP- 12.792 0.257 0.97 4.35 84 97-295 TL = 7.507PF + 1.533 0.101 0.99 2.68 FL = 6.631 PF + 4.306 0.102 0.98 2.95 84 97-295 Atlantic mackerel TL = 5.203 BD + 37.203 0.122 0.97 3.67 FL = 4.641 BD + 39.790 0.118 0.97 3.89 56 158-330 TL = 33.706 ED - 81.239 3.027 0.70 10.83 FL = 30.404BZ) - 68.847 2.652 0.71 10.28 56 158-330 TL = 4Q.081CP - 0.837 0.700 0.98 2.78 FL = 35.874CP + 5.213 0.592 0.99 2.61 56 158-330 TL = 9.037 PF- 1.242 0.178 0.98 3.02 FL = 8.068 PF + 5.346 0.172 0.98 3.06 56 158-330 Butterfish TL = 2.880 BD + 1.650 0.038 0.98 3.32 FL = 2.384 BD + 5.638 0.034 0.98 3.57 108 54-193 TL = 2Q.065FD - 21.148 0.635 0.90 7.48 FL = 16.700ED - 13.166 0.544 0.90 7.52 108 54-193 TL = 17.96QCP+ 7.962 0.290 0.97 3.57 FL = 14.950CP + 11.046 0.264 0.97 4.03 108 54-193 TL = 3.76 6PF + 15.966 0.060 0.98 4.33 FL = 3.126PF+ 18.121 0.064 0.98 5.54 63 54-193 Sand lance TL = 7.308 BD + 58.774 0.455 0.78 6.07 FL = 7.0453D + 57.497 0.463 0.76 6.16 75 109-209 TL = 49.467FD - 45.022 4.358 0.64 8.29 FL = 47.921 ED - 43.479 4.313 0.63 8.45 75 109-209 TL = 35.928CP + 8.185 1.562 0.88 4.32 FL = 34.923 CP + 7.599 1.575 0.87 4.48 75 109-209 TL = 14.993PF- 31.272 0.934 0.78 6.52 FL = 14.542PF- 30.374 0.933 0.77 6.75 75 109-209 Red hake TL = 6.398BD + 10.340 0.210 0.96 7.40 45 108-340 TL = 28AUED - 91.519 0.880 0.96 6.37 45 108-340 TL = 26.384CP- 17.165 0.627 0.98 5.18 45 108-340 TL = 6.885PF - 0.693 0.146 0.98 3.92 45 108-340 Silver hake TL = 6.48GBD + 18.219 0.164 0.94 7.00 95 75-300 TL = 21.998SD - 31.654 0.504 0.95 7.05 95 75-300 TL = 22.823 CP + 22.881 0.345 0.98 4.63 95 75-300 TL = 4.946PF + 20.269 0.070 0.99 3.36 60 75-300 Haddock TL = 4.199M) + 21.964 0.173 0.93 5.30 FL = 4.066 BD + 19.660 0.164 0.93 4.83 47 79-202 TL = 20.069 ED - 35.473 1.375 0.83 7.97 FL = 19.525 ED - 36.724 1.291 0.84 7.87 47 79-202 TL = 18.465CP + 18.596 0.699 0.94 4.90 FL = 17.922CP + 14.205 0.638 0.95 4.54 47 79-202 TL = 7.367 PF- 10.712 0.260 0.95 4.69 II 0 ey> 3 1 CO O 0.263 0.94 4.90 47 79-202 Atlantic cod TL = 4.249BD + 21.367 0.133 0.97 4.03 31 52-140 TL = 19.393 ED - 19.490 1.053 0.92 7.33 31 52-140 TL = 17.999CP + 9.806 0.480 0.98 3.72 31 52-140 TL = 6.728PF + 7.681 0.185 0.98 3.48 31 52-140 possess only two clearly definable margins and are curved along the entire posterior margin (Fig. II). Cfeithra In higher teleost fishes, the cleithrum is the ventral most bone of the pectoral girdle, which consists of the supracleithrum, the cleithrum, and the postclei- thrum. The cleithrum attaches directly to the scap- ulocoracoid, which articulates with the base of the pectoral fin. Together, the left and right cleithra form the frame of the body wall directly posterior to the branchial cavity. The cleithra of the fishes examined here can be differentiated by their general shape as 582 Fishery Bulletin 96(3), 1998 well as the numbers, locations, and shapes of the processes. The majority of the cleithra described here have a gently sloping curvature along the dorsoven- tral axis, whereas others are clearly more sharply curved. Most families examined here possess cleithra with dorsal and ventral processes that end sharply; others possess varying numbers and shapes of pos- terior and anterior processes. The cleithra of silver hake and red hake are gen- tly curved along the dorsoventral axis and each has two posterior processes located directly ventral to the dorsal process (Fig. 2, A and B). The ventral process of the red hake ends in a sharper point and is not as broad as that of silver hake, and the posterior pro- cesses are broader than those of silver hake. Atlan- tic cod and haddock have gently curved cleithra with a general shape similar to that of the cleithra of sil- ver and red hake (Fig. 2, C and D). However, the cleithra of Atlantic cod and haddock contain only one posterior process that is broader than the posterior processes of the hake cleithra, and the dorsal pro- cesses of Atlantic cod and haddock cleithra end sharply and are much more pronounced than those of the hake cleithra. The main difference between the cleithra of Atlantic cod and haddock is in the shape of the posterior process; the process is distinctly more hook-shaped in Atlantic cod than in haddock. The cleithra of alewife and Atlantic herring are each sharply curved and possess an anterior process lo- cated medially along the dorsoventral axis, which is Table 3 Least squares regression equations relating measurements of opercle length (OP), cleithrum length (CL), and dentary length (DN) to total length (TL) and fork length (FL) for ten prey species in the Northwest Atlantic. All measurements are in millime- ters. sb = standard error of the regression coefficient; r2 = coefficient of determination; %PE = mean percent prediction error; n = number of fish measured; Range = size range in total length. Fork lengths were not measured for red hake, silver hake, or Atlantic cod. Total length Fork length Species Equation S6 r2 °lc PE Equation sb r2 %PE n Range Alewife TL = 9.8510P- 1.119 0.126 0.99 2.80 FL = 8.6170P + 0.307 0.118 0.99 3.15 66 73-282 TL = 7.115CL - 3.450 0.097 0.99 2.80 FL = 6.217CL - 1.579 0.097 0.98 3.22 66 73-282 TL = 11.610DJV- 23.205 0.163 0.99 3.23 FL = 10.144LW - 18.824 0.163 0.98 3.74 66 73-282 Blueback herring TL= 10.0140P + 6.641 0.547 0.90 3.07 FL = 9.1980P + 4.259 0.555 0.88 3.46 38 83-134 TL = 7.491CL + 0.161 0.354 0.93 3.09 FL = 6.893 CL - 1.863 0.362 0.91 3.40 38 83-134 TL = 11.743LW- 11.125 0.653 0.90 2.96 FL= 10.87 1DN- 12.840 0.621 0.90 3.04 38 83-134 Atlantic herring TL = 11.732OP-0.427 0.168 0.99 2.30 FL = 10.4050P + 2.028 0.168 0.98 2.65 71 97-269 TL = 8.752 CL - 10.804 0.126 0.99 2.19 FL = 7.771CL - 7.353 0.117 0.98 2.41 71 97-269 TL = 10.692LW- 17.263 0.180 0.98 3.02 FL = 9.495LW - 13.112 0.164 0.98 3.10 71 97-269 Atlantic mackerel TL= 1 1.1 140P + 0.379 0.179 0.99 2.01 FL = 9.867 OP + 7.462 0.163 0.99 2.06 44 158-330 TL = 7.789 CL - 6.990 0.067 0.99 1.26 FL = 6.916CL + 0.889 0.062 0.99 1.27 44 158-330 TL = 12.564LW - 30.649 0.212 0.99 2.53 FL = 11.148DA - 19.970 0.201 0.99 2.44 44 158-330 Butterfish TL = 7.9560P + 5.184 0.299 0.92 7.64 FL = 6.5980P + 9.279 0.266 0.91 8.48 63 54-193 TL = 5.583 CL - 27.802 0.091 0.98 3.23 FL = 4.653 CL - 18.698 0.081 0.98 3.43 63 54-193 TL = 13.643ZW + 6.884 0.484 0.93 7.49 FL = 11.309LW + 10.725 0.436 0.92 8.25 63 54-193 Sand lance TL = 18.903 OP + 25.655 1.042 0.82 5.46 FL = 18.352 OP + 24.729 1.044 0.81 5.60 75 109-209 TL = 18.487CL- 1.156 0.719 0.90 3.81 FL = 17.938CL - 1.228 0.741 0.89 4.12 75 109-209 TL = 2Q.785ZW- 33.880 0.772 0.91 3.89 FL = 20.300LW - 34.146 0.752 0.91 3.92 75 109-209 Red hake TL = 27.956 OP - 4.679 0.612 0.98 4.53 45 108-340 TL = 9.061CL- 12.185 0.123 0.99 2.81 45 108-340 TL = 13.870ZW - 11.769 0.291 0.98 4.37 45 108-340 Silver hake TL = 17.6660P- 11.964 0.212 0.99 3.20 60 75-300 TL = 7.87501-5.271 0.071 0.99 2.57 60 75-300 TL = 8.427LW - 5.149 0.087 0.99 2.63 60 75-300 Haddock TL = 23.0400P- 15.425 1.053 0.91 5.52 FL = 22.3510P- 16.804 0.984 0.92 4.78 47 79-202 TL = 6.567 CL - 2.688 0.176 0.97 3.25 FL = 6.2150,-1.202 0.258 0.93 4.87 47 79-202 TL = 12.348PW - 3.688 0.496 0.93 5.07 FL= 11.680OV- 2.075 0.606 0.89 6.29 47 79-202 Atlantic cod TL = 19.019OP + 7.733 0.705 0.96 5.43 31 52-140 TL = 6.590CL + 1.118 0.101 0.99 2.50 31 52-140 TL = 10.458LW - 5.592 0.181 0.99 2.29 31 52-140 Scharf et al.: Diet analysis of piscivorous fishes 583 unique among the fishes examined here (Fig. 2, E and F). The shape of the anterior process differs be- tween the two genera of Clupeidae. In alewife, the anterior process is sickle-shaped, whereas in Atlan- tic herring, the anterior process is more symmetri- cal, resembling a mushroom. Additionally, the tip of the ventral process is distinctly sharper in the cleithrum of alewife than in the cleithrum of Atlan- tic herring. Atlantic mackerel cleithra are gently curved and relatively thin and one posterior process is located in the extreme dorsal region of the bone (Fig. 2G). The dorsal process of the Atlantic mack- erel cleithrum is relatively small and rounded, whereas the ventral process is broad and does not end sharply as in the cleithra of the gadids. The cleithra of butterfish are only slightly curved along the dorsoventral axis and contain a broad, rounded posterior process. Butterfish cleithra can be distin- guished among the families examined here by the presence of a distinctly sharp, thin dorsal process (Fig. 2H). The cleithra of sand lance have a general Table 4 Independent variables included in stepwise multiple re- gression models estimating original fish length. Abbrevia- tions for independent variables are those given in Tables 2 and 3. * indicates that forward and backward models were not identical. Variables Variables included included in forward in backward Species stepwise model stepwise model Alewife CP, PF, OP CL CP, PF, OP, CL Blueback herring PP CL PF, CL Atlantic herring BD, PP OP, DN BD, PF, OP, DN Atlantic mackerel* CL, DN CL Butterfish* BD, CP, PF, CL ED, CP, PF, CL Sand Lance CP, DN CP, DN Red hake BD, ED, CL BD, ED, CL Silver hake* OP, CL, DN CL Haddock ED, CP, PF, CL ED, CP, PF, CL Atlantic cod BD, PF, CL, DN BD, PF, CL, DN Table 5 Least squares regression equations relating measurements (in millimeters) of body depth (BD), eye diameter (ED), caudal pe- duncle depth (CP), pectoral fin length (PF), opercle length (OP), cleithrum length (CL), and dentary length (DN) to total weight (W) in grams for ten prey species in the Northwest Atlantic. sb = standard error of the regression coefficient; r 2 = coefficient of determination; %PE = mean percent prediction error; n = number of fish measured; Range = size range in total length. Species Equation sb r2 %PE n Range Alewife W = 0.73 xlO-3BD2 940 0.041 0.97 10.74 137 73-282 W= 2.62x10 -3£D4 305 0.133 0.89 24.00 137 73-282 W= 10.18 xlO“3CP3 194 0.037 0.98 8.87 137 73-282 W= 1.25 xlO-W5 229 0.053 0.98 11.31 66 73-282 W = 2.95 xlO-3OP3 331 0.052 0.98 11.73 66 73-282 W = 0.81 xlO_3CL3 383 0.051 0.99 10.65 66 73-282 W= 1.05x10 -3ZW3-739 0.055 0.99 10.66 66 73-282 Blueback herring W = 0.64 xl0~3BD3 066 0.103 0.96 6.83 38 83-134 W = 6.68 xlO_3EZ)4 049 0.340 0.80 15.73 38 83-134 W = 13.80 xlO-3CP3 061 0.154 0.92 9.48 38 83-134 W = 3.94 xlO-W2 903 0.158 0.90 9.88 38 83-134 W = 10.40 xlO -3OP2 851 0.160 0.90 9.84 38 83-134 W= 2.38 xl0“3CL3 027 0.139 0.93 8.15 38 83-134 W = 3.00 xlO -3DN3 383 0.201 0.89 9.96 38 83-134 Atlantic herring W = 2.96 xlO"3#/)2 734 0.033 0.99 8.64 84 97-295 W= 7.14xlO-3£D4012 0.092 0.96 16.58 84 97-295 W = 7.96 xlO“3CP3 438 0.060 0.98 11.81 84 97-295 W = 3.76 xlO -3P£2 935 0.065 0.97 12.06 84 97-295 W = 10.53 xlO -3OP3 030 0.063 0.97 10.01 71 97-295 W = 2.20 xl0_3CL3 193 0.069 0.97 10.01 71 97-295 W = 3.52 xl0-3Z)AF 210 0.084 0.96 12.87 71 97-295 Atlantic mackerel W = 2.79 xl0_3BZ)2 884 0.066 0.97 11.73 56 158-330 W = 2.20 xlO-3£Z)4-762 0.377 0.75 39.17 56 158-330 W = 247.52 xl0_3CP3 373 0.061 0.98 9.12 56 158-330 W = 1.49 xlO-W3 399 0.078 0.97 12.07 56 158-330 W = 3.90 xlO -3OP3 318 0.049 0.99 6.77 44 158-330 continued 584 Fishery Bulletin 96(3), 1998 butterfly shape that is clearly unique among the families examined here (Fig. 21). Dentaries The dentary is the largest bone of the lower jaw and bears teeth in many fishes. The posterior region of the dentary is attached directly to the angular and articular bones. The left and right dentaries are fused anteriorly at the mandibular symphysis. The dentaries of the fishes examined here can be differ- entiated by the presence or absence of teeth and by the relative lengths of the body of the dentary and the dorsal and ventral processes. The dentaries of silver hake and red hake each possess a row of separate conical teeth extending slightly onto the dorsal process. Teeth are curved inwards and spaced relatively far apart (Fig. 3, A and B). The ratios of the lengths of the dorsal and ventral processes to the length of the body of the Table 5 (continued) Species Equation sb r2 %PE n Range Atlantic mackerel, W = 0.76 xlCHCL3 419 0.048 0.99 5.97 44 158-330 continued W = 0.99 xlO^DN3 760 0.093 0.98 10.79 44 158-330 Butterfish W= 0.18xl0-3RD3138 0.026 0.99 7.18 108 54-193 W = 7.05 xlO~3ED4 076 0.113 0.92 24.59 108 54-193 W = 94.11 xlO_3CP2 977 0.038 0.98 10.79 108 54-193 W = 2.14 xlO-W2 795 0.039 0.99 10.55 63 54-193 W = 3.94 xlO-3OP3 232 0.098 0.95 21.39 63 54-193 W = 0.06 xlO_3CL3 905 0.056 0.99 10.83 63 54-193 W = 25.57 xlO -3Z1A3 197 0.085 0.96 19.67 63 54-193 Sand lance W = 35.99 xl0~3BD2 245 0.063 0.95 10.12 75 109-209 W = 11.58 xl( r3ED4 925 0.372 0.71 20.83 75 109-209 IV = 95.52 xlO“3CP3 400 0.105 0.94 9.35 75 109-209 IV = 0.19xl0-3PP4 366 0.274 0.78 20.26 75 109-209 W = 34.61 xl0_3OP3 017 0.177 0.80 18.38 75 109-209 W = 5.07 xlO~3CL3 618 0.111 0.94 9.84 75 109-209 IV = 0.93 xlO~3£W4 258 0.242 0.81 17.08 75 109-209 Red hake IV = 2.82 xlO-3BD2 865 0.129 0.92 23.72 45 108-340 W = 1.26 xlO-3PZ>4 492 0.153 0.95 19.21 45 108-340 IV = 44.67 xl0-3CP3 286 0.091 0.97 15.93 45 108-340 W= 1.23x10 -3PP3127 0.064 0.98 11.85 45 108-340 W= 101.46 xlO~3OP3 073 0.077 0.97 13.48 45 108-340 W= 1.58x10 -3CL3'261 0.042 0.99 7.34 45 108-340 IV = 7.61 xlO^DN3193 0.080 0.97 13.75 45 108-340 Silver hake W = 2.39 xlO-3PP2 971 0.053 0.97 18.47 95 75-300 W = 3.11 xlO“3P£)4 081 0.114 0.93 30.13 95 75-300 IV = 211.32 xl0_3CP2 647 0.037 0.98 13.88 95 75-300 W = 2.47 xlO'W2 753 0.035 0.99 11.11 95 75-300 W= 8.67 xlO“3OP3 458 0.044 0.99 11.54 95 75-300 W = 0.84 xlO-3CL3-354 0.045 0.99 12.40 95 75-300 W= 1.10x10 ^DN3 343 0.043 0.99 11.61 95 75-300 Haddock W = 2.60 xl0-3BZ)2 736 0.079 0.96 11.69 47 79-202 W = 3.32 xlO-3PD4 078 0.289 0.82 25.98 47 79-202 W = 103.33 xlO“3CP2 865 0.100 0.95 15.08 47 79-202 IV = 0.57 xlO_3PP3 523 0.123 0.95 14.97 47 79-202 W = 23.77 xl0_3OP3 609 0.162 0.92 17.41 47 79-202 IV = 0.92 xlO~3CL3 296 0.085 0.97 11.11 47 79-202 IV = 7.16 xlO^PA3 298 0.140 0.92 17.77 47 79-202 Atlantic cod IV = 4.89 xlO-3BD2 558 0.062 0.98 12.09 31 52-140 W = 5.05 xl0~3ED4 042 0.233 0.91 25.68 31 52-140 W = 67.74 xlO-3CP2-984 0.075 0.98 11.67 31 52-140 IV = 3.30 xlO_3PP2 987 0.078 0.98 12.21 31 52-140 W = 73.85 xlO -3OP2 983 0.119 0.96 16.51 31 52-140 IV = 1.15 xl0-3CL3 273 0.052 0.99 6.96 31 52-140 W = 2.48 xlO^DN3 498 0.071 0.99 9.39 31 52-140 Scharf et al.: Diet analysis of piscivorous fishes 585 dentary of silver hake are much smaller relative to the same ratios for the red hake dentary. In addi- tion, the teeth of red hake are much less prominent, and the ventral process of the dentary of red hake is considerably broader than that of silver hake. Simi- lar to the dentaries of the hakes, the dentaries of Atlantic cod and haddock also possess a row of sepa- rate conical teeth that extend slightly farther onto the dorsal process (Fig. 3, C and D). The length of the dorsal process of the dentaries of Atlantic cod and haddock is much smaller than the length of the Table 6 Independent variables included in stepwise multiple re- gression models estimating original fish weight. Abbrevia- tions for independent variables are those given in Tables 2 and 3. * indicates that forward and backward models were not identical. Variables Variables included in included in forward backward Species stepwise model stepwise model Alewife BD, CP, CL BD, CP, CL Blueback herring BD, CP, PF, CL BD, CP, PF, CL Atlantic herring BD, CP, OP BD, CP, OP Atlantic mackerel OP, CL OP, CL Butterfish BD, CP, PF, CL BD, CP, PF, CL Sand Lance BD, CP, CL, DN BD, CP, CL, DN Red hake CP, PF, CL CP, PF, CL Silver hake* BD, CP, PF, OP BD, ED, CP, PF, OP, DN Haddock BD, CP, CL BD, CP, CL Atlantic cod BD, CP, PF, CL BD, CP, PF, CL ventral process, which contrasts with the dentaries of the hakes. The dentaries of Atlantic cod and had- dock can be distiguished by the teeth located just posterior to the mandibular symphysis; the teeth are longer in haddock than in Atlantic cod. Further, the ventral process is considerably broader in haddock than in Atlantic cod. The dentaries of alewife and Atlantic herring each consist of a long, slender ven- tral process that is considerably longer than the broadly expanded dorsal process, and are toothless (Fig. 3, E and F). The only noticeable difference be- tween the dentaries of the alewife and Atlantic her- ring is a more gradual incline along the dorsal mar- gin of the dorsal process and a slight hump located anteriorly in Atlantic herring. Similar to the dentaries of the gadids, Atlantic mackerel dentaries possess a row of separate conical teeth (Fig. 3G). However, the teeth of Atlantic mackerel extend well onto the dorsal process, are spaced relatively uni- form distances apart, and are not curved inwards. The dentaries of butterfish possess a continuous row of separate teeth that are squared off at the tips, which is a unique feature among the fishes exam- ined here (Fig. 3H). The ventral process of the but- terfish dentary extends from the body of the dentary at a considerable angle. In sand lance, the dentaries have a long, thin ventral process that ends sharply and is much longer than the dorsal process, and are toothless (Fig. 31). The dorsal process of the dentary of sand lance is curved along its margin, similar to the dorsal processes of the dentaries of alewife and Atlantic herring. However, the dorsal process of the dentary of sand lance is much less broad than that found in the dentaries of alewife and Atlantic herring. Table 7 Least squares regression equations relating measurements (in millimeters) of total length (TL) and fork length (FL) to total weight (W) in grams for ten prey species in the Northwest Atlantic. sfc = standard error of the regression coefficient; r2 = coeffi- cient of determination; %PE = mean percent prediction error; n = number of fish measured; Range = size range in total length. Fork lengths were not measured for red hake, silver hake, or Atlantic cod. Total length Fork length Species Equation sb r2 %PE Equation h r2 %PE n Range Alewife W= 2.74 xlO^TL3 213 0.031 0.99 7.12 W = 3.29xlQ-6FL3-254 0.035 0.99 7.82 137 73-282 Blueback herring W = 6.24 xlO-6TL2 "3 0.101 0.96 6.78 W = 13.78 xlO_6FL2'887 0.113 0.95 7.16 38 83-134 Atlantic herring W = 4.10 xHHTL3 111 0.037 0.99 8.10 W = 4.53 xlO_6FL3-156 0.040 0.99 8.33 84 97-295 Atlantic mackerel W= 1.09x10 -6TL3-354 0.039 0.99 5.86 W = 0.85 xlO_6FL3-451 0.041 0.99 5.84 56 158-330 Butterfish W= 8.14 xlO^TL3 094 0.024 0.99 6.78 W = 8.43 xlO_6FL3180 0.032 0.99 8.36 108 54-193 Sand lance W = 0.32 xlO-6TL3-449 0.112 0.93 12.74 W = 0.41 xlO_6FL3 421 0.119 0.92 13.63 75 109-209 Red hake W = 4.41 xlO^TL3 053 0.034 0.99 5.75 45 108-340 Silver hake W = 1.91 xlO^TL3213 0.027 0.99 8.96 95 75-300 Haddock W = 3.44 xlO^TL3186 0.063 0.98 7.50 W = 3.78 xlOAFL3 196 0.080 0.97 9.12 47 79-202 Atlantic cod W = 1.99 xlO^TL3-308 0.090 0.99 5.53 31 52-140 586 Fishery Bulletin 96(3), 1 998 Discussion Each of the morphometric measurements, including those from diagnostic bones, were significantly re- lated to total length, fork length, and weight. Mea- surements taken from diagnostic bones, especially cleithrum length, appear to be highly reliable pre- dictors of original size for prey fish species in the Northwest Atlantic. Internal hard parts of fishes have historically proven accurate for estimating original sizes of prey fishes recovered from the stomachs of several freshwater piscivores (Knight et al., 1984; Trippel and Beamish, 1987; Hansel et al., 1988). Our results suggest that examination of diagnostic bones should also benefit diet analyses of marine piscivores. The ability to estimate original length and weight of common prey fishes from a suite of morphometric measurements should result in a considerable in- crease in the amount of size-specific diet informa- tion for several predatory fishes in the Northwest Atlantic. Accurate information on the sizes of prey consumed across temporal and spatial scales is criti- cal in order to calculate predator consumption rates and to determine predator impact on prey popula- tions. Moreover, many piscivorous fishes feed selec- tively on specific sizes of prey (Juanes, 1994), sug- gesting that predatory impact may be greatest on a small range of prey sizes. Therefore, knowledge of prey sizes consumed is necessary in order to deter- mine the extent of size-selective feeding patterns in Northwest Atlantic piscivores and their role in struc- turing marine fish populations. Finally, several au- thors have demonstrated the importance of body size in determining the outcome of predator-prey inter- actions among fishes (Werner and Gilliam, 1984, Miller et al., 1988, Stein et al., 1988). More complete information on predator-prey size relationships in the Northwest Atlantic may improve our understanding of the effects of body size on prey vulnerabilities to predation during various life history stages. Clearly, the estimation of original prey sizes from diagnostic bone dimensions is not as susceptible to measurement error as estimates from external mor- phological features, where measurements are usu- ally associated with soft tissues. Moreover, external morphology can often be altered during digestive processes, which may cause some or all external fea- tures to yield unreliable measurements. However, for recently consumed prey fishes, external morphologi- cal measurements may be highly reliable estimators of original prey size and represent a suitable alter- native to diagnostic bones. For example, the linear relations of both eye diameter and caudal peduncle depth with total length were found to be reliable for reconstructing original prey lengths of six prey spe- cies consumed by juvenile bluefish (Scharf et al., 1997). Similarly, Serafy et al. (1996) found the lin- ear relation between eye diameter and total length of red drum to be particularly useful in estimating the size of angled fish. In addition to cleithrum length, maximum body depth and caudal peduncle depth were consistent predictor variables for esti- mating original weight of prey fishes examined in this paper. Differences in fish condition factor may not be closely correlated with differences in other morphological measurements. Therefore, indices of fish girth, such as body depth and caudal peduncle depth, should be reliable indicators of condition factor and should be most closely related to body weight. Further, the reconstruction of original prey sizes from external morphological measurements has the advantage of being comparatively less labor in- tensive than dissection and removal of internal hard parts. The reconstruction of original prey sizes from di- agnostic bones or external morphological measure- ments does, however, have some important limita- tions. The potential effects of preservatives on bone dimensions need to be considered if stomach contents are stored in a chemical preservative. The fish used in this study were not exposed to preservatives but were frozen and thawed before measured. The use of boiling water to aid in the removal of soft tissue may cause shrinkage and deformation of extracted bones if sufficient time elapses between the boiling process and the completion of bone measurements. However, bone measurements for our analyses were completed immediately after dissection and removal. In addi- tion to the potential effects of preservation, previous studies indicate that estimation of prey sizes from partial remains will likely underestimate the pro- portion of small prey fish in the diet, because bones and external features of larger prey fish should be more resistant to digestion and be recovered from piscivore stomachs more frequently (Trippel and Beamish, 1987; Hansel et al., 1988). Lastly, allomet- ric relationships may not remain constant with changes in fish size. Therefore, regression equations should not be applied to fish outside the size range used for their generation. The results of our study indicate that, for the fami- lies of fishes examined, the opercle, cleithrum, and dentary represent diagnostic tools that can poten- tially aid in the identification of prey fishes recov- ered from the stomachs of Northwest Atlantic piscivores. The diagnostic features unique to each of the three bones appear to perform equally well in distinguishing family taxa from one another. How- ever, within the families Gadidae and Clupeidae, cleithra and dentaries are better suited to illustrate Scharf et a I.: Diet analysis of piscivorous fishes 587 differences between genera relative to opercles. Newsome (1977) and Hansel et al. (1988) revealed similar findings, demonstrating the limitations of opercles for identifying prey fishes beyond the fam- ily level. Moreover, Hansel et al. (1988) found that cleithra and dentaries were more resistant to diges- tion than opercles or pharyngeal arches, when ex- amining several freshwater fishes, and concluded that, for young prey fish, the cleithrum may be the most reliable bone because of its large size and early development in relation to other diagnostic bones. None of the three diagnostic bones appear useful for differentiating between the two species of the genus Alosa examined in this study. Similarly, Hansel et al. (1988) found that for certain congeneric fresh- water fishes, identification was inconsistent and not dependable when opercles, cleithra, dentaries, and pharyngeal arches were used. The size range of fishes used in this study included juveniles and young adults (Table 1) and is typical of size ranges exam- ined in previous studies (Trippel and Beamish, 1987; Hansel et al., 1988). Therefore, for these life stages, opercles, cleithra, and dentaries may be potentially useful as tools for identifying prey fishes to the fam- ily and, in some cases, the generic taxonomic level, but likely are not adequate for distinguishing be- tween species of the same genus. Our descriptions of the diagnostic bones provide a simple means to distinguish prey fishes based on general bone structure. Indeed, there have been more rigorous efforts directed at identifying the system- atic relationships among several of the fishes exam- ined in this paper. However, the bone descriptions in this paper were provided as a simple guide for field identification of prey fishes commonly recovered from piscivore stomachs, and thus centered on gen- eral overall shape and features that were common to all of the fishes. The differences among the taxa that are outlined here should be readily observable to a large range of workers with varying levels of fisheries training. Therefore, we suggest that the level of analysis and description presented here may compliment existing information and may be ideally suited for identification of prey fishes in field settings. The bone descriptions and regressions presented in this paper clearly have the potential to increase the amount of dietary information obtainable from stomach contents analyses of Northwest Atlantic piscivores. Further, the generation of a series of re- gression equations relating measurements from both internal hard parts and external morphological fea- tures to original fish size may provide a means to cross reference prey size estimates, thus facilitating the identification of erroneous estimates. Acknowledgments We thank the scientists at the National Marine Fish- eries Service (Northeast Fisheries Science Center in Woods Hole, MA) for fish collections and allowing us to participate in research cruises to collect samples. We are especially grateful to Arnie Howe and the Massachusetts Division of Marine Fisheries for col- lecting numerous samples. We thank Brian Hanra- han for providing considerable assistance with fig- ure graphics and William Bemis for assistance with osteological literature. We are grateful to John Boreman, Karsten Hartel, and Rodney Rountree for comments that improved the manuscript. This study was supported by the Cooperative Marine Education and Research Program of the National Oceanic and Atmospheric Administration and the Five College Coastal and Marine Sciences Program. Literature cited Bailey, K. M., and E. D. Houde. 1989. Predation on eggs and larvae of marine fishes and the recruitment problem. Adv. Mar. Biol. 25:1-83. Carss, D. N., and D. A. Elston. 1996. Errors associated with otter Lutra lutra faecal analy- sis. II. Estimating prey size distribution from bones recov- ered in spraints. J. Zool. (Lond.) 238:319-322. Crane, S. A., J. M. Fenaughty, and R. W. Gauldie. 1987. The relationship between eye diameter and fork length in the spiny oreo dory, Allocyttus sp. N.Z. J. Mar. Freshwater Res. 21:641-642. Draper, N. R., and H. Smith. 1981. Applied regression analysis, 2nd. edition. John Wiley & Sons, New York, NY. Feltham, M. J., and M. Marquiss. 1989. The use of first vertebrae in separating, and estimat- ing the size of, trout ( Salmo trutta ) and salmon {Salmo salar) in bone remains. J. Zool. (Lond.) 219:113-122. Grosslein, M. D., R. W. Langton, and M. P. Sissenwine. 1980. Recent fluctuations in pelagic fish stocks of the North- west Atlantic Georges Bank region, in relation to species interactions. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 177:374-404. Hackney, P. A. 1979. Influence of piscivorous fish on fish community struc- ture of ponds. In H. Clepper (ed.), Predator-prey systems in fisheries management, p. 111-121. Sport Fishing In- stitute, Washington, D. C.. Hansel, H. C., S. D. Duke, P. T. Lofy, and G. A. Gray. 1988. Use of diagnostic bones to identify and estimate origi- nal lengths of ingested prey fishes. Trans. Am. Fish. Soc. 117:55-62. Houde, E. D. 1987. Fish early life dynamics and recruitment varia- bility. Am. Fish. Soc. Symp. 2:17-29. Jobling, M., and A. Breiby. 1986. The use and abuse of fish otoliths in studies of feed- ing habits of marine piscivores. Sarsia 71:265-274. Juanes, F. 1994. What determines prey size selectivity in piscivorous 588 Fishery Bulletin 96(3), 1998 fishes? In D. J. Stouder, K. L. Fresh, and R. J. Feller (eds.), Theory and application in fish feeding ecology, p. 79- 100. Univ. South Carolina Press, Columbia, SC. Knight, R. L., F. J. Margraf, and R. F. Carline. 1984. Piscivory by walleyes and yellow perch in western Lake Erie. Trans. Am. Fish. Soc. 113:677-693. Langton, R. W., and R. E. Bowman. 1981. Food of eight northwest Atlantic pleuronectiform fishes. U.S. Dep. Commer., NOAA Technical Report NMFS SSRF 749. Lyons, J., and J. J. Magnuson. 1987. Effects of walleye predation on the population dy- namics of small littoral-zone fishes in a northern Wiscon- sin lake. Trans. Am. Fish. Soc. 116:29-39. McIntyre, D. B., and F. J. Ward. 1986. Estimating fork lengths of fathead minnows, ( Pimephales promelas), from measurement of pharyngeal arches. Can. J. Fish. Aquat. Sci. 43:1294-1297. McMahon, T. E., and J. C. Tash. 1979. Effects of formalin (buffered and unbuffered) and hydrochloric acid on fish otoliths. Copeia 1979:155-156. Miller, T. J., L. B. Crowder, J. A. Rice, and E. A. Marschall. 1988. Larval size and recruitment mechanisms in fishes: toward a conceptual framework. Can. J. Fish. Aquat. Sci. 45:1657-1670. Newsome, G. E. 1977. Use of opercular bones to identify and estimate lengths of prey consumed by piscivores. Can. J. Zool. 55:733-736. Overholtz, W. J., S. A. Murawski, and K. L. Foster. 1991. Impact of predatory fish, marine mammals, and sea- birds on the pelagic fish ecosystem of the northeastern USA. ICES Mar. Sci. Symp. 193:198-208. Pikhu, E. Kh., and E. R. Pikhu. 1970. Reconstruction of the size of fishes swallowed by predators from fragments of their vertebral column. J. Ichthyol. (USSR) 10:706-709. Raven, P. 1986. The size of minnow prey in the diet of young king- fishers Alcedo atthis. Bird Study 33:6-11. Rothschild, B. J. 1991. Multispecies interactions on Georges Bank. ICES Mar. Sci. Symp. 193:86-92. Scharf, F. S., J. A. Buckel, F. Juanes, and D. O. Conover. 1997. Estimating piscine prey size from partial remains: testing for shifts in foraging mode by juvenile bluefish. Env. Biol. Fish. 49:377-388. Serafy, J. E., C. M. Schmitz, T. R. Capo, M. E. Clarke, and J. S. Ault. 1996. Total length estimation of red drum from head dimensions. Prog. Fish Cult. 58:289-290. Sissenwine, M. P. 1984. Why do fish populations vary? In R. M. May (ed.), Exploitation of marine communities, p. 59-94. Springer- Verlag, Berlin. Smith, R. J. 1980. Rethinking allometry. J. Theor. Biol . 87:97-111. StataCorp. 1995. Stata statistical software: release 4.0. Stata Cor- poration, College Station, TX. Stein, R. A., S.T. Threlkeld, C. D. Sandgren, W. G. Sprules, L. Persson, E. E. Werner, W. E. Neill, and S. I. Dodson. 1988. Size-structured interactions in lake communities. In S. R. Carpenter (ed.). Complex interactions in lake com- munities, p. 161-180. Springer- Verlag, New York, NY. Tonn, W. M., C. A. Paszkowski, and I. J. Holopainen. 1992. Piscivory and recruitment: mechanisms structuring prey populations in small lakes. Ecology 73:951-958. Trippel, E. A., and F. W. H. Beamish. 1987. Characterizing piscivory from ingested remains. Trans. Am. Fish. Soc. 116:773-776. Werner, E. E., and J. F. Gilliam. 1984. The ontogenetic niche and species interactions in size- structured populations. Ann. Rev. Ecol. Sys. 15:393-426. 589 Abstract .—Large-mesh tangle nets were used to collect marine turtles in Waccasassa Bay, near the Cedar Keys, Florida, from June 1986 to October 1995. Tagging records were analyzed to determine the species composition, population structure, and seasonal oc- currence of Kemp’s ridley, Lepidochelys kempii, loggerhead, Caretta caretta , and green, Chelonia mydas , turtles. Additional information on local move- ments, morphometries, growth, popu- lation estimation, and diet was pro- vided for Kemp’s ridley turtles. Sub- adult green turtles dominated the catch on the seagrass shoals of Waccasassa Reefs. Subadult Kemp’s ridley turtles and, to a lesser degree, subadult and adult loggerhead turtles were primarily captured near the oyster bars of Cor- rigan Reef. Marine turtles were caught in these nearshore waters from April to November. Recaptures indicate that some Kemp’s ridley turtles remain in the vicinity of Corrigan Reef during their seasonal occurrence and return to this foraging area annually. Seasonal and annual size distributions of Kemp’s ridley turtles were investigated and regression equations were developed for carapace morphometries. Carapace growth averaged 4-5 cm/yr for Kemp’s ridley turtles, but growth analyses were confounded by the extrapolation of an- nual estimates from short-term recap- tures. Population estimates for the Kemp’s ridley mark-recapture data in- dicated a mean annual population size of 159 turtles at Corrigan Reef with presumably high rates of immigration and emigration by larger subadult turtles. Examination of fecal samples indicated that crabs were the primary food items of Kemp’s ridley turtles cap- tured near oyster bars. Manuscript accepted 26 January 1998. Fishery Bulletin 96:589-602 (1998). Marine turtle populations on the west-central coast of Florida: results of tagging studies at the Cedar Keys, Florida, 1 986-1 995 Jeffrey R. Schmid Southeast Fisheries Science Center National Marine Fisheries Service, NOAA 75 Virginia Beach Drive Miami, Florida 33 1 49 and Archie Carr Center for Sea Turtle Research University of Florida 223 Bartram Hall Gainesville, Florida 3261 I E-mail address: jeffrey.schmid@noaa.gov Historical information concerning marine turtles in the coastal waters of Florida is limited to landing sta- tistics and observational data asso- ciated with the commercial turtle fishery (Ehrhart, 1983). During the late 1800s, large-mesh tangle nets were used to catch significant num- bers of green turtles, Chelonia mydas , in the Indian River Lagoon and around the Cedar Keys (True, 1887; Brice, 1896). These turtles were exported to markets in the northeastern United States. Kemp’s ridley , Lepidochelys kempii , and log- gerhead turtles, Caretta caretta , were also captured and used as a food resource in local markets, but the landings of these species were not recorded in fisheries reports (Witzell, 1994a). The Florida fish- ery for marine turtles was greatly reduced by 1900 (Ingle and Smith, 1949). However, there are no quan- titative data to demonstrate accu- rately that depletion had occurred as a result of overfishing (Caldwell and Carr, 1957). The lack of defini- tive data is further complicated by the fact that most of the fishery sta- tistics reported for Florida after 1900 include green turtles imported from Costa Rica and Nicaragua by way of Key West, and no distinction was made between turtles caught in Florida and those imported from the Caribbean (Ingle and Smith, 1949; Caldwell and Carr, 1957; Witzell, 1994a, 1994b). The first scientific investigations and conservation efforts for marine turtles were implemented fifty years after the reduction of the turtle fishery. Ingle and Smith (1949) and Carr (1952) outlined sci- entific data necessary for the pro- tection and restoration of green turtles throughout their range. Carr and Caldwell (1956) conducted a one-year tagging study of green and Kemp’s ridley turtles purchased from Cedar Key fish houses and provided the first details on size ranges, morphometries, local move- ments, growth rates, and popula- tion estimates for these species. Florida enacted legislation in 1956 that prohibited the take of nesting female turtles and their eggs (Cald- well and Carr, 1957; Ehrhart, 1983). 590 Fishery Bulletin 96(3), 1998 Restrictions were imposed on the Florida turtle fish- ery in 1971, which consisted of closed seasons and size limits (Ingle, 1972). By 1978, all species of ma- rine turtles were listed as threatened or endangered in the Endangered Species Act and protected under federal legislation. Listing marine turtles in the Endangered Species Act outlawed their harvest and prompted surveys on nesting beaches and adjacent coastal waters in Florida (Carr et al., 1982). Entanglement nets, de- signed similarly to those formerly used in the turtle fishery, were employed to capture green and logger- head turtles inhabiting the northern Indian River Lagoon System (Ehrhart and Yoder, 1978; Mendonga, 1981, 1983; Mendonga and Ehrhart, 1982; Ehrhart, 1983). These fishery-independent studies provided the first biological data on population size and struc- ture, growth rates, and activity patterns for the pre- viously exploited species of marine turtle in the la- goonal habitat. Trawls associated with the commer- cial shrimp fishery were used to collect turtles oc- curring in and around the Port Canaveral ship chan- nel (Carr et al., 1980; Henwood, 1987; Henwood and Ogren, 1987). Information obtained from these fish- ery-dependent surveys includes size class distribu- tion, seasonal occurrence, and migrations of logger- head, Kemp’s ridley, and green turtles. In 1984, the National Marine Fisheries Service (NMFS) initiated long-term tagging studies of ma- rine turtles occurring in the coastal waters of Florida, with emphasis placed on the critically endangered Kemp’s ridley turtle (Ogren, 1989; Schmid and Ogren, 1990, 1992). Rudloe et al. (1991) reported on the size-class distribution, seasonal occurrence, and variations in carapace length by season and water depth for Kemp’s ridley turtles incidentally captured in commercial fisheries of northwest Florida. Tag- ging records of turtles captured in the east-central Florida shrimp fishery provided additional data on species composition, size-class distribution, seasonal occurrence and migrations, morphometric relation- ships, and growth data for marine turtles along the Atlantic coast (Schmid, 1995). Fishery-independent capture techniques have also been used to collect marine turtles in the nearshore waters of west-cen- tral Florida and preliminary results of these efforts have been given by Schmid and Ogren (1990, 1992). The present paper analyzes tagging records collected near the Cedar Keys, Florida, from 1986 to 1995 in order to determine the species composition, popula- tion structure, and seasonal occurrence of Kemp’s ridley, loggerhead, and green turtles. Additional in- formation on local movements, morphometries, growth, population estimation, and diet are provided for Kemp’s ridley turtles. Materials and methods Study area Marine turtles were collected east of the Cedar Keys in Waccasassa Bay, which is located on the west coast of Florida (Fig. 1). The northern and eastern bound- aries of Waccasassa Bay are saltmarsh coastline in- undated by numerous tidal creeks. The Waccasassa River flows into the northeast corner of the bay and is the main contributor of freshwater to the estua- rine system (Wolfe, 1990). The western edge of the bay is semi-enclosed by the Cedar Keys, whereas the southern portion is open to tidal exchange with the Gulf of Mexico. Corrigan Reef, located in northwest- ern Waccasassa Bay, and Waccasassa Reefs, located in the eastern half of the bay, are the prominent geo- graphic features of this shallow embayment. Netting efforts were concentrated at three sites along Corrigan Reef (Fig. 1; site 1: 29°09'N, 82°58'W; site 2: 29°08'N, 82°58'W; and site 3: 29°07'N, 82°58'W), approximately 5 km east of the Cedar Keys. Corrigan Reef comprises a series of oyster ( Crassostera vir- ginica) beds in the northern region (site 1) and oys- ter shell bars in the southern region (sites 2 and 3). Limestone outcroppings occur among the mud and sand flats and in channels on the periphery of oyster bars. Netting was also conducted at the outer shoal of Waccasassa Reefs (29°06'N, 82°53'W), approxi- mately 12 km east-southeast of the Cedar Keys. Waccasassa Reefs are composed of three seagrass shoals with a broad, deep-channel cutting midway through each shoal. The tides in both of these areas are mixed, with two highs and two lows of variable amplitude. Strong tidal currents flow through the channels, particularly during the new and full moon phases (spring tides). Data collection Seasonal netting was conducted at Corrigan Reef from 1986 to 1991 and at Waccasassa Reefs from 1986 to 1988 (see Table 1 for effort and months fished each year). A full year of sampling was performed at Corrigan Reef in 1992 and 1993. Water temperatures were recorded sporadically from 1986 to 1991 and monthly for 1992 and 1993. Netting surveys were performed for 1-3 days every other week during the neap tides. One or two nylon mesh tangle nets (61- or 51-cm stretch mesh, 20 meshes deep, and 65 m length) were set across a channel at a given site and fished over a 6- to 12-hour tidal cycle. Nets were checked hourly or immediately after an entangled turtle had been sighted. Research efforts shifted from netting surveys to telemetric monitoring during 1994 Schmid: Marine turtle populations on the west-central coast of Florida 591 Table 1 Annual effort (km-h) and CPUE (turtles/km-h) by species for Corrigan Reef and Waccasassa Reefs from 1986 to 1993. Lk -Lepidochelys kempii, Cc =Caretta caretta, and Cm =Chelonia mydas. Corrigan Reef Waccasassa Reefs Year (months) Effort Lk/h Cc/h Cm/h Effort Lk/h Cc/h Cm/h 1986 ( Jun-Nov) 153.9850 0.1104 0.0065 0.0000 9.3275 0.0000 0.0000 0.1072 1987 (May-Dec) 236.0963 0.0720 0.0000 0.0042 19.6300 0.0000 0.0509 0.2547 1988 (May-Sep) 142.3988 0.2388 0.0070 0.0000 1.2675 0.7890 0.0000 0.0000 1989 (Jun-Sep) 177.4500 0.1634 0.0394 0.0113 — — — — 1990 (Jun-Sep) 233.4150 0.1499 0.0129 0.0000 — — — — 1991 (Jun-Oct) 152.5225 0.1246 0.0000 0.0000 — — — — 1992 (May-Dec) 461.5650 0.1148 0.0303 0.0022 — — — — 1993 (Jan-Oct) 124.8000 0.1923 0.0160 0.0000 — — — — and 1995, and tag data collected on these turtles were included in the analyses. The following morphometric measurements (Prit- chard et al., 1983) were recorded for each turtle: to- tal straight-line carapace length (TSCL, anterior most edge of carapace to posterior margin of post- central); standard straight-line carapace length (SSCL, nuchal notch to posterior margin of postcen- 592 Fishery Bulletin 96(3), 1 998 tral); minimum straight-line carapace length (MSCL, nuchal notch to notch between postcentrals); mini- mum curved carapace length (MCCL, nuchal notch to notch between postcentrals); and straight-line carapace width (CW) at the widest point. Straight- line carapace lengths and width were measured to the nearest 0.1 inch with forester’s calipers. Curved carapace length was measured to the nearest 0.1 cm with a flexible fiberglass measuring tape. Weight (WT) was measured to the nearest 0.25 lb with a spring scale. All measurements were performed by the author to avoid individual differences in mea- suring technique (Bjorndal and Bolten, 1988) and were converted to metric units for analysis. Notes on the condition of the turtle were recorded when the animal was injured or deformed (e.g. tag scars, carapace wounds, etc.). Turtles were double tagged on the trailing edge of the fore flippers with no. 681 Inconel tags (June 1986 to May 1988; May 1994 to October 1995), with Jumbo Roto plastic tags (June 1988 to October 1991), or with both (May 1992 to September 1993). Beginning in 1988, two holes were drilled in specific marginal scutes of Kemp’s ridley turtles in order to identify the year of capture (1988 — left postcentral, 1989 — right postcentral, 1990 — left 12th marginal, 1991 — right 12th marginal, 1992 — left 11th marginal, 1993 — right 11th marginal, 1994 — left 10th mar- ginal, and 1995 — right 10th marginal). Passive inte- grated transponder (PIT) tags were applied to the left front flipper of Kemp’s ridley turtles from June 1992 to October 1995. Turtles were immediately re- leased after data collection approximately 100 m down-current from the netting site. Data analysis Kemp’s ridley turtles. Standard straight-line cara- pace length was used in the analyses of size distri- bution and growth. Means are followed by ± one stan- dard deviation unless noted otherwise. Turtle catch per unit of effort (CPUE) was standardized accord- ing to Shaver (1994) with the formula EJ Nets * Length ^ where E = the netting effort in hours fished by a 1- km tangle net; Nets = the number of tangle nets fished; Length = the length (m) of a net; and Hrs = the number of hours fished. Kemp’s ridley turtle morphometric relationships were investigated by regressing carapace width on length and log-transformed weight on length. Con- version formulae for Kemp’s ridley turtle carapace lengths were calculated by regressing paired straight-line and curved carapace lengths. Turtles with carapace wounds or deformities were not in- cluded in regression equations. Yearly growth rates for Kemp’s ridley turtles were calculated with the formula G = ' A Length ' v Days , 365, where G = the growth rate in cm/yr; A Length = difference between the recapture length and the initial length; and Days = the number of days at large. Marine turtle life history stages were defined accord- ing developmental habitats and carapace lengths (Schmid, 1995). The term “juvenile” was reserved for immature turtles in the epipelagic stage of develop- ment. A turtle was considered “subadult” after re- cruiting to its respective coastal-benthic habitat and “adult” when sexually mature. Loggerhead turtles greater than 80 cm (Carr, 1986), green turtles greater than 83 cm (Witherington and Ehrhart, 1989), and Kemp’s ridley turtles greater than 60 cm (Pritchard and Marquez M., 1973) were considered adult based on sizes of nesting females. Capture records were analyzed to evaluate species composition within the Cedar Keys study area, length-frequency distribution of each species, and patterns of seasonal occurrence. Additional analyses of morphometries, growth rates, population esti- mates, and dietary composition were performed for Growth rates were grouped and analyzed in terms of the recapture interval duration, recaptures be- tween versus recaptures within netting seasons, and size classes of recaptured turtles. Growth rates were assigned to 10-cm size classes on the basis of mean of the initial and recapture carapace measurements (Bjorndal and Bolten, 1988). The von Bertalanffy growth interval equation was fitted to the recapture data with a nonlinear least-squares regression pro- cedure (SAS Institute Inc., 1989). The von Berta- lanffy growth interval equation (Fabens, 1965) for recapture data is: CL2 = a - (a - CLx)e~kt , where CL9 = the carapace length at recapture; a = the asymptotic length; CLj = the length at first capture; Schmid: Marine turtle populations on the west-central coast of Florida 593 k = the intrinsic growth rate; and t = the time in years between captures. Kemp’s ridley turtle mark-recapture data for 1986-93 were tallied in a method B table (Krebs, 1989) and analyzed with the computer program JOLLY (Hines, 1988; Pollock et ah, 1990). Sum- mary statistics for the Jolly-Seber computer analy- sis include the total number of turtles captured and released each year ( n ), the number of marked (m) and unmarked ( u ) turtles captured each year, the number of turtles released each year that are captured again later (r), and the number of turtles captured before a given year and captured again later (2). The data were applied to a Jolly-Seber model that assumes that the population param- eters survival rate (O) and capture probability (p) are constant per unit time. Annual estimates of the number of marked turtles in the population (M), population size (AO, and the number of indi- viduals recruited to the population ( B ) were com- puted with the reduced parameter model. Dietary analyses were conducted on Kemp’s rid- ley fecal specimens fortuitously encountered dur- ing the tagging process. Fecal specimens were ini- tially examined in the field and the contents were noted in tagging records. Additional examinations were performed from photographs and samples of feces. Components of the feces were identified to the lowest taxon possible and were analyzed to determine the percentage of specimens contain- ing each component (Burke et al., 1994). Nomen- clature of molluscs was identified by using the field guide of Abbott and Morris (1995). Results Marine turtle captures and effort One (12.5%) Kemp’s ridley, 1 (12.5%) loggerhead, and 6 (75%) green turtles were collected during 64.75 h of netting at Waccasassa Reefs. The Kemp’s ridley turtle measured 47.6 cm SSCL, the loggerhead turtle measured 86.4 cm SSCL, and green turtles ranged from 63.0 to 73.9 cm SSCL (mean=68.0 ±3.9 cm; Fig. 2). Three of the green turtles captured at Waccasassa Reefs exhibited fibropapillomas ( 1-4 cm diameter), primarily in the axillary region of the flippers. Maximum CPUE for green turtles at Waccasassa Reefs was 0.255 turtles/km-h in 1987, and values of CPUE for loggerhead and Kemp’s rid- ley turtles were 0.051 turtles/km-h in 1987 and 0.789 turtles/km-h in 1988, respectively (Table 1). Netting effort was conducted at this location from June Capture Locations: E3 Corrigan Reef KS Waccasassa Reefs Carapace length (cm) Figure 2 Length-frequency distributions for Kemp’s ridley turtles, Lepi- dochelys kempii , loggerhead turtles, Caretta caretta , and green turtles, Chelortia my das, collected in Waccasassa Bay, Florida, from 1986 to 1995. Note the different scale in y axis of upper graph. through November and turtles were captured in July and August. Two hundred and fifty-three (91.7%) Kemp’s rid- ley, 19 (6.9%) loggerhead, and 4 (1.4%) green turtles were collected during 980.00 h of netting at Corrigan Reef. Kemp’s ridley turtles ranged from 26.8 to 58.6 cm SSCL (mean=44.5 ±6.3 cm), loggerhead turtles ranged from 50.0 to 77.4 cm SSCL (mean=65.0 ±8.7 cm), and green turtles ranged from 42.9 to 70.9 cm SSCL (mean=56.8 ±12.9 cm; Fig. 2). Loggerhead turtles greater than 80 cm SSCL were caught at 594 Fishery Bulletin 96(3), 1998 Corrigan Reef but could not be landed for data collection. One such turtle was identi- fied as a male because its tail extended considerably be- yond the posterior marginal scutes. Fibropapillomas were not observed on green turtles captured at Corrigan Reef. Annual CPUE for Kemp’s rid- ley turtles at Corrigan Reef ranged from 0.072 turtles/km-h in 1987 to 0.239 turtles/km-h in 1988 (Table 1). Maximum CPUE for loggerhead and green turtles at Corrigan Reef was 0.039 turtles/km-h and 0.011 turtles/km-h in 1989, respectively. Kemp’s ridley and loggerhead turtles were cap- tured in this area from April to November, whereas green turtles were captured from June to September. Recaptures and local movements Thirty-four Kemp’s ridley turtles (23 with tags and 11 with tag scars), five logger- head turtles (3 with tags and 2 with tag scars), and one green turtle (with a tag scar) were identified as recaptures. All recaptured turtles with tags, with the exception of two NMFS Galveston labora- Figure 3 Photographs of recaptured Kemp’s ridley turtles, Lepidochelys kempii , demonstrating (A) a barnacle-encrusted tag and (B) a flipper scar from tag loss. tory headstart Kemp’s rid- leys, were initially captured and tagged at Corrigan Reef. Thirty-five percent of the re- captured turtles exhibited tag scars, which is indicative of moderate tag loss and may account for the lack of recaptures or recoveries in other areas. Schmid and Ogren ( 1992) identified bar- nacle fouling as a potential problem with the other- wise corrosion-resistant Inconel flipper tag. The in- creased drag and weight produced by the barnacle clusters and the necrosis of flipper tissue by the en- crusted tag (Fig. 3A) resulted in the eventual shed- ding of the tags and the formation of a conspicuous notch in the trailing edge of the flippers (Fig. 3B). Barnacle growth was observed on both Inconel and plastic tags in as little as 14 days and tag loss was observed within 10 months after application. Simi- lar retention times were noted for both types of tags, but a quantitative analysis of retention rates was not performed because of small sample sizes. The use of marginal markings allowed for the identification of tag- scarred turtles originally tagged at the Cedar Keys and the tabulation of recapture data by year classes. PIT tags were successfully used to identify tag-scarred Kemp’s ridley turtles in the later part of the study. Schmid: Marine turtle populations on the west-central coast of Florida 595 o Q. E o o a> N Cl a> > jg ® c: 60- 30- C 40- 20- 0- I 40- mean SSCL = 47.23 cm n = 17 — ! — 10 — I 1 — 20 30 40 50 60 — I — 70 1986 80 mean SSCL = 44.06 cm n = 17 10 20 30 40 50 60 — i — 70 1987 80 20- 0 mean SSCL = 45.46 cm n =35 31 1988 0 10 20 30 40 50 60 70 80 60' 30' 0- 60- 30- 0- 60 mean SSCL = 46.96 cm rr n =29 S 1 T 1 1 f ----- — i 1 1989 10 20 30 40 50 60 70 80 mean SSCL = 45.74 cm n =35 m\ i i i ^ — i 1 1990 10 20 30 40 50 60 70 80 mean SSCL = 40.79 cm m n =21 [777 :::: T1 ■ | ^ T | 1991 10 20 30 40 50 60 70 80 40- 20- 0- 40 20- 0- mean SSCL - 43.57 cm n = 53 [717 •7T7? , bq T j 1992 10 20 30 40 50 60 70 80 mean SSCL = 43.58 cm n =24 10 20 30 40 50 60 Carapace length (cm) — I- 70 1993 80 Figure 4 Annual relative size composition of Kemp’s ridley turtles, Lepidochelys kempii, captured at Corrigan Reef from 1986 to 1993. Kemp’s ridley recaptures ranged from 14 to 839 days and loggerhead recap- tures ranged from 142 to 189 days. The two headstart Kemp’s ridley turtles were recaptured approximately 3-4 years after their release in Texas wa- ters as was determined from the loca- tion of the living tag in the carapace (Fontaine et ah, 1993; Cailiouet et ah, 1995). Seven Kemp’s ridley turtles with tag scars had marginal scute markings indicating the year of initial capture at Corrigan Reef. Four of these were at large for 1 year, one was at large for 2 years, and two were at large for 3 years. Four Kemp’s ridley turtles had multiple recaptures in the vicinity of Corrigan Reef. One turtle was tagged at site 2 in September 1991 and recaptured at this site in October 1991 and May 1992 (0.7 year duration). Another turtle tagged at site 1 in July 1990 was recaptured at this site in June 1991 and June 1992 (1.9-year duration). A turtle tagged at site 2 in October 1991 was recaptured at this site in September 1992 and at site 3 in May 1994 (2.6-year duration). The fourth turtle had a 1991 marginal marking and was recaptured at site 2 in September 1993 and August 1995 ( = 4-year duration). Seasonal and annual size distributions Mean water temperatures at Corrigan Reef were calculated by season (Table 2): winter ( Dec- Feb., spring (Mar- May), summer (Jun-Aug), and fall (Sep-Nov). Turtles were captured in water temperatures greater than 20°C. Carapace lengths for Kemp’s ridley turtles captured from 1986 to 1995 were pooled by season (spring, summer, or fall) according to the month of capture (Table 2). A significant difference (F- 3.76, P-0.025) was detected between the mean SSCL of at least two seasons. Multiple comparisons with the Bonferroni procedure demonstrated no significant difference in mean SSCL between spring and summer, or spring and fall. However, mean SSCL in summer was sig- nificantly larger than that of fall. Analysis of the annual relative composition of Kemp’s ridley turtle carapace lengths from 1986 to 1993 indicated that the 40-50 cm size class domi- nated the catch during all years except 1991 (Fig. 4). The majority of turtles captured during 1991 were in the 30-40 cm size class. The carapace length dis- tributions for 1986 through 1990 were not signifi- cantly different when compared with the Kolmogrov- Smirnov two-sample test. The distribution of cara- pace lengths for 1991 was significantly different from all years except 1987, and the distribution for 1992 was significantly different from all years except 1987 and 1988. The carapace length distribution for 1993 596 Fishery Bulletin 96(3), 1998 Table 2 Seasonal water temperatures and Kemp’s ridley turtle, Lepi- dochelys kempii, carapace lengths at Corrigan Reef from 1986 to 1995 (standard deviations given in parentheses). Season Mean water temperature Mean carapace length n Size range Spring 22.5°C (3.0) 44.1 cm (7.4) 24 26.8-58.6 cm Summer 30.4°C (0.6) 45.5 cm (5.9) 142 28.2-54.4 cm Fall 22.9°C (4.2) 43.1 cm (6.6) 85 27.3—56.6 cm Winter 15.7°C (2.0) 0 was not significantly different from the distributions of 1986 through 1990. The observed shift in size-class distribution for 1991 may have been caused by changes in fishing conditions that year. Beginning in July 1991, a smaller mesh (25.4-cm bar) net was deployed either singly or in combination with the larger mesh (30.5-cm bar) net that was used the pre- vious years. The smaller mesh net may have resulted in the increased capture of 30-40 cm turtles in 1991, although the frequency of 40-50 cm turtles increased in the following years. Also, during August 1991, a massive influx of pelagic Sargassum occurred along the west coast of Florida and the majority of netting effort was conducted in the months following this unusual event. It is not known how this latter condition may have affected the relative frequency of carapace lengths, either by increasing the frequency of 30—40 cm turtles or decreasing the frequency of 40-50 cm turtles. Carapace regression equations There was a strong correlation between carapace width and carapace length (r=0.9883, n= 227) for Kemp’s ridley turtles. Regression of width on length resulted in the equation CW = -3.7415 + 1.0530 ( SSCL ). A strong correlation (r=0.9886, n=225 ) was calculated for the weight-to-length data transformed with the natural logarithm. Regression of these variables re- sulted in the equation In WT = -8.1570 + 2.8128 (In SSCL). Conversion equations were computed between the straight-line and curved carapace length measure- Table 3 Formulae for converting between straight-line and curved carapace measurements of Kemp’s ridley turtles, Lepido- chelys kempii. TSCL = total straight-line carapace length, SSCL = standard straight-line carapace length, MSCL = minimum straight-line carapace length, and MCCL = mini- mum curved carapace length. Converted length Conversion formula n r2 TSCL 1.0118 SSCL + 0.0650 227 0.9990 TSCL 1.0214 MSCL + 0.3800 227 0.9982 TSCL 0.9720 MCCL + 0.2203 107 0.9920 SSCL 0.9874 TSCL -0.0199 227 0.9990 SSCL 1.0094 MSCL + 0.3155 227 0.9990 SSCL 0.9598 MCCL + 0.1941 107 0.9929 MSCL 0.9773 TSCL - 0.2896 227 0.9982 MSCL 0.9897 SSCL -0.2658 227 0.9990 MSCL 0.9495 MCCL - 0.0864 107 0.9933 MCCL 1.0205 TSCL + 0.1368 107 0.9920 MCCL 1.0345 SSCL + 0.1162 107 0.9929 MCCL 1.0462 MSCL + 0.3913 107 0.9933 ments (Table 3). These equations will allow for com- parisons between studies with different measuring techniques. Growth analyses Twenty-one Kemp’s ridley turtles were recaptured a total of 24 times, yielding 24 annual growth rates. However, 83% of the recaptures occurred within a year of initial tagging and extrapolating annual growth rates from short-term recapture intervals will amplify any error associated with the measurements. The removal of short-term recaptures decreased the range of annual growth rates and increased the pre- cision of the mean growth rate estimate (Table 4A). Subsequent analyses of growth by netting seasons and size classes were confounded by short-term re- captures. The mean growth rate of Kemp’s ridley turtles recaptured within netting seasons (see Table 1 for months fished each year) was significantly larger (X2=7.93, df=l, P=0.005) than that of turtles recap- tured between netting seasons (Table 4B). However, the duration of all recaptures within netting seasons was less than 180 days (mean=49.6 ± 44.5 days) and the annual growth rates may have been overesti- mated owing to extrapolation error. Although mean growth rates did not vary significantly when com- pared by size class (F=0.753, P=0.484), Kemp’s rid- ley turtles in the 40-50 cm size class appeared to have a higher mean growth rate than those in the 30-40 cm and the 50-60 cm size classes (Table 4C). Schmid: Marine turtle populations on the west-central coast of Florida 597 Deletion of data with recapture intervals less than 90 days reduced the mean growth rate of the 40-50 cm size class (4.7±3.0 cm/yr; n= 9, range: 2.9-12.3 cm/yr). The von Bertalanffy growth interval equation was fitted to each of the recapture interval data treat- ments. Estimates of asymptotic length ranged from 77.3 to 91.4 cm and estimates of intrinsic growth rate ranged from 0.0852 to 0.1167 (Table 5). The growth interval equation for all Kemp’s ridley turtles recap- tured at Cedar Key had the lowest residual mean square, a standard that has been used to select the best fit growth model (Dunham, 1978). However, the estimated asymptotic length for this model (a=91.4 cm) is considerably larger than the average carapace length reported for nesting females (65 cm; Marquez M., 1994) and should therefore be considered biologi- cally unrealistic (Frazer et al., 1990). The estimated asymptotic length for recapture intervals greater than 180 days (a= 77.3 cm) would be more appropri- ate if this latter criterion is used. This model has the least amount of error from short-term recaptures, but suffers from a reduced sample size and a trun- cated range of carapace lengths. Population estimations The computer program JOLLY computed a survival rate of 0.41 (± 0.07 SE) and a capture probability of 0.18 (±0.05 SE) for Kemp’s ridley turtles at Corrigan Reef. Population estimates ranged from 98.05 turtles in 1987 to 262.79 turtles in 1992 (Table 6). For 1987 through 1993, the mean annual population size was 158.50 (±112.40 SE) turtles and there was a mean of 15.35 (±11.58 SE) marked turtles in the population (10% of the estimated mean population size). For 1987 through 1992, there was a mean annual recruit- ment of 102.71 (±48.23 SE) turtles (65% of the esti- mated mean population size). Food Fecal specimens from 12 Kemp’s ridley turtles were examined during the course of this study. Crab com- ponents were identified in all specimens. In addition, two (17%) of the fecal specimens contained mollusc shells and two (17%) specimens contained a portion of undigested seagrass (. Halodule wrightii in one and Halophila engelmannii in the other). Seven (58%) of the fecal specimens contained unidentified crab frag- ments. Five (42%) of the turtles had consumed cheli- peds of stone crab, Menippe spp., and three (25%) had consumed chelipeds of blue crab, Callinectes sapidus. Two (17%) Kemp’s ridley turtles had con- sumed shark eye shells ( Polinices duplicata), one of which also consumed a common eastern nassa shell Table 4 Mean annual growth rates of Kemp’s ridley turtles, Lepi- dochelys kempii, by recapture interval, netting season, and size class (standard deviations given in parentheses). Turtles were assigned to size classes by mean of initial and recapture SSCL. Mean SSCL growth rate Range of growth rates Data treatments n (cm/yr) (cm/yr) Recapture interval All recaptures 24 5.4 (3.3) 1.2-13.0 Recaptures > 90 days 16 4.5 (2.6) 1.2-12.3 Recaptures > 180 days 13 3.6 (1.2) 1. 2-5.4 Netting season Within season 10 7.7 (3.6) o co rH 1 t> Between seasons 11 3.3 (1.1) 1.2-4. 7 Size class 30-40 cm 7 4.6 (2.8) 1. 2-9.4 40-50 cm 13 6.2 (3.7) 2.9-13.0 50-60 cm 4 4.6 (2.5) 2. 2-7. 9 Table 5 Estimated values of asymptotic length (a) and intrinsic growth rate ( k ) from nonlinear regression of von Berta- lanffy growth interval equation for Kemp’s ridley turtles, Lepidochelys kempii (one asymptotic standard error in pa- rentheses). Data treatment n a k All recaptures 24 91.4 cm 0.0852 (41.9) (0.0720) Residual mean square error = 1.3872 All recaptures > 90 days 16 90.9 cm 0.0858 (51.1) (0.0892) Residual mean square error = 2.1179 All recaptures > 180 days 13 77.3 cm 0.1167 (29.2) (0.0957) Residual mean square error = 2.0085 ( Nassarius vibex), that contained hermit crabs (Paguridae). The two turtles that consumed hermit crabs also ingested mollusc components. Cancellate cantharus shells ( Cantharus cancellarius) and oys- 598 Fishery Bulletin 96(3), 1998 Table 6 Summary statistics and estimated parameters for the Jolly-Seber analysis of Kemp’s ridley turtle, Lepidochelys kempii , mark- recapture data (standard errors given in parentheses). Descriptions of the notation are as follows: n = total number of turtles captured and released each year; m = number of marked turtles captured each year; u = number of unmarked turtles captured each year; r = number of turtles released each year that were captured again later; 2 = number of turtles captured before a given year and captured again later; M = annual estimate of the number of marked turtles in the population; N = annual estimate of population size; and B = annual estimate of the number of individuals recruited to the population. Summary statistics Estimated parameters Year n u m r 2 M N B 1986 17 17 0 3 — — — — 1987 17 16 1 3 2 10.91 (5.49) 98.05 (32.11) 139.87 (42.90) 1988 35 31 4 1 1 18.20 (7.41) 187.03 (55.83) 96.38 (32.86) 1989 29 28 1 3 1 7.30 (4.70) 159.79 (50.22) 107.25 (34.87) 1990 32 29 3 2 1 14.17 (6.72) 172.66 (52.53) 45.54 (24.72) 1991 20 18 2 6 1 11.32 (6.05) 109.35 (35.92) 201.60 (58.21) 1992 49 43 6 2 1 28.61 (10.55) 262.79 (75.59) 25.64 (28.35) 1993 22 19 3 16.34 (9.53) 119.81 (39.34) ter shell fragments were identified in both of the fe- cal specimens. Furthermore, one of these specimens contained hooked mussels (Ischadium recurvus ) at- tached to an oyster shell fragment. Discussion The results of this study indicate the importance of seagrass beds and oyster reefs as developmental habitats for Kemp’s ridley, loggerhead, and green turtles. Furthermore, these species may be prefer- entially utilizing the two habitat types on the basis of their respective feeding strategies. The extensive seagrass flats along the west coast of Florida have been identified as foraging habitat for the herbivo- rous green turtle (True, 1887; Carr and Caldwell, 1956; Caldwell and Carr, 1957). Netting effort at the seagrass shoals of Waccasassa Reefs resulted in cap- tures of mid- to late subadult green turtles, compa- rable to the size class of green turtles reported by Carr and Caldwell (1956). The Kemp’s ridley turtle is primarily cancivorous (Shaver, 1991; Burke et al., 1994), and the distribution of this species can be cor- related to areas with abundant crab populations (Ogren, 1989). Intertidal oyster bars provide refuge for stone crabs (McRae, 1950; Bender, 1971; Wilber and Herrnkind, 1986), whereas the mud bottom ad- jacent to these bars is the preferred substrate of blue crabs (Evink, 1976; Wolfe, 1990). Subadult Kemp’s ridley turtles dominated the aggregation of marine turtles captured in the vicinity of Corrigan Reef and the food items for these turtles were typical of the macroinvertebrate fauna inhabiting nearshore oys- ter bars. Subadult and adult loggerhead turtles were also captured at Corrigan Reef; this species also feeds on benthic invertebrates, particularly molluscs (Dodd, 1988). The possibility of competition for food resources between loggerhead and Kemp’s ridley turtles is unknown and could be investigated by com- paring fecal specimens of both turtles captured in the same location. Tagging studies of Kemp’s ridley turtles have re- vealed reproductive migrations of females in the Gulf of Mexico (Pritchard and Marquez M., 1973), sea- sonal migrations of subadults along the Atlantic coast (Henwood and Ogren, 1987; Schmid, 1995), and an east-west movement of subadults in the northern Gulf (Ogren, 1989). However, there are no mark-re- capture data to indicate a seasonal migration of sub- adult turtles in the eastern Gulf of Mexico. As stated by Carr (1980) and observed in the present study, turtles apparently immigrate to the nearshore wa- ters of the Cedar Keys in April and emigrate to some Schmid: Marine turtle populations on the west-central coast of Florida 599 unknown locality in November, presumably in re- sponse to changes in water temperature. Ogren (1989) suggested a seasonal offshore movement of Kemp’s ridley turtles in the northern Gulf on the basis of capture of subadult turtles in deeper waters off Apalachicola Bay during the winter (Rudloe et al., 1991). Satellite telemetry has demonstrated that Kemp’s ridley turtles on the Atlantic coast respond to a decrease in water temperature by moving to warmer waters southward or offshore (Renaud, 1995). Marine turtles in the Cedar Keys area could be moving westward to deeper waters offshore or southward within the shallow coastal waters. Alter- natively, some Cedar Key fishermen believe that turtles overwinter in the remote coastal waters by “burying up” in mud bottom holes (Carr and Caldwell, 1956; Schmid and Ogren, 1990). Loggerhead turtles have exhibited this behavior in the Port Canaveral ship channel at water temperatures below 15°C (Carr et ah, 1980; Ogren and McVea, 1982). Water tem- peratures as low as 12-14°C were recorded at the Cedar Keys study area from December to February. The possibility of winter dormancy or migration (or both) by west coast turtles requires additional infor- mation (Ogren and McVea, 1982) and could be in- vestigated by attaching satellite transmitters to turtles during the fall. Recaptures of Kemp’s ridley turtles tagged and released in the northeastern Gulf of Mexico have provided information on their use of coastal forag- ing grounds. Carr and Caldwell ( 1956) observed that Kemp’s ridley and green turtles released from Ce- dar Key returned to the Withlacoochee-Crystal River fishing grounds within a short period of time. The authors implied that the turtles may be exhibiting a homing behavior and maintaining home ranges at the site of initial capture. Schmid and Ogren (1990) suggested that Kemp’s ridley turtles in the Florida panhandle region were transitory because of the lack of long-term recaptures (Rudloe et al., 1991). By com- parison, recaptures in the Cedar Keys area were in- dicative of a more residential aggregation. Although the majority of Kemp’s ridley turtles tagged near the Cedar Keys were recaptured within a year of initial capture, almost equal numbers of turtles were re- captured within netting seasons and between net- ting seasons. Recaptures within a netting season suggest that some turtles remain in the vicinity of Corrigan Reef during their seasonal occurrence in this region. Recaptures between netting seasons indicate that some turtles return to the previously utilized oys- ter bar habitat annually and may do so for up to four years. Efforts are currently underway to determine the activity patterns and habitat associations of Kemp’s ridley turtles in the Cedar Keys area (Schmid, 1994). Kemp’s ridley turtles were numerous in the coastal waters of Florida prior to the 1950s (Carr, 1980). Data provided by Carr and Caldwell (1956; p. 21, Fig. 3) indicated that approximately 1% of the Kemp’s rid- ley turtles captured at the Withlacoochee-Crystal River fishing grounds were early to mid-subadults (20-40 cm), 88% were mid- to late subadults (40-60 cm), and 11% were adult (60+ cm). There was also an unconfirmed report of a vitellogenic female weigh- ing 42 kg with an estimated length of 75 cm. The presence of adult turtles in this 1955 survey corre- sponds to a period when there were relatively large, though declining, nesting aggregations of Kemp’s ridley turtles (U.S. Fish and Wildlife Service and National Marine Fisheries Service, 1992). The lack of 20-40 cm turtles may be indicative of lower sub- adult recruitment resulting from the intensive egg harvesting that was occurring at this time (Hilde- brand, 1982). In contrast, the catch at the Cedar Keys from 1986 to 1995 comprised 24% early to mid-sub- adults and 76% mid- to late subadults. Observations and captures of adult Kemp’s ridley turtles at sea have become extremely rare owing to the greatly re- duced nesting population, whereas the higher fre- quency of 20-40 cm turtles may suggest higher sub- adult recruitment as a result of nesting beach pro- tection (Ogren, 1989). Carr and Caldwell (1956) also described an appar- ent, though not statistically significant, seasonal shift in the mean carapace length of Kemp’s ridley turtles taken in the commercial turtle fishery. Larger turtles (mean=54.9 cm) were captured early in the April to November fishing season, whereas smaller turtles predominated the mid- and late season catch (mean=50.3 cm and 52.1 cm, respectively). The sea- sonal mean carapace lengths reported in this earlier study were 5-10 cm greater than those recently re- corded at the Cedar Keys. The authors’ description of their measurement technique (“. . . from the cen- ter of the anterior end of the carapace and the great- est posterior projection of the carapace.”) corresponds to the standard straight-line carapace length. Fur- thermore, the entanglement nets used in the present study were the same mesh size as those used in the commercial turtle fishery. The perceived difference could be due to preference by the former turtle fish- ermen for larger, higher-priced turtles. Alternatively, the smaller mean carapace lengths of the present study may be indicative of an increased aggregate of smaller Kemp’s ridley turtles along the west coast of Florida. Despite numerous tagging studies, there is very little information available on the growth rates of wild Kemp’s ridley turtles (Marquez M., 1994). The data treatments used to analyze Kemp’s ridley turtle 600 Fishery Bulletin 96(3), 1998 growth at the Cedar Keys indicated an average in- crease of 4-5 cm/yr in carapace length. Growth rates of 6-9 cm/yr were obtained for Kemp’s ridley turtles at Cape Canaveral with the same data treatments (Schmid, 1995). The higher growth rates observed for east coast turtles are possibly due to measure- ment errors identified in the Cape Canaveral study. Error was minimized in the Cedar Key study because all measurements were determined by the author using the same equipment and techniques. Assum- ing a constant growth rate of 4-5 cm/yr, a Kemp’s ridley turtle would require 8-10 years to grow from a 20-cm postpelagic subadult (Ogren, 1989) to a 60- cm adult. An estimate of 10-12 years to maturity is calculated by combining the duration of the subadult stage with the estimated 2-year pelagic juvenile stage (Schmid and Ogren, 1990). This calculated age to sexual maturity is in agreement with Kemp’s ridley growth models computed from skeletochronological age estimates (Zug and Kalb, 1989; Zug, 1990) and the combination of recapture data for Cape Canaveral and Cedar Keys (Schmid and Witzell, 1997). The Kemp’s ridley turtle aggregation at the Cedar Keys was considered an open “population” with re- cruitment in terms of postpelagic turtles and sub- adult immigrants from other locations and with losses in terms of death and permanent emigration to other subadult or adult aggregations. Annual es- timates from the Jolly-Seber analysis were indica- tive of the catchable turtle population at Corrigan Reef within the months fished, which may or may not be representative of the entire aggregation in this area (Krebs, 1989). The relatively low number of re- captures, and corresponding low estimated capture probability, reduced the precision of the population estimates as evidenced by their high standard er- rors (Pollock et al., 1990). Nonetheless, general com- ments can be made concerning the estimates of re- cruitment and survival of Kemp’s ridley turtles in the Cedar Keys area. The majority of turtles cap- tured in this locality were mid- to late subadults; there were very few captures of postpelagic turtles. Therefore, provided that sampling bias due to the large-mesh nets was minimal, the high level of re- cruitment estimated from the Jolly-Seber analysis was presumably a result of immigration by larger subadult turtles. The low estimated survival rate was probably a function of emigration rather than high turtle mortality. However, there were no recaptures of turtles tagged at Cedar Key to indicate emigra- tion to other localities and there was no systematic sampling of turtle strandings to demonstrate the extent of mortality in this region. In conclusion, tagging studies conducted at the Cedar Keys are models for characterizing foraging populations of marine turtles and these efforts must be expanded to include regions not yet sampled in order to accurately manage these threatened and endangered species (Magnuson et al., 1990; Thomp- son et al., 1990; U.S. Fish and Wildlife Service and National Marine Fisheries Service, 1992). Areas where marine turtle congregate need to be identi- fied through anecdotal information, historical records, incidental captures, and stranding data. In- water sampling programs should be conducted over an extended period of time to establish the distribu- tion and abundance of turtles in areas of aggrega- tion. After implementing a mark-recapture study, supplementary research activities may include the following: holding turtles for fecal sample collection; sampling blood for stress response, sex determina- tion, and genetic analyses; monitoring local move- ments via radio and sonic telemetry; discerning mi- grations via satellite telemetry; and developing GIS models for marine turtle habitat associations. Acknowledgments This project was initiated and managed in part by Larry Ogren, who was assisted by the late Junior McCain on earlier netting surveys. I am indebted to Edgar and Rosa Campbell for their expert advice on netting marine turtles and for the unrestrained use of their facilities. Thanks are also due to Kenny Collins, Lloyd Collins, and Tracey Collins for the use of their fishing vessels and for their assistance in capturing turtles; Richard Byles and Charles Caillouet for supplying PIT tags; Jamie Barichivich, Mike Cherkiss, Kris Fair, Lisa Gregory, Debbie Weston, and numerous student volunteers for their assistance in the field during the later stages of this project; and Frank Maturo for confirming the iden- tification of mollusc shells. Alan Bolten, Larry Ogren, Wayne Witzell and two anonymous reviewers pro- vided constructive comments during manuscript prepa- ration. This study was supported by the NMFS Panama City and Miami laboratories and through NMFS grants to the Archie Carr Center for Sea Turtle Research. Literature cited Abbott, R. T., and P. A. Morris. 1995. A field guide to shells of the Atlantic and Gulf coasts and the West Indies, fourth ed. Houghton Mifflin Co., New York, NY, 350 p. Bender, E. S. 1971. Studies of the life history of the stone crab, Menippe mercenaria (Say), in the Cedar Key area. M.S. thesis, Univ. Florida, Gainesville, FL, 110 p. Schmid. Marine turtle populations on the west-central coast of Florida 601 Bjorndal, K. A., and A. B. Bolten. 1988. Growth rates of immature green turtles, Chelonia mydas, on feeding grounds in the southern Bahamas. Copeia 1988:555-564. Brice, J. J. 1896. The fish and fisheries of the coastal waters of Florida. Rept. U.S. Comm. Fish Fish. 22:263-342. Burke, V. J., S. J. Morreale, and E. A. Standora. 1994. Diet of the Kemp’s ridley sea turtle, Lepidochelys kempii, in New York waters. Fish. Bull. 92:26-32. Caillouet, C. W., Jr., C. T. Fontaine, S. A. Manzella-Tirpak, and D. J. Shaver. 1995. Survival of head-started Kemp’s ridley sea turtles (Lepidochelys kempii) released into the Gulf of Mexico or adjacent bays. Chelonian Conserv. Biol. 1:285-292. Caldwell, D. K., and A. Carr. 1957. Status of the sea turtle fishery in Florida. Trans. 22nd N. Am. Wildl. Conf., p. 457-463. Carr, A. 1952. Handbook of Turtles. Comstock Publ. Assoc., Cornell Univ. Press, Ithaca, NY, 542 p. 1980. Some problems of sea turtle ecology. Am. Zool. 20:489-498. 1986. New perspectives on the pelagic stage of sea turtle development. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFC-190, 36 p. Carr, A., and D. K. Caldwell. 1956. The ecology and migrations of sea turtles: 1. Results of field work in Florida, 1955. Am. Mus. Nov. 1793:1-23. Carr, A., A. Meylan, J. Mortimer, K. Bjorndal, and T. Carr. 1982. Surveys of sea turtle populations and habitats in the western Atlantic. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFC-91-82, 46 p. Carr, A., L. Ogren, and C. McVea. 1980. Apparent hibernation by the Atlantic loggerhead turtle Caretta caretta off Cape Canaveral, Florida. Biol. Conserv. 19:7-14. Dodd, C. K., Jr. 1988. Synopsis of the biological data on the loggerhead sea turtle Caretta caretta (Linnaeus 1758). U.S. Fish Wildl. Serv., Biol. Rep. 88(14), 110 p. Dunham, A. E. 1978. Food availability as a proximate factor influencing individual growth rates in the iguanid lizard Sceloporus merriami . Ecology 59:770-778. Ehrhart, L. M. 1983. Marine turtles of the Indian River Lagoon System. Fla. Sci. 46:337-346. Ehrhart, L., and M. Yoder. 1978. The marine turtles of Merritt Island National Wild- life Refuge, Kennedy Space Center, Florida. In G. E. Henderson (ed. ), Proceedings of the Florida interregional con- ference on sea turtles. Florida Mar. Res. Publ. 33, p. 25-30. Evink, G. L. 1976 Some aspects of the biology of the blue crab, Callinectes sapidus Rathbun, on Florida’s Gulf coast. M.S. thesis, Univ. Florida, Gainesville, FL, 68 p. Fabens, A. J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. Fontaine, C. T., D. B. Revera, T. D. Williams, and C. W. Caillouet Jr. 1993. Detection, verification and decoding of tags and marks in Kemp’s ridley sea turtles, Lepidochelys kempii. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFC-334, 40 p. Frazer, N. B., J. W. Gibbons, and J. L. Greene. 1990. Exploring Faben’s growth interval model with data on a long-lived vertebrate, Trachemys scripta (Reptilia: Testudinata). Copeia 1990:112-118. Henwood, T. A. 1987. Movements and seasonal changes in loggerhead turtle, Caretta caretta , aggregations in the vicinity of Cape Canaveral, Florida ( 1978-84). Biol. Conserv. 40:191-202. Henwood, T. A., and L. H. Ogren. 1987. Distribution and migrations of immature Kemp’s rid- ley turtles (Lepidochelys kempi) and green turtles (Chelo- nia mydas) off Florida, Georgia, and South Carolina. Northeast Gulf Sci. 9:153-159. Hildebrand, H. H. 1982. A historical review of the status of sea t urtle popula- tions in the western Gulf of Mexico. In K. A. Bjorndal (ed.), Biology and conservation of sea turtles, p. 447- 453. Smithson. Inst. Press, Wash., D.C. Hines, J. E. 1988. Program “JOLLY”: user instructions (draft). U.S. Fish and Wildlife Serv., Patuxent Wildlife Res. Center, Laurel, MD. 7 p. Ingle, R. M. 1972. Florida’s sea turtle industry in relation to restrictions imposed in 1971. In Summary of Florida commercial marine landings, Florida. Dep. Natural Resources 1971: 55-62. Ingle, R. M., and F. G. W. Smith. 1949. Sea turtles and the turtle industry of the West Indies, Florida, and the Gulf of Mexico, with annotated biblio- graphy. Univ. Miami Press, Coral Gables, FL, 107 p. Krebs, C. J. 1989. Ecological methodology. Harper and Collins Pub- lishers, New York, NY, 654 p. Magnuson, J. J., K. A. Bjorndal, W. D. DuPaul, G. L. Graham, D. W. Owens, C. H. Peterson, P. C. H. Pritchard, J. I. Richardson, G. E. Saul, and C. W. West. 1990. Decline of the sea turtles: causes and preven- tion. National Academy Press, Washington, D C., 259 p. Marquez M., R. 1994. Synopsis of biological data on the Kemp’s ridley turtle, Lepidochelys kempi (Garman, 1880). U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFSC-343, 91 p. McRae, E. D., Jr. 1950. An ecological study of the xanthid crabs in the Cedar Key area. M.S. thesis, Univ. Florida, Gainesville, FL, 73 p. Mendonqa, M. T. 1981. Comparative growth rates of wild, immature Chelo- nia mydas and Caretta caretta in Florida. J. Herp. 15:447- 451. 1983. Movements and feeding ecology of immature green turtles ( Chelonia mydas) in Mosquito Lagoon, Florida. Copeia 1983:1013-1023. Mendonqa, M. T. and L. M. Ehrhart. 1982. Activity, population size, and structure of immature Chelonia mydas and Caretta caretta in Mosquito Lagoon, Florida. Copeia 1982:161-167. Ogren, L. H. 1989. Distribut ion of juvenile and subadult Kemp’s ridley turtles: results from the 1984-1987 surveys. In C. W. Caillouet Jr. and A. M. Landry Jr. (eds.), Proceedings of the first international symposium on Kemp’s ridley sea turtle biology, conservation and management, p. 116- 123. Texas A&M Univ. (TAMUl-SG-89-105. 602 Fishery Bulletin 96(3), 1998 Ogren, L., and C. McVea Jr. 1982. Apparent hibernation by sea turtles in North Ameri- can waters. In K. A. Bjorndal (ed. ), Biology and conser- vation of sea turtles, p. 127-132. Smithson. Inst. Press, Wash., D.C. Pollock, K. H., J. D. Nichols, C. Browne, and J. E. Hines. 1990. Statistical inference for capture-recapture experi- ments. Wildl. Monogr. 107:1-97. Pritchard, P., P. Bacon, F. Berry, A. Carr, J. Fletemeyer, R. Gallagher, S. Hopkins. R. Lankford, R. Marquez, M., L. Ogren, W. Pringle Jr., H. Reichart, and R. Witham. 1983. Manual of sea turtle research and conservation tech- niques, second ed. K A. Bjorndal and G. H. Balazs (eds. ), Cen- ter for Environmental Education, Washington, D.C., 126 p. Pritchard, P. C. H., and R. Marquez M. 1973. Kemp’s ridley turtle or Atlantic ridley, Lepidochelys kempi. International Union for Conservation of Nature and Natural Resourses (IUCN) Monograph 2:1-30. Renaud, M. R. 1995. Movements and submergence patterns of Kemp’s rid- ley turtles ( Lepidochelys kempii). J. Herpetol. 29:370-374. Rudloe, A., J. Rudloe, and L. Ogren. 1991. Occurrence of immature Kemp’s ridley turtles, Lepi- dochelys kempi , in coastal waters of northwest Florida. Northeast GulfSci. 12:49-53. SAS Institute Inc. 1989. SAS/STAT user’s guide, version 6, fourth ed., vol. 2. SAS Institute Inc., Cary, NC. 846 p. Schmid, J. R. 1994. A GIS model for the analysis of marine turtle habi- tat associations. In K. A. Bjorndal, A. B. Bolten, D. A. Johnson, and P. J. Eliazar (compilers). Proceedings of the fourteenth annual symposium on sea turtle biology and conservation, p. 279-282. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFSC-351. 1995. Marine turtle populations on the east-central coast of Florida: results of tagging studies at Cape Canaveral, Florida, 1986-1991. Fish. Bull. 93:139-151. Schmid, J. R., and L. H. Ogren. 1990. Results of a tagging study at Cedar Key, Florida, with comments on Kemp’s ridley distribution in the southeast- ern U.S. In T. I. Richardson, J. I. Richardson, and M. Donnelly (compilers). Proceedings of the tenth annual work- shop on sea turtle biology and conservation, p. 129-130. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFC-278. 1992. Subadult Kemp’s ridley sea turtles in the southeast- ern U.S.: results of long-term tagging studies. In M. Salmon and J. Wyneken (compilers). Proceedings of the eleventh annual workshop on sea turtle biology and con- servation, p. 102-113. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFSC-32: Schmid, J. R. and W. N. Witzell. 1997. Age and growth of wild Kemp’s ridley turtles (Lepidochelys kempi): cumulative results of tagging stud- ies in Florida. Chelonian Conserv. Biol. 4:532-537. Shaver, D. J. 1991. Feeding ecology of wild and head-started Kemp’s rid- ley sea turtles in South Texas waters. J. Herpetol. 25:327- 334. 1994. Relative abundance, temporal patterns, and growth of sea turtles at the Mansfield Channel, Texas. J. Herpetol. 28:491-497. Thompson, N., T. Henwood, S. Epperly, R. Lohoefner, G. Gitchlag, L. Ogren, J. Mysing, and M. Renaud. 1990. Marine turtle habitat plan. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFC-255, 20 p. True, F. W. 1887. The turtle and terrapin fisheries. In G. B. Goode, The fisheries and fishery industries of the United States, sec. 5, vol. 2, part 19, p. 493-54. U.S. Comm. Fish Fish. U.S. Fish and Wildlife Service and National Marine Fisheries Service. 1992. Recovery plan for the Kemp’s ridley sea turtle ( Lepi- dochelys kempii). National Marine Fisheries Service, St. Petersburg, FL, 40 p. Wilber, D. H. and W. F. Herrnkind. 1986. The fall emigration of stone crabs Menippe mercen- aria (Say) from an intertidal oyster habitat and temper- ature’s effect on locomotory activity. J. Exp. Mar. Biol. Ecol. 102:209-221. Witherington, B. E. and L. M. Ehrhart. 1989. Status and reproductive characteristics of green turtles (Chelonia mydas) nesting in Florida. In L. Ogren, F. Berry, K. Bjorndal, H. Kumpf, R. Mast, G. Medina, H. Reichart, and R. Witham (eds.), Proceedings of the second western Atlantic turtle symposium, p. 351-352. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFC-226. Witzell, W. N. 1994a. The origin, evolution, and demise of the U.S. sea turtle fisheries. Mar. Fish. Rev. 56:8-23. 1994b. The U.S. commercial sea turtle landings. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFSC-350, 130 p. Wolfe, S. H. (ed.). 1990. An ecological characterization of the Florida Springs coast: Pithlachascotee to Waccasassa Rivers. U.S. Dep. Interior Bio. Rep. 90(21), 323 p. Zug, G. R. 1990. Estimates of age and growth in Lepidochelys kempii from skeletochronological data. In T. I. Richardson, J. I. Richardson, and M. Donnelly (compilers), Proceedings of the tenth annual workshop on sea turtle biology and con- servation, p. 285-286. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFC-278. Zug, G. R., and H. Kalb. 1989. Skeletochronological age estimates for juvenile Lepidochelys kempii from Atlantic coast of North America. In S. A. Eckert, K. L. Eckert, and T. H. Richardson (com- pilers), Proceedings of the ninth annual workshop on sea turtle biology and conservation, p. 271-273. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFC-232. 603 Abstract.— Hypoxia in Chesapeake Bay has increased during the past cen- tury, coincident with the disappearance of Atlantic sturgeon spawning stocks. We hypothesized that Atlantic sturgeon young-of-the-year (YOY) might be more susceptible than other estuarine fishes to high temperature and low oxygen conditions, now prevalent in Chesa- peake Bay. Atlantic sturgeon ( 10-70 g) were reared under conditions of hy- poxia (2-3 mg/L dissolved oxygen) and normoxia (6-7 mg/L) at 19°C and 26°C for 10 days. High-temperature hypoxia resulted in lower survival (mean=6.3%) and respiration rate (mean=0.136 mg OgAg-h)). Low-temperature hypoxia re- sulted in a mean survival of 78% and mean respiration rate of 0.212 mg/(g h). Under hypoxia, mean weight-specific growth rate was 1.27%/d, ca. threefold less than growth under normoxia. Tem- perature alone did not significantly af- fect growth rates. When sturgeon were denied access to the surface, growth rates were significantly diminished in both normoxic and hypoxic treatments. At low ambient oxygen levels and high temperature, denial of surface access was fully lethal within 30 hours. We conclude that increased incidence of summertime hypoxia during this cen- tury has degraded sturgeon nursery habitats in Chesapeake Bay. Manuscript accepted 6 October 1997. Fishery Bulletin 96:603-613 ( 1998). Effects of hypoxia and temperature on survival, growth, and respiration of juvenile Atlantic sturgeon, Acipenser oxyrinchus * David H. Secor Troy E. Gunderson Chesapeake Biological Laboratory Center for Environmental and Estuarine Studies The University of Maryland System PO. Box 38, Solomons, Maryland 20688-0038 E-mail address: secor@cbl.umces.edu An economically important popula- tion of Atlantic sturgeon, Acipenser oxyrinchus, once inhabited Chesa- peake Bay. During the late nine- teenth century, Chesapeake Bay supported the second greatest caviar fishery in the eastern United States (Murawski and Pacheco, 1977). In the early 1900s, the population col- lapsed. In Maryland, fishery land- ings declined from 74,500 kg, in 1904, to 320 kg in 1920 (Hildebrand and Schroeder, 1928). Atlantic stur- geon have not recovered in Chesa- peake Bay (Spier and O’Connell, 1996). The last fish legally har- vested in Chesapeake Bay, a mature female, was captured in 1970 from the Potomac River. The spawning population of Atlantic sturgeon may have been extirpated from Chesa- peake Bay (Speir and O’Connell, 1996; Grogan and Boreman* 1). State, federal, academic, and non- profit organizations have begun to mobilize public interest and support for an aquaculture-based restora- tion program for Atlantic sturgeon in Chesapeake Bay. A principal as- sumption for Atlantic sturgeon res- toration is that deleterious conditions that led to the extirpation of the popu- lation (e.g. loss of habitat or over fish- ing) are now abated. Therefore, it is critical to evaluate possible causes for Atlantic sturgeon extirpation from Chesapeake Bay before state and fed- eral agencies move forward with a large-scale aquaculture-based resto- ration program. We hypothesize that increased hypoxia in Chesapeake Bay re- sulted in reduced habitat for Atlan- tic sturgeon and contributed to their decline. During this century, peri- odic increases (albeit small) in At- lantic sturgeon abundances have occurred in the southern and north- ern extent of the species’ range (Murawski and Pacheco, 1977). However, no evidence exists for pe- riodic recoveries of Atlantic stur- geon in Chesapeake Bay. The period of population decline and low abun- dance in Chesapeake Bay corre- sponds to a period of poor water quality, from 1950 to present, caused by increased nutrient load- ing and increased spatial and tem- poral frequency of hypoxia (Officer et al., 1984; Mackiernan, 1987; Jor- dan et al., 1992; Kemp et al., 1992; Cooper and Brush, 1993). * Contribution 3060 of the Chesapeake Bio- logical Laboratory, Center for Environmen- tal Science, University of Maryland, Solo- mons, MD 20688-0038. 1 Grogan, C. S., and J. Boreman 1997. De- termining the probability that historical populations of fish species are now extir- pated. Unpubl. manuscr., U. Mass. Am- herst, MA, 25 p. 604 Fishery Bulletin 96(3), 1 998 The goal of the study was to investigate the effects of dissolved oxygen and temperature on growth, sur- vival, and respiration of juvenile (young-of-the-year) Atlantic sturgeon. High temperatures are known to amplify negative effects of hypoxia on growth and survival of estuarine fishes (Coutant, 1985). Habi- tats in Chesapeake Bay that satisfy both tempera- ture and dissolved oxygen (DO) preferences of At- lantic sturgeon may be limiting during the summer as have been demonstrated for striped bass (Coutant and Benson, 1990; Brandt and Kirsch, 1993). In this study we define hypoxia as oxygen concentrations <4.0 mg/L. Hypoxia has been defined previously as <2 mg/L for Chesapeake Bay (Phil et al., 1991; Harding et al., 1992), an ambient oxygen level that is detrimental to benthic infaunal production. This definition, however, may be too stringent for fishes because oxygen concentrations at this level are of- ten lethal (Brett, 1970; Jordan et al., 1992). A nested-multifactorial experiment was designed to investigate the effects of temperature and dis- solved oxygen on growth, respiration, and survival of young-of-the-year Atlantic sturgeon. During the course of our experiments we observed that fish in hypoxic conditions frequently surfaced, often break- ing the surface of the water with their snout. There- fore we included surface access as a third factor in our investigation — whether this behavior might ben- efit growth and survival under conditions of oxygen stress. grams wet weight (>10 cm total length) and large enough to be handled with little or no stress for scute- clips (see below). Nested growth and survival experiments A nested multivariate experiment (Fig. 1) evaluated survival and growth rates in relation to access to surficial water (sealed or unsealed tank), tempera- ture (~20°C and ~26°C), dissolved oxygen (~3 mg/L or ~7 mg/L), and tank replication. The nested design directed the analysis of variance to occur in hierar- chical order at four levels. The model of the nested design for growth rate was yijklm - L1 + Ti + Pj(i) + Yk(ij) + 0\(,ijk) + £(ijkl)m, where yijklm = V , Ti Pj(i) ~ Y k(ij) - ® l(ijk) ~ £(ijkl)m ~ the growth rate response by indi- vidual juveniles; overall mean; the effect of the ith surface access category; the effect of the yth temperature; the effect of the Mh oxygen level; the effect of the Zth tank (replicate); and the random error component. The model of the nested design for survival rate was sijki = fJ + l'i + fij(i) + Ykdj) + £iujk). Methods Experimental material Juvenile Atlantic sturgeon were obtained from the U.S. Fish and Wildlife Service, Northeast Fishery Center, Lamar, Pennsylvania (Hendrix, 1995). Dur- ing June 1995, Center personnel collected a large female (2.4-m total length) and three male Atlantic sturgeon from the Hudson River near Hyde Park (River km 135). Fish were transported to the Center for artificial spawning and for larval rearing. Lar- vae and early juveniles were reared at the Center at 17°C and 1 ppt salinity. A failure in the water heat- ing system at the Center caused juveniles, age 45 to 60 days after hatching, to experience low water tem- peratures, ca. 10°C. At 60 days after hatching, 500 juveniles (0. 3-2.0 g wet weight) were transported in an oxygenated container to Chesapeake Biological Laboratory and acclimated to 19°C and 2 ppt salin- ity over a 10-d period. Juveniles were reared in six- teen 40-liter tanks and fed Biokyowa©fry feed (700- 2000 pm diameter) ad libitum until they were ca. 5 where Sijk[ = the arcsine-transformed survival rate for each replicate. Statistical significance for factors was accepted at a = 0.05 (type-I sum of squares for type-I error). To remove possible effects due to differences in fish size among experiments and experimental levels, initial fish weight was included in ANOVAs as a covariate. Calculations of variances and significance tests were performed by using PC-SAS, PROC NESTED (SAS, 1982). Survival data were arcsine transformed to meet assumptions of normally distributed error. Four 10-day experiments were conducted under the four combinations of surface access, temperature, and dissolved oxygen, each replicated twice. In the high- temperature hypoxia treatments, high mortality was observed. Because we wished to have greater confi- dence in associating low survival with high-tempera- ture hypoxia, we repeated the treatments with surface access and high temperature (at both low and high DO levels) for a total of four replicates (Fig. 1). Experimental tank dimensions were 78-cm diam- eter, 46-cm height, and 220-liter volume. External Secor and Gunderson: Effects of hypoxia and temperature on Acipenser oxyrinchus 605 No surface Surface Figure 1 Nested experimental design. Replicates are presented as layers. Dissolved oxygen rates (3 or 7 mg/L) were nested within temperature and surface access levels. stand-pipes for flow-through water allowed sturgeon to use the entire bottom surface. During the experi- ments clear plexiglass lids were attached to the tanks with clamps and duct tape. A large hole (5-cm diam- eter) in each lid, fitted with a stopper, permitted ac- cess to the tank for feeding of fish, for periodic checks of wafer quality, and for removal of dead individu- als. Hoses for air and water supply also passed through the lid. In the tanks that permitted access to the air-water interface, water level was maintained at 5 cm below the lid. In treatments designed to limit access to the surface, lids were sealed to the tank and tanks were filled completely. Two head tanks delivered water at either 19°C or 26°C to experimen- tal tanks, each maintained at a rate of 1.2 -1.5 L of flow-through water/min. Dissolved oxygen was con- trolled by maintaining hypoxic conditions in head- tanks and by adding aeration directly to normoxic treatment tanks. Hypoxic levels (2. 5-3.0 mg/L) were maintained in head-tanks by mixing hypoxic well wafer (<2 mg/L) with oxygenated Patuxent estuary water (6 mg/L). Thus, salinity varied slightly between experiments but was always within the range 1.5-3 ppt. When necessary, water was aerated to bring oxygen levels to 2.5 mg/L in the head-tank. Normoxic treatment water was a mixture of well water and Patuxent estuary water. Dissolved oxygen levels were well mixed (homogenous) in the tanks owing to their shallow design, flow-through water, and constant swimming of YOY sturgeon. Lighting was provided on a 12:12 h lightrdark cycle. With the exception of the 26°C and no-surface-ac- cess treatment, temperature and oxygen concentra- tions were maintained within 10% of their prescribed levels (19 or 26 C. 3 or 7 mg/L). In the 26 C sealed experiment, aeration supplied to the head-tank was insufficient to attain oxygen conditions close to 7 mg/L for the normoxic treatment { mean replicate dissolved oxygen concentrations were 5.10 and 5.25 mg/L). In the hypoxic sealed treatment (mean replicate dissolved oxygen concentrations were 3.76 and 4.44 mg/L), dis- solved oxygen was deliberately held above 3 mg/L because this level with high temperature had been observed previously to be lethal in unsealed tanks. Despite these elevated “hypoxic” conditions, the com- bination of high temperature and low oxygen was fully lethal in the sealed tanks (see below). Juveniles (8 to 30 grams wet weight) were accli- mated to experimental conditions over a 4-d period. On day 0 of each experiment, lengths and wet weights (juveniles weighed in water on a top-loading balance) were recorded and a dorsal scute(s) clipped to iden- tify each individual. In preliminary trials, we found that scute clips did not significantly affect growth 606 Fishery Bulletin 96(3), 1 998 rate. Regeneration of the scute did not obscure the clip during the 10-d experiments. Six to eight juve- niles were placed in each experimental tank and fed the formulated diet (2-mm pellets) at 2.5% body weight per day. Fish were fed at 06:00, 10:00, 14:00, 18:00, and 20:00 h. Water quality was checked during feed- ing times and any remaining food or feces were siphoned prior to feeding. Food amounts were adjusted when mortalities occurred. On day 10 of the experiment, in- dividual weights and lengths were recorded. Weight- specific absolute and instantaneous growth rates were determined for the experimental period according to Ricker (1975). In instances where juveniles died before the end of the experiment, individuals surviving >3 days of experimental conditions were included in the analy- sis of treatment effects on growth. peratures. For the hypoxic treatments, oxygen was provided at significantly higher levels (P=0.001) at 26°C (4.09 ±0.07 mg/L) than at 19°C (2.51 ±0.05 mg/L). This was intentional because survival and growth experi- ments had shown that DO levels <3.5 mg/L in con- cert with high temperature were lethal. Despite this precaution, all fish at 26°C and at low oxygen level perished within 24 hours. At the high-level DO treat- ments, oxygen conditions were ca. 1 mg/L higher at 19°C than at 26°C. Total biomass of the 7 fish per tank ranged from 157.6 to 248.2 grams (Table 1). Respiration (i.e. oxygen uptake) was estimated (Cech, 1990) as R = (Me - Mc)/B Me.c = ((C/ ~ C0 ) V Respiration experiment Respiration was estimated over a 42-h period for four combinations of temperature (19 or 26°C) and dis- solved oxygen (-3 or ~7 mg/L) each replicated once. Juveniles were acclimated for a 4-day period and starved 12 h prior to the start of respiration mea- sures. Seven juveniles were weighed (in water) and placed in each tank “respirometer.” The tanks were sealed (no air gap). Oxygen levels were maintained by aerating the head-tanks, rather than each experi- mental tank. There was no feeding during the 42-h experiments. Inflow and outflow temperature, salin- ity, oxygen content, and flow rate were measured in experimental and control “blank” tanks (Cech, 1990) at 06:00, 10:00, 14:00, 18:00 and 22:00. Experimental conditions of temperature were maintained within 1°C of the prescribed treatment level (19 or 26°C) (Table 1). Oxygen levels varied substantially from the prescribed levels between tem- where R M e.c B Ci C'c V weight-specific 02 consumption rate (mg 02/(g-h); 02 consumption rate in experimental E or control C tanks (mg Og/h); combined wet weight (biomass) of stur- geon juveniles (g); 02 concentration in inflowing water ( mg 02/L); 02 concentration in outflowing water (mg 02/L); and water flow rate (L/h). Results Survival Deaths were observed only in hypoxic treatments (Table 2). At hypoxic level, survival was substantially lower at 26°C (mean=6.3% survival) than at 19°C Table 1 Replicate tank environmental conditions and respiration rates (mean ± standard error) for respiration experiment. Dissolved oxygen (DO) levels, low or high, refer to prescribed levels of 3 mg/L and 7 mg/L, respectively. Inflow oxygen and tank temperature refer to actual conditions provided to tanks. Biomass is the total initial weight of the seven sturgeon used in each replicate. Temperature level DO level Inflow oxygen mg/L Tank temperature Biomass (g) Respiration rate mg 02/g h 26°C Low 3.762 + 0.163 25.20 ± 0.04 216.6 0.175 ± 0.042 4.436 ± 0.209 25.34 ± 0.02 181.3 0.103 ± 0.030 High 6.310 ± 0.075 25.54 ± 0.02 188.1 0.245 ± 0.028 6.259 ± 0.088 25.47 ± 0.02 179.6 0.307 ± 0.020 19°C Low 2.536 ± 0.029 19.51 + 0.02 196.3 0.202 ± 0.028 2.495 ± 0.020 19.43 ± 0.02 157.6 0.214 ± 0.022 High 7.361 ± 0.060 19.73 ± 0.06 189.5 0.228 ± 0.028 7.273 ± 0.058 19.67 ± 0.04 249.2 0.207 ± 0.020 Secor and Gunderson: Effects of hypoxia and temperature on Acipenser oxyrinchus 607 Table 2 Replicate tank deaths during the nested survival and growth experiment. Experiments are labeled according to the temperature treatment and whether tanks were sealed or unsealed (unsealed tanks permitted access by sturgeon to surface water). Dissolved oxygen (DO) levels, low and high, refer to prescribed levels of 3 mg/L and 7 mg/L, respectively. Rep. = replicate(s). Experiment DO level Rep. Experimental day Survival (%) 1 2 3 4 5 6 7 8 9 10 26°C unsealed 1 Low 1 6 1 1 0 2 2 4 2 0 3 4 1 2 1 0 4 1 1 1 1 50 26°C unsealed 2 High 1 100 2 100 3 100 4 100 26°C sealed Low 1 8 0 2 7 1 0 High 1 100 2 100 19°C unsealed Low 1 1 1 75 2 1 1 75 High 1 100 2 100 19°C sealed Low 1 1 88 2 1 1 75 High 1 100 2 100 (mean=78.3% survival). No significant difference was found in overall survival rate between sealed and unsealed tanks (P=0.54). In unsealed tanks, deaths were distributed throughout the 10-d experimental period. In the 26°C sealed-hypoxic level tanks, all individuals succumbed within the first 30 hours of the experiment. Moribund sturgeon were observed at the air-water interface in unsealed tanks, or just below the lid in sealed tanks. Fin margins of dead individuals were perfused with blood, an indicator of oxygen deprivation (Jobling, 1995). Growth Across replicates, growth rates ranged in weight from 0.3% to 5.1% per day (Table 3). Sturgeon experienced positive growth in weight and length under all ex- perimental conditions. Initial mean weights and lengths varied substantially among experiments; range was 10.94 to 69.20 g in weight and 14.59 to 26.60 cm in length. Absolute growth in weight was positively related to initial size (regression analysis; P<0.01); tanks with fish having initial mean weights greater than 50 g showed the highest absolute growth rates (>1.0 mg/d). Because initial size covaried with growth rate, initial weight was used as a covariate in statistical analyses of treatment effects. Analysis of variance of instantaneous growth rate showed significant effects due to surface access and oxygen level (Table 4); temperature, replicate, and individual fish did not explain significant variance, although at 7 mg/L DO there was a trend for lower growth rates at 26°C than at 19°C (Fig. 2, Table 3). Mean growth rates were 2.9 times less at 3 mg/L (1.27%/d) than at 7 mg/L (3.62%/d). Sealed tanks showed consistently lower growth rates than those with surface access. The effect was greatest at 3 mg/L oxygen. Only 0.3%/d weight-specific growth was ob- served in the 19°C sealed tanks, and all fish perished before growth determinations could be made in the 26°C sealed tanks. Absolute growth in weight was significantly affected by surface access, temperature, surface access in combination with temperature in- teraction, and by oxygen level. Absolute growth rate was higher when surface access was allowed and at the higher DO level; absolute growth rate was in- versely related to temperature. Respiration Respiration rates measured over the entire experi- ment ranged from 0.05 to 0.55 mg 0,>/(g-h) (Fig. 3). Respiration rates were normally distributed for the high oxygen treatment with a mean of 0.245 ± 0.013 608 Fishery Bulletin 96(3), 1998 Table 3 Growth rates (mean ± standard error) of Atlantic sturgeon during the laboratory experiment. Growth rates in weight and length are instantaneous rates. Absolute growth rates in weight is presented under the heading mg/d. Experiments are labeled accord- ing to the temperature level and whether tanks were sealed or unsealed (unsealed tanks permitted access by sturgeon to surface water). Dissolved oxygen (DO) levels, low and high, refer to prescribed levels of 3 mg/L and 7 mg/L, respectively, n = number of individuals in each tank. Rep. = replicate(s). Experiment DO level Rep. Growth rate in weight (per day) mg/d Growth rate in length (per day) n 26°C unsealed Low 1 0.023 ± .006 0.345 ± 0.094 ~0 2 2 0.013 ± 0.003 0.202 + 0.056 ~0 6 3 0.006 ± 0.005 0.093 ± 0.062 0.003 ± 0.001 4 4 0.029 ± 0.019 0.324 ± 0.192 0.003 ± 0.001 8 High 1 0.037 ± 0.004 0.524 ±0.076 0.010 ± 0.001 8 2 0.037 ± 0.006 0.643 + 0.116 0.012 + 0.002 8 3 0.036 ± 0.007 0.587 + 0.112 0.012 ± 0.002 8 4 0.036 ± 0.004 0.789 ± 0.117 0.015 ± 0.001 8 26°C sealed Low 1 Lethal 0 2 Lethal 0 High 1 0.029 ± 0.001 2.220 ± 0.059 0.006+ 0.001 6 2 0.025 ± 0.001 2.015 ± 0.178 0.007 ± 0.001 6 19°C unsealed Low 1 0.013 ± 0.002 0.185 ± 0.029 0.005 ± 0.001 7 2 0.011 ± 0.003 0.129+ 0.037 0.005+ 0.001 7 High 1 0.050 ± 0.005 0.721 ± 0.057 0.014+0.001 8 2 0.045 ± 0.002 0.719+0.044 0.015+0.001 8 19°C sealed Low 1 0.003 ± 0.001 0.072+ 0.030 0.004 ± 0.001 8 2 0.003 ± 0.001 0.071 ± 0.031 0.004+ 0.001 8 High 1 0.028 ± 0.003 1.887 ± 0.253 0.009 + 0.001 6 2 0.031 ± 0.002 2.270 + 0.197 0.008 ± 0.001 6 Table 4 Nested analysis of variance of surface access, temperature, oxygen level, replicate and individual effects on growth rate. Initial weight was used as a covariate in the analyses. Sealed refers to whether tanks were sealed or unsealed (unsealed tanks permitted access by sturgeon to surface water). Designated temperature and oxygen levels were 19°C and 26°C, and 3 mg/L and 7 mg/L, respectively. Factors in parentheses indicate the nesting procedure. For example “Oxygen (Sealed-Temperature)” refers to vari- ance explained by oxygen level nested within combinations of surface access and temperature. Type of Sum of Significance variable Variable df squares level ( P ) Instantaneous growth rate in weight Covariate Initial weight 1 0.3892 0.0001 Class variables Sealed 1 0.0392 0.019 Temperature (sealed) 2 0.0086 0.54 Oxygen (sealed-temperature) 3 0.0981 0.0039 Tank (sealed-temperature-oxygen) 7 0.0097 0.99 Individual (sealed-temperature-oxygen-tank) 14 0.0077 0.99 Absolute growth rate in weight Covariate Initial weight 1 0.0839 0.0001 Class variables Sealed 1 0.0445 0.0001 Temperature (sealed) 2 0.0523 0.0002 Oxygen (sealed-temperature) 3 0.7734 0.0001 Tank (sealed-temperature-oxygen) 7 0.0293 0.17 Individual (sealed-temperature-oxygen-tank) 14 0.0194 0.92 Secor and Gunderson: Effects of hypoxia and temperature on Acipenser oxyrinchus 609 3 mg/L 7 mg/L 3 mg/L 7 mg/L Figure 2 Effects of dissolved oxygen concentration on instantaneous growth rates in weight (per day) (mean ± 95% confidence interval) of young-of-the-year Atlantic sturgeon at two temperatures and different tank configurations. Stippled bars indicate treatments denying surface access. Replicate tanks were combined for each treatment level comination. The asterisk indi- cates that complete mortality occurred for that treatment. mg Og/fg-h). In relation to the high oxy- gen treatment, hypoxic oxygen treatment respiration rates were skewed towards lower rates with a mean of 0.174 ±0.016 mg OgAg-h). Two individual tank respi- ration rates for the hypoxic treatment were exceptionally high (Fig. 3). These data were measured at 0 hours from tanks at 19°C and 26°C and may have reflected an insufficiently long period of acclimation prior to the experiments. Therefore, we chose to exclude measures taken at 0 hours in the analysis of vari- ance on respiration rates. Respiration rates were significantly influenced by oxygen level CP=0.001), by the interactions between temperature and oxygen (P-0.04), and by the inter- action among temperature, oxygen, and replicate (P=0.04). Mean respiration rates were 0.187 ± 0.012 and 0.247 ± 0.015 mg O^lg-h) under hypoxia and normoxia, respectively (Table 1). At high oxygen levels, 26°C respiration rates (mean=0.281 ±0.023 mg 0I?/(g-h)) tended to be higher than 19°C respira- tion rates (mean-0.210 ±0.021 mg O ^ (g h)). However, the converse was true at hypoxic levels; mean respiration rate at 26°C (0.136 ±0.027 mg O^g-h)) was significantly lower than at 19°C (0.212 ±0.021 mg OgAg-h)). For all but the 26°C and hypoxic-level com- bination, replicate rates were similar. Respiration for the second replicate for the 26°C and hypoxic-level combination was substantially lower than other rep- licates that may have produced the significant in- teraction among temperature, oxygen, and replicate factors in the analysis of variance. In a procedure to reduce bias associated with deviations of actual in- flow DO levels from those prescribed, an analysis of variance was conducted for which the inflow oxygen concentration was a covariate. After viariance asso- ciated with individual tank DO conditions was re- moved, the effect of oxygen level (high or low) re- mained significant (P=0.04). Discussion Juvenile Atlantic sturgeon were vulnerable to con- ditions of high temperature and low oxygen. In five out of six replicates (sealed and unsealed tanks com- bined) at 26°C and -3 mg/L DO, all juveniles died. All sturgeon that died showed a perfusion of blood along the margins of their fins, indicative of oxygen deprivation (Jobling, 1995). Reduced oxygen levels resulted in a threefold reduction in growth rate and a 50% reduction in routine respiration rate. At 19°C, respiration rates were similar between hypoxic and normoxic treatments. But, at the 26°C hypoxic treat- ment, mean respiration rates dropped below 2 mg 09/(g h) and all sturgeon died. We speculate that these fish were unable to supply sufficient oxygen to their tissues at this level of reduced respiration. Despite reduced survival and respiration in condi- tions of low dissolved oxygen, feeding continued and fish grew. Apparently, Atlantic sturgeon were able to reduce activity but still feed and allocate some en- ergy to growth. In unsealed tanks, weight gain ranged from 1.1% to 2.9%, and from 3.6% to 5.0% body weight per day, at ~3 and ~7 mg O./L, respec- tively. Cech et al. (1984) also observed continued growth by juvenile white sturgeon ( Acipenser transmontanus) (ca. 0.5 to 5 g) under conditions of hypoxia. Daily weight-specific growth rates of white sturgeon varied between 1.6% ( 15°C) and 2.9% (25°C) under normoxic conditions and between 0.6% (15°C) and 2.3% (25°C) under hypoxic conditions. Growth rates measured by Cech et al. (1984) under hypoxia were substantially higher than those that we oh- 610 Fishery Bulletin 96(3), 1998 High oxygen treatments Low oxygen treatments Routine respiration rate (mg OJ[q • h) Figure 3 Frequency histogram of respiration rates for young-of-the-year Atlantic sturgeon exposed to either high (7 mg/L) or low levels (3 mg/L) of dissolved oxygen. Mea- sures taken at the initiation of the experiment (hour 0) are stippled. served. This may be attributable to a higher designated oxygen level for hypoxic treatments (<5 mg/L), differences between the two spe- cies, or an ontogenetic effect. Ju- venile white sturgeon were substan- tially smaller (initial weight 0.5 grams) than those used in our study (mean initial weight=23.7 grams). Our study was unique in exam- ining the effects of long-term hy- poxia on routine metabolism. Other studies have examined the effects of hypoxia on sturgeon res- piration in short-term respirom- etry studies (Table 5). Investiga- tions on white sturgeon and Sibe- rian sturgeon ( Acipenser baeri) have indicated reduced rates of respiration under hypoxic condi- tions. An ontogenetic trend of de- creasing routine metabolic rates with increased mass, which is typi- cal in fishes, was also suggested in the comparison of studies. Meta- bolic rates under normoxic condi- tions ranged from 0.9 mg d/lg-h) (1-2 g fish) to 0.055 mg O^fg-h) (1800-g fish). Metabolic rates (0.2 Table 5 Summary of studies on hypoxia and routine metabolism for sturgeons. Oxygen concentrations and consumption rates have been converted to common units (mg OJh or mg O^g-h) from reported units (e.g. mm Hg or pmol Oj/kglg h) for several of the studies. Wt = wet weight; Hyp = hypoxia treatment; Norm = normoxia treatment; Swim = velocities were provided in respirometer to induce swimming velocities; S = salinity; and T = temperature. Species Wt (g) Treatments Routine metabolism mg/(g-h) S (ppt) T (°C) Study duration Reference Acipenser transmontanus 0.5-5 Hyp: 4.7 - 5.7 mg/L Norm: 6.8 - 8.2 mg/L Not measured 0 15, 20, 25 30 d Cech et al., 1984 A. baeri 1-2 High density Low density 0.3 to 0.5 0.4 to 0.9 0 22-24 2-12 h Khakimullin, 1987 A. baeri 3-7 Routine Swim: 5-30 cm/sec 0.3 to 0.7 0.4 to 3.6 0 22-24 3 h Khakimullin, 1988 A. oxyrinchus 12-69 Hyp: 2. 5-4. 4 mg/L Norm: 6. 3-7. 4 mg/L 0.1 to 0.2 0.2 to 0.3 1.8-2. 5 19, 26 10 d This study A. transmontanus 950 Hyp: 1.7-6. 3 mg/L Norm: 9.8 mg/L ~0 to 0.015 0.079 0 15 4.5 h Burggren and Randall, 1978 A. baeri 1800 Hyp: 1.3-3. 8 mg/L Norm: 8.2 mg/L 0.023 0.055 0 15 3.5 h Nonnotte et ah, 1993 A. transmontanus 2000 Hyp: 4. 7-5. 9 mg/L Norm: 7.7-8.9mg/L 0.090 0.098 0 0 18 24 h Ruer et ah, 1987 Secor and Gunderson: Effects of hypoxia and temperature on Acipenser oxyrinchus 61 1 to 0.3 mg Og/Cg-h) and fish sizes (12 to 69 g) in our ex- periments were both intermediate within this range. Surface behavior In the nested growth and survival experiment, sur- face access influenced both growth and survival rates. Eliminating surface accesses in tanks reduced growth rates by ca. 35%, and fivefold at high and low levels of DO, respectively. The effects of denying surface access under “hypoxia” at 26°C was fully lethal within a 30-h period. In the 26°C hypoxic treatments that allowed surface access, the majority of juveniles sur- vived the first 5 days of exposure, although they too eventually died. Many fishes surface in hypoxic environments to convey relatively oxygen-rich water, located at the air-water interface, across their gills. In laboratory experiment, access to surface waters may have in- creased the effective level of DO above nominal lev- els, resulting in improved growth and survival. Al- ternatively, aerial respiration cannot be ruled out for sturgeon that are physostomous. Histological stud- ies should be undertaken to investigate whether the swim bladder of Atlantic sturgeon contains a vascu- lar structure, apart from the gas gland, which meets the criteria for an aerial respiratory organ. No such structure has been identified in any sturgeon species. Reasons for decline of Atlantic sturgeon in Chesapeake Bay We conclude that increased frequency of hypoxia in Chesapeake Bay during this century (Officer et ah, 1984; Cooper and Brush, 1993) was detrimental to Atlantic sturgeon production. Recent water quality monitoring has shown that during summer months (mid-June through mid-September), temperatures >25°C and DO levels <4.0 mg OJh are prevalent in Chesapeake Bay benthic habitats (Breitburg, 1990, 1992; Phil et ah, 1991). Our laboratory experiments showed that juvenile Atlantic sturgeon were less tol- erant of summertime hypoxia than were other juve- nile estuarine species. Young-of-the-year spot, Leiostomus xanthurus (total length 10-20 cm), sur- vived long-term (>1 week) experimental exposure of 2. 4-3.0 mg/L at 25°C, but 0. 8-1.0 mg/L DO was fully lethal (Phil et ah, 1991). Juvenile and adult hog- chokers, Trinectes maculatus (Phil et ah, 1991), and naked gobies, Gobiosoma bosc (Brietburg, 1992), can tolerate several-day periods of 0. 5-1.0 mg/L DO. Our laboratory experiments did not consider be- haviors that can 1) reduce exposure to hypoxic wa- ters and 2) compensate for reduced dissolved oxygen levels. Phil et ah (1991) and Brietburg (1992) have provided field evidence that fish will escape hypoxic conditions through local migrations. These include vertical or shoalward emigrations from hypoxic or anoxic bottom habitats. Following hypoxic events, bottom habitats are recolonized. Further, short-term episodic hypoxia may benefit bottom-feeding fish. Burrowing macrobenthic prey will emerge at DO lev- els <2 mg/L, increasing their vulnerability to preda- tion by fish that can tolerate short-term excursions into hypoxic waters (Phil et ah, 1992). If unable to escape hypoxic conditions, sturgeon may be able to compensate by either surfacing to exploit higher oxy- gen concentrations in surficial water or in the atmo- sphere or by adjusting their metabolic rate (e.g. through reduced swimming [Cech et ah, 19843). Hudson River “strain” Atlantic sturgeon, used in our experiment, might have exhibited a different response to hypoxia than a strain native to Chesa- peake Bay. The Hudson River rarely becomes hypoxic (Cooper et ah, 1988). Therefore, Hudson River At- lantic sturgeon may not have been adapted to hy- poxic conditions. An aquaculture study by Serov et ah (1988) on stellate sturgeon (A. steliatus) showed that heterozygosity in the LDH gene conferred sur- vival advantages to hypoxia and high temperature. Therefore, it is conceivable that selection of Chesa- peake Bay Atlantic sturgeon to hypoxic conditions could have occurred over several generations. How- ever, because generation time is extremely high in Atlantic sturgeon (c.a. 29 years [Stevenson and Secor, 1996]) and because hypoxia increased rapidly dur- ing this century in the Chesapeake Bay, Chesapeake Bay Atlantic sturgeon may not have been able to re- coup historical abundances by dint of selection to low- oxygen conditions. In addition, Hudson River Atlan- tic sturgeon juveniles >80 cm TL are known to visit Chesapeake Bay during summer months (Dovei and Berggren, 1983). Presumably, these fish could have adapted to Chesapeake Bay conditions. Serov et al.’s (1988) observation that water quality experienced by juveniles in culture influences genotypic frequencies has important implications for the use of hatchery- produced sturgeon in restoration programs and mer- its additional research. Scientists and managers are now considering a restoration program for Atlantic sturgeon in Chesa- peake Bay and elsewhere (St. Pierre, 1994; Secor, 1995). The feasibility of a sturgeon restoration pro- gram must address the same issues that led to the sturgeon’s decline. If these conditions persist in Chesapeake Bay, a restoration program cannot be easily justified. Necessary conditions for population recovery must include increased population abun- dance, and improvement in the quality and size and number of essential habitats. Population abundance 612 Fishery Bulletin 96(3), 1 998 can be increased by deliberately releasing artificially produced progeny (parentage from an outside popu- lation like the Hudson River population), by impos- ing a moratorium on sturgeon harvests, and by mak- ing efforts to reduce sturgeon taken as bycatch in other fisheries. A critical and unresolved question is whether habi- tat quality remains sufficient at present in the Chesa- peake Bay to support Atlantic sturgeon growth, sur- vival, and reproduction. An encouraging finding has been a trend in improved water quality and macrobenthic production in Chesapeake Bay tribu- tary nursery habitats, the apparent result of nutri- ent abatement programs (Dauer, 1995). Acknowledgments This research was supported by Maryland Depart- ment of Natural Resources Chesapeake Bay Research and Monitoring Division through funds made avail- able by the National Marine Fisheries Service (Anadromous Fish and Great Lakes Conservation Act). We are especially grateful to Mike Hendrix and Jerre Mohler at U.S. Fish and Wildlife Service (Northeast Fishery Center, Lamar PA) for their as- sistance in providing hatchery-produced young-of-year Atlantic sturgeon used in our experiments. Jill Stevenson and Erik Zlokovitz (Chesapeake Biological Laboratory) provided assistance in the laboratory. We thank Ed Houde, Ron Klauda, and Paul Miller for their comments on an earlier draft of this manuscript. Letty Fernandez assisted with preparation of this report. Literature cited Brandt, S. B., and J. Kirsch. 1993. Spatially explicit models of striped bass growth po- tential in Chesapeake Bay. Trans. Am. Fish. Soc. 122:845- 869. Breitburg, D. L. 1990. Near-shore hypoxia in the Chesapeake Bay: patterns and relationships among physical factors. Estuarine Coastal Shelf Sci. 30:593-609. 1992. Episodic hypoxia in Chesapeake Bay: interacting ef- fects of recruitment, behavior, and physical disturbances. Ecol. Monogr. 62:525-546. Brett, J. R. 1970. The metabolic demand for oxygen in fish, particularly salmonids, with a comparison with other vertebrates. Respir. Physiol. 14:151-170. Berggren, W. W., and D. J. Randall. 1978. Oxygen uptake and transport during hypoxic expo- sure in the sturgeon Acipenser transmontanus. Respir. Physiol. 34:171-183. Cech, J. J. 1990. Respirometry. In C. B. Schreck and P. B. Moyle (eds.), Methods for fish biology, p. 335-362. Am. Fish. Soc., Bethesda, MD. Cech, J. J., S. J. Mitchell, and T. E. Wragg. 1984. Comparative growth of juvenile white sturgeon and striped bass: effects of temperature and hypoxia. Estu- aries 7:12-18. Cooper, J. C., F. R. Cantelno, and C. E. Newton. 1988. Overview of Hudson River Estuary. Am. Fish. Soc. Monogr. 4:11-24. Cooper, S. R., and G. S. Brush. 1993. A 2,500-year history of anoxia and eutrophication in Chesapeake Bay. Estuaries 16:617-626. Coutant, C. C. 1985. Striped bass, temperature, and dissolved oxygen: a speculative hypothesis for environmental risk. Trans. Am. Fish. Soc. 114:31-61. Coutant, C. C., and D. L. Benson. 1990. Summer decline in habitat suitability for striped bass in Chesapeake Bay: reflections on a population decline. Trans. Am. Fish. Soc. 119:757-778. Dauer, D. M. 1995. Long-term trends in macrobenthos of the lower Chesapeake Bay (1985-92). In P. Hill and S. Nelson (eds.), Toward a sustainable watershed: the Chesapeake experi- ment. Chesapeake Research Consortium Publ. 149, Edge- water, MD. Dovel, W. L., and T. J. Berggren. (1983). Atlantic sturgeon of the Hudson Estuary, New York. NY Fish and Game J. 30:140-172. Harding, L. W., M. Leffler, and G. E. Mackiernan. 1992. Dissolved oxygen in the Chesapeake Bay: a scien- tific consensus. Maryland Sea Grant, College Park, MD, 18 p. Hendrix, M. A. 1995. Update — U.S. Fish and Wildlife Service and Atlantic sturgeon. Sturgeon Notes 3:10-11. Hildebrand, S. F., and W. C. Schroeder. 1928. Fishes of the Chesapeake Bay. U.S. Bureau of Fish- eries, Washington, D.C., 388 p. Jobling, M. 1995. Environmental biology of fishes. Chapman and Hall, New York, NY, 455 p. Jordan, S., C. Stenger, M. Olson, R. Batiuk, and K. Mountford. 1992. Chesapeake Bay dissolved oxygen goal for restora- tion of living resource habitats. Maryland Department of Natural Resources, Annapolis, MD, 81 p. Kemp, W. M., P. A. Sampou, J. Garber, J. Tuttle, and W. R. Boynton. 1992. Seasonal depletion of oxygen form bottom waters of Chesapeake Bay: roles of benthic and planktonic respira- tion and physical exchange processes. Mar. Ecol. Prog. Ser. 85:137-152. Khakimullin, A. A. 1987. Intensity of gas exchange in individuals and groups of juvenile Siberian sturgeon, Acipenser baeri. Vop. Ikht. 6:978-983. 1988. Intensity of gas exchange of hatchery Siberian stur- geon, Acipenser baeri, under muscular load. Vop. Ikht. 1988:282-288. Mackiernan, G. B. 1987. Dissolved oxygen in the Chesapeake Bay: processes and effects. Maryland Sea Grant, College Park, MD, 177 p. Murawski, S. A., and A. L. Pacheco. 1977. Biological and fisheries data on Atlantic sturgeon, Secor and Gunderson: Effects of hypoxia and temperature on Acipenser oxyrinchus 613 Acipenser oxyrhynchus (Mitchill). National Marine Fish- eries Service, NOAA, Highlands, NJ, 69 p. Nonnotte, G,, V. Maxime, J. P. Truchot, P. Williot, and C. Peyraud. 1993. Respiratory responses to progressive ambient hypoxia in the sturgeon, Acipenser baeri. Respir. Physiol. 91:71-82. Officer, C. B,, B. B. Biggs, J. L. Taft, L. E. Cronin, M. A. Tyler, and W. R. Boynton. 1984. Chesapeake Bay anoxia: origin, development, and significance. Science 223:22-27. Phil, S. P. Baden, and R. J. Diaz. 1991. Effects of periodic hypoxia on distribution of demer- sal fish and crustaceans. Mar. Biol. 108:349-360. Phil, L., S. P. Baden, R. J. Diaz, and L. C. Schaffner. 1992. Hypoxia-induced structural changes in the diet of bot- tom-feeding fish and crustaceans. Mar. Biol. 112:349-361. Ricker, W. E. 1975. Computation and interpretation of biological statis- tics of fish populations. Canada Fisheries and Oceans, Ottawa, Ontario, 382 p. Ruer, P. M., J. J. Cech and S. I. Doroshov. 1987. Routine metabolism of the white sturgeon, Acipenser transmontanus : effect of population density and hypoxia. Aquaculture 62:45-52. SAS. 1982. Statistical analysis system user guide: statistics. SAS Institute, Inc., Cary, NC, 585 p. Secor, D. H. 1995. Chesapeake Bay Atlantic sturgeon: current status and future recovery. Chesapeake Biological Laboratory, Solomons, MD, 10 p. Serov, D. V., S. I. Kikonorov, and E. S. Asadullaeva. 1988. Genetic selection in young stellate sturgeon. Rybn. Khov. 4: 64-65. Speir, H., and T. O’Connell. 1996. Status of Atlantic sturgeon in Maryland’s Chesapeake Bay. MD Dep. Natur. Resources, Annapolis, MD, 7 p. St. Pierre, R. A. 1994. The national strategy for the management and con- servation of paddlefish and sturgeon species in the United States. In V. J. Birstein ( ed. ), The international confer- ence on sturgeon biodiversity and conservation. Hudson River Foundation, New York. Stevenson, J. T., and D. H. Secor. 1996. Demographic analysis of Atlantic sturgeon fisheries in the New York Bight. Chesapeake Biological Labora- tory, Solomons, MD, 17 p. 614 Abstract.-striped bass, Morone saxatihs, in the Coos River, Oregon, are derived from natural colonists from San Francisco Bay, which in turn were in- tentionally transplanted from the Hudson River. Because of founder ef- fects, this unusually well-documented colonization sequence should have re- sulted in diminished genetic variabil- ity in the penultimate and ultimate populations, which may have been fur- ther compounded in the Coos River population by subsequent drastic re- ductions in its abundance. To test whether these sequential bottlenecks reduced genetic diversity we surveyed both nuclear DNA (nDNA) and mito- chondrial DNA (mtDNA) variation in the Coos River population and in both populations along the historical path- way that led to its founding. There was no evidence of reduced nDNA diversity among these populations at the three loci examined. However, the number of mtDNA haplotypes revealed decreased from 8 in the original Hudson River population, to 5 in the San Francisco Bay population, to only 1 in the Coos River population. This pattern of con- served nDNA diversity and reduced mtDNA diversity is consistent with a recent population bottleneck. Coos River striped bass have shown increas- ing levels of pathological hermaphro- ditism. We speculate that the reduced genetic diversity of the Coos River striped bass population may have led to a depensatory cascade involving her- maphroditism that inhibited reproduc- tion and recruitment, followed by in- creased levels of inbreeding as the population declined. Manuscript accepted 20 October 1997. Fishery Bulletin 96:614-620 (1998). Multiple population bottlenecks and DNA diversity in populations of wild striped bass, Morone saxatilis John R. Waldman Hudson River Foundation for Science and Environmental Research 40 West 20th Street, New York, New York, 10011 E-mail address: john@hudsonriver.org Reese E. Bender Oregon Department of Fish and Wildlife 4475 Boat Basin Drive, Charleston, Oregon 97420 Isaac I. Wirgin Institute of Environmental Medicine New York University Medical Center Long Meadow Road, Tuxedo, New York 10987 Population bottlenecks are often invoked to explain lower than ex- pected levels of genetic diversity in wild populations of fishes (e.g. Bernatchez et al., 1989; Brown et al., 1992; Richardson and Gold, 1997), but rarely is there detailed information available on the degree and duration of the bottlenecks. Striped bass ( Morone saxatilis ) of- fer an exception because sequentially established populations (Fig. 1) in historical times have experi- enced unusually well documented bottlenecks. Striped bass were introduced to the Pacific coast at San Francisco Bay in 1879 and 1882 (Stevens et al., 1987). The two plantings totaled approximately 430 individuals ( 132 in 1879; approximately 300 in 1882). All were yearlings collected in the Navesink and Shrewsbury rivers, New Jersey. The Navesink and Shrewsbury rivers are minor systems that do not support repro- duction by striped bass and that are located near the mouth of the Hudson River; the transplants were almost certainly part of the proxi- mal Hudson River striped bass population. The transplanted year- lings rapidly established a popula- tion in San Francisco Bay which reproduced in its two main tributar- ies: the Sacramento and San Joaquin rivers. The introduction of striped bass to San Francisco Bay has been viewed as one of the few highly suc- cessful introductions of non-native fishes (Raney, 1952); within 10 years of the original introduction striped bass were available in commercial quantities in California waters. The first striped bass captured in Oregon waters were two adults taken in Coos Bay in 1914 (Morgan and Gerlach1 ). These fish were va- grants (or less likely, the offspring of vagrants) from the San Francisco Bay population, the only possible source along the Pacific coast. Since 1914, reproducing populations of striped bass became established in Oregon in the Coos, Coquille, Ump- qua, Smith, and Siuslaw estuaries. Of these, the Coos River population was the largest and most studied. 1 Morgan, A. R., and A. R. Gerlach. 1950. Striped bass studies on Coos Bay, Oregon in 1949 and 1950. Oregon Fish Commis- sion Contribution 14, 31 p. Waldman et a!.: Population bottlenecks and DNA diversity in Morone saxatilis 615 By the mid-1920s, striped bass were being commer- cially harvested in Coos Bay, and in 1945, annual landings from Coos Bay reached a high of 231,000 lb. The adult population also appeared to peak in 1945 at about 69,000 individuals (>age-3), based on catch- per-unit-of-effort sampling. Pathological hermaph- roditism was noted, but it was rare among Coos River striped bass during this period; Morgan and Gerlach1 reported a 3% (n=124) incidence in 1950. Since 1945, the Coos River striped bass popula- tion has crashed, whereas the incidence of hermaph- roditism has increased dramatically. Between 1950 and 1975, population estimates of adults ranged from as many as 43,000 in 1963 to as few as 7800 in 1973. No adult population estimate is available for 1980, but in that year Moser et al. (1983) found 11 of 42 (26%) wild fish to be hermaphrodites. Population size of Coos River adult striped bass was not evaluated again until 1988 and 1989, when estimates of be- tween 1000 and 3000 for both years were obtained. Estimates to date for the 1990s are of an adult popu- lation size under 1000. Furthermore, virtually no natural recruitment appears to be occurring (but supplemention is occurring through stocking of hatchery-produced offspring of San Francisco Bay broodstock). A standardized seine-haul survey of ju- venile production begun in 1978 showed a decline in catch-per-unit-of-effort from 2.9 in 1978 to between 0.3 and 0.1 until 1986, and then only infinitesimal levels or zero through 1995. Recent estimates are that hermaphrodites make up 30% of the naturally pro- duced adult population in the Coos River system (Reimers et al.2 ). Additionally, during 1993, an angling 2 Reimers, P. E., R. E. Bender, J. A. Johnson, T. Rumreich, D. J. Van Dyke, T. A. Confer, R. C. Smith, J. A. Hutado, and R. S. Boots. 1990. Tenmile-Coos-Coquille Fish District: a review of stocks of concern. State conservation department report, Oregon Department of Fish and Wildlife, 41 p. guide captured a hermaphroditic striped bass from the Umpqua River, the first reported from that system. We hypothesized that Coos River striped bass would show reduced genetic diversity because both the history of the population’s establishment and its subsequent demographics favored inbreeding. To test this hypothesis, we surveyed both nuclear DNA (nBNA) and mitochondrial DNA (mtBNA) variation in the Coos River population, the similarly non-na- tive San Francisco Bay population (the source of the Oregon populations), and the source for the San Fran- cisco Bay population, the Hudson River, New York. We also examined mtBNA variation in a second Or- egon population (Umpqua River). Methods Coos River striped bass (both wild fish and fish that were originally hatchery-cultured) were captured in gill nets during spring 1993. These fish were distin- guished on the basis of size; significant stocking of hatchery-cultured striped bass (age-0 only) did not begin until 1989. Only wild Oregon striped bass were used as broodstock until 1991, when broodstock from California were used. Umpqua River specimens were collected in 1992 by angling. San Francisco Bay samples were collected by means of gillnetting in the lower Sacramento and San Joaquin rivers during 1991 and 1992. Collections of striped bass from the Hudson River are described in Wirgin et al. (1990, 1993). Total BNA was isolated from livers or blood by the CTAB method (Saghai-Maroof et al., 1984; Wirgin et al., 1990), phenol-chloroform extractions, and etha- nol precipitations. To determine mtBNA haplotypes, BNAs were digested with Acc I, Hind III, and Rsa I, electrophoretically separated in 1.2% agarose gels, and visualized in Southern blot analyses by using 32P 616 Fishery Bulletin 96(3), 1 998 Table 1 Genotypic frequencies for three single copy, nuclear DNA loci. Heterozygosity values in parentheses. Locus 25 Dra I Locus 27 EcoR V Locus 22 Hinf I Population N AA AB BB N AA AB BB N AA AB BB Hudson River 35 0 11 24 123 6 42 75 80 47 24 9 (0.314) (0.342) (0.300) San Francisco Bay 64 6 23 35 49 3 22 24 65 41 20 4 (0.359) (0.449) (0.308) Coos River (wild) 27 1 12 14 32 3 14 15 34 11 9 14 (0.444) (0.438) (0.265) Coos River (hatchery) 23 1 11 11 25 6 7 12 19 6 5 8 (0.478) (0.280) (0.263) radiolabelled DNA probes (Feinberg and Vogelstein, 1983). Each of the three enzymes generates a diag- nostic fragment which was used to characterize mtDNA major length-variant haplotypes (defined as differences of more than 100 base pairs; Wirgin et al., 1990). In addition, Acc I digestion revealed an in- formative, single base substitution. Probes were either highly purified mtDNA isolated from a single striped bass liver (Wirgin et al., 1990) or a gel-purified 1.7 kb PCR product containing the striped bass mtDNA con- trol region (Wirgin et al., 1995). To determine nDNA genotypes, DNAs (10 pg) were digested with Dra I, EcoRV, and Hinf I, and analyzed in Southern blot analysis with the single copy probes developed from a striped bass genomic DNA library: DSB 25, DSB 27, and DSB 22, respectively (Wirgin and Maceda, 1991). Each of these enzyme-probe combinations revealed a single restriction site polymorphism with two alleles. Analysis of controlled laboratory matings demonstrated the Mendelian inheritance and nonlinkage of loci. Genotypic frequencies derived from nDNA analy- sis were tested for deviations from Hardy-Weinberg equilibrium with the disequilibrium coefficient ap- proach (Weir, 1990). Mitochondrial DNA haplotype diversity was calculated with the formula of Nei and Tajima ( 1981). Chi-square significance (P<0.05) of the differences between striped bass populations in nDNA allele frequencies and mtDNA haplotype fre- quencies was tested by using the randomization ap- proach of Roff and Bentzen ( 1989). Results Nuclear DNA River (wild and hatchery) did not deviate significantly (P>0.05) from Hardy-Weinberg equilibrium; however, locus 22 did differ significantly from Hardy-Weinberg equilibrium for the Coos River wild (P<0.01) and hatch- ery (P<0.05) samples. Allelic frequencies for the three nDNA loci did not differ significantly (P>0.05) between the Hudson River and San Francisco Bay collections. Of the three nDNA loci, only locus 22 differed signifi- cantly in allelic frequencies between the San Francisco Bay and Coos River collections (%2=19.21; P<0.0001). Mitochondrial DNA Mitochondrial DNA haplotypic diversity (based on mtDNA length variants) showed a clear pattern of re- duction (0.810 to 0.0) among the striped bass popula- tions along the historical path that led to and includes the wild Coos River population (Table 2). The number of mtDNA haplotypes revealed decreased from 8 among Hudson River specimens to 5 in the San Francisco Bay collection, to 2 in the Umpqua, and 1 in the wild and hatchery-cultured Coos River samples. The third most common haplotype (C-l) found in the Hudson River collection (17%) was observed in the great majority of San Francisco Bay specimens (81%) and in all Coos River specimens. Also, the three least common haplotypes in Hudson River striped bass were those absent in striped bass from San Francisco Bay. Mitochondrial DNAhap- lotype frequencies were significantly different between the Hudson River and San Francisco Bay samples (^2=71.47; P<0.0001) and between the San Francisco Bay and Coos River samples (%2=8.21; P<0.05). Discussion Genotypic frequencies (Table 1 ) at two loci of samples Extensive allelic surveys of Atlantic coast striped from the Hudson River, San Francisco Bay, and Coos bass have shown extreme monomorphism at the pro- Waldman et al.: Population bottlenecks and DNA diversity in Morone saxatilis 617 Table 2 Mitochondrial DNA composite haplotype frequencies (letters represent length polymorphisms; numerals represent site polymor- phisms; percentages in parentheses) and genotypic diversity indices. Haplotype Population N A-l A-2 B-l B-2 C-l C-2 D-l D-2 cjunuuypic diversity Hudson River 110 3 (0.03) 2 (0.02) 25 (0.23) 7 (0.06) 19 (0.17) 32 (0.29) 17 (0.15) 5 (0.04) 0.810 San Francisco Bay 63 4 (0.06) 4 (0.06) 51 (0.81) 2 (0.03) 2 (0.03) 0.340 Coos Bay (wild) 38 38 (1.0) 0.0 Coos Bay (hatchery) 27 27 (1.0) 0.0 Umpqua River 12 11 (0.92) 1 (0.08) 0.167 tein level both within and among populations (re- viewed in Waldman et al., 1988). For example, Otto (1995) reported mean heterozygosity levels of Atlan- tic striped bass of approximately 1.0%, and 0.75% for the Hudson River population. Very low genetic diversity for Atlantic coast striped bass has also been shown by several mtDNA studies (Chapman, 1990; Wirgin et al., 1990, 1993; Waldman and Wirgin, 1994). For example, Wirgin et al. (1990) estimated the proportion of nucleotides that differed for the most divergent individuals among mid-Atlantic striped bass stocks at 0.0004, one of the lowest val- ues for any animal species. What mtDNA variation does exist in striped bass primarily is length, rather than site variation (Waldman and Wirgin, 1995). Four mtDNA major length variants have been found in striped bass from the Hudson River (Wirgin et al., 1990, 1993; Waldman and Wirgin, 1994). Thus, the substrate of genetic variation available among striped bass from the Hudson River and other Atlantic coast estuaries was extremely low in com- parison with most fishes (e.g. Waldman et al., 1996). From this unusually narrow gene pool some 430 or so yearlings were collected and transplanted in 1879 and 1882 to San Francisco Bay. It is not known how many of the female yearlings survived to reproduce as founders of all Pacific coast striped bass, but about 215 represents an approximate upper limit (assum- ing an unrealistic 100% survival rate), and 5 the lower limit, given the 5 haplotypes detected. Esti- mates of annual expectation of death from natural causes of the San Francisco Bay striped bass popu- lation obtained during a period of exploitation (sum- marized in Westin and Rogers3 ) ranged between 0.31 and 0.12, and averaged about 0.2. Moreover, female striped bass have variable maturation schedules (Berlinsky et al., 1995); in San Francisco Bay, it has been found that females mature at ages 4 and 5 (Stevens et al., 1987). Therefore, a more reasonable estimate for the number of transplanted females that survived to found the San Francisco Bay striped bass population is about 100. Striped bass then appeared in Coos Bay some 35 years after their introduction to San Francisco Bay. It is not possible to determine the number of founders of the Oregon populations nor their initial genetic makeup. Present Oregon striped bass are restricted to mtDNA length haplotype C. It is not known what the mtDNA haplotype frequencies of the San Francisco Bay stock were circa 1915, but if their present haplotype frequencies are used as an approximation, then the chance of a single female founder having a haplotype other than the C haplotype is only 16%. We believe that given the historical scarcity of striped bass in north Pacific coastal waters (Forrester et al., 1972) and the concordance between the dominant haplotype of the San Francisco Bay stock and the single (major length variant) haplotype of the Coos River and Umpqua River stocks, it is reasonable to assume that the Oregon popu- lations were founded by one or a very low number of female striped bass with the C-haplotype. Founding of the Oregon striped bass populations by a limited number of females from California, fol- lowing the initial bottleneck of transplantation from the Atlantic, would have resulted in a greatly reduced level of genetic variation in comparison with the 3 Westin, D. T., and B. A. Rogers. 1978. Synopsis of biological data on the striped bass, Morone saxatilis (Walbaum) 1792. Technical Report 67, Graduate School of Oceanography, Univ. Rhode Island, 154 p. 618 Fishery Bulletin 96(3), 1998 Hudson River stock. However, genetic variation may have been pared further, i.e. the subsequent history of the Coos River population suggests that some of the demographic and life history factors that con- tribute to low levels of genetic variation among na- tive populations of striped bass were pronounced in the Coos River population. The Coos River population has experienced its own bottleneck; recent estimates suggest a reduction in its order of magnitude from 104 * to 103 or 102. Fluctu- ating levels of annual spawning success also reduce the effective population size (Ne). Over generations N is approximated by the harmonic mean of each generation and strongly reflects periods of low abun- dance (Crow and Kimura, 1970). Many striped bass populations are sustained by occasional, extremely successful or “dominant” year classes (Raney, 1952). Dominant year classes have been rare but important for the Coos River population, occurring in 1940 and 1958 (McGie and Mullen4). Other factors that may have contributed to a re- duced N for the Coos River population are intrinsic to all populations of the species. Among these is skewed sex ratios. Males greatly outnumbered fe- males on the Coos River spawning grounds (Morgan and Gerlach1). Estimates of male to female ratios on spawning grounds of other systems range from about 10:1 to 100:1 (Chapman, 1990). Another factor is variance in progeny production among females, i.e. nonrandom family size (Gall, 1987). Large female striped bass can produce on the order of 106 eggs per year. Because of variable environmental conditions within a spawning season, some cohorts of eggs may show low or no survival whereas other cohorts may flourish (Dey, 1981; Secor and Houde, 1995). Thus, in- ordinate success by a few females would cause particu- lar genotypes to be overrepresented (Chapman, 1990). The significant difference in allele frequencies for nDNA locus 22 between the San Francisco Bay and Coos River collections indicates either a founder ef- fect or subsequent genetic drift. Also, the deviation from Hardy-Weinberg equilibrium of one of the three loci in the Coos River population is suggestive of small Ne or some other violation of the assumptions that lead to Hardy-Weinberg frequencies (Weir, 1990). In general, however, it appears that nDNA diversity was not strongly affected by the multiple bottlenecks that the Oregon populations experienced. In contrast, we have shown a stark decrease in mtDNA diversity. Our findings concerning nDNA and mtDNA diversity in Oregon striped bass are congru- 4 McGie, A. M., and R. E. Mullen. 1979. Age, growth, and popu- lation trends of striped bass, Morone saxatilis, in Oregon. Or- egon Department of Fish and Wildlife Information Report Se- ries, Fisheries 79-8, 57 p. ent with their having experienced a recent popula- tion bottleneck. A single breeding pair of diploid ani- mals contains four nuclear genomes and one trans- missible mtDNA; thus, a population that goes through an extreme bottleneck can lose all of its mtDNA; variability while still retaining a significant fraction of its nuclear variability (Wilson et al., 1985). A pattern of highly reduced mtDNA diversity and little altered nDNA diversity is consistent with a re- cent and unprolonged population bottleneck, as we hypothesize to have occurred during the establishment and history of the Coos River striped bass population. There is an intriguing inverse relationship between genetic diversity (reflected in mtDNA) of the striped bass populations investigated and the frequency of hermaphroditism. Inbreeding depression has been firmly associated with detrimental effects in captive vertebrate populations (e.g. Kincaid, 1976, Laikre and Ryman, 1991), and recently, strong evidence of reduced fitness and reproductive impairments due to inbreed- ing depression in wild vertebrates has emerged (e.g. Jimenez et al., 1994; Keller et al., 1994; O’Brien, 1994). Moser et al. (1983) investigated the phenomenon of hermaphroditism in Coos River striped bass in some detail. Protandry was suspected because young hermaphrodites had ripe, motile sperm and imma- ture eggs, whereas older hermaphrodites (ages 7 to 10) had normal appearing eggs and only small patches of testes. Reproductive impairment of older hermaphrodites was evident. Hermaphrodites with small testes and large ovaries showed annual accre- tions of eggs and one or more ovarian ducts blocked by adhesions. Each egg mass, representing previous spawning seasons, became progressively more degen- erated toward the interior of the gonad. Moser et al. (1983) also found that the oldest hermaphrodites had more constricted stomachs and intestines and more swollen abdomens than normal prespawning females. The four oldest hermaphrodites, 10 years old, had re- tained their eggs for up to 5 years. Also, it is possible that the absence of hermaphrodites among striped bass greater than 10 years of age (n= 7) indicates earlier mortality of hermaphroditic individuals, perhaps as a consequence of numerous annual egg mass accretions. We are aware of only a single observation of her- maphroditism among Atlantic coast striped bass. Westin (1978) reported one hermaphrodite among wild individuals from Chesapeake Bay that were sacrificed after being held in captivity for one year. However, in addition to surveying Coos River striped bass for evidence of hermaphroditism, Moser et al. (1983) also examined striped bass from San Fran- cisco Bay. Of more than 500 individuals, two were hermaphrodites. Thus, hermaphroditism in striped bass appears to be exceedingly rare in native, out- Waldman et al.: Population bottlenecks and DNA diversity in Morone saxatilis 619 bred populations, barely detectable in a non-native, outbred, but somewhat genetically constrained popu- lation, and pronounced in a non-native, inbred popula- tion. Moreover, the inverse association between abun- dance and hermaphroditism of the Coos River popula- tion suggests (but does not demonstrate) that a de- pensatory relationship exists because of these factors. That is, demographic influences that continued to re- duce genetic diversity of the Coos River population may have promoted this pathological reproductive response, which then hindered reproduction, further reducing abundance and leading to higher levels of inbreeding. Furthermore, no strong competing hypotheses have emerged to account for hermaphroditism of the Coos River striped bass stock. Chemical contamina- tion is one conceivable cause. However, the Hudson River — the original source of Coos River striped bass — is a heavily polluted estuary, yet hermaphro- ditism has not been observed in its population. Also, the Coos River is home to American shad ( Alosa sapidissima), similarly transplanted from the Atlan- tic (in 1871 to the Sacramento River; Mansueti and Kolb, 1953). In its native east coast rivers, including the Hudson River, American shad overlap with striped bass both spatially and temporally on their spawning runs. Nonetheless, whereas the Coos River striped bass population has approached extinction, its American shad population is flourishing. It is clear that populations of striped bass in Or- egon would benefit from broader genetic diversity (although there is considerable controversy concern- ing maintenance of non-native fish populations; e.g. Courtenay, 1995). If perpetuation of the Coos River population is desired, consideration should be given to the introduction of striped bass to Oregon waters from one or more additional Atlantic coast stocks. A combination of intention and serendipity resulted in the establishment of striped bass in California and then Oregon; however, the ultimate effect was to pro- mulgate northern, hypothermal, nonmigratory popu- lations from individuals originally obtained from a midlatitude, mesothermal, migratory population. There is some evidence that striped bass from north- ern latitudes are better adapted to the ambient en- vironmental conditions of those latitudes (Conover, 1990). We suggest that Atlantic coast striped bass from northerly latitudes such as the Canadian Maritimes would be more preadapted to Oregon habitats. Acknowledgments We are grateful to Don Stevens for providing tissue samples from California striped bass and to Terry Jarmain for Umpqua River samples. Literature cited Berlinsky, D. L., M. C. Fabrizio, J. F. O’Brien, and J. L. Specker. 1995. Age-at-maturity estimates for Atlantic coast female striped bass. Trans. Am. Fish. Soc. 124:207-215. Bernatchez, L., J. J. Dodson, and S. Boivin. 1989. Population bottlenecks: influence on mitochondrial DNA diversity and its effect in coregonine stock discrim- ination. J. Fish Biol. 35 (suppl. A):233-244. Brown, J. R., A. T. Beckenbach, and M. J. Smith. 1992. Influence of Pleistocene glaciations and human in- tervention upon mitochondrial DNA diversity in white stur- geon (Acipenser transmontanus ) populations. Can. J. Fish. Aquat. Sci. 49:358-367. Chapman, R. W. 1990. Mitochondrial DNA analysis of striped bass popula- tions in Chesapeake Bay. Copeia 1990:355-366. Conover, D. O. 1 990. The relationship between capacity for growth and length of growing season: evidence for and implications of counter- gradient variation. Trans. Am. Fish. Soc. 119:416-430. Courtenay, W. R., Jr. 1995. The case for caution with fish introductions. Am. Fish. Soc. Symp. 15:413-424. Crow, J. F., and M. Kimura. 1970. An introduction to population genetics theory. Har- per and Row, New York, NY, 591 p. Dey, W. P. 1981. Mortality and growth of young-of-the-year striped bass in the Hudson River estuary. Trans. Am. Fish. Soc. 110:151-157. Feinberg, A. P., and B. Vogelstein. 1983. A technique of radiolabeling DNA restriction endo- nuclease fragments to high specific activity. Anal. Biochem. 132:6-13. Forrester, C. R., A. E. Peden, and R. M. Wilson. 1972. First records of the striped bass, Morone saxatilis , in British Columbia waters. J. Fish. Res. Board Can. 29: 337-339. Gall, G. A. E. 1987. Inbreeding. In N. Ryman and F. Utter (eds.), Popu- lation genetics and fishery management, p. 47-87. Univ. Washington Press, Seattle, WA. Jimenez, J. A., K. A. Hughes, G. Alaks, L. Graham, and R. C. Lacy. 1994. An experimental study of inbreeding depression in a natural habitat. Science (Wash., D.C) 266:271-273. Keller, L. F., P. Areese, J. N. M. Smith, W. Hochachka, and S. C. Stearns. 1994. Selection against inbred song sparrows during a natu- ral population bottleneck. Nature (Lond.) 372:356-357. Kincaid, H. 1976. Effects of inbreeding on rainbow trout popula- tions. Trans. Am. Fish. Soc. 105:273-280. Laikre, L., and N. Ryman. 1991. Inbreeding depression in a captive wolf (Canis lu- pus) population. Conserv. Biol. 5:33-40. Mansueti, R., and H. Kolb. 1953. A historical review of the shad fisheries of North America. Chesapeake Biological Laboratory Publication 97, 293 p. Moser, M., J. Whipple, J. Sakanari, and C. Reilly. 1983. Protandrous hermaphroditism in striped bass from Coos Bay, Oregon. Trans. Am. Fish. Soc. 112:567-569. 620 Fishery Bulletin 96(3), 1998 Nei, M., and F. Tajima. 1981. DNA polymorphism detectable by restriction endonucleases. Genetics 97:145-163. O’Brien, S. J. 1994. Genetic and phylogenetic analyses of endangered species. Ann. Rev. Genet. 28:467-489. Otto, R. S. 1975. Isozyme systems of the striped bass and congeneric per- cichthyid fishes. Ph.D. diss., Univ. Maine, Orono, ME, 67 p. Raney, E. C. 1952. The life history of the striped bass, Roccus saxatilis (Walbaum). Bull. Bingham Ocean. Coll. 14:5-97. Richardson, L. R., and J. R. Gold. 1997. Mitochondrial DNA diversity in and population struc- ture of red grouper, Epinephelus morio, from the Gulf of Mexico. Fish. Bull. 95:174-179. Raff, D. A., and P. Bentzen. 1989. The statistical analysis of mitochondrial DNA poly- morphisms: x2 and the problem of small samples. Mol. Biol. Evol. 6:539-545. Saghai-Maroof, M. A., K. M. Soliman, R. A. Jorgneson, and R. W. Allard. 1984. Ribosomal DNA spacer-length polymorphisms in bar- ley: Mendelian inheritance, chromosomal location, and popu- lation dynamics. Proc. Natl. Acad. Sci. USA 81:8014—8018. Secor, D. H., and E. D. Houde. 1995. Temperature effects on the timing of striped bass egg production, larval viability, and recruitment potential in the Patuxent River (Chesapeake Bay). Estuaries 18:527-544. Stevens, D. E., H. K. Chadwick, and R. E. Painter. 1987. American shad and striped bass in California’s Sac- ramento-San Joaquin River system. Am. Fish. Soc. Symp. 1:66-78. Waldman, J. R., J. Grossfield, and I. Wirgin. 1988. Review of stock discrimination techniques for striped bass. N. Am. J. Fish. Manage. 8:410-425. Waldman, J. R., K. Nolan, J. Hart, and I. I. Wirgin. 1996. Genetic differentiation of three key anadromous fish populations of the Hudson River. Estuaries 19:759-768. Waldman, J. R., and I. I. Wirgin. 1994. Origin of the present Delaware River striped bass population as shown by analysis of mitochondrial DNA. Trans. Am. Fish. Soc. 123:15-21. 1995. Comment: mitochondrial DNA stability and striped bass stock identification. Trans. Am. Fish. Soc. 124:954- 956. Weir, B. S. 1990. Intraspecific differentiation. In D. M. Hillis and C. Moritz (eds.), Molecular systematics, p. 373-410. Sinnauer Associates, Sunderland. Westin, D. T. 1978. Serum and blood from adult striped bass, Morone saxatilis. Estuaries 1:126-128. Wilson, A. C., R. L. Cann, S. M. Carr, M. George, U. B. Gyllensten, K. M. Helm-Bychowski, R. G. Higuchi, S. R. Palumbi, E. M. Prager, R. D. Sage, and M. Stoneking. 1985. Mitochondrial DNA and two perspectives on evolu- tionary genetics. Biol. J. Linn. Soc. 26:375-400. Wirgin, I., B. Jessop, S. Courtenay, M. Pedersen, S. Maceda, and J. R. Waldman. 1995. Mixed-stock analysis of striped bass in two rivers of the Bay of Fundy as revealed by mitochondrial DNA. Can. J. Fish. Aquat. Sci. 52:961-970. Wirgin, 1. 1., and L. Maceda. 1991. Development and use of striped bass-specific RFLP probes. J. Fish Biol. 39 (suppl. A):159-167. Wirgin, I., L. Maceda, J. R. Waldman, and R. N. Crittenden. 1993. Use of mitochondrial DNA polymorphisms to estimate the relative contributions of the Hudson River and Chesa- peake Bay striped bass stocks to the mixed fishery on the Atlantic coast. Trans. Am. Fish. Soc. 122:669-684. Wirgin, I. I., P. Silverstein, and J. Grossfield. 1990. Restriction endonuclease analysis of striped bass mitochondrial DNA: the Atlantic coastal migratory stock. Am. Fish. Soc. Symp. 7:475-491. 621 Can scales be used to sex winter flounder, Pleuronectes americanus? Allen J. Bejda Beth A. Phelan James J. Howard Marine Sciences Laboratory Northeast Fisheries Science Center National Marine Fisheries Service, NOAA Highlands, New Jersey 07732 E-mail address (for A. J. Bejda): allen.bejda@noaa.gov nearest millimeter total length. Using the method described by Lux and Porter ( 1963), we palpated the blind side of the caudal peduncle noting either roughness (males) or smoothness (females). The fish were then dissected to determine (visually) maturity and sex. All immature fish were eliminated from the study. When possible, a sample of ten or more scales was taken from the palpated area. In the laboratory, the scales were ex- amined with a dissecting micro- scope to determine their type and relation to texture. We tested the accuracy of a tech- nique to rapidly determine the sex of mature winter flounder, Pleuro- nectes americanus * in the field with- out sacrificing individuals. Al- though a number of techniques have been developed for sexing fish by examining reproductive anatomy (Moen, 1959; Driscoll, 1969; Martin et al., 1983; Ross, 1984), they are not well suited for easy and rapid determinations in the field. Exter- nal characteristics, such as features of the urogenital region (Sigler, 1948; McComish, 1968; Flickinger, 1969; Lebeau and Pageau, 1989) or body shape (Snow, 1963), have been successfully applied in the field for other species but are not applicable to winter flounder which exhibit no obvious sexual dimorphism. Some studies have suggested that male and female winter flounder can be differentiated by the texture of the scales on the blind side; males al- legedly have ctenoid scales which feel rough, whereas females alleg- edly have cycloid scales which feel smooth ( Perlmutter, 1947; Lux and Porter, 1963). Unfortunately, the lit- erature is inconclusive and none of the studies have carefully tested the relationship of texture as related to scale type and sex. MacPhee (1978) citing Norman (1934) claimed that scales on the blind side of males are ctenoid rather than cycloid when, in fact, Norman stated that for the species in general the scales are ctenoid on the ocular side and cyc- loid on the blind side. Methods From March through September, 1989, winter flounder were col- lected with an otter trawl in New Jersey waters. Fish were collected at 14 inshore sites in Raritan and Sandy Hook Bays (Phelan, 1992) and 22 offshore sites 22.2 km east of Sandy Hook (Studholme et al., 1995). Fish were measured to the Results and discussion A total of 730 mature fish, ranging from 12.9 to 39.7 cm, were exam- ined. External palpation resulted in a significantly higher (%2=51.528, PcO.OGl) sex ratio (females to males) of 2.7:1 than the actual ra- tio of 1.2:1 (Fig. 1). The difference in the ratios was due primarily to males being identified as females. Forty six percent (155 of 332) of the Manuscript accepted 10 December 1997. Fishery Bulletin 96:621-624 (1998). ■Q E □ 600 500 E3 Misidentified male □ Misidentified female Females Males True identification Females Males identification by palpation Figure 1 Number of female and male winter flounder (Pleuronectes americanus ) identified by gonadal inspection and palpation of the blind side scales. 622 Fishery Bulletin 96(3), 1998 Table 1 Sex ratio of winter flounder as determined by external palpation and gonad inspection. Sex ratio (female:male) Method Number of fish Location Source 2.5:1 External 3711 New England and New York Perlmutter, 1947 2.3:1 External 601 Rhode Island Saila, 1961 2.6:1 External 1431 New England Lux and Porter, 1963; Lux, 1969 2.3:1 External 12,151 Massachusetts Howe and Coates, 1975 1.5:1 Internal 940 Rhode Island Saila, 1962 1.2:1 Internal 1569 Rhode Island Berry et al., 1965 1:1 Internal 227 Newfoundland Kennedy and Steele, 1971 1:1 Internal 1465 Massachusetts Haedrich and Haedrich, 1974 H Female - cycloid ED Female - ctenoid □ Male - ctenoid E3 Male - cycloid True Palpation Female True Palpation Male Figure 2 Number of female and male winter flounder ( Pleuronectes americanus) identified by gonadal inspection and palpation in relation to scale type. males were misidentified, and six percent (23 of 398) of the females were identified as males. Similar differences between sex ratios derived from palpation of the scales and from gonadal inspection have been reported in the literature (Table 1). In the four studies that used palpation to de- termine sex, ratios ranged from 2.3 to 2.6:1 compared with ratios of 1:1 to 1.5:1 when gonads were used. Scales were collected and examined from 672 of the 730 fish. Sex ratios for this subset also differed significantly (X2=44.203, PcO.OOl) between the two sexing methods: 2.5:1 for palpation and 1.2:1 for gonadal inspection (Fig. 2). In- dividual fish exhibited either cycloid (479 fish) or ctenoid ( 199 ) scales, a finding that differs from Norman’s ( 1934) observation that scales on the blind side were cyc- loid. Female fish exhibited primarily cy- cloid scales (Fig. 2) which were present on 352 (98%) of the 361 females examined. Scale type for male fish, however, was variable; 120 (39%) fish exhibited cycloid scales and 191 (61%) ctenoid (Fig. 2). Discrepancies between texture determinations and scale type (observer error) occurred in 33 (5%) of the fish examined (Fig. 2). It is therefore apparent that the use of texture ac- cording to scale type for sexing winter flounder (Perlmutter, 1947; Lux and Porter, 1963; MacPhee, 1978) is not accurate, resulting in a significantly fe- male-biased sex ratio. Literature cited Berry, R. 30 mm cara- pace length) tended to struggle and often managed to escape the electric current. Once escaped, these animals could usually be re-entrained and induced to reach the anode by increasing the current strength and thus the tail flicking reaction. Overstimulation resulted in the animal lying motionless for several minutes after the current had been switched off. All animals that were affected in this way revived after a few minutes, apparently without ill affects. Small EBP (<10 mm) animals nearly always became mo- tionless, usually lying on their backs, upon reaching the anode when it was situated on the substrate. If the electrode was elevated above the substrate, EBP animals gathered below it and remained motionless, but upright. When the current was switched off, they quickly escaped. EBP animals always reached the anode more quickly than large animals. In many cases the reaction was immediate and the anode was reached in less than a second. Mortalities of experi- mental animals (pre- and postexperimental) that could not be attributed to accidental mishandling were negligible (total of 3) during the period of ob- servation (60 d). We concluded that American lobsters exhibit a true electrotaxis, i.e. where an animal in a DC field is compelled to swim to the anode through involuntary muscular contractions. Electrotaxis is well known in fish (Lamarque, 1990), but has rarely been described in crustaceans. Saila and Williams (1972) observed tail muscle contractions in American lobsters sub- jected to currents (<38 V input), but no taxis was evident. Stewart (1974) concluded that similar tail 630 Fishery Bulletin 96(3), 1998 Figure 2 Laboratory setup (bottom) used to determine electrotaxis in early benthic-phase American lobsters Homarus americanus and capture efficiency of five electrode configurations (top). flicking in N. norvegicus was not involuntary, but rather the animal’s natural escape reaction induced by the electric stimulation. We have found only two reports of electrotaxis in crustaceans in which tail flicking and movement toward the anode were ob- served: one for the penaeid shrimp Penaeus duorarum (Higman, 1956) and the other for the rock lobster, Panulirus cygnus (Phillips and Scolaro, 1980). It was apparent from these observations that the strong electrotactic response of EBP lobsters could be used to develop a quadrant-like field-sampling device. Table 1 gives the results of experiments with five different electrode configurations and EBP ani- mals sheltered under cobble on sand substrate. These tests confirmed the unidirectional nature of the elec- trotaxis. Of the 81 (59%, n = 137) animals that emerged from the cobble shelters, all moved directly to the anode. There was a significant difference be- tween electrode configurations in the mean number of animals caught (ANOVA, P<0.001). The best cap- ture rate (2.6 animals per trial, or 85% of the total population) was obtained with a semicircular cath- ode, straight anode and horizontal configuration (con- figuration 3 in Fig. 2), and the worst capture rate (0.8 animals, 25% of the population) was obtained with the circular cathode, plate anode, and vertical configuration (configuration 4 in Fig. 2). There was a significant difference between electrode configura- tions in mean capture time for the first animal (ANOVA, P=0.008, LSD post-hoc test) but no signifi- cant differences in capture times for the other two animals in each trial. Animals tended not to emerge at the same time but in sequence, with the first, sec- ond, and third animals emerging after an overall average of 26.3, 51.6, and 64.5 seconds. In all trials combined, one animal was caught in 91%, two in 60%, and all three in 23% of the trials. The lower and slower capture rate of the second and third animals in each test is probably related to the position of the animals in the rock pile. The EBP lobsters placed in close proximity exhibit intense aggressive behavior (Lawton and Lavalli 1995), which would tend to re- sult in an overdispersed distribution within a bounded habitat like the test rock pile. In this situa- tion some animals will be closer to the anode, and NOTE Koeller and Crowell: Electrotaxis in Homarus americanus 631 these would tend to be caught more quickly than those farther away. Animals farther from the anode have a lower probability of a clear passage through crevices and a greater probability of being caught in a “dead end” before emerging from the rock pile. It is apparent that electrofishing is a potentially useful method for sampling EBP lobsters. However, some additional research is necessary before a prac- tical field sampler can be developed and tested suc- cessfully. Our study, although using the most com- mon EBP lobster habitat, did not examine other habi- tats (e.g. eelgrass, mud) or variations of the common habitat, such as different rock sizes and associated crevices, or different thicknesses of cobble. Although the tedium associated with the existing sampling methods could be reduced significantly by using a diver-operated electrosampler, the most cost-effective and efficient sampler would be operated from a small boat without the need for divers. Conceivably, a sam- pling device that combines the electrotactic response of EBP lobsters with a bottom-to-surface water pump (e.g. Bergstedt and Genovese, 1994) could fulfill these requirements. Finally, it should be noted that the device described was designed only for use in test tanks insulated from the operator. The use of electro- fishing devices in saltwater by divers, although quite safe (Stewart and Cameron, 1974), does require spe- cial precautions from an engineering as well as a personal safety perspective. Literature cited Bergstedt, R. A., and J. H. Genovese. 1994. New technique for sampling sea lamprey in deep- water habitats. N. Am. J. Fish. Manage. (14) 2:449-452. Higman, J. B. 1956. The behaviour of the pink grooved shrimp Penaeus duo- rarum Burkenroad, in a direct current electric field. Tech. Ser. Mar. Lab., Univ. Miami 16, 24 p. Hudon, C. 1987. Ecology and growth of post-larval and juvenile lob- sters, Homarus americanus , off isle de la Madeleine (Quebec). Can. J. Fish. Aquat. Sci. 44: 1855-1869. Lamarque, P. 1990. Electophysiology of fish in electric fields. In I. G. Cowx and P. Lamarque (eds.), Fishing with electricity — applications in freshwater fisheries management. Fishing News Books, Oxford, 248 p. Lawton, P„ and K. L. Lavalli. 1995. Postlarval, juvenile, adolescent, and adult ecology. In J. R. Factor (ed.), Biology of the lobster Homarus americanus , Academic Press, New York, NY, 538 p. Phillips, B. F., and A. B. Scolard. 1980. An electrofishing apparatus for sampling sublittoral benthic marine habitats. J. Exp. Mar. Biol. Ecol. 47:69-75. Saila, S. B., and C. E. Williams. 1972. An electric trawl system for lobsters. Mar. Technol. Soc. J. 6:25-31. 632 Fishery Bulletin 96(3), 1998 Stewart, P. A. M. 1974. An investigation into the effects of electric fields on Neph- rops norvegicus. J. Cons. Int. Explor. Mer 35(3):249-257. Stewart, P. A. M., and G. C. Cameron. 1974. The safe use by divers of a high current pulse gen- erator in studies of behaviour of marine fish in electric fields. J. Cons. Int. Explor. Mer 36:62-70. Wahle, R. A., and R. S. Steneck. 1992. Recruitment habitats and nursery grounds of the American lobster ( Homarus americanus Milne Edwards). Mar. Ecol. Prog. Ser. 69:231-243. 633 Changes in the probability density function of larval fish body length following preservation Pierre Pepin Fisheries and Oceans RO. Box 5667 St. John's, Newfoundland, Canada A! C 5X! E-mail address: pepin@athena.nwafc.nf.ca John F. Dower William C. Leggett Department of Biology, Queen's University Kingston, Ontario, Canada K7L3N6 The influence of body size on physi- ological and developmental pro- cesses during the early life history of fishes has been clearly demon- strated (Miller et ah, 1988; Houde, 1989; Pepin, 1991). Because of this and to provide a direct comparison of field collections with laboratory observations that often use mea- surements of fresh specimens, nu- merous studies have quantified the effects of preservation and han- dling on the length of larval fish (Table 1). Most of the research has used laboratory-reared animals for which changes in length are as- sessed using either comparisons among treatments (i.e. AN OVA) or departures from a one-to-one rela- tionship (i.e. regression) (Table 1). Shrinkage in length is the predomi- nant response to preservation (Table 1), and the amount of shrink- age is influenced by handling (Theilacker, 1980, 1 966: Theilacker and Porter, 1995; Fox, 1996). Changes in larval body length due to preservation are relatively small (3-15%) although variations of up to 1 mm are not uncommon (Table 1). The contrast among spe- cies led Jennings (1991) to suggest that specific correction factors would be required. An alternative was Hjorleifsson and Klein-Mac- phee’s (1992) simple model of the relative change in larval lengths, based on a review of previous stud- ies, which showed clear evidence of the effects of body length and pres- ervation time on shrinkage. Of course, as with any such review, the effects of unaccounted for or con- founding variables are unknown (Pepin and Miller, 1993). However, Hjorleifsson and Klein-Macphee ( 1992 ) predicted a maximum mean shrinkage of -15% for the smallest larvae (-2 mm) followed by an ex- ponential decrease in relative shrinkage with increasing length. This raises an important point. Al- though differences in mean larval body length may appear significant from a statistical perspective, the importance of correcting for the ef- fects of preservation are most pro- nounced at the level of the indi- vidual because morphological mea- surements are used in the analysis of each larva’s physiological condi- tion (e.g. Theilacker and Porter, 1995) to assess the relative state of a population. Even small changes in length resulting from shrinkage can create important biases at this level of analysis. In contrast, the small relative effects of preserva- tion on body length (Hjorleifsson and Klein-Macphee, 1992) is unlikely to substantially influence estimates of length-frequency distributions. Fundamental issues yet to be addressed in the assessment of lar- val shrinkage include 1) the effect of preservation on the distribution of larval length measurements within a given length interval rather than the mean, and 2) the contribution of investigator-in- duced error on that distribution. The former point is of particular importance because the application of correction factors to individual larvae must maintain the status of that animal in relation to others within the population. Investigator- induced error may be equally impor- tant. Few studies have attempted to quantify the variance in repeated estimates for a given operator (Jen- nings, 1991; Hjorleifsson and Klein- Macphee, 1992; Fox, 1996) and none have contrasted bias and variance among operators. Our failure to address these questions to date probably results from 1) the small sample sizes (i.e. generally <100) presented in most studies of the effects of preservation (Table 1), and 2) the fact that usually only a single investigator is involved in making the measurements of the larvae. Nonetheless, these issues are important because they provide the basis for narrowing the possible sources of error in interpretations of physiological processes that in- fluence ichthyoplankton population dynamics. In this study we report on an evaluation of the shrinkage effects of preservation on several species of larval fish collected as part of a field study. We consider changes in both the mean and variance of the distribution of individuals within narrow length intervals and assess the null hypothesis that there are no differences among species. We also Manuscript accepted 11 September 1997. Fishery Bulletin 96:633-640 (1998). 634 Fishery Bulletin 96(3), 1 998 NOTE Pepin et al .: Changes in the distribution of larval fish body length following preservation 635 evaluate the magnitude ofintra- and interoperator dif- ferences in performance for repeated measurements of the same specimens. Materials and methods The study was conducted on Conception Bay, Canada (47°45'N, 53°00'W), during the period of 12 July to 4 August 1995. Sampling was performed daily from CSS Shamook (23-m boat length) at a single site near the head of the bay. Larvae from a number of species were obtained from vertical plankton hauls made with a square trawl 4 m long with a 4-m2 mouth fitted with 333-pm mesh nitex and an oversize codend 20 cm in diam- eter and 30 cm long. The net was lowered to a depth of 20-30 m and retrieved at a rate of 1 m/s. Net de- sign and deployment protocol were chosen to mini- mize trauma to larvae. On deck, the net was washed and the codend contents poured into a 20-L plastic bucket. Live and freshly dead ichthyoplankton were immediately sorted with flexible forceps and trans- ferred to petri dishes filled with chilled seawater. Moribund larvae, indicative of death or extremely poor condition prior to capture (O’Connell, 1981; Otto and Boggs, 1983; Takizawa et al., 1994), were ex- cluded from our samples to avoid possible bias. Each larva was assigned an indentification number, ten- tatively identified to the lowest taxonomic level pos- sible, and recorded on videotape with a camera mounted on a Wild M3C dissecting microscope (S- type mount, 0.5x objective). Individual larvae were immediately transferred into 1.5 mL microcentrifuge tubes filled with 2% buffered formaldehyde. Identi- fications were confirmed in the laboratory and stan- dard lengths were determined to the nearest 0.1 mm using an Optimas® image analysis system. Measure- ments of the video-recorded fresh standard lengths were performed at the Northwest Atlantic Fisheries Centre by an operator with more than 10 years of experience in the study of larval fish. All laboratory analyses performed on preserved larvae were con- ducted approximately five months after the sampling cruise at Queen’s University by a novice operator with less than 1 year of experience in the study of larval fish. Each preserved larva was extracted from its microcentrifuge tube, videotaped in a manner identical to that employed for recently captured lar- vae, and measured (standard length) with an Optimas® image analysis system. All measurements were performed by clicking on a series of points through the centre of the head and along the notochord. To contrast the precision and accuracy of length measurements, the experienced operator repeated a set of measurements on a sample of randomly se- lected larvae from the initial videotape records. In addition, a random subsample of 72 preserved cape- lin larvae ( Mallotus villosus) were selected from the collection for replicate measurements by the experi- enced and novice operators. In each instance, both operators independently videotaped each larva for both the initial and repeated measurements. In ad- dition, the experienced operator performed duplicate measurements on approximately one half of the samples measured by both operators. All measure- ments were performed with only knowledge of speci- men number and magnification level. We evaluated individual differences in body length following preservation (Al = lpreserved - lfresh)- For analysis, data were binned into 1-mm length inter- vals for fresh specimens (e.g. 1 < x < 2). The hypoth- esis that there was no significant difference in mean body length ( H0:Al =0) was evaluated by using a two-sided /-test. Comparison among species was ac- complished by using anANOVA( H0: A lx = for species 1 to n). A species was included in the analy- sis only if there were at least three specimens within a length interval. Homogeneity of the variances in fresh (f) and preserved (p) larval lengths within each length interval was assessed to/ using a one-sided F-test of the null (H0\Sf>=sp) and alternative (HA:sf0.05). Comparison of measurements by experienced and novice operators revealed no significant bias on the part of either operator although the results did ap- proach significance (0.1>P>0.05) and the slope was not significantly different from 1 (Table 2). However, the residual variance about the novice-experienced operator regression was significantly greater than the residual variance of repeated measurements 636 Fishery Bulletin 96(3), 1 998 made by the experienced operator (F72 38=4.61, P<0.001). Initial measurement (mm) Initial measurement (mm) Experienced operator’s measurement (mm) Figure 1 Comparison of repeated measurements of the experienced operator on fresh (top panel) and preserved (center panel) speci- mens as well as the comparison between experienced and novice operators (bottom panel). Solid lines represent least squares regressions (Table 2). Dotted lines show the 1:1 relationship. Dashed lines in the lower panel show the 95% prediction intervals. Changes in body length A total of 1179 larvae representing 9 different spe- cies were used in our analyses. Within each length interval, the total number of specimens available ranged from 10 to 281 individuals, representing 1 to 5 species (Table 3). Within 1-mm length intervals, preservation re- sulted in significant increases in body length for in- dividuals 3-6 mm fresh standard length and signifi- cant decreases in body length for individuals >7 mm fresh standard length (Fig. 2, Table 3). In six in- stances there were also significant differences among species in preserved body length (Table 3). For fresh length intervals between 5 and 9 mm there was con- sistency in the order of species; preserved specimens of Hippoglossoides platessoides were larger for a given length interval than Ulvaria subbifurcata and Mallotus villosus. However, above and below the 5- 9 mm length intervals there was variation in the or- der of species. In the larger size classes (>12 mm), preserved Stichaeus punctatus and Gadus morhua were larger than Clupea harengus and Hippogloss- oides platessoides. The initial variance of fresh standard lengths within each length interval ranged from 0.043 to 0.088 mm2, whereas the variance of preserved stan- dard lengths of these same individuals was significantly greater (Table 3) and ranged from 0.18 to 3.5 mm2 (Fig. 3). Furthermore, the variance of the preserved length measurements increased significantly with in- creasing fresh length (^=0.70, P<0.01, rc=15) (Fig. 3). Kendall’s correlation coefficient (x) revealed that individual larvae remained at the same rank about 1 time in 3 (Fig. 4). This effect increased (i.e. 1 time in 4 or 5) with increasing fresh length. ! Discussion Changes in body length of larval fish due to han- dling and preservation are neither uniform nor con- sistent among individual animals: there is substan- tial variation in reaction to both handling (no mat- ter how gentle) and preservation and variation is greatest in absolute terms for larger larvae. No previous study (Table 1) has employed sample sizes large enough to permit the evaluation of changes in the distribution or rank of individual lar- vae within small length intervals. We found clear evidence that changes in body length can be more substantial than previously estimated by means of analyses applied to a broad range of sizes. Despite significant changes in body length for most length intervals, our results show that 1) variation about NOTE Pepin et a I.: Changes in the distribution of larval fish body length following preservation 637 Table 2 Results of the comparison of initial ( I ) and repeated (R) measurements on fresh and preserved larvae by the experienced ( E ) and novice (N) operators. Numbers in square brackets represent the standard error of the intercept (a) and slope ((3). The sixth and seventh columns show the test of the hypotheses that the intercept is not significantly different from 0 and that the slope is not significantly different from 1. Treatment Initial Repeat Regression (y = a + fix) F- Value f-value Hu : a = 0 t-value tf0:p=l Residual variance Fresh E E R = 0.074 [0.128] + 0.983 [0.0229] / F1 40=1848,P< 0.001 0.58, P> 0.05 0.74, ns 0.020 Preserved E E R = 0.113 [0.126] + 0.977 [0.0228] / Fl ,38=183 3. P < 0.001 0.90, P> 0.05 1.01, ns 0.018 Preserved E N R = -0.271 [0.149] + 0.959 [0.0263]/ F1 72=1414,P< 0.001 1.81,0.1 >P> 0.05 1.56, ns 0.083 4 -i o o o -4 -| — i — 1 — i — | — i — | — i — | — i — |-2h — | — i — | — i — |- 0 2 4 6 8 10 12 14 16 Fresh length (mm) Figure 2 Box plot of the 25th, 50th, and 75th percentile of the dif- ference in mean individual length of larval fish within each millimeter fresh length interval. Capped whiskers show the 10th and 90th percentiles and the open circles show the distribution of observations outside those confines. The dotted lines represent the 95% confidence intervals of the residual population variance from the comparison of mea- surement error between operators. the mean can be substantial and that 2) the relative position of an individual larva within a length inter- val may shift substantially. Our findings are consistent with those of Hay (1982), Theilacker (1986), and Fox (1996) who all found that postpreservation changes in body length were greatest for large larvae. However, the marked increase in variance of preserved larval lengths in relation to the initial measurements was not caused by operator error. The residual variance of repeated 1 r hico^iniostooio'-ort ^oicl)4ul)(f!)Nc6j7VTI’ O ■*“ CN Length interval (mm) Figure 3 Variance in body length of fresh (circles) and pre- served (squares) specimens within each millimeter interval. measurements among operators was of the same or- der as the variance in fresh lengths within individual length intervals and was comparable to that obtained by Jennings ( 1991), Hjorleifsson and Klein-Macphee (1992), and Fox (1996). However, Fox (1996) found some evidence of increased variance with increased length, in contrast with our results. We conclude that remeasurement of larvae per se is unlikely to be a major contributor to our observation that the order of individual larvae, in relation to population esti- mates, is not maintained following preservation. Our finding of postpreservation increases in body length, for individuals <6 mm SL, contrasts with most studies of the effects of formaldehyde preservation on larval fish. Fox (1996) and studies of other meth- 638 Fishery Bulletin 96(3), 1998 Table 3 Species specific mean differences in length following preservation within 1-millimeter length intervals. The number of specimens is indicated by n. Subscripts in the fourth and seventh columns refer to the species for which information was available within each length interval (t= 1, . . . , n). The standard deviation of the species specific difference in length (SD) is also provided. A positive difference in length indicates expansion whereas a negative value indicates shrinkage. The last three columns present the analytical results for overall changes in length following preservation, differences among species, and homogeneity of vari- ance between fresh and preserved specimens within each length interval. Length Interval (mm) Species n Z, SD (A/,) Z-value H0 : A Z = 0 f-value H0 '■ A Z i = F-value H0\sj >= Sp 1-2 Pleuronectes americanus ii 0.18 0.43 1.86, P> 0.05 0.02, P> 0.5 3.21, P< 0.01 Tautogolabrus adspersus 5 0.15 0.19 2-3 Pleuronectes americanus 13 0.12 0.46 0.92, P> 0.2 N/A 2.45, P < 0.05 3-4 Liparis sp. 8 0.07 0.41 9.74, P< 0.001 24.6, P< 0.001 4.49, P < 0.001 Mallotus villosus 73 0.63 0.39 Pleuronectes americanus 6 -0.39 0.35 4-5 Hippoglossoides platessoides 3 -0.33 0.75 6.96, P< 0.001 2.78, P< 0.05 4.87, P< 0.001 Liparis sp. 15 -0.07 0.99 Mallotus villosus 136 0.26 0.56 Pleuronectes americanus 3 0.09 0.77 Ulvaria subbifurcata 75 0.37 0.47 5-6 Hippoglossoides platessoides 5 0.48 0.71 3.38, P< 0.001 4.38, P< 0.01 4.33, P < 0.001 Liparis sp. 7 -0.76 0.43 Mallotus villosus 119 0.14 0.53 Pleuronectes americanus 3 -0.10 0.49 Ulvaria subbifurcata 147 0.14 0.67 6-7 Hippoglossoides platessoides 9 0.23 0.52 0.33, P> 0.5 0.76, P> 0.2 6.63, P < 0.001 Mallotus villosus 51 -0.07 0.65 Ulvaria subbifurcata 122 -0.01 0.70 7-8 Hippoglossoides platessoides 6 0.73 0.67 2.70, P < 0.01 4.61, P< 0.05 10.1, P< 0.001 Mallotus villosus 33 -0.41 1.10 Ulvaria subbifurcata 96 -0.19 0.76 8-9 Hippoglossoides platessoides 4 0.34 0.66 2.13, P< 0.05 3.59, P< 0.05 6.80, P < 0.001 Mallotus villosus 6 -0.79 1.54 Ulvaria subbifurcata 77 -0.14 0.59 9-10 Clupea harengus 3 -0.12 1.08 1.71, P> 0.05 0.13, P> 0.5 5.35, P < 0.001 Hippoglossoides platessoides 7 -0.06 0.65 Ulvaria subbifurcata 36 -0.20 0.66 10-11 Hippoglossoides platessoides 4 -1.00 2.11 2.68, P < 0.05 0.42, P> 0.5 19.9, P < 0.001 Ulvaria subbifurcata 28 -0.56 1.18 11-12 Hippoglossoides platessoides 4 -0.68 0.69 2.58, P < 0.05 0.03, P> 0.5 16.3, P< 0.001 Ulvaria subbifurcata 8 -0.58 0.93 12-13 Clupea harengus 3 -0.13 0.73 2.75, P< 0.05 1.26, P> 0.5 42.7, P< 0.001 Hippoglossoides platessoides 3 -1.25 0.39 Ulvaria subbifurcata 4 -1.09 1.29 13-14 Clupea harengus 10 -0.13 0.78 0.1, P> 0.5 2.64, P> 0.1 16.9, P< 0.001 Hippoglossoides platessoides 3 -0.63 1.44 Stichaeus punctatus 4 0.88 0.87 14-15 Clupea harengus 11 -0.81 0.97 0.8, P> 0.2 4.74, P< 0.05 20.2, P< 0.001 Gadus morhua 3 0.87 0.72 Stichaeus punctatus 3 0.72 1.50 15-16 Clupea harengus 8 -0.80 1.00 3.78, P < 0.01 0.82, P > 0.2 10.8, P< 0.001 Hippoglossoides platessoides 4 -1.30 0.58 ods of preservation (e.g. alcohol, freezing [Theilacker, 1990; Hjorleifsson and Klein-Macphee, 1992] ) have 1980; Fowler and Smith, 1983; Kruse and Dailey, shown that individual animals may increase in NOTE Pepin et a!.: Changes in the distribution of larval fish body length following preservation 639 0.45 0.40 0.35 § 0.30 (O *D | 0.25 0.20 0.15 0.10 0 2 4 6 8 10 12 14 16 Length interval (mm) Figure 4 Rank correlation (Kendall’s x) of preserved versus fresh lengths, within millimeter intervals of fresh lengths. Open symbols represent values not signifi- cantly different from 0, grey symbols present sig- nificant values (P<0.05), and black symbols repre- sent highly significant values (P<0.01). 9 o A •• ? i • \ • • \ 0 0 0 0 . O length following preservation. It is important to note, however, that although our results show that the mean change in length for larvae <6 mm SL was posi- tive, there were also numerous individuals in these same length intervals that either shrank or showed no change in body length after preservation. This finding underscores our contention that responses to preservation vary significantly at the level of the individual as well as among species. It is possible that the capture, sorting, and preser- vation approaches used in this study may have led to variations in the general pattern of changes in body length. Following capture, sorting on deck may have caused a change in water temperatures. This proce- dure may have precipitated the death of individual larvae and resulted in contraction of body tissue in the “fresh length” measurement and consequently may have produced an apparent increase in body length following remeasurement. The source of lar- vae may also be an important feature to consider. Most preservation studies are based on laboratory- reared animals rather than on field-caught speci- mens (Table 1). Laboratory-reared animals show greater changes in length than field caught speci- mens (Table 1), although there is some variation about this general pattern (e.g. Theilacker, 1986). Change in body length of marine fish larvae fol- lowing capture and preservation has been attributed mainly to a breakdown of the osmoregulatory capac- ity of larvae, leading to a net loss of water ( Ahlstrom, 1976). Simple laboratory-based studies of shrinkage have concluded that preservation results in relatively small changes in body length (see Hjorleifsson and Klein-Macphee, 1992) but that the impact of net-cap- ture and handling is likely to have the greatest im- pact on changes in body length (Theilacker, 1980; Jennings, 1991; Fox, 1996). Most laboratory simula- tions of capture used fixed handling times; however, under field conditions it is unknown when a larva enters the net. Thus correction factors for a fixed handling time may introduce substantial error. Our sampling methods differed dramatically from those normally employed in field situations in both the duration of the residence time of larvae in the net and in the aggressiveness of capture and handling. It is therefore likely that, in comparison with other studies, the effect of capture and handling in our study was small. However, despite the possible importance of capture and handling, the interpretation of the con- dition of individual larvae or the distribution of popu- lation characteristics must be approached with caution. Our results clearly show that the relative position of an individual within the size distribution of the sample is likely to change considerably following handling and preservation. Thus, attempts to cor- rect for the effects of preservation or handling on fresh larval lengths can, in certain types of analysis (e.g. individual-based approaches to the use of mor- phometric measurements), introduce significant and unpredictable bias in both the data and their inter- pretation. There appears to be no simple solution to this problem. Conclusions as to the overall vulner- ability of a fractional portion of a population to a given factor must consider the inherent variability in the distribution of measurements caused by op- erator, handling, and preservation-induced error. Acknowledgments We thank Tim Shears and Sarah Donald for providing valuable technical assistance. Additional assistance was provided by Steph Carter, Dave Bell, and the officers and crew of CSS Shamook. C. J. Fox and two anony- mous referees provided valuable criticisms of this study. Funding for this project was provided by the Depart- ment of Fisheries and Oceans (PP) and the Natural Sciences and Engineering Research Council (WCL). Literature cited Ahlstrom, E. H. 1976. Maintenance of quality in fish eggs and larvae col- lected during plankton hauls. In H. F. Steedman (ed.), Zooplankton fixation and preservation: monographs on 640 Fishery Bulletin 96(3), 1998 oceanographic methodology. UNESCO Press, Paris, vol. 4, p. 313-318. Bailey, K. M. 1982. The early life history of Pacific hake. Fish. Bull. 80:589-598. Blaxter, J. H. S. 1971. Feeding and condition of Clyde herring larvae. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 160:128-136. Farris, B. A. 1963. Shrinkage of sardine ( Sardinops caerulea ) larvae upon preservation in buffered formalin. Copeia 1963:185- 186. Fowler, G. M., and S. J. Smith. 1983. Length changes in silver hake (Merluccius bilinearis) larvae: effects of formalin, ethanol, and freezing. Can. J. Fish. Aquat. Sci. 40:866-870. Fox, C. J. 1996. Length changes in herring (Clupea harengus ) larvae: effects of capture and storage in formaldehyde and alcohol. J. Plankton Res. 18:483-493. Hay, D. E. 1982. Fixation shrinkage of herring larvae: effects of sa- linity, formalin concentration and other factors. Can. J. Fish. Aquat. Sci. 39:1138-1143. Hjorleifsson, E., and G. Klein-Macphee. 1992. Estimation of live standard length of winter floun- der Pleuronectes americanus larvae from formalin-pre- served, ethanol-preserved and frozen specimens. Mar. Ecol. Prog. Ser. 82:13-19. Houde, E. D. 1989. Comparative growth, mortality, and energetics of marine fish larvae: temperature and implied latitudinal effects. Fish. Bull. 87:471-495. Jennings, S. 1991. The effect of capture, net retention and preservation upon lengths of larval and juvenile bass, Dicentrarchus labrax ( L. ). J. Fish. Biol. 38:349-357. Johnston, T. A., and J. A. Mathias. 1993. Length reduction and dry weight loss in frozen and formalin-preserved larval walleye, Stizostedion vitreum (Mitchill). Aquatcult. Fish. Manage. 24:365-371. Karjalainen, J. 1992. Effects of different preservation methods on total length and weight of larval vendance ( Coregonus albula L ). Nordic J. Freshwater Res. 67:88-90. Kruse, G. H., and E. L. Dailey. 1990. Length changes in capelin, Mallotus villosus (Muller), larvae due to preservation in formalin and anhydrous alcohol. J. Fish Biol. 36:619-621. Miller, T. J., L. B. Crowder, J. A. Rice, and E. A. Marschall. 1988. Larval size and recruitment mechanisms in fishes: toward a conceptual framework. Can. J. Fish. Aquat. Sci. 45:1657-1670. O’Connell, C. P. 1981. Percentage of starving northern anchovy, Engraulis mordax , larvae in the sea as estimated by histological methods. Fish. Bull. 78:475-489. Otto, R. G., and J. A. Boggs. 1983. Fish eggs and larvae of the Chesapeake Bay: obser- vations on the problem of identifying field mortalities of stripped bass (Morone saxatilis ) larvae from preserved samples. Estuaries 6:167-171. Pepin, P. 1991. Effect of temperature and size on development, mor- tality, and survival rates of the pelagic early life history stages of marine fish. Can. J. Fish. Aquat. Sci. 48: 503-518. Pepin, P., and T. J. Miller. 1993. Potential use and abuse of general empirical models of early life history processes in fish. Can. J. Fish. Aquat. Sci. 50:1343-1345. Schnack, D., and H. Rosenthal. 1978. Shrinkage of Pacific herring larvae due to formalin fixation and preservation. Meeresforsch. 26:222-226. Takizawa, K., Y. Fujita, Y. Ogushi, and S. Matsumoto. 1994. Relative change in body length and weight in sev- eral fish larvae due to formalin fixation and preser- vation. Fish. Sci. 60:355-359. Theilacker, G. H. 1980. Changes in body measurements of larval northern anchovy, Engraulis mordax, and other fishes due to han- dling and preservation. Fish. Bull. 78:685-692. 1986. Starvation-induced mortality of young sea-caught jack mackerel, Trachurus symmetricus, determined with histological and morphological methods. Fish. Bull. 84: 1-17. Theilacker, G. H., and S. M. Porter. 1995. Condition of larval walleye pollock, Theragra chalco- gramma, in the western Gulf of Alaska assessed with his- tological and shrinkage indices. Fish. Bull. 93:333-344. Tucker, J. W., and A. J. Chester. 1984. Effects of salinity, formalin concentration and buffer on quality of preservation of southern flounder (Paralichthys lethostigma) larvae. Copeia 1984:981-988. Yin, M. C., and J. H. S. Blaxter. 1986. Morphological changes during growth and starvation of larval cod ( Gadus morhua L.) and flounder (Platichthys flesus L.). J. Exp. Mar. Biol. Ecol. 104:215-228. 641 Sampling juvenile skipjack tuna, Katsuwonus pelamis, and other tunas, Thunnus spp., using midwater trawls in the tropical western Pacific Toshiyuki Tanabe Tohoku National Fisheries Research Institute 3-27-5, Shinhama, Shiogama, Miyagi, 985-0001 Japan E-mail address: katsuwo@myg.affrc.goJp Kodo Niu Kasumi Senior High School 40-1, Yada, Kasumi, Kinosaki, Hyogo, 669-6563 Japan Skipjack tuna, Katsuwonus pelamis, a highly migratory species, is one of the most important stocks for commercial fisheries in the western Pacific. The spawning ground of this species extends throughout equatorial and subtropical waters but larvae are more abundant in lower than higher latitudes (Nishi- kawa et al., 1978). Young and pre- adult skipjack tuna migrate great distances seasonally between tropi- cal and temperate waters, yet little information is available for juvenile stages of this species. Exploitation of skipjack tuna in the western Pacific has increased in recent years; an annual catch of 940,000 metric tons was taken in 1993 (FAO, 1995). Stock assess- ment and management of the spe- cies require detailed biological in- formation on each life stage but, largely because of sampling difficul- ties, data on juveniles of skipjack tuna and other tunas are relatively limited compared with those on adults. Juveniles swim well and thus escape many gear: ring nets, beam trawls, or small pelagic trawls (see Methot, 1986). Other gear and light traps will not pro- vide adequate spatial or temporal coverage (Thorrold, 1993). Consequently development of appropriate sampling gear has been an important part of our study of the early life history of skipjack tuna and other tunas. Since 1992, we have worked to develop a new method for capturing juvenile skip- jack tuna and tuna spp., namely a midwater trawl net with a large mouth opening, capable of high- speed towing. This appeared to be an excellent gear for sampling ju- veniles. In a trial survey, we cap- tured large numbers of juvenile skipjack tuna and other tuna by using the midwater trawl net in offshore waters of the tropical west- ern Pacific. This study is based on 1992-94 research cruises and con- firms the effectiveness of the new midwater trawling gear. We de- scribe the specifications of this net and its effectiveness for collecting juvenile skipjack tuna and other tuna. Materials and methods The midwater trawl net Tansyu has a large mouth opening and was de- signed to be towed at high speed. It was developed in cooperation with Tohoku National Fisheries Research Institute and Nichimo Co. Ltd., Shimonoseki, Yamaguchi. Tar- gets for the Tansyu included skip- jack tuna and other tuna, ranging in size from 10 to 200 mm standard body length (SL). The goal was to capture sufficient numbers to char- acterize distribution and relative abundance. The Tansyu design was based on the midwater trawl net Yoko-2, which has been used for sampling sardine by Seikai Na- tional Fisheries Research Institute, Nagasaki (Takeshita et al., 1988). The total length of the net was 7 1.6 m and that of the headrope and foot- rope was 38.6 m (see Fig. 1). The net was composed of 8 wing pan- els, 12 body panels, and 6 codend panels; all were made of 380 D fine polyethylene ropes. Twine diameter was between 1.91 mm (P-30) and 5.06 mm (P-210). The stretched mesh size was 1000 mm at the wings and the first panels of the body, diminishing successively to 57 mm in the seventh panels of the body. The dimension of the fishing circle was 144 x 1000 mm (where 144 = no. of meshes at mouth of net and 1000 m = mesh size at first panel of net). A codend of 60- mm mesh size was attached to the end of the body. An inner net of 30-mm and 8-mm mesh size was put inside of the codend to collect samples. At- tached to the headrope were 74 floats of 300-mm diameter. A steel chain of 28-mm diameter was at- tached to the footrope. Buoyancy of the floats and weight of the chain were 762.4 kg and 660.4 kg, respec- tively. The bridles to the bottom, center, and upper wings were 18 mm in diameter and 100 m long. The expected mouth opening was approximately 20 m wide, 18 m high, at a towing speed of 4.5 knots. The net was designed for a maxi- mum towing speed of 5 knots. The Tansyu was designed to be used with 1.7 m x 2.8 m otter doors. Manuscript accepted 10 November 1997. Fishery Bulletin 96:641-646 ( 1998). 642 Fishery Bulletin 96(3), 1998 Figure 1 A schematic diagram of the midwater trawl net Tansyu that was devel- oped to collect juvenile skipjack tuna and other tunas. The numerals on the left side indicate the length of each panel. The numerals within the diagram indicate twine size and mesh size of the net. A scale model of the Tansyu was con- structed and tested in a tank at Nichimo Co. Ltd. to test its design and performance. Research cruises were carried out from late October to early December, 1992-94. The survey covered an area from offshore of Palau past the Calorine Islands in the tropical western Pacific (Fig. 2). A 400-ton class stern trawler was chartered by the Fisheries Agency, Ministry of Agriculture, Forestry, and Fisheries, Tokyo, to conduct the mid- water trawl trials. Midwater trawling was typically con- ducted four times daily; standard du- ration of towing was approximately one hour. Towing time was defined as the duration from beginning of the net tow to commencement of warp hauling. Ten strata were occupied from near surface to 200 m, in 20-m increments of water depth; separate tows in each stratum were carried out. Depth of the net was established by noting the warp length, and towing speed was usually set 4 to 5 knots against the currents. A net-depth recorder (Furuno FNR-200) was used to monitor continuously the mouth open- ing and depth of the net throughout the trawling operation. As soon as the net was brought on deck, the collection from the codend was weighed as whole wet weight; skipjack tuna and other tunas were sorted from the collection. Samples were fixed in 10% buffered formalin and preserved in 80% ethanol. All skipjack tuna speci- mens were identified after the cruise following the methods of Matsumoto et al. (1984). The species of tunas were identified as Thunnus spp. and samples of skipjack tuna and other tunas were measured individually to the nearest 0.1 mm SL with calipers. In this report, the early life stages of skipjack tuna were defined as follows: larva: from hatching to less than 10 mm SL; juvenile: from 10 mm SL to less than 100 mm SL; and young: from 100 mm SL to less than 300 mm SL. Results The new trawl was deployed a total of 327 tows throughout the three years of the study. The target taxa, skipjack tuna and other tunas, were captured in 163 tows; skipjack tuna and other tunas co-oc- curred in 66 tows. Over four thousand skipjack tuna and other tunas were collected (Table 1). The inci- dence of juveniles collected remained high in all years. The maximum number of specimens collected in a single tow was over 100 juveniles of both taxa. The diel distribution of sampling effort was 41.8% during daytime, 58.2% at nighttime (Table 2). For skipjack tuna, the mean daytime occurrence was about 5% higher than that at nighttime, and the mean daylight catch per tow was about four indi- viduals greater than that at nighttime; a similar pattern was observed for other tuna, but the differ- ences were not statistically significant. The whole wet weight for nighttime usually tended to be larger NOTE Tanabe and Niu: Sampling juvenile Katsuwonus pelamis and Thunnus spp 643 than that for daytime (Table 2). Other dominant species collected at nighttime were myctophids, cephalopods, and euphausids; those taken in daylight were engraulids and acanthurids. Size classes of skipjack tuna and other tunas were widely represented; life stages from postlarva to early young were apparent. Where SL could be measured, specimens represented 92.2% of skipjack tuna, 93.5% of other tunas. Skipjack tuna SL ranged from 7.1 mm to 171.6 mm (Fig. 3); other tunas ranged from 8.0 mm to 139.8 mm (Fig. 4). The mean SL was 25.7 mm for skipjack tuna, 27.4 mm for other tunas, with modes at 10-20 mm and 20-30 mm, respectively. Juveniles were abundant and were the dominant life stage of both species in our samples. The composition of larvae, juvenile, and young was 1.6%, 97.6%, and 0.8% for skipjack tuna, 0.7%, 98.6%, and 0.7% for other tunas. Specimens of both taxa col- lected at night tended to be larger than those collected during daylight. For skip- jack tuna, the size of specimens during daytime ranged from 7.1 mm to 81.8 mm (average 23.8 mm); at nighttime lengths ranged from 8.1 mm to 171.6 mm (aver- age 27.7 mm). Tunas captured in daytime ranged from 8.0 mm to 54.7 mm (aver- age 22.6 mm), at night from 11.0 mm to 139.8 mm (average 31.7 mm). There was a significant difference in the average SL of both species between daytime and nighttime (Cochran-Cox t-test, P<0.05). Young skipjack tuna and other tunas were collected only at night and the num- ber of specimens per tow was less than six. Juveniles were as just as abundant at day as at night. Figure 2 The sampling area. Stations (•) on the cruise tracks ( — ) indicate loca- tions where the midwater trawls were deployed. 0 20 40 60 30 100 120 140 160 180 Standard length (mm) Figure 3 Length -frequency distribution of K. pelamis collected by the midwater trawl net in 1992-94 trial surveys. Discussion The midwater trawl net Tansyu made it possible to collect large numbers of juvenile skipjack tuna and other tunas during both day and night; this was not possible with the sampling gears previously avail- able. The results of our survey demonstrate that the Tansyu was effective in collecting juveniles of skip- jack tuna and other tunas. In previous studies, various sampling gears were used for sampling larvae of skipjack tuna and other tunas (Table 3), the results of which demonstrate that Table 1 Results of collection of juveniles and young K. pelamis and Thunnus spp. by the midwater trawl net in 1992-94. Inds. = individuals. Occurrencei Total catch Catch per tow2 Species % (inds.) (inds.) K. pelamis 49.8 3218 9.8 Thunnus spp. 27.5 1074 3.3 1 Number of tows in which juveniles were caught/total tows x 100. 2 Total catch of juveniles/total tows. 644 Fishery Bulletin 96(3), 1998 the maximum size of skipjack tuna and other tunas that could be collected with smaller plankton nets was about 12 mm; the usual size was 3 to 5 mm. Thus skipjack tuna and other tunas larger than 10 mm could easily escape from these gears. Davis et al. (1990) pointed out that net avoidance by larvae should be considered in calculating estimates of abundance of larval tunas. Sampling efforts to collect juveniles (larger than 10 mm) have been reported in a few studies; King and Iversen ( 1962) tried to collect juvenile tunas with four kinds of trawl nets in the central Pacific, but only six juvenile tunas, size 18 to 60 mm, were col- lected. Higgins (1970) tried to collect juvenile tunas in Hawaiian waters using a midwater trawl net with a mouth opening 12 m x 8 m. He collected 578 skip- jack tuna and 417 other tunas. Most samples rarely had juvenile stages of skipjack tuna and other tu- nas. Takuno and Ueyanagi (1978) tried to capture juveniles with a small pelagic trawl net in the tropi- cal western Pacific, collecting 20 skipjack tuna and six yellowfin tuna, from 6 mm to 31 mm. These vari- ous results suggest that juvenile skipjack tuna and other tunas avoided the small trawl nets. The mouth opening and the towing speed of the Tansyu is much greater than those of small trawl nets used in previ- ous studies. We therefore expected that the size of speci- mens that would be collected would be larger than those collected in previous studies if the survey area and pe- riod were chosen appropriately. Indeed, the size range of skipjack tuna and other tunas collected with the Tansyu was much greater than that collected with smaller nets. In addition, large numbers of juvenile skipjack tuna and other tunas could be collected. Skipjack tuna and other tunas that reach the young stage are able to swim much faster than juveniles. In this study, the number of young skipjack tuna and other tunas collected was low, and these fish were caught only at night. These results indicate that net avoidance by young fish affected the number of speci- mens captured. The results of previous studies on swimming speeds of scombrids demonstrate that the “burst swimming speed” of adult tunas is usually from 10 to 20 fork lengths (FL) per second (Mag- nuson, 1978). If the burst swimming speed of young tunas (which might swim slower than adults) is, for discussion, assumed to be 10 FL per sec- ond, the speed at 20 cm FL would be estimated at approximately 3.9 knots. Therefore, the nets would need to be towed faster than 4 knots in order to sample young tunas of 20 cm FL. How- ever, few young skipjack tuna and other tunas were collected with the Tansyu with its maximum towing speed of 5 knots. If large numbers of young skip- jack tuna and other tunas were able to be collected, we would learn more about the swimming speed of young life stages of these fish and also their ecological char- acteristics, for example their swimming behavior and vertical distribution. Other methods exist for collecting ju- venile and young tunas, i.e. by means of light traps (Thorrold, 1993) or from stomach contents of tunas and billfish. 50 0 20 40 60 80 100 120 140 160 180 Standard length (mm) Figure 4 Length-frequency distribution of Thunnus spp. collected by the midwater trawl net in 1992-94 trial surveys. Table 2 Comparison of catches between daytime and nightime tows foriC pelamis and Thunnus spp. and total wet weight (g) of collections. K. pelamis Thunnus spp. Average catch (g) Occurrence Catch per tow (%) (inds.) Occurrence Catch per tow (%) (inds.) Day Night 5405.5 7010.1 52.9 12.5 48.1 8.1 28.7 3.7 27.0 3.0 NOTE Tanabe and Niu. Sampling juvenile Katsuwonus pelamis and Thunnus spp. 645 Table 3 References for sampling larvae and juvenile skipjack tuna and other tunas. Author(s) Sampling gear Size of sample Strasburg, 1960 1-m ring net 3-12 mm larvae Nishikawa et al., 1978 2-m ring net and 1.4-m ring net <12 mm larvae Davis et al., 1990 0.7-m bongo net and 0.7 ring net 3-11 mm larvae Boehlert and Mundy, 1994 1-m2 MOCNESS 2-8 mm larvae2 King and Iversen, 1962 1-m ring net 6-ft Isaccs-Kidd trawl 6-ft Isaacs-Kidd trawl 10-ft Isaacs-Kidd trawl 18-60 mm juveniles Higgins, 1970 anchovy no. 2 Cobb pelagic trawl 7-47 mm larvae and juveniles Takuno and Ueyanagi, 1978 midwater trawl 6-31 mm larvae and juveniles Present study midwater trawl 7.1-171.6 mm larvae, juveniles, and young fish Thorrold, 1993 light trap 10-30 mm juveniles ‘ Boehlert, G. 1997. Natl. Mar. Fish. Serv., NOAA, 1352 Lighthouse Ave., Pacific Grove, CA 93950. Personal commun. Yoshida (1971) studied the early life history of skip- jack tuna using 1742 juveniles collected from the stom- achs of billfishes in Hawaiian waters and the central South Pacific. Mori (1972) reported on the geographi- cal distribution and the relative abundance of juve- nile and young skipjack tuna based on collections from stomachs of tunas and billfishes in the Pacific, Indian, and Atlantic Oceans. This method was con- venient and easy for sampling but produced results that should be carefully considered because they were not based on direct sampling from the habitat. Iizuka et al. ( 1989) using pelagic gill nets conducted research to collect young skipjack in southern Micronesian wa- ters over a period of three years, collecting 49 young skipjack, 120 to 280 mm. The research resulted in small numbers of specimens of young skipjack tuna, and it was not clear whether the number of samples indicated the abundance of young stages. The results of our study, in light of previous stud- ies sampling skipjack tuna and other tunas, show that a wider size range of juveniles can be collected by using a trawl net with a larger mouth opening and by using a high towing speed, as with the Tansyu. A small ring net is appropriate for research based on sampling larvae because of its simplicity in opera- tion. However, if it is necessary that diel patterns of vertical distribution of juvenile skipjack and other tunas be investigated, then the appropriate time and depth of towing should be selected, because ecologi- cal characteristics drastically affect the quantity of collections. Sampling of young skipjack tuna and other tunas, on the other hand, should be conducted at night with a trawl net with a large mouth open- ing or with a net that would minimize net avoidance by young-stage fish. Acknowledgments This research was funded by the Fisheries Agency of Japan with cooperation of the Palau Maritime and Micronesian Maritime Authorities. The authors thank the crew of RV Tanshu Maru and Omi Maru for their assistance in collecting samples. Y. Nishi- kawa, S. Cho, and S. Ueyanagi helped identify juve- niles of skipjack and tunas and A. Naganuma and Y. Watanabe provided many helpful suggestions during the period of this research. We thank M. Ogura, K. Teshima, and G. Boehlert for reviewing the manuscript. Literature cited Boehlert, G. W., and B. C. Mundy. 1994. Vertical and onshore-offshore distributional patterns of tuna larvae in relation to physical habitat features. Mar. Ecol Prog. Ser. 107:1-13. Davis, T. L. O., G. P. Jenkins, and J. W. Young. 1990. Diel patterns of vertical distribution in larvae of southern bluefin Thunnus maccoyii , and other tuna in the East Indian Ocean. Mar. Ecol. Prog. Ser. 59: 63-74. FAO (Food and Agriculture Organization). 1995. FAO yearbook, fishery statistics, catches and land- ings, vol. 7. Food and Agriculture Organization of the United Nations, Rome, 677 p. Higgins, B. E. 1970. Juvenile tunas collected by midwater trawling in Hawaiian waters, July-September 1967. Trans. Am. Fish. Soc. 99:60-69. Iizuka, K., M. Asano, and A. Naganuma. 1989. Feeding habits of skipjack tuna ( Katsuwonus pelamis Linnaeus) caught by pole and line and the state of young skipjack tuna distribution in the tropical seas of the West- ern Pacific Ocean. Bull. Tohoku Reg. Fish. Res. Lab. 51:107-116. 646 Fishery Bulletin 96(3), I 998 King, J. E, and R. T. B. Iversen. 1962. Midwater trawling for forage organisms in the cen- tral Pacific, 1951-1956. Fish. Bull. 62:271-321. Magnuson, J. J. 1978. Locomotion by scombrid fishes:hydromechanics, mor- phology, and behavior. In W. S. Hoar and D. J. Randall (eds.), Fish physiology, vol. 7, Locomotion, p. 239- 313. Academic Press, New York, NY. Matsumoto, W. M., R. A. Skillman, and A. E. Dizon. 1984. Synopsis of biological data on skipjack tuna, Katsu- wonus pelamis. U.S. Dep. Commer, NOAA Tech. Rep. NMFS Circ. 451, 92 p. Methot, R. D. 1986. Frame trawl for sampling pelagic juvenile fish. CalCOFI Rep. 27:267-278. Mori, K. 1972. Geographical distribution and relative apparent abundance of some scombrid fishes based on the occur- rences in the stomachs of apex predators caught on tuna longline. 1: Juvenile and young of skipjack tuna (Katsu- wonus pelamis). Bull. Far Seas Fish. Res. Lab. 6:111-157. Nishikawa, Y., S. Kikawa, M. Honma, and S. Ueyanagi. 1978. Distribution atlas of larval tunas, billfishes and re- lated species: results of larval surveys by R/V Shunyo Maru, and Shoyo Maru, 1956-1975. Far Seas Fish. Res. Lab. S. ser. 9, 99 p. Strasburg, D. 1960. Estimates of larval tuna abundance in the central Pacific. Fish. Bull. 60: 231-255. Takeshita, K., N. Ogawa, T. Mitani, R. Hamada, E. Inui, and K. Kubota. 1988. Acoustic survey of spawning sardine, Sardinops melanosticta in the coastal waters of west Japan. Bull. Seikai Reg. Fish. Res. Lab. 66:101-117. Takuno, H., and S. Ueyanagi. 1978. Results of midwater trawl investigations for larval fishes in the tropical Western Pacific. Far Seas Fish. Res. Lab., Shimizu, 55 p. Thorrold, S. R. 1993. Post-larval and juvenile scombrids captured in light traps - Preliminaly results from the central Great Barrier Reef lagoon. Bull. Mar. Sci. 52:631-641. Yoshida, H. 1971. The early life history of skipjack tuna, Katsuwonus pelamis, in the Pacific Ocean. Fish. Bull. 69:545-554. 647 Entanglement and mortality of bottlenose dolphins, Tursiops runcatus, in recreational fishing gear in Florida Randall S. Wells Chicago Zoological Society, c/o Mote Marine Laboratory 1600 Ken Thompson Parkway, Sarasota, Florida 34236 E-maii address, rwells@mote.org Suzanne Hofmann Mote Marine Laboratory 1600 Ken Thompson Parkway, Sarasota, Florida 34236 Tristen L. Moors Chicago Zoological Society, c/o Mote Marine Laboratory 1600 Ken Thompson Parkway, Sarasota, Florida 34236 Effects of fishing gear interactions are among the most pressing issues currently being addressed by ma- rine mammal management agen- cies in the United States. Informa- tion is needed on the rates and fates of entangled dolphins for evalua- tion of mortality caused by differ- ent fisheries. Virtually the entire emphasis of management agencies has been on marine mammal inter- actions with commercial fisheries; little notice has been taken of the impacts of recreational fisheries. Along the central-west coast of Florida, near Sarasota, ongoing re- search initiated in 1970 on a resi- dent community of about 100 bottlenose dolphins, Tursiops trun- catus (Scott et al., 1990; Wells, 1991) has resulted in opportunities to examine closely the effects of in- teractions between dolphins and human activities, including both commercial and recreational fish- eries (e.g. Wells and Scott, 1994; 1997). We report here on one case in which we removed a large quan- tity of fishing line trailing from a free-swimming bottlenose dolphin, and we include information on the behavior and health of the animal over the first year following en- tanglement and removal of the gear. We also compare the extent of commercial vs. recreational fish- ing interactions with dolphins. On 4 June 1996, a seven-year-old female bottlenose dolphin (“FB03”) was observed traveling alone near shore in the Gulf of Mexico off Anna Maria Island (27°33'N, 82°45'W), trailing large quantities of thin line that extended in clumps to 5 m be- hind the animal. The line was wrapped around the tail flukes and peduncle but did not appear to af- fect noticeably the dolphin’s swim- ming ability. The dolphin (FB03) was known from birth as a mem- ber of a four-generation maternal lineage (Wells et al., 1987; Wells, 1991). She had been seen 146 times in this area prior to this incident, and was within her well-estab- lished home range. She had been observed six days earlier without any line. This dolphin was seen next at 10:28 on 6 June 1996, in Palma Sola Bay (27°29'N, 82°40'W), several miles from her 4 June location and still within her home range. She was alone and surfacing slowly, still trailing the line. She was swim- ming more slowly than during the previous observation (as measured in relation to boat speed), and the clumps of line extended farther be- hind her, to a distance of about 8 m. Ventilations occurred at average intervals of 35.7 sec (±12.71 sec SD, n= 58), comparable to previously measured rates for Sarasota dol- phins (Irvine et al., 1981; Waples, 1995 ). The line was cutting into tis- sues of her flukes and peduncle. We decided to try to remove the line, assembled a team by 12:47, and approached the dolphin in our 6-m long research vessel powered by a 115-hp outboard engine. Because the trailing line followed the dol- phin to the surface at each breath, we were able to reach and retrieve most of it. At 13:10 we secured the line with a boat hook, and cut all but about 2 or 3 m of line from the animal as she continued swimming. The dolphin swam rapidly and “porpoised” for about 3.7 km, left the bay, and continued to the south. We approached her again in shal- low water (<2 m deep) and she maintained position, riding slowly beneath the bow of our vessel for the next 38 minutes; this is the longest recorded instance of bow- riding in the history of our research program. Her position beneath the bow afforded us an opportunity to examine closely the wounds and remaining line. Cuts around the peduncle at the insertion of the flukes and through the anterior edges of the flukes were evident. The remaining line circled the pe- duncle once and two freely trailing ends were draped across and ex- tended about 1.3 m behind the Manuscript accepted 26 February 1998. Fishery Bulletin 96:647-650 (1998). 648 Fishery Bulletin 96(3), 1998 Figure 1 (A) Dorsal view of the scars on peduncle and flukes of the entangled dolphin (FB03) approximately one year after removal of the fishing line. (B) Ventral view of the scars on peduncle and flukes of the same entangled dolphin (FB03) approximately one year after removal of the fishing line. flukes. We tried unsuccess- fully a number of times to grasp the remaining line with the boat hook as the dol- phin rode the bow wave. Throughout these attempts, the dolphin held her position directly below the bow, clearly watching the boat hook as it moved within several cm of her flukes, and she made no effort to move away. The next observation of the dolphin was on 18 June 1996, on Sister Key Flats (27°27'N, 82°39'W ), near where we had left her on 6 June. She was swimming normally and no line was evident, although in- dications of scarring were seen where the line had been. Through 1 July 1997, she was observed 33 more times and her behavior was consid- ered normal each time. She was alone each time that she was seen while entangled; she was seen alone in only 24% of her unentangled sight- ings. Prior to entanglement, she was found in groups of 4.9 dolphins on average (SD=4.32, n= 73). Following removal of the line, she was in groups of 4.4 dolphins on average (SD= 4.56, n= 30). On 13 June 1997, this dol- phin was examined by veteri- narians as part of a dolphin health assessment program. The wounds from the line were well healed, but deep scars remained (Fig. 1). Her health was comparable to that at the time of her previ- ous examination on 17 June 1994. Her weight (132 kg) was 19 kg greater than three years before. Veteri- nary staff considered her blood chemistry and he- matology values to be within acceptable limits, and results of ultrasonic examination of organ condition were unremarkable. Thus, it appears that no long- term effects beyond scarring resulted from this rela- tively brief entanglement event. The depth of scars, however, suggests that her flukes could potentially have been severed if the line had not been removed. The line was determined to be 80-pound (approx.) test “squidding line,” a floating-core braided Dacron line often used locally for tarpon fishing from late spring to early summer. Tarpon fishing occurs in the immediate Gulf coastal waters, especially in the area where this dolphin was first seen with the line around her flukes. A large number of snarls and cuts in the 484 m of line NOTE Wells et a!.: Entanglement and mortality of Tursiops truncatus 649 removed suggested that it had been removed from a reel and discarded, rather than it had been actively involved in fishing at the time of entanglement. Recreational fishing activities pose a largely ig- nored threat to bottlenose dolphins near Sarasota. Of 11 carcasses of resident Sarasota Bay dolphins recovered during 1993-1996, monofilament fishing line was implicated as a contributing factor in three deaths (27%). Large quantities of fishing line were found around two dependent calves (MML 9314=1.5 yr old and MML 9417=0.3 yr old). In one of these cases (MML 9314), growth of invertebrates on the line indicated that the line had been discarded prior to entanglement. The third case involved ingestion of line by an adult female (MML 9514) that had con- sumed a hooked sheepshead ( Archosargus probato- cephalus), as reported by Gorzelany (in press). Wells and Scott (1994) reported that a dispropor- tionately large number of subadult Sarasota Bay resident dolphins involved in entanglements of all kinds had been recorded through 1989 on the basis of scars observed during veterinary examinations as part of a capture-release program. More than half of the cases involved subadults; the balance were re- corded from adults, but the entanglement events had occurred at an undetermined younger age. Dolphin FB03, a subadult at the time of entanglement; and the two dependent calves described above lend fur- ther support to this pattern. Similarly, Mann et al. (1995) reported on four cases of infant bottlenose dolphins becoming entangled in Shark Bay, Austra- lia. In three of these cases, it was possible to remove the line from the semiprovisioned calves; in the fourth case, the line came free from the dolphin without human assistance. For many young animals, curios- ity, inexperience, and unrefined motor skills place them at greater risk of entanglement through play- ful and exploratory behavior, the occurrence of which declines as the animal matures. Play may be strongly related to foraging behavior, for example, and may allow younger animals to practice, learn, and develop other behaviors that will be essential to their sur- vival. Such behaviors may have an important role in the development of adult behaviors, but they may also be a costly practice for newborns or subadults who lack the experience of an adult. In many parts of the world, dolphins are killed in gillnet and other fisheries (Perrin et ah, 1994). It is interesting to note that in Sarasota Bay, Florida, an area of heavy recreational fishing activity, the num- bers of deaths or serious injuries resulting from rec- reational fishing could exceed historical levels from small-scale gillnet fisheries. Gill nets were used ex- tensively in this area prior to a state-wide ban, July 1995. Three deaths and one rescue of Sarasota resi- dent dolphins entangled in recreational fishing gear occurred during 1993-96, but only one death (FB20 in 1976) and one rescue from gill nets (FB11 in 1985) were recorded during the 20 years prior to the 1995 gillnet ban. It should be noted, however, that strand- ing response coverage was uneven prior to the mid- 1980s: not all deaths resulted in recovered carcasses, and not all possible entanglement events could be clearly identified as a direct cause of death, nor could they necessarily be distinguished as net or line entanglements. Mortality and serious injury to dolphins from rec- reational fisheries have been largely overlooked in management, yet such mortality and injury may be important. Management actions to reduce human- related dolphin mortality are needed to address such issues, particularly in regard to the practice of dis- carding fishing line. Discarded line poses a risk that is somewhat analogous to that of “ghost-fishing” by commercial nets. Increased education of fishermen, through clear descriptions of the documented conse- quences of discarded gear is a logical, important, approach to the issue. Acknowledgments We thank the Center for Field Research ( Earthwatch), the Chicago Zoological Society, Dolphin Quest, Dolphin Biology Research Institute, Mote Marine Laboratory, the Henry Foundation, and the dedicated efforts of our field team members for the rescue, monitoring, and subsequent examination of FB03. Observations were conducted under National Marine Fisheries Service Scientific Research Permit 805; physical ex- amination was conducted under National Marine Fisheries Service Scientific Research Permit 945. Finally, we thank M. Scott, A. Read, K. Urian, and J. Gorzelany for comments on the manuscript. Literature cited Gorzelany, J. F. In press. Unusual deaths of two free-ranging Atlantic bottle- nose dolphins ( Tursiops truncatus ) related to ingestion of rec- reational fishing gear. Marine Mammal Science 14(3). Irvine, A. B., M. D. Scott, R. S. Wells, and J. H. Kaufmann. 1981. Movements and activities of the Atlantic bottlenose dolphin, Tursiops truncatus, near Sarasota, Florida. Fish. Bull. 79:671-688. Mann, J. , R. A. Smolker, and B. B. Smuts. 1995. Responses to calf entanglement in free-ranging bottle- nose dolphins. Mar. Mamm. Sci. 11:100-106. Perrin, W. F., G. P. Donovan, and J. Barlow (eds.). 1994. Gillnets and cetaceans. Rep. Int. Whal. Comm, (spe- cial issue 15), 629 p. 650 Fishery Bulletin 96(3), 1998 Scott, M. D., R. S. Wells, and A. B. Irvine. 1990. A long-term study of bottlenose dolphins on the west coast of Florida. In S. Leatherwood and R. R. Reeves (eds.), The bottlenose dolphin, p. 235-244. Academic Press, San Diego, CA, 653 p. Waples, D. M. 1995. Activity budgets of free-ranging bottlenose dolphins ( Tursiops truncatus) in Sarasota Bay. M.Sc. thesis, Univ. California, Santa Cruz, CA, 61 p. Wells, R. S. 1991. The role of long-term study in understanding the so- cial structure of a bottlenose dolphin community. In K. Pryor and K. S. Norris (eds.), Dolphin societies: discover- ies and puzzles, p. 199-225. Univ. California Press, Ber- keley, CA, 397 p. Wells, R. S., and M. D. Scott. 1994. Incidence of gear entanglement for resident inshore bottle- nose dolphins near Sarasota, Florida. In W. F. Perrin, G. P. Donovan, and J. Barlow, (eds. ), Gillnets and cetaceans, p. 629. 1997. Seasonal incidence of boat strikes on bottlenose dol- phins near Sarasota, Florida. Mar. Mamm. Sci. 13:475- 480. Wells, R. S., M. D. Scott, and A. B. Irvine. 1987. The social structure of free-ranging bottlenose dolphins. In H. Genoways (ed.). Current mammalogy, vol. 1, p. 247-305. Plenum Press, New York, NY, 519 p. 651 Publications Awards, 1 996 National Marine Fisheries Service, NOAA The publications Advisory Committee of the National Marine Fisheries Service is pleased to announce the awards for best publication authored by NMFS scientists and published in the Fishery Bulletin, volume 94, and in the Murine Fisheries Review, volume 58. Eligible papers are nominated by the Fisheries Science Centers and Regional Offices and are judged by the NMFS Editorial Board. Only articles that significantly contribute to the understanding and knowledge of NMFS-related studies are eligible. We offer congratulations to the following authors for their outstanding efforts Fishery Bulletin , volume 94 Outstanding publication Marine Fisheries Review, volume 58 Outstanding publication Collette, Bruce B., and Christopher R. Aadland. Revision of the frigate tunas (Scombridae, Auxis ), with de- scriptions of two new subspecies from the eastern Pacific. Fish. Bull. 94(31:423-441. MacKenzie, Clyde L., Jr. History of oystering in the United States and Canada, fea- turing the eight greatest oyster estuaries. Mar. Fish. Rev. 58(41:1-78. 652 Fishery Bulletin 96(3), 1998 Superintendent of Documents Publications Order Form *5178 □yes, please send me the following publications: Subcriptions to Fishery Bulletin for $35.00 per year ($43.75 foreign) The total cost of my order is $ . Prices include regular domestic postage and handling and are subject to change. (Company or Personal Name) (Please type or print) (Additional address/attention line) (Street address) (City, State, ZIP Code) Charge your order. It’s Easy! (Daytime phone including area code) (Purchase Order No.) Please Choose Method of Payment: I I Check Payable to the Superintendent of Documents j GPO Deposit Account [ | | | | | | ~| — Q i ] VISA or MasterCard Account To fax your orders (202) 512-2250 (Credit card expiration date) (Authorizing Signature) Mail To: Superintendent of Documents P.O. Box 371954, Pittsburgh, PA 15250-7954 Thank you for your order! Fishery Bulletin Guide for Contributors Content Articles published in Fishery Bulletin de- scribe original research in fishery marine science, engineering and economics, and the environmental and ecological sciences, in- "bluding modeling. Articles may range from relatively short to extensive; notes are re- ports of 5 to 10 pages without an abstract and describing methods or results not sup- ported by a large body of data. Although all contributions are subject to peer review, responsibility for the contents of papers rests upon the authors and not upon the editor or the publisher. It is therefore im- portant that the contents of the manuscript are carefully considered by the authors. Submission of an article is understood to imply that the article is original and is not being considered for publication elsewhere. Manuscripts should be written in English. Authors whose native language is not En- glish are strongly advised to have their manuscripts checked by English-speaking colleagues prior to submission. Preparation Title page should include authors’ full names and mailing addresses, the corre- sponding author’s telephone, FAX number, and E-mail address, and a list of key words to describe the contents of the manuscript. Abstract should not exceed one double- spaced typed page. It should state the main scope of the research but emphasize its con- clusions and relevant findings. Because ab- stracts are circulated by abstracting agencies, it is important that they represent the research clearly and concisely. Text must be typed double-spaced through- out. A brief introduction should portray the broad significance of the paper; the remain- der of the paper should be divided into the following sections: Materials and meth- ods, Results, Discussion (or Conclusions), and Acknowledgments. Headings within each section must be short, reflect a logical sequence, and follow the rules of multiple subdivision (i.e. there can be no subdivision without at least two items). The entire text should be intelligible to interdisciplinary readers; therefore, all acronyms, abbrevia- tions, and technical terms should be spelled * U S. Go v. Printing Office 1998 689-0 1 5/80002 out the first time they are mentioned. The scientific names of species must be written out the first time they are mentioned; sub- sequent mention of scientific names may be abbreviated. Follow the U.S. Government Printing Office Style Manual (1984 ed.) and the CBE Scientific Style and Format (6th ed.) for editorial style, and the most current issue of the American Fisheries Society’s Common and Scientific Names of Fishes from the United States and Canada for fish nomenclature. Dates should be written as follows: 11 November 1991. Measurements should be expressed in metric units, e.g., metric tons as (t); if other units of measure- ment are used, please make this fact explicit to the reader. The numeral one (1) should be typed as a one, not as a lower-case el (1). Use of appendices is discouraged. Text footnotes should be numbered with Arabic numerals. Footnote all personal communications, unpublished data, and un- published manuscripts with full address of the communicator or author, or, as in the case of unpublished data, where the data are on file. Authors are advised to avoid ref- erences to nonstandard (gray) literature, such as internal, project, processed, or ad- ministrative reports. Where these refer- ences are used, please include whether they are available from NTIS (National Techni- cal Information Service) or from some other public depository. Literature cited comprises published works and those accepted for publication in peer- reviewed literature (in press). Follow the name and year system for citation format. In the text, cite as follows: Smith and Jones (1977) or (Smith and Jones, 1977). If there is a sequence of citations, list by year: Smith, 1932; Smith and Jones, 1985; Smith and Allen, 1986. Abbreviations of serials should conform to abbreviations given in Serial Sources for the BIOSIS Previews Database. Authors are responsible for the accuracy and completeness of all citations. Tables should not be excessive in size and must be cited in numerical order in the text. Headings should be short but ample enough to allow the table to be intelligible on its own. All unusual symbols must be ex- plained in the table legend. Other inciden- tal comments may be footnoted with italic numerals. Use the asterisk only to indicate probability in statistical data. Because ta- bles are typeset, they need only be submit- ted typed and formatted, with double-spaced legends. Zeros should precede all decimal points for values less than one. Figures must be cited in numerical order in the text. The senior author’s last name and the figure number should be written on the back of each one. Hand-drawn illustrations should be submitted as originals and not as photocopies. Submit photographs as glossy prints or slides that show good contrast, otherwise we cannot guarantee a good final printed copy. Graphic illustrations should be submitted as laser-printed copies, not as photocopies. Label all figures with Helve- tica typeface and capitalize the first letter of the first word in axis labels. Do not use boldface. Italicize species name and vari- ables in equations. Use zeros before all deci- mal points. Use uppercase Times Roman bold typeface to label the parts of a figure, e.g. A, B, C, etc. Send original figures to the Scientific Editor when the manuscript has been accepted for publication. Each figure legend should explain all symbols and ab- breviations in the figure and should be dou- ble-spaced and placed at the end of the manuscript. Copyright law does not cover government publications; they fall within the public do- main. If an author reproduces any part of a government publication in his work, refer- ence to source is appreciated. Submission Send printed copies (original and three cop- ies without staples) to the Scientific Editor: Dr. John B. Pearce, Scientific Editor Northeast Fisheries Science Center National Marine Fisheries Service 166 Water Street Woods Hole, MA 02543-1097 Once the manuscript has been accepted for publication, you will be asked to submit a software copy of your manuscript to the Man- aging Editor. The software copy should be submitted in WordPerfect or Microsoft Word text format and placed on a 3.5-inch disk that is double-sided, double or high density, and that is compatible with either DOS or Apple Macintosh systems. A copy of page proofs will be sent to the au- thor for final approval prior to publication. Copies of published articles and notes are available free of charge to the senior author (50 copies) and to his or her laboratory (50 copies). Additional copies may be purchased in lots of 100 when the author receives page proofs. tvs- §40 8 c o O More-2 Rock-1 Rock-2 Bow-2 Area More-2 Rock-1 Rock-2 Bow-2 Area Li More-2 Rock-1 Rock-2 Bow-2 Area Mg (F=14 43, P<0.0001) 22 17 c 12 8 c o o More-2 Rock-1 Rock-2 Bow-2 Area Mn (F=5.68, P<0.002) „ 3 CD a E More-2 Rock-1 Rock-2 Bow-2 Area Na (F=20.90, P<0.0001) More-2 Rock-1 Rock-2 Bow-2 Area P (F=1 7.21 , P<0 0001) More-2 Rock-1 Rock-2 Bow-2 Area 150 TO O) ■§• 130 c 110 More-2 Rock-1 Rock-2 Bow-2 Area S (F=2.20, N.S.) „ 580 CT) I ■§• 500 f 420 Sr (F=1.44, N.S.) _ 2600, 2400 c 2200i o O More-2 Rock-1 Rock-2 Bow-2 Area More-2 Rock-1 Rock-2 Bow-2 Area Figure 2 Mean elemental concentrations (± standard deviation) of school mackerel otoliths sampled from the different areas and F-values determined from ANOVA showing significant elements among samples (df=3, 80; N.S. nonsignificant result). Minimum elemental detection limits of the inductively coupled plasma atomic emission spectroscopy (ICP-AES) (all units are in mg/Kg, except Ca which is measured as a %): Ba = 0.78; Ca = 544; Fe = 23.9; K = 353; Li = 15.0; Mg =10.9; Mn = 1.10; Na = 1250; P = 54.1; S =191; Sr = 3.24. ing a strong effect of age at collection in measure- ment of the elemental composition of otoliths in this area (HSD, P<0.05). Sodium showed the most sig- nificant differences in these univariate analyses (Fig. 2) and was selected as the common variable for scatterplots in Figure 3. Scatterplots of the relation- ships between Na and the other significant elemen- tal concentrations revealed similar spatial patterns 658 Fishery Bulletin 96(4), 1 998 160 140 a. 120 100 80 o o® ° o o . o % ei 7.75 7.85 7.95 8.05 8.15 8.25 Na 3.6 3.2 2.8 2.4-1 2.0 ■ ■■ ■ ■*fV" o a 7.75 7.85 7.95 8.05 8.15 8.25 Na 3.2 590 2.5- ■ O." ■ 480 4 O O 0 8 o 1.8 cm" °°o . * 3701 D °°°o° * 260 O o A 00:°* °mV o ° ° i £ * 1.1 8*°„f^ ■ □ A ■ A A I# ■ 0.4 t , ° ° ■ ‘ , , 150- A 7.75 7.85 7.95 8.05 8.15 8.25 Na 7.75 7.85 7.95 8.05 8.15 8.25 Na Figure 3 Scatterplots of significant elemental ratios showing grouping patterns of area and age school mack- erel samples. Open circle = Bowen 2-year-old fish; open square = Moreton Bay 2-year-old fish; solid square = Rockhampton 1-year-old fish; and solid triangle = Rockhampton 2-year-old fish. Mg and Na are logt,-transformed concentrations. in the otolith chemical composition among samples from the different areas (Fig. 3). The ratio of Na to P concentrations provided the best discrimination among areas, with Bowen samples distinguishable from Rockhampton and Moreton Bay, and samples from Rockhampton and Moreton Bay once again over- lapping in their elemental composition (Fig. 3). Principal component analysis provided further support for the strength of these proposed grouping patterns (Fig. 4). Bowen samples were distinguished from those from the other areas mainly on the first principal component (PC I) where most of the varia- tion was explained by differences in P concentration, whereas Na and Mg contributed most to the varia- tion in the second principal component (PC II). These two components described almost 57% of the total variation in the data. Rockhampton and Moreton Bay samples showed a similar distribution of PC scores, in contrast to samples from Rockhampton that showed some degree of separation between 1- and 2- year-old fish. AN OVA supported these patterns with significant differences detected for both the PC I (F=24.89, df=3, 80, PcO.0001) and PC II scores (F=6.30, df=3, 80, P<0.0007) among the samples. than those from Rockhampton and Moreton Bay, and Bowen samples had significantly higher PC I scores no differences were found between 2-year-old fish -0.5 0.5 PC I score Figure 4 Principal component (PC) analysis of school mackerel otolith elemental concentrations indicating grouping patterns. Open circle = Bowen 2-year-old fish; open square = Moreton Bay 2-year-old fish; solid square = Rockhampton 1-year-old fish; and solid triangle = Rockhampton 2-year-old fish. Begg et at: Stock discrimination of Scomberomorus queenslandicus and 5. munroi 659 Canonical variate I Bowen-2 Moreton Bay-2 Rockhampton-1 Rockhampton-2 Figure 5 Discrimination between school mackerel samples based on the concentrations of 7 trace elements (± 95% confidence ellipses around the sample mean for each area). from Rockhampton and Moreton Bay. The samples of 1-year-old fish from Rockhampton had PC I scores that were significantly different from all the 2-year-old samples, and PC II scores that also were significantly dif- ferent from the other samples, but not from Moreton Bay (HSD, P<0.05). Analyses indicated that school mackerel samples are best separated into three groups: 1) Bowen 2-year- old fish; 2) Moreton Bay and Rock- hampton 2-year-old fish; and 3) Rockhampton 1-year-old fish (Wilks’s lambda=0.195) (Fig. 5). Similar dis- criminant patterns were observed when the “significant” elements (K, Mg, Mn, Na, P; Wilks’s lambda=0.214) were used in isolation. Significant dif- ferences in the discriminant scores were found between these groups for both the first canonical variate (CV I) ( ANOVA, F= 74. 10, df=3, 80, PcO.OOOl) and the second (CV II) (ANOVA, F=6.61, df=3, 80, P<0.0001). Like the results of the PCA analyses, Bowen samples had significantly different CV I scores than those from Rockhampton and Moreton Bay; no differ- ences were found between Rockhampton and Moreton Bay 2-year-old fish; and Rockhampton 1-year-old samples had CV I scores significantly different from all the other groups. The Rockhampton 1-year-old samples also had CV II scores that were significantly different from the other groups, with the exception of fish from Bowen (HSD, P< 0.05). Variation in Na, P, and Mg concentrations were primarily responsible for the separations apparent among the samples (89%) according to the first ca- nonical variate (Table 2). Bowen fish tended to have lower Na and Mg and higher P concentrations than Moreton Bay and Rockhampton samples (Fig. 2). Approximately 62% of the school mackerel samples were classified into their correct groupings, indicating an overlap in otolith composition of Rockhampton and Moreton Bay samples (Table 3). Overall classification success increased to 77% when these areas were pooled, and individual classification rates improved for Bowen 2-yr-old fish (86%), Moreton Bay and Rockhampton 2-year-old fish (70%), and Rockhampton 1-year-old fish (80%), providing further support for the notion of three distinct groups in elemental composition of otoliths. Spotted mackerel The same suite of 11 elements that were used for school mackerel were measured and analyzed for Table 2 Discrimination between samples of school mackerel deter- mined by the pooled within-group correlations of elemen- tal concentrations with the significant (P< 0.05) canonical variates I and II, and the cumulative proportion of the ex- plained variance accounted for by each function for seven elements (Wilks’s lambda=0.195). Canonical variate Element I II Na 0.53 0.07 P -0.48 0.02 Mn 0.17 -0.72 K -0.20 -0.40 Mg 0.42 -0.38 S -0.13 0.24 Sr -0.13 -0.11 Cumulative proportion 0.89 0.97 spotted mackerel (Fig. 6). Iron, Li, and Ca were ex- cluded from statistical analyses for the same reasons as they were for school mackerel. Concentrations of Na (ANCOVA, F=14.02, df=l, 141, P<0.0003; r= -0.0004597) and Sr (ANCOVA, F= 6.03, df=l, 141, P< 0.02; r=-0.9227) in spotted mackerel otoliths were highly correlated with fish length; therefore the data were corrected for length with the respective regres- sion coefficient for the length covariate. Manganese was not used in statistical analyses because a sig- nificant interaction in concentrations existed be- 660 Fishery Bulletin 96(4), 1998 Ba (F=46.85, P0.0001) Ca Fe * o> E More-1Herv-1Herv-3Bow-3 Inn-3 Area K (F=4.58, P<0.002) 43 40 37 12 More-1Herv-1Herv-3Bow-3 Inn-3 Area Li More-1 Herv-1 Herv-3 Bow-3 Inn-3 Area Mg (F=34.89, P<0 0001) 800 600 c 400 S o O Mn 8 c o O More-1Herv-1 Herv-3 Bow-3 Inn-3 Area 28 22 § 4.4 4.6 4.8 5.0 5.2 5.4 P 3.6 3.3-1 3.0 2.7 2.4 A 8°° * -O . ^ n O D O O _ °* * . ° ^ 4.4 4.6 4.8 5.0 5.2 54 P 8.55 8.45 ™ 8.35 8.25 8.15 W • A o^c a,,5a!« . %e ^ hO oOO ■qf» 4.4 4.6 4.8 5.0 5.2 5.4 P 600 540 c/5 480 420 360 AA * A 8 * * qfc*., □ B * o □ □ □ 4.4 4.6 4.8 5.0 5.2 5.4 P Figure 7 Scatterplots of significant elemental ratios showing grouping patterns of area and age spotted mackerel samples. Open circle = Bowen 3-year-old fish; open square = Hervey Bay 1-year-old fish; solid square = Hervey Bay 3-year-o!d fish; solid triangle = Innisfail 3-year-old fish; and solid circle = Moreton Bay 1-year-old fish. K, Mg, Na, and P are logf transformed concentrations. Table 5 Jack-knifed cross-validation classification matrix of the frequency of assigned cases in each area (and age) used to differentiate spotted mackerel samples. Classification of individual spotted mackerel by area Area Correct (%) Bowen (3) Hervey (1) Hervey (3) Innisfail (3) Moreton ( 1 ) Bowen (3) 58 18 0 9 4 0 Hervey (1) 63 0 17 0 0 10 Hervey (3) 34 9 0 10 10 0 Innisfail (3) 68 2 0 7 19 0 Moreton (1) 63 0 11 0 0 19 Total 57 29 28 26 33 29 Begg et al. : Stock discrimination of Scomberomorus queenstandicus and S. munroi 663 Discussion Limitations of elemental analysis of whole otoliths We found evidence of two separate groups of school mackerel and one group of spotted mackerel in the study region based on the variation in the mean el- emental composition of individual fish otoliths of the same age. Such variation has been commonly pre- sumed to reflect prolonged separation of the popula- tions and ultimately stock divergence (Edmonds et al., 1991). However, it is not possible to infer from such studies alone what environmental, dietary, or genetic factors cause spatial patterns. Consequently, a number of wholly different but equally plausible hypotheses could be considered in determining the cause of the patterns observed in the elemental com- positions of the otoliths of the two mackerel species, depending on what factors are considered important in determining otolith chemistry. An obvious explanation is that mean otolith com- position differs spatially for school mackerel because their environment (or diet) varies among locations along the east coast, whereas there are no such differences for spotted mackerel because their en- vironment (or diet) is more uniform throughout their range. For example, Proctor et al. ( 1995) concluded that an inability to discriminate among south- ern bluefin tuna ( Thunnus maccoyii ) samples with trace-element analysis was due partly to the uniformity of the pelagic environment. Water chemistry and food sources along the east coast of Queensland might be expected to vary among coastal bays, between inshore and offshore waters and between deep and shallow parts of the water column, especially given the strong influence of freshwater input in the wet-dry tropics (Thorrold and McKinnon, 1995). Indeed, the clupeid and engraulid species that form major parts of the diet of both school and spotted mackerel are gener- ally found only in the southern half of the study area (Begg and Hopper, 1997). At present there is no fishery-inde- pendent information on the cross-shelf distribution and habitat of school and spotted mackerel and only a coarse understanding of the feeding patterns of both species. School mackerel are thought to inhabit mainly inshore waters, whereas spotted mackerel are thought to be more common offshore (Munro, 1943; 4-, O Q. 0- -2- • • 0 / o * On O ® ■ 0 U I I I I I I -2.0 -1.5 -1.0 -0.5 0 0.5 1.0 1.5 PC I score Figure 8 Principal component (PC) analysis of spotted mackerel otolith elemental concentrations indicating grouping pat- terns. Open circle = Bowen 3-year-old fish; open square = Hervey Bay 1-year-old fish; solid square = Hervey Bay 3- year-old fish; solid triangle = Innisfail 3-year- old fish; and solid circle = Moreton Bay 1-year-old fish. Bowen-3 Hervey Bay-1 Hervey Bay-3 lnnisfail-3 Moreton Bay-1 Canonical variate I Figure 9 Discrimination between spotted mackerel samples based on the concentra- tions of the seven (length corrected) trace elements (+95% confidence el- lipses around the sample mean for each area). Collette and Russo, 1984), although they do move to inshore waters to feed when undertaking seasonal migrations along the coastline (Begg and Hopper, 1997; Begg et al., 1997). School and spotted mack- erel are often caught together in the same mixed schools during winter in the shallow northern bays 664 Fishery Bulletin 96(4), 1998 of the Queensland east coast and in summer in the southern bays, but the location and habitats of juveniles is not well known, particularly for spotted mackerel. Complementary biological information from other stock identification methods is therefore essential to help interpret results from whole otolith analyses, but some hypotheses regarding stock structure can- not be tested in the absence of knowledge about the factors governing otolith composition. The resolution of the technique is limited further by volumetric con- siderations and the nature of otolith growth. Even if there were very large consistent differences in com- position among individuals at the otolith core dur- ing larval and postlarval life, these would be virtu- ally undetectable with bulk analysis, whereas rela- tively small differences accumulated during later life before capture would have a disproportionately large effect on mean composition. To overcome these prob- lems— which may be accentuated for pelagic species with complex ontogenetic movements (Proctor et al., 1995) — bulk analysis of whole otoliths from juveniles, or excision and analysis of the core regions of otoliths from adult fish (Dove et al., 1996), may provide other, solution-based alternatives to sectioning and elec- tron- or laser-probe techniques. There may also be temporal variation in stock dis- crimination patterns, particularly for pelagic species living in coastal areas influenced by boundary cur- rents; Edmonds et al. (1995) demonstrated that varia- tion in composition of pilchard (Sardinops sagax ) otoliths among years was significantly greater than the variation among sites. Stock discrimination pat- terns of pilchards were not persistent from one sam- pling era to the next within a decade. Further bias may be introduced by uneven representation of all life his- tory stages within collections (Edmonds et al., 1995). Age-reSated variation Spatial variation in the concentration of trace ele- ments of school and spotted mackerel was strongly influenced by the age of fish in the samples at time of collection. Differential otolith elemental patterns were found between 1- and 2-year-old school mack- erel, and 1- and 3-year-old spotted mackerel through- out the study region. Numerous studies, including this one, support Kalish’s (1989) hypothesis that in- corporation of trace elements into otoliths is related to growth (Grady et al. , 1989; Thresher et al., 1994; Edmonds et al., 1995; Fowler et al., 1995). Conse- quently, spatial variation in elemental composition of otoliths can be difficult to interpret because of bi- ases related to the size and age of fish in the samples from separate areas. School and spotted mackerel of different ages from the same area have significantly different patterns in the elemental composition of their otoliths. One- year-old school and spotted mackerel tended to have lower concentrations of P, S, and Sr, while having higher levels of Mg, Mn, and Na in their otoliths com- pared with older fish. Grady et al. (1989) found simi- lar results for king mackerel where heavy metal con- centrations were generally higher in otoliths of younger fish. Differences in the concentrations of trace elements accumulated in the otoliths of fish of different ages from the same stock are not unex- pected, particularly if irreversible deposition of ele- ments in otoliths is assumed. Distinct chemical pat- terns in otoliths of fish of different ages may reflect exposure to similar environments, but for different accumulation periods, or alternatively, may reflect life history differences, such as younger fish inhab- iting distinct nursery grounds that are separate from adult habitats. Not all factors affecting deposition of elements in otoliths are strictly environmental or ontogenetic, nor do they necessarily act in a simplistic manner that reflects ambient environmental chemistry (Kalish, 1989; Campana et al., 1994). Chemical deposition can be regulated by many interacting factors, including water temperature, salinity, age, physiology, growth rates, and activity levels of individual fish (Kalish, 1989, 1990, 1991; Radtke and Shafer, 1992; Rieman et al., 1994; Fowler et al., 1995). Stock structure and management Although further research is required to determine the mechanisms responsible for elemental deposition in otoliths of Scomberomorus species, particularly the interaction between environmental and genomic con- trols, the validity of using stock and site-specific el- emental “fingerprints” does not rest upon the mecha- nism underlying otolith formation (Campana and Gagne, 1995). Optimal groupings of mean elemental composition of school mackerel otoliths strongly sup- ported the hypothesis of at least two stocks in the study region — a hypothesis that has been developed with complementary stock identification techniques. Localized movements of tagged school mackerel have shown that there is little exchange between adult fish from different areas throughout Queensland east coast waters; most recaptures occur within the same area of release (Begg et al., 1997). Studies of the tim- ing and location of spawning have also shown that school mackerel spawn concurrently at a number of localities along the east coast (Begg, in press). The species also exhibits differences in growth patterns and genetic variation throughout its distribution on Begg et al.: Stock discrimination of Scomberomorus queenslandicus and S, munroi 665 the east coast (Begg and Sellin, 1998; Begg et al., in press). The limited movements of school mackerel indi- cated by tag-recapture data may also explain the overlap observed in the otolith composition of 2-year- old fish collected in Moreton Bay and Rockhampton. Small numbers of tagged school mackerel released in both these locations were recaptured in Hervey Bay between August and January, where they ap- pear to be mixing on a common feeding ground (Begg et al. , 1997; Begg and Hopper, 1997). In contrast, spotted mackerel of the same year class sampled in Queensland east coast waters had simi- lar patterns of elemental composition in their otoliths regardless of the region in which they were collected. This finding strongly supports the hypothesis of a single intermixing stock in the region of sampling derived from previous tagging, genetic, and repro- ductive studies, and seasonal changes in the loca- tion of commercial harvesting. These sources of in- formation suggest an annual large-scale movement of spotted mackerel along the Queensland east coast to southern feeding grounds in summer and a return migration in winter to northern spawning grounds (Begg and Hopper, 1997; Begg et al., 1997; Begg, in press). In addition, similar growth rates and homo- geneous genetic conditions of spotted mackerel throughout the study region support the hypothesis of a single east coast stock (Begg and Sellin, 1998; Begg et al., in press). The longshore East Australian Current possibly provides cues for the migratory cycle of spotted mack- erel and may facilitate larval dispersal from the spawning grounds and ultimately stock homogeneity throughout Queensland east coast waters. Proctor et al. (1995) proposed a similar stock structure for south- ern bluefin tuna on the basis of chemical composition of otoliths from juveniles collected along the major mi- gration route of the species. Identification of a species’ stock structure is an important requirement for effective fisheries man- agement (Rounsefell, 1975). The inability to define stock boundaries could unknowingly prejudice oth- erwise well-designed management protection efforts (Kutkuhn, 1981). Management of spotted mackerel in Queensland would be best addressed at the state level, because fishing effort in areas remote from one another may have an interaction on a single stock. In contrast, more localized (regional-level) manage- ment actions could proceed for school mackerel within Queensland — especially if the stock boundaries can be refined by further tests of the temporal persis- tence of the patterns described in this study. The use of otolith trace element analysis holds great potential for stock discrimination in fisheries for other Scom beromorus species that share common characteristics of migration through multiple juris- dictions and fishery types. However, the use of bulk analysis of whole otoliths allows us to infer only that there has been prolonged separation of fish at some stage in their life history. Without “life history scans” across otoliths, with techniques such as electron- probe microanalysis (Gunn et al., 1992), we cannot explore the possibility that mean elemental compo- sition is dominated by high concentrations specific to any particular life history stage. There is also need for careful selection of samples among year classes and time periods. Replication of sampling at inter- vals separated by several years and representation of all ontogenetic stages need to be incorporated in analyses to test for the consistency of spatial patterns in order to provide a more accurate environmental “fin- gerprint” for discrimination of mackerel stocks. Acknowledgments We would like to thank the commercial and recre- ational fishermen for cooperation and assistance in collection of samples; Steven Campana and four anonymous reviewers for critical comments and sug- gestions; Steve Cadrin for statistical advice; and Ian Brown and Hamish McCallum for reviews of this work. This study formed part of the Queensland De- partment of Primary Industries Fisheries Research and Development Corporation grant FRDC 92/144. Literature cited Begg, G. A. In press. Reproductive biology of school mackerel ( Scomberomorus queenslandicus) and spotted mackerel (S. munroi) in Queensland east-coast waters. Mar. Freshwa- ter Res. 49. Begg, G. A., D. S. Cameron, and W. Sawynok. 1997. Movements and stock structure of school mackerel ( Scomberomorus queenslandicus) and spotted mackerel (S. munroi) in Australian east-coast waters. Mar. Freshwa- ter Res. 48:295-301. Begg, G. A., and G. A. Hopper. 1997. Feeding patterns of school mackerel ( Scomberomorus queenslandicus ) and spotted mackerel ( S . munroi) in Queensland east-coast waters. Mar. Freshwater Res. 48: 565-571. Begg, G. A., C. P. Keenan, and M. J. Sellin. In press. Genetic variation and stock structure of school mackerel and spotted mackerel in northern Australian waters. J. Fish Biol. 52. Begg, G. A., and M. J. Sellin. 1998. Age and growth of school mackerel ( Scomberomorus queenslandicus) and spotted mackerel (S. munroi) in Queensland east-coast waters with implications for stock structure. Mar. Freshwater Res. 49:109-120. 666 Fishery Bulletin 96(4), 1 998 Campana, S. E., A. J. Fowler, and C. M. Jones. 1994. Otolith elemental fingerprinting for stock identifica- tion of Atlantic cod ( Gadus morhua) using laser ablation ICPMS. Can. J. Fish. Aquat. Sci. 51:1942-1950. Campana, S. E,, and J. A. Gagne. 1995. Cod stock discrimination using ICPMS elemental as- says of otoliths. In D. H. Secor, J. M. Dean, and S. E. Campana ( eds. ), Recent developments in fish otolith research, p. 671-691. Univ. South Carolina Press, Columbia, SC. Campana, S. E., J. A. Gagne, and J. W. McLaren. 1995. Elemental fingerprinting of fish otoliths using ID- ICPMS. Mar. Ecol. Prog. Ser. 122: 15-120. Collette, B. B., and J. L. Russo. 1984. Morphology, systematics, and biology of the Spanish mackerels (Scomberomorus, Scombridae). Fish. Bull. 82:545-689. Dove, S. G., B. M. Gillanders, and M. J. Kingsford. 1996. An investigation of chronological differences in the deposition of trace metals in the otoliths of two temperate reef fishes. J. Exp. Mar. Biol. Ecol. 205:15-34. Edmonds, J. S., N. Caputi, and M. Morita. 1991. Stock discrimination by trace-element analysis of otoliths of orange roughy ( Hoplostethus atlanticus ), a deep- water marine teleost. Aust. J. Mar. Freshwater Res. 42:383-389. Edmonds, J. S., N. Caputi, M. J. Moran, W. J. Fletcher, and M. Morita. 1995. Population discrimination by variation in concentra- tions of minor and trace elements in sagittae of two Western Australian teleosts. In D. H. Secor, J. M. Dean, and S. E. Campana ( eds. ), Recent developments in fish otolith research, p. 655-670. Univ. South Carolina Press, Columbia, SC. Edmonds, J. S., M. J. Moran, N. Caputi, and M. Morita. 1989. Trace element analysis of fish sagittae as an aid to stock identification: pink snapper ( Chrysophrys auratus) in Western Australian waters. Can. J. Fish. Aquat. Sci. 46:50-54. Fowler, A. J., S. E. Campana, C. M. Jones, and S. R. Thorrold. 1995. Experimental assessment of the effect of tempera- ture and salinity on elemental composition of otoliths us- ing laser ablation ICPMS. Can. J. Fish. Aquat. Sci. 52:1431-1441. Grady, J. R., A. G. Johnson, and M. Sanders. 1989. Heavy metal content in otoliths of king mackerel ( Scomberomorus cavalla) in relation to body length and age. Contrib. Mar. Sci. 31:17-23. Gunn, J. S., I. R. Harrowfield, C. H. Proctor, and R. E. Thresher. 1992. Electron probe microanalysis of fish otoliths — evalu- ation of techniques for studying age and stock discrimi- nation J. Exp. Mar. Biol. Ecol 158:1-36. Ihssen, P. E., H. E. Booke, J. M. Casselman, J. M. McGlade, N. R. Payne, and F. M. Utter. 1981. Stock identification: materials and methods. Can. J. Fish. Aquat. Sci. 38:1838-1855. Johnson, A. G., W. A. Fable Jr., C. B. Grimes, and L. Trent. 1994. Evidence for distinct stocks of king mackerel, Scomberomorus cavalla , in the Gulf of Mexico. Fish. Bull. 92:91-101. Kalish, J. M. 1989. Otolith microchemistry: validation of the effects of physiology, age and environment on otolith composi- tion. J. Exp. Mar. Biol. Ecol. 132:151-178. 1990. Use of otolith microchemistry to distinguish the prog- eny of sympatric anadromous and non-anadromous salmonids. Fish. Bull. 88:657-666. 1991. Determinants of otolith chemistry: seasonal varia- tion in the composition of blood plasma, endolymph and otoliths of bearded rock cod Pseudophycis barbatus. Mar. Ecol. Prog. Ser. 74:137-159. Kalish, J. M., M. E. Livingston, and K. A. Schofield. 1996. Trace elements in the otoliths of New Zealand blue grenadier (Macruronus novaezelandiae ) as an aid to stock discrimination. Mar. Freshwater Res. 47:537-542. Kutkuhn, J. H. 1981. Stock definition as a necessary basis for cooperative management of Great Lakes fish resources. Can. J. Fish. Aquat. Sci. 38:1476-1478. Mulligan, T. J., F. D. Martin, R. A. Smucker, and D. A. Wright. 1987. A method of stock identification based on the elemen- tal composition of striped bass Morone saxatilis ( Walbaum) otoliths. J. Exp. Mar. Biol. Ecol. 114:241-248. Munro, I. S. R. 1943. Revision of Australian species of Scomberomorus . Mem. Qld. Mus. 12:65-95. Proctor, C. H., R. E. Thresher, J. S. Gunn, D. J. Mills, I. R. Harrowfield, and S. H. Sie. 1995. Stock structure of the southern bluefin tuna Thunnus maccoyii: an investigation based on probe microanalysis of otolith composition. Mar. Biol. 122:511-526. Radtke, R. L., and D. J. Shafer. 1992. Environmental sensitivity of fish otolith micro- chemistry. Aust. J. Mar. Freshwater Res. 43:935-951. Rieman, B. E., D. L. Myers, and R. L. Neilsen. 1994. Use of otolith microchemistry to discriminate Oncorhyncus nerka of resident and anadromous origin. Can. J. Fish. Aquat. Sci. 51:68-77. Rounsefell, G. A. 1975. Ecology, utilization, and management of marine fisheries. Mosby, St. Louis, MO, 516 p. Shaklee, J. B., S. R. Phelps, and J. Salini. 1990. Analysis of fish stock structure and mixed-stock fish- eries by the electrophoretic characterization of allelic isozymes. In D. H. Whitmore (ed.), Electrophoretic and isoelectric focusing techniques in fisheries management, p. 173-196. CRC Press, Boca Raton. FL. Sutter, F. C., Ill, R. O. Williams, and M. F. Godcharles. 1991. Movement patterns and stock affinities of king mack- erel in the southeastern United States. Fish. Bull. 89:315-324. SYSTAT. 1997. SYSTAT 7.0 statistics. SPSS Incorporated, Chicago, IL, 751 p. Thorrold, S. R., and A. D. McKinnon. 1995. Response of larval fish assemblages to a riverine plume in coastal waters of the central Great Barrier Reef lagoon. Limnol. Oceanogr. 40:177-181. Thresher, R. E., C. H. Proctor, J. S. Gunn, and I. R. Harrowfield. 1994. An evaluation of electron-probe microanalysis of otoliths for stock delineation and identification of nursery areas in a southern temperate groundfish, Nemadactylus macropterus (Cheilodactylidae). Fish. Bull. 92:817-840. Trent, L., B. J. Palko, M. L. Williams, and H. A. Brusher. 1987. Abundance of king mackerel, Scomberomorus cavalla , in the southeastern United States based on CPUE data from charterboats 1982-85. Mar. Fish. Rev. 49:78-90. 667 Pelagic sharks associated with the swordfish, Xiphias gladius, fishery in the eastern North Atlantic Ocean and the Strait of Gibraltar Valentin Buencuerpo Santiago Rios Julio Moron Departamento de Biologi'a Animal I (Zoologl a), Facultad de Biologla Universidad Complutense de Madrid 28040 Madrid, Spain E-mail address: vbuencar@eucmax.sim.ucm.es Abstract .—We report on 175 land- ings from 106 longline and 69 gillnet boats operating in the eastern North Atlantic Ocean and Mediterranean Sea, July 1991 to July 1992. Information on the catch and biology of five shark spe- cies (Isurus oxyrinchus , Prionace glauca , Alopias superciliosus, Alopias vulpinus, and Sphyrna zygaena ) is ana- lyzed and contrasted with swordfish (. Xiphias gladius) landings. A total of 51,205 fish were sampled, of which 40,198 were sharks, 9,990 swordfish, and the rest other bony fish. Spatial, temporal, and gear analyses were per- formed to show the importance of shark bycatch in longline and gillnet fisher- ies operating from the south of Spain. We present information on population structures of the shark species, along with hypotheses about shortfin mako movements as suggested by landing data. Manuscript accepted 18 February 1998, Fish. Bull. 96:667-685 (1998). The Spanish longline fleet operates in the eastern North Atlantic (FAO fishing area 27), north-central At- lantic (FAO 34), and the Mediter- ranean Sea (FAO 37). Sharks are the most important bycatch of swordfish (Xiphias gladius Lin- naeus, 1758) fishing in all of these areas (Moreno and Moron, 1992b; Mejuto and Garces1; Mejuto2). The fleets use surface longlines to catch swordfish and some sharks have become target species as well. Three types of longliners operate along the south Spanish coast: “coastal longliners” that fish two to five days in the area of the Strait of Gibraltar and that in some cases use gill nets in spring and summer; “offshore longliners” that spend 15 to 25 days fishing as far south as Senegal; and “long distance freezer longliners” that operate mainly in the tropical Atlantic Ocean (Gulf of Guinea and off Brazil). Little is known about shark fishing off southern Spain (Garcia, 1970; Bravo and Santaella, 1973; Amorin et al., 1979; Garces and Rey, 1983; Moreno and Moron, 1992b); more is known about shark catches from the northern Spanish longline fleet (Mejuto and Garces1; Mejuto2). The biology of eastern Atlantic pelagic sharks has not been well investigated although detailed taxo- nomic studies (Moreno, 1982), morph- ology and growth studies ( Blasco and Munoz-Chapuli, 1981; Munoz- Chapuli and Blasco, 1984; Moreno and Moron, 1992b), reproduction (Munoz-Chapuli, 1984; Moreno et ah, 1989; Moreno and Moron, 1992a), and faunistic studies (Belloc, 1934; Lozano Cabo, 1950; Bravo, 1974; Bravo3) are available. The most common shark species taken by longliners are blue shark, Prionace glauca (Linnaeus, 1758) and shortfin mako shark, Isurus oxyrinchus Rafmesque, 1810, which represent 80% of the total shark bycatch (Garces and Rey, 1983; Moreno and Moron, 1992b). Other sharks caught include the por- beagle, Lamna nasus (Bonnaterre, 1788), the hammerheads, Sphyrna spp., the bigeye thresher, Alopias superciliosus (Lowe, 1839), the com- mon thresher, Alopias vulpinus 1 Mejuto, J., and A. Garces. 1984. Shortfin mako, Isurus oxyrinchus, and por- beagle, Lamna nasus , associated with longline swordfish fishery in NW and N Spain. ICES, Council Meeting 1984/G: 72. 2 Mejuto, J. 1985 Associated catches of sharks, Prionace glauca, Isurus oxyrinchus and Lamna nasus, with NW and N Span- ish swordfish fishery, in 1984. ICES, Coun- cil Meeting 1985/H: 42. 3 Bravo, J. 1973. Elasmobranchii off Ca- nary Islands. ICES, Council Meeting 1973/ J: 17. 668 Fishery Bulletin 96(4), 1998 (Bonnaterre, 1788), and the requiem sharks, Carchar- hinus spp. (Garces and Rey, 1984). Bony fish are caught occasionally, including the northern bluefin tuna, Thunnus thynnus (Linnaeus, 1758), bigeye tuna, Thunnus obesus (Lowe, 1839), albacore tuna, Thunnus alalunga (Bonnaterre, 1788), oilfish, Ruvettus pretiosus Cocco, 1829, escolar, Lepido- cybium flavobrunneum (Smith, 1849), and longbill spearfish, Tetrapturus pfluegeri Robins and de Sylva, 1963, (Mejuto and Garces1; Mejuto2). Materials and methods Between July 1991 and July 1992, 106 longline and 69 gillnet landings (85.4% and 87.3% respectively of total longline and gillnet landings) were sampled at the Algeciras fish market (Cadiz, southern Spain), the largest market in southwestern Spain receiving 80% of the longline catch from this area (Anonymous, 1986). Data on the number of fish landed by species, days of each trip, active fishing days, number of sets and hooks for longliners, and net length and fishing time for gillnetters were collected during interviews with skippers. The longline catch rate was calculated as hook rate (HR=number of fish/1,000 hooks), and gillnet catch rate was calculated as the average net length by trip multiplied by the number of sets. No total catch and effort data were available; catch rates are approxi- mate indicators and do not reflect actual abun- dance. Fishing time was not included in effort calculations because it was constant for every boat and gear. The gears were set at dusk and retrieved before sunrise. Longlines ranged from 18 to 29 km in length and on average had 1500 hooks (range: 475 to 2500), set at a depth of 11 to 55 m. Gill nets were 2.5 km long, 14 m high, and had a mesh size of 40 cm. All fish landed were identified to species level and counted. Most sharks were sexed and mea- sured to the nearest centimeter for total length (TL) or fork length (FL), or both. Total length was measured from the rear tip of the upper cau- dal lobe to the snout tip, along the horizontal line of the body axis. Length data were grouped in 5- cm intervals of FL (shortfin mako and blue sharks) and TL (thresher and hammerhead sharks) for length-frequency analysis. Biometrical relation- ships between sexes were performed by simple linear regression analysis when the sample size was sufficiently large (shortfin mako, blue, and bigeye thresher sharks). Standard length (SL) has been used as the reference length for the relation with fork length (FL), total length (TL), and upper caudal lobe (UCL) (dorsal-caudal margin). The relation between FL and TL was also calculated. The area sampled (Fig. 1) was divided into 5° latitude sectors from 20°N to 40°N. Sectors 4 and 5 have the same latitude (35°N-40°N) but different longitudes, 3°W-8°W for sector 5, 8°W-13°W for sector 4. This division separates the gillnet fishing ground (sector 5) from the longline Atlantic area (sector 4). A chi-square (%2) test was performed to test the fit of the sample to a normal distribution and to observe the variations in species by sector and month. Strong bias was expected because of the effect of discarded fish at sea. Single factor AN OVA was performed to compare the means of the samples by area. Results A total of 40,198 sharks and 11,007 bony fish were sampled from 175 landings from July 1991 to July 1992 (Table 1). The proportion of sharks landed was higher with longlines than with gill nets, but this might be related to the seasonal abundance of sharks and swordfish in the Strait of Gibraltar during the period when gill nets were used (Fig. 2). The most common species landed by longlines in the areas studied were blue shark, shortfin mako shark, and swordfish (Table 2). Other chondrich- thyans were less common, some rarely occurring in more than 10% of the landings. Buencuerpo et al.: Pelagic sharks associated with the swordfish fishery 669 Proportions of swordfish and the five most com- mon shark species landed by longlines were signifi- cantly different (x2=5.41; df=20; P<0.001) in each sector (Table 3). The highest proportion of fish in all the sectors, except one (sector 1), was that of blue sharks, whereas the lowest proportion in all five sec- tors was that of the common thresher shark. Sword- fish occurred with decreasing proportion from south to north in the Atlantic sectors, reaching a minimum in the Mediterranean sector. Shortfin mako sharks steadily decreased from sector 1 to sector 4. The pro- portion of sharks, overall, was greater than the pro- portion of swordfish in all sectors, and sharks were included in at least 80% of the landings in all sec- Table 1 Total number of fish sampled from 175 landings at the Algeciras (Cadiz) fish market and percentage by species from July 1991 to July 1992. Species Number Percentage Isurus oxyrinchus 5947 11.6 Prionace glauca 32,661 63.7 Alopias vulpinus 52 0.1 Alopias superciliosus 557 1.1 Sphyrna zygaena 757 1.4 Other sharks 224 0.4 Total sharks 40,198 78.5 Xiphias gladius 9990 19.5 Other bony fish 1017 2.0 Total 51,205 100.0 tors, where they were slightly greater in numbers than swordfish, except in sector 1. The proportional distribution of fish from gillnet landings was significantly different from that of fish from longline landings in the same sector 0.05) and gillnet (%2=20.5; df=23; P> 0.5) landings. The dis- tribution by size of males varied significantly by sec- tor (F= 28.796; df=5, 1428; P<0.001). The distribu- tion by size of females also varied significantly by sector (F=14.985; df=5, 1429; PcO.OOl). The maximum size of females landed by longlines tended to increase from sector 1 to sector 3 (Fig. 5A). The maximum size of males tended to increase in Atlantic sectors from south to north (Fig. 5B). Small- est maximum sizes for both sexes occurred in the Mediterranean Sea, and in sector 1 for males. Gill nets catch bigger females and smaller males than do longlines in the same sector. The monthly variations Table 4 Monthly longline and gillnet catch distribution for the five most important shark species and for swordfish by number of fish in) and percentage (%). Species codes: 10 = Isurus oxyrinchus\ PG = Prionace glauca\ AV = Alopias vulpi nus\ AS = Alopias superciliosus; SZ = Sphyrna zygaena; TOT-SHX = Total sharks; XG = Xiph ias gladius; IO PG AV AS SZ TOT-SHX XG Total Gear and n % n % n % n % n % n % n % n month Longline Jul 91 224 16.85 992 74.64 2 0.15 31 2.33 21 1.58 1270 95.56 59 4.44 1329 Aug 91 891 19.96 2865 64.19 1 0.02 3 0.07 147 3.29 3907 87.54 556 12.46 4463 Sep 91 422 15.93 2158 81.46 0 0 7 0.26 28 1.06 2615 98.72 34 1.28 2649 Oct 91 584 9.57 3338 54.68 0 0 32 0.52 131 2.15 4085 66.91 2020 33.09 6105 Nov 91 970 14.10 4333 62.99 4 0.06 25 0.36 252 3.66 5584 81.17 1295 18.83 6879 Dec 91 30 12.35 105 43.21 0 0 5 2.06 0 0 140 57.61 103 42.39 243 Jan 92 863 10.02 6117 71.03 4 0.05 110 1.28 30 0.35 7124 82.72 1488 17.28 8612 Feb 92 868 9.07 6988 73.03 2 0.02 135 1.41 79 0.83 8072 84.36 1497 15.64 9569 Mar 92 423 9.04 3208 68.53 1 0.02 54 1.15 1 0.02 3687 78.77 994 21.23 4681 Apr 92 109 10.64 913 89.16 0 0 1 0.10 0 0 1023 99.90 1 0.10 1024 May 92 25 8.01 283 90.71 1 0.32 0 0 0 0 309 99.04 3 0.96 312 Jun 92 114 10.45 783 71.77 1 0.09 8 0.73 2 0.18 908 83.23 183 16.77 1091 Jul 92 70 15.28 366 79.91 0 0 0 0 12 2.62 448 97.82 10 2.18 458 Total 5593 11.80 32449 68.44 16 0.03 411 0.87 703 1.48 39172 82.62 8243 17.38 47415 Gear and IO PG AV AS SZ TOT-SHX XG Total month n % n % n % n % n % n % n % n Gillnet Jul 91 Aug 91 19 7.39 0 0 i 0.39 37 14.40 5 1.95 62 24.12 195 75.88 257 Sep 91 106 22.70 30 6.42 2 0.43 27 5.78 22 4.71 187 40.04 280 59.96 467 Oct 91 Nov 91 53 9.62 2 0.36 0 0 3 0.54 12 2.18 70 12.70 481 87.30 551 Dec 91 Jan 92 Feb 92 Mar 92 6 6.25 0 0 15 15.6 0 0 1 1.04 22 22.92 74 77.08 96 0 0 1 33.33 0 0 0 0 0 0 1 33.33 2 66.67 3 Apr 92 54 24.00 83 36.89 15 6.67 7 3.11 2 0.89 161 71.56 64 28.44 225 May 92 76 11.66 90 13.80 1 0.15 0 0 0 0 167 25.61 485 74.39 652 Jun 92 14 15.22 1 1.09 0 0 6 6.52 4 4.35 25 27.17 67 72.83 92 Jul 92 26 12.62 5 2.43 2 0.97 66 32.04 8 3.88 107 51.94 99 48.06 206 Total 354 13.89 212 8.32 36 1.41 146 5.73 54 2.12 802 ^1.46 1747 68.54 2549 672 Fishery Bulletin 96(4), 1998 Table 5 Longline effort (number of hooks x 1000) and catch rates (number of fish/1000 hooks) by species and sector. Sector Effort I. oxyrinchus P. glauca A. vulpinus A. superciliosus S. zygaena Total sharks X. gladius 1 302.4 3.03 6.17 0.013 0.09 0.62 9.92 7.97 2 534.0 3.09 14.53 0.002 0.50 0.07 18.19 4.54 3 462.9 4.78 30.69 0.015 0.22 0.85 36.55 3.94 4 61.8 3.90 33.69 0.032 0.15 0.03 37.80 3.22 5 297.5 1.94 21.98 0.007 0.02 0.28 24.23 4.66 Total 1658.6 3.37 19.56 0.010 0.25 0.42 23.62 4.97 Table 6 Longline effort (number of hooks x 1000) and catch rates (number of fish/1000 hooks) by species and month. Month Effort 1. oxyrinchus P. glauca A. vulpinus A. superciliosus S. zygaena Total sharks X. gladius Jul 91 67.40 3.32 14.72 0.03 0.46 0.31 18.84 0.88 Aug 91 180.40 4.94 15.88 0.01 0.02 0.81 21.66 3.08 Sep 91 41.30 10.22 52.25 0.00 0.17 0.68 63.32 0.82 Oct 91 195.00 2.99 17.12 0.00 0.16 0.67 20.95 10.36 Nov 91 185.20 5.24 23.40 0.02 0.13 1.36 30.15 6.99 Dec 91 16.80 1.79 6.25 0.00 0.30 0.00 8.33 6.13 Jan 92 296.60 2.91 20.62 0.01 0.37 0.10 24.02 5.02 Feb 92 393.20 2.21 17.77 0.01 0.34 0.20 20.53 3.81 Mar 92 206.90 2.04 15.51 0.00 0.26 0.00 17.82 4.80 Apr 92 5.00 21.80 182.60 0.00 0.20 0.00 204.60 0.20 May 92 12.80 1.95 22.11 0.08 0.00 0.00 24.14 0.23 Jun 92 47.00 2.43 16.66 0.02 0.17 0.04 19.32 3.89 Jul 92 11.00 6.36 33.27 0.00 0.00 1.09 40.73 0.91 Total 1,658.60 3.37 19.56 0.01 0.25 0.42 23.62 4.97 Table 7 Gillnet effort (average net length (number of sets) and catch rates (number of fish/unit of effort) by species and month. Sector Effort 1. oxyrinchus P. glauca A. vulpinus A. superciliosus S. zygaena Total sharks X. gladius Jul 91 Aug 91 118.50 0.16 0.00 0.01 0.31 0.04 0.52 1.65 Sep 91 152.90 0.69 0.20 0.01 0.18 0.14 1.22 1.83 Oct 91 90.90 0.58 0.02 0.00 0.03 0.13 0.77 5.29 Nov 91 — — — — — — — — Dec 91 — — — — — — — — Jan 92 36.50 0.16 0.00 0.41 0.00 0.03 0.60 2.03 Feb 92 — — — — — — — — Mar 92 2.00 0.00 0.50 0.00 0.00 0.00 0.50 1.00 Apr 92 48.00 1.13 1.73 0.31 0.15 0.04 3.35 1.33 May 92 122.50 0.62 0.73 0.01 0.00 0.00 1.36 3.96 Jun 92 30.00 0.47 0.03 0.00 0.20 0.13 0.83 2.23 Jul 92 56.00 0.46 0.09 0.04 1.18 0.14 1.91 1.77 Total 657.30 0.54 0.32 0.05 0.22 0.08 1.22 2.66 Buencuerpo et at: Pelagic sharks associated with the swordfish fishery 673 of maximum, median, and minimum sizes are pre- for this species were very low in relation to other sented in Figure 6, A and B. species, in some months almost negligible (Table 7). In the longline fishery, sex ratio by sector was close Blue shark to 1 male:0.25 females, except in sector 3 where it was almost equal (1 male:0.92 females); gillnet sex Blue sharks are often discarded at sea because they ratio (1 male:2.45 females) was inverse to that ob- are not commercially valuable in the Spanish mar- served for longlines in the same sector. Monthly data ket; real catch rates are biased by this factor, espe- were not sufficient to be analyzed and therefore were daily in the more distant sectors where the tendency grouped by quarter, i.e. July-September 1991, Octo- is to try to improve long trips with more valuable ber-December 1991, January-March 1992, and species, i.e. with swordfish and shortfin mako shark. April-June 1992. Because of the effect of discarded Therefore, sectors 1 and 2 had lower catch rates, fish at sea and inappropriate sex sampling, we were based on landings, than sectors 3, 4, and 5 (Table 5). unable to make significant assertations about sex Longline catch rates by month (Table 6) ranged from ratio variability. 6 to 52 fish/1000 hooks, except for an extremely high The distribution of overall length frequency is pre- rate in April (182 fish/1 000 hooks). Gillnet catch rates sented in Figure 7A. In longline landings, modal Table 8 Biometric relations of standard length ( SL ) with fork length ( FL ), total length (TL), and upper caudal lobe length (UCL); and of fork length with total length by sex for Isurus oxyrinchus, Prionace glauca, and Alopias superciliosus, and both sexes combined for Alopias vulpinus and Sphyrna zygaena. Females Males Total Isurus oxyrinchus FL = 1.086 SL + 1.630 FL = 1.086 SL + 1.409 FL = 1.086 SL + 1.515 r2 = 0.993 n = 852 r2 = 0.993 77 = 911 t-2 = 0.993 77 = 1,763 TL = 0.817 SL + 0.400 TL = 1.209 SL + 0.435 TL = 1.207 SL + 0.971 r2 = 0.986 n = 852 r2 = 0.983 77 = 681 t-2 = 0.985 77 = 1,533 UCL = 3.693 SL + 13.094 UCL = 3.795 SL + 10.452 UCL = 3.758 SL + 11.640 r2 = 0.874 n = 507 r2 = 0.898 77 = 477 r2 = 0.903 7? = 1,054 TL = 1.106 FL + 0.052 TL = 1.111 FL - 0.870 TL = 1.108 FL-0.480 r2 = 0.985 n = 853 r2 = 0.984 77 = 911 T-2 = 0.984 77 = 1,746 Prionace glauca FL = 1.076 SL + 1.862 FL = 1.080 SL + 1.552 FL = 1.079 SL + 1.668 r 2 = 0.993 n = 1,043 t-2 = 0.997 77 = 1,276 r 2 = 0.996 77 = 2,319 TL = 1.249 SL + 7.476 TL = 1.272 SL + 4.466 TL = 1.262 SL + 5.746 r2 = 0.986 n = 1,043 r2 = 0.993 77 = 1,272 r 2 = 0.990 77 = 2,315 UCL = 0.219 SL + 4.861 UCL = 0.316 SL + 2.191 UCL = 0.306 SL + 3.288 r2 = 0.903 n = 1,038 r2 = 0.948 77 = 1,264 r2 = 0.929 77 = 2,302 TL = 1.158 FL + 5.678 TL = 1.117 FL + 2.958 TL = 1.167 FL+ 4.133 r2 = 0.988 n = 1,043 t-2 = 0.992 77 = 1,272 r2 = 0.990 77 = 2,315 Alopias superciliosus FL = 1.075 SL + 3.346 FL = 1.081 SL + 3.324 FL = 1.073 SL + 4.150 r2 = 0.987 n = 90 r2 = 0.980 77 = 76 T-2 = 0.986 77 = 166 TL = 1.939 SL - 11.990 TL = 1.897 SL - 7.359 TL = 1.937 SL -12.630 r2 = 0.976 77 = 77 r2 = 0.946 77 = 70 7-2 = 0.970 7? = 147 UCL = 0.967 SL -9.588 UCL = 0.922 SL -5.353 UCL = 0.966 SL - 10.908 r2 = 0.952 77 = 75 T-2 = 0.869 77 = 71 7-2 = 0.932 77 = 146 TL = 1.775 FL - 13.007 TL = 1.722 FL - 7.295 TL = 1.773 FL- 14.456 t-2 = 0.956 77 = 77 r2 = 0.923 77 = 70 r2 = 0.949 77 = 147 Alopias vulpinus Sphyrna zygaena Relation r2 n Relation r2 n FL = 1.118 SL - 2.29 0.989 22 FL = 0.845 SL + 1.077 0.998 56 TL = 1.930 SL + 9.331 0.956 22 TL = 1.322 SL + 8.397 0.994 56 UCL = 0.926 SL + 25.670 0.949 22 UCL = 0.329 SL + 8.799 0.975 55 TL = 1.687 FL + 20.483 0.932 22 TL = 1.225 FL + 7.528 0.994 56 674 Fishery Bulletin 96(4), 1998 length for males (95 cm FL) was slightly greater than for females (80 cm FL), whereas in gillnet landings it was equivalent (100 cm FL). Mean lengths in the longline landings, including nonsexed specimens, ranged from 85 cm FL in sector 5 to 205 cm FL in sec- tor 1. The mean length for gill nets was the same for g% _ Males =1292 Figure 3 Total length-frequency distribution by sex of the shortfm mako shark longline landings from all sectors from July 1991 to July 1992. 10% Males =142 Fork length (cm) Figure 4 Total length-frequency distribution by sex of the shortfin mako shark gillnet landings from sector 5 from July 1991 to July 1992. Buencuerpo et a!.: Pelagic sharks associated with the swordfish fishery 675 males and females (100 cm FL). The incompleteness of these data did not allow any temporal analyses. Size distribution by sector varied significantly ( 3 87.1 3 0 ; df=5, 8214; PcO.OOl). Total mean size was greatest in sector 1, lowest in sector 5. Sector 3 had minimum-size females and males. The largest size females were found in sector 1; the largest males, in sector 4 (Fig. 7, B and C). Table 8 shows biometrical comparisons by sex. Bigeye thresher shark Longline catch rates for bigeye thresher sharks were very low. However, this species was caught in all sec- tors, sector 2 having the highest rate (Table 5). There was a slight increase in the longline catch rate, from 676 Fishery Bulletin 96(4), 1998 350 300 250 200 150 100 50 A 4 Females 306 n= 1564 i < ► < ► < < ► , < ◄ < < < 1 , 1 1 1 1 1 1 » 1 . ' ' 1 ■ 1 e 5 i i i i L A > i JL91 A91 S91 091 N91 D91 JA92 F92 M92 AP92 MY92 J-92 JL92 £ re c ® 300 250 200 150 100 50 ♦ maximum ■ mean A minimum B < Males 53 n = 1576 ► < ► < < ► < ► < 4 ► ► < < ► - i k i i . i L > 1 i i i 1 ■ I ^ 1 1 1 1 1 6f ' 4- » — 1 1 1 t 1 1 — — * ( 1 4 1 1 ♦ maximum ■ minimum A mean JL91 A91 S91 091 N91 D91 JA92 F92 M92 AP92 MY92 J-92 JL92 66 Month Figure 6 (A) Size range variation of the shortfin mako shark females landed by month (December 91 excluded); (B) size range variation of the shortfin mako shark males landed by month (December 91 excluded). Buencuerpo et a I.: Pelagic sharks associated with the swordfish fishery 677 250 275 Fork length (cm) Figure 7 (A) Total length-frequency distribution of blue shark from longline and gillnet landings from all sectors from July 1991 to July 1992. (B) Size range variation of blue shark females landed by longlines and gill net (5GN) by sector from July 1991 to July 1992. (C) Size range variation of blue shark males landed by longlines and gill net (5GN) by sector from July 1991 to July 1992. fall to winter, dropping during spring. A maximun catch rate was reached in July 1991 (Table 6). In gillnet fishing, catch rates were highest during sum- mer months, with a maximum in July 1992 (Table 7). The sex ratio found in longline fishing was fairly balanced, 1 male to 1.1 females in sector 2 to 1 to 0.93 females in sector 3. Females were more abun- dant than males in gillnet fishing (1 male: 1.43 fe- males). The distribution of sex ratio by sector was not significantly different (%2=9.14; df=5; P>0.1). The distribution of sex ratio varied significantly by month (/2=38.7; df=ll; P<0.001). In longline fishing there were more males than females (1 male:0.66-0.69 fe- males), except in February and March when the ra- tio was 1 male: 1.45 females, and in September, when the proportion of females rose dramatically to 1 male:6. 1 females. Gillnet fishing reflected a similar proportion during the same month (1 male: 13.21 females). The frequency distribution by size of fish from longline landings was quite similar for males and females (Fig. 8). Modal length for all sectors and months was near to 285 cm TL, the mean size ap- proximately 280 cm TL. Minimum size was 180 cm TL for females, 195 cm TL for males. Maximum size was 40 cm greater for females (430 cm TL) than for males (390 cm TL). The frequency distribution by size of fish from gillnet landings differed by sex (Fig. 9), with two modes for males (290 and 330 cm TL) and for females (395 and 410 cm TL). The size range of fish by sector for longline land- ings was greater for females (183-432 cm TL) than for males (195-391 cm TL). Minimum sizes were found in sector 2 for both sexes, whereas the maxi- mum for males was recorded in sector 3 and for fe- males in sector 5. In gillnet landings, females ranged from 236 to 446 cm TL, with a mean size of 352 cm TL, 45 cm larger than the mean size of males (307 cm TL). As in longline landings, males had a smaller size range (246-373 cm TL). The length mode for fe- males was 80 cm larger than for males (410 cm TL compared with 330 cm TL). The size variation over time was not outstanding for males in either fishery, but there was a marked presence of large females during July-September in gillnet landings (288-446 cm TL) corroborated by longline landings in the same period (230-432 cm TL). The period of January to March was when the greatest number of fish were recorded ( 128 ) with the smallest size range ( 183-360 cm TL). Only one preg- nant female (385 cm TL), with one pup, was recorded in gillnet landings in September 1991. 678 Fishery Bulletin 96(4), 1998 Common thresher shark Only 52 common thresher sharks were recorded dur- ing the sampling period; hence the catch rate never went beyond 0.1 fish/1000 hooks (Tables 5 and 6). The overall sex ratio was close to 1 male:2 females, with some differences between longline (1 male:2.16 females) and gillnet (1 male: 1.75 females) landings. The size range of fish was similar for longline and gillnet landings, as was the mean size for both sexes (325 cm TL for males, 350 cm TL for females). Maxi- mum size was similar in longline (360 cm TL for males, 425 cm TL for females) and gillnet landings (355 cm TL for males, 435 cm TL for females). Mini- mum size was comparable for females in both fisher- ies (295 and 285 cm TL in longline and gillnet, re- spectively), but not for males; in longline landings, a 245-cm-TL male was recorded, whereas in gillnet landings the minimum-size male recorded was about the same size as the minimum-size female (280 cm TL). Although biometrical information on this spe- cies is found in Table 8, our small sample size did not permit analyses by sex. Only one pregnant fe- male was recorded, measuring 385 cm TL; it was caught in sector 4 by longline in May and carried 4 embryos (3 males and 1 female), 75 to 80 cm FL. Scalloped hammerhead shark The catch rate for this species was low in all sectors and months, slightly greater for sectors 1 and 3 (0.6 and 0.8 fish/1000 hooks respectively), and never higher than 1 fish/1000 hooks (Tables 5-7). The high- est overall catch rates by month in longline landings was in November 1991 and July 1992 (1.3 and 1.1 fish/ 1000 hooks). In gillnet landings the greatest catch rates were in September-October and June-July (0.1 fish/ unit effort). Males were more abundant in the longline ( 1 male:0.61 females) than in the gillnet fishery, where females were dominant (1 male: 1.37 females). Females had a larger maximum size than males in both fisheries, 320 cm and 305 cm TL for females, Buencuerpo et al.: Pelagic sharks associated with the swordfish fishery 679 280 and 275 cm TL for males, respectively, in longline and gillnet landings. Minimum size was greater in the gillnet fishery (185 cm TL for males, 165 cm TL for females) than in the longline fishery ( 105 cm TL for males, 115 cm TL for females). The mean size from longline landings was slightly greater for fe- males ( 170 cm TL) than for males (150 cm TL). The mean size from the gillnet fishery was greater than from the longline fishery, with no difference between sexes ( 220 cm TL). Although biometrical information can be found in Table 8, the small sample size did not permit analyses by sex. Discussion The total proportion of sharks in relation to sword- fish observed in the present study was similar to that found by Gouveia (1992) for an adjacent area, but the species composition was different (P. glauca, I. oxyrinchus, Dasyatis violacea, Alopias spp., Mustelus mustelus, Sphyrna spp., T. obesus, and Alepisaurus ferox). Sharks comprised a greater proportion of land- ings in the study area than in other Atlantic longline fishing grounds, i.e. off the Florida coast (Berkeley and Campos, 1988), the Caribbean coast (Tobias, 1991), and the northwestern coast of Cuba (Guitart, 1975). “Other bony fish” (the other category of fish) taken by longlines had a similar proportion to that found in data from other Mediterranean longline fish- eries (Rey and Alot, 1984). The abundance of the most common shark species, blue and shortfin mako sharks, also agrees with results presented for the same area by Garces and Rey (1984). Shortfin mako shark This species was very common in the study area. This finding agrees with reports by Munoz-Chapuli ( 1985), although catch rates obtained were slightly lower than estimates of Moreno and Moron (1992b) from the same area. Proportions of shortfin mako sharks, 680 Fishery Bulletin 96(4), 1998 6% Males = 123 Total length Figure 8 Total length-frequency distribution of the bigeye thresher shark by sex from longline landings from all sectors from July 1991 to July 1992. 6% Males = 123 4% • 2% c 2 t: o a. o 0% 75 200 -2% 4 -4% I -6% 4- Total length (cm) Figure 9 Total length frequency distribution of the bigeye thresher shark by sex from the gillnet landings from all the sectors from July 1991 to July 1992. Buencuerpo et a !.: Pelagic sharks associated with the swordfish fishery 681 compared with swordfish, were similar to proportions calculated by Garces and Rey (1983) for the north- eastern Atlantic and Mediterranean Sea but were greater than the estimates of Rey and Alot (1984) for the western Mediterranean Sea and those of Mejuto2 for the northeastern Atlantic just immedi- ately north of the study area. Monthly variation in the longline catch rate in this study was different from that found off the northwestern Spanish coast by Mejuto and Garces1 and Mejuto2, although it co- incides with the increased trend observed during the last few months of the year. Monthly catch rate varia- tion did not agree with results obtained on the other side of the Atlantic Ocean off the northwestern Cu- ban coast (Guitart, 1975). Shortfin mako shark and swordfish catch rates followed a similar decreasing trend from fall to March; in April the catch rate for shortfin mako sharks increased and that for sword- fish dropped (Fig. 10), thus complementing each other. Sex ratio (1 male:0.9 females) is different from that observed by Mejuto and Garces1 and Mejuto2 in the northeastern Atlantic above 40°N, where males were more abundant (approx. 1 male:0.4 females). The increasing presence of males northwards (up to 1 male:0.6 females in sector 4) correlates with results observed by Mejuto and Garces1 and Mejuto2 and may suggest sexual segregation to the north. For the same area of the present study, Munoz-Chapuli (1984) observed a lower proportion of females (1 male:0.35 females) than we did, and although the sample size was small in that study (n=113), there was a similar trend in the increasing number of males northwards. The overall sex ratio observed in this study agrees with results presented by Moreno and Moron ( 1992b) from August 1983 to August 1985 for the same fleet operating in the same area. Mejuto2 reported a trend of increase in percentage of males with increasing size (for size range 105-260 cm FL). This trend was not found in our study (Fig. 11). The length-frequency distribution of the two fish- eries was different from that found by Mejuto and Garces1 and Mejuto2 for the area immediately north. It was also different from the one presented for the east coast of the United States (Pratt and Casey, 1983). In our study, modal values were always lower than those presented by the above-mentioned au- thors. Following the age structure proposed by Pratt and Casey ( 1983), we estimated that most of the fish in this study belonged to age class 1 and 2 for both sexes. The largest size of females corresponded to age class 5 (sector 3), and the largest size of males to [mj 8 O 5 s CO Jul 91 Aug 91 Sep 91 Oct 91 Nov 91 Dec 91 Jan 92 Feb 92 Mar 92 Apr 92 May 92 Jun 92 Jui 92 tl.oxyrinchus 1 .l.Xgladius 9 P.glauca Figure 10 Comparison of longline catch rates (number of fish/1000 hooks) of the shortfin mako shark and swordfish (left “y” axis) and of the blue shark (right “y” axis) by month. 682 Fishery Bulletin 96(4), 1 998 age class 4 (sector 4). The minimum size corresponded to age class 0 in all sectors. Maximum size decreased during March-May (see Fig. 6, A and B). In June and July there was a pos- sible entry of newborn fish (65 cm) with the same birth size as that suggested by Compagno (1984). Mean size decreased during October-November. These changes in size distribution suggest a move- ment of the largest and mean-size shortfin mako sharks out of the study area. For an area just north of the one studied, Mejuto2 showed a large increase in catch rate during the last quarter of the year, which could be attributed partially to entry of fish coming from our study area. Blue shark The abundance of this species in longline fishing has been mentioned by several authors (Bigelow and Schroeder, 1948; De Metrio et al., 1984; Munoz- Chapuli, 1985; Stevens, 1990; Mejuto2). In the present study the proportion of blue sharks, com- pared with swordfish, was half that estimated by Garces and Rey ( 1983) for an area that included our five sectors. However, it was greater than the pro- portion obtained by Rey and Alot (1984) in the west- ern Mediterranean and very similar to that calcu- lated by Mejuto2 for the area just north of our study region. The number of fish landed increased north- ward and was greater than the number estimated by De Metrio et al. (1984) in the Mediterranean. The temporal distribution of fish was also differ- ent for the Mediterranean (De Metrio et al., 1984) and for the northeastern Atlantic (Mejuto2). As in the case of shortfin mako sharks, the highest catch rates correlated with reductions in swordfish catch rates, i.e. April and September (Fig. 10). The overall sex ratio (1 male:0.71 females) was inverse to that found by Rey and Alot (1984) in the Mediterranean. The latitudinal distribution of sexes was inverse to that estimated by Stevens (1990) off southwest Great Britain and Portugal and to that presented by Nakano et al. (1985) in a similar lati- tude in the central north Pacific. Using Pratt (1979) and Cailliet et al’s. (1983) age estimate, we found that frequency distribution of fish by size of landings in sector 1 showed that mature males were more abundant, whereas in the other sectors a greater proportion of immature fish were observed. Bigeye thresher shark The relatively frequent occurrence of this species noted by Munoz-Chapuli (1985) contrasted with the low numbers observed in our study, with the low 65 90 115 140 165 190 215 240 Fork length (cm) Figure 1 1 Proportion of male shortfin mako sharks in landings by size group (grouped by 5 cm FL). Buencuerpo et al.: Pelagic sharks associated with the swordfish fishery 683 numbers noted by Moreno and Moron ( 1992a) in the same area, and with the low numbers noted by Stillwell and Casey (1976) in the northwestern At- lantic. The sexual segregation hypothesized by Munoz-Chapuli (1984) in this area, i.e. males dis- tributed northward and females southward, was op- posite that of our findings. Moreno and Moron ( 1992a) mentioned that most fish caught during the fall were pregnant females, whereas in this study mostly ma- ture males from the iongiine fishery and mostly fe- males from the gillnet fishery were recorded during the same period. Most males in all sectors were adults (>276 cm TL, Moreno and Moron, 1992a). Most females would be have been considered mature in every sector accord- ing to Gubanov’s (1979) first maturity size (>310 cm TL), but only in sector 5 according to Moreno and Moron’s ( 1992a) criteria (340 cm TL). The largest fish was recorded in sector 5, a 432-cm-TL female, in Sep- tember 1991, which, nevertheless, did not exceed the maximum for the species (460 cm TL, Nakamura, 1935). Size distributions by month showed a greater pro- portion of immature males during the July-Septem- ber period, 'whereas mature fish predominated the rest of the year. During January- ••■June and October- December most females were around maturity size limit (300 cm TL, Gubanov, 1979). Most births occur during fall and winter accord- ing to Moreno and Moron (1992a), and as these au- thors have suggested, sector 5 could be a breeding area. A pregnant female was recorded in September 1991, in sector 5, carrying only one pup. However, no newborn fish were recorded, according to birth sizes reported by Bigelow and Schroeder ( 1948), Bass et al. (1975), Gruber and Compagno (1981), and Moreno and Moron (1992a). Common thresher shark The scarcity of this species in the present study dis- agrees with observations of Munoz-Chapuli (1985). The relative abundance of common thresher sharks found in sector 5 was similar to that presented by Moreno et al. (1989) for the same area. The maxi- mum number during spring agreed with the num- ber observed in California (Cailliet and Bedford, 1983) but did not coincide with the peak reported by Moreno et al. (1989) during fall in this area. The sex ratio for this species in the area studied ( 1 male:2 females) differed greatly from estimates by Holts (1988) on the west coast of the United States (1 male:l female). Moreno et al. (1989) noted the absence of males in May, which suggests sexual seg- regation during the reproduction period. A mature male was recorded in May at 330 cm TL, which rep- resents the lower limit of maturity for males accord- ing to Cailliet et al. (1983). Only one female was recorded below size of first maturity (260 cm TL, established by Gubanov, 1972; Cailliet and Bedford, 1983; Cailliet et al., 1983). Moreno et al. (1989) pointed out the probable exist- ence of a breeding area close to the Strait of Gibraltar during spring time. The record of a 425-cm-TL preg- nant female, with four full-term pups in May, sup- ports this theory. Scalloped hammerhead shark According to sex and length data obtained from sec- tors 1, 3, and 5, the overall sex ratio (1 male:0.83 females) differed from the ratio observed by Munoz- Chapuli (1984) in the same area (1 male:6 females). If we follow Compagno’s ( 1984) first maturity size criteria (256 cm TL for males, 304 cm TL for females), only 6% of males and 4% of females landed would be considered mature in our study area. Munoz-Chapuli (1984) observed many pregnant females in the area throughout the year, but the lowest size recorded in our study (114 cm TL) was much greater than birth size (50-60 cm TL, Compagno, 1984). Moreover, the lack of mature females contradicts the assessment that the study area is a breeding area, as suggested by Munoz-Chapuli ( 1985). Conclusions Sharks, because of their low reproduction rate and late sexual maturity, are extremely sensitive to fish- ing pressure (Holden, 1973, 1974, 1977). The vulner- ability of shark populations as bycatch of some tuna and tunalike fishing is comparable to that of marine mammals (Burke and Francis, 1990). In recent years several cases of overexploitation in shark fisheries (Cailliet and Bedford, 1983; Holts, 1988; Vas, 1990; Hanan et al., 1993) and the effect of other fishing activities (recreational fisheries, the routine proce- dure of discarding sharks after removal of fins) on shark populations (Casey and Hoey, 1985; Stevens, 1992) has been described. A lack of fishery statistics about sharks is com- monplace all around the world. Du Buit ( 1989) com- mented on the complete absence of knowledge about most biological factors influencing shark populations off the coast of France and on the difficulty in sam- pling most of these migratory species. Our study shows the importance of shark landings in the Atlantic swordfish Iongiine fishery and points out the large number of immature fish involved. 684 Fishery Bulletin 96(4), 1 998 Shortfin mako may be the shark species most affected by current fishing procedure despite the fact that the most common species caught is blue shark. As Casey and Kohler (1992) suggested with reference to the shortfin mako shark population in western north Atlantic, this species has a continuous distribution throughout the central and northeastern Atlantic Ocean, preferentially in the high production area of Saharian Bank (sector 3). As it has been shown in this study, sharks greatly affect and are greatly affected by swordfish fishing. Therefore, the management of this fishing industry should be reoriented to multispecies models in which the effect of bycatch and the economic implications of bycatch should be included in the management models for a proper approach to the present situa- tion in the industry. Acknowledgments We are indebted to Enrique Majuelos who has been indispensable in collecting data used in this paper and to Emilio Garcia, Florencio Gonzalez, Jesus Gutierrez, and Pedro Giiemes for collaboration in sampling. We would also like to thank Julio Alonso for his help in the statistical analyses. Finally, we want to acknowledge Manuel Alcantara for collabo- ration with the computer analyses and Jaime Mejuto and J. I. Castro for valuable review comments. Beverly Rising helped with manuscript translation. This work has been financed through project 2481 (1990) of the Universidad Complutense of Madrid. Literature cited Amorin, A., C. Arfelli, A. Garces, and J. C. Rey. 1979. Estudio comparative sobre la biologia y pesca del pez espada, Xiphias gladius L. (1758) obtenidos por las flotas espanolas y brasilena. Col. Vol. Sci. Pap. ICCAT 8(2):496- 503. Anonymous. 1986. Anuario de pesca Maritima 1985. Secretaria Gen- eral Tecnica. Ministerio de Agrieultura, Pesca y Alimen- tacion, Madrid, 501 p. Bass, A. J., J. D. D’Aubrey, and N. Kistnasamy. 1975. Sharks of the east coast of southern Africa. IV. The families Odontaspididae, Scapanorhynchidae, Isuridae, Cetorhinidae, Alopiidae, Orectolobidae and Rhiniodon- tidae. Inv. Rep. Oceanogr. Res. Inst., Durban 39:1-102. Belloc, G. 1934. Catalogue illustre des poissons comestibles de la cote occidentale d’Afrique (du Cap Spartel au Cap Vert). Premiere partie, Poissons cartilagineux. Rep. Trav. Off. (Scient. Tech.) Pech. Marit. 7(2):117— 195. Berkeley, S. A., and W. L. Campos. 1988. Relative abundance and fishery potential of pelagic sharks along Florida’s East Coast. Mar. Fish. Rev. 50( 1):9 — 16. Bigelow, H., and W. C. Schroeder. 1948. Fishes of the western north Atlantic. Pt. 1: sharks. Mem. Sears Found. Mar. Res. 1:59-576. Blasco, M., and R. Munoz-Chapuli. 1981. Presencia de Alopias superciliosus en las costas ibericas y datos sobre su morfologia. Arq. Mus. Bocage, ser. B. 1(6):53— 61. Bravo, J. 1974. Contribucibn al conocimiento de los Peces Con- droictios del Archipielago Canario. Bol. Inst. Espa. Oceanogr. 201:1-77. Bravo, J., and E. Santaella. 1973. Observaciones biologico pesqueras en el banco pes- quero sahariano. Bol. Inst. Espa. Oceanogr. 171:1-79. Burke, T. W., and T. C. Francis. 1990. Options for the management of tuna fisheries in the Indian Ocean. FAO Fish. Tech. Pap. 135:1-74. Cailliet, G. M., and D. W. Bedford. 1983. The biology of three pelagic sharks from California waters, and their emerging fisheries: a review. CalcCOFI Rep. 24:57-69. Cailliet, G. M., L. K. Martin, J. T. Harvey, D. Kusher, and B. A. Welden. 1983. Preliminary studies on the age and growth of blue, Prionace glauca, common thresher, Alopias vulpinus, and shortfin mako, Isurus oxyrinchus, sharks from California waters. In E. Prince and L. M. Pulos, (eds.). Proceedings of the international workshop on age determination of oce- anic pelagic fishes: tunas, billfishes, and sharks, p. 179- 188. U.S. Dep. Commer. NOAA Tech. Rep. NMFS 8. Casey, J. G., and J. J. Hoey. 1985. Estimated catches of large sharks by U.S. recreational fishermen in the Atlantic and Gulf of Mexico. U.S. Dep. Commerce, NOAA Tech. Rep. 31:15-19. Casey, J. G., and N. E. Kohler. 1992. Tagging studies on the shortfin mako shark (Isurus oxyrinchus ) in the western north Atlantic. Aust. J. Mar. Freshwater Res. 43:45-60. Compagno, L. J. V. 1984. FAO species catalogue. Vol. 4: Sharks of the world: an annotated and illustrated catalogue of sharks species known to date. Part 2: Carcharhiniformes. FAO Fish. Synop. 125, p. 251-655. De Metrio, G., G. Petrosino, C. Montanaro, A. Matarrese, M. Lenti, and E. Cecere. 1984. Survey on summer-autumn population of Prionace glauca L. Pisces, Chondrichthyes) in the Gulf of Taranto (Italy) during the four year period 1978-1981 and its inci- dence on sword fish (Xiphias gladius L.) and albacore (Thunnus alalunga (Bonn)) fishing. Oebalia 10 N.S.:105- 116. Du Buit, M. H. 1989. L’exploitation des Selaciens en France. Oceanis 15(3):333-344. Garces, A. G., and J. C. Rey. 1983. Analisis de la pesqueria espanola de pez espada, Xiphias gladius, entre los anos 1973 y 1981. Col. Vol. Sci. Pap. ICCAT 18:622-628. 1984. La pesqueria espanola del pez espada (Xiphias gladius ) 1973-1982. Col. Vol. Sci. Pap. ICCAT 20:419- 427. Garcia, C. 1970. La pesca en Canarias y en el banco sahariano. Publ. Cons. Eco. Soc. Sind. Inter. Canarias. Tenerife, 168 p. Buencuerpo et a L Pelagic sharks associated with the swordfish fishery 685 Gouveia, L. 1992. Swordfish (Xiphias gladius, Linnaeus) fishing experi- ment in Madeira EEZ. Col. Vol. Sci. Pap. ICCAT 39(2):477-483. Gruber, S. H., and L. J. V. Compagno. 1981. Taxonomic status and biology of the bigeye thresher, Alopias superciliosus. Fish Bull. 79(41:617-639. Gubanov, Y. P. 1972. On the biology of the thresher shark Alopias uulpinus in the northwest Indian Ocean. J. Ichthyol. 12:75. 1979. The reproduction of some species of pelagic sharks from the equatorial zone of the Indian Ocean. J. Ichthyol. 18:781-792. Guitart, D. 1975. Las pesquerias pelagico-oceanicas de corto radio de accion en la region noroccidental de Cuba. Acad. Cien. Cuba Ser. Oceanol. 31:1-26. Hanan, D. A., D. B. Holts, and A. L. Coan. 1993. The California drift gill net fishery for sharks and swordfish, 1981-82 through 1990-91. Fish Bull. 175: 1-95. Holden, M. J. 1973. Are long-term sustainable fisheries for elasmo- branchs possible? Rapp. P.-V Cons. Int. Explor. Mer 164:360-367. 1974. Problems in the rational exploitation of elasmobranch populations and some suggested solutions. In F. R. Harden-Jones (ed.), Sea Fisheries Research, p. 117— 137. Paul Elek, London. 1977. Elasmobranchs. In J. A. Gulland (ed.), Fish popu- lation dynamics, p. 187-215. John Wiley and Sons, New York, NY. Holts, D. B. 1988. Review of U. S. west coast commercial shark fish- eries. Mar. Fish. Rev. 50(1): 1-8. Lozano Cabo, F. 1950. Datos sobre la reparticidn geografica de especies de peces de la costa de NW Africa. Bol. R. Soc. Espa. Hist. Nat. Sec. Biol. 48( 1 ):5— 14. Moreno, J. A. 1982. Jaquetones. tiburones del genero Carcharhinus del Atlantico Nor-Oriental y Mediterraneo Occidental. Ministerio de Agricultura, Pesca y Alimentacion, Madrid, 205 p. Moreno, J. A., and J. Moron. 1992a. Reproductive biology of the begeye thresher shark, Alopias superciliosus (Lowe, 1839). Aust. J. Freshwater Res. 43:77-86. 1992b. Comparative study of the genus Isurus (Rafinesque, 1810), and description of a form (“Marrajo Criollo”) appar- ently endemic to the Azores. Aust. J. Mar. Freshwater Res. 43:109-122. Moreno, J. A., J. I. Parajua, and J. Moron. 1989. Biologia reproductiva y fenologia d e Alopias vulpinus (Bonnaterre,1788) (Squaliformes: Alopiidae) en el Atlantico nor-oriental y Mediterraneo occidental. Scient. Mar. 53( 1 ):37— 46. Munoz-Chapuli, R. 1984. Ethologie de la reproduction chez quelques requins de l’Atlantique Nord-Est Cybium 8(3): 1-14. 1985. Analisis de las capturas de escualos pelagicos en el Atlantico nororiental ( 15°-40°N). Inv. Pesq. 49( 1 ):67-79. Munoz-Chapuli, R., and M. Blasco. 1984. Tendencias generales del crecimiento relativo en escualos. Inv. Pesq. 48(3):3Q3-317. Nakamura, H. 1935. On the two species of the thresher shark from Formosan waters. Mem. Fac. Sci. Agric. Taihoku Imp. Univ. 14(1): 1-6. Nakano, H., M. Makihara, and K. Shimaziki. 1985. Distribution and biological characteristics of the blue shark in the central north Pacific. Bull. Fac. Fish. Hokkaido Univ. 36(31:99-113. Pratt, H. L. 1979. Reproduction in the blue shark, Prionace glauca. Fish. Bull. 77:445-470. Pratt, H. L., and J. G. Casey. 1983. Age and growth of the shortfin mako, Isurus oxyrinchus, using four methods. Can. J. Fish. Aquat. Sci. 40(11): 1944-1957. Rey, J. C., and E. Alot. 1984. Contribucion al estudio de la pesqueria de palangre del pez espada (Xiphias gladius) en el Mediterraneo Occidental. Col. Vol. Sci. Pap. ICCAT 20:428-434. Stevens, J. D. 1 990. F urther results from a tagging study of pelagic sharks in the north-East Atlantic. J. Mar. Biol. Assoc. (U.K.) 70:707-720. 1992. Blue and mako shark bycatch in the Japanese longline fishery off south-eastern Australia. Aust. J. Mar. Freshwater. Res. 43:227-236. Stillwell, C. E., and J. G. Casey. 1976. Observations on the bigeye thresher shark, Alopias superciliosus, in the western north Atlantic. Fish. Bull. 74( 1 ):22 1—225. Tobias, W. 1991. Billfish bycatch observer data of the U.S. swordfish longline fleet, St. Croix, U.S. Virgin Islands — 1988 and 1989. Col. Vol. Sci Pap. ICCAT 35(2):518-523. Vas, P. 1990. The abundance of the blue shark, Prionace glauca, in the western English Channel. Env. Biol. Fishes 29: 209-225. 686 Abstract. -a cladistic analysis of interrelationships for 53 (of 59) pleuro- nectid species was performed by using 106 morphological and osteological characters. The analysis resulted in 128 equally parsimonious cladograms (heu- ristic search, 403 steps, consistency in- dex=0.33, retention index=0.79). A 50% majority-rule consensus cladogram in- dicated that only five of 47 resolved nodes were observed in less than 100% of the cladograms. These five nodes are restricted to interrelationships within one subfamily. The Pleuronectidae is monophyletic according to ten synapo- morphies. In addition, five subfamilies were defined: Hippoglossinae, Eopset- tinae, Lyopsettinae, Hippoglossoidinae, and Pleuronectinae. The largest sub- family, the Pleuronectinae, was further subdivided into four tribes: Psettich- thyini, Isopsettini, Microstomini, and Pleuronectini. The interrelationships established within Pleuronectidae pro- vide a strong foundation for a simpli- fied yet phylogenetically informative taxonomic nomenclature at the genus- group level. The following genera are reclassified: Atherestes and Reinhard- tius to Reinhardtius', Errex, Glypto- cephalus, and Tanakius to Glypto- cephalus', Embassichthys , and Micros- totnus to Microstom us\ Hypsopsetta and Pleuronichthys to Pleuronichthys ; and Kareius and Platichthy to Platichthys. To preserve the monophyletic status of Eopsetta , E. exilis was reassigned to the genus Lyopsetta (Lyopsettinae). The genus Pleuronectes (as defined by Sakamoto in 1984) was found to be polyphyletic. Monophyly of this genus is established by revising it to include only five species; Pleuronectes glacialis, P. pinnifasciatus, P. platessus , P. putnami, and P. quadrituberculatus . Other species, formerly placed in Pleuronectes, are now reclassified to Isopsetta, Limanda, Parophrys, Pset- tichthys, and Pseudopleuronectes. The monophyletic status of Limanda (six spe- cies) is uncertain because of unresolved relationships between these species and other taxa in the tribe Pleuronectini. Manuscript accepted 10 March 1998. Fish. Bull. 96(4): 686-726 ( 1998). Monophyly and intrarelationships of the family Pleuronectidae (Pleuronectiformes), with a revised classification J. Andrew Cooper National Marine Fisheries Service, Systematics Laboratory Museum of Natural History, Washington, D.C. 20560-0153 E-mail address: cooperandrew@nmnh.si.edu Francois Chapleau Ottawa-Carleton Institute of Biology, Faculty of Science University of Ottawa, PO. Box 450, Station A, Ottawa, Ontario, KIN 6N5 The Pleuronectidae ( sensu Chapleau and Keast, 1988) contains 59 nominal species of right-eyed flatfishes distrib- uted in marine waters of the North- ern Hemisphere. As presently com- posed, this family excludes the sub- families Poecilopsettinae, Samarinae, Rhombosoleinae, and Paralichtho- dinae, formerly included in Norman (1934). It contains many commercial species that have long been harvested in coastal seas off Europe, North America, and Asia. Species such as the Petrale sole ( Eopsetta jordani), Pacific halibut ( Hippoglossus steno- lepis), American plaice ( Hippo - glossoides platessoides), and Dover sole (Microstomus pacificus), to name but a few, are valued for their large size and excellent meat (Hart, 1973; Scott and Scott, 1988). In to- tal, there are over 36 species of flat- fish monitored by the Food and Ag- ricultural Organization of the United Nations, with a total annual catch of 256,353 metric tons (t) in 1995 (FAO, 1997). Nearly one half (18) of these species are classified within the Pleuronectidae (FAO, 1997 ). Total annual commerical har- vest of Pacific halibut (Hippoglossus stenolepis) was estimated at 25,968 t in 1995 a decrease from 40,584 t in 1985 (FAO, 1997). A similar situ- ation has been observed in witch flounder ( Glyptocephalus cyno- glossus ) for which 19,537 t was har- vested in 1995 as compared to 30, 074 tin 1985 (FAO, 1997). With few exceptions, this decline in annual harvest is observed for most pleuro- nectid species (FAO, 1997). The commercial popularity of pleuro- nectid species, coupled with a need to manage these renewable re- sources, strongly emphasizes the necessity for a better understand- ing of relationships within this group, as well as an informative taxonomic nomenclature that will provide a framework for manage- ment policy. Variation in life history traits are observed at many phylo- genetic levels. Populations of the American plaice ( Hippoglossoides platessoides) are recorded to have age of maturity ranging from 3 to 15 years (Roff, 1981). Within the family, variation in maximum length ranges from 220 mm in Dexistes rikuzenius (Sakamoto, 1984b) to over 2500 mm in Hippo- glossus stenolepis (Hart, 1973). An hypothesis of species interrelation- ships can be used as a framework to assess phylogenetic constraint on life history traits versus a species’ ability to respond to changing envi- ronmental conditions (Brooks and McLennan, 1991). An assessment of Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 687 phylogenetic constraint versus environmental influ- ence could provide a more informed understanding of observed changes in life history traits. The objectives of this study are to clarify the mono- phyletic status of the Pleuronectidae, to offer an hy- pothesis of relationships within the group based on adult morphology, and to establish a phylogenetically informative nomenclature. To attain these objectives, a reassessment of morphological evidence in the lit- erature and new morphological characters were com- piled in a matrix analyzed within a cladistic frame- work. A new classification based on the phylogenetic information is offered. Historical classification and diagnosis of Pleuronectidae The Pleuronectidae was regarded by early ichthy- ologists to represent all known flatfishes (Norman, 1934). For example, Cuvier (1816) subdivided the Pleuronectidae into five subfamilies: Hippoglossinae, Pleuronectinae, Platessinae, Soleinae, and Cynoglos- sinae; as did Jordan and Goss ( 1889) who also added the subfamilies: Samarinae and Oncopterinae. Changes to the early classification in flatfish focused on revisions that accommodated new species with- out a complete revision of the entire scheme. Newly discovered species, thought to represent distinct morphological groups, were classified into new sub- groups; but original species, and those that did not have special morphologies, were to remain within the Pleuronectidae. Thus, the Pleuronectidae became a “garbage” group. Jordan and Evermann (1898) raised flatfishes to the suborder Heterosomata with two distinct fami- lies: Pleuronectidae and Soleidae. The Pleuro- nectidae, with three subfamilies: Hippoglossinae, Pleuronectinae, and Psettinae, were characterized by “a more or less distinct preopercular margin (i.e. not hidden by the skin and scales of the head); eyes large, well separated; mouth moderate or large; teeth present” (Jordan and Evermann, 1898). The Soleidae were subdivided into two subfamilies, Soleinae and Cynoglossinae, and were characterized by “an adnate preopercular margin, hidden by the skin and scales of the head; eyes small, situated close together; mouth very small, much twisted; teeth rudimentary or wanting” (Jordan and Evermann, 1898). Regan (1910) proposed a new classification that raised the Heterosomata to the level of order with two suborders: Psettodoidea and Pleuronectoidea. Within the second suborder, the Pleuronectidae now contained three subfamilies; Pleuronectinae, Samarinae, and Rhombosoleinae. The family was characterized by “having eyes on right side of head, nerve of left eye always dorsal, olfactory lamellae slightly raised, parallel without central rachis and eggs without oil globules” (Regan, 1910). This classification was adopted by Norman ( 1934), who incorporated minor revisions from Regan ( 1920, 1929) and Jordan ( 1923). The Pleuronectidae, at this point containing five subfamilies (Pleuronectinae, Samarinae, Rhombosoleinae, Poecilopsettinae, and Paralichthodinae) were characterized by Norman (1934) as “having eyes on the right side; optic chi- asma monomorphic, the nerve of the left eye always dorsal; dorsal fin extending forward on the head at least to above the eye; all the fin-rays articulated; pelvic of from 3 to 13 rays; mouth usually terminal, with the lower jaw more or less prominent; maxil- lary without a supplemental bone; palatines tooth- less; lower edge of urohyal deeply emarginate, so that the bone appears forked; preoperculum with free margin; nasal organ of blind side usually near edge of head, but sometimes nearly opposite that of ocu- lar side; vertebrae never fewer than 30; on each side a single post-cleithrum; ribs present; egg without an oil-globule in the yolk.” Later classifications removed the genera Br achy pleura and Lepidoblepharon from the Pleuronectidae and placed them in the Citharidae (Hubbs, 1945) but essentially agreed with the clas- sification proposed by Norman (1934). Nelson ( 1984) listed the Poecilopsettinae, Rhombo- soleinae, Samarinae, and Pleuronectinae as subfami- lies in Pleuronectidae on the basis of two character- istics: eyes almost always dextral and no oil globule in yolk of egg. Sakamoto’s (1984a) hypothesis of pleuronectid intrarelationships assumed that the Pleuronectinae, Samarinae, Rhombosoleinae, Poecilopsettinae, and Paralichthodinae were mono- phyletic because both eyes were on right side of the body, optic nerve of the left eye was always dorsal, preopercle had a free margin and fin rays were with- out spines. Hensley and Ahlstrom ( 1984), in a review of flatfish classification, indicated that the evidence for monophyly of Pleuronectidae (sensu Norman, 1934) was not convincing. The diagnostic characters reviewed in Norman (1934) were found to be plesiomorphic for the order or had distributions that were unknown for many pleuronectiform taxa (Hensley and Ahlstrom, 1984). Subsequent cladistic analysis of major taxa within the order supported the hypothesis that the Pleuro- nectidae was not monophyletic and suggested that the subfamilies Pleuronectinae, Samarinae, Rhombo- soleinae, and Poecilopsettinae should be elevated to the family level (Chapleau and Keast, 1988; Chapleau, 1993). This new interpretation of taxo- nomic ranks in right-eyed flounders was recognized 688 Fishery Bulletin 96(4). I 998 by Hensley (1993) and in part by Nelson ( 1994). Spe- cies in Samarinae of Nelson (1984) and Para- lichthodes algoensis were classified to Samaridae (Nelson, 1994). The subfamilies Pleuronectinae, Rhombosoleinae, and Poecilopsettinae, remained in Pleuronectidae. Nelson (1994) argued that without a comprehensive understanding of monophyly for some major groups it is difficult to provide an accu- rate revision of the nomenclature within the order. Although this approach is an obvious attempt to minimize unnecessary changes to the nomenclature, it does not reflect the understanding that only spe- cies in Pleuronectinae possess a bothoid caudal-fin complex that clearly distinguishes these 59 north temperate species as being closely related to Bothidae, Paralichthyidae, Scophthalmidae, and Brachypleura (Hensley and Ahlstrom, 1984; Chap- leau, 1993). The Poecilopsettinae, Rhombosoleinae, and Samaridae do not have this caudal-fin complex and are phylogenetically related to Achiridae, Soleidae, and Cynoglossidae (Chapleau, 1993). Given that Pleuronectinae is the nominotypical subgroup, it is correct to reclassify it to Pleuronectidae. For the species of Poecilopsettinae, and Rhombosoleinae, it is a misrepresentation of the present cladistic frame- work to classify them in Pleuronectidae when the order level phylogeny (Chapleau, 1993) suggests a relationship with Samaridae, Achiridae, Soleidae, and Cynoglossidae. Therefore, only the 59 nominal species of Pleuronectinae are considered in this study and classified as Pleuronectidae ( sensu Chapleau and Keast, 1988). Two tribes in Pleuronectinae (sensu Norman, 1934), the Hippoglossini and Pleuronectini, were clas- sified on the basis of jaw morphological characters (Nelson, 1984). The Hippoglossini identified by “mouth large and symmetrical; maxillae extending to or behind pupil of eyes; teeth well developed on both sides of jaws, contained ten genera (e.g. Atherestes, Eopsetta, Hippoglossoides , Hippoglossus, Lyopsetta, Psettichthys, and Reinhardtius).” The Pleuronectini were identified by “mouth small and asymmetrical; maxillae usually not extending to pupil of eye; teeth chiefly on blind-side of jaw and contained 16 genera (e.g. Embassichthys, Glypto- cephalus, Hypsopsetta, Isopsetta, Lepidopsetta, Limanda, Liopsetta, Microstomus, Parophrys, Platichthys, Pleuronectes, Pleuronichthys , and Pseudopleuronectes)" (Nelson, 1984). Although this classification was effective in identifying two mor- phological types within Pleuronectinae ( sensu Norman, 1934), it was not based on an examination of interrelationships within the group, nor did it ac- curately identify natural groups. The characters de- fining Hippoglossini were all plesiomorphic for the order and the characters defining the Pleuronectini were also observed in many lineages closely related to the Pleuronectinae (sensu Norman, 1934). The 59 nominal species in this group had been his- torically classified in as many as 28 genera, many of which were monotypic. This nomenclature was es- tablished prior to any understanding of phylogeny and reflected the morphological diversity within Pleuronectidae. The number of genera used in iden- tifying pleuronectid species would presumably be used to accommodate new species as they were dis- covered. However, the alpha taxonomy for this group has been well established, and only one new species of Pleuronectidae, Microstomus shuntovi Borets, 1983, has been described in the latter half of this century. Intuitively, a simplified and more informa- tive nomenclature with fewer monotypic genera would seem appropriate. There are few published studies that have dealt with phylogenetic relationships among pleuronectid taxa. The most extensive examination of interrela- tionships and classification within Pleuronectidae (sensu Norman, 1934) was established by Sakamoto (1984a). This phenetic hypothesis of interrelation- ships among 77 species was not aimed at defining relationships within an evolutionary framework; nor was it aimed at determining taxonomic structure on the basis of natural groups. Sakamoto (1984a) con- cluded his study with a reclassification of several genera within the Pleuronectinae (sensu Norman, 1934). In revision, the species of Eopsetta and Lyopsetta, as well as Cleisthenes and Hippoglossoides, were re- classified into Eopsetta and Hippoglossoides. Glypto- cephalus zachirus became Errex zachirus. All species of Isopsetta, Parophrys, Lepidopsetta, Limanda, Pseudopleuronectes, Pleuronectes, and Liopsetta were regrouped under the genus Pleuronectes. Finally Paralichthodes algoensis, previously classified in its own subfamily, Paralichthodinae (Nelson, 1994), was placed within the Pleuronectinae on the basis of over- all similarity and the presence of the first neural arch, a symplesiomorphy for the order (Hensley and Ahlstrom, 1984). Chiu ( 1990) examined the relationships among four glyptocephaline species (formerly classified in Glyptocephalus and Tanakius). The results of this phenetic analysis of body shape were similar to Sakamoto’s (1984a) results with respect to the rela- tionships between Glyptocephalus cynoglossus, G. stelleri, and G. zachirus. The limited scope of this analysis does not provide adequate information to infer relationships beyond these three species. Sakamoto’s (1984a) nomenclatural changes and classification were adopted in the American Fisher- ies Society checklist for flatfish species (Robins et Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 689 al., 1991). The adoption of this reclassification was a recognition of the first and only study that attempted to define intrarelationships in the Pleuronectidae ( sensu Norman, 1934). This new classification has not been widely accepted (Wheeler, 1992; Rass, 1996). It is argued here that because the Pleuronectidae used by Sakamoto (1984a) was determined to be poly- phyletic (Chapleau, 1993) and given the phenetic nature of the analysis, it is unlikely that the nomen- clatorial revisions summarized in that work repre- sent natural groups. The dubious nature of this most recent reclassifi- cation (i.e. Sakamoto, 1984a), the uninformative na- ture of the previous classification (i.e. Norman, 1934), and the commercial importance of this group require that a more comprehensive examination of pleuro- nectid intrarelationships be based on natural groups. A cladistic analysis based on structural variation within Pleuronectidae, in contrast with characters observed in closely related outgroups will establish an hypothesis of genealogical descent (Wiley, 1981). Outgroup hypothesis for Pleuronectidae Relationships within the order reveal that only the Pleuronectidae ( sensu Chapleau and Keast, 1988) have a caudal skeletal complex, synapomorphic with taxa belonging to the Paralichthyidae, Scoph- thalmidae, Brachypleura, and Bothidae (Chapleau, 1993). These taxa have been identified as the bothoid lineage within Pleuronectiformes (Hensley and Ahlstrom, 1984). This lineage is supported in one of 18 equally parsimonious trees observed in a cladis- tic analysis for the order (Chapleau, 1993). The other taxa formerly in Pleuronectidae (Samaridae, Rhombosoleidae, and Poecilopsettidae) were placed in a clade that included the soles Achiridae, Soleidae, and Cynoglossidae (Chapleau, 1993). Paralichthodes algoensis was not included in Chapleau’s ( 1993) cla- distic revision but has recently been determined to be the sister lineage to these other taxa formerly in Pleuronectidae (Cooper and Chapleau, 1998). Con- sequently, the most likely outgroup for Pleuro- nectidae would be represented by species within the bothoid lineage. There is a wide range of morphological types within the bothoid lineage for which the intrarelationships are not resolved (Chapleau, 1993). In the absence of synapomorphies to determine the sister relationship of Pleuronectidae with other bothoid taxa, a compari- son of jaw structure can be used to determine the most likely candidates. It is assumed that the outgroup for the Pleuronectidae should have large symmetrical jaws and pointed teeth. Within Pleuronectiformes, the evolutionary trend for jaw structure and feeding strategy may be considered unidirectional. Symmetry of jaw and dentition found in piscivorous flatfishes, like Psettodidae and Citharidae (de Groot, 1971), is considered to be the plesiomorphic condition. The osteological characters observed in these two taxa are most similar to the generalized acanthopterygian structure (Yazdani, 1969). Taxa with symmetrical jaw structure are hy- pothesized to have given rise to groups with more specialized dentition types and jaw asymmetry, as observed in the Achiridae, Soleidae, and Cyno- glossidae (Yazdani, 1969; Chapleau, 1993), but the reverse situation is not indicated in any study of re- lationships. Within subgroups of Pleuronectiformes, the same evolutionary trend is assumed to occur. Left-eyed flounders within the bothoid lineage have large, nearly symmetrical jaws for piscivory, or a more spe- cialized, asymmetrical jaw structure that accommo- dates capture of benthic prey (Yazdani, 1969; de Groot, 1971). Likewise, the Pleuronectidae contains both piscivores with nearly symmetrical jaws, Hippoglossus stenolepis and Reinhardtius hippoglossoides, as well as more specialized predators with asymmetrical jaws, such as Glyptocephalus stelleri (de Groot, 1971). As- suming that evolutionary trends in pleuronectid jaw structure are consistent with trends observed in the order, the ancestral pleuronectid would have near symmetrically developed jaws. The bothoid family, Paralichthyidae, appears to be one of the most plesiomorphic groups of left-eyed flounders and is chosen as the outgroup for the Pleuronectidae. In addition, Psettodes and Lepidoblepharon are also chosen as secondary outgroups. These assumptions are only valid if the Pleuronectidae is monophyletic. If Pleuronectidae is not monophyletic, then multiple outgroup taxa with either symmetrical or asymmetri- cal jaw structures may account for the variation ob- served within the Pleuronectidae. Materials and methods Fifty-three of 59 pleuronectid species were examined. Five outgroup taxa, chosen from the families Psetto- didae (Psettodes sp.), Citharidae (Lepidoblepharon ophthalmolepis), and Paralichthyidae ( Citharichthys arenaceus , Paralichthys lethostigmus , and P. squamilentus ) were also examined. The following cleared and stained specimens were dissected and examined for osteological characters. Nomenclature follows the conclusions of this analysis with the pre- vious classification of Sakamoto ( 1984a) indicated in 690 Fishery Bulletin 96(4), I 998 parentheses. Changes in species-group nomenclature recognize gender status of the genus (Eschmeyer, 1990) as specified by the International Code of Zoo- logical Nomenclature, article 31 (Ride et al., 1985). Institutional abbreviations follow Leviton et al. (1985). Length, in mm, is standard length (Hubbs and Lagler, 1970). Radiographs for all specimens, as well as radiographs of the specimens listed in Leipertz (1987), were also examined. Psettodidae Psettodes sp. Bennet;ANSP 145394, 65 mm. Citharidae Lepidoblepharon ophthalmolepis Weber; AMS 1.20118-012, 122 mm. Paralichthyidae Citharichthys arenaceus Evermann and Marsh; USNM 00203510, 69 mm. Paralichthys lethostigmus (Jordan and Gilbert); ANSP 143209, 90 mm. Paralichthys squamilentus Jordan and Gilbert; ANSP 150694, 50 mm. Pleuronectidae Acanthopsetta nadeshnyi Schmidt; USNM 77122, 89 mm. UW 22792, 224 mm; Cleisthenes ( -^Hippoglossoides ) herzensteini (Schmidt); USNM 051441, 85, 89 mm. Cleisthenes ( =Hippo - glossoides ) pinetorum Jordan and Starks; UMMZ 159566, 76 mm. Clidoderma asperrimum (Temminck and Schlegel); not examined. Dexistes rikuzenius Jor- dan and Starks; UMMZ 159662, 122 mm. Eopsetta grigorjewi (Herzenstein); UMMZ 159590, 66, 88 mm. Eopsetta jordani (Lockington). NMC 81-1015, 68 mm. Glyptocephalus cynoglossus (Linnaeus); NMC 77- 1087, 89, 114 mm. Glyptocephalus kitaharai (Jordan and Starks); UMMZ 141741, 139 mm. Glyptocephalus stelleri (Schmidt); UMMZ 159566, 125 mm. Glypto- cephalus i=Errex) zachirus Lockington; NMC 65- 0211, 95 mm; NMC 81-1027, 133 mm. Hippo- glossoides dubius Schmidt; not examined. Hippo- glossoides elassodon Jordan and Gilbert; NMC 61- 0117, 71, 82 mm. Hippoglossoides platessoides (Fab- ricius); NMC 80-0601, 31, 61, 63, 77 mm; ROM 504CS, 73 mm; ROM 786CS, 70, 83, 87, 90 mm. Hippoglossoides robustus Gill and Townsend; ANSP 105133, 110 mm. Hippoglossus hippoglossus (Linnaeus); ARC 8808487, 148 mm. Hippoglossus stenolepis Schmidt; NMC 61-0072, 100 mm; UW 22743, 54, 56, 67 mm. Isopsetta ( ^Pleuronectes ) isolepis (Lockington); UMMZ 63214, 119, 123 mm. Lepidopsetta ( ^Pleuronectes ) bilineata (Ayres); NMC 61-0050, 51 mm; NMC 81-1027, 56 mm. Lepidopsetta {-Pleuronectes ) mochigarei Snyder; UMMZ 159575, 113 mm. Limanda (= Pleuronectes ) aspera (Pallas); NMC 66-0016, 156 mm. Limanda (= Pleuronectes ) ferruginea (Storer); NMC 80-0217, 102, 107 mm; ROM 560CS, 49 mm. Limanda (= Pleuronectes ) limanda (Linnaeus); MNHN 1959-560, 119 mm; Limanda ( ^Pleuronectes ) proboscidea Gilbert. USNM 268496, 141 mm; UW 22742, 98, 115 mm. Limanda (= Pleuronectes ) punctatissima (Stein- dachner); HUMZ 93958, 135 mm. Limanda ( -Pleuro - nectes) sakhalinensis Hubbs; HUMZ 60455, 138 mm. Lyopsetta ( =Eopsetta ) exilis (Jordan and Gilbert); NMC 60-0501, 111, 115 mm . Microstomus achne (Jor- dan and Starks); UMMZ 159434, 145 mm. Microsto- mus ( =Embassichthys ) bathybius (Gilbert); UW 22791, 167 mm. Microstomus kitt (Walbaum); FMNH 35527, 135 mm. Microstomus pacificus (Lockington). NMC 81-1027, 109 mm. Microstomus shuntovi Borets; not examined. Parophrys (= Pleuronectes ) vetula Girard; FMNH 97128, 97, 128 mm; NMC 81- 1121, 39, 61, 66, 68 mm; NMC 85-0025, 53, 67 mm. Platichthys i-Kareius) bicoloratus (Basilewsky); UMMZ 159667, 117 mm . Platichthys flesus (Linnaeus); ANSP 93141, 78 mm. Platichthys stellatus (Pallas); NMC 61-0044, 53, 85, 90 mm. Pleuronectes glacialis Pallas; NMC 62-0352, 72 mm. Pleuronectes pinni- fasciatus (Kner) Steindachner and Kner. HUMZ 75681, 150 mm. Pleuronectes platessus (Linnaeus); ANSP 93145, 66, 88 mm. Pleuronectes putnami (Gill); ROM 23214, 28 mm, ROM 556CS, 104, 110 mm. Pleuronectes quadrituberculatus Bean; USNM 064042, 85 mm. Pleuronichthys coenosus Girard; not examined. Pleuronichthys cornutus (Temminck and Schlegel); UMMZ 159618, 91 mm. Pleuronichthys decurrens Jordan and Gilbert; CAS 23703, 58 mm. Pleuronichthys ( =Hypsopsetta ) guttulatus Girard; NMC 74-0242, 64 mm. Pleuronichthys ocellatus Starks and Thompson; CAS 82189, 105 mm. Pleuronichthys ritteri Starks and Morris; CAS 11403, 46 mm. Pleuronichthys verticalis Jordan and Gilbert; CAS 34728, 83 mm. Psettichthys melanostictus Girard; NMC 62-2158, 58, 63 mm. Pseudopleuro- nectes (= Pleuronectes) americanus (Walbaum); ANSP 105133, 43, 66 mm; NMC 82-0016, 61, 73 mm; ROM 670CS , 27, 35, 46, 49, 54, 60 mm. Pseudopleuronectes (= Pleuronectes ) herzensteini (Jordan and Snyder); UMMZ 159631, 88 mm. Pseudopleuronectes {- Pleuro- nectes) obscurus (Herzenstein); not examined. Pseudopleuronectes {-Pleuronectes) schrenki (Schmidt); HUMZ 75697, 123 mm. Pseudopleuronectes ( =Pleuro - nectes) yokohamae (Gunther); UMMZ 159548, 58, 83 mm; UMMZ 220249, 74 mm; USNM 056359 86 mm. Reinhardtius {-Atheresthes) evermanni (Jordan and Starks); not examined. Reinhardtius hippoglossoides (Walbaum); NMC 64-0756, 103, 119 mm. Rein- hardtius {=Atheresthes) stomias (Jordan and Gilbert); NMC 65-0262, 97 mm; NMC 66-0022, 39 mm; NMC 80-0073, 80, 117 mm; NMC 80-1024, 119 mm. Verasper moseri [Jordan and Gilbert] Jordan and Evermann; USNM 056385, 78 mm. Verasper variegatus (Temminck and Schlegel); USNM 056375, 91 mm. All characters used in the study are described in the Appendix. This list also includes the number of steps and character consistency index (cci) for each character in the analysis. The character matrix Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 691 (Table 1), illustrating the distribution of 106 charac- ters for 58 taxa, combines information from the lit- erature (Norman, 1934; Batts, 1964; Amaoka, 1969; Ahls trorn et ah, 1984; Hensley and Ahlstrom, 1984; Sakamoto, 1984a) as well as 67 new morphological features observed through examination of cleared and stained material, whole preserved specimens, and radiographs. Morphological states obtained from the literature are indicated by the citation immedi- ately following the italic character description (Ap- pendix). Morphological states based on meristic counts represented the modal value of the sample population. Characters at each node of the cladogram are described and numbered in order of presenta- tion in the text. Given the unresolved nature of the interrelation- ships for Pleuronectidae and other bothoid taxa within the order (Chapleau, 1993), the establishment of character polarity by outgroup comparison, as outlined by Watrous and Wheeler (1981) and Maddison et al. (1984), was not possible. Character polarity was determined through a direct examina- tion of states observed in outgroup taxa (Table 1). For each character, the majority state observed in the three bothoid taxa Citharichthys arenaceus, Paralichthys lethostigmus, and P. squamilentus was assumed to represent the plesiomorphic condition. This decision was only overruled if there was het- erogeneity in the distribution of states within these three taxa and if both secondary outgroup taxa Lepidoblepharon ophthalmolepis and Psettodes sp. possessed the alternative state. One exception to this rule is stated in character 82 (Appendix). Character states hypothesized as plesiomorphic for the family are coded as zero (0). Heuristic search methods The matrix was analyzed with all combinations of heuristic search parameters available in PAUP 3.1. 1.1 An exhaustive search of the most parsimoni- ous tree with this many taxa (53) would have re- quired the analysis of an estimated 2.84 x TO82 bifur- cating trees (Felsenstein, 1978) and is not possible within the current standard of computational time. For these same reasons, a branch and bound search technique proved inadequate to resolve relationships for more than 25 taxa (Forey et al., 1992). A two-step procedure (a sequential addition of taxa to produce a cladogram that minimizes homoplasy followed by the subsequent branch-swapping of this addition tree to search for more parsimonious cladograms) was 1 Swofford, D. L. 1991. PAUP Version 3.1.1. Unpublished software documentation. used to search for the most parsimonious result. The random addition sequence in combination with all of the branch-swapping algorithms was a nonrigorous means of assessing the efficiency of the heuristic methods (Forey et al., 1992). If 50 random replicates give the same set of tree topologies, then it is likely that the maximally parsimonious trees have been found. However, if after 100 replicates shorter cla- dograms are still being found, then it is likely that more trees remain (Forey et al., 1992). To minimize confounding effects of local minima, the “keep” option was used on successive searches to allow swapping on nonminimal trees (Forey et. al., 1992). Both the “MULPARS” and “swap on all trees” options were employed during each heuristic search to minimize the effect of plateau (Forey et. al., 1992). All searches assumed that the ingroup was mono- phyletic, and all uninformative characters were ig- nored. Character optimization was set for acceler- ated transformation ( ACCTRAN ). The five outgroup taxa were not included in the analysis. Instead, an- cestral states for all characters were set as zero ac- cording to established character polarity to repre- sent a hypothetical outgroup. All most parsimonious trees were saved from each search and combined (without duplication) to establish a 50% majority- rule consensus of the equally parsimonious results. Character analysis, character consistency index (cci), tree statistics, and tree presentations were gener- ated with MacClade version 3.04 (Maddison and Maddison, 1992). Results and discussion Phylogenetic analysis The heuristic searches found multiple trees of equal length (Table 2). The most parsimonious trees were found to be 403 steps from a minimum of 131 steps, with a consistency index (ci) of 0.33, excluding unin- formative characters, and a retention index (ri) of 0.79. Additional rounds of heuristic search allowed swapping on nonminimal trees up to 410 steps but did not resolve cladograms shorter than 403 steps. The “simple” and “closest” addition sequences were biased by taxa with large numbers of unknown char- acter states. Criteria for establishing the initial tree and the addition of taxa are strongly influenced by unknown character states in these two algorithms. As a result, Reinhardtius evermanni, Clidoderma asperrimum, Hippoglossoides dubius, Microstomus shuntoui, Pleuronichthys coenosus, and Pseudo- pleuronectes obscurus were not included in the heu- ristic searches. The “as is” and “simple” addition se- 692 Fishery Bulletin 96(4), 1998 . Table 1 Matrix of 106 morphological characters for 5 outgroup taxa, representing the Psettodidae, Citharidae, and Paralichthyidae and 53 ingroup! Blanks represent the hypothesized ancestral state (state 0). Question marks (?) indicate an unknown state. Character states deemed not Character number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 4243 44 45 46 47 48 49 Psettodidae Psettodes sp. ? 1 1 1 na 1 1 1 na na na 1 1 Citharidae Lepidoblepharon ophthalmolepis 1 na 1 1 na na na 1 1 1 Paralichthyidae Citharichthys arenaceus 1 1 2 1 1 1 1 1 1 2 1 Paralichthys lethostigmus 1 1 1 1 1 1 P. squamilentus Pleuronectidae Acanthopsetta nadeshnyi 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 2 1 1 1 1 1 1 Cleisthenes herzensteini 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 1 1 1 1 1 1 1 C. pinetorum 1 1 ? 1 1 1 1 1 1 1 1 ? 1 1 1 1 2 1 1 1 9 1 1 ? 1 1 ? ? ? 1 1 1 1 Dexistes rikuzenius 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 1 1 1 1 1 1 1 Eopsetta gngorjewi 1 1 1 1 1 1 1 1 1 1 1 1 1 E. jordani 1 1 1 1 1 1 1 1 1 1 ? 1 ? ? ? ? ? ? ? ? ? 1 ? ? ? 1 Glyptocephalus cynoglossus 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 G. kitaharai 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 G. stelleri 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 G. zachirus 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 2 1 1 1 1 1 1 1 1 1 Hippoglossoides elassodon 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 1 1 1 1 H. platessoides 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 1 1 1 H. robustus 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 1 1 1 Hippoglossus hippoglossus 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 H. stenolepis 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 Isopsetta isolepis 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 2 1 1 1 1 1 1 Lepidopsetta bilineata 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 ? 1 1 1 1 9 1 L. mochigarei 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 1 1 1 Limanda ctspera 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 2 1 1 1 1 1 1 1 1 5 L. ferruginea 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 2 1 1 1 1 1 1 L. limanda 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 1 1 1 1 1 L. proboscidea 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 2 1 1 1 1 1 1 l L. punctatissima 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 L. sakhalinensis 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 2 1 1 1 1 1 1 1 1 Lyopsetta exillis Microstomus achne 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l 1 M. bathybius 1 1 1 1 1 1 1 1 ? 1 9 2 1 1 1 2 1 1 2 1 1 1 1 ? 1 9 1 1 1 1 1 M. kitt 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M. pacificus 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Parophrys vetula 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 2 1 1 1 1 1 1 1 1 1 5 Platichthys bicoloratus 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 2 2 1 1 1 1 1 2 1 ? ? 1 i : P. flesus 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 2 2 1 1 1 2 1 1 i ; P. stellatus 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 2 1 2 2 1 1 1 2 1 1 i : Pleuronectes glacialis 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 2 1 1 1 2 1 1 1 1 1 1 2 1 1 i : P. pinnifasciatus 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 2 1 1 1 1 1 1 1 2 1 1 i ; P. platessus 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 2 1 1 1 2 1 1 1 2 1 1 1 2 1 1 1 i ; P. putnami 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 2 1 1 1 2 1 1 1 1 1 1 1 2 1 1 i P. quadrituberculatus 1 1 1 1 1 1 1 1 1 1 1 1 1 9 1 2 2 1 1 1 ? 1 1 ? ? ? ? 1 ? 1 1 2 ? 1 ? ? 1 1 Pleuronichthys cornutus 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ? 1 P. decumns 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 P. guttulatus 1 1 1 1 1 1 1 1 1 9 1 1 1 1 1 1 1 1 1 2 1 1 P. ocellatus 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 P. ritteri 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 P. verticalis 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Psettichthys melanostictus 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 1 1 2 1 1 1 1 1 1 Pseudopleuronectes americanus 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 2 1 1 1 1 1 1 2 1 1 l : P herzensteini 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 2 1 1 1 1 1 1 1 1 1 l i P. schrenki 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 2 1 1 1 2 1 1 1 1 1 1 2 1 1 1 ! P. yokohamae 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 2 1 1 1 1 1 1 2 1 1 l : Reinhardtius hippoglossoides 1 1 1 1 1 1 1 1 1 1 ? 2 1 1 2 1 1 1 1 l R. stomias 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l Verasper moseri 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 1 1 2 1 l V. variegatus 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 1 1 2 1 1 l Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 693 axa representing the Pleuronectidae. Numbers in the matrix represent character state for each morphology as described in the Appendix, tpplicable due to extreme variation in morphology are coded “na.” 50 51 52 53 54 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 2 na 1 1 1 1 1 I 2 4 II 2 4 2 1 1 1 ? 1 1 9 1 1 1 1 1 1 2 ? 9 1 1 l l 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 l l 1 1 l 1 1 ? ? 9 1 ? i 9 1 9 1 9 2 1 1 1 1 1 1 1 1 1 2 1 l 1 1 3 2 1 1 1 1 1 1 1 1 1 1 1 1 2 1 l 1 1 4 2 1 1 2 1 1 1 1 1 1 1 1 1 2 1 l 1 1 1 3 2 ? 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 2 1 l 1 2 1 3 2 1111 1 1 1 1 1 1 1 l 1 1 1 1 1 1 l 1 1 1 1 1 ? 1 1 i l 1 1 1 1 1 1 1 l 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 i : i l 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 l : i l 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l l 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l i l 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l l l 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l l l 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l l l 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l i i 1 1 2 1 1 1 1 1 2 2 1 3 1 1 l 1 1 l 1 1 2 1 1 1 1 1 1 2 1 1 ? 1 1 2 2 1 2 l 9 1 1 2 2 1 ? 2 1 1 1 1 1 2 2 1 3 1 l 1 l 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 2 1 l l 1 l 1 1 2 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 2 l 1 1 1 3 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 l l l 1 1 1 1 1 3 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 l l l 1 1 1 1 1 1 1 3 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 l l l 1 1 1 1 1 3 1 1 1 1 1 1 1 2 1 1 1 2 2 2 l l l 1 1 1 1 3 1 1 1 1 1 1 1 2 2 1 1 2 2 2 l l l 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 2 2 2 l l i 1 1 1 1 2 3 1 1 1 1 1 1 1 2 1 1 1 2 2 2 l l l 1 1 1 1 1 ? 1 1 1 1 1 1 1 1 2 2 1 1 2 1 1 l l l 1 1 1 2 1 1 1 1 1 1 2 2 i i 1 1 . 1 1 1 1 1 1 1 1 1 1 1 2 2 l l 1 1 l 1 1 1 1 2 1 1 I 1 1 1 1 2 1 i i 1 1 . 1 1 1 1 1 1 1 1 2 2 9 l l 1 1 . 1 1 1 1 1 1 1 1 1 1 2 2 l l 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 2 2 l l 1 1 . 1 1 1 1 1 1 2 1 1 1 . 1 1 l i 1 1 1 2 1 1 1 1 1 1 1 1 2 2 1 2 1 1 l i i l l 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 2 1 1 l l l 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 2 1 1 l i i l l 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 2 1 1 l 1 i l i 1 1 1 1 1 1 1 1 1 1 1 ^97 Fishery Bulletin 96(4), 1998 Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 693 Table l Matrix of 106 morphological characters for 5 outgroup taxa, representing the Psettodidae, Citharidae, and Paralichthyidae and 53 ingroup Blanks represent the hypothesized ancestral state (state 0). Question marks (?) indicate an unknown state. Character states deemed not taxa representing the Pleuronectidae. Numbers in the matrix represent character state for each morphology as described in the Appendix, applicable due to extreme variation in morphology are coded “na.” Character number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 4243 44 45 46 47 48 43 50 51 52 53 54 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 Psettodidae Psettodes sp. ? 1 1 1 na 111 na na na 1 i Citharidae Lepidoblepharon ophthalmolepis 1 na 1 1 na na na 1 l ] Paralichthyidae Citharichthys arenaceus 112 111 11 12 1 Paralichthys lethostigmus 1 1 1111 P. squamilentus 1 1 1 Pleuronectidae Acanthopsetta nadeshnyi 1111111111111112 1112 1 1 1 1 1 Cleisthenes herzensteini 1 1 L 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 1 11 1 1 | i C. pinetorum 1 1 ? 1 1 1 1 1 1 1 1 ? 1 1 1 1 2 1 1 1 ? 1 1 ? 1 1 ? ? ? 1 1 l j Dexistes rikuzenius 1111111111111112 1112 1 11 1 11 1 1 1 1 Eopsetta grigorjewi 1111111111 11 1 E.jordani 11111111117 1? ? ? ? ? ? ? ? 7 17? ? l Glyptocephalus cynoglossus 1111111111 11111 11 11 2 1 11 1 1 1 1 i G. kitaham 11111111111111 111 11111 1 11 111 G.stelleri 1111111111 111111111 2 111 1 111 1 G. zachirus 1111111111 1112 1111 2 111 1 1111 i Hippoglossoides elassodon 11111111111111112 1112 11 11 111 H. platessoides 1111111111111112 1112 11 1 111 H.robustus 1111111111111112 1112 11 1 111 Hippoglossus hippoglossus 111111111 1 111 11111111 1 H. stenolepis 111111111 1 12 1 111111 1 Isopsetta isolepis 111 111111111211211121 11 1 1 1 Lepidopsetta bilineata 11111111111111112 1117 1 11 1 7 1 L. mochigarei 11111111111111112 1112 1 11 1 11 Limanda aspera 1111111111112 112 1112 1 1 1 111 112 L. ferruginea 111 11111111211211121 1 1 1 1 12 L. limanda 11111111111 11112 1112 1 1 1 11 1 112 L. proboscidea 1111111111112 112 1112 1 1 1 11 1 12 L. punctatissima 111 1111111211211111 1 1 11 1 112 L. sakhalinensis 11111111111 1211211121 1 l ll l 112 Lyopsetta exillis 111111111111111 1 1111 1 Microstomus achne 11111 111111 2 112 11 1 1 11 l l l 111 M. bathybius 111111 11? 1? 21 11211 2 1 11 1? 1? 1 1111 M.kitt 11111111111 2 112 11 1111 1 1111 111 M.pacificus 11111 11111 1 2 112 11 1 1 11 1 l l l 1 1111 Parophrys vetula 111 111111111211211121 11 1 1 1 1 112 Platichthys bicoloratus 11111111111112112111 1 221 1 1112 1 7? I'3 P flesus 11111111111111112 1112 11 1 2 2 1 112 1 1 1 3 P. stellatus 111111111111211211121 22 1 112 1 l'3 Pleuronecles glacialis 1111111111111212211121 11 l 112 1 113 P. pinnifasciatus 1111111111111 12 2 1112 11 11 1 112 1 1 1 3 P. platessus 11111111111112 12 2 1112 11 12 1 112 1 1 1 1 3 P putnami 11111111111112 12 2 1112 11 11 1 112 1 113 P quadrituberculatus 11111111111117 12 2 1117 11 fill 1 7 1 1 2 7 1 7 7 1 1 2 Pleuronichthys cornutus 11111111 1 1 1111 111 11 1111 l i 1 1 7 1 Pdecurrens 111111 1111 1211 111 ll 1 1 1 1 1 1 1 1 1 Pguttulatus 11111111 17 1111 11 111 211 Pocellatus 11111111111 1211 111 ll ill 1 11 1 PritUri 111111 111 1211 111 ll ill 111 1 P vertical* 11111111111 1211 111 ll 1 1 1 1 1 1 1 1 1 Psettichthys melanostidus 111 1111111121211121 1 111 1 Pseudopleuronectes americanus 1 1 1 1 1 1 1 1 1 1 11 12 2 1112 1 11 1 112 1 1,2 P herzensteini 1111111111111 12 2 1112 1 1 11 1 11 1 1 1 2 R schrenkl 11111111111112 12 2 1112 1 11 1 112 1 1 1 2 P yokohamae 1111111111111 12 2 1112 1 11 1 112 1 1,2 Reinhardtius hippoglossoides 111111111 1 ’ 2 1 1 2 1 111 1 R;slmm 11 1111 I 1111111 1 1 1 1 Verasper mosen 11111111111 i , , 11221 11112 1 1 v^s ..1.1,11,1,, , , \\\i\ j;;;; ; 2 na 1 1 1 2 4 1 ? na 1 1 1 1 1 2 4 2 1 1 i 1 1 1 1 1 ? 1 1 1 1 1 7 111 11 12 111 1 1 7 7 111 1 11 111112 11111 11 1 1 1 111 1 ? ? 7 1? 17 17 1 7 1 2 11 1111 1112 1 1 1 13 2 111111 1 I 111112 1 1 1 14 2 11 1 2 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 3 2 7 1 1 1 1 1 1 1 II 1111 1112 1 1 1 2 13 2 1111111 1 111 111 11 111 111 11 111 7111 1 1 1111 11 1 1111 1 111111111111 11 1 1 2 111 1111111111111 1 1 12 111 1111111111111 1 1 12 11 11111111111111 1 1 11 1 I ! 2 1 1 11111111111111 11 1 * 1 11 1 112 11 11111111111111 11 1 1 112 11 11111111111111 11 1 1 11111 112 11 11111111111111 11 1 1 11111 1 12 11 11111111111111 11 1 1 1 1 1 1 2 1 1111 2 2 13 1 1 11 12 1111 1 1 12 1 17 112 2 12 1 9 1 12 2 1 ? 1 2 1 1111 2 2 13 1 1 11 11111 1 1 1 1111 2 2 12 111 11 12 111 1 12 11111111112 112 12 1 1 1 1 13 11 111111111112 2 1 11 1 1 111 11 13 11 111111111112 11 11 11 11111 11 13 11 111111111112 11 11 11 1 111 11 3 1 1 1 1 1 1 1 2 1 1 1 2 2 2 1 1 1 1 1 1 1 1 1 3 1 1 1 1 1 1 1 2 2 1 1 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 2 2 2 1 1 1 1 1 1 1 2 1 3 1 1 1 1 1 1 1 2 1 1 1 2 2 2 1 1 1 1 1 111 1 ’ll 11111122112111 11 1 112 1 1 1 1 1 1 1 2 2 11111111 1 1 1 1 1 1 11 2 2 111111112 1 11111112 1 11111 1 1 1 1 1 1 1 1 2 2? 11111111 1 111 11 22 111111112 1 1 1 1 1 1 1 1 11 2 2 11111111 1 1 1 2 1 1 1 1 1 11 1 1 1 1 2 1 1 1 1 1111 2 2 12 111111 1 111 1 1 1 1 1 1111 2 2 112 111 11 1 1 1 ! 1 1 11 111112 2 112 111111 1 111 1 1 1 1 1 1 1111112 2 12 111111 1 1 111 1 1 1111 11 1 111 1 111 1 111 1 694 Fishery Bulletin 96(4), 1 998 Table 2 Heuristic search results for an analysis of 106 characters for 53 ingroup taxa. Addition sequence Branch- swapping algorithm Tree length (steps) Number of trees found Cumulative total of unique trees as is tbr 403 112 112 as is spr 403 80 112 as is nni7 405 16 128 closest tbr 403 16 128 closest spr 403 16 128 closest nni 403 16 128 simple tbr 403 112 128 simple spr 403 16 128 simple nni7 409 48 128 random2 tbr 403 128 128 random2 spr 403 128 128 random2 nni 403 128 128 7 Trees found by this method were not of minimum length and not added to the total number of trees observed. 2 One hundred replicates for each random addition sequence. quences using the “nearest neighbour interchange” branch-swapping were the only combinations that did not find trees with 403 steps (Table 2). The re- sults with 100 “random” addition replicates were not different from those observed with the three other methods (Table 2). This nonrigorous test suggests that heuristic search combinations were effective in finding all of the most parsimonious cladograms. A quantitative estimate of data decisiveness (Golo- boff, 1991) is not possible for an analysis of 53 taxa which requires calculation of mean length for all possible cladograms. However, a generalized state- ment of data decisiveness suggests that data for this analysis are decisive. “Data are strongly decisive if one or more cladograms explaining them is very much shorter than others, and only weakly decisive if all possible cladograms are not very different for each other in length” (Forey et al., 1992). In total, only 128 unique cladograms trees were observed (Table 2), which represents a minute fraction of possible trees, and only a slightly larger fraction of those ac- tually examined during the search. Assuming that all trees of 403 steps were found, there must be many trees that have more than 403 steps. Although the frequency distribution of trees with number of steps cannot be determined, it is assumed that the num- ber of trees with more than 403 steps must increase dramatically given that the maximum number of steps is 1384. The majority-rule consensus of these trees (Fig. 1) illustrates clades observed in 50% or more of the 128 results. Clades found in less than 100% of the trees are indicated in parentheses as the percent of trees observed at this node. Only 5 of 47 resolved nodes were observed in less than 100% of the trees and only 2 were observed in less than 75% of the trees. Ex- amination of character distribution in the tree re- veals the homoplastic nature of many characters used in the analysis. This is reflected by the low consis- tency index (ci=0.33) observed in the 128 most par- simonious cladograms (Rohlf, 1982). Low consistency index is expected in studies of interrelationships for large numbers of taxa (Sanderson and Donoghue, 1989). The expected con- sistency index for a study of 53 taxa would be 0.14 with an equation of linear regression derived from 60 previous cladistic studies (Sanderson and Donoghue, 1989). This consistency index suggests that there is less homoplasy describing the interre- lationships of the Pleuronectidae than in other stud- ies of this size. The retention index (ri=0.79) indi- cates that homoplasy is occurring at terminal nodes and not internal nodes, which, in turn, suggests that relationships at the level of subfamily, tribe, and gen- era are not based purely on homoplastic morpholo- gies (Forey et al., 1992) and that the strength of this analysis can be used to determine relationships at this level. The retention index also suggests that this analysis is not effective at determining relationships near terminal ends, such as interrelationships among species within a genus. Consequently, the character analysis is restricted to the level of subfamily and tribe, and intragenera analysis will be explored only for relationships well corroborated by uniquely de- rived character states. Monophyly of the Pleuronectidae Ten synapomorphies define the Pleuronectidae. Dis- tributions of these states were also surveyed in the literature for taxa within the bothoid lineage (Hensley and Ahlstrom, 1984) and basal lineages within Pleuronectiformes (Chapleau, 1993). 1 Ocular-side frontal articulated with mesethmoid. Outgroup taxa with ocular-side prefrontal sepa- rating frontal from mesethmoid. Distribution of this structure within other pleuronectiform taxa reveals that the frontal on the ocular side is not articulated with the mesethmoid in Psettodidae, Citharidae, Paralichthyidae, Taeniopsettinae, and some genera of Bothinae ( Arnoglossus , Psettina, Asterorhombus, Japonolaeops, Laeops, Komoharaia, Neolaeops, and C hascanopsetta) and is observed to be in contact with the mesethmoid only in the Bothinae genera ( Parabothus , Cooper and Chap!eau: Monophyly and intrareiationships of the family Pleuronectidae 695 696 Fishery Bulletin 96(4), 1998 Engyprosopon, Tosarhombus, Crossorhombus, and Bothus [Amaoka, 1969]). 2 Ocular-side preorbital sensory canal absent, with exceptions only in Reinhardtius hippoglossoides and Acanthopsetta nadeshnyi. In the bothoid group, as well as in the Citharidae and Psetto- didae, this sensory canal is present (Amaoka, 1969). 3 Ventral margin of metapterygoid flattened (Fig. 2, B-D), with exception in Reinhardtius stomias, which has a distinct curvature along the ventral margin of this bone. Outgroup taxa also possess this ventral curvature of the metapterygoid (Fig. 2A). 4 First and second basibranchials loosely joined by cartilage with exceptions observed in Eopsetta grigorjewi, Isopsetta isolepis, Limanda ferruginea, L. punctatissima, Parophrys vetula, Psettichthys melanostictus, and Reinhardtius stomias in which basibranchials are sutured. Outgroup taxa also have a suture between first and second basibran- chials (Fig. 3A). 5 Second and third basibranchial loosely joined by cartilage (Fig. 3, B-D) with exceptions in Limanda punctatissima and Psettichthys melano- stictus. Outgroup taxa have a suture between sec- ond and third basibranchials (Fig. 3A). 6 Posteriormost abdominal vertebrae lack haema- pophysis (Fig. 40, with exceptions in Eopsetta grigorjewi , Microstomus achne, M. kitt, M. pacificus, and Reinhardtius stomias in which haemapophysis are present (Fig. 4D). Outgroups have fused parapophysis forming a haemal arch on the posteriormost abdominal vertebrae (Fig. 4, A and B). 7 Accessory processes on caudal vertebrae absent (Fig. 5, B and C), with exceptions in Hippoglossus hippoglossus, H. stenolepis, Microstomus bathy- bius,Pleuronichthys decurrens , P . guttulatus , and P. ritteri. Outgroups have accessory processes on ventral surface of centrum for all caudal verte- brae (Fig. 5A). 8 Ocular-side infraorbital bones present with ex- ception in Microstomus bathybius. Outgroup taxa, except Psettodes, do not have infraorbital bones on the ocular side. Psettodes has four infraorbital bones on the ocular side. Presence of infraorbital bones in Pleuronectidae, although not unique in comparison with all outgroup taxa, is likely a re- versal in the bothoid lineage and synapomorphic for the family. 9 Oil globules in egg absent, with exceptions of one oil globule found in Pleuronichthys cornutus, P. guttulatus , and P. ritteri (Ahlstrom et al., 1984). Outgroups have at least one oil globule (Ahlstrom et ah, 1984). Distribution of this character was not confirmed in the specimens used for this analysis; however the source for this information (Ahlstrom et al., 1984) confirms this distribution in 46 of the 59 pleuronectid species. 10 Olfactory laminae are parallel without a central rachis, with exception in Reinhardtius evermanni and R. stomias. As in the outgroup, species of Reinhardtius have laminae that radiate from a central rachis (Norman, 1934). Distribution of this structure was not confirmed in this analysis. His- torically, this structure has been used to diagnose the Pleuronectinae ( sensu Norman, 1934) and appears to be unique in flatfishes. There has not been any evidence to suggest that this occurs in any other flatfish species. Exceptions in distribution of these ten synap- omorphies do not have a common phylogenetic pat- tern and do not corroborate exclusion of any of the 53 species examined in this analysis. The exceptions suggest independent cases of reversed characters (reversals) for either the species or their immediate ancestors, and these will be discussed in the context of species interrelationships within the Pleuro- nectidae. It is noted that Reinhardtius stomias has four exceptions to the ten synapomorphies in Pleuronectidae. The most parsimonious explanation places this species in the Hippoglossinae on the ba- sis of seven additional synapomorphies. Other char- acters (48, 77, 83, and 104) appear to be synapo- morphies for the Pleuronectidae. However, distribu- tion of these characters was examined in only the 53 ingroup taxa and in the five outgroup taxa used for this analysis. This limited survey is not sufficient to provide a full understanding of the distribution of these character states within the order. Therefore these character states are not presented as synapo- morphies for the family and are only presented in a phylogenetic context within Pleuronectidae. Future analysis examining higher-level relationships within Pleuronectiformes should include these morphologi- cal characters. Intrarelationships of the Pleuronectidae The phylogenetic analysis reveals various monophyl- etic lineages within the Pleuronectidae, four of which illustrate the interrelationships of five newly defined subfamilies: Hippoglossinae, Eopsettinae, Lyopset- tinae, Hippoglossoidinae, and Pleuronectinae (Fig. 6). These subfamilies are separated by a gradation of characters (11 to 22) such that the first three, Hippoglossinae, Eopsettinae, and Lyopsettinae, con- tain species with a large proportion of plesiomorphic Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 697 698 Fishery Bulletin 96(4), 1998 D Figure 3 Dorsal aspect of branchial apparatus. (A) Paralichthys lethostigmus, ANSP 143209; (B) Lyopsetta exilis , NMC 60-0501; (C) Limanda ferruginea , NMC 80- 0217; (D) Pleuronectes putnami, ROM 556. bb=basibranchial; bh = basihyal; cb = ceratobranchial; eb = epibranchial; gr = gill raker; hp = hypobranchial. The fourth basibranchial is cartilagenous and illustrated in only Fig. 3A. Scale bars are 10 mm. character states. As a result, interrelationships within these groups are supported by few synapo- morphies. Serving as successive outgroups (Stiassny and de Pinna, 1994), the position of these three sub- families determines the polarity and relationships within the Hippoglossoidinae and the diversified Pleuronectinae. The first lineage (I) distinguishes Hippoglossinae from all other pleuronectid taxa. The second lineage (II) contains all species classified in Eopsettinae, Lyop- settinae, Hippoglossoidinae, and Pleuronectinae. This second lineage is supported by two synapomorphies (Fig. 6): ocular-side pterosphenoid and prootic join to form the dorsal margin of anterior prootic foramen (11, Fig. 7, B and C); and first epibranchial not bifurcated at its distal end (12, Fig. 3, C and D). Exceptions to the distribution of these two charac- ters are observed in Pleuronectinae and Verasper Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 699 (Hippoglossinae). The pterosphenoid and prootic do not unite to form the dorsal margin of the anterior prootic foramen (Fig. 7A) in the following pleuronectine spe- cies: Glyptocephalus cynoglossus, G. stelleri, G. zachirus, Limanda ferruginea, L. proboscidea, L. punctatissima, Platichthys stellatus, Pleuronichthys guttulatus, and Pseudopleuronectes americanus. Verasper variegatus, not classified in this second lineage, has the dorsal margin of this foramen formed by the pterosphenoid and prootic on both ocular and blind side. The first epibranchial is bifurcated in the pleuronectine species Dexistes rikuzenius, Limanda aspera, L. limanda , L. sakhalinensis, Glyptocephalus , Microstomus, and Pleuronichthys. Both species of Verasper were observed to have a bifurcated first epibranchial. The pattern of these exceptions is similar in these two characters, but the analysis did not indicate an alternative topology that would exclude any of the previously mentioned species, or include Verasper within the second lineage. The pattern does suggest that these exceptions are in- stances of reversal or convergence and may determine phylogeny within these other groups. The third pleuronectid lineage (III) indicates a com- mon ancestor for Lyopsettinae, Hippoglossoidinae, 700 Fishery Bulletin 96(4), 1998 and Pleuronectinae. Three synapomorphies defining this lineage (Fig. 6) are spines absent on gill rakers (13, Fig. 3, C and D); anterior margin of upper orbit complete with an overlap between mesethmoid and prefrontal of the blind side ( 14, Fig. 8, C and D); and first anal pterygiophore broadly thickened (15). Exceptions to the distribution in these synapo- morphies are found in Pleuronectinae and Hippo- glossinae. The anterior margin of the upper orbit is incomplete in Microstomus achne, M. bathybius, M. kitt, M. pacificus, Pleuronectes pinnifasciatus, Pseudopleuronectes americanus, P. herzensteini, and P. yokohamae. Reinhardtius stomias, which also has the derived state for this character, is excluded from this third lineage. The first anal pterygiophore is not thickened in Microstomus achne, M. bathybius, M. kitt , and M. pacificus , whereas in Hippoglossus and Verasper (Hippoglossinae) the first anal ptery- giophore is broadly thickened. The fourth lineage (IV) includes all species of Hippoglossoidinae and Pleuronectinae. The sister re- lationship between these two subfamilies is deter- mined by seven synapomorphies (Fig. 6): dentition of uniform size (16, Fig. 9, B-D); interorbital process reduced or completely absent (17, Fig. 10, B, D, and E); hyomandibula broadened anteriorly (18, Fig. 2D); dentition on third epibranchial absent (19, Fig. 3, C and D); bony plates absent on branchial arches (20, Fig. 3, C and D); two rows of teeth present on fourth ceratobranchial (21, Fig. 3, C and D); and dorso- posterior margin of operculum fimbriated (22, Fig. 2, B-D). Exceptions to these character distributions are found in only two species of Hippoglossoides and in Microstomini. Hippoglossoides platessoides and H. robustus have larger anterior teeth (16) that were historically termed as “canines” (Norman, 1934). Glyptocephalus kitaharai, G. zachirus, Microstomus achne, M. kitt, M. pacificus, and all species in Pleuronichthys have an interorbital process (17, Fig. 10C). The anterior margin of the hyomandibular (18) is not broadened in Glyptocephalus and Microstomus (Fig. 20. Dentition on the third epibranchial and bony plates on branchial arches are observed in Pleuronichthys guttulatus . The number of rows of teeth on the fourth ceratobranchial are reduced to only one in Glyptocephalus , Platichthys bicoloratus, and Pleuronichthys and are absent in Limanda punctatissima. Fimbriation of the operculum (22) is also observed in Flippoglossinae (Fig. 2B) but is ab- sent in the Eopsettinae and Lyopsettinae. This last exception suggests that fimbriation of the operculum may be synapomorphic for Pleuronectidae because only Eopsetta grigorjewi, E. jordani, and Lyopsetta Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 701 B Figure 7 Lateral aspect of crania. (A) Hippoglossus hippoglossus , ARC 8808487; (B) Pleuronichthys cornutus, UMMZ 159618; (C) Limanda limanda, MNHN 1959-560. Left is ocular side; right is blind side, bo = basioccipital; eo = exoccipital; epo = epiotic; frb = frontal (blind side); fro = frontal (ocular side); ic = intercalar; ip = interorbital process; me = mesethmoid; pa = parietal; pfb = prefrontal (blind side); pro = prootic; ps = parasphenoid; pt = pterotic; pts = pterosphenoid; pv = prevomer; sp = sphenotic. Scale bars are 10 mm. exilis do not show this condition. This alternative hypothesis would require three evolutionary steps, one more than is presently hypothesized. An equivo- cal alternative (two steps) would require a single reversal in character 22 to define a monophyletic group of Eopsetta and Lyopsetta. However, this to- pology was not observed in any of the 128 most par- simonious trees. Subfamily Hippoglossinae The first pleuronectid lineage is classified as Hippoglossinae, with eight species (6 examined) in four genera: Clidoderma ( incertae sedis), Hippoglossus, Reinhardtius, and Verasper (Fig. 11). This subfamily, as well as the intrarelationships of its species, were observed in all of the most parsimonious cladograms. The Hippo- glossinae are hypothesized to be monophyletic ac- 702 Fishery Bulletin 96(4), 1 998 Figure 8 Dorsal aspect of anterior margin of upper orbit, illustrating the structures for mesethmoid and suture between mesethmoid and prefrontal of the blindside. (A) Eopsetta grigorjewi , UMMZ 159590; (B ) Hippoglossoides elassodon, NMC 61-0117; (C ) Lepidopsetta mochigarei , UMMZ 159575; (D) Pleuronectes putnami, ROM 232214. me = mesethmoid; pfb = prefron- tal (blind side); pfo = prefrontal (ocular side). Scale bars are 2 mm. cording to three synapomorphies (Fig. 6): increase in number of abdominal vertebrae to more than 12 (23); lunate-shaped caudal fin (24); and presence of fimbriation on dorsoposterior opercular margin (22, Fig. 2B). These characters are found to have a high degree of homoplasy within the Pleuronectidae. The fimbriation pattern of the opercular margin (22 ) ap- pears in all other species of Pleuronectidae, except in Eopsetta and Lyopsetta that comprise the next two lineages. An increase in the number of abdominal vertebrae (23) is also observed in other subfamilies and the lunate-shaped caudal fin ( 24 ) is not observed in all members of this lineage; a reversal is hypoth- esized for Verasper. Monophyletic status of this clade is suspect. The low proportion of derived character states inherent for basal lineages provides little support for monophyly and intrarelationships for these taxa. Analysis of morphological characters that are homoplastic at the familial level, but not so at lower levels of universality, might clarify both the monophyletic status of this subfamily and its intrarelationships. Intrarelationships of Hippoglossinae Genus Reinhardtius This genus contains three species: R. hippoglossoides, R. evermanni (not exam- ined), and R. stomias , characterized by four synapomorphies (Fig. 11): gill rakers absent on fourth ceratobranchial (21); migrated eye is situated near dorsal midline of cranium, so that it is visible from blind side (25); more than 35 caudal vertebrae (26); and dentary foramen absent on blind side (27). Flomoplasy observed for these characters is ob- served in unrelated pleuronectid taxa and does not Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 703 corroborate an alternative hypothesis. The absence of gill rakers on fourth ceratobranchial is observed in Limanda punctatissima (Pleuronectinae). The position of the migrated eye in relation to the dorsal midline is a reversal also observed in Cleisthenes pinetorum (Hippoglossoidinae). An increase in the number of cau- dal vertebrae is also observed in Glyptocephalus and Microstomus, a clade of nine species whose monophyl- etic status is strongly supported within the Pleuro- nectinae. Absence of the dentary foramen on the blind side is only observed in four other pleuronectid species including Hippoglossus kippoglossus. Reinhardtius euermanni was not examined, but morphological characters reported in the literature suggest monophyly for R. evermanni, R. hippo- glossoides, and R. stomias. Reinhardtius evermanni also has an increased number of caudal vertebrae, ranging from 37 to 40 (Sakamoto, 1984a). Rein- hardtius evermanni and R. stomias share two unique structures: olfactory lamellae that radiate from a central rachis; and jaw and pharyngeal teeth with barbed tips (Norman, 1934). If these structures are considered synapomorphic within the genus, then Reinhardtius evermanni and R. stomias are sister species, with R. hippoglossoides immediately basal to them. The sister relationship between Hippoglossus and Verasper is supported by six synapomorphies (Fig. 11): 704 Fishery Bulletin 96(4), 1998 first anal pterygiophore broadly thickened (15); ocu- lar-side palatine reduced and not articulated with pterygoid (28, Fig. 2B); gill rakers on first epi- branchial absent (29) with exception in Hippoglossus stenolepis; gill rakers on second and third epi- branchials (30, 31) reduced ( Hippoglossus ) or absent (Verasper ); and gill rakers on first hypobranchial re- duced (32). These states are homoplastic in other lineages of the Pleuronectidae. Thickening of the first anal pterygiophore (15) unites taxa in the third pleuro- nectid lineage. Reduction of the palatine on the ocu- Cooper and Chapfeau: Monophyly and intrarelationships of the family Pfeuronectidae 705 lar side (28, Fig. 2, B and C) also defines the lineage uniting Glyptocephalus, Microstomas, and Pleuro- nichthys within the Pleuronectinae. Absence or re- duction of gill rakers on first, second, and third epibranchials (29, 30, 31) is found to define basal lin- eages within the Pleuronectinae, whereas reduction of gill rakers on the first hypobranchial is observed in only some species of Pleuronichthys and Platichthys bicoloratus. Genus Hippoglossus This genus contains two species: II. hippoglossus and H. stenolepis and is monophyletic with three synapomorphies (Fig. 11): presence of subdivisions in hypural and parahypural plates, an autapomorphy for the genus (33); pres- ence of accessory processes on ventral margin of cau- dal vertebrae (7, Fig. 5A), a reversal for the family; and metapterygoid articulated with the blind-side entopterygoid (34, Fig. 2C). Only Microstomas bathybius and three species of Pleuronichthys have accessory processes on the caudal vertebrae. The metapterygoid is also articulated with the entop- terygoid of the blind side in Reinhardtius hippo- glossoides and species within Pleuronectinae. Genus Verasper This genus, containing V. mos- eri and V. variegatus is monophyletic with nine synapomorphies (Fig. 11): presence of a large fora- men between mesethmoid and blind-side prefrontal (35), autapomorphic for Verasper ; first epibranchial bifurcated (12); caudal fin is rounded (24), a reversal in this subfamily; gill rakers absent on both second and third epibranchials (30, 31); gill rakers reduced on second hypobranchial (36); sphenotic process not forming dorsal margin of hyomandibular socket (37, Fig. 7, B and C); groove present along supraoccipital crest for insertion of pterygiophores (38, Fig. 10, C- E); and cardiac apophysis of urohyal bifurcated (39, Fig. 120. Many character states found within Verasper are also observed in the Pleuronectinae, but the strength of the hypothesis placing Verasper within Hippo- glossinae exceeds the characters mentioned above and illustrates the convergent evolution of these structures. Genus Clidoderma C. asperrimum was not avail- able for analysis. This species has modified scales on the ocular side to form distinct bony tubercles very similar to those observed in Platichthys . However, this unique species appears to be more closely re- lated to Verasper than to Platich thys (Norman, 1934). This species has subsymmetrical jaws and a mix of pointed and bluntly conical teeth, not uniform in length, set in multiple rows on both upper and lower jaws. These features are plesiomorphic for the fam- 706 Fishery Bulletin 96(4), 1998 ily and only exclude Clidoderma from the fourth pleuronectid lineage. This species has 14 abdominal vertebrae found to be synapomorphic for Hippo- glossinae; it also has a thickened first anal pterygiophore suggesting a common ancestry with Hippoglossus and Verasper. A rounded caudal fin, observed in Verasper is also observed in Clidoderma. Phylogenetic position of this species could be further ascertained by an examination of the accessory pro- cesses on caudal vertebrae, the palatine structure, and the structure of gill rakers on first, second, and third epibranchials. Subfamily Eopsettinae The Eopsettinae consists of E. grigorjewi and E. jordani. This analysis defines this subfamily with two synapomorphies (Fig. 6): presence of gill rakers on fourth epibranchial (40); and a single row of teeth on lower jaw (41). Gill rak- ers on the fourth epibranchial are observed in Cleisthenes herzensteini, C. pinetorum (Hippo- glossoidinae), and Psettichthys melanostictus (Psettichthyini). The number of rows of teeth on the lower jaw were found to be much more homoplastic. The latter feature was also observed in Reinhardtius hippoglossoides , Verasper moseri, and V. variegatus (Hippoglossinae) as well as in many species of the Pleuronectinae. Genus Eopsetta This genus was described in Norman (1934) by a number of plesiomorphic char- acters. The presence of distinct canines on the upper jaw was suggested in Norman (1934) as diagnostic for Eopsetta. However, distribution of this character in the Pleuronectidae is not well defined. Members of the Hippoglossinae also have teeth of irregular lengths (16, Fig. 9A). The longer teeth can also be interpreted as canines in these species. From this analysis, data supporting monophyly of Eopsetta are not conclusive, but no other interpretation is avail- able owing to insufficient information. Subfamily Lyopsettinae Genus Lyopsetta This lineage contains only L. exilis. Its position as a monotypic lineage within the Pleuronectidae is determined by five derived char- acter states (Fig. 6): 12 to 14 abdominal vertebrae (23); barbed teeth present on dentaries and premax- illae (42); supratemporals on both ocular and blind sides are jointed at anterior ends of their bifurcation (43, 44, Fig. 13B); and presence of scales on eye sur- faces (45). These structures are also distributed within other pleuronectid taxa. An increase in abdominal verte- brae is found in Hippoglossinae, some species of Hippoglossoidinae, and in two separate lineages of Pleuronectinae. Barbed teeth are also observed in Reinhardtius. Bifurcation of the supratemporals is also observed in Reinhardtius hippoglossoides, Cleisthenes , Hippoglossoides , Glyptocephalus cynoglossus, G. stelleri, G. zachirus, Microstomus pacificus , Limanda aspera, and Pleuronectes platessus. Presence of scales on the eye surfaces is also found in Reinhardtius stomias, Acanthopsetta nadeshnyi, Dexistes rikuzenius, Glyptocephalus kitaharai, and Microstomus. The distribution of these character states within the family does not indicate an alternative hypothesis of relationships between Lyopsetta exilis and other pleuronectid taxa. Subfamily Hippoglossoidinae This subfamily con- tains seven species (6 examined) in three genera: Acanthopsetta, Cleisthenes, and Hippoglossoides. This group is characterized by four synapomorphies (Fig. 6): absence of supraoccipital plate extending posteroventrally between epiotics (46, Fig. 10B); pterosphenoid and prootic of blind side join to form dorsal margin of anterior prootic foramen (47, Fig. 7, B and C); pterosphenoid of blind side is reduced and does not form posterior margin of orbit (48, Fig. 7B), a reversal, except in Cleisthenes; and two uniform rows of teeth present on fifth ceratobranchial (49). Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 707 Distribution of these states within other pleuro- nectid taxa indicates a degree of homoplasy for these characters but does not refute the monophyly of Hippoglossoidinae. The supraoccipital plate is also absent in Microstom us (Microstomini). Pterosphenoid and prootic of the blind side are also united to form the dorsal margin of the anterior prootic foramen in Verasper variegatus and in most species of the Pleuronectinae. Reduction of the blind-side ptero- sphenoid is also observed in Pleuronichthys (Micro- stomini) and may be a reversal within the Pleuro- nectidae. Two rows of uniform teeth are also present in Dexistes rikuzenius (Microstomini) and the Pleuronectini. Intrarelationships of Hippoglossoidinae Genus Acanthopsetta A. nadeshnyi is the sister species to Cleisthenes and Hippoglossoides (Fig. 14). Placement of A. nadeshnyi as a distinct lineage within Hippoglossoidinae is supported by two struc- tures not present in other Hippoglossoidinae: sen- sory canal on ocular-side preorbital present (2); and presence of scales on eye surfaces (45). The presence of a sensory canal on the ocular-side preorbital is a reversal for the family that also occurs in Rein- hardtius hippoglossoides . Scales present on the sur- face of each eye have a more homoplastic distribu- tion as they are found in Reinhardtius stomias, Lyopsetta exilis, Dexistes rikuzenius, Glyptocephalus kitaharai, and Microstomus. The sister relationship between Cleisthenes and Hippoglossoides is supported by three synapo- morphies (Fig. 14): ocular-side supratemporal jointed at anterior end of bifurcation (43, Fig. 13B); nasal bones of the blind side are absent ( 50 ); and more than seven infraorbital bones on ocular side (51). These morphological states are not unique within the family. The bifurcation of the supratemporal is observed in Lyopsetta exilis on both ocular and blind side, but not all species of Cleisthenes and Hippo- glossoides have the supratemporal bifurcation on the blind side. The absence of nasal bones on the blind side and number of infraorbitals are shared with some species in the Pleuronectinae, the latter being also observed in Hippoglossus stenolepis. Genus Cleisthenes This genus contains two spe- cies, C. herzensteini and C. pinetorum, and is diag- nosed by four synapomorphies (Fig. 14): migrated eye is near dorsal midline (25); gill rakers on fourth epibranchial present (40); a double crest or groove present on supraoccipital and blind-side frontal (38, Fig. 10, C-E); and crest extending from supraoccipi- tal to blind side reduced (52, Fig. 10, D and E). These morphological characters are distributed within other pleuronectid taxa but are not observed in any other Hippoglossoidinae. The position of the migrated eye is also observed in Reinhardtius . Pres- ence of gill rakers on the fourth epibranchial has a limited distribution in Eopsetta grigorjewi, E. jordani, and Psettichthys melanostictus. The double crest on the supraoccipital and blind-side frontal is recurrent throughout the family but unique within Hippoglossoidinae. The reduced crest on the blind- side frontal is also observed in Reinhardtius stomias, Pleuronichthys verticalis, and Pleuronectini. Genus Hippoglossoides This genus contains 4 species: H. dubius (not examined), H. elassodon, H. platessoides, and H. robustus, defined by two synapomorphies (Fig. 14). The structure of the ante- rior margin of the mesethmoid is complex but con- sistent in Hippoglossoides . In this genus, the “thin plate” structure of the mesethmoid (53, Fig. 8B) is distinct in relation to other members of Hippo- glossoidinae who have a thickened triangular-shaped mesethmoid (Fig. 8C), or the plesiomorphic structure of an open canal found in Cleisthenes pinetorum (Fig. 8A). In addition, there are 12 to 14 abdominal verte- brae (23), an increase from 11 or fewer in other taxa. The interrelationships within Hippoglossoides are not fully resolved by this analysis. Hippoglossoides dubius has 13 abdominal vertebrae (Norman, 1934; Sakamoto, 1984a) and is assumed to have common ancestry with other species of Hippoglossoides . Subfamily Pleuronectinae The Pleuronectinae is the largest subfamily within the Pleuronectidae with 40 species (38 examined in 708 Fishery Bulletin 96(4), 1998 this analysis). Seven synapomorphies define this group as monophyletic (Fig. 6): absence of a dentary fossa (55, Fig. 9, C and D); and absence ofceratohyal foramen (56) are both autapomorphic for Pleuro- nectinae; mesethmoid and blind-side prefrontal are sutured but without a foramen between these bones (14, Fig. 8D); a double supraoccipital crest forms a groove for insertion of anterior dorsal-fin ptery- giophores (38, Fig. 10, C-E); a single row of teeth on lower jaw (41); intercalar in contact with basioccipi- tal (57, Fig. 70; and presence of a posterior exten- sion of supratemporal branch of the lateral line (58). Some of these character states are not found in all pleuronectines, but the occurrence of these states at basal lineages and their predominance within the Pleuronectinae indicates that the absence of these synapomorphies within the subfamily are instances of evolutionary reversal. For example, the presence of a supratemporal branch (58) is observed in 11 pleuronectine taxa and is hypothesized to arise at this node with two secondary losses observed in more ad- vanced lineages within Microstomini and Pleuronectini. Intrarelationships of Pleuronectinae The Pleuro- nectinae is classified into 4 tribes: Psettichthyini, Isopsettini, Microstomini, and Pleuronectini (Fig. 6). Branchial structure and characters associated with jaw asymmetry determine the interrelationships of these tribes. Tribe Psettichthyini Genus Psettichthys Mono- typic with only P. melanostictus , this lineage is unique within the Pleuronectinae, having six distinct char- acters (Fig. 6): dorsal fin rays are elongated beyond dorsal-fin membrane, an autapomorphy for the spe- cies (59); second and third basibranchials are sutured (5, Fig. 3A), a reversal within Pleuronectidae; teeth are not uniform in length (16); greater than seven infraorbital bones (51); gill rakers on fourth epibranchial present (40); and one row of teeth on upper jaw (60). These morphological states are shared with taxa both within and prior to the Pleuronectinae. This evi- dence clearly positions Psettichthyini as a basal tribe of the Pleuronectinae. A suture between the second and third basibranchials is observed elsewhere only in Limanda punctatissima. Tooth length is uniform in other pleuronectine taxa. An increase in infraorbital number is observed in two other pleuronectine species, Lepidopsetta bilineata and Pleuronichthys decurrens, as well as in Cleisthenes, Hippoglossoides , and Hippoglossus stenolepis. The presence of gill rakers on the fourth epibranchial is found elsewhere only in Cleisthenes and Eopsetta. A single row of teeth on the upper jaw is observed in all species of the Pleuronectini, and in Glyptoeephalus and Microstomus. The second lineage contains the newly defined tribes Isopsettini, Microstomini, and Pleuronectini. Taxa within this lineage are characterized by 11 synapomorphies (Fig. 6): one gill raker at proximal base of second and third epibranchials (30, 31, Fig. 3D); blind-side premaxilla protruding past the sag- ittal axis at its symphysis with that of ocular side (61, Fig. 9, C and D); ocular-side premaxilla much longer than that of blind side (62, Fig. 9, C and D); ventral posterior curvature on blind-side premaxilla is present (63, Fig. 9, C and D); asymmetry in space between dentary and articular such that blind-side space is larger than on ocular side (64, Fig. 9, C and D); dorsoposterior process of ocular-side dentary larger than its blind-side counterpart (65, Fig. 9C); teeth on both ocular-side premaxilla and dentary reduced (66, 67, Fig. 9, C and D); epiotic processes present (68, Fig. 10, D and E); and ocular-side entopterygoid larger than that of blind side (69, Fig. 2, C and D). Distribution of these character states is not without exceptions or homoplasies. Reduction of gill rakers on the second epibranchial was not observed in Limanda (Fig. 30, and a reduction of gill rakers on the third epibranchial was not observed for Pleuronectes quadrituberculatus . These reductions are homoplastic in Hippoglossus and Verasper (Hippoglossinae). The third lineage indicating a sister relationship between the tribes Microstomini and Pleuronectini, is based on six synapomorphies (Fig. 6): within this lineage there is an evolution of dentition, from pointed or bluntly conical teeth to incisorlike or even molariform teeth with uniform cutting edges (70); sphenotic process positioned high on sphenotic (37, Fig. 7, B and C); urohyal with strongly bifurcate car- diac apophysis (39, Fig. 12, D-F); blind-side pterosphenoid and prootic form dorsal margin of anterior prootic foramen (47, Fig. 7, B and C); me- dial margin of fifth ceratobranchial slightly curved (71, Fig. 3C); and teeth on fifth ceratobranchial bluntly pointed (72, Fig. 30. Exceptions to the distribution of character states within the third lineage are observed in few species and appear to be cases of reversal. They do not con- tradict the sister relationship between Microstomini and Pleuronectini. The sphenotic process is not posi- tioned high on the sphenotic in Glyptoeephalus . A strongly bifurcated cardiac apophysis on the urohyal was not found in Limanda punctatissima, Pleuro- nichthys ritteri, P. ocellatus, and Parophrys vetula. These two morphological characters are homoplas- tic in Verasper (Hippoglossinae). The sphenotic forms the dorsal margin of the anterior prootic foramen on the blind side (47, Fig. 7A) in Limanda ferruginea, L. proboscidea, Microstomus achne, M. kitt, Glypto- Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 709 cephalus cynoglossus, G. stelleri, G. zachirus, and Pleuronichthys guttulatus. However, the blind-side pterosphenoid and prootic uniting to form the dorsal margin of the anterior prootic foramen is homoplas- tic in Hippoglossoidinae and Verasper variegatus (Hippoglossinae). The medial curvature of the fifth ceratobranchial is not observed in Microstomus, Glyptocephalus, and Pleuronichthys, and teeth on the fifth ceratobranchial are sharply pointed in Micros- tomus bathybius, Pleuronichthys , and Lepidopsetta. Tribe Isopsettini Genus Isopsetta The tribe Isopsettini is monotypic with Isopsetta isolepis. Four character transformations identify the lineage (Fig. 6): blind-side metapterygoid is not articulated with entopterygoid (34, Fig. 2A); anterior margin of mesethmoid forming an open canal (53, Fig. 8A); haemal spines of anteriormost caudal vertebrae broadly attached to centrum (73, Fig. 5B); and epiotics are sutured along dorsal posterior margin of skull (74). Three of these four character states are reversals within the Pleuronectinae. The absence of an articu- lation between the blind-side metapterygoid and entopterygoid is a reversal of the structure observed in the Psettichthyini and in most species of Microstomini and Pleuronectini, except Lepidopsetta, Pleuronichthys guttulatus, Limanda aspera, and L. ferruginea. The open canal on the anterior margin of the mesethmoid is a reversal that is also observed in the pleuronectines, Glyptocephalus, Microstomus, and Pleuronichthys (except P. verticalis), and in Cleisthenes pinetorum (Hippoglossoidinae). Epiotics sutured along the dorsal posterior margin are ob- served only in one other pleuronectid species, Mi- crostomus pacificus , and the broad attachment of haemal spines to the anteriormost caudal vertebrae is a reversal recurrent throughout the family, indi- cating a homoplastic structure with a complex dis- tribution. Despite the reversals noted for this lin- eage, placement of Isopsettini within Pleuronectinae is supported by the eight synapomorphies for Pleuro- nectinae and the 11 derived morphological charac- ters for the second lineage in Pleuronectinae. Tribe Microstomini The tribe Microstomini con- tains 19 species (17 examined) classified in five gen- era: Lepidopsetta, Dexistes, Pleuronichthys, Glypto- cephalus, and Microstomus (Fig. 15). Although the placement of this tribe within the Pleuronectinae is supported by 25 character transformations presented for monophyly and intrarelationships of the subfam- ily, the status of this tribe, as well as its intra- relationships, are determined mostly by character reversals. The monophyly of this tribe is character- ized by four character transformations. All are re- versals within the Pleuronectinae (Fig. 6): suture between mesethmoid and blind-side prefrontal is ei- ther incomplete or complete with a small foramen present (14, Fig. 8, A-C); single crest on supraoccipi- tal (38, Fig. 10, A and B); lower jaw with multiple rows of teeth (41); and reduced or absent process on dorsoposterior edge of epiotics (68, Fig. 10, A-C). The few exceptions to the distribution of these states within Microstomini and the occurrence of these same reversals outside of Microstomini indi- cate the homoplastic nature of these structures. Multiple rows of teeth on the lower jaw were not ob- served in Microstomus and Glyptocephalus, which is homoplastic within Hippoglossinae and recurrent for basal lineages within the family but which has only this one instance of reversal in Pleuronectinae. The absence or reduction of an epiotic crest is a reversal also found in Pseudopleuronectes herzensteini and P. yokohamae . Intrarelationships of Microstomini Three lineages of Microstomini are defined by 16 character trans- formations; ten are reversals (Fig. 15). Genus Lepidopsetta The first lineage of Micro- stomini, consists of Lepidopsetta containing two spe- cies, L. bilineata and L. mochigarei (Fig. 15). This ge- nus is diagnosed by the presence of demersal eggs (75), a feature observed in only four other pleuronectid spe- cies: Pseudopleuronectes americanus, P. schrenki, P. yokohamae , and P. obscurus (not examined). The second lineage indicates common descent for Dexistes, Pleuronichthys, Microstomus, and Glypto- cephalus. Four character states unite these genera (Fig. 15): first epibranchial bifurcated at distal end (12, Fig. 3A); cardiac apophysis simple at tip with a bifurcation positioned anteriorly (39, Fig. 12, E and F), except in Pleuronichthys ocellatus and P. ritteri; intercalar not in contact with basioccipital (57, Fig. 7B), a reversal in Pleuronectinae; and less than seven teeth on ocular-side premaxilla (66). Exceptions to character states within the second lineage of Microstomini were not congruent and did not provide alternative topologies that either ex- cluded taxa observed to be exceptions to this distri- bution or included taxa that were homoplastic with these structures. Bifurcation of the first epibranchial is a reversal in the Pleuronectidae. Shape of the car- diac apophysis of the urohyal was observed only in one other pleuronectid, Limanda aspera. The intercalar is in contact with the basioccipital in Glyptocephalus cynoglossus (57, Fig. 7C). The reduc- tion of teeth on the ocular-side premaxilla (less than 7) is not observed in Pleuronichthys guttulatus and Glyptocephalus. In Glyptocephalus , variation in tooth number on the ocular-side premaxilla ranged from 8 to 16 in G. zachirus, but in all species there is a 710 Fishery Bulletin 96(4), 1998 smaller number of teeth on the ocular-side premax- illa than the number observed in previous lineages of Pleuronectinae (Norman, 1934). Glyptocephalus cynoglossus is reported to have 8 to 15 teeth on the ocular-side premaxilla and 17 to 26 on the associ- ated blind-side premaxilla. G. stelleri has only seven teeth on the ocular-side premaxilla and 20 to 27 teeth on the blind side. G. tanakius has 12 to 14 teeth on the ocular-side premaxilla and 14 to 16 on the blind side, and G. zachirus has 12 to 16 teeth on the ocu- lar-side premaxilla and 20 to 27 teeth on the blind side (Norman, 1934). The third lineage of Microstomini indicates mono- phyly for Pleuronichthys, Microstomus, and Glypto- cephalus. It is based on ten character transforma- tions, four synapomorphies, and six reversals (Fig. 15). Synapomorphies include increased number of abdominal vertebrae to between 12 and 14 (23); lips thickened or fleshy (76); palatine of ocular side re- duced and not attached to pterygoid (28, Fig. 20; and teeth on ocular-side dentary reduced to fewer than six (67). The reversals are exoccipital and prootic in contact with each other (77, Fig. 7B); presence of an interorbital process (17, Fig. IOC); one row of gill rakers on fourth ceratobranchial (21, Fig. 3, A and B); haemal spine broadly attached to centrum (73, Fig. 5B); space between both ocular- and blind-side dentary and articular equal in size (64, Fig. 9D); and dorsoposterior process of similar size on ocular- and blind-side dentaries (65, Fig. 9D). Exceptions and homoplasy in this distribution do not support an alternative hypothesis. Glypto- cephalus kitaharai has only 11 abdominal vertebrae (Sakamoto, 1984a). Reduction of the ocular-side pa- latine was also observed in Hippoglossus and Verasper (Hippoglossinae). The exoccipital and prootic are not joined, and secondary reduction of teeth on the dentary of the ocular side are not ob- served in Glyptocephalus . These species all have pro- portionally fewer teeth on the ocular-side dentary, but the exact number ranges from 11 to 15 in Glyptocephalus kitaharai to 10 to 18 in G. zachirus (Norman, 1934). Reversals in this lineage reproduce plesiomorphic states observed prior to the fourth lineage of Pleuronectidae or states that are plesiomorphic within the Pleuronectinae. The contact between exoccipital and prootic, may be a reversal for the fam- ily. The lost interorbital process in the fourth pleuronectid lineage is observed (at least partially) in these two pleuronectine genera. The single row of gill rakers on the fourth ceratobranchial is a rever- sal of the two rows that define the fourth pleuronectid lineage. The broad attachment of the haemal spine to the centrum is a reversal of a narrower attach- ment defining the fourth pleuronectid lineage. Sym- Cooper and Chapleau: Monophyly and intrarelationships of the family Pleuronectidae 71 I metry in the space between dentary and articular, as well as the dentary process, are reversals for Pleuronectinae. These reversals are also observed in some species of the Pleuronectini (Table 1). The fourth lineage in Microstomini represents the sister relationship between Microstomus and Glyptocephalus. Eight synapomorphies support this hypothesis (Fig. 15): intestine extending posteriorly into body cavity (87), unique for this lineage; more than 20 caudal-fin rays, an increase from 18 or 19 (88); more than six branched caudal-fin rays (89); 36 to 41 caudal vertebrae, an increase from 25 to 35 (26); less than five infraorbital bones, a reduction from between five and seven (51); teeth incisorlike (70); single row of teeth on upper jaw (60); and hyomandibula without a broad anteroventral margin (18, Fig. 20. Exceptions to the distribution of these synapo- morphies do not indicate exclusion of any species defined in Glyptocephalus or Microstomus, nor does homoplasy suggest the inclusion of additional taxa. An increase in caudal-fin rays is only observed in one other pleuronectid, Pleuronectes platessus, which usually has 20 caudal-fin rays. An increase in the relative number of branched caudal-fin rays is also observed in Pleuronichthys decurrens , P. ritteri , and Pseudopleuronectes yokohamae . An increase in the number of caudal vertebrae is homoplastic in Reinhardtius (Hippoglossinae). A reduction in the number of infraorbital bones is not observed in Glyptocephalus kitaharai and Microstomus pacificus but is reduced in Pseudopleuronectes americanus. The incisorlike tooth structure is unique within Microstomini but is observed in Parophrys vetula, Pleuronectes, and Pseudopleuronectes . A single row of teeth on the upper jaw is also unique within Microstomini but is also observed in Psettichthys melanostictus (Psettichthyini) and Pleuronectini. The absence of a broadened hyomandibula is a reversal unique within this lineage of Glyptocephalus and Microstomus . All other taxa within the fourth lin- eage of Pleuronectidae have a hyomandibular with a broadened anterior margin (Fig. 2D). Genus Dexistes The monotypic genus contains D. rikuzenius (Fig. 15). The monotypic status of this species is based on the morphologlical characters examined for intrarelationships of Microstomini and the presence of an ocular-side postocular ridge (78). The latter character has an additional evolutionary step within the sister tribe Pleuronectini, uniting a clade comprising Limanda, Platichthys, Pleuronectes, and Pseudopleuronectes. Genus Pleuronichthys This genus contains seven species: Pleuronichthys coenosus (not examined), P. cornutus, P. decurrens, P. guttulatus, P. ritteri, P. ocellatus, and P. verticalis. This genus is identified by eight character transformations (Fig. 15): pres- ence of villiform teeth, autapomorphic for Pleuro- nichthys (79); large foramen in blind-side prefrontal (80); gill rakers on second hypobranchial reduced to one at proximal base (36); lateral process present on ocular-side frontal (81, Fig. 7B, IOC); 25 or less cau- dal vertebrae (82); ocular-side pterosphenoid reduced and not forming posterior margin of orbit (83, Fig. 7B); blind-side pterosphenoid is similarily reduced (48, Fig. 7B); and ocular-side entopterygoid similar in size to that of blind side (69, Fig. 2, A and B). Pleuronichthys coenosus is assumed to be a member of this genus on the basis of presence of villiform teeth, reduced number of caudal vertebrae (24 to 25), and presence of a lateral process on ocular-side fron- tal (Sakamoto, 1984a). There are few exceptions to the distribution of char- acter states within Pleuronichthys, and instances of homoplasy do not indicate an alternative hypothesis. The large foramen in the blind-side prefrontal was observed in only one other species, Glyptocephalus stelleri. Reduction of gill rakers on the second hypo- branchial was not observed in Pleuronichthys ocellatus, which has at least two gill rakers on the second hypobranchial. The presence of a lateral pro- cess on the ocular-side frontal is also observed in Microstomus achne, M. kitt, and M. pacificus. How- ever, it is not nearly as distinct as that in Pleuro- nichthys. Only Platichthys flesus, P. stellatus, and Pleuronectes putnami (Pleuronectini) have also fewer than 25 caudal vertebrae. Reduction of the ocular- side pterosphenoid is not observed in Pleuronichthys ritteri, where the pterosphenoid separates frontal and parasphenoid to form the posterior margin of the orbit. Reduction of the blind-side pterosphenoid is also observed in Acanthopsetta and Hippoglossoides (Hippoglossoidinae). These two character states are reversals in the Pleuronectidae. The symmetry be- tween ocular-side and blind-side entopterygoids is a reversal of the asymmetrical structure observed prior to the Isopsettini. This reversal only occurs in Pleuronichthys and Glyptocephalus kitaharai . The intrarelationships of Pleuronichthys are not fully resolved. Pleuronichthys guttulatus is the sister spe- cies to all other species of Pleuronichthys (Fig. 15). Four synapomorphies unite P. cornutus, P. coenosus, P. decurrens, P. ocellatus, P. ritteri, and P. verticalis : dor- sal fin originates on blind side of head (84); reduced or absent cartilaginous interspace between blind-side pre- frontal and parasphenoid (85, Fig. 7B); mesethmoid forms only part of anterior margin of upper orbit (86); and ocular-side metapterygoid articulated with entopterygoid (54). This last feature is homoplastic because it is also observed in Reinhardtius hippo- glossoides and Limanda punctatissima. 712 Fishery Bulletin 96(4), 1 998 Genus Microstomas This genus contains five species: M. achne, M. bathybius, M kitt ., M . pad ficus , and M. shuntovi (not examined) and is defined by three synapomorphies (Fig. 15): posterior extension of supraoccipital absent (46, Fig. 10B); first anal pterygiophore thin ( 15); teeth in both upper and lower jaws uniform in length forming a continuous cutting edge (16, Fig. 9D). Descriptions of Microstomus shuntovi indicate that this species is very similar to M. kitt (Borets, 1983). The teeth are described as “chisel-like” (Borets, 1983), a term that similarly describes teeth in other species of Microstomus. An examination of the supraoccipital and the first anal pterygiophore would verify this classification. Absence of the posterior extension on the supraoc- cipital has evolved independently in the Hippo- glossoidinae. The thin structure of the first anal pterygiophore is a reversal of a thickened structure defining the third lineage. The continuous cutting edge of the teeth is an advanced state, also present in Glyptocephalus zachirus, Pleuronectes glacialis, P . pinnifasciatus , P . putnami , and Pseudopleuronectes schrenki. Microstomus bathybius is the sister species to M. achne, M. kitt , and M. pacificus (Fig. 15). This deep- sea species is unique in being the only pleuronectid with a series of infraorbital bones on the ocular side (8). It is the only species of Microstomus in which the blind-side nasal bone is absent (50) and is the only pleuronectid in which the anteroventral tip of the pelvis is anterior to the cleithrum (90). The other species of Microstom us share four character states (Fig. 15): ocular-side scales with radii completely surrounding the focus (91); margins of intero- perculum and suboperculum fimbriated (92, 93, Fig. 2C); and haemapophysis present on most posterior abdominal vertebrae (6, Fig. 4D). This last feature is a reversal for Pleuronectidae, and only Glypto- cephalus zachirus has also a fimbriated subopercular margin. Genus Glyptocephalus This genus contains four species: G. cynoglossus, G. kitaharai, G. stelleri, and G. zachirus. Four synapomorphies are hypothesized at this node (Fig. 15): more than 21 caudal-fin rays (88); cleithra inserted by tip of urohyal (94); pres- ence of two to four pyloric appendages and two or three on upper intestine, an increase from two to three and one on upper intestine (95); sphenotic pro- cess positioned low on sphenotic and forming dorsal roof of hyomandibular socket (37, Fig. 7A). This last character is a reversal of that observed in all other Microstomini and Pleuronectini. Glyptocephalus kitaharai is the sister species to G. cynoglossus, G. stelleri, and G. zachirus. It is unique within this clade and with all other pleuro- nectids in having 23 caudal-fin rays, the highest number observed in the family (88). The other spe- cies are united by three character states (Fig. 15): blind-side nasal bones larger than those of ocular side (96); development of large mucous cavities on blind- side head (97), both characters unique for these three species; and presence of an interpterosphenoid bar (98, Fig. 16), not observed in any other species of Glyptocephalus or immediate lineages, but homoplas- tic in Hippoglossoides, Acanthopsetta, Psettichthys, Limanda aspera, L. ferruginea, L. proboscidea, L. punctatissima, Pseudopleuronectes americanus, P. yokohamae, and Pleuronectes. The monophyletic sta- tus of these three species is in agreement with rela- tionships inferred through an analysis of body shape (Chiu, 1990). Tribe Pleuronectini The tribe Pleuronectini con- tains 20 species (19 examined). Four synapomorphies define this tribe (Fig. 6): at least two regular rows of teeth on fifth ceratobranchial (49, Fig. 3, C and D); postocular ridge present on ocular side (78); upper jaw teeth in single row (60); and dorsal crest extend- ing anteriorly from supraoccipital to blind-side fron- tal reduced or absent (52, Fig. 10, D and E). Exceptions and homoplasy did not indicate an al- ternative hypothesis for Pleuronectini; all 19 species examined were grouped in all 128 of the equally par- fro i j { \ , frt> i \ ^ iPts bar pts /t vffl Ip #12% were again read in- dependently by each reader without knowledge of previous annulus counts. The count with the great- est difference from the mean of all counts was then discarded and replaced with a new count made by the reader whose reading was discarded. This proto- col was repeated twice, and if the CV for a given otolith remained >12%, the otolith was rejected from our analysis. The von Bertalanffy (1957) growth equation TLf = LJ \-e(~K,t~to))) was fitted to observed age-length data with nonlinear regression procedures. Age was esti- mated as the mean of all six annulus counts. We did not attempt to backcalculate the length at which the last annulus was formed because of the variability among annulus counts for most otoliths. Our esti- mates of length at age thus include some seasonal growth that occurred after the formation of the final annulus. To estimate mean length at age, we rounded all ages to the nearest integer, but in growth and maturity models the mean age was not rounded to an integer and fractional ages were used. Length- weight regressions were calculated by linear regres- sion of log10-transformed data. Age validation Measurements for marginal-increment analysis were made with a digital image-processing system on an axis extending along the sulcal ridge from the otolith’s core to the proximal margin of the section. Because many black grouper otoliths were difficult to read, we selected only otoliths for which there was 100% agreement among all readings for marginal- increment analysis. In addition, we restricted mea- surements to otoliths with from one to seven annuli because it was difficult to measure the more closely spaced outer annuli of older fish. We expressed the distance from the final annulus to the otolith’s edge (marginal increment) as a percentage of the distance between the last two annuli formed on the otolith. For all grouper, the distance between the otolith core and the first annulus (r x ) was typically much greater than the distance between the first and second an- nuli (r2-r1). For this reason, we divided the distance between the first and second annuli by the distance between the otolith’s core and the first annulus for each otolith measured and then calculated the mean of this number for the entire sample: n X((,2 ~rl)lrl)i = 0 353 (SE = 0.0065) n We then estimated the expected distance between the first and second annulus for each age-1 black grouper otolith as a function of the distance between the otolith’s core and the first annulus. The percent marginal increment for age-1 fish was then calcu- lated as [MI / (0.353 xr,))x 100, where MI = the marginal increment. We then plotted the percent marginal increments as a function of capture month for 1995-96, the period dur- ing which we made regular monthly collections. Reproduction Histological sections of gonads from 857 black grou- per that ranged in length from 155 to 1518 mm were prepared and assessed for reproductive state. Gonad samples were processed histologically with ' Modifi- cation of the periodic acid SchifF s (PAS) stain for gly- col-methacrylate sections, with Weigert’s iron-hema- toxylin as a nuclear stain, and with metanil yellow as a counterstain (Quintero-Hunter et al., 1991). Oocytes were staged and counted from histologi- cal preparations at lOOx with a compound microscope attached to a digital image-processing system. Four oocyte stages were recognized in black grouper ova- ries: primary growth, cortical alveolar, vitellogenic, and oocytes in the final stages of maturation (Wallace and Selman, 1981). The final stages of oocyte matu- ration included yolk coalescence, germinal vesicle migration, germinal vesicle breakdown, and hydra- tion. At least 300 oocytes per slide were staged and counted in arbitrarily chosen fields, and frequencies were expressed 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. Gonads were classified on the basis of the matu- rity criteria of Moe ( 1969 ). F emale grouper were con- sidered sexually mature if ovaries contained vitel- logenic oocytes or contained evidence of widespread atresia consistent with gonadal recrudescence. Ma- ture females included Moe’s classes 2, 3, and 4. Monthly median gonadosomatic indices (GSI) were plotted to show seasonal reproductive patterns. Gonadosomatic indices were calculated for 201 sexu- 738 Fishery Bulletin 96(4), 1 998 ally mature female grouper ranging in length from 565 to 1310 mm and for 43 sexually mature male grouper ranging in length from 947 to 1518 mm as GSI = (GW / (TW - GW)) x 100, where GW = total gonad weight (g); and TW = total fish weight (g). We measured the diameters of whole oocytes from six females whose ovaries contained numerous hy- drated oocytes. Samples of ovarian tissue that were preserved in 10% formalin were rinsed briefly in water and transferred to 33% glycerin for measure- ments. Diameters of at least 300 oocytes for each fe- male were measured with a digital image-prossessing system. Only oocytes larger than 0.2 mm in diam- eter were measured. We estimated the length at which 50% of the fe- males in the population reached sexual maturity by fitting a logistic function to the percentage of females that were sexually mature and their respective lengths. A similar function was fitted to the percent- age of the population made up by females to esti- mate the length at which 50% of the females in the population transformed into males. The parameter b in Table 1 is the inflection point of the logistic curves and is the estimate of the length or age at which 50% of the females in the population were sexually ma- ture or the length or age at which 50% of the popula- tion had undergone transition from female to male. Sometimes distinguishing between the gonads of sexu- ally immature grouper and the regressed gonads of mature fish was difficult, and we eliminated 29 females that ranged in length from 586 to 894 mm from our analysis of maturation because we were uncertain of their maturity. Hunter et al. ( 1992) recommended that only fish collected early in the spawning season be used to estimate the length at 50% maturity, but because we found evidence of some spawning during most of the year, we included all months in our analysis. Results The 1166 black grouper we examined ranged from 155 to 1518 mm in length. Among black grouper that we sexed, females ranged from 155 to 1310 mm in length (mean=696.3, SD=204.53, t?=834) and males ranged from 947 to 1518 mm (mean=1254.9, SD=126.19, n=54; Fig. 1). Our sample contained 276 grouper that were not sexed, usually because the grouper had been eviscerated at sea by fishermen. The sex ratio of our sample was 1:15.4 (male:female); greatly skewed towards females. Black grouper Table 1 The relations between standard length, fork length, and total length, between length and weight, between the per- centage of females that were sexually mature and length, and between the percentage of the population made up by females and length of black grouper, Mycteroperca honaci , from South Florida waters. TL = total length (mm), FL = fork length (mm), SL = standard length (mm), WT = whole weight (g), GWT = gutted weight (g), OTOWT = otolith weight (g), and AGE = age (years). The total length range for all length-length regressions was 238-1518 mm and for the length-weight regression was 177-1518 mm; the total weight range for the total weight-gutted weight re- gression was 467-61,462 g; the age range for the age-otolith weight regression was 1-30.5 years. Y = a +bX a b Y X n (1 SE) (1SE) r2 SL FL 1,134 -23.712 0.883 0.999 (0.6852) (0.0008) SL TL 1,141 -21.611 0.858 0.999 (0.7743) (0.0009) FL SL 1,134 27.607 1.131 0.999 (0.7514) (0.0011) FL TL 1,150 1.641 0.973 0.999 (0.5223) (0.0006) TL SL 1,141 26.186 1.164 0.999 (0.8757) (0.0012) TL FL 1,150 -1.317 1.028 0.999 (0.5378) (0.0006) WT GWT 638 81.519 1.056 0.999 (12.9253) (0.0014) log10WT log10TL 772 -5.457 3.218 0.995 (0.0323) (0.0115) Y = a + bx -~[x OTOWT AGE 389 -0.1026 0.1210 0.952 (0.0036) (0.0014) Y = (l/(l + e( a Total length (mm) E Estimated age (yr) Z Total length (mm) Estimated age (yr) Figure 2 Total lengths (mm) and ages (years) of black grouper, Mycteroperca bonaci , landed in the Florida Keys (A) and in Pinellas County on Florida’s Gulf coast (B). 740 Fishery Bulletin 96(4), 1998 ratio of black grouper landed in Pinellas County by commercial fishermen was 1:3.1. The pooled length-weight equa- tion for sexed and unsexed fish and the relations between SL, FL, and TL are presented in Table 1. Age and growth Figure 3 (A) Sectioned sagitta from a 3-year-old black grouper, Mycteroperca bonaci (508 mm TL), captured in May 1996. The third annulus is on the edge of the otolith. (B) Sectioned sagitta from a 3+ year-old black grouper (671 mm TL) captured in January 1996. Note the wide marginal increment (m) subsequent to the third annulus. (C) Sectioned sagitta from a 33-year-old male black grouper (1427 mm TL) captured in June 1996. Note the close spacing of annuli that caused difficulty in age estimation. Scale bars = 500 microns. When viewed with reflected light, black grouper otoliths have narrow, opaque (bright) annuli that alternate with broad translucent (dark) zones (Fig. 3, A and B). Proceeding from the otolith’s core towards the otolith’s proximal mar- gin, these translucent zones become in- creasingly opaque in appearance as the otolith grows. In the outer portion of the otoliths from older grouper ( > 10 years old), the dark translucent zones are narrow. In some individuals, the annuli are indistinct and irregular in appear- ance, which makes age estimation dif- ficult. In older grouper, the annuli be- come closely spaced near the edge of the otolith, and this also makes age estima- tion difficult (Fig. 3C). Often otoliths of older grouper ( > 1 5 years) were more easily read under transmitted light and compound microscopes; otoliths from younger fish were often more easily read under reflected light. Overall, the otoliths from black grouper are similar in appear- ance to those of other grouper species we have examined in our laboratory. Marginal-increment analysis of otoliths from grouper 1-7 years old sug- gested that one annulus had been formed during April-June each year (Fig. 4). Median marginal increments had a consistent seasonal minimum during May-July and a maximum dur- ing December-March in both 1995 and 1996. Marginal increments during April- June tended to be either large (>75%) or small (<25%), suggesting that most individuals either had wide mar- gins, as expected just prior to annulus formation, or had just formed an annulus and had either no mar- ginal increment or a narrow margin. By June or July, individuals with wide margins were no longer present, suggesting that annulus formation was completed. Of 1059 otoliths that we examined, 132 (12.5%) were rejected because of disagreements among read- ings. We had total agreement among annulus counts for 403 otoliths, but all of these otoliths were from grouper estimated to be less than 10 years old. The length-frequency distribution of fish whose otoliths were rejected because they were unsuitable for age estimation was not significantly different from that of fish whose otoliths were readable (Kolmogorov- Smirnov two-sample test, two-sided test statistic = 1.319, P=0.062), but the otolith weight distribution of fish whose otoliths were unreadable was signifi- cantly different from that of fish whose otoliths were readable (Kolmogorov-Smirnov two-sample test, two- sided test statistic=2.017, P<0.001). Otolith weight Crabtree and Bullock: Life history of Mycteroperca bonaci 74 1 1995 1996 Month Figure 4 Percent marginal increments (*) and medians (+ and line) for black grouper, Mycteroperca bonaci, ages 1-7 captured from South Florida waters (n=330). was closely related to fish age (r2=0.952, Table 1, Fig. 5A), and the median weight (0.2437 g) of otoliths re- jected as unreadable was significantly greater than that of otoliths that were readable (0.1571 g; Mann- Whitney W test; W=12, 492.0, P<0.001). Black grouper growth was rapid until an age of about 10 years and then slowed considerably (Table 2, Fig. 5B). Most of the fish in our sample were from 2 to 10 years old, and the most abundant age classes were 2-6 years old (Fig. 6). The two oldest black grouper exam- ined were a 33-year-old (1325-mm) fish that was not sexed and a 33-year-old ( 1427-mm) male (Table 2). The oldest female was a 1275-mm fish estimated to be 23 years old (Fig. 6); the largest female was 1310 mm long (Fig. 1) and was estimated to be 12 years old. Only three female black grouper were estimated to be over 20 years old. The smallest and youngest male was a 947-mm fish estimated to be 6 years old (Fig. 1 and 6). Black grouper landed in Pinellas County on Florida’s Gulf coast by commercial fishermen were older overall than those landed in the Florida Keys, where only one fish older than 15 years was exam- ined (Fig. 2). Estimates of von Bertalanffy growth model parameters are presented in Table 3. Sexual maturation and transition Black grouper, like most epinepheline serranids, are protogynous hermaphrodites. Protogyny was sug- gested by the presence of peripheral sperm-collect- ing sinuses rather than the centrally located sinuses typical of gonochorists. Furthermore, a membrane- lined central lumen that was not used for sperm transport was present in testes and was a structural remnant of the ovarian lumen (Sadovy and Shapiro, 1987). Additional evidence of hermaphroditism was the presence of atretic vitellogenic oocytes in a tran- sitional gonad containing proliferating testicular tis- sue (Fig. 7A). A second grouper that we would have considered transitional except for the presence of some mature sperm in peripheral sperm sinuses also contained degenerating vitellogenic oocytes (Fig. 7, B and C); this fish was considered to be a functional male. Transitional grouper also contained numerous PAS-positive melanomacrophage centers (Ravaglia and Maggese, 1995) referred to as “yellow-brown bodies” by Sadovy and Shapiro ( 1987). When stained with the PAS stain, these structures are brilliant purple. Melanomacrophage centers are thought to be active in degrading atretic oocytes, postovulatory follicles, and residual cells of the spermatogenic cycle (Chan et al., 1967; Ravaglia and Maggese, 1995). Many mature males contained scattered primary growth stage oocytes throughout the testis. We also noted that small amounts of testicular tissue were often present in functional ovaries that showed no evidence of undergoing transition. Finally, the dif- ferences in male and female length-frequency distri- 742 Fishery Bulletin 96(4), 1998 Estimated age (yr) Figure 5 (A) The otolith weight-age relation for black grouper Mycteroperca bonaci, col- lected from South Florida waters. The equation for the relation is presented in Table 1. (B) Observed and predicted total lengths (mm) from the von Bertalanffy growth model for sexed and unsexed black grouper, Mycteroperca bonaci. butions and the absence of small or young males are consistent with our hypothesis of protogynous her- maphroditism (Fig. 1). We estimated that 50% of the females in the popu- lation reached sexual maturity by 826 mm and an age of 5.2 years (Table 1, Fig. 8). The smallest sexu- ally mature female in our sample was 508 mm, and the youngest sexually mature female was 2 years old. All of the ovaries we examined contained primary growth stage oocytes. Cortical alveolar stage oocytes occurred only in ovaries from grouper larger than about 500 mm and older than 2 years, and they were common only among grouper larger than about 600 mm and older than 3 years (Fig. 9A). Vitellogenic oocytes were found only in ovaries from fish larger than about 600 mm and older than 2 years and were common only among grouper larger than 800 mm and older than 5 years (Fig. 9B). The length and age at which vitellogenic oocytes were commonly found agrees well with our estimate of the length and age at which 50% maturity was reached, suggesting that misclassification of regressed gonads did not greatly bias our estimates. Transition from female to male was also partially a function of length and age. By a length of 1214 mm and an age of 15.5 years, 50% of the females in the population had transformed into males (Table 1, Fig. 8). We found only one transitional fish ( 1030 mm long; Fig. 7 A) and a second grouper (947 mm long) that appeared to have recently completed transition (Fig. 7, B and C). We did not find any immature males (Moe’s class 6) in our sample, suggesting that males become sexually active soon after transition. Most (91%) of the males we examined had ripe testes (Moe’s class 9) that contained mature sperm. Fin pigmentation differed between sexes in black grouper (Fig. 10). Ontogenetic color change indepen- dent of sex was ruled out because relatively small males (<1000 mm) displayed the male color phase, but the oldest and largest females (e.g., 1255 mm) did not. The pectoral fins, anal fin, and caudal fin were dark colored in females, but these fins were jet Crabtree and Bullock: Life history of Mycteroperca bonaci 743 black in males (Fig. 10, A-C). The jet black pigmentation of the pectoral fin was most apparent on the medial side. The dorsal interspinous mem- brane of females was yellow to dusky (Fig. 10D). In males, this membrane was usually dark black but was oc- casionally yellow or yellowish with black tips. Because this membrane can be yellow in both males and fe- males, the coloration of this fin was the least reliable indicator of sex. The distal one-third of the soft dorsal fin was dark black in both sexes and was not useful in determining sex. Seasonality of gonad development The frequencies of the four oocyte stages we counted in black grouper ovaries had a seasonal pattern that was repeated in both 1995 and 1996 ( Fig. 1 1, A and B ). Vitellogenic oocytes were present in greatest number in January-March in both 1995 and 1996 and were present in all months except September-November 1996. Cortical alveolar stage oocytes were also present in all months except September-November 1996 and were most abundant during December- April. Primary growth stage oocytes were present during all months and made up at least 55% of the total number of oocytes present. Oocytes in the final stages of maturation were not abundant, but they were most common during December-May (Fig. 11B). We examined six females with hydrated ovaries: one was caught during October, one in December, and four in March. The ovaries from these Table 2 Average observed and predicted total lengths (mm) for sexed and unsexed black grouper, Mycteroperca bonaci, and average observed lengths for females and males. The average observed length at age includes some seasonal growth that occurred after the formation of the final annulus. Values in parentheses are standard error and sample size. Age (yr) Sexed and unsexed Females Males Average observed Predicted Average observed Average observed 0 155 (1) 159 155 (1) 1 338 (8.7;41) 337 339(8.9; 39) 2 486 (5.7;113) 487 486(5.7:112) 3 606 (5.7;136) 614 607 (5.7;134) 4 733 ( 5.5; 155 ) 722 734(5.7:147) 5 814 (5.8; 1 13 ) 813 814 (6. 1;97) 6 871 (7.7;70) 889 857 (10.2;41) 947 (1) 7 972 (9.1;40) 954 966 (13.6;20) 1062 (1) 8 1008(11.6;39) 1009 998 (20.5:15) 1006 (46.0;2) 9 1038(12.3:31) 1055 1024 (19.8;14) 1058 (1) 10 1074 (12.3;16) 1094 1078 (14.6;5) 1167(1) 11 1116(13.1:33) 1127 1147 (27.0;10) 1130(22.8:5) 12 1138(14.4;21) 1155 1131 (29.0;9) 13 1164(11.6;11) 1178 1182(16.0:6) 1172(1) 14 1175(30.6;11) 1198 1192(18.5:2) 1222 (30.5;2) 15 1239 (14.4;7) 1215 1213 (1) 1266 (26.0;3) 16 1165(10.5;2) 1229 17 1244 (21.2;6) 1241 1270 (1) 1283 (47.5;2) 18 1252 (30.3;10) 1251 19 1226 (35.5:6) 1260 1285 (1) 20 1248 (12.9;9) 1267 1220 (1) 21 1264 (20.2;15) 1273 1195(1) 1315 (38.7;4) 22 1278(29.6:5) 1278 23 1298 (27.3;6) 1283 1275 (1) 1326 (68.5;2) 24 1320 (26.6;5) 1286 1370(1) 25 1307 (72.2;4) 1289 1390 (128.0:2) 26 1305 (44.5;4) 1292 27 1309 (23.3;7) 1294 1291 (38.3;4) 28 1300 (1) 1296 29 1349 (28.3;3) 1298 1349 (28.3;3) 30 1307 (55.9;3) 1299 31 1350 (1) 1300 1350 (1) 32 1301 33 1376 (51.0;2) 1302 1427(1) grouper contained a group of vitellogenic oocytes that ranged from 0.5 to 0.7 mm in diameter and a clutch of hydrated oocytes that ranged from 0.8 to 1.2 mm in diameter (Fig. 12). Most of the oocytes <0.2 mm in diameter, which we did not measure, were primary growth stage or cortical alveolar stage oocytes. Postovulatory follicles were not recognized in any female examined. Females with the greatest GSIs (>5) were captured during February-March (Fig. 13). Male GSIs (range 0.086-0.399, n=43) were generally much less than fe- male GSIs (range 0.032-10.108, n=201) and had no Table 3 Parameter estimates for the von Bertalanffy growth model for black grouper, Mycteroperca bonaci, collected from the waters of South Florida. Values in parentheses are stan- dard errors. n Lm(mrn) K t0 adjusted r 2 927 1306.2 0.169 -0.768 0.941 (8.05) (0.0037) (0.0640) 744 Fishery Bulletin 96(4), 1998 Estimated age (yr) Figure 6 Age-frequency distributions for male and female black grouper, Mycteroperca bonaci, from South Florida waters. seasonal trend (Fig. 13). Testes from sexually mature males ranged in weight from 14 to 197 g and were much smaller overall than ovaries from sexually mature fe- males, which ranged in weight from 1 to 1354 g. Discussion We obtained black grouper from a variety of fishery- dependent and fishery-independent sources, and it is difficult to assess the extent to which the length and age structure of our sample reflects that of the population or the degree to which different fishing gears biased the various samples. Although we sampled grouper that were landed in two geographi- cally separate locations, the Florida Keys and in Pinellas County on Florida’s Gulf coast, most of the grouper landed in both locations were caught in the waters off the Florida Keys. Most of the large (>1000 mm) fish and most of the males that we examined came from commercial fish houses on Florida’s Gulf coast, and most of these grouper were caught on the Tortugas Banks west of the Florida Keys. Fish ob- tained from Keys headboats, fish houses, and spear- fishermen were smaller and younger overall than grouper sampled from fish houses in Pinellas County (Fig. 2). Most black grouper landed in Pinellas County fish houses were caught with commercial longlines in depths greater than 37 m (20 fathoms); most of the fish landed in the Keys were captured by spear fishermen, principally in water 6-50 m deep. The greater depths fished and the greater distances traveled by Gulf-coast commercial fishermen to pre- sumably more remote fishing areas probably account for the differences between the two samples. Age and growth Black grouper appear, on the basis of marginal-in- crement validation for grouper 1-7 years old, to form a single annulus each year in late spring or early summer. We were unable to validate that annuli are annual marks in older grouper, and the accuracy of our age estimates for older fish is unknown. The annuli we counted on otolith sections from older grou- per were similar in appearance to validated annuli; Crabtree and Bullock: Life history of Mycteroperca bonaci 745 Figure 7 (A) A transitional gonad from a 1030-mm-TL black grouper, Mycteroperca bonaci , captured in September 1994 with prolif- erating spermatogenic tissue in sperm cysts (S), primary growth stage oocytes (PG), degenerating vitellogenic oocytes (VO), and PAS-positive melanomacrophage centers (PAS). (B) A 947-mm- TL black grouper captured in November 1995 with proliferat- ing spermatogenic tissue in sperm cysts (S), primary growth stage oocytes ( PG), degenerating vitellogenic oocytes (VO ), and PAS-positive melanomacrophage centers (PAS). (C) The same ovary as in B showing the presence of mature sperm in periph- eral sperm sinuses (SS). Scale bars = 200 microns. however, as annuli became progressively more closely spaced, age estimation became more dif- ficult. Additional work is needed to assess the accuracy of our age estimates for older grouper; consequently, our age estimates predicted lengths at age, and growth model parameters should be used with caution. From their analyses of mar- ginal increments, Manooch and Mason (1987) also reported that black grouper from the Florida Keys form a single annulus each year. They found that annuli were usually formed during March- May, about a month earlier than we estimated. This difference is probably a result of differences in interpretation of the appearance of an annu- lus present on the otolith margin. Other conge- ners also form annuli during late spring and early summer. Collins et al. (1987) reported that gag, M. microlepis, off the southeastern U.S. coast form annuli during late spring to late summer, and Hood and Schlieder (1992) suggested that gag form annuli during summer in the eastern Gulf of Mexico. Bullock and Murphy (1994) re- ported that yellowmouth grouper, M. inter- stitialis, form opaque bands during late summer and early fall. Matheson et al. (1986) found that scamp, M. phenax, in North Carolina waters form an annual mark during April or May. We rejected 12.5% of the otoliths we sectioned as unreadable, and this is a possible source of bias to our growth model parameters. Otolith weight is a useful predictor of age of black grou- per (r2=0.952), and the otoliths that we rejected were generally heavier and thus probably older than most of those that we considered readable. In addition, although the length-frequency dis- tribution of fish whose otoliths were rejected was not significantly different from that of all fish whose otoliths were readable, the significance level of the test (P=0.062) was close enough to 0.05 to cause us to suspect that the distributions could have been different. If so, we may have re- jected as unreadable more otoliths from large grouper than from small grouper. We may also have tended to exclude slower-growing older grouper from our sample of aged fish, and this could have biased our growth models. However, when we recalculated growth models and in- cluded all fish regardless of CV, the resulting growth-parameter estimates were all within one standard error of those in Table 3. Thus any bias to the growth model caused by our rejection of otoliths from older grouper appears to be negli- gible. Our choice of a threshold CV of 12% was arbitrary, but growth model parameters did not appear to be sensitive to the choice of CV thresh- 746 Fishery Bulletin 96(4), 1998 Age (yr) Figure 8 Logistic curves describing the percentage of female black grouper, Mycteroperca bonaci, collected from South Florida waters that were sexually mature and the percentage of our black grouper Sample that was made up by females as a function of both total length (mm) and age (years). M-0 is the length or age predicted by the logistic regression at which 50% of the female black grouper were sexually mature, and P50 is the length or age predicted by the logistic regression at which 50% of the black grouper in our sample were females. olds. When we recalculated growth models with CV thresholds of 10% and 5%, our estimates of Lo and K remained within two standard errors of the estimates at a CV of 12%, but the numbers of otoliths elimi- nated from our analysis increased substantially at lower CV thresholds. The two oldest black grouper in our sample were both estimated to be 33 years old, considerably older than Manooch and Mason’s ( 1987) oldest grouper ( 14 years). Manooch and Mason sampled hook-and-line- caught black grouper from Keys headboats during 1978-85 and speculated that the maximum age at- tained by black grouper was probably about 19 years. The presence of older fish in our sample is probably because our sample contained black grouper that were larger (largest 1518 mm) than those examined by Manooch and Mason (largest aged 1180 mm; larg- est measured 1259 mm). We measured 125 black grouper longer than 1200 mm and aged 94 of these. Of the 98 grouper that we estimated to be older than 14 years, 84% were longer than 1200 mm and were longer than any fish aged by Manooch and Mason. Most of the large black grouper in our sample were caught by commercial longliners and landed in Pinellas County; the lengths and ages of the grouper in our sample that were landed in the Keys were simi- lar to those of Manooch and Mason (1987). Our estimates of length at age are similar to those of Manooch and Mason (1987), and the predicted growth curves from both data sets are in general agreement (Fig. 14). In some cases, such as age classes 5-9, our estimates of mean length at age are Crabtree and Bullock: Life history of Mycteroperca bonaci 747 20 20 A A 15 ■ 15 : * * 1 «« ■ . * 10 5 % » /iV-v* ■ ’ . 1 <*.. • ** * * ... *»* ■*« ■* 10 5 1 • * j *; . -c-y v* •* a * * # * M M 0 — •_ * 0 -T * ■21" S a isgSp&swia scsi m mm* m >> U a 3 cr a 0 300 600 900 1200 Total length (mm) 1500 >> o G 0 5 10 15 Age (yr) 20 25 u. 50 tL, 50 . 40 B 40 B ; * * % ' 30 • ■ i * 30 * " « * * 20 * « ** ’ * * « » ■■ '■ 20 * * * 10 v *.»v • . ■ 10 V d ,V * •* m M * N * * * * * • - 0 m m* * * iS 0 * , » " * * i 1 ■ ■ ■ ms* * *m • m * 0 300 600 900 1200 1500 0 5 10 15 20 25 Total length (mm) Figure 9 Age (yr) The percent frequency of occurrence of (A) cortical alveolar oocytes and (B) vitellogenic oocytes in black grouper. Mycteroperca bonaci , ovaries as a function of total length (mm; n =772) and age (years; n=609). greater than those of Manooch and Mason, but their estimates for these age classes fall below not only our predicted lengths but also their predicted lengths. The von Bertalanffy growth model parameters esti- mated by Manooch and Mason were = 1,352 mm, greater than our estimate of Lj= 1,306 mm, and K = 0.1156, less than our estimate of K = 0.169. Black grouper, though not as large as jewfish or warsaw grouper, are among the largest species of grouper, and it is not surprising that they attain ages over 30 years. Estimates of the longevity of other grouper species are similar to our estimates for black grouper. Hood and Schlieder (1992) and Collins et al. (1987) estimated a maximum age of 21-22 years for gag, Bullock et al. (1992) estimated a maximum age of 37 years for jewfish ( E . itajara), Bullock and Murphy ( 1994) estimated a maximum age of 28 years for yellowmouth grouper, Manooch and Mason (1987) reported that warsaw grouper ( E . nigritus) reach 41 years, and Moe (1969) reported a red grouper ( E . morio) 25 years old. The growth rate (K-0. 169/year) of black grouper is similar to that estimated for gag (K=Q. 166/year, Hood and Schlieder, 1992), but is greater than the growth rates estimated for many other grouper species: jewfish K = 0.126/year (Bul- lock et al., 1992), yellowmouth grouper K - 0.08/year (Bullock and Murphy, 1994), and warsaw grouper K = 0.054/year (Manooch and Mason, 1987). Sexual maturation and transition Our histological analysis of black grouper gonads is consistent with the diagnostic criteria of Sadovy and Shapiro (1987) for a monandric protogynous her- maphrodite. Furthermore, the absence of small males and a sex ratio highly skewed towards females are consistent with our diagnosis of protogynous her- maphroditism. Although the length and age distri- butions of males and females overlapped, males oc- cupied the largest and oldest length and age classes and were unrepresented in smaller and younger length and age classes (Figs. 1 and 6). This is an important difference between black grouper and Nassau grouper (Epinephelus striatus), which has recently been diagnosed as a gonochorist with po- tential for sex change (Sadovy and Colin, 1995). In 748 Fishery Bulletin 96(4), 1998 3«-SSb G«rrn> . Figure 1 0 Differences in fin pigmentation in male (above) and female (below) black grouper, Mycteroperca bonaci. (A) Medial surface of pectoral fin, (B) anal fin, and (C) caudal fin jet black in males. (D) Dorsal interspinous membrane yellow or dusky in females, usually black in males. Nassau grouper, both males and females can develop directly from juveniles and there is little difference in the lengths of males and females. Furthermore, there is little difference in the length at which sexual maturation occurs in both sexes, in contrast to black grouper. The scarcity of transitional black grouper and the absence of immature males in our samples suggest that transition occurs quickly and that tes- tes become active soon after transition. The presence of large females in the population suggests that some females may not transform into males, but even the largest females we observed were considerably smaller and younger than the largest and oldest males (Figs. 1 and 6). The current 20-inch (508-mm) minimum size limit is well below our estimate of the size at which 50% of females reach sexual maturity (826 mm and 5.2 years) and is not adequate to allow females to repro- duce before recruiting to the fishery. Similar esti- mates of length at maturity were made by Garcia- Cagide and Garcia (1996) who reported that the smallest sexually mature female black grouper they examined from Cuban waters was 570 mm and that most mature females were 850-1 100 mm. Huntsman et al. (1994) used an estimate of 5.04 years as the age at sexual maturity in yield per recruit and spawn- ing stock per recruit models for black grouper. They did not present data to support this estimate, but it is within two standard errors of ours. Female black grouper reached sexual maturity at a relatively large size compared to other grouper species. Female gag reach 50% sexual maturity at 600-650 mm and 3-4 years (Hood and Schlieder, 1992). Female yellowmouth grouper mature between 400 and 450 mm and 2-4 years (Bullock and Murphy, 1994). Yellowedge grouper, E. flavolimbatus, reach 50% maturity at 568 mm (Bullock et al., 1996). Larger grouper such as jewfish reach sexual maturity as females at even larger lengths of 1200-1350 mm and greater ages of 6-7 years than black grouper do (Bul- lock et al., 1992). The length and age at which 50% of our sample consisted of females was 1215 mm and 15.5 years. Garcia-Cagide and Garcia (1996) reported that the smallest male black grouper from Cuban waters was 980 mm long and that most males were 1000 to 1100 mm long, similar to our findings even though their sample appeared to contain fewer large grouper than ours did. Our estimates of the length and age at tran- sition for black grouper are both larger and older than estimates for other grouper species. Eastern Gulf of Mexico gag populations are 50% male at about 1050 mm and age 11 (Hood and Schlieder, 1992). Gag in the South Atlantic Bight undergo transition from female to male at 750-950 mm and age 5-11 (Collins et al., 1987). Transitional yellowmouth grouper ex- amined by Bullock and Murphy ( 1994) were 505-643 mm and 5-14 years old. Yellowedge grouper popula- Crabtree and Bullock. Life history of Mycteroperca bonaci 749 1995 1996 Month Figure 1 1 (A) Monthly mean percent frequency of occurrence and standard errors of oocyte stages in black grouper, Mycteroperca bonaci, ovaries (aj =774) for 1995-96. * = primary growth stage oocytes, = cortical alveolar oocytes, and = vitellogenic oocytes. (B) The percent frequency of occurrence of oocytes in the final stages of maturation as a function of month in black grouper ovaries. tions are 50% females at 817 mm (Bullock et al., 1996). Other large grouper such as Warsaw grouper and jewfish have not been confirmed to be protogy- nous hermaphrodites. The sex ratio of our sample (1:15.4, male:female) may not resemble that of the population because many black grouper that we examined, especially large fish, were eviscerated and could not be sexed. Most of these large fish were probably males, so we may have underestimated their numbers. We are not able to assess the extent of this bias. Furthermore, sex ratios appear to vary widely depending on the depth and location sampled. It is possible that fish- ing mortality has reduced the numbers of large males in the population, as has been reported for gag from the eastern Gulf of Mexico, where male:female sex ratios as extreme as 1:76.6 have been reported (Coleman et al., 1996). Unfortunately, there are no historical estimates of black grouper sex ratios with which to compare our estimates. Garcia-Cagide and Garcia (1996) reported a male:female sex ratio of 1:30.3 for black grouper from Cuban waters, more skewed towards females than our overall sex ratio but less skewed than the sex ratio of the black grouper we sampled that were landed in the Florida Keys (1:58.7). Hydrated black grouper oocytes ranged from 0.8 to 1.2 mm in diameter and are similar in size to the hydrated oocytes reported for other groupers (Moe, 1969; Colin et al., 1987; Carter et al., 1994; Sadovy et al., 1994b). We usually did not know the time of day when fish were caught, so we could not estimate the time at which hydration began. Vitellogenic oo- cytes reached a diameter of about 0.6 mm before under going the final stages of maturation. A dis- tinct group of vitellogenic oocytes with a mean di- ameter of about 0.6 mm, along with a clutch of larger hydrated oocytes, was present in most of the six grou- per for which we measured oocytes. This is consis- tent with a group-synchronous mode of reproduction (Wallace and Selman, 1981). 750 Fishery Bulletin 96(4), 1998 n =300 673 mm TL oM IWlff Tn m a : 1 n ["I Oocyte diameter (mm) Figure 12 Oocyte diameter (mm) frequency distributions for six hydrated female black grouper, Mycteroperca bonaci , from South Florida waters. Hydrated oocytes were 0.8-1. 2 mm in diameter. Only oocytes with diameters larger than 0.2 mm were measured, n = the number of oocytes measured. Seasonality of gonad development Reproduction in black grouper was seasonal, with peak gonad development during December-March, but females with vitellogenic oocytes were present during all months and females with elevated GSIs occurred in most months. In Cuban waters, Garcia- Cagide and Garcia ( 1996) also found that black grou- per spawn during winter and spring. Other eastern Gulf of Mexico grouper species have similar seasonal patterns of gonad development, and in several spe- cies, the spawning season appears to be prolonged. In gag, gonad development occurs during December- May, and peak gonadal activity occurs during Feb- ruary and March (Hood and Schlieder, 1992). Yellow- mouth grouper gonads are active throughout the year, and peak gonadal activity occurs during April- May (Bullock and Murphy, 1994). Jewfish have de- veloped gonads from June-December, with peak ac- tivity during July-September (Bullock et al., 1992). Yellowedge grouper gonads are developed during January-October, with peak activity during May- September (Bullock et al., 1996). Many grouper species are known to form seasonal spawning aggregations at specific times and locations each year (Sadovy, 1994). Aggregations of black grou- per have been reported off Central America during January and February (Carter, 1989; Fine, 1990; Carter et al., 1994), the same time of year that we observed peak gonadal development and the great- est incidence of oocytes in the final stages of matu- ration. Among other species of Mycteroperca , aggre- gations of gag and scamp have been reported (Gilmore and Jones, 1992), and spawning by an ag- gregation of tiger grouper, M. tigris, was reported by Sadovy et al. ( 1994a). Spawning aggregations of black grouper have not been documented in Florida wa- ters, but it is clear from the presence of oocytes in the final stages of maturation in our sample that spawning black grouper are landed in South Florida. Whether black grouper form spawning aggregations in Florida waters and the extent of these aggrega- tions is unknown. Crabtree and Bullock: Life history of Mycteroperca bonaci 751 V) O Month Figure I 3 Gonadosomatic indices (GSI, *) and medians (+ and lines) for sexually mature female and male black grouper, Mycteroperca bonaci , plotted by month. Conclusions The status of black grouper stocks is unclear. Our sample contained many old and large grouper, most of which came from the commercial longline fishery. Large grouper are apparently still abundant in some areas, and their continuing presence in the commer- cial fishery could reflect the expansion of fishing ef- fort into deeper and more remote areas. In the shal- low waters adjacent to the Florida Keys, large and old black grouper were rare, probably a consequence of fishing mortality. The current minimum size does not protect sexually mature females, and the fishery could substantially reduce spawning success in black grouper. In addition, size-selective fishing mortality may selectively remove males from the population. Sex ratios are currently skewed towards females, and it is unknown if populations can compensate by re- ducing the size of transition as males are selectively removed. Finally, additional research is needed to document the existence of black grouper spawning aggregations in the eastern Gulf of Mexico and in the waters off the Florida Keys and to evaluate the extent to which these aggregations have been affected by fishing mortality. Acknowledgments We thank D. BeMaria and K. DeMaria for collecting black grouper from the Florida Keys. We also thank C. Boniface and M. Puig for their assistance in sam- pling in the Keys. Many commercial fishermen coop- erated by providing ungutted black grouper, and Dick’s Seafood, Madeira Beach Seafood, and Captain’s Finest Seafood allowed sampling of grou- per during their busy schedules. Captains C. Sullivan, B. Banks, and G. Haring deserve special thanks for saving gonads with their catch. We ac- knowledge the following FMRI personnel for their assistance: M. Cladas, L. French, G. Gerdeman, D. Harshany, M. Norris, D. Merryman, H. Patterson, T. Sminkey, F. Stengard, and C. Stevens. A. Collins, NMFS, Panama City, FL, provided samples from the 752 Fishery Bulletin 96(4), 1 998 Estimated age (yr) Figure 14 Mean length at age and von Bertalanffy growth curves for black grouper, Mycteroperca bonaci, based on data from our study (* and solid line) and from Manooch and Mason (1987; and broken line). Florida Panhandle. J. Leiby, R. McBride, M. Murphy, J. Quinn, and D. Winkelman made helpful comments that improved the manuscript. Lastly, we thank C. Manooch and M. Burton for providing us with the data described in Manooch and Mason (1987). This project was funded by a grant from the National Marine Fisheries Service, U.S. Department of Com- merce, Award NA57FF0060, L.H. Bullock and R.E. Crabtree, principal investigators. Literature cited Bannerot, S. P., W. W. Fox Jr., and J. E. Powers. 1987. Reproductive strategies and the management of snap- pers and groupers in the Gulf of Mexico and Caribbean. In J. J. Polovina and S. Ralston (eds. ), Tropical snappers and groupers: biology and fisheries management, p. 561- 603. Westview Press, Boulder, CO. Bullock, L. H., M. F. Godcharles, and R. E. Crabtree. 1996. Reproduction of yellowedge grouper, Epinephelus flavolimbatus , from the eastern Gulf of Mexico. Bull. Mar. Sci. 59:216-224. Bullock, L. H., and M. D. Murphy. 1994. Aspects of the life history of the yellowmouth grou- per, Mycteroperca interstitialis, in the eastern Gulf of Mexico. Bull. Mar. Sci. 55:30-45. Bullock, L. H., M. D. Murphy, M. F. Godcharles, and M. E. Mitchell. 1992. Age, growth, and reproduction of jewfish Epinephelus itajara in the eastern Gulf of Mexico. Fish. Bull. 90: 243-249. Bullock, L. H., and G. B. Smith. 1991. Seabasses (Pisces: Serranidae). Mem. Hourglass Cruises, Vol. 8, part 2, 243 p. Carter, J. 1989. Grouper sex in Belize. Nat. Hist. 10:60-69. Carter, J., G. J. Marrow, and V. Pryor. 1994. Aspects of the ecology and reproduction of Nassau grouper, Epinephelus striatus, off the coast of Belize, Cen- tral America. Proc. Gulf. Caribb. Fish. Inst. 43:65-111. 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. (Lond.) 153:527-539. Coleman, F. C., C. C. Koenig, and L. A. Collins. 1996. Reproductive styles of shallow-water groupers (Pisces:Serranidae) in the eastern Gulf of Mexico and the consequences of fishing spawning aggregations. Environ. Biol. Fish. 47:129-141. Colin, P. L., D. Y. Shapiro, and D. Weiler. 1987. Aspects of the reproduction of two groupers, Epinephelus guttatus and E. striatus in the West Indies. Bull. Mar. Sci. 40:220-230. Collins, M. R., C. W. Waltz, W. A. Roumillat, and D. L. Stubbs. 1987. Contribution to the life history and reproductive bi- ology of gag, Mycteroperca microlepis (Serranidae), in the South Atlantic Bight. Fish. Bull. 85:648-653. Fine, J. C. 1990. Groupers in love. Sea Frontiers. Jan-Feb:42-45. Garcia-Cagide, A., and T. Garcia. 1996. Reproduccion de Mycteroperca bonaci y Mycteroperca Crabtree and Bullock: Life history of Mycteroperca bonaci 753 venenosa (Pisces: Serranidae) en la plataforma Cubana. Rev. Biol. Trop. 44:771-780. Gilmore, R. G., and R. S. Jones. 1992. Color variation and associated behavior in the epinepheline groupers Mycteroperca microlepis (Goode and Bean) and M. phenax (Jordan and Swain). Bull. Mar. Sci. 51:83- 103. Hood, P. B., and R. A. Schlieder. 1992. Age, growth, and reproduction of gag, Mycteroperca microlepis (Pisces:Serranidae), in the eastern Gulf of Mexico. Bull. Mar. Sci. 51:337-352. Hubbs, C. L., and K. F. Lagler. 1964. Fishes of the Great Lakes region. Univ. Michigan Press, Ann Arbor, MI, 213 p. Hunter, J. R., B. J. Macewicz, N. C. Lo, and C. A. Kimbrell. 1992. Fecundity, spawning, and maturity of female Dover sole Microstomus pad ficus, with an evaluation of assump- tions and precision. Fish. Bull. 90:101-128. Huntsman, G. R., J. Potts, and R. W. Mays. 1994. A preliminary assessment of the populations of seven species of grouper ( Serranidae, Epinephelinae ) in the west- ern Atlantic Ocean from Cape Hatteras, North Carolina to the Dry Tortugas, Florida. Proc. Gulf Caribb. Fish. Inst. 43:193-213. Manooch, C. S., III, and D. L. Mason. 1987. Age and growth of the Warsaw grouper and black grouper from the southeast region of the United States. Northeast Gulf Sci. 9:65-75. Matheson, R. H. Ill, G. R. Huntsman, and C. S. Manooch III. 1986. Age, growth, mortality, food and reproduction of the scamp, Mycteroperca phenax, collected off North Carolina and South Carolina. Bull. Mar. Sci. 38:300-312. Moe, M. A., Jr. 1969. Biology of the red grouper Epinephelus morio (Valenciennes) from the eastern Gulf of Mexico. Fla. Dep. Nat. Resour. Mar. Res. Lab. Prof. Pap. ser. 10, 95 p. Quintero-Hunter, I., H. Grier, and M. Muscato. 1991. Enhancement of histological detail using metanil yellow as counterstain in periodic acid SchifF s hematoxy- lin staining of glycol methacrylate tissue sections. Biotechnol. 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, Synbranchidae): histological and histochemical characteri- zation. J. Fish Dis. 18:117-125. Sadovy, Y. 1994. Grouper stocks of the western central Atlantic: the need for management and management needs. Proc. Gulf Caribb. Fish. Inst. 43:43-64. Sadovy, Y., and P. L. Colie. 1995. Sexual development and sexuality in the Nassau grouper. J. Fish Biol. 46:961- 976. Sadovy, Y., P. L. Colin, and M. L. Domeier. 1994a. Aggregation and spawning in the tiger grouper, Mycteroperca tigris (Pisces: Serranidae). Copeia 1994: 511-516. Sadovy, Y., A. Rosario, and A. Roman. 1994b. Reproduction in an aggregating grouper, the red hind, Epinephelus guttatus. Environ. Biol. Fish. 41:269- 286. Sadovy, Y., and D. Y. Shapiro. 1987. Criteria for the diagnosis of hermaphroditism in fishes. Copeia 1987:136-156. Secor, D. H., J. M. Dean, and E. H. Laban. 1992. Otolith removal and preparation for microstructural examination. In D. K. Stevenson and S. E. Campana (eds.), Otolith microstructure examination and analysis, p. 19-57. Can. Spec. Publ. Fish. Aquat. Sci. 117. Smith, C. L. 1971. A revision of the American groupers: Epinephelus and allied genera. Bull. Am. Mus. Nat. Hist. 146(2):67-242. 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 growth in teleosts. Am. Zool. 21:325-343. 754 Abstract —We examined stomach contents of 385 bonefish that ranged in length from 228 to 702 mm FL. Rela- tively few prey species made up most of the diet by weight — xanthid crabs (29.9%), gulf toadfish, Opsanus beta ( 17.2%), portunid crabs ( 10.9%), alpheid shrimp (9.2%), and penaeid shrimp (7.7%) together made up 74.9%. A vari- ety of gastropods (17 families and 24 species) and bivalves (9 families and 16 species) were eaten, but gastropods made up only 2.7% of the diet by weight and bivalves made up only 2.5%. Poly- chaetes, represented by at least seven families, were important in the diet nu- merically (37.1%) but made up little of the diet by weight (1.4%). Cluster analysis and ordination of stomach con- tents permitted bonefish to be grouped according to length. Large bonefish (>439 mm FL) ate more xanthid crabs, alpheid shrimp, Callinectes spp., and O. beta than did small bonefish; penaeid shrimp were more important in the diet of small bonefish (<440 mm FL). The stomach contents of bonefish caught in Florida Bay were significantly differ- ent from those of bonefish caught on the ocean (Florida Straits) side of the Florida Keys, but the differences were slight and the same prey taxa domi- nated the diet in both areas. Xanthid crabs, alpheid shrimp, O. beta, penaeid shrimp, and Callinectes spp. together made up over 50% of the dissimilarity in diet of bonefish between the two ar- eas. Some seasonal effects on diet were found, but variable sample sizes among seasons in the respective sampling ar- eas made it difficult to detect seasonal trends. Bonefish fed selectively on some prey groups, but other common prey groups were not selected and were less common in stomachs than in the prey environment. The suite of epibenthic crustaceans and fishes found in bone- fish stomachs was significantly differ- ent from that available as prey in the environment. Our results suggest that teleosts, mainly O. beta, are more im- portant in the diet of bonefish than re- ported in previous studies. Manuscript accepted 10 March 1998. Fish. Bull. 96:754-^66 (1998). Feeding habits of bonefish, Albula vulpes, from the waters of the Florida Keys Roy E. Crabtree Connie Stevens Derke Snodgrass Fredrik J. Stengard Florida Marine Research Institute, Department of Environmental Protection 100 Eighth Avenue SE, St. Petersburg, Florida 33701-5095 E-mail address (for R. E. Crabtree): crabtree_r@epic7. dep.state.fi. us Bonefish, Albula vulpes, are the basis of an economically important recreational fishery in the Florida Keys and many parts of the Carib- bean. In the Florida Keys, fishing for bonefish is a year-round activ- ity and provides an important source of income to professional fishing guides. Most bonefish are caught in relatively shallow (<2 m) water over either seagrass or sandy bottom, and it is common for bone- fish to forage in water less than 0.3 m deep, where their tails and dorsal fins can often be seen extending from the water as they feed on benthic and epibenthic prey. Bone- fish are known for their wariness when approached in shallow water and for their strong fighting abili- ties when hooked. In Florida, the commercial sale of bonefish is pro- hibited, and regulations on the rec- reational fishery include a bag limit of one fish per angler per day and a minimum total length of 457 mm (390 mm fork length). Bonefish are not considered a food fish in Florida, and therefore most bonefish caught are released. Crabtree et al. (1996, 1997) re- cently described the age, growth, and reproduction of bonefish from South Florida waters and found that bonefish can attain ages of 19 years. In the Florida Keys, 50% of male bonefish reach sexual matu- rity at 418 mm and an age of 3.6 years, and 50% of female bonefish reach sexual maturity at 488 mm FL and an age of 4.2 years. Bonefish gonadal activity in the Florida Keys is seasonal and spawning occurs during November-May. Feeding habits of bonefish have been studied by Warmke and Erd- man (1963) in Puerto Rico, by Bruger (1974) in the Florida Keys, and by Colton and Alevizon (1983) in the Bahamas; however, none of these studies have adequately de- scribed the diet of bonefish. Warmke and Erdman (1963) identified only mollusks, Bruger (1974) presented frequency of occurrence data for crustaceans but did not quantify noncrustacean prey, and Colton and Alevizon (1983) sorted prey into 10 broad taxonomic categories but did not quantify the abundance of each prey species. Consequently, the relative importance of each prey species in the diet of bonefish is unknown. This information is needed to evaluate the effects of habitat changes on Keys bonefish populations and is particularly im- portant considering the recent seagrass die-offs that have been documented in Florida Bay (Rob- blee et al., 1991; Carlson et ah, 1994; Durako, 1994; Butler et al., 1995; Matheson et al.1 ). If changes in the benthic epifauna and infauna Crabtree et al.: Feeding habits o f Albula vulpes 755 have resulted from the seagrass die-off, these changes could potentially affect both feeding and occurrence of bonefish in Florida Bay. In this article, we describe the feeding habits of bonefish from waters off the Florida Keys. We consider both length-related and sea- sonal changes in bonefish diet. In addition, we com- pare the diets of bonefish collected from two important Keys areas: Florida Bay (including parts of Everglades National Park and adjacent waters) and ocean-side (Florida Straits) fishing areas off the Florida Keys. Methods Collections We examined stomach contents of 385 bonefish col- lected from South Florida waters from December 1991 to April 1995. Most of these bonefish were caught with hook-and-line gear in waters off the Florida Keys, Florida Bay, and Biscayne Bay during daylight hours either by biologists or by a single pro- fessional bonefish guide and his anglers. Supplemen- tal collections of small bonefish (<425 mm FL, n- 22) were made with seines and gill nets of various sizes in waters off the Florida Keys. Bonefish were placed on ice immediately after capture. Fork length (FL) was later measured to the nearest millimeter (mm), stomachs were removed, and the contents preserved in 10% buffered formalin. Contents of individual stomachs were sorted and identified to the lowest possible taxon. Fragments of prey organisms were counted as one, unless countable parts such as eye lenses were found. Weights of prey organisms were measured by blotting prey items on filter paper and weighing them on an analytical balance. The num- ber of individuals of each food type as a percentage of the total number of identifiable prey items (per- cent numerical abundance, N), the percentage of stomachs containing prey in which a particular prey taxa occurred (frequency of occurrence, F), and wet weight as a percentage of the total weight of all prey items (percent weight, W) were determined. For the larger and more abundant prey taxa (alpheid shrimp, penaeid shrimp, portunid crabs, xanthid crabs, and Opsanus beta), we measured prey size to examine the relation between predator and prey size. We measured total length (TL, tip of the rostrum to tip of the uropod) of shrimp, carapace width of crabs, and standard length (SL) of O. beta. 1 Matheson, R. E., D. A. Camp, S. M. Sogard, and K. A. Bjorgo. 1998. Changes in seagrass-associated fish and crustacean com- munities on Florida Bay mud banks: the effects of recent eco- system changes? Manuscript in review. We compared the abundance of prey found in bone- fish stomachs with the abundance of benthic and epibenthic crustaceans and fishes from typical Florida Keys bonefish habitat. Information on the abundance of potential prey in Florida Bay is based on meter-square throw-trap collections by Matheson et al.1 during 1994-96. We used data for Buchanan Bank (n=30 collections) in the Atlantic suben- vironment as described by Zieman et al. (1989) and followed by Matheson et al.1 The Atlantic suben- vironment, and specifically the Buchanan Bank area sampled by Matheson et al.,1 is an area where many of our Florida Bay bonefish were caught. Data on prey abundance from ocean-side (Florida Straits) areas of the Florida Keys are from 54 samples that we collected following the methods of Sogard et al. (1987) and Matheson et al.1 with meter-square throw traps. Ocean-side samples were taken during Sep- tember 1996 (n= 14) and January 1997 (n=40) at vari- ous locations from the middle Keys north to Elliot Key. We sampled areas where we had previously caught bonefish and that appeared to be representa- tive of typical ocean-side bonefish flats. Throw-trap samples were collected over a different time period (1994-97) than that for our bonefish specimens ( 1991-95), and we assumed for our comparisons that prey availability did not change over this time. If prey abundance changed during 1991-97, this change could have biased our comparisons. Data analysis Nonparametric multivariate techniques were used to analyze feeding data. Similarity matrices were constructed with pairwise Bray-Curtis similarity coefficients (Bray and Curtis, 1957). Square-root- transformed, percent-standardized prey-weight data were used to generate similarities. Prey weight was used for all comparisons except feeding selectivity comparisons, because this measure more closely re- flects the energetic importance of a prey species in the diet than does either frequency of occurrence or percent numerical abundance. Percent numerical abundance was used in feeding selectivity compari- sons because we were interested in the relative abun- dance of prey in stomachs and in the environment. Hierarchical agglomerative cluster analysis that in- corporated a group-average linking method was used to search for groups among bonefish stomach con- tents. A nonparametric ordination technique, nonmetric multidimensional scaling (MDS), was used to ordinate sites on the basis of the similarity ma- trix. The contribution of the various prey categories to the percentage similarity within groups and the differences among groups were estimated with a simi- 756 Fishery Bulletin 96(4), 1998 larities percentage (SIMPER) procedure (Clarke, 1993; Clarke and Warwick, 1994). All multivariate analyses were performed with Plymouth Routines in Marine Environmental Research (PRIMER) pro- grams (copyright M. R. Carr and K. R. Clarke, Ma- rine Biological Laboratory, Plymouth, UK; Clarke and Warwick, 1994). Separate analyses were performed to compare stomach contents between various length groups of bonefish, between seasons, and between areas (Table 1). To detect length-related differences in feeding, we pooled bonefish into 20-mm length intervals and used cluster and MDS analyses to compare stomach con- tents. Stomach contents of all 20-mm length groups within the 480- to 699-mm length range had a level of similarity >55%, so we restricted all other com- Table 1 Sample sizes, and collection months and years for data included in ANOSIM comparisons. Numbers in parenthe- ses are the number of samples collected during a particu- lar month. Bonefish, Albula vulpes, fork lengths for all comparisons ranged from 480 to 699 mm. Comparison n Months Years Area Florida Bay 50 Jan (5), Feb (3), Mar (2), Apr (5), May (7), Jun (1), Jul (4), Aug (4), Sep (3), Oct (6), Nov (7), Dec (3) 1991-1995 Ocean side 50 same monthly sample sizes as Florida Bay 1991-1995 Season Ocean side Jan-Mar 39 1991-1995 Apr-Jun 43 Jul-Sep 6 Oct-Dec Florida Bay 33 1991-1995 Jan-Mar 8 Apr-Jun 25 Jul-Sep 70 Oct-Dec 18 Stomach throw trap Ocean side Stomachs 39 Jan. (7), Feb ( 13), Aug (4), Sept (3), Oct (12) 1991-1995 Throw traps Florida Bay 54 Jan (40), Sep ( 14) 1996-1997 Stomachs 45 Mar (2), May (9), Jun ( 10), Sep (24) 1991-1995 Throw traps 30 Mar (6), May (6), Jun (6), Sep (12) 1994-1996 parisons to this length group in order to minimize length-related dietary shifts that could have con- founded comparisons of areas and seasons. We com- pared stomach contents of bonefish from two areas, Florida Bay and the ocean (Florida Straits) side of the Keys, using the analysis of similarity ( ANOSIM) permutation test (Clarke, 1993; Clarke and Warwick, 1994). We did not include the lower Keys or Biscayne Bay in our area comparisons because we examined relatively few bonefish stomachs from these two ar- eas. Bonefish move seasonally between Florida Bay and ocean-side areas; thus for any given month sample sizes were rarely the same for the two areas. To reduce confounding from seasonal effects that could result from unequal seasonal representation of the two areas, we eliminated some stomachs from our analysis to achieve equal monthly sample sizes for the two areas for each month. We pooled samples from all years, totaled the number of samples by month for each area, and then randomly eliminated stomachs for each month from the area with the greatest sample size so that, for any given month, both areas had equal sample sizes. The resulting sample of 50 stomachs from each area contained equal sample sizes from both areas for each month, but the total sample size varied from month to month (Table 1). To detect seasonal dietary shifts, collec- tions were pooled into four seasonal groupings: Janu- ary-March, April-June, July-September, and Octo- ber-December. We used ANOSIM to make seasonal comparisons of the diets of bonefish separately for the two principal sampling areas. Six pairwise com- parisons were made among seasons for each area. No adjustment of significance levels exists for ANOSIM to account for the increased possibility of type-1 error associated with multiple comparisons (Clarke and Warwick, 1994). To determine feeding selectivity, we used ANOSIM to compare the species of crustaceans and fishes found in the stomachs of bonefish 480-699 mm long to those found in samples collected in the potential- prey environment. We included only bonefish col- lected during the same seasons as those when the throw-trap collections were made. For Florida Bay comparisons, bonefish collected during March, May, June, and September were included; for ocean-side comparisons, bonefish collected during January, Feb- ruary, August, September, and October were included (Table 1). Significant differences in the suite of crus- taceans and fishes found in bonefish stomachs and the potential prey available in the environment would imply selective feeding. We used SIMPER to indi- cate the percentage of dissimilarity contributed by each prey species and thus show which prey were selected or not selected. Taxa that were not selected Crabtree et al.: Feeding habits of Albula vulpes 757 could have been avoided, not preferred, or only inci- dentally ingested by bonefish. Alternatively, taxa that were not selected could have been preferred, but were able to evade capture. Results Stomachs of 385 bonefish that ranged in length from 228 to 702 mm contained prey (Fig. 1) consisting mostly of small benthic and epibenthic organisms (Table 2). Stomachs of 67 of the bonefish we exam- ined were empty. Decapods and teleosts dominated the diet by weight, but gastropods and bivalves were among the most speciose prey categories. Relatively few prey species made up most of the diet by weight; xanthid crabs ( W=29.9%), the gulf toadfish, Opsanus beta (W=17.2%), portunid crabs (W=10.9%), alpheid shrimp ( W=9.2%), and penaeid shrimp (W=7.7%) to- gether made up 74.9% of the diet. At least 17 fami- lies and 24 species of gastropods and 9 families and 16 species of bivalves were recognized, but gastro- pods made up only 2.7% of the diet by weight and bivalves made up only 2.5%. Polychaetes, represented by at least seven families, were important numerically (Af=37. 1%) but made up only 1.4% of the diet by weight. Both the cluster analysis and the MDS ordination grouped bonefish stomach contents according to fish length (Fig. 2). Cluster analysis organized the 23 length groups into two principal clusters that were linked at a level of similarity greater than 20% and one outlying group of bonefish 280-299 mm long (group 2) that was linked to the other clusters at a Fork length (mm) Figure 1 Length-frequency distribution of 384 bonefish, Albula vulpes, whose stomachs contained recognizable prey. The tail of one of the 385 bonefish whose stomach contents were examined was eaten by a shark during capture, so that fish was not measured. level of similarity less than 10%. One principal clus- ter contained bonefish 260 to 439 mm long, and a second principal cluster contained mostly larger fish 400 to 702 mm long. Bonefish in the 400-419 mm length interval (group 8) were classified with larger fish, and bonefish in the 420-439 mm length inter- val (group 9) were classified with smaller bonefish. In Table 3, stomach contents are summarized sepa- rately for small bonefish (<440 mm) and large bone- fish (>439 mm) on the basis of cluster analysis, but we reassigned bonefish stomachs in the 400-419 mm length interval (group 8) into the <440 mm group to avoid any overlap in the table between lengths of small (<440 mm) and large (>439 mm) bonefish. In the SIMPER analysis (Table 4), bonefish stomachs were grouped according to the cluster analysis shown in Figure 2, and no groups were reassigned. Levels of similarity among stomach contents of the 20-mm length groups within the 260-439 mm cluster were less overall than the levels of similarity of stomach contents of the 20-mm length groups of bonefish in the 400-699 mm cluster. Stomach contents of bone- fish length groups ranging from 480 to 699 mm had a high level of similarity (>55%) and were tightly grouped in the MDS; we chose this length group as the basis for all other statistical comparisons. SIMPER analysis suggested that much of the dis- similarity between the two principal length clusters was due to xanthid crabs, penaeid shrimp, alpheid shrimp, and O. beta (Table 4). The large values of the ratios (5; /SD(5;)) between the mean contribution (S-) to the overall level of dissimilarity and the stan- dard deviation (SD) of the 5( values across all stom- achs suggest that these taxa consistently contributed to the level of dissimilarity, and so they are probably reliable discriminating prey taxa characteristic of one or the other length clusters. Bonefish longer than 439 mm consumed more decapods (alpheid shrimp, xanthid crabs, and Callinectes spp.) and teleosts than smaller bonefish (Tables 3 and 4). The most striking difference was in the consumption of teleosts, prin- cipally O. beta, which was not eaten by bonefish shorter than 440 mm but made up 17.8% of the diet of bonefish longer than 439 mm. Penaeid shrimp made up a larger proportion of the diet of small bone- fish (W=40.5) than large bonefish (W=6.7), but they were eaten by bonefish of all sizes. Portunid crabs were eaten in nearly equal amounts by both length groups of bonefish, but this finding is misleading because all the Portunus spp. eaten by small bone- fish were eaten by a single individual collected in Florida Bay; thus the importance of portunid crabs in the diet of small bonefish is probably less than what is suggested in Table 3. No crabs of the genus Callinectes were eaten by small bonefish. 758 Fishery Bulletin 96(4), 1998 Table 2 Food items found in stomachs of bonefish, Albula vulpes , caught in the waters of the Florida Keys (n=385). W = percent weight, F = percent frequency of occurrence, N = percent numerical abundance. Taxon and prey item W F N Taxon and prey item W F N Plant material 0.6 34.5 Olividae Algae <0.1 0.5 — Unidentified Olividae <0.1 0.8 <0.1 Halimeda <0.1 8.1 — Jaspidella jaspidea 0.2 8.3 0.7 Unidentifiable seagrass 0.1 8.1 — Marginellidae Halodule <0.1 7.0 — Unidentified Marginellidae <0.1 1.3 <0.1 Syringodium <0.1 1.6 — Prunum apicinum 0.3 8.1 1.7 Thalassia 0.3 19.2 — Prunum sp. <0.1 1.3 <0.1 Annelida Volvarina avena <0.1 1.6 0.1 Total Polychaeta 1.4 39.5 37.1 Cystiscidae Unidentified Polychaeta 0.2 6.8 1.7 Persicula catenata <0.1 1.6 <0.1 Amphinomidae 0.3 1.3 <0.1 Persicula pulcherrima <0.1 0.3 <0.1 Lumbrinereidae Conidae Lumbrinereis spp. <0.1 0.8 0.7 Conus stearnsi <0.1 0.3 <0.1 Opheliidae 0.7 30.1 33.0 Bullidae Spionidae <0.1 0.8 0.4 Bulla striata <0.1 2.3 0.7 Orbiniidae <0.1 1.3 0.9 Nudibranchia <0.1 0.3 <0.1 Capitellidae Total Bivalvia 2.5 24.9 2.9 Unidentified Capitellidae <0.1 1.6 0.1 Unidentified Bivalvia 0.5 5.2 0.4 Dasybranchus spp. <0.1 0.5 <0.1 Mytilidae Sabellidae <0.1 1.0 0.1 Brachidontes modiolus <0.1 0.5 0.3 Mollusca Pteriidae Unidentified Mollusca 1.6 15.6 0.8 Pinctada imbricata <0.1 0.5 <0.1 Total Gastropoda 2.7 31.2 5.9 Limidae Unidentified Gastropoda 0.3 8.6 0.6 Limaria pellucida 0.1 3.9 0.6 Acmaeidae Pectinidae Unidentified Acmaeidae <0.1 0.3 <0.1 Unidentified Pectinidae <0.1 0.8 <0.1 Patelloida pustulata <0.1 0.3 <0.1 Argopecten irradians 0.5 2.1 0.1 Trochidae Argopecten spp. <0.1 0.8 <0.1 Tegula fasciata <0.1 0.3 <0.1 Lucinidae Turbinidae Unidentified Lucinidae 0.2 1.6 0.1 Turbo castanea <0.1 0.3 <0.1 Codakia orbicularis <0.1 0.3 <0.1 Eulithidium affine <0.1 1.3 0.5 Codakia orbiculata <0.1 0.3 <0.1 Turritellidae Carditidae Torcula acropora <0.1 0.3 <0.1 Carditamera floridana <0.1 1.3 <0.1 Modulidae Cardiidae Modulus modulus <0.1 1.8 0.1 Unidentified Cardiidae <0.1 0.5 <0.1 Cerithiidae Americardia media <0.1 0.3 <0.1 Unidentified Cerithiidae <0.1 1.0 <0.1 Laevicardium mortoni <0.1 0.3 <0.1 Cerithium eburneum <0.1 0.8 <0.1 Trachycardium muricatum <0.1 0.3 <0.1 Cerithium muscarum <0.1 0.5 <0.1 Tellinidae Triviidae Unidentified Tellinidae <0.1 1.0 <0.1 Trivia quadripunctata <0.1 0.5 <0.1 Strigilla carnaria <0.1 0.3 <0.1 Naticidae Tellina fausta <0.1 0.3 0.1 Natica canrena <0.1 0.8 <0.1 Tellina similis <0.1 3.9 0.3 Columbellidae Tellina tampaensis <0.1 0.5 0.1 Unidentified Columbellidae <0.1 0.3 <0.1 Tellina spp. <0.1 1.3 <0.1 Anachis avara <0.1 1.0 <0.1 Veneridae Columbella rusticoides <0.1 0.3 <0.1 Chione cancellata 0.6 6.5 0.4 Zafrona taylorae <0.1 0.5 <0.1 Transennella conradina <0.1 0.3 <0.1 Nassariidae Crustacea Nassarius sp. <0.1 0.3 <0.1 Unidentified Crustacea <0.1 1.6 <0.1 Fasciolariidae Copepoda <0.1 0.3 <0.1 Leucozonia nassa <0.1 0.3 <0.1 Total Stomatopoda 2.0 10.9 1.0 Fasciolaria tulipa 1.5 6.8 0.6 Unidentified Stomatopoda 1.0 5.7 0.6 Fasciolaria spp. <0.1 0.5 <0.1 Pseudosquilla ciliata 1.0 5.2 0.3 continued Crabtree et al.: Feeding habits of Albu la vulpes 759 There was a significant positive correlation between prey size and bonefish length for xanthid crabs (n=286, r=0.262, PcO.OOl) and portunid crabs (rc=58, r=0.465, P<0.001), but not for alpheid shrimp (P^O.630), penaeid shrimp (P=0.063), or O. beta (P=0.782). The largest prey consumed were O. beta (largest 113 mm SL), the portunid crab Callinectes sapidus (largest 106 mm cara- pace width), and penaeid shrimp (largest 100 mm TL; Fig. 3). Most of these relatively large animals were con- sumed only as juveniles by bonefish. The stomach contents of bonefish caught in Florida Bay were significantly different from those of bone- fish caught on the ocean side of the Keys ( ANOSIM, P=0.034; P=0.013). The statistic R can range from -1 to 1 with a value of 1 indicating that all replicates within a sample are more similar to each other than to any replicates from the other samples and with a value of 0 indicating that the similarities between and within samples are on average equal (Clarke and Warwick, 1994). Although the R value of 0.034 was Table 2 (continued) Taxon and prey item W F N Taxon and prey item W F N Total Decapoda 67.8 88.6 42.1 Echinodermata Unidentified Dendrobranchiata 0.2 8.3 0.9 Ophiuroidea 0.5 1.8 1.0 Penaeidae Holothuroidea 0.5 2.9 0.2 Unidentified Penaeidae 0.9 4.9 1.2 Ascidiacea <0.1 0.3 <0.1 Penaeus spp. 3.8 17.9 2.9 Chordata Penaeus duorarum 3.0 5.5 1.1 Total Teleostei 20.5 44.9 4.9 Palaemonidae Unidentified teleostei 0.5 10.1 0.6 Unidentified Palaemonidae <0.1 8.6 1.1 Ophicthidae Brachycarpus biunguiculatus <0.1 0.8 <0.1 Unidentified Ophicthidae 0.2 1.3 0.1 Alpheidae Ahlia egmontis 0.7 2.9 0.2 Unidentified Alpheidae 0.9 9.4 1.8 Myrophis punctatus 0.7 2.6 0.1 Alpheus floridanus <0.1 1.0 <0.1 Engraulidae Alpheus normanni 7.6 35.3 13.2 Unidentified Engraulidae <0.1 0.3 <0.1 Alpheus spp. 0.6 6.8 0.9 Anchoa sp. <0.1 0.3 <0.1 Hippolytidae Batrachoididae Unidentified Hippolytidae 0.1 4.7 1.2 Opsanus beta 17.2 29.1 3.1 Thor spp. 0.3 19.7 5.6 Bythitidae Tozeuma spp. <0.1 0.5 <0.1 Ogilbia cayorum <0.1 0.8 <0.1 Palinuridae Cyprinodontidae Panulirus argus 0.6 0.5 <0.1 Floridichthys carpio <0.1 0.3 <0.1 Unidentified Brachyura 3.8 17.7 1.0 Lucania parva <0.1 1.0 <0.1 Majidae Syngnathidae Unidentified Majidae 3.0 6.2 0.6 Unidentified Syngnathidae <0.1 1.6 <0.1 Pitho mirabilis 0.2 0.3 <0.1 Hippocampus zosterae <0.1 0.3 <0.1 Pitho spp. 0.7 2.3 0.2 Hippocampus sp. <0.1 0.3 <0.1 Portunidae Syngnathus floridae <0.1 0.5 <0.1 Unidentified Portunidae 2.9 8.1 0.6 Syngnathus scovelli <0.1 0.3 <0.1 Callinectes ornatus 1.3 1.0 0.2 Syngnathus spp. 0.1 1.6 0.1 Callinectes sapidus 2.5 1.8 0.2 Lutjanidae Callinectes spp. 3.1 7.0 0.5 Lutjanus griseus 0.1 0.3 <0.1 Portunus spp. 1.1 1.0 0.5 Scaridae 0.2 0.8 <0.11 Xanthidae Gobiidae Unidentified Xanthidae 29.9 49.6 8.2 Unidentified Gobiidae <0.1 1.3 <0.1 Neopanope sp. <0.1 0.3 <0.1 Gobiosoma robustum <0.1 0.3 <0.1 Panopeus spp. 0.1 0.8 0.1 Balistidae Grapsidae Monacanthus ciliatus <0.1 0.3 <0.1 Unidentified Grapsidae <0.1 1.0 0.1 Monacanthus hispidus 0.2 0.8 <0.1 Sesarma sp. <0.1 0.3 <0.1 Ostraciidae Palicidae Lactophrys sp. 0.1 0.3 <0.1 Unidentified Palicidae <0.1 0.3 <0.1 Miscellaneous material — 28.1 — Mysidae <0.1 0.5 <0.1 Nonfood material Tanaidacea <0.1 2.9 4.0 sandy debris — 1.8 — Isopoda <0.1 0.3 <0.1 coral rock — 0.3 — 760 Fishery Bulletin 96(4), 1 998 significantly different from zero, the difference was small and thus the differences between the stomach contents of Florida Bay and ocean-side bonefish were probably slight. Xanthid crabs, alpheid shrimp, O. beta, penaeid shrimp, and Callinectes spp. together made up over 50% of the dissimilarity between the two areas (Table 5). Although these taxa contributed to the overall level of dissimilarity, the ratios (8 / SD(8;)) between the mean contribution (5;) to the overall level of dissimilarity and the standard devia- tion of the 8 values across all stomachs were low for each prey taxa. Thus, these taxa did not consistently contribute to the level of dissimilarity, and are prob- ably not reliable discriminating prey taxa character- istic of either area. In both areas, the same prey taxa dominated the diet (Table 6). A seasonal effect on feeding was found for ocean-side bonefish Bray-Curtis similarity 0. 20. 40. 60. 80. FL GP n 290 2 2 310 3 6 270 1 1 330 4 5 350 5 11 370 6 10 390 7 10 430 9 5 410 8 10 450 10 7 470 11 10 490 12 10 510 13 12 530 14 17 550 15 22 670 21 24 590 17 48 610 18 41 570 16 33 650 20 41 630 19 43 690 22 15 710 23 2 100. Figure 2 Dendrogram and multidimensional scaling (MDS) ordination showing similari- ties in diet among bonefish, Albula vulpes, grouped into 20-mm length intervals. Stress for the MDS plot = 0.13. FL = median fork length of the 20-mm length intervals, GP = group number used on the MDS ordination, and n = the number of bonefish in each 20-mm length interval. The vertical bars to the right of the cluster show the two principal length groupings referred to in the text and the 480-699 mm length grouping used for all statistical comparisons. The circled groups in the MDS ordination correspond to the principal groupings shown on the cluster dendrogram and referred to in the text. (ANOSIM, 72=0.069; P=0.002) but not for bonefish collected from Florida Bay (ANOSIM, 72=0.066; P=0.080). For both tests, the R val- ues were close enough to zero to suggest that any seasonal differ- ences in diet were slight. On the ocean side of the Keys, pairwise tests between bonefish caught dur- ing January-March and those caught during all other seasons were significant at P<0.05; no other pairwise seasonal compari- sons were significant. Xanthid crabs, alpheid shrimp, brachyuran crabs (excluding xanthids, por- tunids, and majids), O. beta, penaeid shrimp, and stomatopods accounted for most of the dissimi- larity between stomach contents of bonefish collected during Janu- ary-March and those of bonefish collected during other seasons (Table 7). The ratios (8-/SD(8.)) be- tween the mean contribution (8;) to the overall level of dissimilarity and the standard deviation of the values across all stomachs were low for each prey taxa. No taxa consistently contributed to the level of dissimilarity, and there were no reliable discriminating prey taxa characteristic of any par- ticular season. Variable sample sizes between seasons in both ar- eas reduced the power of our sea- sonal comparisons; most ocean- side bonefish were caught during January-May, and most Florida Bay bonefish were caught during June-December. Only six stom- achs were examined from ocean- side bonefish captured during July-September. Although the stomach contents of these six bone- Crabtree et a!.: Feeding habits of Albu !a vulpes 761 Tab!e 3 Food items found in stomachs of bonefish, Albula vulpes , caught in the waters of the Florida Keys ( m =384 ) by bonefish length interval. W = percent weight, F = percent frequency of occurrence, N = percent numerical abundance. <440 mm FL fn=611) >439 mm FL (n=323) laxon ana prey item W F N W F N Annelida Polychaeta 3.06 19.67 9.54 1.34 43.34 40.08 Mollusca Unidentified Mollusca 2.29 19.67 1.82 1.56 14.24 0.67 Gastropoda 4.99 21.31 13.18 2.53 32.82 5.07 Bivalvia 4.64 22.95 3.51 2.39 25.70 2.81 Crustacea Stomatopoda 2.26 3.28 2.52 2.03 12.38 0.79 Decapoda Penaeidae 40.45 36.07 11.50 6.68 26.01 4.37 Alpheidae 0.35 8.20 1.12 9.36 51.70 17.52 Hippolytidae 1.87 11.48 7.99 0.33 26.01 6.76 Majidae 3.28 8.20 0.70 3.92 8.98 0.89 Portunidae 15.64 8.20 3.37 10.91 19.50 1.83 Xanthidae 7.43 13.11 1.68 31.49 56.97 8.94 Chordata Teleostei 3.90 19.67 3.09 21.10 49.54 5.07 Batrachoididae Opsanus beta 0.00 0.00 0.00 17.83 34.67 3.43 fish were significantly different from those of bone- fish collected in January-March ( ANOSIM, /?=0.284, P=0.018), we have little confidence in this test be- cause of the small sample size, and these results are not included in Table 7. Bonefish fed selectively on some prey groups but did not select others. The suite of epibenthic crusta- ceans and fishes found in bonefish stomachs was sig- nificantly different from that collected with throw traps both on the ocean side of the Florida Keys (ANOSIM, i?=0.261, P<0.001) and in Florida Bay (ANOSIM, f?=0.419, P<0.001). Bonefish on the ocean side of the Florida Keys fed selectively on alpheid shrimp, xanthid crabs, P. duorarum , and O. beta , whereas they did not select the small but abundant crustaceans Thor spp. and Periclimenes americanus (Table 8). Similarly, Florida Bay bonefish fed selec- tively on xanthid crabs, alpheid shrimp, O. beta, P. duorarum, and Callinectes spp. but did not select the abundant but small crustaceans Thor spp., Hippolyte zostericola, and P. americanus, as well as the abun- dant goby Gobiosoma robustum (Table 9). Discussion A variety of factors could have biased our descrip- tion of the diet of bonefish. Some prey, particularly soft-bodied prey, may have been digested more rap- idly than others with bony or chitinous skeletons. Consequently, we may have underestimated the im- portance of soft-bodied organisms such as polychae- tes. Furthermore, bonefish have massive pharyngeal tooth plates capable of crushing shells and other hard structures. If bonefish are able to expel the crushed shells of mollusks and swallow only the soft-bodied organism, then we could have underestimated the importance of mollusks. This might explain why mollusks were relatively unimportant in our samples. Our sample consisted principally of bonefish caught with hook-and-line gear; therefore most of the fish we examined were probably actively foraging or they would not have consumed the bait presented by an- glers. We do not believe that the number of fish with empty guts in our sample reflects the number of fish in the area that were not feeding, therefore we did not attempt to evaluate temporal feeding patterns. We cannot eliminate the possibility that some bone- fish regurgitated prey during capture trauma; if some prey taxa were more likely to be regurgitated than others, this could have biased our results. Most of the bonefish in our sample came from relatively shal- low (<2 m) grass, sand, or hard-bottom flats, but be- cause the fish were caught by anglers, we did not have corresponding data on bottom type for each fish. Colton and Alevizon (1983) found differences in the 762 Fishery Bulletin 96(4), 1998 Table 4 Breakdown into the most important prey groups of the mean dissimilarity between stomach contents (percent weight) of bonefish, Albula vulpes, from the two principal clusters shown in Figure 2. Small bonefish ranged from 260 to 439 mm FL and large bonefish ranged from 400 to 699 mm FL. Prey groups are listed in order of decreasing contribution to the overall dissimilarity between the two bonefish length groups. is the mean contribution of the ith species to the dissimilarity between the two groups, 5 / SD( 5; ) is the ratio between the mean contribution of the ith species (5;) and the standard deviation of the values for that species [SD(8;)], 5: % is the contribution to the total dissimilarity scaled as a percentage, and Cum 8; % is the cumulative contribution to the total dissimilarity scaled as a percentage. Taxa that are likely to be reliable discriminators of the two length groups are indicated by ** in the 5 /SD(8 ) column. Taxa proportionally more important in the diet of large bonefish than small bonefish are shown in bold type. Species 8, S/SD(8.) 8, % Cum 8^ % Penaeidae 8.13 1.94** 10.47 10.47 Xanthidae 7.97 2.48** 10.26 20.73 Alpheidae 5.29 1.84** 6.81 27.54 O. beta 5.28 1.84** 6.80 34.34 Portunidae (unidentified) 4.62 1.26 5.95 40.29 Brachyura7 3.68 1.37 4.73 45.02 Majidae 3.25 1.05 4.19 49.21 Portunus spp. 2.80 0.45 3.60 52.81 Stomatopoda 2.51 1.16 3.23 56.03 Callinectes spp. 2.50 0.97 3.23 59.26 1 Excluding xanthids, portunids, and majids. stomach contents of Bahamian bonefish collected over different bottom types, and this variation prob- ably also occurs in the Florida Keys. There was also no evidence that bonefish do not feed in deeper wa- ters than those traditionally fished by anglers; prey availability and bonefish feeding may be quite dif- ferent at greater depths than in the shallow waters we sampled. Most (77%) of the fish in our sample were longer than 500 mm (Fig. 1); consequently, the diet of large bonefish is better described by our data than that of small bonefish. The inadequacy of our description of the diet of small bonefish is reflected in the low lev- els of similarity among 20-mm length intervals of bonefish smaller than 480 mm (Fig. 2). Many of the length intervals smaller than 500 mm contained few fish and resulted in greater variation in diet among length intervals and probably caused the lower lev- els of similarity among 20-mm length groups of small bonefish than among large fish. The changes in diet as length of bonefish increased in general reflect the expansion of the diet to include Table 5 Breakdown into the most important prey groups of the mean dissimilarity between stomach contents (percent weight) of bonefish, Albula vulpes (480-699 mm FL), caught on the ocean side of the Florida Keys (n=50) and in Florida Bay (n= 50). Prey groups are listed in order of de- creasing contribution to the overall dissimilarity between the two study areas. Taxa proportionally more important in the diet of ocean-side bonefish than Florida Bay bone- fish are shown in bold type. The low values of 8( /SD(8;) suggest that the data were variable and that no taxa were reliable discriminators of either area. Symbols are ex- plained in the legend of Table 4. Species 5, 8/SD(8,) 8; % Cum 8; % Xanthidae 13.47 1.15 17.17 17.17 Alpheidae 8.56 1.06 10.91 28.08 O. beta 8.04 0.82 10.24 38.32 Penaeidae 7.49 0.82 9.55 47.87 Callinectes spp. 4.53 0.46 5.78 53.65 larger prey such as O. beta and crabs of the genus Callinectes. In some cases, for example xanthid and portunid crabs, prey size was positively correlated with predator length. In other cases, for example O. beta, prey were not eaten at all by small bonefish, and there was no correlation between prey size and predator length among large bonefish. Small but abundant crustaceans such as Thor spp., H. zoster- icola, and P. americanus were not important in the diet of the bonefish we examined. These small crusta- ceans may be eaten by smaller bonefish (<228 mm), but they are apparently outside of the size range of prey typically consumed by the size of bonefish we considered. Bruger (1974) examined the stomachs of 129 bone- fish ranging from 221 to 679 mm FL (reported as 210 to 656 mm SL) collected from the waters off the Keys and reported the frequency of occurrence of crustaceans in the diet. Of these 129 stomachs, 19 were empty. We recalculated his frequency of occur- rences on the basis of only the number of stomachs with prey (n = 110) to compare with our frequency data: the recalculated results were 85% crustaceans, 33% mollusks, and 17% teleosts. These results are in general agreement with our findings of 89% crus- taceans, 51% mollusks, and 45% teleosts, except that our samples contained more teleosts than Bruger’s. Among crustacean prey, Bruger’s results resemble ours; penaeid shrimp, alpheid shrimp, portunid crabs, and xanthid crabs were the most frequently occurring crustacean prey. Although Bruger did not quantify by species the fishes found in stomachs, he did not include O. beta in his list of teleosts eaten by Crabtree et al.: Feeding habits of Albula vulpes 763 £ 6 3 z Total length (mm) Total length (mm) Carapace width (mm) Standard length (mm) Carapace width (mm) Figure 3 Size-frequency distributions of some prey important in the diet of bonefish, Albula vulpes. bonefish. In contrast, O. beta was the teleost most frequently eaten by the bonefish we examined (F= 29.1%). Most of Bruger’s bonefish were collected in the lower Keys between Marathon and Key West, but most of our bonefish were captured in the upper Keys from Marathon north to Key Biscayne and in- cluding Florida Bay. Although we have no data on prey availability in the lower Keys, habitat differ- ences between the two study areas could account for some of the differences between our results and Bruger’s. Colton and Alevizon ( 1983) examined the stomach contents of 365 Bahamian bonefish ranging from 268 to 652 mm FL (reported as 256 to 630 mm SL). Bivalves made up 39.2% of the diet of Bahamian bone- fish by dry weight, but they made up only 2.5% of the diet of Keys bonefish by weight. Bivalves were the most important prey of Bahamian bonefish both in terms of dry weight (39.2%) and frequency of oc- currence (66.3%); portunid (W=20.1%; F=40.5%) and xanthid crabs ( W=15.0%; F=24.8%) were also impor- tant. Teleosts (O. beta and Bathygobius soporator ; pooled W=4.9%), alpheid shrimp ( W=4.6%), Pseudos- quilla ciliata (W=3.2%), polychaetes (W=3.2%), gas- tropods ( W=2.4%), and Penaeus duorarum ( W=1.6%) occurred in 15-25% of the guts that Colton and Alevizon examined but made up little of the diet in terms of dry weight. The most notable difference between Keys and Bahamian bonefish diets was the greater importance of O. beta in the diet of Keys bon- efish ( W=17.2%). Colton and Alevizon ( 1983 ) reported length-related changes in the diet of Bahamian bonefish that were similar to those that we observed in the Florida Keys. They found that bonefish larger than 416 mm FL (400 mm SL) ate more xanthid and majid crabs, alpheid shrimp, and teleosts than smaller bonefish did. Teleosts (gobiids, batrachoidids, ophichthids, and small lutjanids) were found principally in stomachs from bonefish larger than 575 mm FL (555 mm SL). In contrast to our conclusions, Colton and Alevizon ( 1983 ) found that small bonefish (<416 mm ) ate more portunid crabs (Callinectes ornatus ) than large bone- fish did; we found no Callinectes spp. in any bone- 764 Fishery Bulletin 96(4), 1998 Table 6 Food items found in stomachs of bonefish, Albula vulpes , caught in Florida Bay (rc=130) and on the ocean side (n = 144) of the Florida Keys. Stomachs from all bonefish 480 to 699 mm FL collected during all months are included. W = percent weight; F = percent frequency of occurrence; N = percent numerical abundance. Taxon and prey item Florida Bay Ocean side W F N W F N Annelida Polychaeta 0.62 34.62 34.22 1.63 52.08 32.72 Mollusca Gastropoda 1.86 33.85 4.94 2.83 32.64 3.52 Bivalvia 2.68 28.46 3.00 2.02 24.13 1.85 Crustacea Stomatopoda 0.35 3.85 0.24 3.85 19.44 0.73 Decapoda Penaeidae 8.30 35.38 6.72 4.30 15.97 1.35 Alpheidae 5.13 40.77 11.93 14.12 60.42 13.16 Hippolytidae 0.38 32.31 11.38 0.33 23.61 2.81 Majidae 1.97 4.62 0.28 4.45 9.72 0.67 Portunidae 17.06 23.85 2.96 5.00 14.58 0.60 Xanthidae 33.53 63.08 10.94 27.78 52.78 4.64 Chordata Teleostei 21.25 56.92 7.43 24.87 43.75 2.31 Batrachoididae Opsanus beta 18.94 48.46 5.85 19.87 25.00 1.19 fish smaller than 440 mm, although crabs of the genus Portunus were eaten in large numbers by one 435-mm bonefish. There was evidence of a seasonal effect on diet, but small sample sizes during some seasons in each of the respective sampling areas reduced our ability to detect signifi- cant differences. Colton and Alevizon (1983) also found seasonal differences in feeding in Bahamian bonefish. Bivalves were eaten more during the summer by bonefish of all lengths, whereas small bone- fish (<416 mm) ate more portunid crabs during the winter. They also noted habi- tat-related differences in bonefish feeding. Penaeid shrimp were eaten almost exclu- sively by bonefish caught over grassy bot- tom and not by those caught over sandy bottom. Bonefish caught over sandy bottom ate relatively more crabs and bivalves than did bonefish caught over grassy areas. Warmke and Erdman (1963) examined the stomach contents of 56 bonefish rang- ing from 292 to 663 mm FL (reported as 0.75 to 10.25 pounds) from Puerto Rican waters and, like Colton and Alevizon Table 7 Breakdown into the most important prey groups of the mean dissimilarity between stomach contents (percent weight) of bonefish, Albula vulpes (480-699 mm FL), col- lected on the ocean side of the Florida Keys during Janu- ary-March (n=39), April-June (n=43), and October-Decem- ber (n= 33). Prey groups are listed in order of decreasing contribution to the overall dissimilarity between the sea- sonal samples. Taxa proportionally more important in the diet of bonefish collected during January-March than dur- ing other seasons are shown with bold type. The low val- ues of 8-/SD(8-) suggest that the data were variable and that no taxa were reliable discriminators of any particular season. Symbols are explained in the legend of Table 4. Species 5, 8/SD<8,) 5, % Cum 5, % Jan-Mar vs. Apr-Jun Xanthidae 12.44 1.16 16.02 16.02 Alpheidae 11.77 1.28 15.16 31.18 Brachyura7 6.71 0.69 8.64 39.81 O. beta 5.41 0.62 6.97 46.78 Stomatopoda 5.18 0.64 6.67 53.45 Jan-Mar vs. Oct- -Dec Alpheidae 11.52 1.27 14.52 14.52 Xanthidae 10.99 0.99 13.86 28.38 Brachyura7 7.26 0.75 9.16 37.54 Penaeidae 4.97 0.76 6.27 43.81 O. beta 4.90 0.56 6.18 50.00 1 Excluding xanthids, portunids, and majids. (1983), found that mollusks were the most important prey. Stomachs of Puerto Rican bonefish contained 56% mollusks, 42% crustaceans, and 2% other prey types by volume. In contrast, we found that mollusks accounted for only about 7% of the diet of Keys bonefish by weight and that crustaceans ac- counted for about 70% of the diet. Teleosts were part of Warmke and Erdman’s “other” classification and made up less than 2% of the diet in their study; in the Keys, teleosts made up over 20% of the diet by weight. Warmke and Erdman identified only mol- lusks to species. The most important mollusk they found was the bivalve Codakia costota, which oc- curred in 62% of the stomachs they examined. Codakia orbicularis and C. orbiculata occurred in stomachs from Keys bonefish but made up less than 1% of the diet by numbers or weight. Codakia costata was not found in Keys bonefish stomachs and was not reported to occur in Florida Bay by Turney and Perkins (1972). There were slight but significant differences be- tween the diets of bonefish from Florida Bay and those from the ocean side of the Keys. These differ- ences may reflect differences in prey availability in the two areas, but overall the dominant prey eaten by bonefish was the same in the two areas. The fauna of Florida Bay has been characterized as Gulf-Caro- linian in nature, whereas that of the Keys ocean side is Antillean (Sogard et al., 1987; Holmquist et al., Crabtree et al.: Feeding habits of Albula vulpes 765 Table 8 Breakdown into the most important prey groups of the mean dissimilarity between stomach contents (percent number) of bonefish, Albula vulpes (480-699 mm FL), caught on the ocean side of the Florida Keys (/r =39 ) and throw-trap samples (n= 54) from the ocean side of the Florida Keys. Prey groups are listed in order of decreasing contribution to the overall dissimilarity between the two samples. Taxa proportionally more important in the diet of bonefish than suggested by their proportional abundance in throw-trap samples are shown with bold type. The low values of 5- /SD( 8; ) suggest that the data were variable and that no taxa were reliable discriminators of either sample source. Symbols are explained in the legend of Table 4. Species 5/SD(8,) 5, % Cum 8- % Alpheidae 10.58 1.31 13.87 13.87 Xanthidae Periclimenes 7.89 1.13 10.35 24.23 americanus 6.83 0.93 8.96 33.19 Thor spp. 6.11 0.97 8.02 41.20 P. duorarum 5.95 0.88 7.81 49.01 O. beta 4.61 0.80 6.05 55.06 1989a, 1989b). Our sampling effort was over a large and diverse area, and this limited our ability to re- solve area-specific differences in bonefish diet. Some Florida Bay areas that we sampled were near passes leading to ocean-side flats and may have more closely resembled ocean-side areas than some of the more remote areas in Florida Bay where we occassionally caught bonefish. Larger sample sizes, more inten- sive sampling of specific areas along with site-spe- cific descriptions of habitat types, and sampling of prey availability concurrent with bonefish collections are needed to better describe spatial variation in the diet of Keys bonefish. Comparisons of the stomach contents of bonefish collected in Florida Bay and ocean-side areas as well as seasonal comparisons were complicated by the variable monthly sample sizes from the two areas. We excluded over half of the bonefish in our sample from our area comparisons because seasonal sample sizes from the two areas were greatly unequal. The variable sample sizes from the two areas reflect gen- eral seasonal trends in bonefish availability in the two areas. Bonefish are typically most abundant in Florida Bay during summer and fall. Winter cold fronts tend to reduce Florida Bay temperatures more than ocean-side temperatures (Hudson et al., 1976; Roberts et al., 1982; Chiappone, 1996), and many productive summer-fall fishing areas in Florida Bay rarely hold bonefish during winter and spring be- cause bonefish move to ocean-side areas with more moderate temperatures and closer proximity to deep Table 9 Breakdown into the most important prey groups of the mean dissimilarity between stomach contents (percent number) of bonefish, Albula vulpes (480-699 mm FL), caught in Florida Bay (n= 45) and throw-trap samples (tj=30) from Florida Bay (Matheson et al.1). Prey groups are listed in order of decreasing contribution to the overall dissimilarity between the two samples. Taxa that are likely to be reliable discriminators of the two samples are indi- cated by ** in the 5/SD(5;) column. Taxa proportionally more important in the diet of bonefish than suggested by their proportional abundance in throw-trap samples are shown with bold type. Symbols are explained in the leg- end of Table 4. Species 5[ S/SDtS,) 8| % Cum 8| % Thor spp. 18.13 1.97** 23.03 23.03 Xanthidae 9.63 1.17 12.24 35.27 Alpheidae 7.61 1.23 9.67 44.94 O. beta 7.51 1.15 9.55 54.49 P. duorarum Hippolyte 5.17 0.81 6.57 61.06 zostericola Periclimenes 5.00 1.65** 6.36 67.42 americanus 4.55 1.85** 5.78 73.20 Callinectes spp. Gobiosoma 3.23 0.48 4.10 77.30 robustum 3.03 1.98** 3.85 81.15 water. Thus, most of our Florida Bay bonefish were captured during summer and fall, and most ocean- side bonefish were caught during winter and spring. Seagrass die-offs have recently been documented in Florida Bay (Robblee et al., 1991; Carlson et al., 1994; Durako, 1994; Butler et al., 1995). Anecdotal evidence suggests that changes in the Everglades ecosystem have caused a decline in the quality of Fish- ing in Florida Bay and the waters of the Florida Keys (Chiappone and Sulka, 1996). If changes in the benthic epifauna and infauna have resulted from the seagrass die-off, these changes could potentially af- fect feeding and occurrence of bonefish in Florida Bay. Data on the species composition and abundance of epifaunal crustaceans and fishes collected subse- quent to the sea grass die-off and the studies of Sogard et al. (1987, 1989) and Holmquist et al. (1989a, 1989b) prior to the seagrass die-off suggest little evidence of declines in populations of impor- tant bonefish prey species (Matheson et al.1). One significant change reported by Matheson et al.1 was an increase in the abundance of O. beta in some ar- eas of Florida Bay since the 1980s. Whether the in- creased abundance of O. beta compared to that found in previous studies accounts for its greater promi- nence in stomachs of the bonefish we sampled is unknown. 766 Fishery Bulletin 96(4), 1998 Acknowledgments We thank Capt. John Kipp, who provided us with most of the bonefish examined in this study and with- out whose efforts this work would not have been pos- sible. We thank Capt. Mike Collins of the Florida Keys Fishing Guides Association, the staff of the Islamorada bonefish tournaments for their support, and Chris Harnden, Jim Colvocoresses, John Hunt, and others at the South Florida Regional Labora- tory for their cooperation. We also thank David Camp, Bill Lyons, Ed Matheson, and Tom Perkins for confirming identifications of bonefish prey and Gil McRae for help with the PRIMER software pack- age. Llyn French, Judy Leiby, Ed Matheson, Gil McRae, Jim Quinn, and Dana Winkelman made help- ful comments that improved the manuscript. This work was supported in part by funding from the De- partment of the Interior, U.S. Fish and Wildlife Ser- vice, Federal Aid for Sportfish Restoration, Project Number F-59. Literature cited Bray, J. R., and J. T. Curtis. 1957. An ordination of the upland forest communities of southern Wisconsin. Ecol. Monogr. 27:325-349. Bruger, G. E. 1974. Age, growth, food habits, and reproduction of bone- fish, Albula uulpes, in South Florida waters. Fla. Mar. Res. Publ. 3, 20 p. Butler, M. J., IV, J. H. Hunt, W. F. Herrnkind, M. J. Childress, R. Bertelsen, W. Sharp, T. Matthews, J. M. Field, and H. G. Marshall. 1995. Cascading disturbances in Florida Bay, USA: cyanobacteria blooms, sponge mortality, and implications for juvenile spiny lobsters Panulirus argus. Mar. Ecol. Prog. Ser. 129:119-125. Carlson, P. J. Jr., L. A. Yarbro, and T. R. Barber. 1994. Relationship of sediment sulfide to mortality of Thalassia testudinum in Florida Bay. Bull. Mar. Sci. 54:733-746. Chiappone, M. 1996. Oceanography and shallow-water processes of the Florida Keys and Florida Bay. Site characterization for the Florida Keys National Marine Sanctuary and environs, vol. 2. Farley Court Publ., Zenda, WI, 86 p. Chiappone, M., and R. Sulka. 1996. Fishes and fisheries. Site characterization for the Florida Keys National Marine Sanctuary and environs, vol. 2. Farley Court Publ., Zenda, WI, 149 p. Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol. 18:117-143. Clarke, K. R., and R. M. Warwick. 1994. Change in marine communities: an approach to sta- tistical analysis and interpretation. Natl. Environ. Res. Council, U.K., 144 p. Colton, D. E., and W. S. Alevizon. 1983. Feeding ecology of bonefish in Bahamian waters. Trans. Am. Fish. Soc. 112:178-184. Crabtree, R. E., C. W. Harnden, D. Snodgrass, and C. Stevens. 1996. Age, growth, and mortality of bonefish, Albula uulpes, from the waters of the Florida Keys. Fish. Bull. 94:442- 451. Crabtree, R. E., D. Snodgrass, and C. W. Harnden. 1997. Maturation and reproductive seasonality in bonefish, Albula uulpes, from the waters of the Florida Keys. Fish. Bull. 95:456-465. Durako, M. J. 1994. Seagrass die-off in Florida Bay (USA): changes in shoot demographic characteristics and population dynamics in Thalassia testudinum. Mar. Ecol. Prog. Ser. 110:59-66. Holmquist, J. G., G. V. N. Powell, and S. M. Sogard. 1989a. Decapod and stomatopod communities of seagrass- covered mud banks in Florida Bay: inter- and intra-bank heterogeneity with special reference to isolated subenviron- ments. Bull. Mar. Sci. 44:251-262. 1989b. Decapod and stomatopod assemblages on a system of seagrass-covered mud banks in Florida Bay. Mar. Biol. 100:473-483. Hudson, J. H., E. A. Shinn, R. B. Halley, and B. Lidz. 1976. Sclerochronology, a tool for interpreting past environments. Geology 4:361-364. Robblee, M. B., T. R. Barber, P. R. Carlson, M. J. Durako, J. W. Fourqurean, L. K. Muehlstein, D. Porter, L. A. Yarbro, R. T. Zieman, and J. C. Zieman. 1991. Mass mortality of the tropical seagrass Thalassia testudinum in Florida Bay (USA). Mar. Ecol. Prog. Ser. 71:297-299. Roberts, H. H., L. J. Rouse, Jr., N. D. Walker, and J. H. Hudson. 1982. Cold-water stress in Florida Bay and the Bahamas: a product of winter cold-air outbreaks. J. Sediment. Petrol. 52:145-155. Sogard, S. M., G. V. N. Powell, and J. G. Holmquist. 1987. Epibenthic fish communities on Florida Bay banks: relations with physical parameters and seagrass cover. Mar. Ecol. Prog. Ser. 40:25-39. 1989. Spatial distribution and trends in abundance of fishes residing in seagrass meadows on Florida Bay mud- banks. Bull. Mar. Sci. 44:179-199. Turney, W. J., and B. F. Perkins. 1972. Molluscan distribution in Florida Bay: sedimentia III. Rosenstiel School of Marine and Atmospheric Science, Miami, FL. 37 p. Warmke, G. L., and D. S. Erdman. 1963. Records of marine mollusks eaten by bonefish in Puerto Rican waters. The Nautilus 76:115-120. Zieman, J. C., J. W. Fourqurean, and R. L. Iverson. 1989. Distribution, abundance and productivity of seagrasses and macroalgae in Florida Bay. Bull. Mar. Sci. 44:292-311. 767 AbStraCt-Variation in restriction sites of mitochondrial (mt)DNA was ex- amined from 444 greater amberjack ( Seriola dumerili) sampled from 11 off- shore localities in the northern Gulf of Mexico (Gulf) and along the U.S. south- east Atlantic coast (Atlantic). A total of 49 mtDNA haplotypes (genotypes) were detected. Percent nucleotide sequence divergence among the haplotypes ranged from 0.156 to 2.623 (mean ±SE=0.980 ±0.015). Nucleon diversity within samples ranged from 0.845 to 0.906, and inirapopu laticnai mtDNA diversities ranged (mean ±SD) from 0.483 ±0.370 to 0.619 ±0.419. The lat- ter did not differ significantly from one another. Homogeneity tests of mtDNA haplotype frequencies, FST values, analy- sis of molecular variance (AMOVA), and comparisons of pairwise dy r distances were consistent with the hypothesis that (at least) two subpopulations (stocks) of greater amberjack exist in U.S. waters: one in the northern Gulf and one along the U.S. Atlantic coast. The latter subpopulation includes in- dividuals from the Florida Keys. There was no evidence of phylogeographic structure among mtDNA haplotypes or among sample localities, suggesting either that restricted gene flow between the subpopulations is fairly recent or that gene flow between the two is rela- tively infrequent. No significant spatial autocorrelations of mtDNA haplotypes was found among samples of greater amberjack from the Gulf, indicating continuous gene flow across the north- ern Gulf. Long-term effective (female) population sizes of both subpopulations were estimated to be in the range of 90,000-95,000 individuals. The esti- mates were commensurate with esti- mates in other, economically important marine fish. Based on suggested differ- ences in mtDNA evolutionary rates be- tween homeothermic and poikilother- mic vertebrates, the effective (female) population sizes of both stocks of greater amberjack could be in the range of 500,000 to 1,000,000 individuals. Manuscript accepted 13 March 1998 Fish. Bull. 96:767-778 (1998). Population structure in greater amberjack, Seriola dumerili, from the Gulf of Mexico and the western Atlantic Ocean* John R. Gold Linda R. Richardson Center for Biosystematics and Biodiversity Department of Wildlife and Fisheries Sciences Texas A&M University College Station, Texas 77843-2258 E-mail address: goldfish@tamu.edu Greater amberjack, Seriola dumerili, constitute an increasingly important component of the U.S. Gulf and At- lantic commercial and recreational fisheries. Commercial landings in the Atlantic rose over twentyfold between 1981 and 1991, from un- der 100,000 pounds annually to nearly two million pounds (Cum- mings-Parrack* 1). Commercial land- ings in the Gulf rose similarly, reaching a peak of nearly 2.5 mil- lion pounds in 1988 that was then followed by a >50% decline in 1990 and a further decline in 1991 (McClellan and Cummings2). An increase in commercial landings in 1993 was followed by declines in both 1994 and 1995 (McClellan and Cummings2). The commercial inter- est in greater amberjack appears to be driven by: 1) consumer accep- tance of greater amberjack as an edible fish; 2) replacement of red drum in the blackened fish market; and 3) effort displacement of king mackerel and reef-fish fishermen, especially off the coasts of Florida (Cummings-Parrack1). Recreational landings are smaller in both the Gulf and Atlantic and appear to have declined in recent years (Cum- mings-Parrack1; McClellan and Cummings2). Problems that con- found landing statistics and fishery analysis of the greater amberjack resource include species misidenti- fication and the possibility that re- ported landings may be small in proportion to total removals (Cum- mings-Parrack1). Current manage- ment of the greater amberjack re- source in U.S. waters is based on a two-stock (management unit) model. One stock resides along the U.S. southeastern coast and includes all fishing areas north of Cape Hatteras, NC (35°0G'N) southward to the Dry Tortugas and the Florida Keys (24°35'N); the other includes fish- ing grounds in the Gulf of Mexico off the Dry Tortugas and Florida Keys north and then westward to the U. S. Mexico border (26°00'N) * This paper represents xx in the series “Ge- netic studies in marine fishes” and contri- bution number 67 of the Center for Bio- systematics and Biodiversity at Texas A&M LTniversity, College Station, Texas 77843-2258. 1 Cummings-Parrack, N. 1993. The ex- ploitation status of the Atlantic amberjack fisheries through 1991. Miami Laboratory, SE Fisheries Sci. Center, Natl. Mar. Fish. Serv., Cont. MIA-92/93-30, Miami, FL, 98 p. 2 McClellan, D. B., and N. J. Cummings. 1996. Stock assessment of Gulf of Mexico greater amberjack through 1995. Miami Laboratory, SE Fisheries Sci. Center, Natl. Mar. Fish. Serv., Cont. MIA-96/97-03, Mi- ami, FL, 69 p. 768 Fishery Bulletin 96(4), 1 998 (Cummings and McClellan3). Movement patterns inferred from mark-and-release experiments carried out between 1959 and 1994 (Cummings and McClellan3) are consistent with the two-stock hypoth- esis. In brief, over 1400 recaptures from approxi- mately 14,000 releases revealed cyclical tag-return patterns that suggested resident stocks or subpopu- lations of greater amberjack along both the eastern coast of Florida and the northern Gulf. Exchange rates between the two stocks was estimated as ap- proximately 1.5% by Cummings and McClellan3 al- though, as noted by these authors, the rate estimates were not adjusted for fishing pressure or for poten- tial biases due to mortality, tag shedding, lack of re- porting, and fishing effort. It also was clear from a few tag returns that greater amberjack can migrate considerable distances, e.g. from near Charleston, South Carolina, to Texas or from northwest Florida to Virginia (Cummings and McClellan3). Biological information on greater amberjack is lim- ited to studies reported in Berry and Burch (1978), Shipp (1986), Manooch (1988), and GMFMC (1989). Thompson et al.4 presented data on greater amber- jack age, growth, and reproduction, and Cummings- Parrack1 and McClellan and Cummings2 summa- rized most of the available information on landings and other fishery statistics. Direct or indirect infor- mation on genetic stock structure is even more lim- ited. Johnson5 carried out a pilot study of nuclear- gene (allozyme) variation among 225 greater amber- jack sampled from the Atlantic (zz =60 ), eastern Gulf (n=84), and western Gulf ( /z =8 1 ). Of 72 putative loci examined, only one polymorphic (and nonin- formative) system was found. On the surface, these data do not support the concept of separate stocks. However, genetic homogeneity ( sensu stricto) does not unequivocally establish the existence of a single breeding population (stock), but rather is simply con- sistent with the hypothesis that samples are drawn from a population with the same parametric allele frequencies. In addition, the almost total absence of variation effectively precluded rigorous testing of the null hypothesis (i.e. the interpretation of genetic homogeneity among samples is potentially compro- mised by virtue of the absence of significantly vari- able nuclear-gene loci). 3 Cummings, N. J., and D. B. McClellan. 1996. Movement pat- terns and stock interchange of greater amberjack Seriola dumerili , in the southeastern U.S. Miami Laboratory, SE Fish- eries Sci. Center, Natl. Mar. Fish. Serv., Cont. MIA-95/96-14, Miami, FL, 24 p. 4 Thompson, B. A., C. A. Wilson, J. H. Render, H. Beasley, and C. Cauthron. 1992. Age, growth, and reproductive biology of greater amberjack and cobia from Louisiana waters. Final Rep., Marfin Prog., U.S. Dep. Comm., Coop. Agreement NA90AA-H-MF722, 77 p. Alternatively, the apparent absence of genetic variation raises considerable concern about the ef- fective size of greater amberjack populations. Com- pared with other marine finfish (Smith and Fujio, 1980; Waples, 1987; Bohlmeyer and Gold, 1991), lev- els of nuclear-gene variation in greater amberjack (as reported by Johnson5) are low. Richardson and Gold (1993) examined mitochondrial (mt)DNA varia- tion among 59 greater amberjack sampled primarily from the west coast of Florida. Levels of mtDNA variation in greater amberjack were low in compari- son with red drum and several clupeid species (e.g. Atlantic menhaden), but higher than those found in black drum, red snapper, and red grouper (Camper et al., 1993; Gold et al., 1993; Richardson and Gold, 1993). Estimates of long-term, effective female popu- lation size (computed directly from levels of mtDNA variation) paralleled levels of mtDNA variation, sug- gesting that effective (female) population sizes of Gulf greater amberjack were not atypically low. Concerns regarding greater amberjack fisheries in the Gulf and Atlantic include the following: 1) pre- sumed decreases in average individual size in both Gulf and Atlantic fisheries; 2) apparent declines in size of the presumed Atlantic stock; 3) a trend of de- clining yield in both commercial and headboat fisher- ies in the Gulf; and 4) apparent highly erratic recruit- ment where success of individual year classes is quite variable (Cummings-Parrack1; Cummings and McClellan3). These concerns have intensified as the economic importance of greater amberjack has grown (GMFMC, 1989; Cummings and McClellan3). In this study, we employed variation in restriction sites in mi- tochondrial (mt)DNA of greater amberjack to determine if significant population structure (separate genetic stocks) occurs in U.S. waters, i.e. in the northern Gulf of Mexico and along the U.S. southeastern Atlantic coast. The rationale for this study is the need for accurate geographic definition when conducting stock assessments (Hilborn, 1985; Sinclair et al., 1985), in this case for greater amberjack in U.S. waters. Materials and methods Appropriate tissues (heart and white muscle) were obtained from a total of 444 greater amberjack sampled from 11 offshore localities in U.S. waters (Table 1; Fig. 1). With exception of a sample of seven individuals from near Gulfport, MS, sample sizes 5 Johnson, A. G. 1990. Progress report: electrophoretic exami- nation of greater amberjack ( Seriola dumerili). Panama City Laboratory, SE Fish. Sci. Center, Natl. Mar. Fish. Serv., Panama City, FL, 34 p. Gold and Richardson: Population structure of Seriola dumerih 769 ranged from 24 to 58 individuals. Individuals were sampled variously from charter boats, headboats, and commercial fishing boats, and from catches of recre- ational fishers at tournaments. Tissues were removed from each specimen, quickly frozen in liquid nitro- gen, and transported to Texas A&M University where they were stored at -80°C in an ultracold freezer. Methods (including DNA extraction, precipitation, and storage) used to assay restriction-enzyme frag- ment patterns of mtDNA molecules of individual fish followed those described in Gold and Richardson (1991). Sixteen, type-II restriction-endonuclease en- zymes were used to digest 1.0-1. 5 pg of DNA in 40 pL reactions according to manufacturer’s specifica- tions. Enzymes used were ApaLI, Apal, EcoRI, EcoKV, iTmdIII, Hpal, Ncol , Pstl, Pvull, Seal, Smal, Spe I, Sspl, Sstl, Stul, and Xbal. Methods of DNA digestion, agarose electrophoresis, transfer to nylon filters (after Southern, 1975), hybridization, and autoradiography also followed those in Gold and Richardson (1991). Hybridization employed a 12.5 kilobase (kb) fragment of greater amberjack mtDNA cloned into lambda bacteriophage (Richardson and Gold, 1993). Bacteriophage lambda DNA digested with restriction enzyme Hindlll was employed as a molecular weight marker on each gel. MtDNA frag- ments produced by single digestion of greater am- berjack mtDNA were sized by fitting migration dis- tances to a least-squares regression of lambda DNA- Hi TzdIII fragment migration distances. Single digestion mtDNA-fragment patterns were used to TabSe 1 Sample localities of greater amberjack (Seriola dumerili). Localityj Number of individuals Port Aransas, Texas (PA) 58 Freeport, Texas (FP) 44 Port Fourchon, Louisiana (PF) 43 Gulfport, Mississippi (GP) 7 Pensacola, Florida (PN)2 24 Panama City, Florida (PC) 50 St. Petersburg, Florida (SP) 42 Sarasota, Florida (SR)2 32 Florida Keys (FK) 55 New Smyrna Beach, Florida (NS) 53 South Carolina (SC) 36 Total 444 1 Localities represent dock sites where individuals sampled off- shore were off-loaded. Individuals from the Florida Keys were off-loaded at Islamorada; individuals from South Carolina were off-loaded at several localities. 2 Previously examined by Richardson and Gold (1993). generate composite digestion patterns (mtDNA haplotypes). Analysis of mtDNA data was facilitated by the Restriction Enzyme Analysis Package (REAP) of McElroy et al. (1992). Genotypic (nucleon) diversity within sample localities was calculated following Nei and Tajima (1981) and was based on the total num- ber of mtDNA haplotypes identified within a local- 770 Fishery Bulletin 96(4), 1998 ity. This value represents the probability that any two individuals drawn at random will differ in mtDNA haplotype. Intrapopulational nucleotide se- quence diversity within sample localities also was estimated after Nei and Tajima (1981). This value represents the average nucleotide sequence differ- ence between any two individuals drawn at random from a given sample or locality. Homogeneity testing of mtDNA haplotype frequen- cies among sample localities was carried out by us- ing 1) a randomization (bootstrap) procedure devel- oped by Roff and Bentzen (1989), 2) a log-likelihood ( G ) test (Sokal and Rohlf, 1969), and 3) V tests that employed arcsine, square-root-transformed haplotype frequencies (DeSalle et al., 1987). Significance lev- els for multiple tests were adjusted after Rice (1989). Fst values, a measure of the variance in mtDNA haplotype frequencies, were calculated after Weir and Cockerham (1984) by using algorithms described in Weir ( 1990). Significance of EgT values was tested by the randomization procedure in version 1.4 of the analysis of molecular variance ( AMOVA) program of Excoffier et al. ( 1992). The latter program (AMOVA) was used primarily to examine the distribution of variance in mtDNA haplotypes. AMOVA analysis generates estimates of (genetic) variance components and a set of hierarchical E-statistic analogs (<7> sta- tistics) that are tested for significance through ran- dom permutation methods. The permutation ap- proach avoids the parametric assumptions of normal- ity and independence normally not met by molecu- lar distance measures (Excoffier et al., 1992). Sample localities were nested into regional groupings, i.e. Gulf and Atlantic, that were input into AMOVA. In one AMOVA run, the sample from the Florida Keys was included with the Gulf group, whereas in a sec- ond run it was included in the Atlantic group. Restriction-site presence and absence matrices for individual mtDNA haplotypes were constructed with the GENERATE program in REAP by inferring restric- tion site gains and losses for each enzyme. Maximum- parsimony analysis of restriction-site presence and absence matrices representing all haplotypes employed MULPARS and CONTREE options in version 3.1 of the Phylogenetic Analysis Using Parsimony (PAUP) program of Swofford (1991). Autapomorphic and symplesiomorphic characters were removed prior to PAUP runs. Nucleotide-sequence divergence values among sample localities (interpopulational divergence) were generated following Nei and Tajima (1981) and Nei and Miller ( 1990). MtDNA-based similarity among sample localities was assessed with the neighbor join- ing method (Saitou and Nei, 1987) and employed the NEIGHBOR program in Phylogenetic Inference Pack- age (PHYLIP), version 3.4 of Felsenstein (1989). The spatial distribution of mtDNA haplotypes was investigated by means of spatial autocorrelation analysis (SAAP; Wartenberg, 1989). This analysis determines whether haplotype frequencies at any sample locality are independent of haplotype frequen- cies at neighboring sample localities. Correlograms that plot autocorrelation coefficients (Moran’s I val- ues) as a function of geographic distance between pairs of localities were used to summarize patterns of geographic variation of haplotype frequencies. To minimize “noise” generated by low-frequency haplotypes, Moran’s 7 values were calculated only for haplotypes occurring in eight or more individuals ( 10 haplotypes total). These included haplotypes 1-4, 6, 11, 13, 23, 37, and 48. Results Single digestions of mtDNA molecules from the 444 individuals surveyed produced variable fragment patterns for the 16 restriction enzymes. The major- ity of fragment patterns observed are given in Ap- pendix Table A2 of Richardson and Gold (1993). New fragment patterns revealed in our study are as fol- lows (fragment sizes in base pairs; asterisks repre- sent fragments assumed to exist but not covered by the mtDNA probe): EcoRI, pattern C (8700, 5025, 3175*); Hpal, pattern C (8800, 5550, 2550); Ncol, pattern C (11050, 5850); Sstl, pattern C (11700, 3450, 1750); Neal, pattern C (7500, 3600, 3100, 2700) and pattern D (10000, 3600, 3300); Smal, pattern D (10700, 4250, 1500, 450); Spe I pattern D (7900, 5800, 1200, 1200*, 800); Sspl, pattern C (6600, 6200, 4100); and Xbal, pattern D (7300, 5000, 4600) and pattern E (7300, 4600, 2600, 2275, 125*). The mean genome size of all apparently complete digestion patterns was 16.9 ±0.2 kilobase pairs. No evidence of mtDNA size variation or heteroplasmy was observed. All frag- ment patterns for each restriction enzyme were con- sistent with the hypothesis of single nucleotide sub- stitutions. A total of 72 unique restriction sites was inferred from the digestion patterns. A total of 49 mtDNA haplotypes was identified among the 444 individuals surveyed (Table 2). Four haplotypes (8, 14, 16, and 18) are not listed in Table 2; these were listed in Richardson and Gold (1993) and were identified by three restriction enzymes (Clal,Nsil, and Puul) not employed in our study. Four haplotypes (1, 4, 6, and 13) were abundant, occur- ring in 77, 91, 54, and 70 individuals, respectively. Two haplotypes (2 and 3) occurred in 20 and 22 indi- viduals, respectively, whereas the remainder oc- curred in 10 or fewer individuals. Twenty haplotypes were found in only one individual each. Estimates of Gold and Richardson: Population structure of Seriola dumerili 771 Table 2 Spatial distribution of mitochondrial (mt)DNA haplotypes of greater amberjack (Seriola dumerili). Letters (from left to right) are digestion patterns for ApaLI, Apal, EcoRI, EcoRW, Hindlll, Hpal , Ncol, Pstl, Pvull , Seal, Smal, Spel, SspI, Sstl, Stul , and Xbal. Haplotype no. Composite mtDNA genotype Samples PA FP PF GP PN PC SP SR FK NS SC i AAAAAAAAAAAAAAAA 7 10 8 2 5 8 12 7 4 8 6 2 AAAABBAAAAAAAAAA 4 4 3 — 1 2 3 — 3 2 1 3 AAAAABAAAAAAAABA 3 2 3 — 3 1 1 — 4 4 1 4 AAAAABAAAAAAAAAA 15 9 10 1 5 11 4 10 10 9 7 5 AABBABAAAABAABAA — 1 — — 3 — — 1 — — — 6 AAAAAAAAAAAAAABA 8 4 2 1 — 4 4 5 10 6 10 7 AAAAAAAAAAABAAAA — — — — — — — 1 — — — 9 BABAAAAAAACAAABA — — — 1 — — — 1 1 — — 10 BAAAABBBAAAAAAAA — — 1 — — 2 — 1 — — — 11 BABAAAAAAAAAAABA 1 1 2 — — 1 1 1 — 1 1 12 AAAAAAAAABACAAAA — — — — — — — 1 — 2 — 13 AAAAAAAAABABAAAA 7 4 4 2 1 12 7 2 10 15 6 15 AAAAAAAAAAAABABA — — 1 — — 1 — 1 — — — 17 AAAAABAABAAAAAAC — — — — 1 — — — — — — 19 AAAAABACAAAAAAAA — - — — 1 — — — — — — 20 AAAAAAAAAAAAAABB — — — — 1 — — — — — — 21 AAAAABABAAAAAAAA — 1 — — 1 1 — — — — — 22 ABAACAAACBAAAABA — — — — 1 — — — — — — 23 AABAABAAAAAAAAAA 3 1 2 — 1 1 — — — — — 24 BAAAAAAAAAAABABA — 1 25 AAAAAAAAACABCAAA — — — — — — — — 2 — — 26 BAAABAAAAAAAAAAA — 1 27 ABAACAAACAAAACBA 1 28 AAAAAAAAABAAAAAA — 2 — — — — 1 — — — — 29 AAAACAAAAAAAAABA — 1 30 AAAABBAAAAAAACAA — — 1 — — — — — — — — 31 AAAAABBAAAAAAAAA — 1 32 AAAACAAAADDDAABD — 1 1 — — — — — — — — 33 AAABAAAAAAAAAAAA 1 — — — — 1 — — — — — 34 AACAAAAAABABAAAA 2 35 AAAAABAAAAAAAACA 1 — — — — — 1 1 1 — — 36 BABAABAAAAAAAABA 1 1 37 ABAACAAACAAAAABA 1 — 2 — — — 2 — 4 1 — 38 AAAAAAAAACABAAAA 1 — — — — 1 — — — 1 — 39 AABACCAAADDDAABD 1 — 40 AAAAABAAAAAACAAA — — — — — — 1 — — 1 — 41 ABAAAAAAABABAAAA 1 — 42 AAAAABAAAAAABAAA — — — — — — — — 1 — — 43 AAAAABAAAAAAAAAE — — — — — 1 — — — — — 44 AAAAABCAAAAAAABA — — 1 — — 1 — — — — — 45 AAAAABCAAAAAAAAA — — 1 — — 1 — — — — — 46 AAAACCAAADDDCABD 1 — — — — 1 — — — — — 47 AAAAABAAAABAABAA — — — — — — — — 1 — — 48 AAAAABAAAAAAACAA 1 2 — — — — 2 — 2 1 — 49 AAAAABAAABABAAAA — — 1 — — — — — — — — 50 AAAAABAAABAAAAAA — — — — — — 1 — — — — 51 AAAACAAACAAAAABA — — — — — — 1 — — — 1 52 AABAABAAAABAABAA 1 53 AAAACAAAAAADAAAD — — — — — — 1 — — — — the percentage nucleotide-sequence divergence among the 49 haplotypes ranged from 0.156 to 2.623 (mean ±SE=0.980 ±0.015). MtDNA nucleon diversity was 0.905, and intrapopulational nucleotide sequence diversity was 0.548 ±0.412 (mean ±SD). The latter estimates were based on all 444 individuals surveyed. Nucleon diversity within samples (Table 3) ranged from 0.845 in the sample from Sarasota, Florida, to 772 Fishery Bulletin 96(4), I 998 Table 3 Mitochondrial (mt)DNA nucleon and intrapopulational nucleotide sequence diversities among samples of greater amberjack ( Seriola dumerili) from the Gulf of Mexico and Atlantic Ocean Locality Number of individuals Number of haplotypes Nucleon diversity Nucleotide sequence diversity (+SD)7 Port Aransas, TX 58 17 0.886 0.561 ± 0.443 Freeport, TX 44 18 0.899 0.515 ± 0.387 Port Fourchon, LA 43 16 0.901 0.579 ± 0.435 Gulfport, MS 7 5 0.905 0.587 ± 0.415 Pensacola, FL 24 12 0.906 0.601 ± 0.449 Panama City, FL 50 17 0.872 0.525 ± 0.415 St. Petersburg, FL 42 15 0.879 0.519 ± 0.395 Sarasota, FL 32 12 0.845 0.483 ± 0.370 Florida Keys 55 14 0.893 0.619 ± 0.419 New Smyrna, FL 53 14 0.861 0.537 ± 0.427 South Carolina 36 10 0.846 0.502 ± 0.342 Total 444 49 0.905 0.548 ± 0.412 1 In per cent. 0.906 in the sample from Pensacola, Florida. Intrapopulational nucleotide sequence diversity within samples (Table 3) ranged (mean ±SD) from 0.483 ±0.370 in the sample from Sarasota, Florida, to 0.619 ±0.419 in the sample from the Florida Keys. The latter values are all within one standard error (estimated from bootstrap analysis) of one another, indicating that levels of mtDNA variation are essen- tially identical throughout the geographic area sur- veyed. Levels of mtDNA variability, as measured by nucleon diversity and intrapopulational nucleotide sequence diversity, are commensurate with those estimated for several other marine fish of commer- cial or recreational value (Gold et ah, 1993). Results of bootstrap analyses and log-likelihood tests of spatial homogeneity in mtDNA haplotype frequencies are shown in Table 4. Tests were carried out 1) among all sample localities, 2) among samples from the Gulf, and 3) between defined groups where samples were pooled. In the last category, defined groups were based on region, i.e. Gulf and Atlantic. Gulf samples included the following: Port Aransas, TX; Freeport, TX; Port Fourchon, LA; Gulfport, MS; Pensacola, FL; Panama City, FL; and Sarasota, FL. Atlantic samples included New Smyrna, FL, and South Carolina. Two sets of pooled comparisons were carried out: in one, the sample from the Florida Keys was included with Gulf samples; in the other, the sample from the Florida Keys was included with At- lantic samples. We used this approach because the sample from the Florida Keys is located on the geo- graphic boundary between the two putative stocks of greater amberjack (Cummings and McClellan3), Table 4 Tests for spatial homogeneity in mtDNA haplotype frequen- cies among greater amberjack (Seriola dumerili) from the Gulf of Mexico and U.S. southeastern Atlantic. FST is a measure of variance in haplotype frequencies. Number of samples is in parentheses. Number of haplotypes Homogeneity tests7 Test group r RB ^ST All samples (11) 49 0.158 >0.050 0.005 Gulf samples (8) Pooled comparisons Gulf + Florida Keys 45 0.250 >0.050 0.003 vs. Atlantic (2) Atlantic + Florida 49 0.627 >0.050 0.004 Keys vs. Gulf (2) 49 0.042 -0.002 0.0092 ; PRB is probability from randomization (bootstrap) approach of Roff and Bentzen (1989); PG is probability from log-likelihood (G) test (Sokal and Rohlf, 1969). 2 Value differs significantly (P=0.007) from 0.00; all other FST val- ues are nonsignificant. and we could not place a priori the sample from the Florida Keys into either stock before testing the two- stock hypothesis. Significant heterogeneity was found only in the pooled comparison of samples from the Atlantic and Florida Keys versus samples from the Gulf; tests of homogeneity among all samples, among samples from the Gulf, and between pooled samples from the Atlantic versus those from the Gulf plus the Florida Keys were nonsignificant (Table 4). Estimates of FgT revealed the same pattern; the FgT Gold and Richardson: Population structure of Seriola dumerili 773 TabSe 5 Hierarchical analysis of molecular variation (AMOVA) among mtDNA haplotypes of greater amberjack (Seriola dumerili) from the Gulf of Mexico and U.S. southeastern Atlantic. Variance component Observed partition Variance % total P1 0 values Gulf + Florida Keys vs. Atlantic Between regions 0.00127 0.29 0.259 d>CT= 0.003 Among samples within regions 0.00160 0.36 0.190 0SC = 0.004 Within samples 0.43974 99.35 0.121 ST = 0.006 Atlantic + Florida Keys vs. Gulf Between regions 0.00398 0.90 0.001 0rT= 0.009 Among samples within regions 0.00011 0.03 0.444 0SC = 0.000 Within samples 0.43974 99.07 0.113 0ST= 0.009 ' Probability of finding a more extreme variance component by chance alone (5000 permutations). value of 0.009 in the pooled comparison of samples from the Atlantic and Florida Keys versus samples from the Gulf differed significantly from zero, whereas FST values in all other comparisons were nonsignificant (Table 4). For AMOVA, where the variance in mtDNA haplotypes was partitioned by region, we also per- formed two separate analyses: one where the sample from the Florida Keys was included with Gulf samples, and one where the sample from the Florida Keys was included with Atlantic samples. In both analyses, the majority of the variance (>99%) was distributed within samples, and in both, the among- samples-within-groups component (<7>sc) was nonsig- nificant (Table 5). The between-region component ( &CT) in the comparison of Atlantic and Florida Keys versus the Gulf was highly significant (P<0.001); whereas d>CT in the comparison of Atlantic versus the Gulf plus Florida Keys was not (Table 5). These re- sults paralleled results of the homogeneity tests and the Fst values. Finally, we employed pairwise CT value) in AMOVA, when regions were defined as the Gulf versus the Atlan- tic (and the Florida Keys); and 4) com- parison of pairwise <2>ST “distances” where samples from the Gulf or from the Atlantic were at least three times more similar to one another than were samples from the Gulf versus those from the Atlantic (and the Florida Keys). Neither maximum-parsimony (MP) analysis of individual mtDNA haplo- types nor neighbor-joining analysis of a matrix of interpopulational (inter- sample) nucleotide-sequence distances revealed phylogeographic patterns in- dicative of population structuring. A few, well-supported clades of indi- vidual haplotypes were detected in MP analysis, but in no case were haplo- types within individual clades from the same or geographically proximate a 3 c r o> Haplotype 6 Haplotype 13 Figure 2 Frequency of mtDNA haplotypes 6 and 13 among greater amberjack (Seriola dumerili) sampled from 11 localities in the northern Gulf of Mexico and southeastern (U.S.) Atlantic. Acronyms are defined in Table 1. Pensacola, FL Sarasota, FL — New Smyrna, FL Panama City, FL Port Fourchon, LA — Florida Keys Freeport, TX St. Petersburg, FL ■ Port Aransas, TX South Carolina Gulfport, MS Figure 3 Neighbor-joining topology generated from matrix of pairwise, interpop- ulational ( mtDNA) nucleotide sequence divergence among samples of greater amberjack, Seriola dumerili. Gold and Richardson: Population structure of Senola dumerili 775 12 3 4 Correlograms based on frequencies of mtDNA haplotypes found in eight or more individuals among samples of greater ambeijack, Seriola dumerili , from the northern Gulf of Mexico. Abscissas: distance classes 1-4 (left to right); ordinates: mean auto- correlation coefficients (Moran’s I values) for each distance class (±SE). (A) Equal frequencies/distance class; (B) equal distances between distance classes. sample localities. Absence of structure in MP analy- sis is not inconsistent with results of homogeneity testing, FgT values, and AM OVA analysis, in that 1) presumed restrictions in gene flow between subpopu- lations could be relatively recent, i.e. there has been insufficient time for haplotypes (e.g. haplotype 6 ) that differ in frequency between subpopulations to become reciprocally monophyletic, or 2) there may be lim- ited gene flow between subpopulations. Finally, there was no indication of spatial autocorrelation (positive or negative) in common haplotypes among samples from the Gulf. This finding indicates the absence of an isolation-by-distance effect among greater amber- jack in the Gulf and is consistent with the hypoth- esis of continuous gene flow across the northern Gulf. Divergence in mtDNA between Gulf and Atlantic subpopulations has been documented for a variety of marine species. In some (e.g. American oysters, toadfish, black sea bass, and to a lesser extent, horse- shoe crabs), major phylogeographic discontinuities between Gulf and Atlantic subpopulations were found, leading to the hypothesis that the similar vicariant patterns may have stemmed from episodic changes in environmental conditions during Pleis- tocene glaciation (Avise, 1992). In addition, the pres- ence of phylogeographic structure was taken to indi- cate that current-day gene flow between subpopula- tions is very restricted, if it occurs at all. In species such as red drum (Gold et al., 1993), king mackerel (Gold et al., 1997), and greater amberjack (our study), mtDNA differences between Gulf and Atlantic sub- populations are documented but are limited to ei- ther a frequency difference in a single mtDNA hap- lotype, a small (but significant) difference in mtDNA haplotype distribution, or both. The relatively small genetic differences observed between subpopulations in these species, along with the absence of phylogeo- graphic structure, both between regions and among haplotypes, suggest either that limited gene flow occurs between subpopulations or that separation between subpopulations is fairly recent. In the case of greater amberjack, the former is consistent with mark-and-recapture data that suggest exchange rates of 1.5% between Gulf and Atlantic stocks (Cummings and McClellan3). For greater amberjack, the boundary between Gulf and Atlantic subpopulations appears to be between the Florida Keys (included in the Atlantic subpopu- lation) and somewhere off the central-western Florida coast, possibly the Florida Middle Ground, a series of north-south reef structures located about 150 km south of the north Florida coast and about 160 km northwest of Tampa Bay (Hopkins et al.6), and the primary source for amberjack fishermen located in Sarasota and St. Petersburg. Separation between Gulf and Atlantic subpopulations (stocks) of greater amberjack could stem from a number of causes that involve historical or recent interactions between dispersal capability and impediments to gene flow, or both. Among present-day alternatives are 1) offshore currents that are not conducive to unrestricted movement between regions; and 2) ab- sence of suitable habitat or difference in ecological (biogeographic) provinces between regions. A third (historical) possibility might be that subpopulations were separated (e.g. during Pleistocene glaciations) and have only recently (in geological time) begun exchanging genes. Rates of approach to genetic ho- mogeneity under this last hypothesis are, in part, time-dependent, and one could speculate that there has been insufficient time for accumulated genetic differences to disappear. 6 Hopkins, T. 1981. Florida Middle Ground. In R. Rezak and T. J. Bright (eds.), Final report: northern Gulf of Mexico topo- graphic features study, p. 1-5. Tech. Rep. No. 81-2-T, Dep. Oceanography, Texas A&M University, College Station, TX. 776 Fishery Bulletin 96(4), 1998 There is at least suggestive evidence for each of the above three possibilities in greater amberjack. First, measurements of offshore current velocities are consistent with increased passive movement of indi- viduals from the Florida Keys into the Atlantic rather than northward along the west Florida coast. Mean current velocities ( 145 m below the surface) eastward from the Keys through the Florida Straits are ap- proximately 75 cm/sec, and along the southeast Florida coast, these currents can be as fast as 95 cm/ sec (SAIC ' ). Because current velocities are expected to be greater nearer the surface, and because greater amberjack have been observed spawning in the Florida Keys (Cummings and McClellan3), buoyant eggs and larvae would likely be impacted and have little opportunity to move northward along the west coast of Florida. Second, in terms of habitat, the southern half of the Florida peninsula represents a transition zone between temperate and tropical forms, where the southern ranges of many temper- ate species terminate in tropical southern Florida (Briggs, 1974). There also is relatively little reef habi- tat or sharp topography between the Florida Middle Ground and the Keys; most of the area has a rela- tively smooth bottom comprising shell, sand, and quartz (Rezak et al., 1985; Rezak and Bright7 8). Be- cause greater amberjack exhibit a marked preference for reefs, rock outcrops, and wrecks (Shipp, 1986; Manooch, 1988), it is possible that the paucity of sig- nificant reef or other major structure on the Outer Continental Shelf off southwestern Florida may in- hibit movement of greater amberjack between the Florida Middle Ground and the Keys. Finally, al- though relatively little is known about the early life history of greater amberjack, spawning is thought to occur from mid-spring to early summer both along the southeastern coast of Florida (including the Florida Keys) and in the northern Gulf off Louisiana (Cummings and McClellan3; Thompson9). This spawning time suggests that warmer water tempera- tures may trigger onset of reproductive activity. One might then hypothesize that during the late Pleis- tocene Epoch, when waters of the northern Gulf were much cooler (Rezak et al., 1985), subpopulations of 7 SAIC (Science Applications International Corporation). 1992. Straits of Florida physical oceanographic field study, final in- terpretative report, volume II: technical report. OCS Report/ MMS 92-0024. U. S. Dep. Interior, Minerals Mgmt. Serv., Gulf of Mexico OCS Regional Office, New Orleans, LA, 179 p. 8 Rezak, R., and. T. J. Bright (eds.). 1981. Final report: north- ern Gulf of Mexico topographic features study. Tech. Rep. No. 81-2-T, Dept. Oceanography, Texas A&M University, College Sta- tion, TX, 150 p. 9 Thompson, B. A. 1997. Coastal Fisheries Institute, Louisi- ana State University, Baton Rouge, LA 70803. Personal commun. greater amberjack were isolated in warm-water refu- gia, perhaps off the Florida Keys (or in the Carib- bean) and off the Yucatan Peninsula (Campeche Banks) where considerable reef habitat exists (Rezak et al., 1985). Following glacial retreat, the (putatively) iso- lated subpopulations could then have returned to the northern Gulf. Because the rate of approach to genetic homogeneity would be partly a function of time and gene flow, genetic differences between present-day sub- populations of greater amberjack could simply be his- torical artifacts, reflecting insufficient time (or re- stricted gene flow) relative to genetic homogenization. Our finding that the sample of greater amberjack from the Florida Keys was included in a grouping (subpopulation) with samples from the Atlantic dif- fers from a recent study in king mackerel (Gold et al., 1997), where a sample from the Florida Keys was placed in a grouping with samples from the Gulf. King mackerel are highly migratory, and the sample of king mackerel from the Florida Keys was taken during late winter when the majority of king mack- erel in the Keys are thought to be from the Gulf Mi- gratory Unit or stock (Williams and Godcharles10 * *). The sample of greater amberjack from the Keys, how- ever, was obtained during late March-early April, a time when the majority of king mackerel in the Keys are considered to be from the Atlantic Migratory Unit or stock (Williams and Godcharles10). Although greater amberjack are not migratory in the same way as king mackerel, it would be of more than passing interest to examine winter samples of greater amberjack from the Florida Keys and ask whether the Florida Keys consti- tute a mixing zone in greater amberjack as in king mackerel. Along these lines, it also would be important to examine both summer and winter samples of greater amberjack in the area between the Florida Middle Ground and the Florida Keys. A better definition of the geographic limits of the two subpopulations is critical for assessment and allocation of the greater amberjack resources along the west coast of Florida. Avise et al. (1988) presented models that allow estimation of evolutionary (long-term) effective fe- male-population size (N^g) values) based on mtDNA intrapopulational nucleotide sequence diversities. Assuming that the generation time in greater am- berjack is three years (Wilson11), we estimated N^e) for the two subpopulations (stocks) of greater am- berjack to be 90,000 (Gulf of Mexico) and 93,500 (U.S. southeastern Atlantic, including the Florida Keys). 10 Williams, R. O., and M. F. Godcharles. 1984. Completion re- port, king mackerel tagging and stock assessment. Project 2- 341-R. Florida Dep. Natural Resources., St. Petersburg, FL. 11 Wilson, C. A. 1997. Coastal Fisheries Institute, Louisiana State University, Baton Rouge, LA 70803. Personal commun. Gold and Richardson: Population structure of Seriola dumerili 111 The two estimates do not differ significantly from one another and are larger than those reported in a more limited (in terms of sample size and geographic coverage) study of greater amberjack (Richardson and Gold, 1993). They also are commensurate with N„ , estimates for several other marine fish of com- fie) mercial or recreational value (Gold et ah, 1993). Al- though a positive correlation exists between effec- tive population sizes and census sizes, the latter are generally an order of magnitude or two larger than the former (Avise et ah, 1988). One reason for this difference, at least in estimates of Nf}g) values for poikilothermic vertebrates, is that the models of Avise et al. (1988) employ (estimated) mtDNA evolution- ary rates for homeothermic vertebrates. Estimated mtDNA evolutionary rates for poikilothermic verte- brates may be 5-10 times less than those for homeo- thermic vertebrates (Martin and Palumbi, 1993), suggesting that long-term effective population sizes of each subpopulation of greater amberjack could be on the order of 500,000 to 1,000,000 females. According to results of this project, current stock boundaries for assessment and allocation of greater am- beijack resources in U.S. waters appear appropriate except possibly for the west coast of Florida, south of the Florida Middle Ground. Depending on present or future importance of the greater amberjack fishery in this area, it would be useful to know both the geographic limits of the two subpopulations and whether the Florida Keys constitute a mixing zone for greater am- beijack as for species such as king mackerel. Acknowledgments We thank J. Bielawski, D. Codella, C. Denis, K. Dunn, D. Fable, J. Franks, Jer. Gold, Jes Gold, E. Gricius, C. Grimes, E. Heist, P. Hood, J. Magursky, C. Ragland, and B. Thompson for assistance in procur- ing specimens, and N. Cummings and T. Turner for comments on the manuscript. Work was supported by the Saltonstall-Kennedy Program of the U.S. De- partment of Commerce (Award NA57FD-0-069-01), as administered by the National Marine Fisheries Service, and by the Texas Agricultural Experiment Station under Project H-6703. Part of the work was carried out in the Center for Biosystematics and Biodiversity, a facility funded, in part, by the Na- tional Science Foundation (Award DIR-89-07006). Literature cited Avise, J. C. 1992. Molecular population structure and the biogeographic history of a regional fauna: a case history with lessons for conservation biology. Oikos 63:62-76. Avise, J. C., R. M. Ball, and J. Arnold. 1988. Current versus historical population sizes in verte- brate species with high gene flow: a comparison based on mitochondrial DNA lineages and inbreeding theory for neutral mutations. Mol. Biol. Evol. 5:331-344. Berry, F. H., and R. K. Burch. 1978. Aspects of the amberjack fisheries. Proc. Gulf Caribb. Fish. Inst. 31:179-194. Bohlmeyer, D. A., and J. R. Gold. 1991. Genetic studies in marine fishes. II. A protein elec- trophoretic analysis of population structure in the red drum Sciaenops ocellatus. Mar. Biol. 108:197-206. Briggs, J. C. 1974. Marine zoogeography. McGraw Hill, New York, NY, 475 p. Camper, J. D., R. C. Barber, L. R. Richardson, and J. R. Gold. 1993. Mitochondrial DNA variation among red snapper ( Lutjanus campechanus ) from the Gulf of Mexico. Mol. Mar. Biol. Biotechnol. 3:154-161. DeSalle, R., A. Templeton, I. Mori, S. Pletscher, and J. S. Johnston. 1987. Temporal and spatial heterogeneity of mtDNA poly- morphisms in natural populations of Drosophila merca- torum. Genetics 116:215-223. Excoffier, L., P. E. Smouse, and J. M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mi- tochondrial DNA restriction data. Genetics 131:479-491. Felsenstein, J. 1989. PHYLIP (Phylogeny Inference Package), version 3.4. Cladistics 5:164-166. Gold, J. R., and L. R. Richardson. 1991. Genetic studies in marine fishes. IV. An analysis of population structure in the red drum ( Sciaenops ocellatus) using mitochondrial DNA. Fish. Res. 12:213-241. Gold, J. R., A. Y. Kristmundsdottir, and L. R. Richardson. 1997. Mitochondrial DNA variation in king mackerel ( Scomberomorus cavalla ) from the western Atlantic Ocean and Gulf of Mexico. Mar. Biol. 129:221-232. Gold, J. R., L. R. Richardson, C. Furman, and T. L. King. 1993. Mitochondrial DNA differentiation and population structure in red drum ( Sciaenops ocellatus ) from the Gulf of Mexico and Atlantic Ocean. Mar. Biol. 116:175-185. GMFMC (Gulf of Mexico Fishery Management Council). 1989. Amendment number 1 to the reef fish fishery man- agement plan. Gulf of Mexico Fishery Management Coun- cil, Tampa, FL, 356 p. Hilborn, R. 1985 Apparent stock-recruitment relationships in mixed stock fisheries. Can. J. Fish. Aquat. Sci. 42:718-723. Manooch, C. S., III. 1988. Fisherman’s guide: fishes of the southeastern United States. NC State Mus. Nat. Hist., Raleigh, NC, 362 p. Martin, A. P., and S. R. Palumbi. 1993. Body size, metabolic rate, generation time, and the molecular clock. Proc. Natl. Acad. Sci. (USA) 90:4087- 4091. McElroy, D., P. Moran, E. Bermingham, and I. Kormfield. 1992. REAP-The restriction enzyme analysis package. J. Hered. 83:157-158. Nei, M., and J. C. Miller. 1990. A simple method for estimating average number of nucleotide substitutions within and between populations from restriction site data. Genetics 125:873-879. 778 Fishery Bulletin 96(4), 1 998 Nei, M., and F. Tajima. 1981 . DNA polymorphism detectable by restriction endonu- cleases. Genetics 97:145-163. Rezak, R„, T. J. Bright, and D. W. McGrail. 1985. Reefs and banks of the northwestern Gulf of Mexico. John Wiley & Sons, New York, NY, 259 p. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43: 223-225. Richardson, L. R., and J. R. Gold. 1993. Mitochondrial DNA variation in red grouper ( Epinephelus morio) and greater amberjack ( Seriola dumerili ) from the Gulf of Mexico. ICES J. Mar. Sci. 50:53-62. Roff, D. A., and P. Bentzen. 1989. The statistical analysis of mitochondrial polymor- phisms: chi-square and the problem of small samples. Mol. Biol. Evol. 6:539-545. Saitou, N., and M. Nei. 1987. The neighbor joining method: a new method for re- constructing phylogenetic trees. Mol. Biol. Evol. 4:406- 425. Shipp, R. L. 1986. Dr. Bob Shipp’s guide to the fishes of the Gulf of Mexico. 20th Century Printing Co., Mobile, AL, 256 p. Sinclair, M., V. C. Anthony, T. D. lies, and R. N. O’Boyle. 1985. Stock assessment problems in Atlantic herring ( Clupea harengus) in the northwest Atlantic. Can. J. Fish. Aquat. Sci. 42:888-898. Smith, P. J., and Y. Fujio. 1980. Genetic variation in marine teleosts: high variabil- ity in habitat specialists and low variability in habitat generalists. Mar. Biol. 69:7-20. Sokal, R. R., and F. J. Rohlf. 1969. Biometry: the principles and practice of statistics in biological research. Freeman and Sons, San Francisco, CA, 776 p. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503-517. Swofford, D. L. 1991. PAUP: Phylogenetic analysis using parsimony. Us- ers manual. Illinois Nat. Hist. Surv., Champaign, IL, 179 p. Waples, R. S. 1987. A multispecies approach to the analysis of gene flow in marine shore fishes. Evolution 41:385-400. Wartenberg, D. 1989. SAAP: a spatial autocorrelation analysis pro- gram. Dep. Environ. Community Med., Robert Wood Johnson Medical School, Univ. of Medicine and Dentistry of New Jersey, Piscataway, NJ, 28 p. Weir, B. S. 1990. Genetic data analysis. Sinauer Assoc., Inc., Sunder- land, MA, 377 p. Weir, B. S., and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38:1358-1370. 779 Abstract .“The age, growth, and re- production of the tropical Indo-Pacific ommastrephid squid Nototodarus hawaiiensis was studied on the North West Slope of Australia. The weight, mantle length ( ML ), gonad weight, and nidamental gland length were mea- sured for 37 males and 52 females cap- tured in January and February 1992 and ranging in size from 42 mm ML to 214 mm ML. Statolith increments were counted, as a proxy for age. The num- ber of statolith increments, counted on 42 of the squid, ranged from 49 to 195. The relation between increment num- ber (i.e. age) and ML was linear for both sexes. The relation between increment number and ovary weight, and between increment number and testis weight, had greater variability than did ML versus ovary weight, and ML versus testis weight, indicating a large range in age at maturity in individuals of similar size. Some statoliths showed two prominent zones, the origins of which are uncertain. Back-calculated hatch dates indicated that all squid hatched between July and December 1991 and that the majority hatched between August and October (Austral spring). Manuscript accepted 7 January 1998. Fish. Bull. 96:779-787 (1998). Age, growth, and reproduction of the tropical squid Nototodarus hawaiiensis (Cephalopoda: Ommastrephidae) off the North West Slope of Australia George D. Jackson Department of Marine Biology, James Cook University Townsville, Queensland 4811, Australia Present address: University of Tasmania Institute of Antarctic and Southern Ocean Studies Hobart, Tasmania 7000, Australia E-mail address: george.jackson@jcu.edu.au Vicki A. Wadley CSIRO Division of Marine Research GPO Box 1538, Hobart Tasmania 7001, Australia Nototodarus hawaiiensis (Berry, 1912) was thought originally to be restricted to the Hawaiian Islands in the central Pacific (Roper et al., 1984). However, studies by Dunning (1988a, 1988b) revealed that this species is also distributed in the Indo-Pacific including the Philip- pines (previously referred to as N. sloani philippinensis [Voss, 1962]), northern Australia, South China Sea, western Indian Ocean, and off Chile (Dunning and Forch, in press). In Australian waters, post- paralarval N. hawaiiensis inhabit slope waters at depths between 200 and 500 m off the North West Slope and between 100 and 600 m off the north east coast to southern Queens- land (Dunning, 1988b) in bottom temperatures of 12.4°C. Recent trawl surveys on the North West Slope of Australia have shown that N. hawaiiensis is the dominant spe- cies of cephalopod in commercial catches from demersal trawlers (Wadley, 1993). The recent identifi- cation and dominance of N. hawai- iensis in waters off the North West Slope of Australia necessitates under- standing the biology and role of this species in deepwater marine commu- nities in tropical Australian waters. This study was carried out to ob- tain preliminary age, growth, and reproductive parameters for N. hawaiiensis not previously reported. Statolith age analysis, a valuable tool in squid growth and life history studies (Jackson, 1994), was carried out as an indication of age and growth. Previous statolith ageing work on tropical Australian squid species has been restricted to shal- low water loliginids that predomi- nantly complete their life span in less than 200 d (Jackson, 1990; Jackson and Choat, 1992; Jackson and Yeat- man, 1996). However, no growth in- formation has been available to date for deepwater tropical species of other families. It was therefore of in- terest, for comparative purposes, to obtain life history parameters of this deepwater tropical ommastrephid. Statoliths of AT hawaiiensis were studied to assess if periodic incre- ments might be useful for ageing this species. Owing to the difficulty of obtaining live specimens from deep water, age information has to be inferred (as in this study), rather than validated on living squid. Incre- ment structure was considered in re- lation to the validation evidence available for other ommastrephids. 780 Fishery Bulletin 96(4), 1998 Materia! and methods Squid collection and analysis Individuals were collected day and night, 20 Janu- ary to 14 February 1992, off the North West Slope of Australia, between about 12°S and 22°S (Wadley, 1993) and 385 to 555 m. This study was composed of 89 individuals of which 42 individuals were used for age analysis. Specimens were obtained from the RV Southern Surveyor and commercial vessels, which trawled with demersal gear (45-mm mesh net) on soft sandy substrates. The fresh N. hawaiiensis were identified by using field characters from Wadley (1990). Other omma- strephids were captured (Wadley, 1993) but Noto- todarus gouldi was absent in the area. Specimens were sampled at random from the range of sizes avail- able in the trawl catches. The squid were frozen at sea and subsequently defrosted in the laboratory for reproductive and statolith analysis. Measurements taken included mantle length (ML), total wet weight (W), testis weight for males and ovary weight and nidamental gland length (NGL) for females. Hectocotylisation of the ventral arm and presence of spermatophores in Needham’s sac were used as in- dicators of sexual maturity in males. In females, de- velopment of the nidamental gland and presence of mature oocytes in the ovaries were used as indica- tors of sexual maturity. Statolith removal and analysis Statoliths were removed through an incision in the cephalic cartilage, placed singly on microscope slides, and rinsed with 100% ethanol. The dried statoliths were mounted in thermoplastic cement (Crystal Bond, Jackson, 1990). Statoliths were taken for age analysis from 42 individuals selected from the full size range available. To observe the increment structure, statoliths were ground and polished, usually on both the anterior and posterior surfaces. The complete in- crement sequence was usually visible only on the posterior (convex) surface, which was preferred for routine counting. In many of the statoliths, the crystal structure obstructed a view of the incre- ments in the nuclear region of the anterior sur- face. Some statoliths were ground from the ante- rior (concave) surface through the obstructing crys- tals until the nucleus was revealed and all the in- crements could be observed (terminology accord- ing to Lipinski et al., 1991). Statoliths were viewed on a computer monitor with a video camera attached to a compound mi- croscope. Increments were counted by following the increment sequence with a cursor while using a hand counter. The number of increments on each statolith was counted at least three times and the mean was taken. Counts that varied more than 10% from the mean were repeated or rejected. The increments were similar in structure to daily statolith increments ob- served in other ommastrephid species (e.g. Hurley et al., 1985; Villanueva, 1992; Arkhipkin, 1996). Results Length-weight relationship Weight and length parameters were measured for 37 males and 52 females. Males ranged in size from 42 mm ML and 6.6 g to 176 mm and 211.7 g; females ranged from 48 mm ML and 5.3 g to 214 mm ML and 381.6 g (Fig. 1). Increment number The relation between the number of statolith incre- ments and ML was linear for both males and females over the size range available, although there was considerable variability in the data (Fig. 2). The re- gression equations were Jackson and Wadley. Age, growth, and reproduction of Nototodarus hawaiiensis 781 y = 0.95.x - 7.75 (r2 = 0.75) and y - 1.08.x - 14.915 (r2 = 0.78), for males and females respectively, where x = increment number; and y = mantle length (mm). The increment count for males ranged from 71 (42 mm ML) to 192 (164 mm ML). Statolith increments in the largest male collected (176 mm ML) could not be counted owing to overgrinding of the statolith. The increment count for females ranged from 49 (48 mm ML) to 195 (183 mm ML). The largest female (214 mm ML) had an increment count of 179. Male maturation patterns The largest immature male was 127 mm ML and had 167 statolith increments. Males appeared to mature as early as 100 d and 90 mm ML. Mature males showed considerable range in weights of testis (from 0.92 g to 2.84 g, Fig. 3A). On the basis of ML alone, testis weight appeared to increase rapidly with growth (Fig. 3A), and there was some evidence of testis regression (on the basis of reduced testis weight relative to ML) at larger sizes. However, when testis growth was compared with increment number, there was no clear pattern. Plotting testis weight against increment number showed that mature individuals varied widely in age and individuals of similar age varied widely in testis weight (Fig. 3B). Female maturation patterns The largest immature female was 166 mm ML and had 192 statolith increments, whereas the smallest mature female was 136 mm ML and had 146 sta- tolith increments. All females larger than 166 mm ML were mature on the basis of the presence of ma- ture oocytes. On the basis of ML (Fig. 4A), females reached full maturity over a narrow length range between approximately 136 and 170 mm ML. How- ever, as with testis weight for the males, ovary weight plotted against increment number (Fig. 4B) showed considerably more variability than did ovary weight plotted against ML (Fig. 4A). The youngest mature female was 125 d. It appeared that there was little growth in the ovary before 100 d. The largest immature female ( 166 mm ML) had a NGL of 48 mm ML. Whereas the smallest mature female (136 mm ML) had a NGL of 51 mm, all ma- ture females had a NGL greater than 38 mm. All immature individuals had a NGL less than 50 mm. The trend in growth of the nidamental gland fol- lowed a similar pattern for either ML or incre- ment number (Fig. 5, A and B). Hatching dates On the assumption that statolith increments are formed daily, hatching dates were backcalculated. All individuals that were aged hatched between July and December 1991; the majority (83%) hatched between August and November, and 7 1% hatched in August, September, and October (Fig. 6). Most squid hatched during the austral spring (September-November), although there was some hatching in late winter (July) and early summer (December). Statolith microstructure and growth zones The statoliths of N. hawaiiensis generally had clear increments that could be counted from the nucleus to the outer margin of the dorsal dome (Fig. 7). There was a pattern in the zonation in nine of the 42 statoliths examined (Fig. 7, A and B). Some individuals had a distinct inner opaque zone followed by an outer translucent zone (Fig. 782 Fishery Bulletin 96(4), 1 998 7). No obvious opacity was observed in any of the juvenile (<110 mm ML, n-lA) statoliths. Opaque zone increment counts on the nine individuals ranged between 89 and 135. The crystal structure obscuring increments in some statoliths of N. hawaiiensis resembled crystals in other squid statoliths. They appeared similar to structures in Illex illecebrosus, described as “nodules” by Lipinski ( 1981 ) or “occulting crystals” by Dawe et al. (1985). Similar crystals were also observed in the statolith microstructure of the Antarctic squid Mastigoteuthis psychrophila (Jackson and Lu, 1994). Discussion Maturation Nototodarus hawaiiensis matures at a relatively young age (<150 d). According to our study of trawl- caught specimens, N. hawaiiensis may have a life span of less than 200 d. This is in contrast to its tem- perate-water conspecifics Nototodarus gouldi and N. sloanii, which do not reach maturity until 200 d in New Zealand waters (Uozumi et al., 1995). Matura- tion in N. hawaiiensis appears to be more closely tied Jackson and Wadley. Age, growth, and reproduction of Nototodarus hawaiiensis 783 to body size than to age, suggesting that there is a minimal physical or physiological size threshold to be reached before maturity can take place, regard- less of age. This pattern has also been found in shal- low-water Photololigo sp. (referred to as Loligo chinensis in Jackson, 1993a) and Lolliguncula brevis (Jackson et al., 1997), as well as in males of the deepwater onychoteuthid Moroteuthis ingens (Jack- son, 1997). Guerra and Castro (1994) found that fe- male reproductive organs of Loligo gahi generally required a minimum body size before increasing sub- stantially in size. Such a trend may be a common 90 A :• 80 • • 70 V. • 60 •• • 50 - •o o 40 - • o 30 - o 20 - 0 0 o° o E e, io - XI □> 0 - c u — 0 50 100 150 200 250 c Mantle length (mm) CD 03 •£ 90-i a> E B nj 80 ■ • • 70 • • • 60 - *o - • 50 - • o 40 - 30 - o o 20 - o 10 - P) O o 0 - C 25 50 75 100 125 150 175 200 Increment number Figure 5 The relation between (A) mantle length and nidamental gland length and (B) increment number and nidamental gland length for Nototodarus hawaiiensis. Filled circles represent mature individuals; hollow circles represent immature individuals. strategy in squids. Illex argentinus in the South At- lantic (Rodhouse and Hatfield, 1990) likewise shows considerable variability in the timing of maturation in relation to age. Rodhouse and Hatfield ( 1990) pos- tulated that for I. argentinus, maturity and gonad growth does not occur at the expense of somatic growth. This also appears to be the case for N. hawaiiensis. However, this pattern of maturation contrasts with the deepwater onychoteuthid squid Moroteuthis ingens, which undergoes degradation of somatic tissues with maturation (Jackson and Mladenov, 1994). Size of the nidamental gland relative to body length in N. hawaiiensis appears to be a useful indication of female maturity because growth of this organ is closely associated with growth of the ovary (Ikeda et al., 1991, Collins et al., 1995). Uozumi et al. (1995) found a close association in growth of the nidamental gland and ovary size and ovulation in the closely re- lated Nototodarus sloanii and N. gouldi in New Zealand. Hatching Ageing data suggest that some ommastrephids have extended spawning periods (e.g. Illex argentinus, Arkhipkin, 1993; Todarodes paeificus, Nakamura and Sakurai, 1993; Todarodes sagittatus, Nototodarus sloanii, Uozumi and Ohara, 1993; Ommastrephes bartramii. Bower, 1996). Other ommastrephids, such as Illex illecebrosus, which hatches predominantly in spring (Dawe and Beck, 1997), appear to have peaks of spawning. In some instances, spawning peaks may be regionally influenced; Martialia hyadesi captured on the Patagonian Shelf Edge had Ju! Aug Sep Oct Nov Dec Month Figure 6 The hatching-date distribution of all individuals (n=4'2) of Nototodarus hawaiiensis aged in this study. 784 Fishery Bulletin 96(4), 1998 Figure 7 Statolith microstructure of Nototodarus hawaiiensis (A) dorsal dome region of statolith male (132 increments, 127 pm ML) — note the distinct inner opaque zone followed by the outer translucent zone, scale bar = 100 pm (B) dorsal dome region of statolith of female (166 pm ML, increment count not taken) — note the less distinct opaque and translucent zones, scale bar = 100 pm (C) dorsal dome region of statolith of female (146 increments, 136 pm ML) — note lack of any obvious opaque zone and the unusual overlapping increment structure shown to the right of the photograph, scale bar = 100 pm (D) nuclear region of statolith of juvenile female (61 increments, 57 pm ML), scale bar = 25 pm. a spring peak, whereas individuals captured at the Antarctic Polar Front hatched in winter (Rodhouse et al., 1994). This preliminary data indicates a late winter-spring to early summer hatching for N. hawaiiensis in this study. Wadley ( 1993) collected 50- mm-ML juveniles of N. hawaiiensis off the North West Slope in August and 70-80 mm ML juveniles in April, suggesting hatching at different times of the year. From plankton surveys and collection of mature females in North West Slope waters, Dun- ning (1988a) concluded that N. hawaiiensis spawned year-round because paralarvae and juveniles were captured in September and October, and mature fe- males were captured in February, April, August, and late September. Statolith zones The origin of the opaque and translucent zones within the statolith microstructure of some of the N. Jackson and Wadley: Age, growth, and reproduction of Nototodarus hawaiiensis 785 hawaiiensis specimens is unclear. Similar zones that occur in statoliths of the deepwater onychoteuthid Moroteuthis ingens might be related to a habitat shift from a pelagic to demersal environment (Jackson, 1993b). The number of increments in the opaque zone was similar for the two species. No zonation was ob- served in any of 43 statoliths examined from Nototodarus sloanii in New Zealand (senior author’s personal observ.) which is predominantly a pelagic species. Nototodarus hawaiiensis adults are trawled day and night on the seafloor and might be demer- sal. Pelagic tows at discrete depths would be useful for establishing the ontogenetic descent in this species. Age and life span of Nototodarus hawaiiensis The deepwater habitat of N. hawaiiensis may pre- vent validation of statolith increment periodicity because of the difficulty of carrying out experiments on live individuals from this habitat. However, some evidence is available on the periodicity of statolith increments in other ommastrephids. Experimental maintenance with chemical markers has shown that statolith increments are laid down daily in the tem- perate north Atlantic Illex illecebrosus (Dawe et al., 1985; Hurley et al., 1985) and in Todarodes pacificus in the north Pacific (Nakamura and Sakurai, 1990, 1991). Furthermore, increment counts on successive cohorts suggest that statolith increments are laid down daily in both Illex argentinus in the south At- lantic (Uozumi and Shiba, 1993) and in Nototodarus sloanii in southern New Zealand waters (Uozumi and Ohara, 1993). We therefore assume that statolith increment periodicity is also daily in N. hawaiiensis. If the assumption of daily periodicity in increment formation for N. hawaiiensis is correct, this suggests that the squid matures early and has a short life span. The species reaches maturity in less than 200 d off the Australian North West Slope, considerably ear- lier than its conspecifics Nototodarus gouldi and N. sloanii in New Zealand waters, which live for about a year (Uozumi and Ohara; 1993, Uozumi et al., 1995). However, specimens obtained from trawls in this study were smaller than the maximum size re- corded for N. hawaiiensis. The largest individuals recorded from Australia were 248 mm ML for a fe- male captured off the Northwest Shelf, and 215 mm ML for a male off southern Queensland (Dunning, 1988b; Dunning and Forch, in press). Based on the regressions in Figure 2, and assuming linear growth throughout the life span, the total number of incre- ments even in these larger individuals would be less than 250. However, A. hawaiiensis females have been reported to reach 290 mm ML in the western Indian Ocean and 318 mm ML in the southeastern Pacific. In contrast, individuals of N. hawaiiensis are much smaller in the Hawaiian and Philippine waters, with maximum sizes of 180 mm and 160 mm ML, respec- tively (Dunning and Forch, in press). There thus appear to be regional differences in maximum size (and possibly age) attained by this species. On the basis of statolith analysis of A. hawaiiensis in Australia, this species may complete its life cycle in less than one year. It has a growth rate and life span comparable to tropical Australian shallow-wa- ter loliginids, which complete their life cycle in less than 200 d (Jackson, 1990; Jackson and Choat; 1992; Jackson and Yeatman 1996). Nototodarus hawaiien- sis spends a considerable proportion of its adult life in deeper, cooler waters compared with the loliginids. However, many oceanic squids spend a proportion of their early life phase in the epipelagic zone (Roper and Young, 1975; Vecchione, 1987; Bigelow, 1992) Therefore, a considerable proportion of the life span of A. hawaiiensis is probably spent in warmer, epi- pelagic waters. The youngest individual captured at depth in this study was 49 d, which suggests that perhaps the first 50 days (approximately 25% of the life span) might be spent in the epipelagic zone. Forsythe (1993) proposed a model of squid growth that predicted that increased temperature during a squid’s early growth phase can dramatically increase its growth rate, resulting in a much larger adult size. This model has recently been validated by seasonal growth data for Lolliguncula brevis (Jackson et al., 1997). Nototo- darus hawaiiensis may therefore reach a larger size more quickly than if it spent most of its life span in cooler waters at depth. The statolith analysis of A. hawaiiensis on the North West Shelf of Australia suggests a much shorter life span than that of other ommastrephids in tropical waters, e.g. Todarodes angolensis in the northern Benguela upwelling (Villanueva, 1992) and Sthenoteuthis pteropus in the tropical Atlantic (Arkhipkin and Mikheev, 1992) which have estimated life spans of around one year. Acknowledgments We thank M.C. Dunning for advice on maturity indi- cators in A. hawaiiensis. We are also thankful for comments from three anonymous reviewers. Literature cited Arkhipkin, A. I. 1993. Age, growth, stock structure and migratory rate of pre-spawning short-finned squid Illex argentinus based on 786 Fishery Bulletin 96(4), I 998 statolith ageing investigations. Fisheries Res. 16: 313-338. 1996. Geographical variation in growth and maturation of the squid ///ex coindetii (Oegopsida, Ommastrephidae) off the north-west African coast. J. Mar. Biol. Assoc. U.K. 76:1091-1106. Arkhipkim, A. I., and A. Mikheev. 1992. Age and growth of the squid Sthenoteuthis pteropus (Oegopsida: Ommastrephidae) from the Central-East Atlantic. J. Exp. Mar. Biol. Ecol. 163:261-276. Bigelow, K. A. 1992. Age and growth in paralarvae of the mesopelagic squid Abralia trigonura based on daily growth increments in statoliths. Mar. Ecol. Prog. Ser. 82:31-40. Collins, M. A., G. M. Burnell, and P. G. Rodhouse. 1995. Recruitment, maturation, and spawning of Loligo forbesi Steenstrup (Cephalopoda: Loliginidae) in Irish waters. ICES J. Mar. Sci. 52:127-137. Bower, J. R. 1996. Estimated paralarval drift and inferred hatching sites for Ommastrephes bartramii (Cephalopoda: Omma- strephidae) near the Hawaiian Archipelago. Fish. Bull. 94:398-411. Dawe, E. G., and P. C. Beck. 1997. Population structure, growth, and sexual maturation of short-finned squid (///ex illecebrosus ) at Newfound- land. Can. J. Fish. Aquat. Sci. 54:137-146. Dawe, E. G., R. K. O’Dor, P. H. Odense, and G. V. Hurley. 1985. Validation and application of an ageing technique for short-finned squid (///ex illecebrosus). J. Northwest Atl. Fish. Sci. 6:07-116. Dunning, M. C. 1988a. Distribution and comparative life history studies of deepwater squid of the family Ommastephidae in Australasian waters. Ph.D diss., Univ. Queensland, Queensland, 288 p. 1988b. First records of Nototodarus hawaiiensis (Berry, 1912) (Cephalopoda, Ommastrephidae) from Northern Australia. Mem. Mus. Viet. 49:59-168. Dunning, M. C., and E. C. Forch. In press. A review of the systematics, distribution and bi- ology of arrow squids of the genus Nototodarus Pfeffer, 1912 (Cephalopoda, Ommastrephidae). Smith. Cont. Zool. 586. Forsythe, J. W. 1993. A working hypothesis on how seasonal temperate change may impact the field growth of young cephalopods. In T. Okutani, R. K. O’Dor, and T. Kubodera (eds.), Recent advances in cephalopod fisheries biology, 133-143. Tokai Univ. Press, Tokyo. Guerra, A., and B. G. Castro. 1994. Reproductive-somatic relationships in Loligo gahi (Cephalopoda: Loliginidae) from the Falkland Islands. Antarctic Science 6:175-178. Hurley, G. V., P. H. Odense, R. K. O’Dor, and E. G. Dawe. 1985. Strontium labelling for verifying daily growth incre- ments in the statolith of the short finned squid (///ex illecebrosus). Can. J. Fish. Aquat. Sci. 42:380-383. Ikeda, Y., Y. Sakurai, and K. Shimazaki. 1991. Development of female reproductive organs during sexual maturation in the Japanese common squid Toda- rodes pacificus. Nippon Suisan Gakkaishi 57:2243-2247. Jackson, G. D. 1990. Age and growth of the tropical near shore loliginid squid Sepioteuthis lessoniana determined from statolith growth ring analysis. Fish. Bull. 88:113-118. 1993a. Seasonal variation in reproductive investment in the tropical loliginid squid Loligo chinensis and the small tropical sepioid Idiosepius pygmaeus. Fish Bull. 91: 260-270. 1993b. Growth zones within the statolith microstructure of the deepwater squid Moroteuthis ingens (Cephalopoda: Onychoteuthidae): evidence for a habitat shift? Can. J. Fish. Aquat. Sci. 50:2366-2374. 1994. Application and future potential of statolith incre- ment analysis in squids and sepioids. Can. J. Fish. Aquat. Sci. 51:2612-2625. 1997. Age, growth and maturation of the deepwater squid Moroteuthis ingens (Cephalopoda: Onychoteuthidae) in New Zealand waters. Polar Biol. 17:268-274. Jackson, G. D., J. Forsythe, R. F. Hixon, and R. T. Hanlon. 1997. Age, growth and maturation of Lolliguncula brevis (Cephalopoda: loliginidae) in the Northwestern Gulf of Mexico with a comparison of length-frequency vs. statolith age analysis. Can. J. Fish. Aquat. Sci. 54:2907-2919. Jackson, G. D., and J. H. Choat. 1992. Growth in tropical cephalopods: an analysis based on statolith microstructure. Can. J. Fish. Aquat. Sci. 49:218-228. Jackson, G. D., and C. C. Lu. 1994. The statolith microstructure of seven species of ant- arctic squid captured in Prydz Bay, Antarctica. Antarctic. Sci. 6:195-200. Jackson, G. D., and P. V. Mladenov. 1994. Terminal spawning in the deepwater squid Moro- teuthis ingens (Cephalopoda: Onychoteuthidae). J. Zool. Lond. 234:189-201. Jackson, G. D., and J. Yeatman. 1996. Variation in size and age-at-maturity in Photololigo (Mollusca: Cephalopoda) from the North West Shelf of Australia. Fish. Bull. 94:59-65. Lipinski, M. R. 1981. Statoliths as a possible tool for squid age deter- mination. Bull. Acad. Pol. Sci. (Sci. Biol.). 28:569-82. Lipinski, M. R., E. E. Dawe and Y. Natsukari. 1991. Introduction. In P. Jerev, Rs. Ragonese, and S.Von Boletzky (eds.), Squid age determination using statoliths, p. 77-81. Proceedings of an international workshop, 9-14 Oct. 1989, Mazara del Vallo, Italy. N.T.R.-I.T.P.P. Special Publication 1. Nakamura, Y., and Y. Sakurai. 1990. On the daily formation of growth increments in the statoliths of Japanese common squid, Todarodes paci- ficus. Bull. Hokkaido Nat. Fish. Res. Inst. 54:1-7. 1991. Validation of daily growth increments in statoliths of Japanese common squid Todarodes pacificus. Nippon Suisan Gakkaishi 57:2007-2011. 1993. Age determination from daily growth increments in statoliths of some groups of the Japanese common squid Todarodes pacificus. In T. Okutani, R. K. O’Dor, and T. Kubodera (eds.), Recent advances in cephalopod fisheries biology: 337-342. Tokai Univ. Press, Tokyo. Rodhouse, P. G., and E. M. C. Hatfield. 1990. Dynamics of growth and maturation in the cephalo- pod ///ex argentinus de Castellanos, 1960 (Teuthoidea: Ommastrephidae). Philo. Trans. R. Soc. Lond. B Biol. Sci. 329:229-241. Rodhouse, P. G., K. Robinson, S. B. Gajdatsy, H. I. Daly, and M. S. J. Ashmore. 1994. Growth, age structure and environmental history in the cephalopod Martialia hyadesi (Teuthoidea: Omma- strephidaw) at the Antarctic Polar Frontal Zone and on the Patagonian Shelf Edge. Antarct. Sci. 6:259-267. Jackson and Wadley: Age, growth, and reproduction of Nototodarus hawaiiensis 787 Roper, C. F. E., M. J. Sweeney, and C. E. Nauen. 1984. Cephalopods of the world: an annotated and illus- trated catalogue of species of interest to fisheries. FAO Fish. Synop. 125(3), 277 p. Roper, C. F. E., and R. E. Young. 1975. Vertical distribution of pelagic cephalopods. Smithsonian Cont. Zool. 209, 51 p. Uozumi, Y., S. Koshida, and S. Kotoda. 1995. Maturation of arrow squids. (Nototodarus gouldi and N. sloanii) with age in New Zealand waters. Fish. Sci. 61:559-565. Uozumi, Y., and H. Ohara. 1993. Growth and age composition of Nototodarus sloanii (Cephalopoda: Oegopsida) based on daily increment counts in statoliths. Nippon Suisan Gakkaishi 59:1469-1477. Uozumi, Y., and C. Shiba. 1993. Growth and age composition of Illex argentinus (Cephalopoda: Oegopsida) based on daily increment counts in statoliths. In T. Okutani, R. K. O’Dor and T. Kubodera (eds.), Recent advances in cephalopod fisheries biology, 591-605. Tokai Univ. Press, Tokyo. Vecchione, M. 1987. Juvenile ecology. In P R. Boyle (ed.) Cephalopod life cycles, vol. II, p. 61-80. Academic Press, London. Villanueva, R. 1992. Interannual growth differences in the oceanic squid Todarodes angolensis Adam in the northern Benguela up- welling system, based on statolith growth increment analysis. J. Exp. Mar. Biol. Ecol. 159:157-177. Voss, G. L. 1962. Six new species and two new subspecies of cephalo- pods from the Philippine Islands. Proc. Biol. Soc. Wash. 75:169-176. Wadley, V. A. 1990. Squid from the west and North West Slope trawl fishery. CSIRO Australia, 12 p. 1993. Cephalopods from demersal trawling on Australia’s North West Slope. In T. Okutani, R.K. O’Dor and T. Kubodera (eds.), Recent advances in cephalopod fisheries biology, p. 607-617. Tokai Univ. Press, Tokyo. 788 Description of pelagic larval and juvenile grass rockfish, Sebastes rastrelliger (family Scorpaenidae), with an examination of age and growth Thomas E. Laidig Keith M. Sakuma Tiburon Laboratory, Southwest Fisheries Science Center National Marine Fisheries Service Tiburon, California 94920 E-mail address (forT E, Laidig): toml@tib.nmfs.gov Abstract.— Pigmentation patterns, meristic characters, morphometric measurements, and head spination were recorded and illustrated for the developmental series of larval and pe- lagic juvenile ( 10.0-27.7 mm SL) grass rockfish, Sebastes rastrelliger. Larvae were identified by the presence of strong dorsal and postanal ventral mid- line pigment and by the presence of pig- ment on the lateral midline. Juveniles became highly pigmented over most of their body and fin surfaces at a length of 27.7 mm. Pigment patterns were suf- ficiently distinct and consistent to dif- ferentiate larval and juvenile S. ras- trelliger from other Sebastes species occurring off central California. Otolith characters were also useful in identifi- cation of this species. Larval and juve- nile S. rastrelliger grew at a rate of 0.36 mm/day, similar to the growth rates observed in other species of Sebastes. Manuscript accepted 11 February 1998 Fish. Bull. 96:788-796 (1998). Rockfishes ( Sebastes spp.) support important commercial and recre- ational fisheries in the northeast- ern Pacific Ocean. During 1995, they accounted for approximately 12.6% of the commercial groundfish catch in weight ( PFMC1 ) and 19.4% of all species in the recreational landings off California, Oregon, and Washington (Witzig et al., 1992). Management of rockfishes has been difficult owing to the large number of species within this genus. Sixty- one species of rockfish are reported from California alone (Eschmeyer et ah, 1983). The need to identify and separate adult rockfish landed by the fishery has long been recognized (Phillips, 1964; Chen, 1971). Larvae and juveniles are far more difficult to differentiate than adults, but the need for their accurate identifica- tion is growing with their use in bio- mass estimates and recruitment studies (Moser and Butler, 1987; Hunter and Lo, 1993; Ralston et al.2). Here we provide means to identify the larvae and juveniles of grass rockfish, S. rastrelliger. Sebastes rastrelliger is a thick- bodied, medium-size (maximum size, 56 cm), bottom-dwelling rock- fish that inhabits shallow, rocky areas (maximum depth, 46 m) from central Baja California to Yaquina Bay, Oregon (Eschmeyer et al., 1983). This species accounts for less than 1% of the commercial ground- fish catch by weight (Pearson3). However, in recent years landings have increased from 1.5 metric tons (t) in 1991 to 52 t in 1995. This may be attributed to a relatively new live fish fishery for this species in south- ern California (Love, 1996). Addi- tionally, Sebastes rastrelliger is also an important part of the nearshore recreational catch (Love, 1996). 1 PFMC (Pacific Fishery Management Council). 1996. Status of the Pacific coast groundfish fishery through 1994 and recommended acceptable biological catches for 1997. Pacific Fishery Management Council, Portland, OR, 168 p. 2 Ralston, S., J. R. Bence, M. B. Eldridge, and W. H. Lenarz. 1993. Estimating the spawning biomass of shortbelly rockfish ( Sebastes jordani ) in the region of Pioneer and Ascension Canyons using a larval pro- duction method, 32 p. Southwest Fisher- ies Science Center, Natl. Mar. Fish. Serv., NOAA, 3150 Paradise Drive, Tiburon, CA 94920. 3 Pearson, D. E. 1994. Southwest Fish. Sci. Center, Natl. Mar. Fish. Serv., 3150 Paradise Drive, Tiburon, CA, 94920. Per- sonal commun., 1997. Laidig and Sakuma: Description of larval and juvenile Sebastes rastrelliger 789 Only a few studies have been conducted on the development of S. rastrelliger. Moreno (1990) de- scribed the pigmentation patterns of laboratory- reared larvae up to 53 days old and 7.2 mm in length. Laroche4 described the pigmentation of a 27.0-mm individual. The purpose of this study is to describe the development of S. rastrelliger from larvae to the pelagic juvenile stage and to examine the age and growth of larvae and juveniles. Methods Specimens of pelagic larval and juvenile S. rastrel- liger were obtained from research cruises conducted aboard the NOAA RV David Starr Jordan. Specimens were collected with a 26 x 26-m midwater trawl (12.7- mm stretched mesh codend liner). Surveys were con- ducted in the spring of 1990, 1992-94, and 1996. Specimens from midwater trawls were frozen for later analysis. All samples were collected off central California between Cypress Point (36°35'N) and Salt Point (38°35'N). We examined pigmentation patterns and physical characteristics of 18 S. rastrelliger larvae and pe- lagic juveniles. Standard length (SL) was measured for each individual, and sizes ranged from 10.0 mm to 27.7 mm. Specimens greater than 19.9 mm were identified by meristic characters (Chen, 1986; Matarese et al., 1989; Moreland and Reilly, 1991; and Laroche4), and pigment patterns were recorded. Specimens less than 20 mm were initially identified from pigment patterns developed from a size series based on the patterns observed by Moreno ( 1990) and the pigment patterns of the smallest, positively iden- tifiable individuals with complete meristic charac- ters. Whenever possible, dorsal, anal, and pectoral- fin ray counts and the number of gill rakers on the first gill arch were recorded and subsequently used in identifications. Gill-raker counts were obtained only from fish larger than 15 mm in length. Snout to anus length, head length, snout length, eye diameter, body depth at the pectoral-fin base, body depth at anus, and pectoral-fin length were measured on 17 specimens ranging from 10.0 to 27.7 mm. Terminology for morphometries follows Richard- son and Laroche (1979). Fifteen specimens (ranging from 10.0 to 27.7 mm) were stained with alizarin red-S and examined for head spination. Terminology for head spination fol- lows Richardson and Laroche (1979). 4 Laroche, W. A. 1987. Guide to larval and juvenile rockfishes ( Sebastes ) of North America. Box 216, Enosburg Falls, VT 05450. Unpubl. manuscript, 311 p. Otoliths were removed from 16 larvae and juve- niles (10.0 to 27.7 mm) and ages were determined following the procedures in Laidig et al. (1991). Ad- ditionally, extrusion check radius was measured fol- lowing the procedures in Laidig and Ralston ( 1995). Transformation from the larval stage to the juvenile stage was ascertained by the presence of secondary primordia in the otolith (Laidig et al., 1991). Results General development At 10.0 mm, S. rastrelliger larvae had already un- dergone flexion, and a full complement of adult fins rays was present (13 dorsal, 6 anal, and 19 pectoral- fin rays), although the spinous dorsal rays were not quite fully developed (Table 1; Fig. 1C). Meristic counts were similar to those reported by Laroche4 and Moreland and Reilly (1991). In our sample, the earliest signs of transformation were observed in a specimen 20.9 mm in length, and all specimens 27.7 mm or larger had completed transformation. Lateral line pores were present only in the largest individual; therefore a full complement was never observed dur- ing this study. Morphometries Notable changes in body proportion for S. rastrelliger occurred between lengths of 10.0 and 13.8 mm and during juvenile transformation (Table 2). At 10.0 mm, larvae had large heads in relation to their body size, whereas larger larvae became more elongate and thus decreased the proportion of head size in relation to body Table 1 Frequency of occurrence of dorsal, anal, and pectoral-fin ray, and gill-raker counts in larval and juvenile grass rock- fish, Sebastes rastrelliger. Frequency of Percent Character Count occurrence occurrence Dorsal-fin rays 12 2 Hi 13 16 88.9 Anal-fin rays 6 18 100 Pectoral-fin rays 18 2 11.8 19 15 88.2 Gill rakers 21 2 16.7 22 6 50.0 23 3 25.0 24 1 8.3 790 Fishery Bulletin 96(4), 1998 Figure 1 Developmental series of grass rockfish, Sebastes rastrelliger. (A) 4.6 mm SL yolksac preflexion larva; (B) 7.2 mm SL flexion larva; (C) 10.0 mm SL larva; (D) 13.8 mm SL larva; (E) 18.3 mm SL larva; (F) 21.6 mm SL transforming larva; (G) 27.7 mm SL pelagic juvenile. Illustrations A and B (4.6 and 7.2 mm SL, respec- tively) from Moreno, 1993. length. After transformation, the fish became more thick-bodied, as evidenced by the increase in body depth at the pectoral-fin ray base and at the anus. Head spination Several head spines had already developed by a length of 10.0 mm (Table 3). The nasal, postocular, parietal, nuchal, pterotic, inferior posttemporal, supracleithral, operculars, and preoperculars were all well formed, whereas the 1st inferior and 1st su- perior infraorbital were barely perceptible at 10.0 mm. By 13.8 mm, the preocular, 2nd inferior infraor- bital, and the superior posttemporal were developed. At 18.8 mm, the 4th superior infraorbital was formed. The tympanic formed at 19.7 mm. The pterotic be- came overgrown and embedded by 21.6 mm. The supraocular, coronal, 3rd inferior infraorbital, and 2nd and 3rd superior infraorbital were not apparent at any length examined. Body pigmentation A distinct pattern of heavy dorsal and postanal ven- tral midline pigment, along with strong lateral mid- line pigment, was characteristic at a length of 10.0 Laidig and Sakuma: Description of larval and juvenile Sebastes rastrelliger 791 Figure 1 (continued) mm (Fig. 1C; Table 4). Pigment was also evident along the anterior tip of the lower jaw, on the snout anterior to the eye orbit, on the top of the cranium, on the operculum, and along the dorsal and poste- rior margin of the gut cavity. Pigmentation patterns at a length of 13.8 mm were similar to, but more intensely developed than, those at 10.0 mm (Fig. ID). Pigment advanced in the ante- rior direction along the postanal ventral and lateral midline, and to a lesser degree along the dorsal mid- line. In addition, pigment along the posterior-ven- tral portion of the eye orbit had begun to form. At a length of 18.3 mm, the dorsal half of the fish had become highly pigmented (Fig. IE). Nape pig- ment had developed and merged with the dorsal mid- line pigment, forming a continuous band from the parietal and nuchal spines to the beginning of the caudal fin at this size. Snout and cranial pigment also had merged to form a solid line of pigment from the upper jaw to the parietal and nuchal spines at a 792 Fishery Bulletin 96(4), 1 998 Table 2 Morphometric measurements (in mm) of grass rockfish, Sebastes rastrelliger. SL (mm) Snout-anus length Head length Snout length Eye diameter Body depth at pectoral-fin ray base Body depth at anus Pectoral fin length 10.0 6.4 3.7 1.2 1.2 2.9 2.4 1.7 13.8 7.9 4.5 1.5 1.6 3.4 3.1 2.9 16.7 8.4 5.0 1.7 1.8 3.8 3.3 3.4 16.7 8.4 5.0 1.7 1.8 3.8 3.3 3.4 16.8 8.9 5.4 1.8 1.9 3.8 3.4 3.4 17.4 10.2 5.8 2.1 1.9 3.9 3.3 3.6 18.0 10.5 5.9 2.2 2.0 4.3 3.4 3.8 18.3 10.5 6.2 2.2 2.0 4.3 3.9 3.7 18.8 10.8 6.3 2.3 2.1 4.5 3.9 3.8 19.2 10.9 6.5 2.4 2.1 4.6 3.7 4.0 19.3 11.1 6.8 2.5 2.2 4.6 3.8 4.2 19.7 11.4 6.9 2.5 2.2 4.8 3.9 4.3 20.9 11.7 7.2 2.7 2.3 5.1 4.4 4.5 21.6 12.8 7.5 2.9 2.4 5.9 5.2 4.8 22.7 13.7 7.9 3.0 2.5 6.2 5.5 5.2 24.3 14.0 8.7 2.5 2.7 6.7 5.7 5.8 27.7 15.9 9.0 2.6 2.9 7.6 6.5 6.7 length of 18.3 mm. An unpigmented section was con- spicuous between the nape and cranial regions. Pig- ment on the flanks of the dorsal body intensified, especially in between the myomeres. Lateral mid- line pigment extended from the caudal peduncle to the gut cavity. Pigment on the lateral midline near the caudal peduncle coalesced into a large pigmented area, which decreased in intensity anteriorly. Much of the dorsal half of the operculum was covered with pigment. At 18.3 mm, the ventral half of the body remained mostly unpigmented, except for the presence of a series of melanophores on the ventral surface ante- rior to the gut cavity and between the branchiostegals (Fig. IE). Pigment at the anterior tip of the lower jaw remained similar in intensity to that of the 13.8- mm specimen. Postanal ventral midline pigment became reduced, especially at the base of the anal fin. Pigment was also absent between the ventral myomeres. By 21.6 mm, pigmentation increased over the en- tire body surface (Fig. IF). The dorsal midline pig- ment merged with the nape, head, snout, and ante- rior tip of the upper jaw pigment to form a solid line along the dorsal surface of the body from the mouth to the caudal fin. Pigment formed on the ventral half of the membranes of the spinous dorsal fin. Pigment increased around the ventral surface of the eye or- bit. Nape, head, and operculum pigment merged to form a vertical bar from the operculum to the dorsal midline, except for a small unpigmented area near the nuchal and parietal spines. Hypural pigment began to form, and the caudal peduncle became highly pigmented. The ventral surface of the body remained only lightly pigmented. A few pigment spots were observed at the base of the pelvic fin. Juvenile S. rastrelliger became highly pigmented throughout most of their body surface by 27.7 mm (Fig. 1G). The entire dorsal surface of the body was completely pigmented. Pigment also covered much of the ventral surface of the body. The gut cavity, the ventral surface of the head, and the area just above the anal fin still remained largely unpigmented. Some pigment was observed at the base of the pecto- ral fin. A line of light pigment occurred along the lat- eral midline. Hypural pigment became intense, but an area of decreased pigment was evident just ante- rior to the hypural area. The spinous and soft dorsal fins were highly pigmented, except for the membrane between spine 11 and 12 which was completely de- void of pigment. A few pigment spots occurred on the anal and caudal fins. Otolith examination A linear model provided a good fit for the relation- ship of total otolith radius versus SL (r2=0.94)(Fig. 2). The growth rate of S. rastrelliger was estimated by a linear model fitted to the relationship of SL and age (coefficient of determination, r2=0.88)(Fig. 3). This model estimated a growth rate of 0.36 mm/day and an extrusion SL of 0. 15 mm. The extrusion check radius for S. rastrelliger ranged from 13.3 to 14.6 pm, averaging 14.0 pm (SD=0.39). Secondary primordia Laidig and Sakuma: Description of larval and juvenile Sebastes rastrelliger 793 Development of head spines in Table 3 grass rockfish, Sebastes rastrelliger. “1” means spine present and “0” means spine absent. Standard length (mm Spines 10.0 13.8 16.7 17.4 18.0 18.3 18.8 19.2 19.3 19.7 20.9 21.6 22.7 24.3 27.7 Nasal 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Preocular 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Supraocular 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Postocular 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Coronal 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Tympanic 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 Parietal 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Nuchal 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Pterotic 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 Posttemporals Superior 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 Inferior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Supracleithral 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Operculars Superior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Inferior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Preoperculars 1st anterior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2nd anterior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3rd anterior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1st posterior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2nd posterior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3rd posterior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 4th posterior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 5th posterior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Infraorbitals 1st inferior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2nd inferior 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3rd inferior 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1st superior 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2nd superior 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3rd superior 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4th superior 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 Standard length (mm) Figure 2 Plot of total otolith radius and standard length of grass rockfish, Sebastes rastrelliger (n = 16). Solid line indicates predicated values from linear model. 794 Fishery Bulletin 96(4), 1998 Table 4 Pigment occurrence at various pigment loci for grass rockfish, Sebastes rastrelliger. SL = standard length in mm. Definitions of pigment loci are given below. 0.0 = no pigment, 1.0 = some pigment present, and 2.0 = area heavily pigmented. LJ = anterior tip of the lower jaw, EYE = posterior-ventral edge of the eye orbit, HEAD = cranial surface, FACE = dorsal surface anterior to the eyes, OPER = operculum, CHK = radiating cheek bars, NAPE = nape pigment, DORS = dorsal body surface, VENT = ventral body surface, MID = along the lateral midline, HYP = hypural region, DFIN = spinous dorsal fin, AFIN = anal fin, PEC = blade of the pectoral fin. SL LJ EYE HEAD FACE OPER CHK NAPE DORS VENT MID HYP DFIN AFIN PEC 10.0 2.0 0.0 2.0 2.0 2.0 0.0 1.0 2.0 2.0 1.0 0.0 0.0 0.0 0.0 13.8 2.0 2.0 2.0 2.0 2.0 0.0 1.0 2.0 2.0 2.0 1.0 0.0 0.0 0.0 16.7 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 0.0 0.0 0.0 16.7 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 0.0 0.0 0.0 16.8 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 0.0 0.0 0.0 17.4 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 0.0 0.0 0.0 17.5 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 0.0 0.0 0.0 18.0 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 0.0 0.0 0.0 18.3 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 0.0 0.0 0.0 18.8 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 1.0 0.0 0.0 19.2 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 0.0 0.0 0.0 19.3 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 1.0 0.0 0.0 19.7 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 0.0 0.0 0.0 0.0 20.9 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 1.0 1.0 0.0 0.0 21.6 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 1.0 2.0 0.0 0.0 22.7 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 1.0 2.0 0.0 0.0 24.3 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 1.0 2.0 0.0 0.0 27.7 2.0 2.0 2.0 2.0 2.0 0.0 2.0 2.0 2.0 2.0 2.0 2.0 1.0 0.0 first appeared in the otoliths of the 20.9-mm speci- men and was subsequently observed in the otoliths of all specimens larger than 20.9 mm. By 27.7 mm, the otolith was completely encircled by secondary primor- dia, signaling the end of juvenile transformation. Discussion Sebastes rastrelliger have developed a unique pig- ment pattern by 10 mm (Fig. 1C), which allows them to be distinguished from other Sebastes spp. At this size, the presence of dorsal, ventral, and lateral body midline pigment is distinctive for only three species (S. dallii [Moser and Butler, 1981], S. rufus [Moser and Butler, 1987], and S. saxicola [Laidig et al., 1996]) that co-occur with S. rastrelliger off Califor- nia. All four species have anterior lower jaw, snout, and opercular pigments, which could add to the con- fusion between these species. One major difference in pigmentation between these species is that S. dallii and S. rufus have pectoral-fin pigment. We found no pectoral-fin pigmentation in S. rastrelliger larger than 10.0 mm, and Moreno (1990) found no pectoral-fin pigment for individuals smaller than 8 mm. Sebastes saxicola also lacks pectoral-fin pig- ment in fish smaller than 20 mm (Laidig et al., 1996). Sebastes saxicola and S. rastrelliger can exhibit simi- lar pigmentation at 10 mm. This similarity occurs until approximately 18 mm, when S. saxicola devel- ops hypural and dorsal fin pigment and when sad- dling along the dorsal surface eventually changes into a barred pattern. Sebastes rastrelliger was also found to have a relatively small pectoral fin for its SL (Table 2) in comparison with other Sebastes spp. (Moser et al., 1977; Richardson and Laroche, 1979; Laroche and Richardson, 1980, 1981; Sakuma and Laidig, 1995; Laidig et al., 1996) Meristics can aid in differentiating S. rastrelliger from other species. The average counts of 13 dorsal, 6 anal, and 19 pectoral-fin rays (Table 1) are similar to those observed by Moreland and Reilly (1991) and Laroche.4 Moreland and Reilly (1991) found this com- bination of fin-ray counts typical for only two spe- cies: S. rastrelliger and S. babcocki. Sebastes babcocki has a distinct barred pigmentation pattern and typi- cal gill-raker counts of 30-31. Sebastes rastrelliger had no barred pattern and average gill-raker counts of 22. This count is lower than that observed by Moreland and Reilly (1991) (an average of 25) and Laroche4 (an average of 24). Gill-raker counts may be low in our study because the gill rakers may not have fully developed in some of our specimens (Sakuma and Laidig, 1995). Therefore, the use of pigment patterns in conjunction with meristics should aid in the identification of S. rastrelliger. Laidig and Sakuma: Description of larval and juvenile Sebastes rastrelliger 795 Otolith characters can also be useful in separat- ing Sebastes rastrelliger from some other Sebastes species. Laidig et al. (1996) showed that the extru- sion check radius of S. rastrelliger ( 14.0 pm) was sig- nificantly different from those of S. saxicola, S. maliger, S. atrovirens , and the copper complex (S. carnatus, S. caurinus, and S. chrysomelas). It was not different from those of S. auriculatus or S. semicinctus. However, these latter two species have distinctly different meristic counts and pigmentation patterns. In addition, the extrusion check radii found by Laidig and Ralston (1995) for S. paucispinis, S. flavidus, S. entomelas, and S. mystinus were smaller (10.93-12.20 pm) than that for S. rastrelliger, whereas the extrusion check radii for S.jordani and S.goodei were much larger (15.15-16.96 pm). Postflexion Sebastes rastrelliger, from 10.0 to 27.7 mm, grew at a rate of 0.36 mm/day. Moreno (1990) found that the growth rate for reared preflexion S. rastrelliger less than 8 mm was 0.07 mm/day. This slow growth rate may be attributed to laboratory rearing of the fish. However, slow growth rates have been shown for the first few weeks in S. goodei (Sakuma and Laidig, 1995), S. saxicola (Laidig et al., 1996), and S.jordani (Laidig et al., 1991). Sakuma and Laidig ( 1995 ) observed growth rates of 0. 135 mm/ day in S. goodei less than 40 days old whereas Woodbury and Ralston (1991) noted growth rates of 0.399-0.555 mm/day in specimens 35-170 days old. Laidig et al. (1996) observed growth rates of 0.125 mm/day in S. saxicola less than 40 days old, whereas larvae and juveniles older than 40 days exhibited in- creased growth rates of 0.367 mm/day. Laidig et al. (1991) determined that in S.jordani there were large fluctuations in growth rates associated with flexion, with relatively slow growth prior to flexion, almost no growth during flexion, and relatively rapid growth af- ter flexion. Because the S. rastrelliger in this study had already undergone flexion, we expected a faster growth rate than that of smaller fish as recorded by Moreno (1990). For this reason, our growth curve does not ac- curately express the growth in the first few weeks of life; this lack of fit is consistent with the smaller than expected estimate ofSL at extrusion (0.15 mm). There- fore, extrapolating growth rates for flexion and preflexion larvae from our model would not be advised. Acknowledgments We would like to thank the officers and crew of the RV David Starr Jordan and the scientific personnel onboard who assisted in the collection of larval and juvenile fish. We also thank Mary Nishimoto for help in identifications during this study. Literature cited Chen, L. 1971. Systematics, variation, distribution, and biology of rockfishes of the subgenus Sebastomus (Pisces, Scor- paenidae, Sebastes). Univ. California Press, Berkeley, CA, 115 p. 1986. Meristic variation in Sebastes (Scorpaenidae), with an analysis of character association and bilateral pattern and their significance in species association. U. S. Dep. Commer., NOAATech. Rep. NMFS 45, 17 p. Eschmeyer, W. N., E. S. Herald, and H. Hammann. 1983. A field guide to Pacific coast fishes. Houghton Mifflin Company, Boston, MA, 336 p. Hunter, J. R., and N. C. -H. Lo. 1993. Ichthyoplankton methods for estimating fish biomass introduction and terminology. Bull. Mar. Sci. 53:723-727. Laidig, T. E., and S. Ralston. 1995. The potential use of otolith characters in identifying larval rockfish ( Sebastes spp.). Fish. Bull. 93:166-171. Laidig, T. E., S. Ralston, and J. R. Bence. 1991. Dynamics of growth in the early life history of shortbelly rockfish, Sebastes jordani. Fish. Bull. 89: 611-621. Laidig, T. E., K. M. Sakuma, and M. M. Nishimoto. 1996. Description of pelagic larval and juvenile stripetail rockfish, Sebastes saxicola (family Scorpaenidae), with an examination of larval growth. Fish. Bull. 94:289-299. Laroche, W. A., and S. L. Richardson. 1980. Development and occurrence of larvae andjuveniles of the rockfishes Sebastes flavidus and Sebastes melanops (Scorpaenidae) off Oregon. Fish. Bull. 77(4):901-922. 1981. Development of larvae andjuveniles of the rockfishes Sebastes entomelas and S. zacentrus (family Scorpaenidae) and occurrence off Oregon, with notes on head spines of S. mystinus , S. flavidus, and S. melanops. Fish. Bull. 79(2): 231-257. Love, M. 1996. Probably more than you wanted to know about the fishes of the Pacific coast. Really Big Press, Santa Bar- bara, CA, 381 p. Matarese, A. C., A. W. Kendall Jr., D. M. Blood, and B. M. Vinter. 1989. Laboratory guide to early life history stages of north- east Pacific fishes. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 80, 652 p. Moreland, S. L., and C. A. Reilly. 1991. Key to the juvenile rockfishes of central Cali- fornia. In T. E. Laidig and P. B. Adams (eds.), Methods used to identify pelagic juvenile rockfish (genus Sebastes) occurring along the coast of central California, p. 59- 180. U.S. Dep. Commer., NOAA Tech. Memo., NOAA-TM- NMFS-SWFSC-166, 180 p. Moreno, G. 1990. Description of the larval stages of five northern Cali- fornia species of rockfishes (Family Scorpaenidae) from rearing studies. M.S. thesis, California State Univ., Stanislaus, CA, 68 p. 1993. Description of early larvae of four northern Califor- nia species of rockfishes (Scorpaenidae: Sebastes) from rear- ing studies. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 116, 18 p. Moser, H. G., and J. L. Butler. 1981. Description of reared larvae and early juveniles of the calico rockfish, Sebastes dallii. Calif. Coop. Oceanic. Fish. Invest. Rep. XXIL88-95. 796 Fishery Bulletin 96(4), 1998 1987. Descriptions of reared larvae of six species of Sebastes. In W. H. Lenarz, and D. R. Gunderson (eds.), Widow roekfish. Proceedings of a workshop, Tiburon, Cali- fornia, December 11-12, 1980, p. 19-29. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 48. Moser, H. G., E. H. Ahlstrom, and E. M. Sandknop. 1977. Guide to the identification of scorpionfish larvae (Fam- ily Scorpaenidae) in the eastern Pacific with comparative notes on species of Sebastes and Helicolenus from other oceans. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 402, 71 p. Phillips, J. B. 1964. Life history studies on ten species of roekfish. Calif. Dep. Fish Game Fish Bull. 126, 70 p. Richardson, S. L., and W. A. Laroche. 1979. Development and occurrence of larvae and juveniles of the rockfishes Sebastes crameri, Sebastes pinniger, and Sebastes heluomaculatus (Family Scorpaenidae) off Oregon. Fish. Bull. 77(l):l-46. Sakuma, K. M., and T. E. Laidig. 1995. Description of larval and pelagic juvenile Sebastes goodei with an examination of larval growth. Fish. Bull. 93:720-730. Witzig, J. F., M. C. Holliday, R. J. Essig, and D. L. Sutherland. 1992. Marine recreational fisheries statistics survey, Pa- cific coast, 1987-1989. National Marine Fisheries Service, Silver Springs, MD, 367 p. Woodbury, D. P., and S. Ralston. 1991. Interannual variation in growth rates and back-cal- culated birthdate distributions of pelagic juvenile rock- fishes ( Sebastes spp. ) off the central California coast. Fish. Bull. 89:523-533. 797 Changes in the sex ratio and size at maturity of gag, Mycteroperca microlepis, from the Atlantic coast of the southeastern United States during 1 976-1 995* John C. McGovern David M. Wyanski Oleg Pashuk Marine Resources Research Institute South Carolina Department of Natural Resources RO. Box 12559, Charleston, South Carolina 29422-2559 E-mail address (forJ.C. McGovern). mcgovernj@mrd. dnr.state.se. us George R. Sedberry Marine Resources Research Institute South Carolina Department of Natural Resources RO. Box 12559, Charleston, South Carolina 29422-2559 Abstract.-Gag, Mycteroperca microlepis, is a large, slow-growing, protogynous grouper that probably makes annual migrations to specific locations to aggregate for spawning. During 1976-82, male gag constituted 19.6% of the sexually mature individu- als taken during fishery-dependent and fishery-independent sampling along the southeast coast of the United States. A similar percentage of males was found in the Gulf of Mexico from 1977 to 1980; however, males made up only 1.9% of the population in the Gulf of Mexico during 1992. To assess the current sex ratio of gag along the south- east U.S. coast, an emergency rule was enacted by the Department of Com- merce in January 1995 that required commercial vessels from North Caro- lina to southeast Florida to land gag with gonads intact. Histological exami- nation of 2613 gonads of sexually ma- ture gag collected from 18 January through 18 April 1995 revealed that 5.5% of the gag from the southeast At- lantic were male. There was a weak trend indicating that females reached maturity at a smaller size in 1994-95 than in 1976-82. Very few transitional specimens were collected during the spawning season. Most transitional in- dividuals (79%) were taken during April through June immediately after the 1995 spawning season. Gag in spawning condition were landed dur- ing December through mid-May by fish- ermen working offshore from North Carolina to southeast Florida. In addi- tion, gag in spawning condition were taken during research cruises docu- menting the occurrence of spawning north of Florida (off South Carolina and Georgia at depths ranging from 49 to 91 m). Manuscript accepted 13 January 1998. Fish. Bull. 96:797-807 (1998). Charles S. Manooch II Southeast Fisheries Science Center National Marine Fisheries Service, NOAA Beaufort, North Carolina 28516-9722 Gag, Mycteroperca microlepis, is a large, slow-growing serranid asso- ciated with inshore-reef and shelf- break habitats in the western At- lantic from New York to Brazil and in the Gulf of Mexico (Smith, 1971; Huntsman, 1976; Hardy, 1978; Collins et al., 1987). Gag are proto- gynous and probably make annual late-winter migrations to specific lo- cations to form spawning aggrega- tions (Collins et al., 1987; Keener et al., 1988; Van Sant et al., 1994). Recent evidence indicates that males may be selectively removed because they are the largest and most aggressive individuals in a spawning aggregation and are the first to be taken by fishing gear (Gilmore and Jones, 1992). Effects of fishing on spawning aggregations in other grouper species have been found to be deleterious on popula- tion size, sex ratio, genetic diversity, and behavior of individuals (Nelson and Soule, 1987; Smith et al., 1991; Carter et al., 1994; Coleman et al., 1996). Histological examination of go- nads of 498 gag collected through- out the year during 1976-82 along the southeast coast of the United States revealed that 84% were fe- males, 15% males, and 1% were un- dergoing transition from female to male (Collins et al., 1987). Hood and Schlieder (1992) found similar sex ratios (14.0% male) for Gulf of Mexico gag collected from 1977 to * Contribution 409 of the South Carolina Resources Center, South Carolina Depart- ment ofNatural Resources, Charleston, SC 29422-2559. 798 Fishery Bulletin 96(4), I 998 1980. In both of these earlier studies, immature fe- males were included in the sex ratio. In 1992, Coleman et al. (1996) determined that males made up a smaller percentage (1.9%) of the Gulf of Mexico adult (sexually mature) population than did males in the earlier study by Hood and Schlieder (1992). No current estimate of sex ratio is available for gag along the southeast coast of the United States. In 1993, the South Atlantic Fishery Management Council (SAFMC) and the Florida Marine Fish Com- mission considered a closure for fishing of gag along the southeast coast during the gag spawning season to prevent the possibility that fishing on aggrega- tions could severely decrease spawning output and affect sex ratios for future spawning seasons and generations (SAFMC, 1993). Much of the impetus for the proposed closure was the heavy fishing pressure exerted by recreational and commercial fishermen on spawning aggregations along the narrow Florida shelf. However, the SAFMC elected to take no action because it was unknown if gag spawned north of Florida and because the impacts of a spawning sea- son closure for the entire southeast coast might be an unnecessary hardship on fishermen. In addition, no recent data were available on the sex ratio of gag along the southeast coast. Owing to the paucity of data on the current reproductive status of gag, the U.S. Department of Commerce followed recommen- dations from SAFMC and the National Marine Fish- eries Service (NMFS) and enacted an emergency rule requiring all commercial vessels along the Atlantic coast of the southeastern United States to land gag in whole condition so that sex ratios and reproduc- tive condition could be assessed. In this paper we compare current sex ratio and maturity data ( 1994- 95) with historical information (1976-82). An addi- tional objective of this work was to determine if gag spawn north of Florida. Methods Gag reproductive data were collected during 1976- 82 and 1994-95 by the Marine Resources Monitor- ing Assessment and Prediction (MARMAP) program at South Carolina Department of Natural Resources (SCDNR). Data collected from 1976 to 1982 were obtained from the commercial hook and line fishery or research cruises conducted by the MARMAP pro- gram. These data (1976-82) were published by Collins et al. (1987), but raw data were available to us for comparison with recent information. During 1994-95, the collection of most gag resulted from an emergency rule that required commercial fishermen along the southeast Atlantic coast to land all gag with gonads intact from 18 January through 18 April 1995. The rule was extended through 17 July 1995 but applied only to gag larger than 889 mm (TL). Upon reaching the dock, fishermen were instructed to con- tact a NMFS or state port sampler assigned to the region. The port samplers recorded date, port, ap- proximate area of fishing, and total length (TL) and weight of specimens. All length measurements in our study refer to TL. Gonads were obtained from each specimen, packed in ice or preserved in formalin, and shipped to personnel of the MARMAP program. Ad- ditional samples were collected by MARMAP during 1994 and 1995 with chevron traps and hook-and-line gear at randomly chosen reef areas off the southeast- ern United States (Collins and Sedberry, 1991; Cuellar et al., 1996) and through port sampling efforts. To assess sex and reproductive state, all gonads collected during 1994-95 were prepared for histologi- cal examination. The posterior portion of the gonad was removed and preserved in 10% formalin buff- ered with seawater. After a 2-6 week fixation, the tissue samples were transferred to 50% isopropanol, processed and vacuum infiltrated in a modular vacuum tissue processor, and blocked in paraffin. The resulting imbedded samples were sectioned at 7 pm, stained with double-strength Gill hematoxy- lin, and counter-stained with eosin-y. A similar tech- nique was used for the 1976-82 samples (Collins et al., 1987). Sex and reproductive stage were assessed by one reader according to histological criteria (Table 1). Sections from 100 randomly selected specimens were examined by a second reader early in the study to verify the assessments. Specimens with developing, ripe, spent, or resting gonads were considered to be sexually mature. For females, this definition of sexual maturity included specimens with oocyte develop- ment at or beyond the cortical granule (alveoli) stage and specimens with beta, gamma, or delta stages of atresia. To ensure that females were correctly as- signed to the immature and resting categories, the length-frequency histogram for immature females was compared to the histograms for resting females and females with evidence of certain maturity (e.g. developing, ripe, or spent). If there was little or no overlap between the two histograms representing mature individuals and the histogram for immature females, it was assumed that immature and resting individuals were not being confused. To produce a sex ratio for the adult portion of the population, only data for mature gag were included in analyses. Data from Collins et al. (1987) and Hood and Schlieder (1992) were analyzed again to deter- mine a sex ratio based on mature gag so that the results of these studies and the present study could McGovern et al.: Changes in the sex ratio and size at maturity of Mycteroperca microlepis 799 Tabie 1 Histological criteria used to determine reproductive state in gag, Mycteroperca microlepis (based on Moe, 1969; Hunter and Macewicz, 1985; Hunter al. 1986; Collins et al., 1987; West, 1990; Ferreira, 1993; Shapiro et al., 1993a, 1993b). Reproductive state Male Female Immature No primary males found. “Immature males” in Moe (1969) considered late transitional because sex transition is not yet completed. No evidence of atresia. In comparison to resting female, oogonia more abundant along the margin of lamellae, most previtellogenic oocytes <80 pm in diameter, area of transverse section of ovary is smaller, lamellae lack muscle and connective tissue bundles, lamellae are not as elongate, and ovarian wall is thinner. Developing Development of cysts containing primary and secondary spermatocytes through some accumulation of spermatozoa in lobular lumina and peripheral sinuses within gonadal wall. Oocytes undergoing cortical granule (alveoli) formation through nucleus migration and partial coalescence of yolk globules. Running ripe Predominance of spermatozoa in lobules and peripheral sinuses. Little or no occurrence of spermatogenesis. Completion of yolk coalescence and hydration in the most advanced oocytes. Zona radiata becomes thin. Postovulatory follicles sometimes present. Spent No spermatogenesis. Some residual sperma- tozoa in lobules and peripheral sinuses. More than 50% of vitellogenic oocytes in alpha or beta stage of atresia. Resting Little or no spermatocyte development. Empty lobules and sinuses evident. Traces of atresia. In comparison to immature female, oogonia less abundant along margin of lamellae, most previtellogenic oocytes >80 pm in diameter, area of transverse section of ovary is larger, lamellae have muscle and connective tissue bundles, lamellae are more elongate and convoluted, and ovarian wall is thicker. Developing, recent spawn Not assessed. Developing stage as described above plus presence of postovulatory follicles. Mature specimen, state unknown Mature, but inadequate quantity of tissue or postmortem histolysis prevent further assessment of reproductive state. Mature, but inadequate quantity of tissue or postmor- tem histolysis prevents further assessment of reproduc- tive state. Transitional Proliferation of spermatogonia through limited spermatogenesis within lamellae of resting ovary and development of peripheral sinuses in musculature of ovarian wall. be compared. The sex ratio for 1995 was based on mature individuals that were collected during the early portion of the emergency rule (18 January to 18 April 1995). Recent data were analyzed by region: North Carolina, South Carolina, Georgia, northern Florida, and southern Florida (south of New Smyrna Beach). Plots of reproductive seasonality were based on all individuals that were collected during 1994 and 1995. Females that possessed hydrated oocytes or postovulatory follicles were considered to be in spawning condition. Data for all females collected during the two peri- ods were used in size-at-maturity analyses. The probit procedure (SAS, 1990) was used to fit a gompit model with the Gompertz distribution function to maturity data in 5-cm-TL intervals. This procedure provided estimates of length at 50% maturity (L50 ) and a comparison of size at maturity between peri- ods. Analysis of variance (AN OVA) and the Duncan multiple range test were used to determine if there was a significant difference in the mean size of gag collected by month for mature individuals. AN OVA and the Duncan multiple range test were also used to determine if there was a significant difference in the monthly mean size of gag between 889 and 1050 mm TL to reduce the potential for bias that might result from selection of larger fishes during the lat- ter part of the emergency rule. The results of all sta- tistical tests were considered significant if P was <0.05. 800 Fishery Bulletin 96(4), 1 998 Results Histological examination of 2606 sexually mature gag (size range=5 17-1275 mm TL) taken during 18 Janu- ary through 18 April 1995 revealed that there were significantly (%2=66.852; P <0.001) fewer males (5.5%) along the southeast coast of the United States than during 1976-82 ( 19.6%; Tables 2 and 3). There were no significant differences in the percentage of males captured in 1995 off North Carolina, South Carolina, Georgia, and southern Florida (/2 =4.926; P> 0.05; Table 4). However, the percentage of males taken in northern Florida (14.9%) was significantly greater than in all other regions combined (/2 =133.160; P<0.001). Four collections accounted for over 35% of the males that were collected in northern Florida. The size range of gag captured during 1994 and 1995 was similar to those caught from 1976 to 1982; however, there were some differences in size between periods for both sexes (Tables 3 and 5). Larger ma- ture female gag were collected in 1994-95, although the mean lengths for the two periods were similar (1976-82: x TL=831 mm, SD=92; 1994-95: x TL = 832 mm, SD=82). The smallest mature female was Table 2 Sex ratio of gag, Mycteroperca microlepis, (M=Male; F=Female; T=Transitional) taken during various studies in the northeast Gulf of Mexico and southeastern United States between 1976 and 1995. Sex ratios based on sexually mature individuals. Study Dates Area No. of M % M No. ofF % F No. of T % T This study 1976-1982 SE Atlantic 104 19.6 419 78.9 8 1.5 Hood and Schlieder (1992) 1977-1980 NE Gulf of Mexico 125 15.6 659 82.5 15 1.9 Coleman et al. (1996) 1992 NE Gulf of Mexico 9 1.9 457 97.3 3 0.6 This study 18 Jan-18 April 19957 SE Atlantic 143 5.5 2468 94.4 2 0.1 This study 1994-1995 SE Atlantic 256 6.4 3720 92.7 39 0.9 ' Period of emergency rule when fishermen were required to land all gag with gonads intact. Table 3 Number of sexually mature gag, Mycteroperca microlepis , collected by when fishermen were required to land all gag with intact gonads (TL size class during 1976- =total length). Trans = -82 and 18 January to 18 April 1995 transitional specimens. Size (mm TL) 1976-1982 1995 Female Male Trans Female Male Trans 450-499 500-549 2 — — 2 — — 550-599 1 — — 12 — — 600-649 10 — — 19 — — 650-699 14 — — 79 — — 700-749 31 — — 171 — — 750-799 64 1 2 437 — 800-849 94 1 1 709 — 1 850-899 79 1 2 580 5 1 900-949 42 4 1 271 3 — 950-999 7 13 1 99 26 — 1000-1049 13 14 — 48 48 — 1050-10°9 5 27 — 27 36 — 1100-1149 — 14 — 6 13 — 1150-1199 — 2 — 1 7 — 1200-1249 — 1 — — 5 — 1250-1299 No length 47 25 1 7 — — Total 419 104 8 2468 143 2 Percent 78.9 19.6 1.5 94.4 5.5 0.1 McGovern et al.: Changes in the sex ratio and size at maturity of Mycteroperca microlepis 801 Table 4 Number of sexually mature gag, Mycteroperca microlepis, females (F), males (M), and transitionals (T) taken in North Carolina, South Carolina, Georgia, northern Florida, and southern Florida during 18 January to 18 April 1995 when fishermen were required to land all gag with intact gonads (TL=total length). North Carolina South Carolina Georgia North Florida South Florida Size (mm TL) F M T F M T F M T F M T F M T 450-499 500-549 — — — 2 — — — — — — — — — — — 550-599 1 — — 10 — — — — — 1 — — — — — 600-649 3 — — 12 — — 1 — — — — — 3 — — 650-699 13 — — 39 — — 16 — — 4 — — 7 — — 700-749 37 — — 87 — — 35 — — 6 — — 6 — — 750-799 53 — — 199 — — 113 — — 46 — — 26 — — 800-849 64 — — 234 — 1 214 — — 111 — — 86 — — 850-899 64 — — 160 3 — 148 1 — 155 — — 53 1 1 900-949 23 — — 66 1 — 43 1 — 112 — — 27 1 — 950-999 19 1 — 11 3 — 16 1 — 44 18 — 9 3 — 1000-1049 7 3 — 4 5 — 10 4 — 26 33 — 1 3 — 1050-1099 3 1 — 2 4 — 7 3 — 14 27 — 1 1 — 1100-1149 — — — 3 1 — — 1 — 2 9 — 1 2 — 1150-1199 — — — — 3 — 1 1 — — 3 — — — — 1200-1249 — 1 — — 1 — — 2 — — 1 — — — — 1250-1299 No lengths 2 2 2 1 Total 287 6 — 831 21 1 606 14 — 523 91 — 221 11 1 Percent 97.9 2.1 — 97.4 2.5 0.1 97.7 2.3 — 85.1 14.9 — 94.8 4.7 0.5 508 mm in 1976-82 and 517 mm in 1994-95. All gag >683 mm in the 1976—82 samples and those >750 mm in the 1994-95 samples were sexually mature. Dur- ing 1994-95, 266 males ( x TL=1041 mm, SD=68) were collected and no gag less than 875 mm was male (Table 5). During 1976-82, 104 males ( * TL=1041 mm, SD=73) were collected with the smallest speci- men at 795 mm. The transitional specimens collected during 1994-95 were larger than those from 1976 to 1982. The length at 50% maturity (L50 ) was similar dur- ing 1978-82 (641 mm; 95% confidence interval (Cl) =616-658 mm) and 1994-95 (622 mm; 95% CI=611- 631 mm (Fig. 1); however, the probit analysis with period and total length as independent variables (Pearson yy, P= 0.9953, df=27) indicated that the size at maturity was likely decreasing (Table 6). The nega- tive coefficient for period indicated that overall a smaller proportion of specimens in each size class was mature during 1976-82. The overlap in the his- tograms of female gag that were developing, ripe, or spent and females that were resting indicated that reproductive tissue was correctly assigned to the immature and resting categories (Fig. 2). Female gag were in spawning condition from De- cember 1994 through mid May 1995 on the basis of Table 5 Number of mature gag, Mycteroperca mi during 1994-95 (TL=total length). Trans specimens. crolepis, taken = transitional Size (mm TL) Female Male Trans 450-499 0 500-549 2 — — 550-599 16 — — 600-649 35 — — 650-699 121 — — 700-749 267 — — 750-799 650 — 1 800-849 1021 — 2 850-899 790 7 5 900-949 386 14 12 950-999 132 46 10 1000-1049 70 82 8 1050-1099 33 68 — 1100-1149 7 28 — 1150-1199 2 15 — 1200-1249 — 5 1 1250-1299 1 — — No Lengths 41 1 — Total 3574 266 39 Percent 92.1 6.9 1.0 802 Fishery Bulletin 96(4), 1998 occurrence of hydrated oocytes or postovulatory follicles (Fig. 3). Spawning individuals were caught from North Carolina to Florida by commercial fisher- men and fishery-independent sampling (MARMAP). Ap- proximate fishing locations provided by commercial fish- ermen indicated that gag were spawning from the border be- tween North Carolina and South Carolina to a point east of Ft. Lauderdale, Florida. Spawning individuals were captured on research cruises off South Carolina and Geor- gia at depths ranging from 49 to 91 m. The spawning season appeared to be more pro- tracted in southern Florida than in other areas, with spawning beginning in De- cember and extending into mid May. Peak spawning oc- curred from March through mid-April (Fig. 4). When the data were stratified by area, the highest percentage of spawning females was found in northern Florida during March and April (Fig. 4). Overall, spawning of gag declined rapidly after 8 April. Individuals undergoing sexual transition were extremely rare during most of the spawning season; only one transitional specimen was documented dur- ing February and March 1994-95, out of 2147 fish examined (Table 7). There was a sharp increase in the number of transitional individuals immediately after the spawning season. The majority of transi- tional fish (n=26) were collected during mid-April through mid-June when 1473 gag gonads were ex- Table 6 Results of probit analysis comparing the proportions of mature females in 5-cm length intervals collected during two periods: 1976—82 and 1994-95 (TL=total length). Parameter Estimate SE P Intercept -12.304 0.871 <0.0001 Period -0.378 0.176 0.0320 cm TL 0.192 0.013 <0.0001 * = significant at P< 0.05; = significant at P<0.0001. amined. AN OVA and the Duncan multiple range tests indicated that there were significant differences in the mean lengths of gag sampled between months. When the number of males, females, and transi- tionals between 889 mm and 1050 mm was tabulated Table 7 Total number sampled (n ), mean total length ( mm TL; sexes combined), standard deviation (SD), number of males (M), number of females (F), and number of transitional speci- mens (T) by month for mature gag captured during 1994 and 1995. Means with same letter are not significantly different (P>0.05). Month n x TL SD M F T Jan 243 821.0bcd 85.5 10 229 4 Feb 754 829. 7*"' 87.0 27 726 1 Mar 1142 859. 06 99.2 87 1055 — Apr 1027 844.8fcc 86.4 50 965 12 May 297 873. 2°6 110.8 48 235 14 Jun 48 928.1° 110.7 11 32 5 Jul 77 925.9“ 124.1 19 56 2 Aug 2 717. ty 106.1 — 2 — Sep 33 774. 8°f 84.1 2 31 — Oct 24 759. Tf 83.6 — 24 — Nov 77 818.5* 86.8 — 76 — Dec 89 788. 5C* 89.7 2 86 1 McGovern et al.: Changes in the sex ratio and size at maturity of Mycteroperca microlepis 803 to reduce the potential for bias that might result from the selection of larger fishes after 18 April 1995, the Duncan multiple range test showed little difference in the mean size of gag collected each month (Table 8), and it was still apparent that transitional fish did not appear in the collections until immediately after the spawning season. Discussion Analysis of reproductive data in- dicates that the gag population along the southeast coast of the United States is stressed. The per- centage of males in the adult popu- lation has decreased from 19.6% (1976-82) to 5.5% (1994-95). How- ever, the percentage of males cur- rently in the population may ac- tually be less than 5.5%; more re- cent data were collected during the spawning season when male gag are most vulnerable to fishing gear (Coleman et al., 1996). In contrast, data were collected throughout the year during 1976-82 rather than primarily during the spawning season. Had the 1995 emergency rule been put into place at a time of the year when gag were not spawning, it is possible that fewer males would have been captured. The relative abundance of males and the length of the spawning period were greater off Florida than off other southeastern states. The highest percent- age of males ( 14.9%) during 1995 was noted off north- ern Florida. However, the majority of males were taken by one fisherman on 2 March 1995, 17 March 1995, and 1 April 1995 indicating that this fisher- man may have been targeting a relatively unfished spawning aggregation. Spawning aggregations of gag (Coleman et al., 1996) and red hind, Epinephelus guttatus (Shapiro et al., 1993a), contained a higher percentage of males than nonaggregation groups. Commercial fishermen indicate that after the spawn- ing season (May and June), female gag move in groups to shallower water ( - 30 m ) and the larger males become solitary and remain at depths of 50 to 90 m. Probit analysis revealed that female gag may have been maturing at smaller sizes during 1994-95 than during 1976-82 which would further indicate that Total length (cm) Figure 2 Comparison of the size frequency of immature gag, Mycteroperca microlepis , with the size frequencies of resting females and females with evidence of certain maturity. Table 8 Total number sampled ( n ), mean total length ( mm TL; sexes combined), standard deviation (SD), number of males (M), number of females (F), and number of transitional speci- mens (T) by month for mature gag (>889 mm and <1050 mm) captured during 1994 and 1995. Means with same letter are not significantly different (F>0.05). Month n if TL SD M F T Jan 45 932. 4h 41.5 5 38 2 Feb 126 935. 6h 42.5 12 114 — Mar 308 948. 3“6 46.4 51 257 — Apr 219 939. 6"6 43.8 27 183 9 May 92 949.8"fe 43.4 24 54 14 Jun 25 971.0afe 49.7 7 13 5 Jul 41 963. 1"6 46.7 10 30 1 Aug 0 — — — — — Sep 3 986.7" 28.9 2 1 — Oct 2 944.5aft 77.1 — 2 — Nov 10 961.6"6 52.7 1 9 — Dec 11 933.0'’ 47.4 2 8 1 804 Fishery Bulletin 96(4), 1998 North Carolina r=358 Northern Florida n=586 South Carolina n=l,338 J FMAMJ JASOND Southern Florida «=307 J FMAMJ JASOND 100 80 60 40 20 0 J FMAMJ JASOND Georgia n=861 Month = Developing = Spawning = POF present = Spent - Resting Month Figure 3 Reproductive seasonality of female gag, Mycteroperca microlepis, in North Carolina, South Carolina, Georgia, northern Florida, and southern Florida during 1994—95. POF = postovulatory follicles. the gag population is stressed. Although probit analy- sis suggested size at maturity is decreasing, it is dif- ficult to conclude that this is happening in the popu- lation on the basis of estimates of L50 because there is overlap in the confidence intervals. A cautious in- terpretation is also warranted given the small num- ber of females (n= 422) sampled during 1976-82. Coleman et al. (1996) found that female gag collected in the Gulf of Mexico during 1991-93 became sexu- ally mature and underwent transition at smaller sizes than those reported by Hood and Schlieder (1992) for the same region during 1977-80. Changes in life history aspects of gag from the Gulf of Mexico were attributed to steadily increasing fishing pressure. Gag were in spawning condition from December through mid May in southern Florida, January through May in northern Florida, and January through April in South Carolina. Although no speci- mens were collected during January in North Caro- lina and Georgia, individuals in spawning condition were landed during February through April. Peak spawning activity occurred from March through mid- April in all areas. Collins et al. (1987) and Keener et al. (1988) also reported that peak spawning occurred McGovern et al.: Changes in the sex ratio and size at maturity of Mycteroperca microlepis 805 in March and April along the southeast Atlantic coast. Hood and Schlieder ( 1992 ) and Coleman et al. ( 1996 ) indicated that peak spawning occurred during Feb- ruary and March in the Gulf of Mexico. Despite the large number of specimens examined during 1994-95 (n=3879), transitional specimens (n= 39) made up a small percentage (1.0%) of indi- viduals collected. The low number of individuals with transitional gonads may have been due to the rapid nature of sex transition that appears to occur imme- diately after the spawning season. Smith (1965) and Moe ( 1969) suggested that other groupers may have a quick rate of sex transition. Because there was no significant monthly difference in the mean length of gag between 889 and 1050 mm TL landed during January through July, we feel that the peak in the number of transitional specimens during April and May was real and not an artifact of the emergency rule that required fishermen to land only larger fish (>889 mm TL) after 18 April. Coleman et al. (1996) stated that sex transition in gag and other grouper species may be socially mediated either through size ratio or sex ratio cues. Shapiro et al. (1993b) pro- posed that one function of annual spawning aggre- gations in red hind may be that of enabling a female to determine if it should function as a female or a male during the next spawning season. Shapiro et al. ( 1993b) further suggested that if this hypothesis is true, sex- changing individuals should be found soon after the aggregation has dispersed. The high percentage of males and transitionals taken during April and May in this study indicates that aggregations may remain intact for a short time after the spawning season. Van Sant et al. ( 1994) and MARMAP tagging data1 provided evidence that some gag migrate from off- 1 1998. Unpublished data from MARMAP tagging study. Marine Resources Research Institute, Charleston, SC. 806 Fishery Bulletin 96(4), 1998 shore areas of South Carolina and Georgia to the shelf edge off Florida, perhaps to spawn. However, the presence of hydrated oocytes and new postovulatory follicles in female gag taken from North Carolina through Florida, including specific locations off South Carolina and Georgia, shows that spawning is not restricted to Florida. Hunter et al. (1985) reported that the presence of hydrated oocytes in teleosts is an indication that individuals will spawn within 12 hours. Furthermore, Hunter et al. (1986) found that postovulatory follicles are present only in fishes that have recently spawned. Gag off Florida are subject to intense fishing pres- sure by commercial fishermen, sport divers, and an- glers, particularly during the late fall through early spring, because the continental shelf is very narrow and spawning aggregations are easily accessed. Therefore, even if most gag are not migrating to Florida for spawning, the vulnerability of at least part of the stock to winter fishing is of concern. This “funneling” of migratory fishes off southeastern Florida may increase fishing mortality for many spe- cies and requires further investigation. In 1993, the SAFMC decided to take no action on a proposal to close all fishing for gag during their spawning season because insufficient data were available on sex ratios and the spatial extent of spawning along the southeast coast of the United States. The information on sex ratio and size at ma- turity presented here indicates that the gag popula- tion is overfished. Because there has been a reduc- tion in the percentage of males in commercial land- ings and a modest decrease in the size at maturity for females, the total fishing effort should be lim- ited, especially in the removal of males from spawn- ing aggregations. Since gag spawn in offshore areas from North Carolina to southeast Florida, any regu- lation should apply to all southeastern states. How- ever, seasonal closure to fishing may not be suffi- cient to protect gag from overexploitation because males will continue to be lost during the open sea- son. Protected areas with no fishing, such as marine fishery reserves, may be a possible solution to over- fishing of reef fishes along the southeast coast of the United States (PDT, 1990). Marine fishery reserves can protect population age structure, species diver- sity, genetic diversity, and recruitment supply to ex- ploited areas (Bohnsack, 1993). Sedberry et al. (1996) found that marine reserves in Belize, Central America, had a higher diversity of fishes than in ar- eas that were not protected. Top predators, such as various grouper and snapper species, were more abundant and larger in reserves. In addition, popu- lations of herbivorous forage species were reduced to presumed natural levels in the presence of pro- tected predators. Marine reserves in Belize appear to have a natural balance of predators and forage species in relation to fished areas. There are other species in the snapper-grouper complex that are showing signs of overfishing, in addition to gag, in- cluding red porgy (Harris and McGovern, 1997), ver- milion snapper (Zhao and McGovern, 1997; Zhao et al., 1997), and black sea bass.2 In addition, less im- portant species in fisheries (e.g. white grunt, gray triggerfish) are increasing in abundance.2 This trend is likely to continue unless different management regulations are imposed that will protect ecosystems and restore a natural equilibrium community. Acknowledgments We thank personnel associated with cooperative state and federal agencies, for their assistance with the collection of data, and commercial fishermen for pro- viding trip information. In particular, we recognize T. Brandt, L. Bishop, M. Burton, D. Codella, C. Den- nis, R. Roman, and D. Thiesen of the NMFS, G. Rogers and J. Ross of the Georgia Department of Natural Resources, and F. Rohde of the North Caro- lina Department of Environment, Health, and Natu- ral Resources for their efforts. We thank Dr. John Grego (University of South Carolina) for helping us with the probit procedure. The assistance at sea by SCDNR MARMAP personnel and the crew of the RV Palmetto is appreciated. We also thank K. Grimball, T. Kellison, and other MARMAP personnel for pro- cessing gonads that were sent from state and fed- eral port samplers. This study was sponsored by the National Marine Fisheries Service (MARMAP Contract No. 52WCNF6006013PW and Saltonstall-Kennedy Grant NA57FD0030), and the South Carolina Depart- ment of Natural Resources. Literature cited Bohnsack, J. A. 1993. Marine reserves. Oceanus 36:63-71. Carter, J., G. J. Marrow, and V. Pryor. 1994. Aspects of the ecology and reproduction of Nassau grouper, Epinephelus striatus, off the coast of Belize, Cen- tral America. Proc. Gulf Caribb. Fish. Inst. 43:65-111. Coleman, F. C., C. C. Koenig, L. A. Collins. 1996. Reproductive styles of shallow-water groupers (Pi- sces: Serranidae) in the eastern Gulf of Mexico and the consequences of fishing spawning aggregations. Environ. Biol. Fish. 47:129-141. 2 1998. MARMAP unpublished data. Marine Resources Re- search Institute, South Carolina Department of Natural Re- sources, Charleston, SC 29422-2559. McGovern et al. : Changes in the sex ratio and size at maturity of Mycteroperca microlepis 807 Collins, M. R., and G. R. Sedberry. 1991. Status of vermilion snapper and red porgy stocks off South Carolina. Trans. Am. Fish. Soc. 120:116-120. Collins, M. R., C. A. Waltz, W. A. Roumillat, and D. L. Stubbs. 1987. Contribution to the life history and reproductive bi- ology of gag, Mycteroperca microlepis (Serranidae), in the South Atlantic Bight. Fish. Bull. 85:648-653 Cuellar, N„ G. R. Sedberry, and D. M. Wyanski. 1996. Reproductive seasonality, maturation, fecundity, and spawning frequency of the vermilion snapper ,Rhomboplites aurorubens, off the southeastern United States. Fish. Bull. 94:635-653. Ferreira, B. P. 1993. Reproduction of the inshore coral trout Plectropomus maculatus (Perciformes: Serranidae) from the Central Great Barrier Reef, Australia. J. Fish. Biol. 42:831-834. Gilmore, R. G., and R. S. Jones. 1992. Color variation and associated behavior in the epinepheline groupers, Mycteroperca microlepis , (Goode and Bean) and M. phenax Jordan and Swain. Bull. Mar. Sci. 51:83-103. Hardy, J. E., Jr. 1978. Development of fishes of the Mid-Atlantic Bight — an atlas of egg, larval, and juvenile stages, vol. Ill: Aphredoderidae through Rachycentridae. U.S. Fish and Widl. Serv., U.S. Dep. Interior, 394 p. Harris, P. J. and J. C. McGovern. 1997. Changes in the life history of red porgy, Pagrus pagrus, from the southeastern United States, 1972- 1994. Fish. Bull. 95:732-747. Hood, P. B., and R. A. Schlieder. 1992. Age, growth, and reproduction of gag Mycteroperca microlepis, (Pisces: Serranidae), in the eastern Gulf of Mexico. Bull. Mar. Sci. 51:337-352. Hunter, J. R., N. C. H. Lo, and R. J. H. Leong. 1985. Batch fecundity in multiple spawning fishes. In R. Lasker ( ed. ), An egg production method for estimating spawning biomass of pelagic fish: application to the north- ern anchovy ( Engraulis rnordax ), p. 67-77. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 36. Hunter, J. R., and B. J. Macewicz. 1985. Rates of atresia in the ovary of captive and wild north- ern anchovy, Engraulis rnordax. Fish. Bull. 83:119-136. Hunter, J. R., B. J. Macewicz, and J. R. Sibert. 1986. The spawning frequency of skipjack tuna, Katsuwonus pelamis, from the South Pacific. Fish. Bull. 84:895-903. Huntsman, G. R. 1976. Offshore headboat fishing in North Carolina and South Carolina. Mar. Fish. Rev. 38(3): 13-23. Keener, P., G. D. Johnson, B. W. Stender, E. B. Brothers and H. R. Beatty. 1988. Ingress of postlarval gag, Mycteroperca microlepis (Pisces: Serranidae), through a South Carolina barrier is- land inlet. Bull. Mar. Sci. 42:376-396. Moe, M. A., Jr. 1969. Biology of the red grouper, Epinephelus morio (Valenciennes), from the eastern Gulf of Mexico. Fla. Dep. Nat. Resour. Mar. Res. Lab., Prof. Pap. 10, 95 p. Nelson, K., and M. Soule. 1987. Genetic conservation of exploited fishes. In N. Ryman and F. Utter (eds.), Population genetics and fish- ery management, p. 345-368. Univ. Washington Press, Seattle, WA. PDT (Snapper-Grouper Plan Development Team). 1990. The potential of marine fishery reserves for reef fish management in the U.S. South Atlantic. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFC-216, 45 p. SAFMC (South Atlantic Fishery Management Council). 1993. Amendment number 6, regulatory impact review and final environmental impact statement for the snapper grou- per fishery of the South Atlantic region . SAFMC, Charles- ton, SC, 189 p. SAS Institute, Inc. 1990. SAS/STAT® user’s guide, version 6, 4th edition, vol. 2. SAS Institute, Inc, Cary, North Carolina, p. 1325-1350. Sedberry, G. R., J. Carter and P. A. Barrick. 1996. A comparison of fish communities between protected and non-protected areas of the Belize Barrier Reef ecosys- tem: Implications for conservation and management. Proc. Gulf Caribb. Fish. Inst. 45:95-127. Shapiro, D. Y., Y. Sadovy, and M. A. McGehee. 1993a. Periodicity of sex change and reproduction in the red hind, Epinephelus guttatus, a protogynous grouper. Bull. Mar. Sci. 53:1151-1162. 1993b. Size, composition, and spatial structure of the an- nual spawning aggregation of the red hind, Epinephelus guttatus (Pisces: Serranidae). Copeia 1993:399-406. Smith, C. L. 1965. The patterns of sexuality and the classification of serranid fishes. Am. Mus. Novit. 2207:1-20. 1971. A revision of American groupers: Epinephelus and allied genera. Bull. Am. Mus. Nat. Hist. 146:67-242. Smith, P. J., R. I. C. C. Francis, and M. McVeagh. 1991. Loss of genetic variation due to fishing pressure. Fish. Res. 10:309-316. West, G. 1990. Methods of assessing ovarian development in fishes: a review. Aust. J. Mar. Freshwater Res. 41:199-222. Van Sant, S. B., M. R. Collins, and G. R. Sedberry. 1994. Preliminary evidence from a tagging study for a gag ( Mycteroperca microlepis) spawning migration, with notes on the use of oxytetracycline for chemical tagging. Proc. Gulf Caribb. Fish. Inst. 43:417-428. Zhao, B., and J. C. McGovern. 1997. Temporal variation in sexual maturity and gear-spe- cific sex ratio of the vermilion snapper in the South Atlan- tic Bight. Fish. Bull. 95:837-848. Zhao, B., J. C. McGovern, and P. J. Harris. 1997. Age, growth, and temporal change in size at age of the vermilion snapper from the South Atlantic Bight. Trans. Am. Fish. Soc. 126:181-193. 808 Abstract .—Boat surveys along ran- domly placed line transects were con- ducted from June to August 1991 and June to October 1992 to determine dis- tribution and abundance of and habi- tat use by harbor porpoise ( Phocoena phocoena) off the northern San Juan Is- lands, Washington. There were 301 sightings (average 4.4 sightings/h) of 526 harbor porpoise during 73 random boat surveys, with group sizes of 1 to 8 individuals (mean = 1.87, SE=0.06, =278). An estimated 299 harbor por- poise (1.26 porpoise/km2, SE=0.20) were distributed in an aggregated pat- tern within a 237 km2 area (10% of Washington Sound), indicating that a large proportion (30%) of harbor por- poise in Washington Sound occur in the northern San Juan Islands. Harbor porpoise were distributed over a depth range from 20.1 to 235.0 m (mean= 141.6 m, SE=2.43, n= 275) and were observed more than expected (P<0.05) in depths greater than 125 m and over shallow slopes ( < 10%) and observed less than expected (P<0.05) in depths less than 75 m. Porpoise occurred at sea surface temperatures of 10. 1° to 16.3°C and were sighted more frequently than expected (P<0.05) in water tempera- tures of 11° to 12°C. Boat surveys along fixed location transects indicated dis- tribution was similar between 1991 and 1992. The occurrence of harbor porpoise in deep water, at cooler sea surface tem- peratures, over shallow sloping seaf- loor, and in tidally mixed regions (ow- ing to currents and tide rips) within our study area may, collectively, affect prey distribution and associated harbor por- poise distribution. Manuscript accepted 12 March 1998. Fish. Bull. 96:808-822 (1998). Distribution and abundance of and habitat use by harbor porpoise, Phocoena phocoena, off the northern San Juan Islands, Washington Kimberly L. Raum-Suryan James T. Harvey Moss Landing Marine Laboratories RO. Box 450 Moss Landing, California 95039 Present address (for K. L. Raum-Suryan): Alaska Department of Fish and Game Division of Wildlife Conservation 333 Raspberry Road Anchorage, Alaska 995 ! 8 E-mail address (for K. L. Raum-Suryan): kimr@fishgame. state. ak. us Harbor porpoise (Phocoena phocoena) are present year-round off the coast and inland waters of Washington State. Historically, harbor porpoise have been present throughout the Strait of Juan de Fuca, Washington Sound (San Juan Island archi- pelago), and south in Puget Sound. Once considered the most common cetacean in southern Puget Sound (Scheffer and Slipp, 1948), harbor porpoise sightings are now rare (Everitt et al., 1980; Calambokidis et al., 1984, 1985). Although harbor porpoise have not been sighted off the central San Juan Islands in re- cent years (Flaherty and Stark1 ; Calambokidis et al.2 ), sightings off the northern San Juan Islands have been common (Flaherty and Stark1; Calambokidis et al.2). Reasons for the disappearance of harbor porpoise from South Puget Sound are unclear but may be due to reduced availabil- ity of prey (because of changes in en- vironmental conditions), fishing pres- sure, disturbance, net entanglement, or pollution. Many biological (e.g. prey) and physical oceanographic factors (e.g. depth, seafloor relief, tidal currents, and sea surface temperature) affect the distribution of cetaceans. In- creased availability of prey in deep waters may be a factor affecting the distribution of harbor porpoise. Smith and Gaskin (1983) found a significant positive correlation be- tween abundance of mother-and- calf pairs and bottom depth and copepod ( Calanus spp.) density. Abundance of harbor porpoise also was positively correlated with depth and physiographic features that con- centrated Atlantic herring ( Clupea harengus) in near-surface waters (Watts and Gaskin, 1985). In Fish Harbor, New Brunswick, Canada, harbor porpoise were associated with reduced sea surface tempera- tures that coincided with a large in- flux of juvenile herring in the region (Gaskin and Watson, 1985). Tidal state affected movements of harbor porpoise in the Bay of Fundy; har- 1 Flaherty, C., and S. Stark. 1982. Harbor porpoise (Phocoena phocoena ) assessment in “Washington Sound.” Final report for subcontract 80-ABA-3584. National Ma- rine Mammal Laboratory, Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA, 84 p. 2 Calambokidis, J., J. C. Cubbage, J. R. Evenson, S. J. Jeffries, and R. F. Brown. 1993. Abundance estimates of harbor porpoise in Washington and Oregon waters. Final Report to National Marine Mammal Laboratory, Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA, 55 p. Raum-Suryan and Harvey: Distribution and abundance of and habitat use by Phocoena phocoena 809 Random boat transects (n= 73) completed from June to October 1992 within five stratified sections (A, B, C, D, E), northern San Juan Islands, Washington. bor porpoise were observed more often during flood fide than ebb tide (Watson, 1976) and moved inshore during flood tides and offshore during ebb tides (Gaskin and Watson, 1985). Harbor porpoise in northern Puget Sound are vul- nerable to some of the same detrimental effects (dis- turbance, net entanglement, and pollution) that may have caused the disappearance of harbor porpoise in southern Puget Sound. It is important, therefore, to determine the abundance of harbor porpoise and identify habitat variables that may influence their distribution in northern Puget Sound. The main ob- jectives of this study were to determine 1) the spa- tial and temporal distribution, density, and abun- dance of harbor porpoise occurring off the northern San Juan Islands and 2) the relation of harbor porpoise to depth and percentage slope of the seafloor, sea surface temperature (SST), tidal state, and time of day. Methods Study area Washington Sound is located in the northwest cor- ner of Washington State (48°15’ to 48°5G’N and 122°27! to 123°13’W), between the southern portion of Vancouver Island and the mainland, from Fidalgo Island to north of Vancouver, including the Ameri- can and Canadian islands of the San Juan Archi- pelago (Kozloff, 1973). Mean diurnal tide heights are between 1.3 and 2.9 m (NOAA, 1991). Northern Washington Sound (northern San Juan Islands) has numerous islands and reefs with deep channels, strong currents, and tide rips. The study area off the northern San Juan Islands (Fig. 1) was selected on the basis of preliminary boat surveys conducted in 1991 to determine areas of harbor porpoise occur- rence. Additionally, information was obtained from local residents and The Whale Museum, Friday Har- bor, Washington. Random boat surveys Randomly located boat transects (n- 73; Fig. 1) were conducted from 27 June to 2 October 1992 within a study area composed of five strata (lettered A-E; Fig. 1) to determine harbor porpoise distribution, den- sity, abundance, and habitat use. Eight-km transects were located within each approximately equal (42 to 50 km2) stratum by using random starting points and random compass bearings. Strata were originally chosen so that transects would adequately cover the entire study area. Because placement of straight 8-km transects was constrained by the boundaries of strata and islands or reefs within strata, however, some regions of each strata were not adequately sampled. When sea conditions permitted, the five 810 Fishery Bulletin 96(4), 1998 strata, hereafter referred to as sections, were sur- veyed on the same day. Sections were chosen in a random starting order and every attempt was made to complete all sections before surveying the same sections again. Harbor porpoise were surveyed from a 7.3-m alu- minum marine patrol vessel during Beaufort sea state 0 (wind speed=0-1.8 km/h), 1 (wind speed=1.8- 5.6 km/h), or rarely Beaufort 2 (wind speed=7.4-ll km/h). Each transect was completed in approximately 52 min, at an average boat speed of 9 km/h. Date, time, and tidal phase (flood or ebb) were recorded before each transect was surveyed. At the beginning and end of each transect, Secchi disk readings were recorded to the nearest 0.1 m and SST was recorded (by a calibrated thermometer located on the trans- ducer of the survey vessel) to the nearest 0. 1°C. Dur- ing surveys, two observers divided the field of view across the forward 180° of the transect path. Obser- vations were made from the roof of the vessel (height above the waterline=2.68 m, measured to the observ- ers’ eyes in surveying position) with unaided eyes and with Fujinon 7 x 50 reticle and compass bin- oculars. When an individual or group of harbor por- poise was located, an observer recorded time; group size and composition; compass bearing to the por- poise; ocular reticle marks from the horizon to the porpoise; Beaufort sea state; number of boats, birds, and marine mammals within 1 km of vessel; and di- rection of harbor porpoise travel. A group of harbor porpoise was defined as two or more porpoise visible at the water’s surface within three body lengths (5 m) of each other, having nearly synchronous div- ing patterns ( <15 seconds between sightings of each individual). Observers were trained and tested in the use of reticle binocompasses and in calibrating their readings on buoys and points of land and comparing distance accuracy to National Oceanic and Atmo- spheric Administration (NOAA) navigational charts and vessel’s radar. Compasses on binoculars were also tested and were found not to be significantly affected by metal on the survey vessel. Although not common, the horizon was sometimes obscured by land in the observers’ viewing area. To compensate for this, we estimated the number of reticles from the land-water interface (directly beyond the por- poise) to the porpoise sighting, then carefully rotated the binoculars from the land to the horizon. We then determined the number of reticles the horizon was beyond the land and added this amount to the reticle reading of the harbor porpoise sighting. Loran coor- dinates, Beaufort sea state, visibility, SST, and num- ber of boats within 1 km of the vessel were recorded every ten minutes during surveys and for each har- bor porpoise sighting. Depth and seafloor slope (at each harbor porpoise sighting) were determined from NOAA navigational charts and bathymetric charts. Locations of harbor porpoise were determined with the aid of Fujinon 7 x 50 reticle (one reticle=17 min or 0.283°) and compass binoculars. Vertical angle was calculated as the angle between the horizon and the harbor porpoise. Distance to harbor porpoise was calculated as ' tan(G! ) ’ where Dr = the radial distance from the vessel to the porpoise; H = the eyeheight of observers; and a = the vertical angle between the horizon and porpoise. Locations were plotted on NOAA navigational charts by using Loran (latitude and longitude) coordinates of the vessel at the time harbor porpoise were sighted, and distance and bearing to the sighting. Perpendicular distance from the trackline to har- bor porpoise was determined by using Dp = Dr x sin(a), where Dp - perpendicular distance; Dr = the radial distance to the harbor por- poise; and a = the angle off the trackline (the difference between the trackline heading and the bearing to porpoise). Seafloor depth and slope were determined by us- ing a NOAA navigational chart and bathymetric map. Percentage slope was calculated as dz % slope = — x 100%, ds where dz = the difference between the two closest depths (m) printed on the chart on ei- ther side of a harbor porpoise location (with contour lines drawn among depths ); and ds - the distance (m) between those two depths. Bathymetric charts with contour intervals of 10 m were used to verify angle of slope between depths. To determine if harbor porpoise occurred over depths and slopes in proportion to available depths Raum-Suryan and Harvey: Distribution and abundance of and habitat use by Phocoena phocoena and slopes in the study area, eight random points were plotted within a 2-km strip along the length of each 8-km transect {n= 73 transects, 584 points) and depth and slope were determined for each point. The number of random points chosen was determined by plotting precision (standard deviation/mean) against sample size until there was little variability in this measure (i.e. a plateau and subsequent leveling of the curve). Density and abundance estimates were calculated by using the line transect method as described by Burnham et al. (1980) and the computer program DISTANCE (Laake et al., 1993). Each transect was considered a replicate. Density and variance esti- mates of harbor porpoise sightings (rc=250) were cal- culated by replicate for each section (n = 12 to 15 transects) and by replicate for all sections combined (n= 70 transects). Transects with Beaufort sea state of 2 (n= 3) were deleted from analyses because sight- ing rates of harbor porpoise in Beaufort 2 are less than Beaufort 0 or 1 (Barlow, 1988). Density was calculated as D = n x f( 0) x s 2 L where n /TO) s L number of individual harbor porpoise sightings; the probability density function of dis- tances from the trackline evaluated at zero distance; average group size of harbor porpoise sightings, and total length of the trackline. Abundance was calculated as density multiplied by area of each section (A-E) and all sections (237 km2). The parameter /TO) is essentially a measure of sight- ing efficiency and should not vary with porpoise abun- dance as long as sighting conditions (e.g. Beaufort sea state, visibility) remain the same. Because we surveyed only during optimal sighting conditions (Beaufort <1, no rain or fog) within all sections and because relatively large sample sizes are required to estimate /TO) accurately, values of /TO) for each sec- tion were estimated by pooling all sightings in all sections. Effective strip width is defined as 1//T0), which equals one-half the transect width, such that as many objects are detected outside the strip as re- main undetected within it (Buckland et al., 1993). Because group size was independent of distance from the trackline (determined through size-bias regres- sion analysis with DISTANCE software), average group size was used to calculate density. Average group size was estimated by section and for all sightings combined. Uniform, half-normal (hermite), hazard rate, and negative exponential models were compared with the frequency distributions of perpendicular sighting distance of harbor porpoise to trackline with DIS- TANCE. Several groupings and truncation points were investigated to achieve the best model fit. Buckland et al. (1993) recommend truncating 5 to 10% of objects detected at the greatest distances from the trackline. The half-normal (hermite) model, grouped into 50 m intervals and truncated at 750 m (deleting 5% of sightings), was chosen on the basis of lowest Akaike Information Criterion (AIC; Buckland et al., 1993) score for all sections combined. The probability of detection at zero perpendicular distance, g( 0), was assumed to be one (all harbor porpoise on the trackline were assumed to be seen) because we were unable to estimate perception bias (bias resulting from animals available to be seen but that were not; Marsh and Sinclair, 1989). We did not have an independent observer to watch the trackline for porpoise that were missed by our two observers, therefore, a correction was not applied tog(0). It is likely g(0) was less than one but it is probably high (slow boat speed and excellent sighting conditions). Because g(0) was constant over the survey time pe- riod, the habitat correlations are valid; however, the abundance estimate is underestimated by an un- known amount. Seafloor depth and slope available in the study area in relation to areas of harbor porpoise occurrence were compared by using chi-square goodness-of-fit analyses. To test whether the frequency of occurrence of harbor porpoise was independent of frequency of tidal currents and surface temperature, we also used chi-square goodness-of-fit analyses. More surveys were conducted during flood tide (n=52) than during ebb tide (n=17); therefore, the number of harbor por- poise observed per minute during flood or ebb tide was used to standardize the data. A Mann-Whitney U, nonparametric two-sample test was conducted to examine differences in number of harbor porpoise observed per minute for each transect (n- 73) during flood and ebb tides. Power analyses (Cohen, 1988) were conducted on nonsignificant categories of chi-square goodness-of- fit analyses. Randomization statistics with the pro- gram Resampling Stats (Resampling Stats, 1995) were performed to assess the probability of detect- ing a difference between flood and ebb tides when the difference was determined to be nonsignificant. Additionally, power analyses were used to estimate the probability of detecting trends in abundance over time (Gerrodette, 1987). 812 Fishery Bulletin 96(4), 1998 Not all times of day were sampled equally; there- fore, abundance of harbor porpoise in relation to time of day was compared by using number of porpoise observed per minute to standardize the data. Mean number of harbor porpoise observed per minute for each hour of daylight was compared with Kruskal- Wallis nonparametric analysis of variance and Kolmogorov-Smirnov goodness-of-fit analyses (Zar, 1984). Nonparametric statistics were used for data with non-normal distributions or unequal variances. Fixed boat surveys To determine temporal changes in harbor porpoise distribution between 1991 and 1992, six 8-km transect lines (hereafter called fixed transects; Fig. 2), placed in areas of harbor porpoise occurrence (pre- liminary harbor porpoise surveys and information from Orca Hotline, The Whale Museum, Friday Har- bor, Washington), were surveyed regularly from 27 July to 26 August 1991 and from 24 July to 28 Au- gust 1992. In 1991, there was only one observer per survey; therefore, only one half the transect (bow out to 90° port or starboard) was completed during each survey. To be consistent in 1992, one observer sur- veyed from bow to 90° port while the other surveyed from bow to 90° starboard so that one half of each transect could randomly be compared to 1991 transects. Harbor porpoise were counted from the same 7.3-m vessel as in random surveys during a Beaufort sea state of 0 or 1. Each fixed transect survey was con- ducted at an average speed of 11 km/h and completed in 40 to 45 minutes. This vessel speed was chosen in 1991, and to be consistent, 1992 fixed transect sur- veys were conducted at the same speed (instead of 9 km/h as in random boat surveys). Harbor porpoise locations were calculated as in random boat survey methods. Mean number of sightings of harbor porpoise per survey between 1991 and 1992 was compared by us- ing a /-test. Because both sides of the vessel were observed during a single survey in 1992, each side could not be considered an independent sample. Therefore, one side of the vessel was randomly cho- sen from each survey in 1992 to compare with 1991. Power tests (Cohen, 1988) were performed when re- sults were not significant. Results Random boat surveys There were 301 sightings of 526 harbor porpoise (Fig. 3) during random boat surveys. Of these, 20 sightings (39 porpoise) were possible resightings (i.e. observer believed the porpoise had already been seen during that survey, given the location and direction of travel of por- poise), therefore, these pos- sible resightings were not used in analyses. An average of 4.4 harbor porpoise sightings were recorded per hour (8.1 harbor porpoise per hour), with group sizes of 1 to 8 (mean=1.87, SE=0.06, n= 278) individuals. Thirteen cow and calf pairs were observed between June and September. Harbor por- poise were sighted during 75% of surveys at a mean perpen- dicular distance of 237 m (SE=13.89, n=250, range: 0 to 1060 m) from the trackline. The half-normal (hermite) model, truncated at 750 m, best fitted the frequency dis- tribution of perpendicular dis- tance of harbor porpoise sighted from the trackline (Fig. 4). Using harbor porpoise sightings (n=250) for all sec- Figure 2 Harbor porpoise locations along fixed boat transects ( 1, 2, 3, 5, 6, 7) in 1991 and 1992 off the northern San Juan Islands, Washington. The “x” denotes locations of harbor porpoise sighted in 1991; the denotes locations in 1992. Raum-Suryan and Harvey: Distribution and abundance of and habitat use by Phocoena phocoena 813 Figure 3 Location of 301 harbor porpoise sightings made during 73 random boat surveys completed from June to October 1992 within five stratified sections (A, B, C, D, E). tions combined, we estimated that the effective half-strip width (ESW) was 337 m (95% CI=307-371 m; coefficient of variation, CV=0.048), with an f( 0) of 2.96/km (SE = 0.14, CI=2.70-3.25, CV=0.048). No significant correlation (r=0.097, n= 250, P=0.938) was detected between harbor por- poise group size and perpen- dicular distance from the track- line. For surveys conducted during Beaufort <1 (re =70) in all sections (A-E), the mean group size was 1.91 harbor porpoise (SE=0.07, n= 250; Table 1) and mean density was 1.26 harbor porpoise/km2 (SE=0.20; Table 1). Harbor porpoise densities were least in section A (0.60 porpoise/km2, SE = 0.21) and greatest in section D (2.3 por- poise/km2, SE=0.74; Table 1). There were an estimated 299 harbor porpoise (CI=2 19-409) in all sections (Table 1), ranging from 30 harbor porpoise (CI= 5-60) in sec- tion A to 116 harbor porpoise (CI=62- 221) in section D (Table 1). The pooled estimate of harbor porpoise abundance (299 porpoise) for all sections yielded the same abundance estimate as add- ing the estimates for each individual strata (A-E; Table 1). If the present surveys were con- ducted annually with a similar sam- pling regime that produced an equally low CV (0.159), there would be suffi- cient power (80%) to detect a 14% an- nual change (a=0.05; 12% change for a=0.10) after five years. Mean number of harbor porpoise per survey was greatest in section D and least in section E (Fig. 5A). Mean depth throughout the study area was 108.1 m (SE=21.68, n=584); section D had the greatest mean depth and sec- tion E the least (Fig. 5B). Harbor por- poise were distributed over a depth range of 20.1 to 235.0 m (mean=141.6 m, SE=2.43, n- 275), with 83% of harbor porpoise sightings occurring over depths greater than 100 m. Significantly (P<0.05) fewer than expected harbor porpoise occurred in depths less than 75 m and sig- nificantly (P<0.05) more than expected in depth cat- egories greater than 100 m (Fig. 6). The effect size (degree to which depths differed among categories) was small and the power to detect a difference was 814 Fishery Bulletin 96(4), 1998 Table 1 Survey effort, line transect model parameters, density, and abundance estimates of harbor porpoise for each section ( A-E) and all sections combined surveyed within the northern San Juan Islands, Washington, from June to October 1992. Parameter Section A Section B Section C Section D Section E All sections Area (km2) Effort (km) Transect lines Truncation width (m) Probability density /)0)/km Sightings of harbor porpoise Mean group size Standard error (SE) of group size Density (porpoise/km2) 95% confidence intervals (porpoise/km2) SE of density % coefficient of variation (CV) of density Abundance 95% confidence interval (abundance) SE of abundance 50.26 43.12 50.40 112 120 120 14 15 15 750 750 750 2.96 2.96 2.96 26 62 67 2.19 1.94 1.81 0.32 0.15 0.11 0.60 1.03 1.55 0.30-1.21 0.60-1.75 0.77-3.10 0.21 0.27 0.52 35.09 26.00 33.86 30 44 77 15-60 26-75 39-155 10.53 11.43 26.07 50.47 42.43 237 112 96 560 14 12 70 750 750 750 2.96 2.96 2.96 80 15 250 1.90 1.80 1.91 0.11 0.26 0.07 2.33 0.76 1.26 .24-4.4 0.35-1.66 0.92-1.73 0.74 0.30 0.20 31.55 39.16 15.86 116 32 299 J— 221 15-70 217-406 36.91 12.53 47.0 low (37%) for categories that were not significantly different (Fig. 6). Given the small effect size, a power of 80% would require 358 locations of harbor por- poise in these three depth categories (there were 122 in this study). It is, therefore, unlikely that harbor porpoise occur in depths within these nonsignificant categories in different proportions than those available. Mean seafloor slope for all sections combined was 9.85% (SE=0.656, rc=584). Section B had the least slope and section C the greatest (Fig. 7). Harbor por- poise were sighted over a mean slope of 6.90% (SE=0.51, n=275, range: 0.37% to 45.75%). The great- est number of harbor porpoise sightings (79%) oc- curred over shallow slopes ( < 10%). There were sig- nificantly (P<0.05) greater numbers of harbor por- poise sightings than expected in category 0 to 2% slope, and significantly (P<0.05) fewer numbers of harbor porpoise than expected in categories 6 to 8%, 18 to 20%, and >26% slope (Fig. 8). The power to detect a difference was fairly high (69%) for catego- ries that were not significantly different (Fig. 8). To increase power to 80%, we would need 151 samples within these ten categories (we had 126 samples). Mean sea surface temperature (SST) recorded dur- ing all transects was 12.6°C (SE=0.081, n=427, range: 10.1° to 17.5°C). Little variability was found among the five sections. Section E had the greatest mean SST (mean=13.5°C, SE=0.22, n=69) and section B the least (mean=12.3°C, SE=0.14, n- 97). Mean SST recorded during harbor porpoise sightings was 12. 1 °C (SE=0.09, n- 267, range: 10.1° to 16.3°C). Harbor porpoise were sighted more frequently than expected (P< 0.05) in water temperatures of 11° to 12°C and less frequently than expected (P<0.05) in water tem- peratures >16° C (Fig. 9). The power to detect a dif- ference was moderate (50%) for categories that were not significantly different (Fig. 9). To increase power to 80%, we would need 180 samples within these ten categories (we had 171 samples). There was no significant difference between num- ber of harbor porpoise observed per minute during flood and ebb tides ( U= 315.5, n=69, P-0.076, a=0.05). Bootstrap estimates (resampling statistics; 10,000 iterations) indicated an 86% chance of correctly re- jecting the null hypothesis that mean number of sightings was equal between flood and ebb tides. Fifty samples in each tide stage (we sampled 52 in flood and 17 in ebb tide) were required to reject the null hypothesis at a = 0.05. Mean Secchi reading for all harbor porpoise sightings was 9.3 m (SE=0.09, n- 275, range: 5.7 to 11.9 m). Greatest numbers of harbor porpoise were ob- served in mid-morning ( 1000 h) and afternoon ( 1400 to 1500 h) throughout the study area (Fig. 10). Fewer harbor porpoise were observed at midday (1100 to 1300 h; Fig. 10), although there was no significant difference (77=10.99, n=274, P=0.276) among mean number of harbor porpoise observed per minute and each hour of daylight surveyed (0900 to 1800 h). Density estimates were calculated over all four months of the survey period rather than by month, which would have yielded too low of a sample size. If abundance estimates of porpoise had varied greatly among months during our survey period, we should have observed a higher CV (ours was relatively low: 0.159). Raum-Suryan and Harvey: Distribution and abundance of and habitat use by Phocoena phocoena 815 Fixed boat surveys Fifty-six sightings of 92 harbor porpoise were recorded during 33 surveys (port or starboard) in 1991, and 69 sightings of 118 harbor por- poise during 24 surveys (both sides of vessel surveyed) in 1992 (Fig. 2). Harbor porpoise were sighted during 79% of surveys in 1991 and 75% in 1992. Mean group size was 1.6 harbor porpoise (SE=0.09, n=56) in 1991 and 1.7 harbor porpoise (SE=0.127, n=69) in 1992. Distribution of harbor porpoise was patchy but similar between 1991 and 1992 (Fig. 2). The greatest number of harbor porpoise sightings recorded per survey were along transects 1, 2, and 5 in 1991 and transects 1 and 5 in 1992 (Table 2; Fig. 2). The least num- ber of sightings were recorded for transects 3 and 6 in 1991 and 1992 (Table 2; Fig. 2). Mean number of harbor porpoise sightings per sur- vey was not significantly different (P>0.05) between 1991 and 1992 for any of the fixed transects (Table 2). Sample sizes for all transects were low because of the limited sur- vey period (July to August). Given our low sample size (Cohen, 1988, requires a sample size of eight or more), we were unable to de- termine power. If eight samples of each fixed transect line had been taken, power to detect a difference in density between 1991 and 1992 would still have been low (power<27% for all fixed transects except transect three which had<6% power). Section A Section B Section C Section D Section E n = 14 n = 11 n = 15 «=I5 n = 12 b) 180 Section A Section B Section C Section D Section E Figure 5 Mean number of harbor porpoise (A) and mean water depth (B) of each section (A-E) determined during random boat surveys ( June- October 1992). Vertical lines represent standard error and “n” rep- resents number of transects completed in each section (A; total=73) and random depth locations plotted within each section (B); to- tal=584). Discussion Population and density estimates of harbor porpoise were based on several assumptions of line transect theory. Relevant assumptions included the following: 1) study area was sampled randomly (transect lines placed randomly with respect to the distribution of objects) or animals were randomly distributed; 2) all animals on the trackline were de- tected; 3) group size was estimated without error; 4) locations were measured accurately for each indi- vidual or group; and 5) animals did not move in re- sponse to the survey vessel or were detected before they moved (Burnham et al., 1980). The first assumption of line transect theory was met by employing a stratified random sampling de- sign within the study area. This design was chosen so that the 8-km transects would adequately cover the entire study area. By using fixed length straight transects, however, certain areas of sections B and C were not adequately sampled. The habitat features of these areas were similar to the rest of the study area, and portions of section C not sampled during random surveys were sampled during fixed transect surveys 5 and 6. A study design that incorporated shorter transects (4 km) would allow more complete coverage of all areas within strata. If this study were replicated, we recommend incorporating 4-km transects to cover the areas that we missed. We do not believe, however, that our study design affected the results of the habitat correlates. By randomly surveying within a defined region off the northern San Juan Islands, we adequately sampled oceano- graphic features of interest (depth, seafloor slope, surface temperature, tides). The assumption that all animals are detected on the trackline is often violated during marine mam- mal surveys. Animals with long durations of submer- 816 Fishery Bulletin 96(4), 1998 gence have a high probability of remaining undetec- ted during the passage of an aircraft or vessel, re- sulting in availability bias (Marsh and Sinclair, 1989). Several studies (e.g. Marsh and Sinclair, 1989) of harbor porpoise have indicated, on the basis of perception bias, that the probability of detecting a harbor porpoise on the trackline, gtO), is less than one (Barlow, 1988; Palka, 1993; Calambokidis3). Using an independent team of three observers, Barlow (1988) reported an estimated 22% of harbor porpoise that surfaced on the trackline were missed by a team of five observers (perception bias) travel- ing on a vessel at 18.5 km/h. Using three observers per survey, Calambokidis3 and Palka (1993) esti- mated the probability of observing a group of harbor porpoise on the trackline, g(0), was less than 0.5. We assumed g(0) was one during our study because we were unable to determine availability or perception bias. It is probable that some porpoise did avoid the vessel and might have been sub- merged for up to five minutes (Raum-Suryan, 1995). It is, therefore, likely thatg(O) is less than one and harbor porpoise abundance is underestimated. The ability to estimate group size can vary by the number of animals within the group and by the species of interest. Data from land- based calibration studies off the Washington coast indi- cated that observers on ves- 3 Calambokidis, J. 1991. Vessel surveys for harbor porpoise off the Washington coast. In H. Kajimura (ed.). Harbor porpoise interactions with Makah Salmon set net fishery in coastal Washington waters, 1988- 89. National Marine Mammal Laboratory Processed Report, Na- tional Marine Mammal Laboratory, Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA 98115. 70 25 50 75 100 125 150 175 200 225 250 Depth (m) i | Observed number of porpoise [~| Expected number of porpoise Figure 6 Depth distribution of 275 harbor porpoise sightings determined from random boat surveys (June to October 1992) in relation to expected distribution of harbor porpoise if they were distributed randomly with depth (as determined from depths at 584 ran- dom locations). An asterisk (*) designates a significant (P< 0.05) difference determined with chi-square goodness-of-fit analyses. Table 2 Mean, standard error (SE), number of sightings, and number of harbor porpoise determined during fixed boat surveys in 1991 and 1992. In 1992, the number of harbor porpoise observed (Obs.) are presented, as are values from four randomly chosen surveys used in analysis (Anal.), comparing mean number of harbor porpoise sighted along each transect in 1991 and 1992). n refers to the number of transect “sides” (bow out to 90° on port or starboard) surveyed. 1992 Transect no. 1991 Mean SE n No. of sightings No. of porpoise Probability Mean SE n No. of sightings No. of porpoise Obs. Anal. Obs. Anal. Obs. Anal. Obs. Anal. Obs. Anal. i 2.20 0.80 5 11 18 2.50 2.00 0.42 0.41 8 4 20 8 31 13 P = 0.843 2 1.83 0.48 6 11 16 1.25 1.25 0.45 0.63 8 4 10 5 22 17 P = 0.474 3 1.00 0.38 7 7 13 0.13 0.25 0.13 0.25 8 4 1 1 1 1 P = 0.200 5 2.83 0.54 6 17 29 2.25 3.00 0.49 0.58 8 4 18 12 25 18 P = 0.844 6 0.60 0.40 5 3 6 0.88 0.50 0.40 0.29 8 4 7 2 12 3 P= 0.853 7 1.75 0.63 4 7 10 1.50 1.00 0.82 0.41 8 4 12 9 26 19 P = 0.356 Raum-Suryan and Harvey: Distribution and abundance of and habitat use by Phocoena phocoena 817 sels underestimated true group size of harbor porpoise, missing up to 60% or more of animals (Calambokidis et al.4). However, these results were based on a very small sample size and the survey vessel traveled at twice the speed of our vessel. In our study, mean group size (1.91 por- poise) was only 5% different from that detected from concurrent shore-based surveys (2.01 por- poise; Raum-Suryan, 1995). We are, therefore, confident that any biases in group size estimates are small. Accurately measuring the locations of marine mammals from vessels can be affected by the height of observers above water (Polacheck and Smith, 1989) and the use of reticle and compass binoculars (Smith, 1982; Barlow and Lee, 1994). From a low height above the water, angles are greatly affected by small deviations in reticle estimates. As radial distances to harbor porpoise decrease, however, errors in reticle estimates (sighting angles) have progressively less effect on distance calculations. The majority (78%) of our radial and perpendicular sighting distances were less than 350 m. At 350 m an error of ±0.1 reticle was equal to 50 m (the range of our data groupings which best fitted the model). Therefore, although the platform height of our survey vessel was low (2.68 m), we were able to obtain accurate sighting data by conducting sur- veys only during optimal sighting conditions (Beaufort <1), and be- cause there was both a lack of ocean swell and the majority of sightings were less than 350 m distant. Harbor porpoise are small, in- conspicuous animals that avoid boats (Amudin and Amudin, 1974; Gaskin, 1977; Prescott and Fiorelli, 1980; Barlow, 1988). Detection of harbor porpoise before they be- come aware of the survey vessel is often difficult without prior knowl- edge of their locations. Polacheck and Thorpe (1990) observed har- bor porpoise swimming away from their survey vessel a significant 90- 80- 70- 60- 50- « 40- 30- 20- 10- n= 120 n = 96 n= 120 = 136 Section A Section B Section C Section D Section E Figure 7 Mean percentage slope of seafloor for each section (A- E) de- termined from random boat surveys. Vertical lines represent standard error and “n” represents the number of random slope locations plotted within each section (total=584). 8 10 12 14 16 18 20 22 24 26 >26 12 14 16 Percentage Slope Observed porpoise □ Expected porpoise Figure 8 Harbor porpoise sightings (n= 275) from random boat surveys (June— October 1992) in relation to expected distribution if porpoise were distributed randomly with slope (as determined from slopes at 584 random locations). An asterisk (*) designates a significant (P<0.05) difference, determined with chi-square goodness-of-fit analyses. 4 Calambokidis, J., S. R. Melin, and D. J. Rugh. 1991. Land- based calibrations of harbor porpoise sightings from a vessel along the northern Washington coast. In H. Kajimura ( ed. ), Harbor porpoise interactions with Makah Salmon set net fish- ery in coastal Washington waters, 1988-89. National Marine Mammal Laboratory Processed Report. National Marine Mam- mal Laboratory, Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA 98115-0070. Fishery Bulletin 96(4), 1 998 proportion of time. Barlow (1988) reported that har- bor porpoise quickly avoid a closely approaching sur- vey vessel. Vessel avoidance by harbor porpoise may result in animals remaining undetected by observ- 10.5 II 11.5 12 12.5 13 13.5 14 14.5 15 15.5 16 >16 Temperature (°C) [ | Frequency of porpoise sightings □ Expected frequencies (boat survey temperatures) Figure 9 Sea surface temperatures (°C) recorded at harbor porpoise sightings (n= 267) in relation to expected distribution of harbor porpoise if they were distributed randomly with temperature (as determined from temperatures at 427 locations along 73 random transect lines). An asterisk (*) designates a significant (P<0.05) difference, determined using chi-square goodness-of-fit analyses. ers or may affect estimates of perpendicular distance of harbor porpoise from the vessel. If the frequency of harbor porpoise sightings were greatest near the trackline and decreased with increasing perpendicu- lar distance during this study, it appeared that most harbor porpoise were detected before po- tentially significant vessel avoidance occurred. If porpoise did avoid the vessel, our abundance estimates would be underestimated. Ship avoid- ance was likely constant throughout the sur- vey period, however, and would not have af- fected results of habitat correlates. Density estimates of harbor porpoise (1.26 porpoise/km2) within the study area (237 km2) were greater than densities reported by Calambokidis et al.2 (0.42 porpoise/km2) for waters off the San Juan Islands and part of the Strait of Georgia (2291 km2) but were similar to density estimates of Flaherty and Stark1 (0.85 to 1.63 porpoise/km2) for the north and west San Juan Islands (1005 km2). Density es- timates reported by Calambokidis et al.2 were based on an initial g( 0) equal to 0.324 (CV= 0.171) multiplied by a correction factor of 3.1 and yielding a g(0) of one. Green et al.5 sur- veyed an extensive area within the 100-m isobath off the coast of Oregon and Washington and also reported a much lower density of har- bor porpoise (0.17 porpoise/km2) than reported here. These differences probably result from Green et al.5 and Calambokidis et al.2 includ- ing regions of high and low harbor porpoise abundance in contrast to our focus on high density areas off the northern San Juan Islands. Prey or habitat requirements often limit distribution of cetaceans to regions that may vary daily, seasonally, or yearly, de- pending on an individual’s foraging, mat- ing, or behavioral requirements. During this study, surveys were conducted only within the summer months (June to Octo- ber) and thus may account for the rela- tively high density estimates of harbor porpoise within our study area. Flaherty and Stark1 sighted harbor porpoise dur- ing all months of the year off the San Juan 5 Green, G. A., J. J. Brueggeman, C. E. Bowlby, R. A. Grotefendt, M. L. Bonnell, and K. T. Balcomb III. 1992. Cetacean distribution and abundance off Oregon and Washington, 1989-1990. In J. J. Brueggeman (ed. ), Final report prepared by Ebasco Environmental and Ecological Consulting, Inc. for Minerals Management Service, Pacific OCS Region. Offshore Continental Shelf (OCS) Study MMS 91-0093, 100 p. Figure 10 Mean number of harbor porpoise observed per minute during random boat surveys (June to October 1992) for all sections (A-E) combined. Vertical lines represent standard error. Raum-Suryan and Harvey: Distribution and abundance of and habitat use by Phocoena phocoena 819 Islands but observed more harbor porpoise in sum- mer months (June to August) than other times of the year. Surveys of harbor porpoise throughout the year along the east and west coasts of the United States have indicated a seasonal pattern among various regions (Neave and Wright, 1968; Gaskin and Watson, 1985; Barlow, 1988; Green et al.5). Results of fixed transect surveys conducted in our study in- dicated no change in distributions of harbor porpoise between July and August 1991 and 1992. Clumped distribution of harbor porpoise along tracklines was likely associated with habitat features (harbor por- poise were sighted most often over deep water). Among island regions, such as the Bay of Fundy, Glacier Bay, Alaska (Taylor and Dawson, 1984), and off the San Juan Islands, harbor porpoise are more often associated with deeper waters than along coastal regions of North America. Most harbor por- poise observed off the coast of California, Oregon, and Washington occurred at shallow water depths, and sightings decreased with increasing depth (Barlow, 1988; Dorfman, 1990; Calambokidis3; LaBarr and Ainley6). Incidental net entanglement of harbor porpoise within Washington waters oc- curred at the bottom of nets, at depths of 73 to 81 m (Scheffer and Slipp, 1948), and near the bottom or in the lower one-half of nets set from 11 to 18 m deep, indicating porpoise were foraging along the bottom or in deeper areas of the net (Gearin et al.7). The depth of water where harbor porpoise were sighted in this study may have been due to occurrence of prey within these areas. The Pacific herring (Clupea pallasi) population in the Strait of Georgia is the largest known in Wash- ington state, and herring are quite abundant in sec- tions of the eastern Strait during summer, fall, and winter (Lemberg, 1978). During this study, harbor por- poise, harbor seals, and a minke whale were observed feeding on a school of Pacific herring. The dominant prey items in stomachs of harbor porpoise taken in a setnet fishery in summer off northern Washington were Pacific herring, market squid (Loligo opalescens), gadids, and osmerids (Gearin and Johnson8). Pacific herring and market squid migrate vertically within the water column, remaining close to the seafloor during 6 LaBarr, M. S., and D. G. Ainley. 1985. Depth distribution of harbor porpoise off central California: a report of cruises in April and May-June 1985. NMFS Contract No. 41 USC 252, 23 p. 7 Gearin, P. J., M. A. Johnson, and S. Joner. 1991. Harbor por- poise interactions with the Makah Chinook Salmon Set-Net Fishery, 1988-89. In H. Kajimura (ed.), Harbor porpoise in- teractions with Makah salmon set net fishery in coastal Wash- ington waters, 1988-89. National Marine Mammal Laboratory Processed Report. National Marine Mammal Laboratory, Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA 98115-0070. the day, and approach the surface at night (Hart, 1973; Blaxter, 1985; Flaherty and Stark1). Characteristics of these prey items and the greater occurrence of porpoise over deep waters may indicate that harbor porpoise feed in deep water during the day. Aggregations of sur- face schooling fish and associated harbor porpoise were rarely ( 1% of surveys) observed within our study area, further indicating that harbor porpoise were likely feed- ing on prey in deep water. In our study, harbor porpoise were sighted most often in shallow sloping areas with little bathymet- ric relief. These results contrast with those of Flaherty and Stark,1 in which 70% of harbor por- poise sighted were found in areas with seafloor re- lief greater than 40%, and with those of Calam- bokidis,3 who observed significantly more harbor porpoise than expected within areas of uneven bot- tom topography off the outer Washington coast. It is likely that slope of the seafloor does not significantly affect the distribution of harbor porpoise or their prey in our study area. We believe that harbor porpoise and their prey are associated with deeper waters in this region which, in general, has shallow slopes. Water temperature may influence the distribution of harbor porpoise. Calambokidis3 reported harbor porpoise sightings in water temperatures ranging from 9° to 16°C off Washington. In the Bay of Fundy, Watts and Gaskin (1985) found a negative correla- tion between harbor porpoise abundance and mean August SST, and Watson (1981) reported that har- bor porpoise occurred in water temperatures less than 15°C in the Bay of Fundy. It is unlikely, how- ever, that SST alone would influence harbor porpoise distribution. Most harbor porpoise entered Fish Har- bor, New Brunswick, Canada, when SST was between 9° and 10°C, a period when large numbers of juve- nile herring were also entering the region (Gaskin and Watson, 1985). Within the Bay of Fundy, Watts and Gaskin (1985) found herring associated with vertically mixed waters and reduced surface tempera- tures. This association was possibly due to increased concentrations of zooplankton, which also occurred along convergent zones (Watts and Gaskin 1985). Sea surface temperatures off the northern San Juan Is- lands, therefore, were possibly related to tidal cur- rents that may be associated with concentrations of harbor porpoise prey. 8 Gearin, P. J., and M. A. Johnson. 1991. Prey identified from stomachs of harbor porpoise and Chinook salmon from the 1988- 89 Makah Salmon Set-Net Fishery. In H. Kajimura (ed ), Har- bor porpoise interactions with Makah Salmon set net fishery in coastal Washington waters, 1988-89. National Marine Mam- mal Laboratory Processed Report. National Marine Mammal Laboratory, Natl. Mar. Fish. Serv., NOAA, 7600 Sand Point Way NE, Seattle, WA 98115-0070. 820 Fishery Bulletin 96(4), 1998 In this study, SST was measured along tracklines and may not have represented water temperatures where harbor porpoise were sighted. Tide rips mix- ing water, or currents moving through the study area could have altered water temperatures by a few degrees between trackline and harbor porpoise locations. Sea surface temperature varied by 5°C from beginning to end of the 8-km tracklines. We assumed, however, less bias was introduced by collecting SST along the trackline than by continu- ously going off transect and potentially disturbing harbor porpoise ahead of the vessel. Because our methods were consistent over the study period, the comparison in use versus availability of SST is likely representative. It is doubtful that time of day had a significant effect on the ability to sight harbor porpoise in our study; therefore, other environmental factors must have affected harbor porpoise distribution in rela- tion to time of day. Occurrence of harbor porpoise appears closely associated with the strength of tidal currents. From shore-based surveys within our study area (Raum-Suryan, 1995), mean number of porpoise observed per minute was greatest two hours before each peak in the maximum flood tide, and signifi- cantly more (P<0.05) porpoise were observed per minute during flood than ebb tides. From June to October 1992, the majority of low tides in the north- ern San Juan Islands occurred in the early morning hours. The relation between the occurrence of har- bor porpoise with tide and time of day indicates that porpoise movements may have been associated with concentrations of prey in flood currents and tide rips. It is possible that harbor porpoise range throughout Washington Sound but continue to return to north- ern San Juan Island waters as a primary foraging area. We found a large proportion of the harbor porpoise population of Washington Sound located within our study area. Calambokidis et al.2 estimated the popu- lation size of harbor porpoise for the San Juan Is- lands (2291 km2) at 960 animals (corrected as in den- sity estimate). In approximately 10% (237 km2) of the area that Calambokidis et al.2 surveyed, we esti- mated 30% (299 porpoise) of the harbor porpoise population. Given that ourg(0) was assumed to be one (thus underestimating the population size), the proportion of harbor porpoise within our study area is likely greater than 30% of the total population within the San Juan Island region. On the basis of pollutants detected in harbor porpoise tissues, por- poise along the west coast do not mix freely between California, Oregon, and Washington (Calambokidis and Barlow, 1991). In addition, Washington and Cali- fornia are considered repositories of genetic diver- sity for harbor porpoise of the Northeast Pacific (Rosel et al., 1995) and also indicate that harbor por- poise ranges may be restricted. In addition to aerial surveys conducted over Washington waters (Calam- bokidis et al.2), our study area appears to be an im- portant site for monitoring trends in distribution and abundance of harbor porpoise in inland water of Washington. It is not clear why harbor porpoise are not as abun- dant in other areas of Washington Sound as they are in our study area. The relatively low abundance out- side our study area may be due to factors other than food availability, such as pollution, fishing pressure, increased boat traffic, or other environmental changes. We believe that harbor porpoise are more abundant in our study area than in other parts of Washington Sound because certain environmental conditions (deep, cool water, and strong tidal mix- ing) influence the distribution of harbor porpoise prey. Future monitoring studies on oceanographic conditions and prey availability associated with har- bor porpoise sightings would greatly assist in deter- mining mechanisms affecting harbor porpoise abun- dance and distribution in this and other areas and help in managing this genetically important stock. Acknowledgments This study would not have been possible without the help of many people. We would especially like to thank Birgit Kriete, Robert DeLong, Steve Jeffries, Steve Osmek, Harriet Huber, and Dave Rugh for lo- gistical support and assistance in obtaining funding in 1991 and 1992. We sincerely thank Jay Barlow and William Broenkow for critical review of an ear- lier version of this manuscript and three anonymous reviewers for helpful comments on the final manu- script. We are indebted to those who assisted in the field, Rob Suryan.Tomo Eguchi, Karen Russel, Doug Huddle, John Raum, and Marilyn Raum. Funding and support for this project was provided by the National Marine Fisheries Service (National Marine Mammal Laboratory), The Washington Department of Wildlife,. Earl H. and Ethel M. Myers/Oceano- graphic and Marine Biology Trust, Lerner-Gray Fund for Marine Research, Packard Foundation, The Whale Museum, and Save The Whales, Inc. Literature cited Amudin, M., and B. Amudin. 1974. On the behaviour and study of the harbour porpoise, Phocoena phocoena, in the wild. In G. Pilleri (ed.), Inves- Raum-Suryan and Harvey: Distribution and abundance of and habitat use by Phocoena phocoena 821 tigations on Cetacea, vol.V, p. 317-328. Hirnanatomische Institut der Universitat, Berne. Barlow, J. 1988. Harbor porpoise, Phocoena phocoena, abundance es- timation for California, Oregon, and Washington: I. Ship surveys. Fish. Bull. 86(31:417-432. Barlow, J., and T. Lee. 1994. The estimation of perpendicular sighting distance on SWFSC research vessel surveys for cetaceans: 1974 to 1991. U.S. Dep. Commer., NOAA Tech. Memo.NMFS, NOAA, La Jolla, CA, 92038, 46 p. Blaxter, J. H. S. 1985. The herring: a successful species? Can. J. Fish. Aquat. Sci. 42 (suppl. 11:21-30. Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 1993. Distance sampling: estimating abundance of biologi- cal populations. Chapman and Hall, New York, NY, 446 p. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wild. Monogr. 72.T-202. Calambokidis, J., and J. Barlow. 1991. Chlorinated hydrocarbon concentrations and their use for describing population discreteness in harbor por- poises from Washington, Oregon, and California. In J. E. Reynolds III and P. K. Odell (eds.l, Marine mammal strandings in the United States: proceedings of the second marine mammal stranding workshop Miami, FL, Dec. 3- 5, 1987, p. 101-110. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 98. Calambokidis, J., J. Peard, G. H. Steiger, J. C. Cubbage, and R. L. DeLong. 1984. Chemical contaminants in marine mammals from Washington state. U.S. Dep. Commer., NOAA Tech. Memo., NOS OMS 6, 167 p. Calambokidis, J., S. M. Speich, J. Peard, G. H. Steiger, J. C. Cubbage, D. M. Fry, and L. J. Lowenstine. 1985. Biology of Puget Sound marine mammals and ma- rine birds: population health and evidence of pollution effects. U.S. Dep. Commer., NOAA Tech. Memo., NOS OMA 18, 159 p. Cohen, J. 1988. Statistical power analysis for the behavioral sci- ences. Lawrence Erlbaum Assoc., Publ., Hillsdale, NJ, 567 p. Dorfman, E. J. 1990. Distribution, behavior, and food habits of harbor por- poises ( Phocoena phocoena) in Monterey Bay. M.S. the- sis, Moss Landing Marine Laboratories, Moss Landing, CA, 57 p. Everitt, R. D., C. H. Fiscus, and R. L. DeLong. 1980. Northern Puget Sound marine mammals. EPA-600/ 7-80-139. National Marine Mammal Laboratory, Natl. Mar. Fish. Serv., NOAA, Seattle, WA, 134 p. Gaskin, D. E. 1977. Harbour porpoise Phocoena phocoena (L.) in the west- ern approaches to the Bay of Fundy 1969-75. Rep. Int. Whal. Comm. 27:487-492. Gaskin, D. E., and A. P. Watson. 1985. The harbor porpoise, Phocoena phocoena, in Fish Harbour, New Brunswick, Canada: occupancy, distribution, and movements. Fish. Bull. 83(31:427-442. Gerrodette, T. 1987. A power analysis for detecting trends. Ecology 68(51:1364-1372. Hart, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can. Bull. 180, 740 p. Kozloff, E. N. 1973. Seashore life of Puget Sound, the Strait of Georgia, and the San Juan Archipelago. Univ. Washington Press, Seattle, WA, 282 p. Laake, J. L., S. T. Buckland, D. R. Anderson, and K. P. Burnham. 1993. DISTANCE user’s guide. Colorado Cooperative Fish and Wildlife Research Unit, Colorado State University, Fort Collins, CO 80523, 84 p. Lemberg, N. A. 1978. Hydroacoustic assessment of Puget Sound herring, 1972-1978. Washington Dep. Fisheries, Tech. Rep. 41, 43 p. Marsh, H., and D. F. Sinclair. 1989. Correcting for visibility bias in strip transect aerial surveys of aquatic fauna J. Wildl. Manage. 53:1017-1024. Neave, D. J., and B. S. Wright. 1968. Seasonal migrations of the harbor porpoise (Phocoena phocoena) and other cetacea in the Bay of Fundy. J. Mamm. 49(21:259-264. National Oceanic and Atmospheric Administration. 1991. High and low water predictions: west coast of North and South America including the Hawaiian Island. U.S. Dep. Commer., NOAA, 238 p. Palka, D. L. 1993. Estimating density of animals when assumptions of line-transect surveys are violated. PhD. diss., Univ. Cali- fornia, San Diego, CA, 69 p. Polacheck, T., and T. D. Smith. 1989. A proposed methodology for field testing line transect theory for shipboard surveys of cetaceans. Rep. Int. Whal. Comm. 39:341-345. Polacheck, T., and L. Thorpe. 1990. The swimming direction of harbor porpoise in rela- tionship to a survey vessel. Rep. Int. Whal. Comm. 40:463-470. Prescott, J. H., and P. M. Fiorelli. 1980. Review of the harbor porpoise (Phocoena phocoena) in the U.S. Northwest Atlantic. U.S. Dep. Commer., Na- tional Tech. Info. Ser., pb-80-176928, 64 p. Raum-Suryan, K. L. 1995. Distribution, abundance, habitat use, and respira- tion patterns of harbor porpoise (Phocoena phocoena) off the northern San Juan Islands, Washington. M.S. the- sis, Moss Landing Marine Laboratories, Moss Landing, CA, 79 p. Resampling Stats, Inc. 1995. Resampling Stats users guide. Resampl. Stats., Inc., Arlington, VA, 128 p. Rosel, P. E., A. E. Bizon, and M. G. Haygood. 1995. Variability of the mitochondrial control region in populations of the harbour porpoise, Phocoena phocoena, on interoceanic and regional scales. Can. J. Fish. Aquat. Sci. 52:1210-1219. Scheffer, V. S., and J. Slipp. 1948. The whales and dolphins ofWashington State with a key to the cetaceans of the west coast of North America. Am. Midi. Nat. 39:257-337. Smith, T. D. 1982. Testing methods of estimating range and bearing to cetaceans aboard the RV PS. Jordan. U.S. Dep. Commer, NOAA Tech. Memo. NMFS, NOAA, La Jolla, CA 92038, 30 p. 822 Fishery Bulletin 96(4), 1998 Smith, G. J. D., and D. E. Gaskin. 1983. An environmental index for habitat utilization by female harbour porpoises with calves near Deer Island, Bay of Fundy. Ophelia. 22:1-15. Taylor, B. and P. K. Dawson. 1984. Seasonal changes in density and behavior of harbor porpoise ( Phocoena phocoena) affecting censusing meth- odology in Glacier Bay National Park, Alaska. Rep. Int. Whal. Comm. 34:479-483. Watson, A. P. 1976. The diurnal behaviour of the harbour porpoise ( Phocoena phocoena L.) in the coastal waters of the west- ern Bay of Fundy. M.S. thesis, Univ. of Guelph, Ontario, Canada, 94 p. 1981. Sea guide to whales of the world. E.P. Dutton, New York, NY, 302 p. Watts, P., and D. E. Gaskin. 1985. Habitat index analysis of the harbor porpoise ( Phocoena phocoena) in the southern coastal waters of the Bay of Fundy. Can. J. Zool. 61( 1 ): 126—132. Zar, J. H. 1984. Biostatistical analysis (second ed.). Prentice-Hall, Inc., Englewood Cliffs, NJ, 718 p. 823 Global phylogeography of mackerels of the genus Scomber Daniel R. Scales* Bruce B. Collette*' ** John E. Graves* * School of Marine Science Virginia institute of Marine Science College of William and Mary Gloucester Point, Virginia 23062 ** National Systematics Laboratory National Marine Fisheries Service, NOAA National Museum of Natural History Washington, DC, 20560 Present address (for D, R. Scoles): Neurogenetics Laboratory, Division of Neurology Cedars-Sinai Medical Center UCLA School of Medicine 8700 Beverly Boulevard Los Angeles, California 90048 E-mail address (for D. R. Scoles): scolesd@csmc.edu Abstract .—Inter- and intraspecific genetic relationships among and within three species of mackerels of the genus Scomber were investigated by restric- tion site analysis of the whole mito- chondrial (mt) DNA genome and direct sequence analysis of the mitochondrial cytochrome b gene. A total of 15 samples, averaging 19 individuals each, were col- lected from geographically isolated populations throughout the ranges of S. scombrus (two samples), S. austra- lasicus (five samples), and S. japonicus (eight samples). Restriction site analy- sis with 12 restriction enzymes re- vealed substantial genetic variation within each species. Sample haplotype diversities ranged from 0.28 to 0.95, and nucleotide sequence diversities from 0.13% to 0.76%. Spatial partition- ing of genetic variation was observed in each of the species. Eastern and western North Atlantic samples of S. scombrus exhibited significant hetero- geneity in the distribution of mtDNA haplotypes, but no fixed restriction site differences were observed between samples. Similarly, no fixed restriction site differences occurred among samples of S. japonicus in the Atlantic Ocean, although there were significant differ- ences in the distribution of haplotypes among samples. In contrast, samples of S. japonicus from within the Pacific Ocean were characterized by fixed re- striction site differences. North and South Pacific samples of S. austra- lasicus were highly divergent, and one of two divergent mtDNA matrilines was restricted to samples from the South Pacific. A 420-bp segment of the cyto- chrome b gene was sequenced for rep- resentatives of each of the major mtDNA lineages identified by restric- tion site analysis. Scomber scombrus differed from S. australasicus and S. japonicus by more than 11% net nucle- otide sequence divergence, considerably greater than the 3.5% sequence diver- gence between S. australasicus and S. japonicus. Levels of interspecific ge- netic divergences based on restriction site data were similar in pattern, but were approximately 20% lower in mag- nitude when based on the cytochrome b sequences. Parsimony analysis and neighbor-joining of restriction site data, and parsimony analysis of cytochrome b sequences showed similar paraphyletic patterns in both S. japonicus and S. australasicus. Levels of divergence among samples of S. japonicus were similar to those between samples of .S'. australasicus and S. japonicus. Com- plete partitioning of halpotypes among some samples of S. japonicus that are morphologically distinct suggests that Atlantic and Indo-Pacific populations of S. japonicus may need to be recognized as separate species. Manuscript accepted 26 January 1998. Fish. Bull. 96: 823-842 (1998). Population structure is largely in- fluenced by the biological character- istics of a species and the attributes of its environment that promote or impede gene flow. In marine fishes, dispersal of planktonic early life history stages and adult vagility can facilitate genetic exchange over great distances (Rosenblatt, 1963; Shulman and Bermingham, 1995). Features of the marine environ- ment that limit gene flow are often equally large in scale, such as current systems, major changes in temperature or salinity, or the presence of large land masses (Sinclair, 1988). Accordingly, many genetic analyses of broadly distrib- uted marine fishes have shown little divergence among conspecific samples from geographically dis- tant locations. Limited population structure has been demonstrated for a variety of marine fishes. Electrophoretic analyses of allozymes have revealed little population divergence among collections of milkfish {Chanos chanos ) separated by up to 10,000 km (Winans, 1980), damselfish, Stegastes fasciolatus, sampled from throughout the 2500 km Hawaiian archipelago (Shaklee, 1984), in twelve species of tropical marine shore fishes sampled from both sides of the Pacific Barrier, a 5000- km expanse of deep ocean separat- ing central and eastern Pacific shal- low water habitats (Rosenblatt and Waples, 1986), or in five species of damselfishes collected from isolated Caribbean reefs (Lacson, 1992 ). Similarly, analyses of mitochondrial DNA have revealed little genetic dif- ferentiation among populations of five species of tropical reef fishes sampled from throughout the Car- ibbean ( Shulman and Bermingham, 1995), or within three cosmopolitan species of tunas (Graves et al., 1984; Graves and Dizon, 1989; Scoles and Graves, 1993). Much less information is avail- able on the population structure of broadly distributed marine species that occur in fragmented or disjunct distributions. Recent studies of striped mullet (Mugil cephalus, 824 Fishery Bulletin 96(4), 1998 Crosetti et al., 1994) and bluefish (Pomatomus saltatrix, Goodbred and Graves, 1996) have demon- strated that significant phylogeographic population structuring can exist in discontinuously distributed marine fishes. The reduced gene flow among isolated populations of such species provide excellent oppor- tunities for studying the effects of historical marine zoogeographical processes and observe incipient speciation. Mackerels of the genus Scomber, like the striped mullet and bluefish, are ideally suited for studies of genetic population structure in cosmopolitan fishes with fragmented distributions and for comparative studies that describe genetic patterns of other more vagile scombrids (Graves et al., 1984; Graves and Dizon, 1989; Scoles and Graves, 1993). Distributional patterns of scombrids vary widely from world-wide species such as the yellowfin tuna, albacore, and skip- jack tunas to supposedly wide-spread species such as the chub mackerel, S. japonicus , which is divided into geographically disjunct populations, to those with limited ranges such as a Spanish mackerel (Scomberomorus munroi) found only in the Gulf of Papua between Australia and New Guinea (Collette and Nauen, 1983). The three species of Scomber (S. japonicus, S. australasicus, and S. scombrus) occur in temperate to subtropical waters, and two (S. japonicus and S. australasicus ) display antitropical distributions. Each species occurs in disjunct popu- lations of various sizes, and their distributions over- lap in several areas (Fig. 1; Collette and Nauen, 1983). The cosmopolitan chub mackerel, S. japonicus, occurs in coastal regions and adjacent seas of the Atlantic, Pacific, and northwest Indian oceans. The spotted chub mackerel, S. australasicus, is restricted to the Pacific Ocean, southeastern Indian Ocean, and the Red Sea. The Atlantic mackerel, S. scombrus, is restricted to the North Atlantic Ocean. The present study emphasizes S. japonicus, which of the three species is the most widely distributed and morpho- logically divergent among its isolated populations (Matsui, 1967). Life histories and population structures The spawning behavior and duration of the larval stage varies among the three species of Scomber. Scomber scombrus is capable of spawning serially up to 30 times in a spawning season (Watson et al., 1992) at any time of day (Walsh and Johnstone, 1992), where S. japonicus spawns on average only 9 times in a season, and does so only at night (Watanabe, Figure 1 Sampling and distributions of Atlantic mackerel, Scomber scombrus (S-MAS and S-ENG), spotted chub mackerel, S. australasicus (A-RED, A-JPN, A-AUS, A-NZL, and A-MEX), and chub mackerel, S. japonicus (J-JPN, J-TWN, J-CAL, J-HAW, J-FLA, J-ARG, J-ISR, J-IVC, and J-SAF), according to Collette and Nauen (1983). S. japonicus distributions off Brazil and Namibia follow Perrotta and Aubone (1991) and Zenken and Lobov (1989), respectively. Scoles et al.: Global phylogeography of Scomber 825 1970). Scomber japonicus and S. scombrus have simi- lar larval stage durations: eggs of S. scombrus and S. japonicus hatch in less than 6 d, and schooling behavior begins when larvae metamorphose (about 15 mm) which occurs at 24-9 d and 22-9 d, respec- tively (Hunter and Kimbrell, 1980; Ware and Lam- bert, 1985). Thus the duration of passive planktonic transport of the early life history stages is usually 29 d or less, but probably extends into the early ju- venile stages as well. Little spawning or larval ecol- ogy data are available for S. australasicus, but it is expected that it has a similar early life history. Exchange between regional populations of S. scombrus in the North Atlantic appears sufficient to maintain genetic homogeneity. Two spawning groups in the northwest Atlantic, the “northern contingent” in the southern Gulf of St. Lawerence, and the “south- ern contingent” between Cape Cod and Cape Hatteras, were identified by Sette ( 1950) by size com- position and tagging data. However, these groups were not discriminated by meristic or growth char- acter analysis (MacKay and Garside, 1969; Simard et al., 1992), or by allozyme analysis (Maguire et al., 1987). In the northeast Atlantic, two spawning groups, the “western stock” south and west of the British Isles, and the “North Sea stock,” were iden- tified by tagging (Hamre, 1980), but these groups were not distinguished by allozymic differences (Jamieson and Smith, 1987). Matsui (1967) showed greater phenotypic varia- tion in S. japonicus than in the other two species of Scomber, possibly because of its wider distribution. Populations of S. japonicus from the eastern and western Atlantic had nonoverlapping distributions of gill-raker counts, and other variable morphologi- cal characters of Atlantic populations were similar, including belly spots, strongly crenulated teeth, and large scales. In contrast, Pacific S. japonicus exhib- ited lightly crenulated teeth, no belly spots, and smaller scales (Matsui, 1967). Analysis of four poly- morphic allozyme loci revealed no divergence among samples of S. japonicus from the southeast Atlantic off Namibia (Zenkin and Lobov, 1989), but heteroge- neity in immunological reactivity suggested popula- tion structure off the northwest African coast (Weiss, 1980). A study of 14 polymorphic allozyme loci showed significant differences between samples from the north- and southeastern Pacific Ocean, which sug- gested reduced gene flow across the tropics in com- parison with other similarly distributed pelagic fishes (Stepien and Rosenblatt, 1996). Differences in growth rates and morphological characters among samples within the southwest Atlantic led Perrotta (1993) to conclude that S', japonicus populations within the region had diverged to the level of subspecies. To evaluate the genetic relationships among the three species of Scomber, and the discontinuous popu- lations within them, we examined mitochondrial (mt) DNA restriction sites, and cytochrome b sequences. Analysis of mtBNA has proven useful in revealing phylogeographic structure in a variety of marine and freshwater fish species (Avise, 1992). Because mtDNA is clonally inherited, information regarding historical phylogenetic relationships is retained, hence analysis of mtDNA can render considerable information on historical relationships among popu- lations and the mtDNA lineages they possess. Materials and methods Specimen collection Samples of 15 to 21 individuals each of S. scombrus, S. australasicus, and S. japonicus were obtained from 15 locations (Table 1, Fig. 1). Specimens of Ras- trelliger kanagurta, a species of the sister group to Scomber (Collette et al., 1984), were obtained from Sri Lanka. Whole fish were frozen after collection and shipped to the laboratory on dry ice. MtDNA preparation and analysis MtDNA-enriched genomic DNA was isolated from 3 g of lateral red muscle, or whole young-of-year fish with digestive tracts removed (A-NZL only) accord- ing to Chapman and Powers (1984) method, modi- fied by the omission of sucrose step gradients and the use of 1.5% sodium dodecyl sulfate for mitochon- drial lysis. DNAs were digested with 10 hexameric (Apal, Bgll, Bsu36l, Oral, Pvull, Seal, Stul, Sspl, Hpal, Spe I) and 2 multi-hexameric (Aval, Hae II) re- striction endonucleases following manufacturers’ (Stratagene and Gibco-BRL) instructions. DNA frag- ments were separated by electrophoresis in 1.0 or 1.5% agarose gels, transferred to nylon membranes by Southern transfer, and hybridized to a biotin-la- beled probe as described previously (Scoles and Graves, 1993). The probe was made by separately cloning four yellowfin tuna, Thunnus albacares, mtDNA fragments that cover the entire mitochon- drial genome in the Pstl site of pBluescript SK- ( Stratagene). MtDNA fragments were visualized by using the BluGene Non-Radioactive Nucleic Acid Detection Kit (Gibco-BRL). A 12-letter composite mtDNA haplotype, indicat- ing the fragment pattern for each enzyme, was de- veloped for each individual. Letters were assigned to restriction fragment patterns as they were encoun- tered, beginning with A’ for the closely related 826 Fishery Bulletin 96(4), 1998 Table 1 Scomber japonicus , S. australasicus, and S. scombrus. Sample name, size, location, haplotype diversity ih), and percent nucle- otide sequence diversity (n). Sample Haplotype Percent nucleotide Sample name size (n) Sample location diversity (h) sequence diversity (tt) Scomber scombrus S-MAS 20 Boston, Massachusetts 0.85 0.29 S-ENG 20 Plymouth, England 0.28 0.07 Total 40 0.58 0.18 Scomber australasicus A-RED 15 Sinai Peninsula south, Red Sea 0.95 0.41 A-AUS 18 New South Wales, Australia 0.86 0.75 A-NZL 19 Wellington, New Zealand 0.75 0.77 A-JPN 21 Tokyo, Japan 0.81 0.30 A-MEX 20 Revillagigedo Islands, Mexico 0.59 0.13 Total 93 0.90 2.18 Scomber japonicus J-FLA 20 Panama City, Florida 0.93 0.40 J-ARG 18 Mar del Plata, Argentina 0.90 0.50 J-ISR 20 Mediterranean coast of Israel 0.90 0.38 J-IVC 20 Abidjan, Ivory Coast 0.91 0.39 J-SAF 20 Cape Town, South Africa 0.81 0.38 J-TWN 20 Kaohsing, Taiwan 0.86 0.29 J-JPN 20 Tokyo, Japan 0.88 0.35 J-CAL 20 San Diego, California 0.64 0.14 Total 158 0.95 2.42 Grand total 291 S.japonicus and S. australasicus group, and proceed- ing through the alphabet, and beginning with ‘Z’ and proceeding in reverse alphabetical order for S. scombrus. For the few S. scombrus restriction frag- ment patterns that also occurred in S. japonicus or S. australasicus, the latter species’ restriction morph letter designation was used. Estimates of nucleotide sequence divergence ( d ) among haplotypes were determined by using the approach of Nei and Li ( 1979) for fragment data, and Nei and Tajima (1981) and Nei and Miller (1990) for site data, with weighting based on the proportion of fragments or sites produced by each enzyme class (Nei and Tajima, 1983). Estimates of nucleon or hap- lotype diversity ih) and nucleotide sequence diver- sity in) were then calculated with the program DA of the statistical package REAP (McElroy et al., 1992). Corrected nucleotide sequence divergences (8) among populations were calculated by using the protocol of Nei and Li ( 1979). The homogeneity of haplotype dis- tributions among samples was evaluated by chi- square analyses using the Monte-Carlo method of Roff and Bentzen (1989) with 1000 randomizations of the data with REAP. Restriction sites were inferred from completely additive fragment patterns, and the information for each individual was coded using a presence-or-ab- sence matrix. Homology of restriction sites could not be assured between S. scombrus and the S. austra- lasicus-S. japonicus group owing to the relatively large genetic divergence between the species. Restric- tion site data were therefore evaluated only for phy- logenetic relationships within the S. australasicus- S. japonicus group, and an estimate of divergence between S. scombrus and the pooled data of S. australasicus and S. japonicus was determined by using restriction fragment data. Relationships among haplotypes were inferred by cluster analysis (un- weighted pair-group method, UPGMA, Norusis, 1988). Neighbor-joining, parsimony analyses and consensus trees were generated with the PAUP* soft- ware program (test version). DNA amplification and sequence analysis Individuals representative of the major Scomber mtDNA matrilines, identified by restriction site analysis, and of Rastrelliger kanagurta, were selected for DNA sequence analysis. A 418-bp region of the cytochrome b gene was amplified from mtDNA-en- riched genomic DNA isolations by the polymerase chain reaction (PCR) with primers L15079 Scoles et al.: Global phylogeography of Scomber 827 (5’GAGGCCTCTACTATGGCTCTTACC3') and H15497 ( 5’GCTAGGGTATAATTGTCTGGGTC GC C 3 ’ ) devel- oped by Finnerty and Block (1992) for blue marlin (Makaira nigricans). Double-stranded DNA amplifi- cations were accomplished in 100 pi of 10 mM Tris- HC1 pH 8.3, 50 mM KC1, 1.5 mM MgCl2, 0.001% (w/v) gelatin, 0.2 mM of each dNTP, 50 pmol of each primer, 50 ng genomic DNA, and 2.5 units Taq polymerase (Perkin-Elmer/Cetus). Cycling parameters were pre- ceded by a 4-min initial denaturation at 94°C and included 38 cycles of 94°C for 1 min, 50°C for 1 min, and 72°C for 1.5 min, followed by a 5-min final ex- tension at 72°C. Primers L15079 and H15497 would not amplify some haplotypes of S. australasicus. Specific primers with conserved 3' ends were de- signed from other sequences: CYT2A (5TACCTTTTCATGGAAACATG3') and CYT2B ( 5’ AAG AGGTTGGG AG AG AAG A3 ' ) . Both strands were fully sequenced by using the Sequenase kit (United States Biochemical) or the CircumVent Cycle Sequencing Kit (New England BioLabs) following the vendors’ protocols. Sequences were aligned by eye to the human mi- tochondrial cytochrome b sequence1 (Anderson et al., 1981). Nucleotide sequence divergences ( <7 ) among cytochrome b sequences were generated with the program K of REAP. Cytochrome b sequences were analyzed by parsimony analysis with rooting to Rastrelliger kanagurta by using ALLTREES in PAUP. Results MtDNA restriction site analysis of 40 Scomber scombrus from two locations revealed a total of 56 restriction sites, of which 16 were polymorphic, de- fining 13 haplotypes. Nucleotide sequence diver- gences among the haplotypes of S. scombrus ranged from <7=0.17% to 0.86%. An average of 294 bp was surveyed, representing 1.75% of the mitochondrial genome. The average size of the S. scombrus mtDNA genome, determined from several restriction frag- ment profiles from each of the 12 restriction enzymes, was 16,784 ±213 bp (SD). Restriction site analysis of 246 S. japonicus and S. australasicus revealed a total of 93 unique haplotypes, including information from 86 restric- tion sites, of which 58 were polymorphic. Among haplotypes of S. japonicus and S. australasicus , nucleotide sequence divergences ranged from <7=0.15% to 2.9%. On average, 310 bp were surveyed per individual, representing 1.85% of the mtDNA genome. The average size of the mtDNA genome of 1 GenBank accession no. V00662. S. japonicus or S. australasicus was estimated to be 16,781 ±195 bp (SD), not statistically different than that of S. scombrus. Genetic diversities MtDNA restriction site analysis revealed that the genetic variability was lower in S. scombrus than the other two species (Table 1). S-ENG was the least variable sample in the study (h= 0.28), consisting of four haplotypes of which three occurred only once (Table 2). Ten haplotypes were revealed in S-MAS, with three at elevated frequencies, resulting in a haplotype diversity of h= 0.85. The overall haplotype diversity for S. scombrus was /? =0.58, and nucleotide sequence diversity was 71=0.18%. Three restriction enzymes, Pauli, Bgll, and Bsu36l, were invariant in S. scombrus, whereas the remaining nine restriction enzymes each revealed from two to five fragment patterns. A wide range of diversities was observed among samples of S. australasicus (h- 0.59-0.95) and S. japonicus (77=0.64-0.93) based on the restriction site data. Eastern Pacific samples of S', australasicus and S. japonicus from Mexico (A-MEX, h =0.59, 71=0.13%) and California (J-CAL, /7=0.64, 71=0.14%) had lower diversities than other samples, possessing only four and six haplotypes, respectively (Tables 1 and 2). The highest haplotype diversity occurred in A-RED (h- 0.95), with five of 10 haplotypes represented twice; the highest nucleotide sequence diversities occurred in A-NZL (ti=0.77%) and A-AUS (71=0.75%) owing to the presence of two divergent mtDNA matrilines within each sample (Tables 1 and 2). With the exception of Bgll in S. australasicus, all restriction enzymes were variable in both species, revealing two to seven fragment patterns in S. australasicus, and two to 15 in S. japonicus. The enzyme Apa I revealed the greatest number of variants in both species. Genetic divergences Restriction fragment analysis showed S. scombrus to be highly divergent from S. australasicus and S. japonicus , with a net nucleotide sequence divergence of 8=11.9% between S. scombrus and the pooled data of S. japonicus and S. australasicus . In comparison, the mean net nucleotide sequence divergence be- tween S. japonicus and S. australasicus based on the restriction site data was only 8=1.17%. Probabilities of Roff and Bentzen chi-square tests among samples that shared haplotypes are given in Table 3. Scomber scombrus Considerable intraspecific di- vergence was evident in S. scombrus from the mtDNA 828 Fishery Bulletin 96(4), 1 998 Table 2 Distributions of mtDNA haplotypes among samples of species of Scomber, grouped by their occurrence in phenetic cluster analy- sis. Letters represent fragment patterns produced by the following enzymes (left to right): Seal, Dra I, Stul, Pvull, Haell, Apal, Aral, Sspl , Bgl I, Hpal, Spel, Rsw36I. Western Pacific Scomber japonicus ID Haplotype J-TWN J-JPN Total 1 BAIABIAGEBAA 4 5 9 2 BAIABLADEBAA 5 4 9 3 BAIABMAGEBAA 5 4 9 4 BAIABJADEBAA 2 1 3 5 BAIABMHGEBAA 2 2 6 BAIABJAGEBAA 1 1 7 BAIABLAGEBAA 1 1 8 BAIABMADEBAA 1 1 9 BAIABMAGEBCA 1 1 10 BAIABNAGEBAA 1 1 11 BAIABOAEEBAA 1 1 12 BAIABOAGEBAA 1 1 13 BAIAGIAGEBAA 1 1 Eastern Pacific Scomber japonicus ID Haplotype J-CAL Total 14 BAIABIADEAAA 12 12 15 BAIABJADEAAA 5 5 16 BAIABIADEAAB 1 1 17 BAIABIDDEAAA 1 1 18 BAIADIADEAAA 1 1 19 B A J AB I ADC AAA 1 1 Atlantic Scomber japonicus. ID Haplotype J-FLA J-ARG J-SAF J-IVC J-ISR Total 20 ABAAABAAAAAA 1 1 3 1 6 21 AAAAABADAAAA 1 1 2 22 ABAAABAHAAAA 3 3 23 ABAAABADAAAA 2 2 24 ABAAABAIAAAA 1 1 25 ABAAADAAAAAA 1 1 26 ABAAAKADAAAA 1 1 27 ABAAEBAAAAAA 1 1 28 ABAAEBADAAAA 1 1 29 ABABABAHAAAA 1 1 30 AAAAABAAAAAA 1 1 31 AAAAAAAAAAAA 6 4 3 2 15 32 AABAAAAAAAAA 4 4 33 AAAAAJAAAAAA 2 2 34 AAAAAAABAAAA 1 1 35 AAAAAPAAAAAA 1 1 36 AAAAAABAABAA 1 1 37 AABAAAAABAAA 1 1 38 AABAAAACAAAA 1 1 39 ABAAAAAAAAAA 1 1 40 BAAAAJAAAAAA 1 1 41 AAGAAAADAAAA 4 9 5 6 24 42 ABGAAAADAAAA 1 2 3 6 43 AAAAAAADAAAA 2 1 1 4 continued on next page Scoles et al.: Global phylogeography of Scomber 829 Table 2 (continuedj Atlantic Scomber japonicus. (continued) ID Haplotype J-FLA J-ARG J-SAF J-IVC J-ISR Total 44 AAGAAAADADAA 1 1 2 45 AAAAAAADAEAA 1 1 46 AAAAAEADAAAA 1 1 47 AAAAAGGDAAAA 1 1 48 AABAACADAAAA 1 1 49 AACBABABAAAA 1 1 50 AAGAAAADBAAA 1 1 51 AAGAAAAEAAAA 1 1 52 AAGAAAAEABAA 1 1 53 AAGAABADAAAA 1 1 54 AAHAAAADAAAA 1 1 55 ABAAAAADAAAA 1 1 56 ABAAAAADAAAB 1 1 57 ABAAAAADBAAA 1 1 58 BAAAAAADAAAA 1 1 Red Sea Scomber australasicus ID Haplotype A-RED Total 59 BADAAAADDEAA 2 2 60 BADAAAADDFAA 2 2 61 BADAFAADAFAA 2 2 62 BADBAAAAAFAA 2 2 63 BADBAAADAFAA 2 2 64 AADBAAADAF AA 1 1 65 BADAAAADAEAA 1 1 66 BADBABADAFAA 1 1 67 BALBAAADAFAA 1 1 68 BALBAJADAFAA 1 1 Unique lineage Scomber australasicus ID Haplotype A-NZL A-AUS Total 69 BCDBBAADABAA 8 6 14 70 BCDBBAADAAAA 1 1 71 BCDBBAEDABAA 1 1 72 BCDBBAFDABAA 1 1 73 BCDBBBADABAA 1 1 74 BCDBBGADABAA 1 1 75 BCDBBHADABAA 1 1 76 BCDBCAADABAA 1 1 77 BDDBBAADABAA 1 1 78 BDDBBAADABBA 1 1 Ubiquitous lineage Scomber australasicus ID Haplotype A-NZL A-AUS A-MEX A-JPN Total 79 BAEABACFABAA 1 1 12 9 23 80 BAEABACEACAA 6 4 10 81 BBEABACFABAA 2 1 3 82 BAEABACFABAA 5 5 83 BAKABACFABAA 3 3 84 BAEABJCFABAA 2 2 continued on next page 830 Fishery Bulletin 96(4), 1998 Table 2 (continued) Ubiquitous lineage Scomber australasicus (continued) ID Haplotype A-NZL A-AUS A-MEX A-JPN Total 85 BAEABABFABAA 1 1 86 BAEABACFAFAA 1 1 87 BAEABSCFABAA 1 1 88 BAEBBACFABAA 1 1 89 BAFABACEACAA 1 1 90 BBDABACFABAA 1 1 91 BBEBBRCFABAA 1 1 92 BEKBBACFABAA 1 1 93 CAEABACFABAA 1 1 Scomber scombrus ID Haplotype S-ENG S-MAS Total 94 ZZZZZZZZZFZZ 17 7 24 95 ZZZZYZZZZFZZ 4 4 96 ZZZZYZZZZFZZ 2 2 97 CZYZZXZZZFZZ 1 1 98 XZZZZZZZZFYZ 1 1 99 ZXZZZZZZZFZZ 1 1 100 ZYZZYZZZZFZZ 1 1 101 ZZZZXZZZZFZZ 1 1 102 ZZZZZVYZZFZZ 1 1 103 ZZZZZWZZZFZZ 1 1 104 ZZZZZZZXZFZZ 1 1 105 ZZZZZZZYZFZZ 1 1 106 ZZZZZZZZZYZZ 1 1 restriction site data. Only one S. scombrus haplo- type (no. 94) was shared between the eastern and western North Atlantic samples, and it occurred at significantly different frequencies in each (0.35 in S- ENG, and 0.85 in S-MAS, Table 2). The distribution of haplotypes between the two samples was highly heterogeneous (P<0.001), although the estimate of net nucleotide sequence divergence between S-ENG and S-MAS was low (5=0.011%), reflecting the close relationship among haplotypes (Fig. 2). Scomber australasicus Restriction site analysis of S. australasicus mtDNA exhibited a range of in- traspecific divergences. Samples collected from Aus- tralia and New Zealand were very similar. Three haplotypes (nos. 69, 79, and 80), representing two genetically divergent matrilines, occurred at similar frequencies in each sample (Table 2), and no hetero- geneity was revealed (P=0.718, 5=-0.019%). The two samples of S. australasicus from the North Pacific (A-JPN and A-MEX) revealed greater divergence, of individuals in both samples (n=9 and n~ 12, re- comprising twelve haplotypes, of which two were spectively). Haplotype 82 occurred in five individu- shared, and one (no. 79) occurred in a large number als from A-MEX but was not present in the A-JPN Scoles et al.: Global phylogeography of Scomber 831 Table 3 Probabilities of significance from Roff and Bentzen’s (1989) chi- square analysis with 1000 randomizations of the data of samples of Atlantic Scomber japonicus. Pacific S. japonicus , S. australasicus, and S. scombrus that shared haplotypes. Comparison X2 Number of simulations exceeding yj Probability Atlantic Scomber japonicus J-ISR J-IVC J-SAF J-FLA J-ARG 191.36 0 <0.001** J-ISR J-IVC J-SAF 48.70 451 0.451ns J-ARG J-IVC J-SAF 60.84 28 0.028* J-ARG J-IVC 22.21 68 0.070NS J-ARG J-SAF 26.89 0 0.001** J-FLA J-ARG 28.37 1 0.001** J-FLA J-ISR J-IVC 58.40 16 0.016* J-FLA J-ISR 32.00 1 0.001*’ J-FLA J-IVC 24.33 14 0.014* Pacific Scomber japonicus J-JPN J-TWN 10.67 658 0.658NS Scomber australasicus A-JPN A-MEX A-AUS A-NZL 128.94 0 <0.001** A-JPN A-MEX 17.75 13 0.013* A-AUS A-NZL 11.67 718 0.718ns Scomber scombrus S-ENG S-MAS 20.17 0 <0.001** Significantly different at a=0.05. ’’Significantly heterogeneous at a=0.01. NSNot significant. Scomber japonicus Of the three species, samples of S. japonicus spanned the greatest geographical area, and mtDNA restriction site analysis of this species revealed the largest range of divergences. In the North Pacific, samples from Japan (J-JPN) and Taiwan (J- TWN) shared four haplotypes, three of which occurred in more than one individual in each sample (nos. 1-3, Table 2), and no significant heterogeneity was observed between the samples (P=0.658, a=0.05, 5=-0.010%). In con- trast, a comparison of the pooled data of J-JPN and J-TWN with S. japonicus from the eastern North Pacific ( J-CAL) revealed one fixed restric- tion site difference, a net nucleotide sequence divergence of 5=0.30%, and genotype distribu- tions were significantly different (PcO.001, cc=0.05). Many closely related haplotypes were ob- served in samples of S. japonicus from the At- lantic and Mediterranean Sea, and several were shared among two or more sample locations (Table 2). No significant differences were ob- served in the distribution of haplotypes among J-ISR, J-IVC, and J-SAF, from the eastern At- lantic (P=0.451, a=0.05, 5=-0.0Q3%-0.015%, Table 4), or between J-IVC and J-ARG across the Atlantic (P=0.07, a=0.05, 5=0.025%). All other tests of haplotype frequencies among At- lantic samples were significant (P< 0.05). sample. Although the net nucleotide sequence diver- gence between A-JPN and A-MEX was small (8=0.021%), the distribution of haplotypes between the samples was significantly heterogeneous (P=0.013, a=0.05). Together, the Australian and New Zealand samples of S. australasicus from the South Pacific were genetically distinct from the combined North Pacific samples (A-JPN and A-MEX). The mean corrected nucleotide sequence divergence between the pooled groups was high (8=0.54%), and only one haplotype was shared between the pooled South and North Pacific samples (no. 79), oc- curring at very different frequencies (0.05 and 0.51, respectively). The Red Sea sample of S. australasicus was dis- tinct from all other collections of this species. The mean net nucleotide sequence divergence between A-RED and the pooled data of all other S. aus- tralasicus was 8=0.86%. The Red Sea sample was less divergent from the S. australasicus samples from Australia and New Zealand (8=0.51%) than were the S. australasicus samples from Japan and Mexico (8=1.2%) Phylogeographic patterns MtDNA restriction site analysis and neighbor-join- ing of S. japonicus and S. australasicus mtDNA haplotypes revealed five major clusters: S. japonicus from the Pacific Ocean, S. japonicus from the Atlan- tic Ocean, S. australasicus from the Red Sea, a “ubiq- uitous” cluster of haplotypes of S. australasicus from all other samples of this species, and a “unique” clus- ter of only New Zealand and Australia S. austral- asicus haplotypes (Fig. 3). Parsimony analyses were used to explore cladistic relationships among the haplotypes of S. japonicus and S. australasicus identified by mtDNA restriction site analysis. Parsimony analysis among all Scomber haplotypes could not be conducted because S. scombrus mtDNA sequences were so divergent that it was not possible to determine homologous restric- tion site characters among the three species. Mainly interested in patterns occurring in S. japonicus, we chose a root in S. australasicus (haplotype no. 92) on the basis of results of parsimony analyses of cyto- chrome b sequences that included S', scombrus and a 832 Fishery Bulletin 96(4), 1 998 0 0 0 1 02 0 3 04 0 5 0 6 0 7 0 8 0 9 1.0 1.1 1.2 1.3 1 4 1 5 1.6 1 7 1 8 1 9 2 0 2.1 22 2 3 2 4 Percent nucleotide sequence divergence Figure 3 Neighbor-joining analysis of mtDNA haplotypes of Scomber japonicus and S. australasicus identified by restric- tion site analysis. Branch lengths are indicated. The terminal nodes labeled ATL are Atlantic haplotypes of S. japonicus 20-58 from J-FLA, J-ARG, J-SAF, J-IVC, and J-ISR. Haplotypes 1-13 labeled JPN or TWN are S. japonicus from J-JPN and J-TWN. Haplotypes 14-9 labeled CAL are S. japonicus from J-CAL. Haplotypes 61-68 labeled RED are S. australasicus haplotypes from A-RED. Those labeled UN are “unique” haplotypes 69-78 ofS. australasicus from A-NZL and A-AUS. Those labeled UB are “ubiquitous” haplotypes 79-93 of S. australasicus from A-NZL, A-AUS A-MEX, and A-JPN. Scoles et a I.: Global phylogeography of Scomber 833 more distant outgroup, Rastrelliger kanagurta, and a preliminary parsimony tree drawing. Analysis of 58 informative characters yielded 1031 equally par- simonious trees. Each of these 1031 trees shared the same topology among the five major matrilines and differed only within groups, with most rearrange- ments occurring among Atlantic S. japonicus haplotypes. Branches defining the five matrilines were supported at 100% in a majority-rule consen- sus that illustrated the relationship of all haplotypes of S. japonicus and S. australasicus and revealed a topology that was not different from that obtained by neighbor-joining (Fig. 4). The Red Sea sample had haplotypes intermediate to S. japonicus and S. australasicus, highlighted by character state reversals. Two reversals occurred among the Red Sea S. australasicus and the Japan- Taiwan S. japonicus clades, and the unique clade of S. australasicus. An Hpal site that is absent in eight of the 10 Red Sea haplotypes (nos. 60, 62-68), nine of the 10 unique S. australasicus haplotypes (nos. 69-73, 75-78), and all of the S. japonicus haplotypes from Japan and Taiwan, is present in all other haplotypes. A Puull site that is absent in six of the 10 Red Sea haplotypes (nos. 63-68) and all haplo- types of the unique S. australasicus clade is present in all other haplotypes. Consequently, corrected nucleotide sequence divergences among the Red Sea and unique S. australasicus haplotypes are lower than those among Red Sea and Atlantic S. japonicus haplotypes (Table 4). Although UPGMA cluster analysis grouped the Red Sea haplotypes with the unique 011 0.006 0.003 0 0 0 0 0.001 0 0 0 0 0 0 0 0 0 0 0 0.011 0 0.002 0 0.022 0 0.007 0 0 0.005 0.003 0 0.011 0 0 0 0.002 0 0 0 0 0 0.001 0.016 0.021 0.006 0.031 0 0 0.054 0.010 0.051 0.011 0.011 0.022 0 0.004 0.102 0.113 0.065 0.229 0.111 0.128 0.271 0.132 0.154 0.272 0.125 0.182 0.011 0.082 0.209 0.241 0.181 0.156 0.227 0.181 0.054 0.182 0.103 0.174 0.163 0.150 0.125 0.249 0.045 0.041 0.102 0 0.065 0.053 0.041 0.068 0.077 0.076 0.067 0.073 0.078 0.051 0.071 0.047 0.009 0.021 0.007 0.011 0.108 0.019 0.167 0.011 0.115 0.095 0.235 0.114 0.262 0.296 0.363 0.292 0.223 0.319 0.203 0.292 0.218 0.185 0.115 0.168 0.444 0.282 0.063 0.036 0.006 0.011 0.014 0.021 0.027 0.013 0.128 0.033 0.106 0.088 0.047 0.028 11 116 0.151 0.127 0.198 0.108 0.149 0.054 0.125 0.038 0.152 0.154 0.120 0.026 0.106 1013 0.011 0 0 0 0.021 0 0.002 0.051 0.011 0 0.018 0.006 0.018 0-003 0.003 0 0 0 0 0 0 0 0 0 0 0.029 0.031 0 0 0.051 0.041 0 0 0.014 0.037 0.011 0.058 0.032 0.026 0 0 0 0 0.071 0.068 0 0 0.095 0.064 0 0.038 0.064 0.015 0 0 "•757 0.735 0.714 0.792 0.763 0.723 *0.703 0.728 0.744 0.696 0.788 0.729 0.657 0.772 0.815 0.801 0.812 0.849 0.796 0.865 0.826 0.850 0.812 0.885 0.864 0.722 0.822 848 Fishery Bulletin 96(4), 1998 Table Uncorrected and corrected (bold type) allele frequencies, observed heterozygosity (Hobs) and expected heterozygosity (Hexp) at Oisl03 loc1 frequencies of the populations in the region, with region names abbreviated as in Table 1. Populations out of Hardy-Weinberg equilibria cies for individual populations in the lower Fraser River and the Thompson River are given in Table 3 of Small et al. (1998). Sample si: , Alleles Pallant 93 Atnarko 87 Kitimat 148 C Coast 235 Cluxewe 38 Stephens 65 Waukwaas 30 NC VI 133 Nitinat 39 Robertson 84 WCVI 123 Quins; , 160 61 0 0 0.007 0.004 0 0.037 0 0.018 0.105 0 0.031 0 0 0.006 0.004 0 0.037 0 0.018 0.080 0 0.024 71 0 0 0 0 0.013 0.022 0 0.014 0 0 0 0 0 0 0 0.013 0.022 0 0.014 0 0 0 75 0.218 0.039 0.081 0.064 0.184 0.265 0.156 0.217 0 0.018 0.012 0.14 0.223 0.041 0.077 0.064 0.162 0.257 0.152 0.205 0 0.017 0.012 0.13 79 0.048 0.006 0.017 0.012 0 0 0.016 0.004 0.053 0.185 0.138 0.03' 0.048 0.006 0.016 0.012 0 0 0.016 0.004 0.050 0.161 0.124 0.03 83 0.021 0 0 0 0.026 0 0 0.007 0 0 0 0.021 0 0 0 0.026 0 0 0.007 0 0 0 86 0.005 0 0.003 0.002 0.013 0.088 0.047 0.058 0.026 0.012 0.016 0.005 0 0.003 0.002 0.013 0.089 0.033 0.053 0.025 0.011 0.016 ■j 90 0 0.006 0.003 0.004 0 0.007 0.016 0.007 0 0.030 0.020 0.03 0 0.006 0.003 0.004 0 0.007 0.016 0.007 0 0.023 0.016 0.03 94 0 0.079 0.013 0.037 0.013 0 0 0.004 0.053 0 0.016 0.05 0 0.075 0.013 0.035 0.013 0 0 0.004 0.039 0 0.012 0.05 98 0.005 0.067 0.010 0.031 0.026 0.059 0.031 0.043 0 0.024 0.016 0.005 0.064 0.010 0.029 0.014 0.059 0.017 0.038 0 0.023 0.016 102 0.117 0.028 0.037 0.033 0.105 0 0 0.029 0.039 0 0.012 0.01 0.117 0.028 0.033 0.031 0.096 0 0 0.026 0.038 0 0.012 0.01 106 0.165 0.017 0.047 0.035 0.026 0.051 0.016 0.036 0.276 0.042 0.110 0.00 0.163 0.017 0.043 0.033 0.026 0.051 0.016 0.036 0.224 0.040 0.095 0.00 110 0.069 0.073 0,044 0.053 0 0.015 0.047 0.018 0.039 0.054 0.047 0.031 0.063 0.073 0.037 0.050 0 0.015 0.033 0.015 0.026 0.037 0.033 0.02 114 0.027 0.051 0.034 0.037 0.013 0.044 0.125 0.054 0.026 0.060 0.043 0.00 0.027 0.042 0.029 0.034 0.013 0.039 0.125 0.052 0.025 0.047 0.040 0.00 117 0.005 0.006 0.060 0.039 0 0.015 0.156 0.043 0 0 0 0.005 0.006 0.056 0.037 0 0.015 0.124 0.034 0 0 0 121 0.011 0 0.013 0.008 0.079 0.007 0.047 0.036 0.053 0.048 0.047 0.011 0 0.013 0.008 0.073 0.007 0.033 0.030 0.039 0.046 0.044 125 0.005 0.006 0.007 0.006 0.092 0.037 0.219 0.094 0.013 0.036 0.028 0.01 0.005 0.006 0.006 0.006 0.092 0.037 0.191 0.087 0.013 0.034 0.028 0,01 129 0.011 0 0.020 0.012 0.066 0 0 0.018 0 0 0 0.02 0.011 0 0.017 0.010 0.054 0 0 0.015 0 0 0 0.02 133 0.053 0.056 0.007 0.025 0 0 0 0 0 0 0 0.02 0.053 0.052 0.006 0.023 0 0 0 0 0 0 0 0.02 136 0.011 0.011 0.003 0.006 0 0 0 0 0.039 0.006 0.016 0.04 0.011 0.011 0.003 0.006 0 0 0 0 0.038 0.006 0.016 0.03 140 0 0 0.003 0.002 0 0.022 0 0.011 0.066 0.006 0.024 0.05 0 0 0.003 0.002 0 0.022 0 0.011 0.052 0.006 0.020 0.05 144 0 0.017 0.010 0.012 0 0.088 0.016 0.047 0.079 0 0.024 0.02 0 0.017 0.007 0.010 0 0.088 0.016 0.047 0.064 0 0.020 0.02 148 0 0.034 0.010 0.019 0.039 0.066 0 0.043 0 0 0 0.01 0 0.034 0.010 0.019 0.039 0.066 0 0.043 0 0 0 0.01 152 0 0.051 0.010 0.025 0.053 0 0 0.014 0 0.024 0.016 o.oc 0 0.039 0.010 0.019 0.053 0 0 0.014 0 0.023 0.016 0.00 156 0 0.062 0.040 0.047 0.053 0.029 0.016 0.033 0 0 0 0.02. 0 0.062 0.039 0.047 0.042 0.029 0.016 0.030 0 0 0 0.02 160 0.021 0.079 0.084 0.080 0 0.015 0 0.007 0.013 0 0.004 0.04 0.021 0.079 0.071 0.075 0 0.015 0 0.007 0.013 0 0.004 0.03 165 0.021 0.022 0.101 0.070 0.105 0.029 0.016 0.047 0 0.006 0.004 0.15 0.021 0.022 0.086 0.063 0.087 0.024 0.016 0.038 0 0.006 0.004 0.14 170 0.037 0.017 0.104 0.070 0.013 0.051 0.031 0.036 0.026 0.018 0.020 0.05 0.037 0.017 0.096 0.067 0.013 0.047 0.031 0.033 0.025 0.017 0.020 0.05 175 0.106 0.028 0.084 0.062 0.013 0 0 0.004 0 0.036 0.024 0.014 0.105 0.023 0.071 0.054 0.013 0 0 0.004 0 0.029 0.020 0.01 180 0.027 0.067 0.027 0.041 0.013 0 0 0.004 0.013 0 0.004 0.02 0.027 0.060 0.026 0.038 0.013 0 0 0.004 0.013 0 0.004 0.02 185 0 0.028 0.020 0.023 0 0 0.016 0.004 0 0.179 0.118 0.021 0 0.028 0.019 0.023 0 0 0.016 0.004 0 0.147 0.097 0.02 190 0 0.039 0.027 0.031 0.013 0.015 0 0.011 0.026 0.208 0.146 0.0^ 0 0.039 0.026 0.031 0.013 0.015 0 0.011 0.025 0.185 0.132 0.04 200 0.005 0.028 0.027 0.027 0.026 0.015 0 0.014 0.013 0.006 0.008 0.0( 0.005 0.018 0.026 0.023 0.026 0.016 0 0.015 0.013 0.006 0.008 0.05 210 0.005 0.062 0.030 0.041 0.013 0.007 0.016 0.011 0 0 0 0.0( 0.005 0.058 0.029 0.040 0.013 0.007 0.016 0.011 0 0 0 0.0< 220 0 0.011 0.010 0.010 0 0.015 0.016 0.011 0 0.006 0.004 0.0< 0 0.011 0.010 0.010 0 0.011 0.016 0.084 0 0.006 0.004 0.0< 230 0 0.011 0.007 0.008 0 0 0 0 0.039 0 0.012 0 0.005 0.006 0.008 0 0 0 0 0.026 0 0.008 250 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 270 0.011 0 0.003 0.002 0 0 0 0 0 0.006 0.004 0.0( 0.006 0 0.003 0.002 0 0 0 0 0 0.006 0.004 0.01 Nul! 0 0 0.032 0.021 0 0 0 0 0.050 0.034 0.039 0.0< 0.006 0.069 0.092 0.079 0.104 0.035 0.136 0.154 0.176 0.124 0.154 Hobs 0.856 0.764 *0.779 0.849 *0.697 0.831 *0.625 0.885 *0.632 *0.714 0.705 *0.7' Hexp 0.777 0.791 0.948 0.979 0.921 0.897 0.906 0.920 0.899 0.885 0.959 0.9: Small et al. : Population and stock indentification of Oncorhynchus kisutch 849 coho salmon populations from British Columbia. The weighted means of allele frequencies for regions are in bold and follow the indicated by * next to the observed heterozygosity. The allele number (in basepairs) is the lower limit of the allele bin. Allele frequen- given underneath the population names. Big alicum 200 EC VI Toboggan 260 164 Cedar 47 Babine 139 Clear- water 48 Sustut 36 Skeena 434 Tseax 40 Zolzap Mexiadin 47 64 Nass 151 TR 1015 L Fr 1043 0 0 0.079 0.020 0.072 0.024 0 0.056 0.038 0.063 0.019 0.039 0.002 0.001 0 0 0.076 0.019 0.065 0.010 0 0.052 0.038 0.055 0.015 0.037 0.001 0.001 0 0 0 0 0.004 0.024 0 0.003 0 0 0 0 0.024 0.027 0 0 0 0 0.004 0.010 0 0.002 0 0 0 0 0.016 0.025 .136 0.136 0.201 0.265 0.040 0.146 0.167 0.144 0.075 0.115 0.144 0.114 0.016 0.086 .114 0.125 0.176 0.208 0.037 0.076 0.132 0.121 0.075 0.103 0.093 0.105 0.011 0.082 .019 0.026 0.028 0.020 0.036 0.012 0 0.025 0.038 0.031 0.010 0.025 0.002 0.010 .017 0.025 0.027 0.009 0.036 0.010 0 0.024 0.038 0.031 0.007 0.025 0.001 0.009 0 0 0.003 0 0.047 0.037 0 0.019 0 0 0 0 0.001 0.023 0 0 0.003 0 0.046 0.021 0 0.018 0 0 0 0 0.001 0.022 .045 0.024 0.097 0.029 0.086 0 0.014 0.067 0.038 0.021 0.067 0.043 0.003 0.018 .038 0.021 0.091 0.028 0.080 0 0.014 0.063 0.038 0.021 0.045 0.040 0.002 0.017 .019 0.025 0.022 0 0.036 0.037 0.014 0.024 0.025 0.042 0.067 0.046 0.001 0.056 015 0.022 0.021 0 0.036 0.021 0.014 0.023 0.014 0.042 0.045 0.043 0.001 0.053 043 0.046 0 0 0 0.024 0 0.002 0 0 0 0 0.029 0.011 .040 0.045 0 0 0 0.010 0 0.001 0 0 0 0 0.020 0.010 029 0.015 0 0 0.014 0 0 0.005 0.038 0 0 0.011 0.003 0.014 .027 0.015 0 0 0.014 0 0 0.005 0.038 0 0 0.011 0.003 0.014 0 0.007 0.003 0.049 0.011 0.012 0 0.011 0.025 0.021 0 0.014 0.002 0.038 0 0.007 0.003 0.038 0.011 0.010 0 0.010 0.025 0.021 0 0.014 0.001 0.036 .005 0.004 0.013 0.010 0.119 0.085 0.069 0.057 0.050 0.010 0.029 0.029 0.049 0.034 .005 0.004 0.012 0.009 0.111 0.042 0.056 0.049 0.050 0.010 0.022 0.029 0.034 0.033 .008 0.019 0.126 0.010 0.194 0.061 0.083 0.121 0.025 0.073 0.029 0.043 0.302 0.031 .007 0.017 0.109 0.009 0.181 0.051 0.070 0.107 0.014 0.066 0.022 0.037 0.209 0.028 .019 0.014 0.006 0.069 0.004 0.110 0.056 0.026 0.013 0.021 0.038 0.025 0.123 0.036 017 0.014 0.006 0.057 0.004 0.063 0.029 0.020 0.014 0.021 0.030 0.025 0.080 0.035 Oil 0.006 0.053 0.049 0 0 0 0.025 0.013 0.021 0.019 0.018 0.025 0.069 010 0.006 0.047 0.038 0 0 0 0.022 0.013 0.021 0.015 0.018 0.018 0.065 .005 0.003 0.006 0 0 0.024 0 0.005 0.038 0.031 0.019 0.029 0.012 0.019 005 0.003 0.006 0 0 0.010 0 0.003 0.038 0.031 0.015 0.029 0.009 0.018 021 0.015 0 0 0 0.037 0 0.003 0.038 0.031 0 0.021 0.035 0.033 017 0.014 0 0 0 0.031 0 0.003 0.038 0.031 0 0.021 0.027 0.033 021 0.021 0 0 0.004 0 0 0.001 0 0 0 0 0.008 0.019 015 0.018 0 0 0.004 0 0 0.001 0 0 0 0 0.005 0.018 003 0.011 0 0 0 0 0 0 0 0 0 0 0 0.006 002 0.011 0 0 0 0 0 0 0 0 0 0 0 0.006 045 0.042 0 0.010 0.011 0 0 0.005 0 0.010 0.010 0.007 0.002 0.037 036 0.037 0 0.009 0.011 0 0 0.005 0 0.010 0.007 0.007 0.002 0.036 024 0.038 0.003 0.010 0.004 0.037 0.014 0.008 0.038 0.063 0.087 0.064 0.001 0.059 022 0.036 0.003 0.009 0.004 0.031 0.014 0.008 0.038 0.066 0.060 0.062 0.001 0.056 016 0.018 0.016 0.010 0.011 0 0 0.010 0 0.042 0.029 0.025 0.006 0.058 013 0.017 0.015 0.009 0.007 0 0 0.009 0 0.042 0.022 0.025 0.003 0.056 008 0.013 0.006 0.020 0.004 0.049 0 0.010 0.038 0.010 0.019 0.021 0.019 0.039 007 0.013 0.006 0.019 0.004 0.021 0 0.008 0.038 0.011 0.008 0.018 0.014 0.037 051 0.031 0.022 0 0.018 0.049 0.042 0.022 0.075 0.021 0.029 0.039 0.017 0.038 041 0.027 0.021 0 0.018 0.031 0.028 0.020 0.067 0.022 0.022 0.037 0.015 0.037 106 0.067 0.035 0.059 0.061 0.085 0.153 0.059 0.038 0.083 0.058 0.061 0.004 0.036 1088 0.060 0.031 0.056 0.058 0.042 0.117 0.052 0.038 0.077 0.038 0.056 0.011 0.036 061 0.050 0.041 0.010 0.036 0.037 0 0.031 0,100 0.031 0.077 0.068 0.008 0.026 1044 0.041 0.037 0.009 0.036 0.031 0 0.030 0.093 0.031 0.046 0.061 0.003 0.025 056 0.100 0.009 0.029 0.004 0 0.014 0.009 0.063 0.021 0.029 0.036 0.020 0.035 044 0.085 0.009 0.028 0.004 0 0.014 0.009 0.063 0.021 0.022 0.036 0.006 0.033 053 0.053 0.003 0.029 0.025 0.073 0.042 0.023 0.025 0.021 0.019 0.021 0.065 0.036 045 0.051 0.003 0.028 0.025 0.042 0.041 0.021 0.025 0.021 0.008 0.018 0.045 0.035 (056 0.035 0.025 0.088 0.004 0.024 0.069 0.029 0.038 0.042 0.010 0.029 0.008 0.040 050 0.034 0.024 0.076 0.004 0.020 0.043 0.025 0.038 0.042 0.007 0.029 0.005 0.040 043 0.032 0.031 0.029 0.018 0.012 0.056 0.026 0.038 0.021 0.048 0.036 0.007 0.019 035 0.029 0.030 0.028 0.015 0.010 0.042 0.024 0.038 0.021 0.037 0.036 0.005 0.019 1005 0.014 0.035 0.010 0.029 0 0.028 0.025 0 0.010 0.058 0.025 0 0.019 (005 0.013 0.034 0.009 0.029 0 0.027 0.025 0 0.010 0.045 0.025 0 0.018 061 0.051 0.085 0 0.014 0 0 0.035 0.087 0.073 0.019 0.057 0 0.005 1050 0.048 0.078 0 0.014 0 0 0.033 0.071 0.066 0.015 0.049 0 0.005 1027 0.042 0.022 0.039 0.018 0 0 0.018 0 0.042 0,010 0.018 0.004 0.004 1023 0.038 0.021 0.037 0.018 0 0 0.018 0 0.042 0.007 0.018 0.002 0.004 003 0.003 0.025 0.010 0.05 0 0.111 0.035 0.013 0.031 0.038 0.029 0.007 0.005 1002 0.003 0.024 0.009 0.048 0 0.097 0.033 0.013 0.031 0.030 0.029 0.004 0.005 1 0 0.001 0.003 0.010 0.029 0 0.069 0.017 0 0 0.019 0.007 0 0.002 0 0.001 0.003 0.009 0.029 0 0.068 0.017 0 0 0.015 0.005 0 0.002 0 0 0 0 0 0 0 0 0 0 0 0 0 0.001 0 0.003 0 0 0 0 0 0 0 0 0 0 0 0.001 005 0.003 0 0.118 0 0 0 0 0 0 0 0 0 0 005 0.004 0 0.064 0 0 0 0 0 0 0 0 0 0 005 0.004 0 0 0 0 0 0 0 0 0 0 0 0.001 1005 0 0 0 0 0 0 0 0 0 0 0 0 0.001 060 0.033 0.030 0.007 0 0.163 0.027 0.041 0 0 0 0 0.193 0.007 ,|l 53 0.115 0.080 0.186 0.052 0.406 0.197 0.137 0.066 0.065 0.300 0.063 0.444 0.048 1750 0.764 0.774 *0.588 *0.737 *0.390 *0.431 0.721 0.837 0.802 *0.721 0.836 0.254 0.875 (1945 0.944 0.915 0.896 0.921 0.938 0.919 0.938 0.950 0.958 0.942 0.957 0.086 0.958 848 Fishery Bulletin 96(4), 1998 Small et al. Population and stock indentification of Oncorhynchus kisutch 849 Table 2 pected heterozygosity (Hexp) at Ots 103 loCUi for coho salmon populations from British Columbia The we ghted means of allele frequencies for regions are in bold and follow the frequencies of the populations cies for individual populations in the region, with region names abbreviated as in Table 1. Populations out ot Hardy-Weinberg equilibria in the lower Fraser River and the Thompson River are given in Table 3 of Small et al. < 1998). Sample siZe5 jre indicated by next to the observed heterozygosity. The allele number (in basepairs) is the lower limit of the allele bin. Allele frequen- ce given underneath the population names. Big Clear- C Coast Cluxewe Stephens Waukwaas NCVI Nitinat Robertson WCVI Qninsam Qualicum ECVI Toboggan Cedar Babine water Sustut Skeena Tseax Zolzap Mexiadin Nass TR L Fr Alleles 93 87 148 235 38 65 30 133 39 84 123 160 200 260 164 47 139 48 36 434 40 47 64 151 1015 1043 61 0 0 0.007 0.004 0 0 0.037 0.037 0 0 0.018 0.018 0.105 0.080 0 0 0.031 0.024 0 0 0 0 0 0 0.079 0.076 0.020 0.019 0.072 0.065 0.024 0.010 0 0 0.056 0.052 0.038 0.038 0.063 0.055 0.019 0.015 0.039 0.037 0.002 0.001 0.001 0.001 0.022 0 0 0 0 0.003 0 0 0 n 0 0.014 0 0 ft 0 0.010 0 0.002 0 0 75 ft 91B 0.265 0.156 0.217 0 0.018 0.147 0.136 0.136 0.265 0.040 0.146 0.167 0.144 0.075 0.115 0.144 0.114 0.016 0.086 n 0.257 0.152 0.205 0 0.139 0.114 0.125 0.132 0.121 0.075 0.103 0.093 79 n't\AO 0 0.016 0.004 0.053 0.185 0.038 0.019 0.036 0.012 0 0.025 0.038 0.031 0.010 0 048 0 0 0.016 0.004 0.050 0.035 0.017 0.036 0.010 0 0.024 0.038 0.031 0.007 0.025 0.001 0.009 83 ft ft91 0 0 0.007 0 0 0.037 0 0.019 0 0 0 0 0.001 0.023 0 0 0.007 0 0 0 0.021 0 0.0 1H 0 0 0 86 0.013 0.088 0.047 0.058 0.026 0.012 0 0.045 0.024 0.029 0.086 0 0.014 0.067 0.038 0.021 0.067 0.043 0.003 0.018 0.013 0.089 0.033 0.053 0.025 0.016 0 0.038 0 0.014 0.063 0.038 0.021 0.045 0.040 0.002 0.017 90 0 0.007 0.016 0.007 0 0.034 0.019 0.022 0.036 0.037 0.014 0.024 0.025 0.042 0.067 0.046 0.001 0 0.007 0.016 0.007 0 0.033 0.015 0.021 0.014 0.023 0.014 0.042 0.045 0.043 0.001 0.053 94 0.013 0 0 0.004 0.053 0 0.053 0.043 0.046 0 0 0.024 0 0.002 0 0 0 0 0.029 0.011 0 0 0.004 0.039 0.051 0.040 0 0.001 0 0 0 0 98 0.026 0.059 0.031 0.043 0 0.024 0.016 0 0.029 0.015 0 0 0.014 0 0 0.005 0.038 0 0 0.011 0.003 0.014 0.014 0.059 0.017 0.038 0 0.023 0 0.027 0.015 0.014 0 0 0.005 0.038 0 0 0.011 0.003 0.014 102 0.105 0 0 0.029 0.039 0 0.016 0 0.007 0.003 0.049 0.011 0.012 0 0.011 0.025 0.021 0 0.014 0.002 0.038 0.096 0 0 0.026 0.038 0 0.012 0.016 0 0.007 0.003 0.038 0.011 0.010 0 0.010 0.025 0.021 0 0.014 0.001 0.036 106 0.026 0.051 0.016 0.036 0.276 0.042 0.110 0.003 0.005 0.004 0.013 0.010 0.119 0.085 0.069 0.057 0.050 0.010 0.029 0.029 0.049 0.034 0.026 0.051 0.016 0.036 0.224 0.040 0.095 0.003 0.005 0.004 0.012 0.009 0.111 0.042 0.056 0.049 0.050 0.010 0.022 0.029 0.034 0.033 110 0 0.015 0.047 0.018 0.039 0.054 0.047 0.034 0.008 0.019 0.126 0.010 0.194 0.061 0.083 0.121 0.025 0.073 0.029 0.043 0.302 0.031 0 0.015 0.033 0.015 0.026 0.037 0.033 0.029 0.007 0.017 0.109 0.009 0.181 0.051 0.070 0.107 0.014 0.066 0.022 0.037 0.209 0.028 114 0.013 0.044 0.125 0.054 0.026 0.060 0.043 0.009 0.019 0.014 0.006 0.069 0.004 0.110 0.056 0.026 0.013 0.021 0.038 0.025 0.123 0.036 0.034 0.013 0.039 0.125 0.052 0.025 0.047 0.040 0.009 0.017 0.014 0.006 0.057 0.004 0.063 0.029 0.020 0.014 0.021 0.030 0.025 0.080 0.035 117 0.006 0.060 0.039 0 0.015 0.156 0.043 0 0 0 0 0.011 0.006 0.053 0.049 0 0 0 0.025 0.013 0.021 0.019 0.018 0.025 0.069 0.056 0.037 0 0.015 0.124 0.034 0 0 0 0 0.01ft 0.006 0.047 0.038 0 0 0 0.022 0.013 0.021 0.015 0.018 0.018 0.065 121 0 0.013 0.008 0.079 0.007 0.047 0.036 0.053 0.048 0.047 0 0.005 0.003 0.006 0 0 0.024 0 0.005 0.038 0.031 0.019 0.029 0.012 0.019 0 0.013 0.008 0.073 0.007 0.033 0.030 0.039 0.046 0.044 0 0.005 0.003 0.006 0 0 0.010 0 0.003 0.038 0.031 0.015 0.029 0.009 0.018 125 0.007 0.006 0.092 0.037 0.219 0.094 0.013 0.036 0.028 0.013 0.021 0.015 0 0 0 0.037 0 0.003 0.038 0.031 0 0.021 0.035 0.033 0.006 0.006 0.092 0.037 0.191 0.087 0.013 0.034 0.028 0.010 0.017 0.014 0 0 0 0.031 0 0.003 0.038 0.031 0 0.021 0.027 0.033 129 0.011 0 0.020 0.012 0.066 0 0 0.018 0 0 0 0.025 0.021 0.021 0 0 0.004 0 0 0.001 0 0 0 0 0.008 0.019 0 0.017 0.010 0.054 0 0 0.015 0 0 0 0.022 0.015 0.018 0 0 0.004 0 0 0.001 0 0 0 0 0.005 0.018 133 0.053 0.056 0.007 0.025 0 0 0 0 0 0 0 0.022 0.003 0.011 0 0 0 0 0 0 0 0 0 0 0 0.006 0.052 0.006 0.023 0 0 0 0 0 0 0 0.022 0.002 0.011 0 0 0 0 0 0 0 0 0 0 0 0.006 136 0.011 0.003 0.006 0 0 0 0 0.039 0.006 0.016 0.041 0.045 0.042 0 0.010 0.011 0 0 0.005 0 0.010 0.010 0.007 0.002 0.037 0.011 0.003 0.006 0 0 0 0 0.038 0.006 0.016 0.038 0.036 0.037 0 0.009 0.011 0 0 0.005 0 0.010 0.007 0.007 0.002 0.036 140 0 0 0.003 0.002 0 0.022 0 0.011 0.066 0.006 0.024 0.056 0.024 0.038 0.003 0.010 0.004 0.037 0.014 0.008 0.038 0.063 0.087 0.064 0.001 0.059 0 0 0.003 0.002 0 0.022 0 0.011 0.052 0.006 0.020 0.054 0.022 0.036 0.003 0.009 0.004 0.031 0.014 0.008 0.038 0.066 0.060 0.062 0.001 0.056 144 © 0.017 0.010 0.012 0 0.088 0.016 0.047 0.079 0 0.024 0.022 0.016 0.018 0.016 0.010 0.011 0 0 0.010 0 0.042 0.029 0.025 0.006 0.058 0 0.017 0.007 0.010 0 0.088 0.016 0.047 0.064 0 0.020 0.022 0.013 0.017 0.015 0.009 0.007 0 0 0.009 0 0.042 0.022 0.025 0.003 0.056 148 0 0.034 0.010 0.019 0.039 0.066 0 0.043 0 0 0 0.019 O.OoS 0.013 0.006 0.020 0.004 0.049 0 0.010 0.038 0.010 0.019 0.021 0.019 0.039 0 0.034 0.010 0.019 0.039 0.066 0 0.043 0 0 0 0.019 0.0(17 0.013 0.006 0.019 0.004 0.021 0 0.008 0.038 0.011 0.008 0.018 0.014 0.037 152 0 0.051 0.010 0.025 0.053 0 0 0.014 0 0.024 0.016 0.009 0.05 1 0.031 0.022 0 0.018 0.049 0.042 0.022 0.075 0.021 0.029 0.039 0.017 0.038 0 0.039 0.010 0.019 0.053 0 0 0.014 0 0.023 0.016 0.009 0.04 1 0.027 0.021 0 0.018 0.031 0.028 0.020 0.067 0.022 0.022 0.037 0.015 0.037 156 0 0.062 0.040 0.047 0.053 0.029 0.016 0.033 0 0 0 0.025 0.106 0.067 0.035 0.059 0.061 0.085 0.153 0.059 0.038 0.083 0.058 0.061 0.004 0.036 0 0.062 0.039 0.047 0.042 0.029 0.016 0.030 0 0 0 0.025 0.038 0.060 0.031 0.056 0.058 0.042 0.117 0.052 0.038 0.077 0.038 0.056 0.011 0.036 160 0.021 0.079 0.084 0.080 0 0.015 0 0.007 0.013 0 0.004 0.041 0.061 0.050 0.041 0.010 0.036 0.037 0 0.031 0.100 0.031 0.077 0.068 0.008 0.026 0.021 0.079 0.071 0.075 0 0.015 0 0.007 0.013 0 0.004 0.036 0.044 0.041 0.037 0.009 0.036 0.031 0 0.030 0.093 0.031 0.046 0.061 0.003 0.025 165 0.021 0.022 0.101 0.070 0.105 0.029 0.016 0.047 0 0.006 0.004 0.159 0.056 0.100 0.009 0.029 0.004 0 0.014 0.009 0.063 0.021 0.029 0.036 0.020 0.035 0.021 0.022 0.086 0.063 0.087 0.024 0.016 0.038 0 0.006 0.004 0.143 0.044 0.085 0.009 0.028 0.004 0 0.014 0.009 0.063 0.021 0.022 0.036 0.006 0.033 170 0.037 0.017 0.104 0.070 0.013 0.051 0.031 0.036 0.026 0.018 0.020 0.056 0.05 -i 0.053 0.003 0.029 0.025 0.073 0.042 0.023 0.025 0.021 0.019 0.021 0.065 0.036 0.037 0.017 0.096 0.067 0.013 0.047 0.031 0.033 0.025 0.017 0.020 0.045 0.051 0.003 0.028 0.025 0.042 0.041 0.021 0.025 0.021 0.008 0.018 0.045 0.035 175 0.106 0.028 0.084 0.062 0.013 0 0 0.004 0 0.036 0.024 0.013 0.056 0.035 0.025 0.088 0.004 0.024 0.069 0.029 0.038 0.042 0.010 0.029 0.008 0.040 0.105 0.023 0.071 0.054 0.013 0 0 0.004 0 0.029 0.020 0.050 0.034 0.024 0.076 0.004 0.020 0.043 0.025 0.038 0.042 0.007 0.029 0.005 0.040 180 0.027 0.067 0.027 0.041 0.013 0 0 0.004 0.013 0 0.004 0.043 0.032 0.031 0.029 0.018 0.012 0.056 0.026 0.038 0.021 0.048 0.036 0.007 0.019 0.027 0.060 0.026 0.038 0.013 0 0 0.004 0.013 0 0.004 0.035 0.029 0.030 0.028 0.015 0.010 0.042 0.024 0.038 0.021 0.037 0.036 0.005 0.019 185 0 0.028 0.020 0.023 0 0 0.016 0.004 0 0.179 0.118 0.005 0.014 0.035 0.010 0.029 0 0.028 0.025 0 0.010 0.058 0.025 0 0.019 0 0.028 0.019 0.023 0 0 0.016 0.004 0 0.147 0.097 0.005 0.013 0.034 0.009 0.029 0 0.027 0.025 0 0.010 0.045 0.025 0 0.018 190 0 0.039 0.027 0.031 0.013 0.015 0 0.011 0.026 0.208 0.146 0.047 0.061 0.051 0.085 0 0.014 0 0 0.035 0.087 0.073 0.019 0.057 0 0.005 0 0.039 0.026 0.031 0.013 0.015 0 0.011 0.025 0.185 0.132 0.045 0.050 0.048 0.078 0 0.014 0 0 0.033 0.071 0.066 0.015 0.049 0 0.005 200 0.005 0.028 0.027 0.027 0.026 0.015 0 0.014 0.013 0.006 0.008 0.063 0.027 0.042 0.022 0.039 0.018 0 0 0.018 0 0.042 0.010 0.018 0.004 0.004 0.005 0.018 0.026 0.023 0.026 0.016 0 0.015 0.013 0.006 0.008 0.058 0.023 0.038 0.021 0.037 0.018 0 0 0.018 0 0.042 0.007 0.018 0.002 0.004 210 0.005 0.062 0.030 0.041 0.013 0.007 0.016 0.011 0 0 0 0.003 0.003 0.025 0.010 0.05 0 0.111 0.035 0.013 0.031 0.038 0.029 0.007 0.005 0.005 0.058 0.029 0.040 0.013 0.007 0.016 0.011 0 0 0 0.002 0.003 0.024 0.009 0.048 0 0.097 0.033 0.013 0.031 0.030 0.029 0.004 0.005 220 0 0.011 0.010 0.010 0 0.015 0.016 0.011 0 0.006 0.004 0.003 0 0.001 0.003 0.010 0.029 0 0.069 0.017 0 0 0.019 0.007 0 0.002 0 0.011 0.010 0.010 0 0.011 0.016 0.084 0 0.006 0.004 0 0.001 0.003 0.009 0.029 0 0.068 0.017 0 0 0.015 0.005 0 0.002 230 0 0.011 0.007 0.008 0 0 0 0 0.039 0 0.012 0 0 0 0 0 0 0 0 0 0 0 0 0 0.001 0 0.005 0.006 0.008 0 0 0 0 0.026 0 0.008 0 0.003 0 0 0 0 0 0 0 0 0 0 0 0.001 250 0 0 0 0 0 0 0 0 0 0 0 0.005 0.003 0 0.118 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.003 0.005 0 0.064 0 0 0 0 0 0 0 0 0 0 270 0.011 0 0.003 0.002 0 0 0 0 0.004 0.005 0 0 0 0 0 0 0 0 0 0 0 0.001 0.006 0 0.003 0.002 0 0 0 0 0 0.006 0.004 0.001 0.005 0 0 0 0 0 0 0 0 0 0 0 0 0.001 Null 0 0 0.032 0.021 0 0 0 0 0.050 0.034 0.039 0.060 0.033 0.030 0.007 0 0.163 0.027 0.041 0 0 0 0 0.193 0.007 0.006 0.069 0.092 0.079 0.104 0.035 0.136 0.154 0.176 0.154 0.062 0.153 0.115 0.080 0.186 0.052 0.406 0.197 0.137 0.066 0.065 0.300 0.063 0.444 0.048 Hobs 0.856 0.764 *0.779 0.849 *0.697 0.831 *0.625 0.885 *0.632 0.705 0.750 0.764 0.774 *0.588 0.737 0.390 *0.431 0.721 0.837 0.802 *0.721 0.836 0.254 0.875 Hexp 0.777 0.791 0.948 0.979 0.921 0.897 0.906 0.920 0.899 0.885 0.959 0.931 0.945 0.944 0.915 0.896 0.921 0.938 0.919 0.938 0.950 0.958 0.942 0.957 0.086 0.958 850 Fishery Bulletin 96(4), 1998 inheritance at all three loci, although a null (nonamp- lifying) allele is present at O/sl03 (Small et ah, 1998). Individual samples that although amplified with both other primer sets yet produced no Ots 103 alleles af- ter the Ots 103 amplification was performed three times, were scored as Ots 103 null homozygotes. Al- lele frequencies at the Ots 103 locus were then cor- rected according to a maximum-likelihood estimate (EM algorithm, Dempster et ah, 1977) by using GENEPOP, version 1.2 (Raymond and Rousset, 1995a). Corrected frequencies were used in the ge- netic distance analysis. For OtslOl and OtsS, fish with a single allele were scored as homozygotes and fish with two alleles were scored as heterozygotes. Allele fre- quency data for individual populations in the Fraser River and Thompson River are from Small et al. ( 1998) and only the weighted means of total allele frequen- cies for these regions are presented here ( Tables 1 and 2). The upper Fraser River population (Bridge River) has been grouped with the Thompson River popula- tions in accordance with its genetic profile. Hardy-Weinberg equilibrium (HWE) was tested with the probability test of HWE (an exact HW test, Guo and Thompson, 1992) with GENEPOP (version 1.2). F-statistics and their standard deviations were calculated according to Weir and Cockerham (1984) using FSTAT (Goudet, 1995). We use the notations Fst, F s and Fj( for Weir and Cockerham’s 9, f, and F, respectively. P-values for a = 0.05 in all data analy- ses were corrected for simultaneous multiple tests (Lessios, 1992). Pairwise comparisons of populations for differences in allele distributions were conducted at all three loci with Fisher’s exact tests (Raymond and Rousset, 1995b). In populations with multiple year classes, annual variability in allele frequencies was tested with 1000 simulations in a Monte Carlo analysis (Roff and Bentzen, 1989) and populations were divided into year classes for a neighbor-joining (NJ) analysis (Saitou and Nei, 1987). A NJ dendro- gram illustrating genetic relationships among popu- lations was constructed with PHYLIP 3.5c software (Felsenstein, 1993). The allele frequency matrix was resampled 500 times and Cavalli-Sforza and Edward’s (1967) chord distances were estimated for population pairs. NJ dendrograms were constructed for each matrix and a consensus NJ dendrogram was generated with CONSENSE (see Felsenstein, 1993). Individual fish were classified to specific populations with a jackknife discriminant analysis (SAS Insti- tute Inc., 1989). Estimation of stock composition Microsatellite allele frequency data were examined for use in estimating stock composition in a mixed- stock fishery. We pooled low-frequency alleles in ad- jacent bins (so that each bin contained at least 6% of the alleles) to reduce the number of genotypes for which frequencies were estimated. This pooling re- sulted in 13 “analysis” bins (91 genotypes) for Ots 101, 10 bins (65 genotypes) for Ots 3, and 13 bins (91 geno- types) for Ofsl03 (Table 3). Genotype frequencies at Ots 101 and Ots 3 were estimated for each population from the allele frequencies by assuming a Hardy- Weinberg distribution of genotypes. Because geno- type frequencies at the Ots 103 locus were generally not in HWE owing to the presence of the null allele, the observed genotype frequencies were used to char- acterize each population for Ots 103. Genotype fre- quencies at all three loci were used as input in a maximum likelihood mixed-stock fishery analysis (Fournier et al., 1984). Two types of hypothetical mixed-stock fishery samples were simulated: single-region fishery samples composed of fish from lower Fraser River and the Thompson River coho salmon populations, and multiregion fishery samples composed of fish from populations from several regions. Populations contributing to the multiregion samples were cho- sen on the basis of known or inferred migration pat- terns. For the Fraser River fishery simulation, only populations from the lower Fraser and Thompson rivers were present in the baseline and in the mixed- stock fishery samples. In the multiregion fishery simulations, all 34 populations were used in the baseline. In all simulations, fishery samples of 200 fish were generated from specified populations in the Table 3 Method of pooling low-frequency alleles for OtslOl, Ots 3, and Otsl03 to reduce the number of genotypes for baseline populations in mixed-stock analyses. Microsatellite allele bin numbers Analysis bin no. OfslOl Ots 3 Ots 103 1 1-7 1 1, 2, 3 2 8, 9 3-11 4-7 3 10 12 8-11 4 11 13 12 5 12 14 13 6 13, 14 15 14, 15 7 15, 16, 17 16 16-20 8 18, 19 17 21, 22 9 20, 21 18 23, 24 10 22 19 25, 26, 27 11 23, 24 28,29 12 25, 26 30-38 13 27-32 39 Small et al. : Population and stock indentification of Oncorhynchus kisutch 851 baseline by sampling randomly, with replacement, thus simulating the randomness present in data col- lection. For each simulated mixed-stock fishery, stock contributions were estimated in 50 independent simu- lations, and means and standard deviations for the es- timated contributions of each population were obtained. Results Allele frequencies and heterozygosity Observed heterozygosity was high in all populations at OfslOl, ranging from 0.72 for Nitinat River to 0.97 for Waukwaas River (Table 1). Heterozygosity was slightly lower at Ots3, ranging from 0.70 for Zolzap River to 0.82 for Cluxewe River (Table 1). At Ots 103, observed heterozygosity varied widely among re- gions, ranging from 0.25 in the Thompson River (TR) to 0.89 in North Coast Vancouver Island populations (NCVI) (Table 2). For Thompson River populations, the apparently low heterozygosity values were due to high frequencies of the null allele. Of the 22 popu- lations with data for multiple year classes, Atnarko was the only one for which allele frequencies differed significantly among year classes at all loci. Allele fre- quencies differed significantly among year classes in the Kitimat, Sustut, and Toboggan populations at Ots3, in the Kitimat River and Pallant Creek popu- lations at OfslOl, and in the Quinsam and Tobog- gan populations at Ots 103. Year-class variation in the Fraser and Thompson River populations are re- ported in Small et al. (1998). Multiple samples from individual populations clustered together in NJ analyses (except for three lower Fraser River popu- lations as noted in Small et al., 1998), indicating that allele frequency differences among year classes were less than those among populations. Thus, all fish collected from the same location in different years were pooled and treated as a single population in subsequent analyses. Hardy-Weinberg equilibrium The degree of deviation from HWE varied substan- tially among the 3 loci (Table 1). After correction for multiple tests (P<0.0015), three populations (Atnarko, Deadman and Upper Pitt) deviated from HWE at Ots 101, and three populations (Kitimat, Big Qualicum, and Sustut) were out of HWE at Ots3 (Table 1). In most populations, observed heterozy- gosity was lower than expected heterozygosity (Table 1), and this lower heterozygosity was reflected in single-locus F[s values of 0.0435 (SD 0.009) for Ots 101 and 0.0617 (SD 0.008) for Ots3 (both P<0.005). HWE was rejected for Ots 103 in 25 out of 34 populations (Table 2), reflecting the presence of the null allele. The single locus Fis value was 0.2738 (SD 0.008). Fish homozygous for the null allele were scored in most populations and corrected Ots 103 allele frequencies were generated for all populations (Table 2). Population differences In pairwise tests, all populations had significantly different allele frequencies at one or more loci, with the exception of two geographically proximate popu- lations in the Thompson River (Lemieux River and Dunn Creek), and two sets of populations from the adjacent Skeena and Nass River systems: (Cedar River [Skeena)]and Zolzap River [Nass]; and Sustut River [Skeena] and Meziadin River [Nass]). The single- and multilocus Fgt values indicated signifi- cant differentiation among populations with values of 0.040 (SD 0.006) for OfslOl, 0.054 (SD 0.009) for Ots3, 0.059 (SD 0.009) for Otsl03 (all P<0.005) and a multilocus value of 0.051 (SD 0.006, P<0.005). Allele frequencies With the exception of Ots 101 allele 96, found only in the Robertson Creek population, all alleles were present in more than one population and region. Thus, population- or region-specific alleles were gen- erally nonexistent, but allele frequencies varied among populations and regions. At OfslOl, lower Fraser River populations had high frequencies of smaller alleles (74% shorter than 143 bp in length) and relatively low frequencies of large alleles, whereas Thompson River populations had low fre- quencies of small alleles (64% longer than 161 bp). In the Thompson River populations, Ots 3 allele 66 was absent and allele 94 was more common than in populations from other regions. Ots3 alleles 104 and 106 were found only in the Skeena and Nass River populations. The most striking differentiation pro- vided by Ots 103 was the high frequency (0.44) of the null allele in Thompson River populations. This was three times the next highest frequency (0.15) that occurred in the north and west coast Vancouver Is- land populations. Population structure The unrooted consensus NJ dendrogram possessed three major branches that provided regional defini- tion among coho salmon populations of British Co- lumbia (Fig. 2). The best defined branch contained the Thompson River (and Bridge River of the upper Fraser drainage) populations, that occurred together 852 Fishery Bulletin 96(4), 1998 in all 500 trees used to construct the consensus tree. The branch containing all lower Fraser River popu- lations was also well supported. Populations from the northern Skeena and Nass rivers were inter- spersed on the third branch, but only the upper Skeena and Nass populations formed a well-sup- ported group (Fig. 2). Lower Skeena and Nass popu- lations were more similar to a diverse central clus- ter of Vancouver Island and mainland coastal coho salmon populations. Within the coastal group, east coast (Big Qualicum and Quinsam), west coast (Robertson and Nitinat), and north coast (Cluxewe, Waukwaas and Stephens) Vancouver Island popula- tions were as well distinguished from each other as they were from the more northern mainland coastal populations or the Queen Charlotte Island popula- tion of Pallant Creek. Estimation of stock composition In the mixed-stock fishery simulated within the Fraser River, estimates of stock composition were accurate and precise, with an aver- age of 2.3% of the mixture incorrectly assigned to each population (Table 4). In general, misassigned portions of the mixture were assigned to closely related populations. For instance, the 30% contribution of Coldwater River fish to the mixture was underesti- mated as 26.4%, but 3.1% of the mix- ture was assigned to the genetically similar Spius River population, al- though no Spius fish were included in the mixture. Fish misassigned to population were almost invariably assigned to the correct region. Less than 1%' of fish were misclassified between the Thompson River and Lower Fraser regions (Table 4). Thus, within the Fraser River drainage, populations contributing to a mixed- stock sample can be identified to re- gion and to individual population, or population group within region, with a high degree of accuracy and precision. Accuracy and precision of stock composition estimates were also high in the simulated mixed-stock samples to which populations from different regions contributed (Table 5). Aver- age error of individual population contributions to the mixed fishery was 1.5% (3% for populations actu- ally contributing to the mixture), and average error of regional contribu- tions was 2.2% (higher for regions actually contributing). In most simu- lations, the contributions of popula- tions present in the mixture were un- derestimated because a small propor- tion of the mixture was allocated to baseline populations not present in the mixture. In general, misassigned fish were allocated to genetically similar populations and thus were Figure 2 Unrooted neighbor-joining tree relating 34 British Columbia coho salmon populations. The tree was constructed from a consensus of Cavalli-Sforza and Edward’s chord distances. Bootstrap values at the tree nodes were computed over 500 replications by resampling the allele frequency matrix. Bootstrap values indicate the percentage of trees in which the populations beyond the node occurred together. All names correspond to population names given in Tables 2 and 3 with the exception of Big Qualicum River which is shortened to Big Qualic. The number of the population (from Fig. 1) is next to the name. Small et ai.: Population and stock indentification of Oncorhynchus kisutch 853 correctly identified to region. Regions with no popu- lations present in the mixture were allocated less than 5% of the fish, except in the mixture composed of Vancouver Island populations (mix 1, Table 5). In that case, 10% of the fish were allocated to the lower Fraser River. The regional misclassification of Van- couver Island coho salmon as fish of lower Fraser origin in mix 1 was due largely to the allocation of 4% of the mixture to Chilliwack River, the lower Fraser population most genetically similar to the northern Vancouver Island populations (Fig. 2). Simi- larly, Vancouver Island and central coast fish were underestimated in mix 5, and some of them attrib- uted to the lower Fraser rather than to other coastal island, mainland, or lower Skeena and Nass popula- tions as might be expected from the dendrogram (Table 5; Fig. 2). Conversely, when lower Fraser coho salmon formed a high proportion of the mixture (mix 3) , they were underestimated and misidentified fish tended to be attributed to Vancouver Island popula- tions (Table 5). In contrast, the genetic distinctiveness of Thomp- son River coho salmon resulted in their accurate iden- tification in mixtures in which they were present (mixes 3 and 5) and little misrepresentation in mix- tures from which they were absent (mixes 1,2, and 4) . Nass and Skeena populations were well-separated (mix 4), and contributions from lower Skeena (Ce- dar and Clearwater) populations were identified as well as those from the upper Skeena-Nass popula- tions. Given the lack of separation of Skeena and Nass in the NJ dendrogram (Fig. 2), this result needs to be confirmed by more extensive sampling and fur- ther mixture analysis for populations within these two watersheds. Individual classification of Nass and Skeena fish (see below), and the lack of significant allele frequency differences between two Nass- Skeena pairs of populations, indicated that additional genetic markers may be required for accurate differ- entiation of Nass and Skeena coho salmon. Never- theless, the results of these preliminary analyses indicate that a mixed-stock sample of coho salmon collected in the field can generally be resolved into its regional components by using microsatellite ge- netic markers. Identifying individuals Accuracy of the classification of individual fish to region with discriminant analysis varied among re- gions (Table 6), but individual classification was gen- erally much less accurate than was estimation of stock composition. An average of 48% of fish, rang- ing from 85% of Thompson to 20% of Nass individu- als, was correctly classified to region. Misclassified Table 4 Accuracy and precision of estimated Fraser River coho salmon population contributions in a simulated mixed- stock sample from a 16-population Fraser River baseline. A 200-fish mixture was generated with replacement from the baseline 50 times and the population composition was estimated for each mixture. The mean percentage of the mixture allocated to each population is given, followed by the standard deviation in parentheses. Regional totals are given in the rows marked TR (Thompson River) and L Fr (lower Fraser River). True Mean SD Coldwater 30 26.4 (4.4) Salmon 0 1.3 (1.6) Eagle 20 15.7 (3.4) Spius 0 3.1 (3.5) Lemieux 0 1.0 (1.5) Dunn 0 0.6 (1.3) Louis 0 1.0 (1.7) Deadman 0 0.5 (0.9) Bridge 0 0.7 (1.3) TR 50 50.2 (1.4) Chilliwack 20 16.8 (4.0) Chehalis 0 1.4 (1.8) Stave 15 12.6 (3.3) Upper Pitt 0 1.8 (2.1) Nicomen 15 12.2 (3.4) Norrish 0 2.9 (2.5) Inch 0 2.0 (2.4) L Fr 50 49.8 (1.4) fish were most commonly attributed to the most ge- netically similar region according to relationships depicted in the NJ dendrogram. The Nass and Skeena populations were interspersed in the NJ analysis (Fig. 2), and an essentially equal proportion of Nass River fish (20%) were identified as Nass and Skeena fish (Table 6). Skeena River fish were identified more accurately than Nass River fish, with 53% and 11% of the Skeena River fish classified as being of Skeena and Nass origin, respectively. If only the Fraser River populations were included in the baseline, Fraser River coho salmon were assigned to region accurately, with 93% correctly identified to either the lower Fraser River or the Thompson River (data not shown). Discussion Microsatellite DNA analysis shows great promise for the elucidation of population structure in coho salmon. Very strong regional structure was appar- ent in the British Columbia coho salmon populations 854 Fishery Bulletin 96(4), 1 998 Table 5 Estimated stock composition of five 200-fish mixtures of coho salmon from several regions in British Columbia using a 34-stock baseline. Each mixture was generated 50 times, and stock composition of the mixture was estimated by randomly resampling each baseline population, with replacement, to derive a new estimation of the fish mixture composition. The mean and standard deviations of 50 estimations are reported for each population. Individual population estimates are followed by the sum of contri- butions from a region (bold type). Mix 1 Mix 2 Mix 3 Mix 4 Mix 5 True Mean SD True Mean SD True Mean SD True Mean SD True Mean SD Pallant 0 0.02 0.02 0 0.01 0.01 0 0.00 0.01 0 0.00 0.01 0 0.01 0.01 Atnarko 0 0.01 0.01 0.13 0.11 0.02 0 0.00 0.01 0 0.02 0.02 0 0.02 0.02 Kitimat 0 0.01 0.02 0.13 0.10 0.02 0 0.01 0.01 0 0.02 0.02 0.20 0.16 0.04 C Coast 0 0.02 0.02 0.26 0.21 0.03 0 0.01 0.01 0 0.04 0.03 0.20 0.18 0.04 Cluxewe 0 0.01 0.01 0 0.01 0.01 0 0.00 0.01 0 0.00 0.01 0 0.01 0.01 Stephens 0.13 0.10 0.02 0 0.00 0.00 0 0.01 0.01 0 0.01 0.01 0 0.01 0.02 Waukwaas 0.17 0.12 0.03 0 0.00 0.00 0 0.01 0.01 0 0.00 0.01 0 0.00 0.01 NCVI 0.30 0.23 0.03 0 0.01 0.01 0 0.02 0.02 0 0.01 0.01 0 0.02 0.02 Nitinat 0.13 0.10 0.02 0 0.00 0.00 0 0.00 0.01 0 0.00 0.00 0 0.00 0.00 Robertson 0.13 0.12 0.03 0 0.00 0.00 0.15 0.13 0.01 0 0.00 0.01 0 0.00 0.00 WCVI 0.26 0.22 0.03 0 0.00 0.01 0.15 0.13 0.01 0 0.00 0.01 0 0.00 0.01 Quinsam 0.17 0.15 0.05 0 0.02 0.03 0 0.01 0.01 0 0.01 0.01 0 0.03 0.04 Big Qualic 0.26 0.22 0.04 0.17 0.14 0.03 0 0.01 0.02 0 0.01 0.01 0.25 0.19 0.04 ECVI 0.43 0.37 0.06 0.17 0.16 0.04 0 0.02 0.02 0 0.02 0.02 0.25 0.22 0.05 Toboggan 0 0.00 0.00 0.26 0.24 0.03 0 0.00 0.00 0.14 0.15 0.03 0 0.01 0.01 Cedar 0 0.01 0.01 0 0.01 0.01 0 0.00 0.01 0.13 0.09 0.03 0.20 0.13 0.03 Babine 0 0.00 0.01 0.17 0.16 0.02 0 0.00 0.00 0.26 0.25 0.04 0 0.00 0.00 Clearwater 0 0.00 0.01 0 0.00 0.01 0 0.00 0.00 0.17 0.14 0.03 0 0.01 0.01 Sustut 0 0.00 0.00 0 0.01 0.01 0 0.00 0.00 0.13 0.10 0.03 0 0.00 0.01 Skeena 0 0.01 0.02 0.43 0.42 0.03 0 0.00 0.00 0.83 0.73 0.04 0.20 0.15 0.03 Tseax 0 0.01 0.01 0 0.00 0.01 0 0.01 0.01 0 0.01 0.01 0 0.01 0.01 Zolzap 0 0.00 0.01 0 0.01 0.01 0 0.00 0.01 0 0.01 0.01 0 0.01 0.01 Meziadin 0 0.01 0.01 0 0.01 0.01 0 0.00 0.00 0.17 0.13 0.03 0 0.01 0.01 Nass 0 0.02 0.03 0 0.02 0.02 0 0.01 0.00 0.17 0.15 0.03 0 0.03 0.02 Coldwater 0 0.00 0.00 0 0.00 0.00 0 0.03 0.02 0 0.00 0.00 0 0.00 0.00 Salmon 0 0.00 0.00 0 0.00 0.00 0 0.00 0.01 0 0.00 0.00 0.10 0.09 0.01 Eagle 0 0.00 0.00 0 0.00 0.00 0 0.01 0.01 0 0.00 0.00 0 0.00 0.00 Spius 0 0.00 0.00 0 0.00 0.00 0.20 0.16 0.03 0 0.00 0.00 0 0.00 0.00 Lemieux 0 0.00 0.00 0 0.00 0.00 0 0.00 0.01 0 0.00 0.00 0 0.00 0.00 Dunn 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 Louis 0 0.01 0.01 0 0.00 0.00 0 0.01 0.01 0 0.00 0.00 0 0.00 0.00 Deadman 0 0.01 0.01 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0 0.01 0.00 Bridge 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 TR 0 0.02 0.01 0 0.01 0.01 0.20 0.21 0.01 0 0.01 0.01 0.10 0.10 0.01 Chilliwack 0 0.04 0.03 0 0.01 0.01 0.10 0.10 0.04 0 0.00 0.01 0 0.02 0.03 Chehalis 0 0.01 0.02 0 0.01 0.02 0.25 0.18 0.03 0 0.01 0.01 0 0.02 0.03 Stave 0 0.01 0.02 0 0.02 0.02 0 0.02 0.02 0 0.01 0.02 0 0.03 0.03 Upper Pitt 0 0.01 0.01 0 0.01 0.01 0 0.01 0.02 0 0.00 0.01 0.10 0.08 0.04 Nicomen 0 0.01 0.02 0 0.01 0.01 0.10 0.10 0.03 0 0.00 0.01 0 0.01 0.02 Norrish 0 0.01 0.02 0 0.01 0.02 0 0.02 0.02 0 0.01 0.02 0 0.03 0.03 Inch 0 0.01 0.01 0.13 0.09 0.03 0.20 0.17 0.03 0 0.00 0.01 0.15 0.11 0.03 L Fr 0 0.10 0.04 0.13 0.16 0.03 0.65 0.60 0.03 0 0.04 0.03 0.25 0.30 0.05 of our study on the basis of only three microsatellite loci, albeit ones selected for high levels of intra- and interpopulation polymorphism. The four major re- gional components of diversity in B.C. coho salmon almost certainly constitute the phylogeographic legacy of the last ice age, which ended 10-15,000 years ago. When the headwaters of the neighboring Columbia River system were in contact with the up- per reaches of the Thompson-Fraser watershed, Co- lumbia River coho salmon recolonized the Thomp- son and upper Fraser River (McPhail and Lindsey, 1986; Wehrhahn and Powell, 1987; Small et al., 1998). Coho salmon from the northern Bering Sea- Yukon River glacial refuge dispersed southward by Small et al.: Population and stock indentification of Oncorhynchus kisutch 855 Table 6 Percentage of classifications of individual fish to regions for 34 populations of coho salmon from British Columbia. In this jack- knife analysis the fish tested was not included in the discriminant analysis sample, which included the rest of the fish. The percentage of fish correctly classified to region is in the “correct” column and the percentage of fish from that region misclassified to other regions is read across the row. The percentage of fish from other regions misclassified to a particular region is read down the region column, n is the number of fish from the region. Region Region n Correct Pallant C. Coast L. Fraser Thompson WCVI ECVI NCVI Nass Skeena Total Pallant 92 48.91 6.52 8.70 5.43 6.52 7.61 13.04 1.09 2.17 100 C. Coast 219 27.85 10.05 5.94 6.85 4.11 17.81 11.42 7.31 8.68 100 L. Fraser 1018 53.73 4.72 5.89 2.85 3.05 8.74 12.87 5.11 3.05 100 Thompson 798 84.96 1.13 2.51 2.63 0.25 3.26 3.01 1.00 1.25 100 WCVI 115 60.00 6.09 6.09 11.30 2.61 5.22 3.48 2.61 2.61 100 ECVI 338 44.08 5.03 10.36 13.61 2.37 5.03 10.06 6.80 2.66 100 NCVI 133 43.61 4.51 5.26 18.05 6.02 1.50 10.53 6.77 3.76 100 Nass 133 20.30 0.75 13.53 14.29 6.02 6.77 8.27 9.02 21.05 100 Skeena 412 52.67 3.40 5.58 6.55 6.07 1.94 6.80 6.31 10.68 100 sea, or traversed the shifting freshwater waterways of northern B.C., to recolonize the upper reaches of major watersheds as far south as the Nass and Skeena rivers (Lindsey and McPhail, 1986). Dispersal of coho salmon from a central B.C. coastal refuge(ia) located on the mainland or unglaciated portions of Vancouver or the Queen Charlotte Islands (or both) (Warner et al., 1982) likely established the hetero- geneous mainland-coastal population group, of which lower Fraser coho salmon populations may be a dis- tinctive offshoot. Of the four regional components of biodiversity in B.C. coho salmon, the Thompson and upper Fraser is the most distinctive. Little introgression has ap- parently occurred between the lower Fraser and Thompson River coho salmon populations despite their common passage through the lower Fraser River on their return to spawning grounds for over 3,000 generations. Coho salmon populations are few and small in the Fraser River drainage above its confluence with the Thompson River, and our data show no evidence of introgression between the coho salmon of the Thompson-Fraser Rivers and the up- per Skeena watersheds such as that postulated for sockeye salmon (Wood et al., 1994). However, our sampling of the Skeena watershed is limited to date, and the current analysis (in which larger, relatively infrequent, alleles have been binned) has limited power for the detection of historical gene exchange through an analysis of rare alleles. The Skeena River watershed has been identified as the southern limit for freshwater fish and sock- eye salmon that dispersed from a northern glacial refuge (Lindsey and McPhail, 1986; Wood et al., 1994; Bickham et al., 1995). Thus, it seems likely that the coho salmon populations of the upper Skeena and Nass watersheds are derived from Beringia. The ge- netic intermediacy of lower Skeena and Nass popu- lations between upper Skeena-Nass and neighboring coastal populations suggests a hybrid nature for the lower river populations. The extent to which the com- posite nature of these populations reflects historical or current gene flow (or both) between the two found- ing groups, and the adaptive consequences of such introgression, has yet to be determined. The coho populations of the lower Fraser drainage basin formed a cohesive genetic group in the NJ analysis of genetic distance in this study. This was in sharp contrast to the heterogeneity observed among other coastal mainland and island populations likely derived from one or more coastal refugia. The heterogeneity of central coast populations may re- flect the existence of several coastal refugia, intro- gression from northern coho populations originating from Beringia that have yet to be well characterized, or simply founder effects in the establishment of in- dividual coastal populations from a single refuge, as postulated for sockeye salmon (Wood et al., 1994). If the heterogeneity does result from an admixture of several founding groups among coastal coho popula- tions, the lower Fraser River populations may bet- ter represent the origin al genetic profile of fish from a single, possibly southern, coastal refuge. Interest- ingly, Vancouver Island coho salmon were more fre- quently misclassified as lower Fraser than as cen- tral coast fish in the mixed-stock fishery simulations of our study, in spite of their apparently greater ge- netic similarity to central coast populations. If the 856 Fishery Bulletin 96(4), 1 998 lower Fraser River populations are characteristic of coho salmon derived from a refuge located on, or to the south of, Vancouver Island, Puget Sound coho salmon populations might be expected to be of simi- lar origin. This is consistent with the inclusion, based primarily on genetic data, of lower Fraser and south- eastern Vancouver Island coho salmon populations in a Puget Sound and Strait of Georgia ESU (Weitkamp et ah, 1995). Further sampling of Vancouver Island populations will be necessary to define the boundaries of the historical groups of coho salmon that likely converge there. The geographic basis for population structure of coho salmon revealed in this study is remarkably similar to that described for sockeye salmon based on allozyme and mtDNA data (Wood et al., 1994; Bickham et ah, 1995). A genetic discontinuity be- tween chinook salmon originating from the Colum- bian and Beringial glacial refugia (Gharrett et ah, 1987; Cronin et ah, 1993) also lies in British Colum- bia, and dispersal from a coastal refuge(ia) can be traced in the microsatellite data for chinook salmon as well (Beacham, unpubl. data). Thus, the phylo- geographic reconstruction of postglacial dispersal based on freshwater fish distributions in British Columbia (McPhail and Lindsey, 1970, 1986; Lindsey and McPhail, 1986) has proven to be taxonomically robust and also provides the foundation for the genetic architecture of anadromous Pacific salmon species. The mixed-stock fishery analyses demonstrated the utility of microsatellite DNA variation for coho salmon stock identification. We obtained accurate and precise estimates both of population contribu- tions in mixed-stock samples from a single drainage and of population and regional contributions in mixed-stock samples drawn from several regions. An important feature of the microsatellite data set is the strong regional structuring of the observed ge- netic variation, which means that contributions from populations present in a mixture sample, but not in the baseline, will be identified correctly to region. Identifying individual fish to correct populations is more difficult than estimating percentage contribu- tions to a stock mixture, because only characteris- tics of individual fish are used in the classification. In general, the microsatellite loci of this study pro- vided a similar level of accuracy in the classification of individual coho salmon to population and region as did minisatellite DNA markers (Beacham et al., 1996). Within the Fraser River drainage, identifica- tion of individual fish was more accurate with microsatellite data (correct identification of 54% of lower Fraser and 85% of upper Fraser Rive coho salmon) than with minisatellite data (correct identi- fication of 30% and 60% of the respective groups). The identification of individual fish is an important enforcement tool, and may improve as more microsatellite loci are added to the database. The results of this study are consistent with a de- piction of population structure in coho salmon as dis- tinct phylogenetic lines composed of geographically based metapopulations (McPhail, 1997). The more consistent regional grouping of coho salmon popula- tions than of sockeye salmon (Wood et al., 1994) may reflect greater, or more recent, gene flow among geo- graphically proximate coho salmon populations than among similar sockeye populations, or may reflect differences between allozyme- and microsatellite- based data sets. Moreover, the increasing power of genetic methods applied to coho salmon data, as dem- onstrated in this and other (Weitkamp et al., 1995; Beacham et al., 1996; Van Doornik et al., 1996; Miller et al., 1996; Miller and Withler, 1997; Small et al., 1998) studies, enable us not only to delineate regional (metapopulation) structure in coho salmon but also to identify the regional contributions of coho from different metapopulations in mixed-stock fishery harvests. Thus, the challenge for metapopulation- based coho salmon management may lie not so much in the delineation of metapopulation structure ( McPhail, 1997 ) as in the evaluation of the biological and social costs associated with the loss, even if only temporary in historical terms, of the less productive components of metapopulation structure during pe- riods of overall low abundance. Additional aspects of metapopulation theory, as applied to Pacific salmonids, need to be addressed before this model of population structure will sup- port practical management decisions. Managers need to know, at any given point in history, how the adap- tive genetic diversity of a metapopulation is likely to be distributed among its subpopulations, and how many subpopulations can be lost before evolution- ary potential is compromised. This depends on, among other things, which model of metapopulation structure is adopted. Are salmonid metapopulations of the “source-sink” variety in which gene flow is basically unidirectional from large source subpopu- lations to ephemeral sink subpopulations (Pulliam, 1988; Pulliam and Danielson 1991)? Or are salmo- nid metapopulations of the “balanced exchange” type, in which gene flow is bidirectional and migration rates are inversely proportional to subpopulation size (McPeek and Holt, 1992; Doncaster et al., 1997)? Current models of metapopulation structure have been more extensively investigated with respect to population dynamics (extinctions and recolonizations among subpopulations) than with respect to popula- tion and evolutionary genetics (the spatial and tern- Small et al.: Population and stock indentification of Oncorhynchus kisutch 857 poral distribution of genetic diversity among sub- populations). The finding in this, and other molecu- lar genetic studies on coho salmon (Beacham et al., 1996; Miller et al., 1996), of temporally stable allele frequency differences at neutral loci among local “stocks” (subpopulations in the metapopulation model) may indicate that there is very little effective gene flow among subpopulations (and few recoloni- zation events in vacant habitat) on time scales of relevance to managers. In summary, we have demonstrated that variation at microsatellite loci can be used both to define the regional and local components of coho salmon popu- lation structure in British Columbia and to identify these elements in stock composition estimation for mixed-stock fishery analysis. Molecular genetic tech- nology will enable delineation of metapopulations or ESUs (or both) in the coho salmon of British Colum- bia, but management tools based on these concepts are lacking. Acknowledgments This work was funded by a National Sciences and Engineering Research visiting postdoctoral fellow- ship to M. P. Small and by the Department of Fish- eries and Oceans. Samples were collected with the assistance of J. Candy and staff at the Habitat and Assessment Branch of the DFO. Technical work was greatly assisted by A. Schulze and D. Tuck. R. J. Nelson, C. C. Wood, and J. Candy were most helpful with advice on laboratory techniques, data organi- zation, analysis, and statistics. Literature cited Angers, B., L. Bernatchez, A. Angers, and L. Desgroseillers. 1995. Specific microsatellite loci for brook charr reveal strong population subdivision on a microgeographic scale. J. Fish Biol. 47, (suppl. A): 177-185. Bartley, D. M., B. Bentley, P. G. Olin, and G. A. E. Gall. 1992. Population genetic structure of coho salmon ( Oncorhynchus kisutch) in California. Calif. Fish and Game 78:88-104. Beacham, T. D., L. Margolis, and R. J. Nelson. 1998. A comparison of methods of stock identification for sockeye salmon ( Oncorhynchus nerka ) in Barkley Sound, British Columbia. N. Pac. Andr. Fish Comm. Bull. 1: 227-239. Beacham, T. D., K. M. Miller, and R. E. Withler. 1996. Minisatellite DNA variation and stock identification of coho salmon. J. Fish Biol. 48:1-19. Beamish, R. J., and D. R. Bouillon. 1993. Pacific salmon production trends in relation to climate. Can. J. Fish. Aquat. Sci. 50:1002-1016. Bickham, J. W., C. C. Wood, and J. C. Patton. 1995. Biogeographic implications of cytochrome b sequences and allozymes in sockeye (Oncorhynchus nerka). J. Hered. 86:140-144. Cavalli-Sforza, L. I.., and A. W. F. Edwards. 1967. Phylogenetic analysis: models and estimation procedures. Am. J. Hum. Gen. 19:233-257. Cronin, M. A., W. J. Spearman, R. L. Wilmot, J. C. Patton, and J. W. Bickham. 1993. Mitochondrial DNA variation in chinook (Oncorhy- nchus tshawytscha) and chum salmon (O. keta) detected by restriction enzyme analysis of polymerase chain reac- tion (PCR) products. Can. J. Fish. Aquat. Sci. 50:708-715. Dempster, A. P., N. M. Laird, and D. B. Rubin. 1977. Maximum likelihood from incomplete data via the EM algorithm. J. Royal Stat. Soc. B 39:1-38. Doncaster, C. P., J. Clobert, B. Doligez, L. Gustafsson, and E. Danchin. 1997. Balanced dispersal between spatially varying local populations: an alternative to the source-sink model. Am. Nat. 150:425-445. Felsenstein, J. 1993. PHYLIP, version 3.4. Department of Genetics, Univ. Washington, Seattle, WA, unpaginated. Fraser, F. J., P. J. Starr, and A. Y. Federenko. 1982. A review of the chinook and coho salmon of the Fraser River. Can. Tech. Rep. Fish. Aquat. Sci. 1126, 130 p. Forbes, S., K. Knudsen, and F. Allendorf. 1993. Genetic variation in DNA of coho salmon from the Lower Columbia River. Final Report to the U.S. Dep. Energy, BPA., Div. Fish and Wildlife, Contract DE-BI179- 92BP 30198, 25 p. Fournier, D. A., T. D. Beacham, B. E. Riddell, and C. A. Busack. 1984. Estimating stock composition in mixed stock fisher- ies using morphometric, meristic, and electrophoretic characteristics. Can. J. Fish. Aquat. Sci. 41:400-408. Galbraith, D. A., P. T. Boag, H. L. Gibbs, and B. N. White. 1991 . Sizing bands on autoradiograms: a study of precision for scoring DNA fingerprints. Electrophoresis 12:210-220. Gharrett, A. J., S. M. Shirley, and G. R. Tromble. 1987. Genetic relationships among populations of Alaskan chinook salmon (Oncorhynchus tshawytscha). Can. J. Fish. Aquat. Sci. 44:765-774. Gill, P., K. Sullivan, and D. J. Werrett. 1990. The analysis of hypervariable DNA profiles: problems associated with the objective determination of the prob- ability of a match. Hum. Genet. 85:75-79. Goudet, J. 1995. Fstat, version 1.2: a computer program to calculate F-statistics. J. Hered. 86:485-486. Guo S. W., and E. A. Thompson. 1 992. Performing the exact test of Hardy -Weinberg propor- tions for multiple alleles. Biometrics 48:361-372. Lessios, H. A. 1992. Testing electrophoretic data for agreement with Hardy- Weinberg expectations. Mar. Biol. 112:517-523. Lindsey, C. C., and J. D. McPhail. 1986. Zoogeography of fishes of the Yukon and Mackenzie basins. In C. H. Hocutt and E. O. Wiley (eds.), Zoogeog- raphy of North American freshwater fishes, p. 639-674. John Wiley & Sons, Inc., New York, NY. McConnell, S. K., P. O’Reilly, L. Hamilton, J. M. Wright, and P. Bentzen. 1995. Polymorphic microsatellite loci from Atlantic salmon ( Salmo salary, genetic differentiation of North American 858 Fishery Bulletin 96(4), 1 998 and European populations. Can. J. Fish. Aquat. Sci. 52:1863-1872. McPeek, M. A., and R. D. Holt. 1992. The evolution of dispersal in spatially and tempo- rally varying environments. Am. Nat. 140:1010-1027. McPhail, J. D. 1997. The origin and speciation of Oncorhynchus revisited. In D. J. Stouder, P. A. Bisson and R. J. Naiman (eds.). Pa- cific salmon and their ecosystems: status and future op- tions, p. 29-38. Chapman and Hall, New York, NY. McPhail, J. D., and C. C. Lindsey. 1970. Freshwater fishes of northwestern Canada and Alaska. Bull. Fish. Res. Board Can. 173, 381 p. 1986. Zoogeography of freshwater fishes of Cascadia (the Columbia system and rivers north to the Stikine). In C. H. Hocutt and E. O. Wiley (eds.), Zoogeography of North American freshwater fishes, p. 615-637. John Wiley & Sons, Inc., New York, NY. Miller, K. M., and R. E. Withler. 1997. Mhc diversity in Pacific salmon: population structure and trans-species allelism. Hereditas 127:83-95. Miller, K. M., R. E. Withler, and T. D. Beacham. 1996. Stock identification of coho salmon ( Oncorhynchus kisutch ) using minisatellite DNA variation. Can. J. Fish. Aquat. Sci. 53:181-195. Nehlsen, W., Williams, J. E., and J. A. Lichatowich. 1991 Pacific salmon at the crossroads: stocks at risk from California, Oregon, Idaho, and Washington. Fisheries 16(2):4-21. Northcote, T. G., and D. Y. Atagi. 1997. Pacific salmon abundance trends in the Fraser river watershed compared with other British Columbia systems. In D. J. Stouder, P. A. Bisson, and R. J. Naiman (eds.), Pacific salmon and their ecosystems: status and future options, p. 199-219. Chapman and Hall, New York, NY. Pulliam, H. R. 1988. Sources, sinks, and population regulation. Am. Nat. 132:652-661. Pulliam, H. R., and B. J. Danielson. 1991. Sources, sinks, and habitat selection — a landscape per- spective on population dynamics. Am. Nat. 137:S50-S66. Raymond, M., and F. Rousset 1995a. GENEPOP (version 1.2): population genetics software for exact tests and ecumenism. Hereditary 86:248-249. 1995b. An exact test for stock differentiation. Evolution 49:1280-1283. Reisenbichler, R. R. 1997. Genetic factors contributing to declines of anadro- mous salmonids in the Pacific northwest. In D. J. Stouder, P. A. Bisson, and R. J. Naiman (eds.), Pacific salmon and their ecosystems: status and future options, p. 223- 224. Chapman and Hall, New York, NY. Ricker, W. E. 1972. Heredity and environmental factors affecting certain salmonid populations. In R. C. Simon and P. A. Larkin (eds.), The stock concept of Pacific salmon, p. 19-60. H. R. MacMillan lectures in fisheries, Univ. British Columbia, Vancouver. Roff, D. A., and P. Bentzen. 1989. The statistical analysis of mitochondrial DNA poly- morphisms: x2 and the problem of small sample sizes. Mol. Biol. Evol. 6:539-545. Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method for recon- structing phylogenetic trees. Mol. Biol. Evol. 4:406-425. Sakai, R. K., S. Scharf, F. Falloona, K. B. Mullis, G. T. Horn, H. A. Erlich, and N. Arnheim. 1985. Enzymatic amplification of Beta-globin genomic se- quences and restriction site analysis for diagnosis of sickle cell anemia. Science (Wash., D.C.) 230:1350-1354. SAS Institute Inc. 1989. SAS/STAT user’s guide, version 6, 4th edition, vol. 1. p. 678-771. SAS Institute Inc., Cary, NC. Scribner, K. T., J. R. Gust, and R. L. Fields. 1996. Isolation and characterization of novel salmon microsatellite loci: cross-species amplification and popu- lation genetic applications. Can. J. Fish. Aquat. Sci. 53:833-841. Small, M. P., T. D. Beacham, R. E. Withler, and R. J. Nelson. 1998. Discriminating coho salmon (Oncorhynchus kisutch ) populations within the Fraser River, British Columbia. Mol. Ecol. 7:141-155. Tautz, D. 1989. Hypervariability of simple sequences as a general source for polymorphic DNA markers. Nuc. Acid Res. 17:6563-6471. Van Doornik, D. M., G. B. Milner, and G. A. Winans. 1996. Transferrin polymorphism in coho salmon, Oncorhy- nchus kisutch , and its application to genetic stock identification. Fish. Bull. 94:566-575. Walters, C. J. 1993. Where have all the coho gone? In L. Berg and P. W. Delaney (eds.), Proceedings of the coho salmon workshop, Nanaimo, B.C., May 26-28, 1992, p. 1-8. Department of Fisheries and Oceans, Nanaimo. Waples, R. S. 1991. Pacific salmon, Oncorhynchus spp., and the defini- tion of “species” under the Endangered Species Act. Mar. Fish. Rev. 53(3):ll-22. Warner, B. G., R. W. Mathewes, and J. J. Clague. 1982. Ice-free conditions on the Queen Charlotte Islands, British Columbia, at the height of late Wisconsin glacia- tion. Science (Wash. D.C.) 218:675-677. Wehrhahn, C. F., and R. Powell. 1987. Electrophoretic variation, regional differences and gene flow in the coho salmon ( Oncorhynchus kisutch) of Southern British Columbia. Can. J. Fish. Aquat. Sci. 44:822-831. Weir, B. S., and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38:1358-1370. Weitkamp, L. A., T. C. Wainwright, G. J. Bryant, G. B. Milner, D. J. Teel, R. G. Kope, and R. S. Waples. 1995. Status review of coho salmon from Washington, Or- egon, and California. U.S. Dep. Commerce, NOAA Tech. Memo. NMFS-NWFSC-24, 258 p. Wood, C. C., B. E. Riddell, D. T. Rutherford, and R. E. Withler. 1994. Biochemical genetic survey of sockeye salmon (Oncorhynchus nerka) in Canada. Can. J. Fish. Aquat. Sci. 51:114-131. 859 Gonadal development and associated changes in plasma reproductive steroids in English sole, Pleuronectes vetulus, from Puget Sound, Washington Abstract .—Gonadal development, and associated changes in reproductive parameters were investigated in adult female and male English sole (Pleuro- nectes vetulus) in Puget Sound to pro- vide baseline data for future research on effects of contaminants on reproduc- tive endocrine function in this species. Changes in gonadal histology, gonado- somatic index (GSI), and plasma repro- ductive steroids (female: 17|3-estradiol [E2], testosterone [T] ; male: T, 11- ketotestosterone [ 1 1KT] ) were moni- tored throughout the spawning season. Female sole sampled July through early September were primarily re- gressed or previtellogenic and had low GSI and plasma steroid levels. GSI and plasma steroid levels were also low in male sole sampled during this time, but the majority had already entered sper- matogenesis and, in some fish, produc- tion of mature sperm was observed. Fish sampled in October were in the early stages of vitellogenesis and sper- miogenesis and showed increases in GSI and plasma steroid levels. By No- vember, about 50% of female fish had entered vitellogenesis and about 30% of males were producing sperm. Propor- tions of vitellogenic females and sperm- producing males continued to increase through January, with significant num- bers of spawning females and males present in February. By late March, the majority of both sexes were spent. Vitellogenic female sole had the high- est plasma E2 levels, vitellogenic sole with hydrated oocytes had the highest GSI, and spawning female sole had the highest plasma T levels. Plasma T, 11KT, and GSI were highest in spawn- ing male sole. Reproductive parameters returned to baseline levels in spawned out female and male sole. A potent maturation-inducing steroid (MIS) in many species of teleosts, 17a, 20(3- dihydroxy-4-pregnen-3-one, was not detected in spawning English sole. Ad- ditional research is needed to identify the MIS in English sole and to under- stand better the hormonal regulation of early gonadal development in male sole. Manuscript accepted 28 January 1998. Fish. Bull. 96:859-870 (1998). Sean Y. Sol O. Paul Olson Daniel R Lomax Lyndal L. Johnson Environmental Conservation Division Northwest Fisheries Science Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, Washington 98112 E-mail address (for S. Y. Sol): Sean.Sol@noaa.gov English sole (Pleuronectes vetulus) is a commercially important flatfish species indigenous to the west coast of North America. It is also a pri- mary sentinel species for a number of environmental monitoring pro- grams on the west coast of the United States, including the Na- tional Benthic Surveillance Project (Myers et ah, 1994) and the Puget Sound Ambient Monitoring Pro- gram (PSWQA1). Studies suggest that female English sole are quite sensitive to environmental contami- nants. For example, female sole from several sites within Puget Sound, WA, with high levels of poly- chlorinated biphenyls (PCBs) and polycyclic aromatic hydrocarbons (PAHs) (in sediment), exhibit vari- ous types of reproductive dysfunc- tion, including depressed sex hor- mone levels, altered or inhibited re- productive development, and re- duced egg and larval viability (Johnson et al., 1988, 1993; Casillas et ah, 1991; Collier et ah, in press). Male English sole may also be at risk for reproductive injury. Recent studies suggest that exposure to certain chemical compounds (e.g. alkylphenols, phthalates, and some PCBs and chlorinated pesticides) can reduce testicular growth and sperm production in male teleosts (Jobling et ah, 1996; Nimrod and Benson, 1996), and alterations in reproductive steroids have been re- ported in male fish exposed to oil and other pollutants (Truscott et ah, 1992a; Idler et ah, 1995). However, no comprehensive studies have been conducted on the effects of en- vironmental contaminants on re- productive function in male English sole. Before such studies can be car- ried out effectively, the normal re- productive endocrine cycle of male English sole must be characterized. The general life history of English sole is well known (Ketchen, 1947; Harry, 1959; Garrison and Miller, 1982), and its cycle of oocyte devel- opment and related changes in es- tradiol-17p (E2) and other plasma parameters have been described 1 PSWQA (Puget Sound Water Quality Au- thority). 1995. 1994 Puget Sound up- date: fifth annual report of the Puget Sound Ambient Monitoring Program. Puget Sound Water Quality Authority, Olympia, WA, 122 p. 860 Fishery Bulletin 96(4), 1 998 (Johnson and Casillas, 1991; Fargo and Tyler, 1994). However, no information exists on other reproduc- tive steroids involved in oocyte development and spawning in female English sole, or on changes in reproductive parameters in male English sole dur- ing gonadal recrudescence. This study was designed to gain a better understanding of reproductive pro- cesses in both female and male English sole. This information can be used to assess the impact of en- vironmental degradation on their reproductive de- velopment and to develop effective techniques for flat- fish mariculture. It may also be useful in evaluating the reproductive status of English sole stocks that are a fisheries resource off the northwest coast of the United States. Wild populations of female and male English sole were sampled during the 1992-93 and 1994-95 spawning cycles at two residential sites as well as at a known spawning ground in Puget Sound, Wash- ington. Several reproductive parameters, including gonadosomatic index (GSI), histologically determined stages of gonadal development, and steroid hormones (female: plasma testosterone [T] and E2; male: T and 1 1-ketotestosterone [11KT] ) were monitored through- out the spawning season. Additionally, 17a, 20p- dihydroxy-4-pregnen-3-one (17a, 20(3-P), a potent maturation inducing steroid (MIS) in many teleosts (see review by Scott and Canario, 1987), was tested as the potential MIS in English sole. Materials and methods Chemicals Tritium-labeled steroids were purchased from DuPont NEN (USA), and Amersham (UK). The anti- bodies for E2 and T were purchased from G. Niswender (Colorado State University, United States). The 11KT antibody was a gift from Y. Nagahama (Japan), and the 17a, 20P-P antibody was a gift from A. P. Scott (United Kingdom). The stan- dards for the steroids were purchased from Stera- loids, Inc. (USA). Collection of samples The sampling times for fish were chosen to coincide with the period of gonadal recrudescence and spawn- ing in English sole (Ketchen, 1947; Harry, 1959; Johnson et al., 1991). Maturing female (>200 mm total length (TL)) and male sole ( >190 mm TL) were collected from Tulalip Bay (October through Janu- ary) and University Point (January through March) by otter trawl during the 1992-93 (October- March) and 1993-94 (October) spawning cycles. Ad- ditional sole were collected monthly at Pilot Point from February 1994 to December 1995. Tulalip Bay and Pilot Point (residential sites) were chosen for their large populations of English sole, and Univer- sity Point was chosen because sole migrate to this area for spawning (Johnson et al., 1991; Collier et al., 1992). Fish were kept in holding tanks aboard the re- search vessel for approximately one hour before sam- pling was carried out. The sole were weighed (to the nearest g) and measured (to the nearest mm), and blood ( 1 mL) was collected from the caudal vein with a heparinized syringe. Blood samples were centri- fuged for 10 minutes at 800 x g, and plasma was col- lected and stored in triplicate bullet vials (250 pL each vial) for subsequent analysis of reproductive steroids (females: E2,T, and 17a, 20(3-P; males: 11KT, and T). Immediately following blood collection, the fish were sacrificed by severing the spinal cord. The gonads were removed and weighed, and samples of gonadal tissue were collected and preserved in Dietrich’s fluid (Gray, 1954) for histological exami- nation. The abdominal contents were removed and the gutted carcass was weighed. Plasma samples were frozen and stored at -80 C until steroid analy- ses could be performed. Sample analyses Plasma T, E2, and 11KT levels were determined by radioimmunoassay (RIA) as described by Sower and Schreck (1982), and 17a, 20(3-P levels were deter- mined by RIA as described by Scott et al. ( 1982). Each RIA (E2, 11KT, and T, and 17a, 20(3-P) consisted of duplicates of standards (ranging from 0 to 2.0 ng / mL), samples, and quality control samples (pooled plasma from male or female English sole). Detection limits of the assays (20-80% binding range of the standard curve) were 0.006-0.3 ng/mL plasma for E2; 0.009-0.4 ng/mL for 11KT; 0.002-0.13 ng/mL for T; and 0.003-0.08 ng/mL plasma for 17a, 20(3-P. For each assay, samples with steroid concentrations out- side the detection limits of the assay were retested by adjusting the sample volume. All samples were corrected for dilution factor. Mean (±SE) of quality control samples were 4.4 ±0.55 ng/mL for E2 (n=15); 0.89 ±0.12 ng/mL for 11KT (n= 6); and 0.29 ±0.04 ng/ mL for T (rc=16). The steroid 17a, 20(3-P was not de- tected in quality control samples. Gonadal tissues collected for histological exami- nation were embedded in paraffin, sectioned, stained with Harris hematoxylin and eosin-phloxine (Luna, 1968) and examined by light microscopy. Ovarian developmental stage (Table 1) was classified accord- Sol et a I.: Gonadal development and changes in plasma reproductive steroids in Pleuronectes vetulus 861 Table 1 Classification scheme for female ovarian developmental stages, modified from histological criteria outlined in Johnson et al. (1991). Stage Features Regressed Primary and secondary oocytes Previtellogenic Vacuolated secondary oocytes with clear peripheral vacuoles; zonal ra- diata present Vitellogenic Yolked oocytes present Vitellogenic with Yolked globules coalescing; hydrated hydrated oocytes oocytes present, no postovulatory fol- licles (POFs) Spawning Hydrated oocytes with POFs Spawned out Many POFs visible, yolked oocytes undergoing resorption, and inflam- matory infiltrate, beta or gamma atretic follicles or macrophage aggre- gates (or both) generally present ing to the criteria modified from Johnson et al. (1991). Testicular developmental stage was classified as modified from Billard ( 1992). GSI was calculated by using the formula GSI = (gonad weight (g) /gutted weight (g)) x 100. only spermatogonia; in early recrudescence, primary and secondary spermatocytes begin to appear; in the late recrudescent stage spermatogonia, spermato- cytes (primary and secondary), and spermatids are evident, but no spermatozoa; in early spermiogen- esis primarily spermatocytes (primary and second- ary), spermatids, and early sperm production are evident; in the spawning stage predominately ma- ture sperm fill the tubular space and sperm ducts; in the spawned-out testis, the tubular space and sperm ducts are largely empty and few sperm remain. Size at sexual maturation The size of English sole captured in this study ranged from 212 to 455 mm for females and from 195 to 345 mm total length for males. Histological analyses of the gonads showed that a majority of the female sole <260 mm TL were immature (regressed-previ- tellogenic), and failed to reach vitellogenesis (Table 2). Male sole <230 mm reached spermatogenesis but most did not fully mature and spawn (Table 3). Fe- male sole <260 mm were excluded from further sta- tistical analyses, but no size limit was used for male sole. Proportions of fish undergoing gonadal maturation and spawning Statistical analysis Mean (±SE) GSI and plasma reproductive steroid levels were calculated by month and gonadal matu- ration stage. Analysis of variance (AN OVA) and sub- sequent multiple comparison testing by Fisher’s pro- tected least significant difference (PLSD) test (95% Cl), were used to test for changes in these factors during the reproductive cycle (Dowdy and Wearden, 1991). The data was normalized by log transforma- tion prior to statistical comparisons. Results Classification of testis The testicular developmental stages for male English sole are shown in Figure 1, A-F. The testis consists of tubules of reproductive cells in various stages of development. Testicular development is categorized into six stages: regressed (Fig. 1A), early recrudes- cence (Fig. IB), late recrudescence (Fig. 1C), early spermiogenesis (Fig. ID), spawning (Fig. IE), and spawned out (Fig. IF). Each stage is characterized as follows: in the regressed stage, tubules contain Female Figure 2 shows the proportion of female sole at different ovarian developmental stages by month. Even among adult sole (>260 mm), regressed females were found all year, and the mean proportion of the regressed animals each month was 24% (±11%, SD). Developing females were found as early as July; 14% of the female sole sampled in July were vitellogenic. The proportion of vitellogenic sole generally increased until January (78%), then decreased in the follow- ing months; by March only 11% of the sole sampled were vitellogenic. Spawned out females were found as early as October (1%), and the proportion in- creased through March (31%). Male Figure 3 shows the proportion of male sole at different testicular developmental stages categorized by month of capture. Animals collected in July con- sisted of regressed males (12%) and early develop- ing males (25% each were at early and late recru- descence), and males in early spermiogenesis (38%). Spawning males were found as early as October (16%), and the proportion generally increased until February (91%). By March, 41% were spawning and 36% were spawned out. Unlike females, in which a certain proportion of fish did not develop during the reproductive cycle, almost all males collected did 862 Fishery Bulletin 96(4), 1998 1 1 m ^ '** 4 %> i * . M Mi &fn ft, >#* V- I re / JIBpgBgi w«k '>, " i. » ^ %# ■ .±f> ‘ % S' #' J^§> -» a _ K&l* <*' J Si * %V * * } ***** / . **, #?. *** f* f3E$SS(»si i vS&V\ N&£ V. Figure 1 Micrographs of testicular developmental stages in male English sole: (A) regressed (510x), (B) early recrudescence (560x), (C) late recrudescence (560x), (D) early spermiogenesis (560x), (E) spawning (540x), and (F) spawned out ( 180x). a=spermatogonia, b=primary spermatocytes, c=secondary spermatocytes, d=spermatids, and e=spermatozoa. enter spermiogenesis. However, only a small propor- tion of males <230 mm were actually producing sperm, and none in this size range were collected on the spawning ground. The smallest male caught at University Point measured 238 mm, suggesting that males <230 mm may not migrate to the spawning ground (University Point) (Table 3). Changes in reproductive parameters during gonadal recrudescence Female Figure 4, A and B, shows reproductive pa- rameters in female sole at each developmental stage. Generally, the immature (regressed and previ- tellogenic) sole had low GSI (1.4 ±0.1, 85), and low Sol et a I.: Gonadal development and changes in plasma reproductive steroids in Pleuronectes vetulus 863 Table 2 Length maturation relationship in female English sole collected during months of gonadal recrudescence, October-March. Num- bers represent percentage of the animals in the length class at each ovarian stage. Length (mm) Regressed Previtellogenic Vitellogenic Vitellogenic with hydrated oocytes Spawning Spawned out n <220 50 50 2 221-230 100 1 231-240 100 11 241-250 78 11 11 9 251-260 69 15 15 13 261-270 31 23 31 7.7 7.7 13 271-280 32 16 37 16 19 281-290 44 11 44 18 291-300 25.7 23 40 2.9 8.6 35 301-310 34 13 44 3.1 6.3 32 311-320 33 12 45 3 6.1 33 >320 26.9 16 47 2.9 7 11 172 o 3 ® (0 cn c cn c © © — CL © 100- (22) (8) (50) (74) (77) (9) (55) (62) £ © © o> co © E w .s* 75- m 50“ 25- l 1 1 I 1 1 □ regressed ^ previtellogenic S vitellogenic 0 vitellogenic w/ hydrated oocytes □ spawning _ spawned * out Month of collection Figure 2 Percentage of female English sole at each stage of ovarian devel- opment by collection month. Numbers in parentheses represent animals sampled each month and histologically analyzed for ova- rian stages. plasma reproductive steroid concentrations (E2: 650 ±110 pg/mL plasma, n= 66; T: 80 ±12 pg/mL plasma, « =66). Increases in GSI and reproductive steroids were observed in ani- mals at the previtellogenic stage, and peak levels of GSI (20.0 ±4.8, n= 9) and plasma E2 levels (7000 ±4700, n = 4) were found in vitellogenic sole with hydrated oocytes, whereas peak plasma T levels (2300 ±620, n= 9) were found in spawning females. All reproduc- tive parameters were reduced in spawned out sole (GSI: 2.8 ±0.5, n=31; E2: 280 ±70, n=21; T: 90 ±50, n- 21). 17a, 20(3-P was not detected in female sole. Figure 5, A-C, shows the mean levels of re- productive parameters in female sole at each month of collection. The sole sampled in July had low GSI (2.14 ±0.45, n= 23) and plasma reproductive steroid levels (E2: 710 ±160, n= 26; T: 100±30, n=27), which remained low until the onset of vitellogenesis (September- October for the majority of fish). The highest GSI (12.0 ±2.9, n=10) and plasma steroid lev- els (E2: 12000 ±5000, n- 3; T: 2400 ±1200, n=4) were observed in January. By March, GSI and plasma reproductive steroid levels were reduced to levels comparable to those found in sole sampled in July. In general, fish from Tulalip Bay and Pilot Point had similar levels of reproductive parameters, which were lower than levels in fish from the University Point. The fish at University Point were either in final maturation or were actively spawning, whereas fish from Tulalip Bay and Pilot Point were migrating to spawning areas to undergo final maturation. Male Figure 6, A and B, shows the reproductive parameters in male sole at each developmental stage. The changes in reproductive parameters were simi- lar to those observed in female sole. Generally, GSI (0.60 ±0.08, n=6) and plasma steroid levels (11KT: 1900 ±920 pg/mL plasma, n= 4; T: 160 ±60 pg/mL plasma, n- 4) were low in regressed animals. In- creases in GSI and reproductive steroids were ob- served in animals at the early recrudescence stage and peaked in spawning fish (GSI: 1.6 ±0.1, n=91; 11KT: 5300 ±610, n= 57; T: 830 ±130, n=64). Repro- 864 Fishery Bulletin 96(4), I 998 Table 3 Length maturation relationship in male English sole collected during months of gonadal recrudescence, October-March. bers represent percentage of the animals in the length class at each testicular stage. Mum- Length (mm) Regressed Early recrudescence Late recrudescence Early spermiogenesis Spawning Spawned out n <220 85 15 13 221-230 6.7 13 6.7 60 13 15 231-240 13 6.7 13 33 27 6.7 15 241-250 6.7 3.3 27 60 3.3 30 251-260 5.6 5.6 0 47 39 2.8 36 261-270 2.9 2.9 49 40 5.7 35 271-280 4.8 43 48 4.8 21 281-290 6.7 40 53 15 291-300 36 64 14 301-310 38 50 13 8 311-320 25 50 25 4 >320 10 10 10 70 10 ductive parameters were reduced in spawned out fish (GSI: 0.8 ±0.1, n=8; 11KT: 83 ±5, n= 8; T: 24 ±7, n= 7), and plasma steroid levels in spawned out fish were significantly lower than levels observed in fish at other stages of tes- ticular development. Figure 7, A-C, shows the mean levels of re- productive parameters in male sole at each month of collection. GSI in sole sampled dur- ing July was low (0.89 ±0.18, n=6), increased to a peak in January (2.6 ±0.7, n= 4), and returned to baseline levels in sole sampled in March (0.8 ±0.8, n- 22). Plasma steroid levels were low in sole sampled in July (11KT: 890 ±140, rc = 13; T: 120 ±30, n = 12) even though a few fish expressed milt. The levels increased to a peak in Febru- ary (11KT: 13000 ±2800, n = 6; T: 1800 ±340, n = 14), and decreased in March (11KT: 1300 ±950, n=18; T: 400 ±310, n = 15). Highest levels of plasma T and 11KT were found in fish from University Point where fish had migrated to spawn. A few fish with comparable levels of 1 1KT were found in March from the residential site at Tulalip Bay. Discussion Typical size-at-maturity for English sole is 260 and 210 mm TL for female and male sole, respectively (Garrison and Miller, 1982). Histological analyses of gonads collected in our study showed the smallest group of vitellogenic females to be 240-249 mm TL, but generally fish less than 260 mm TL were imma- ture. Also, among females >260 mm, which were pre- sumably old enough to undergo sexual maturation, approximately 24% were found in the regressed stage even during the peak of the reproductive season. This finding is consistent with findings by Johnson et al. (1988), in which the percentage of vitellogenic fe- males at reference areas during peak recrudescence was about 80%. In contrast, males at much smaller length (i.e. 195 mm) were maturing, but no spawned out males were found at sizes <230 mm, suggesting that males <230 mm may not migrate to the spawn- ing area. Although fish age was not determined in this study, length-at-age relationships determined for Puget Sound English sole (Myers et al., 1993) showed Sof et a l.: Gonadal development and changes in plasma reproductive steroids in Pleuronectes vetulus 865 B o to ffl O 3 ® <6. 3 ■O 0) cn 3 £U Figure 4 Mean +SE of(A) gonadosomatic index (GSI), and (B) plasma steroid con- centrations in female English sole at each stage of oocyte development. * denotes significantly higher levels than levels observed in regressed and spawned out fish (AN OVA, Fisher’s PLSD, 95% Cl). Numbers in parenthe- ses represent animals at each ovarian developmental stage. that females at 260 mm were three years of age, and males at 195 mm were two years of age. Other studies have shown that females first mature at three years of age, whereas males mature at two years of age (Smith, 1936; Harry, 1959). The pattern and timing of ovarian de- velopment observed in our study were similar to previous reports on English sole (Garrison and Miller, 1982; Kruse and Tyler, 1983; Johnson et al., 1991; Fargo and Tyler, 1994). Gonadal recru- descence in Puget Sound English sole began in early fall (September-October), vitellogenesis and spermiogenesis peaked in early winter (December-January), and spawning took place during late winter months (February and March). However, our findings suggest that small propor- tions of animals did enter vitellogenesis early; vitellogenic sole were found as early as July, whereas spawned out sole were found as early as October. In an earlier study that investigated oocyte development in female English sole (Johnson etah, 1991), the earliest spawn- ing females were found in January. This discrepancy may be due to the fact that the number of fish collected in early fall in the first study (Johnson et al., 1991) was small compared with the present study; therefore, early spawning females may not have been observed. Neverthe- less, changes in reproductive parameters (GSI, and plasma sex steroids) were simi- lar to those described by Johnson et al. ( 1991 ) and to the reproductive endocrine cycles described for other species of fe- male flatfish (e.g. Liu et al., 1991; Methven and Grim, 1991; Harmin et al., 1995). Generally, regressed fish had low values for the various reproductive pa- rameters. The onset of vitellogenesis was accompanied by an increase in GSI and plasma sex steroid levels. After spawn- ing, these levels declined to levels similar to those ob- served in regressed animals. Reproductive parameters at the two residential sites (Pilot Point and Tulalip Bay) were similar and supported the findings of Laroche and Richardson (1979) who observed no apparent latitudi- nal trend in the time of spawning in sole from the Or- egon coast. Reproductive parameters in fish from Uni- versity Point, however, were higher because a major- ity of sole at this site were undergoing final matura- tion or spawning (Johnson et al., 1991). It is well established that E2 stimulates the liver to produce vitellogenin (Ng and Idler, 1983), and in this study, as in Johnson et al. (1991), vitellogenesis in female English sole coincided with a rise in plasma E2 concentrations. Plasma concentrations of T, a bio- synthetic precursor to E2 (Kagawa et al., 1982), changed in parallel with the plasma E2 concentra- tions, although plasma T levels were lower than plasma E2 levels at all stages of gonadal develop- ment. Higher plasma T levels were observed in 866 Fishery Bulletin 96(4), 1998 vitellogenic sole compared with regressed sole, and highest plasma T concentrations were observed in spawning sole. Similar observations have been made in winter flounder (Pleuronectes americanus), where concentrations of T have been correlated with oocyte stages characterized by germinal vesicle migration (Truscott et al., 1992b). In salmonids, estrogens and androgens such as T appear to play an important role in regulating production of pituitary GTH-II (Xiong et al., 1993, 1994), which stimulates the pro- duction of progestins, steroids involved in inducing final maturation (Swanson, 1991). This would be con- sistent with the high concentrations of T observed in spawning female English sole and winter flounder. In male teleosts, 11KT is known to be important in stimulating spermatogenesis, whereas T is known to be important in feedback effects on the pituitary (Borg, 1994). T is a biosynthetic precursor to 11KT (Ozon, 1972). As with E2 in female sole, the levels of T and 11KT covaried in male sole; 11KT levels were significantly higher than T levels at all stages of gonadal development. Similar relations between T and 11KT concentrations have been observed in other flatfish species, such as Pacific halibut (Hippoglossus stenolepis) (Liu et al., 1991) and winter flounder (Harmin et al., 1995). These re- sults are consistent with the role of 11-oxygenated androgens as the dominant regulatory androgens in male teleosts, stimulating secondary sexual characters, reproductive behavior, and spermato- genesis (Borg, 1994). In male winter flounder (Harmin et al., 1995) and plaice (Wingfield and Grimm, 1977), repro- ductive parameters (GSI, plasma 11KT and T con- centrations) increased with the onset of seasonal testicular recrudescence, reached a peak in prespawning and spawning males, then decreased in spawned out males. However, in most male te- leosts that have been studied, the levels of plasma 11KT and T peak during the prespawning period rather than during the spawning period (Borg, 1994). For example, the levels of plasma 11KT and T in Atlantic halibut (Hippoglossus hippoglossus) peaked briefly in sperm-producing fish, then de- clined during the spawning period (Methven and Crim, 1991). The reproductive parameters mea- sured in male English sole were similar to the patterns observed in winter flounder and plaice, where steroid concentrations remained high dur- ing the peak spawning period. Interestingly, the reproductive steroid levels observed in spawned out male English sole were significantly lower than levels observed in re- gressed males that were collected in summer and early fall. A pattern somewhat like this was also observed in male winter flounder and Atlantic halibut. In both male winter flounder (Harmin et al., 1995) and Atlantic halibut (Methven and Crim, 1991), the plasma steroid levels were low in spent males but began to increase about two months af- ter the peak spawning period. Although this study did not measure reproductive parameters in the months immediately following the peak spawning period (i.e. April through June), similar changes Sol et al.: Gonadal development and changes in plasma reproductive steroids in Pleuronectes vetulus 867 B Figure 6 Mean ±SE of (A) gonadosomatic index (GSI) and (B) plasma steroid concentrations in male English sole at each stage of test icular develop- ment. “a” denotes significantly higher levels than levels observed in regressed fish; “b” denotes significantly higher levels than levels ob- served in early recrudescence fish; and “c” denotes significantly higher levels than levels observed in spawned out fish I AN OVA, Fisher’s PLSD, 95% Cl). Numbers in parentheses represent animals at each testicular developmental stage. in plasma steroid levels may have occurred in male English sole and would account for the significant difference in plasma steroid levels between spent and regressed males. A few male English sole sampled in July expressed milt even though reproductive steroid levels were low. Release of milt was also observed in winter flounder (Harmin et al., 1995) and Atlantic halibut (Methven et al., 1991) a few months earlier than vi- tellogenesis in females. Moreover, in win- ter flounder (Harmin et al., 1995), milt remained expressible a few months after the spawning season even though steroid levels were low during this time. Histo- logical analyses of English sole testis showed that initiation of spermiogenesis does begin in July in a substantial propor- tion of sole, in spite of their low plasma steroid concentrations. This phenomenon has also been observed in other male fish, but the mechanism accounting for it is not entirely clear. Borg (1994) suggests that concentrations ofT and 11KT in the testis itself may be quite high in early spermato- genesis, sufficient to stimulate spermio- genesis even when plasma concentrations of T and 11KT are low. The early produc- tion of mature sperm may be typical of most northern latitude, temperate zone teleosts in which fully developed sperma- tozoa are formed before winter and stored for discharge in spring (Lofts, 1987). It is unclear which steroid induces final oocyte maturation in English sole. In all teleosts which have been studied, final maturation is triggered by a surge in plasma concentrations of a maturational gonadotropin (GTH-II). GTH-II stimulates gonadal production of progesterones and related compounds, the C21 steroids, which induce final oocyte maturation and production of sperm (Swanson, 1991). In most teleosts, 17a, 20p-P is believed to be a potent MIS (Scott and Canario, 1987, 1992; Inbaraj et al., 1995). However, in flatfish species that have been studied, low or nondetectable levels of 17a, 20j3-P were found ( Howell and Scott, 1989; Truscott et al., 1992b; Mugnier et al., 1995). Similarly, 17a, 20(3-P was not detected in spawning English sole, and in some flatfish species, other steroids have been suggested as the MIS. For example, in winter flounder 17P-hydroxy-5p-andro- stan-3-one and T were correlated with oocyte stages characterized by germinal vesicle migration (Truscott et al., 1992b), whereas in turbot (Scophthalmus maxi- mus L.) 17a, 21p,21-trihydroxy-4-pregnene-3-one-20 has been shown to induce final maturation (Mugnier et al., 1995). Recent studies also suggest that in pla- ice (Pleuronectes platessa) and Dover sole (Solea solea) 17a, 20P-P actually does induce final matura- tion but is metabolized before it can be detected in 868 Fishery Bulletin 96(4), 1998 the bloodstream with conventional RIAs (Scott and Canario, 1992; Inbaraj et al., 1995). Clearly, addi- tional research is needed in this aspect of English sole reproductive biology. In summary, the reproductive endocrine cycle in female and male English sole from two residential sites and a spawning site in Puget Sound, WA, was characterized by measuring reproductive steroid con- centrations and correlating them with collection time S O N D J Month of collection Figure 7 Mean ± SE of (A) gonadosomatic index (GSI), (B) plasma 11-ketotestosterone ( 11KT) concentrations, and (C) plasma testosterone (T) concentrations in male sole from Tulalip Bay, Pilot Point, and University Point at each month of fish collection. Line represents mean levels at each month of fish collection. Numbers in parentheses represent ani- mals sampled each month at each site. and histological assessment of gonadal stage. This study provides the first description of the reproduc- tive endocrine cycle in male English sole and can be used as a baseline tool in evaluating the effects of con- taminants on reproductive endocrine function in male English sole, or in individuals used in mariculture. Acknowledgments We thank Y. Nagahama, G. Niswender, and A. P. Scott for providing antibodies for RIAs, M. J. Willis and T. Lee for processing of the histology slides, P. Plesha and H. Sanborn for coordinating sampling trips, M. S. Myers and C. M. Stehr for reviewing the manu- script, and others at Northwest Fisheries Science Center for assistance with fishing and necropsy. Literature cited Billard, R. 1992. Reproduction in rainbow trout: sex differentiation, dynamics of gametogenesis, biology, and preservation of gametes. Aquaculture 100:263-297. Borg, G. 1994. Androgens in teleost fishes. Comp. Biochem. Physiol. 109:219-245. Casillas, E. D., L. Misitano, Johnson, L. Rhodes, T. Collier, J. Stein, B. McCain, and U. Varanasi. 1991. Inducibility of spawning and reproductive success of female English sole (Parophrys vetulus) from urban and nonurban areas of Puget Sound, Washington. Mar. Environ. Res. 31:99-122. Collier, T. K., L. L. Johnson, C. M. Stehr, M. S. Myers, and J. E. Stein. In press. A comprehensive assessment of the impacts of contaminants on fish from an urban waterway. Mar. Environ. Res. Collier, T. K., J. E. Stein, H. R. Sanborn, T. Horn, M. S. Myers, and U. Varanasi. 1992. Field studies of reproductive success and bioindi- cators of maternal contaminant exposure in English sole (Parophrys vetulus). Sci. Tot. Environ. 116:169-185. Dowdy, S., and S. Weardon. 1991. Statistics for research. John Wiley and Sons, New York, NY, 629 p. Fargo, J., and A.V. Tyler. 1994. Oocyte maturation in Hecate Strait English sole (Pleuronectes vetulus). Fish. Bull. 90(11:189-197. Garrison, K. J., and B. S. Miller. 1982. Review of the early life history of Puget Sound fishes. Univ. Washington, School of Fisheries Research Institute, Seattle, WA, 729 p. Gray, P. 1954. The microtomists’ formulatory and guide. Blakiston, NY, 808 p. Harmin, S. A., L. W. Crim, and M. D. Wiegand. 1995. Plasma sex steroid profiles and the seasonal repro- ductive cycle in male and female winter flounder, Pleuronectes americanus (Walbum). Mar. Biol. 121: 601-610. Sol et a I.: Gonadal development and changes in plasma reproductive steroids in Pleuronectes vetulus 869 Harry, G. Y. 1959. Time of spawning, length at maturity, and fecundity of English, petrale, and Dover sole (Parophrys vetulus, Eopsetta jordani, and Microstomus pacificus, respec- tively). Fish. Comm. Oreg. Res. Briefs 7:5-13. Howell, R., and A. P. Scott. 1989. Ovulation cycles and post-ovulatory deterioration of eggs of the turbot ( Scophthalmus maximus L.). Rapp. Reun. Cons. Int. Explor. Mer 191:21-26. Idler, D. R., Y. P. So, G. L. Fletcher, and J. F. Payne. 1995. Depression of blood levels of reproductive steroids and glucuronides in male winter flounder (Pleuronectes americanus) exposed to small quantities of Hibernia crude, used crankcase oil, oily drilling mud, and harbour sedi- ments in the 4 months prior to spawning in late May- June. In F. W. Goetz and P. Thomas (eds.), Proc. 5th int. symp. reprod. physiol, fish, p.187. Univ. Texas, Austin, TX. Inbaraj, R. M., A. P. Scott, and E. L. M. Vermeirssen. 1995. Monitoring final oocyte maturation in female plaice (Pleuronectes platessa) using RIAs which detect metabo- lites of 17a, 20(3-dihydroxy-4-pregnen-3-one. In F. W. Goetz and P. Thomas (eds.), Proc. 5th int. symp. repro. physiol, fish, p. 311. Univ. Texas, Austin, TX. Jobling, S., D. Sheahan, J. A. Osborne, P. Matthiessen, and J. P. Sumpter. 1996. Inhibition of testicular growth in rainbow trout (Oncorhynchus my kiss) exposed to estrogenic alkylphenolic chemicals. Environ. Toxicol. 15(2):194-202. Johnson, L. L., and E. Casillas. 1991. The use of plasma parameters to predict ovarian maturation stage in English sole Parophrys vetulus Girard. J. Exp. Mar. Biol. Ecol. 151:257-270. Johnson, L., E. Casillas, T. Collier, B. McCain, and U. Varanasi. 1988. Contaminant effects on ovarian maturation in En- glish sole ( Parophrys vetulus) from Puget Sound, Washington. Can. J. Fish. Aquat. Sci. 45:2133-2146. Johnson, L. L., E. Casillas, T. K. Collier, J. E. Stein, and U. Varanasi. 1993. Contaminant effects on reproductive success in se- lected benthic fish species. Mar. Environ. Res. 35:165- 170. Johnson, L. L., E. Casillas, M. S. Myers, L. D. Rhodes, and O. P. Olson. 1991. Patterns of oocyte development and related changes in plasma 17]3-estradiol, vitellogenin, and plasma chemis- try in English sole Parophrys vetulus Girard. J. Exp. Mar. Biol. Ecol. 152:161-185. Kagawa, H., G. Young, and Y. Nagahama. 1982. Estradiol 17-p production in isolated amago salmon (Onchorhynchus rhodurus) ovarian follicles and its stimula- tion by gonadotropins. Gen. Comp. Endocrinol. 47:361-365. Ketchen, K. S. 1947. Studies on lemon sole development and egg production. Fish. Res. Board Can. Prog. Rep. Prog. Rep. Pac. Coast Sta. 73:68-70. Kruse, G. H., and A. V. Tyler. 1983. Simulation of temperature and upwelling effects on the English sole (Parophrys vetulus) spawning season. Can. J. Fish. Aquat. Sci. 40:230-237. Laroche, J. L., and S. L. Richardson. 1979. Winter-spring abundance of larval English sole, Parophrys vetulus , between the Columbia River, and Cape Blanco, Oregon, during 1972-1975 with notes on occur- rences of three other pleuronectids. Estuarine Coastal Mar. Sci. 8(51:455-476. Liu, H., R. Stickney, and W. Dickoff. 1991. Changes in plasma concentrations in adult Pacific halibut, Hippoglossus stenolepis. J. World Aquacult. Soc. 22(l):30-35. Lofts, B. 1987. Testicular function. In D. O. Norris and R. E. Jones (eds.), Hormones and reproduction in fishes, amphibians and reptiles, p. 283-325. Plenum, New York, NY. Luna, L. G. (ed.). 1968. Manual of histologic staining methods of the Armed Forces Institute of Pathology, third edition. McGraw-Hill, New York, NY, 206 p. Methven, D. A., and L. W. Crim. 1991. Seasonal changes in spermatocrit, plasma sex ste- roids, and motility of sperm from Atlantic halibut (Hippoglossus hippoglossus) . In A. P. Scott, J. P. Sumpter, D. E. Kime, and M. S. Rolfe (eds.), Proc. 4th int. symp. reprod. physiol, fish, p. 170. Univ. East Anglia, Sheffield, UK. Mugnier, C., J. L. Gaigmon, E. Lebegue, C. Fauvel, and A. Fostier. 1995. Maturation inducing steroid in turbot ( Scophthalmus maximus L). In F. W. Goetz and P. Thomas (eds.), Proc. 5th int. symp. reprod. physiol, fish, p. 323. Univ. Texas, Austin, TX. Myers, M. S., C. M. Stehr, O. P. Olson, L. L. Johnson, B. B. McCain, S-L. Chan, and U. Varanasi. 1993. National status and trends program, national benthic surveillance project: Pacific Coast. Fish histopathology and relationships between toxicopathic lesions and exposures to chemical contaminants for cycles I to V ( 1984-88). U.S. Dep. Commerce, NOAA Tech. Memo. NMFS-NWFSC-6, 160 p. 1994. Relationship between toxicopathic hepatic lesions and exposure to chemical contaminants in English sole (Pleuronectes vetulus ), starry flounder (Platichthys stellatus), and white croaker (Genyonemus lineatus) from selected marine sites on the Pacific Coast, USA. Environ. Health Perspect. 102(2):200-215. Ng, T. G., and D. R. Idler. 1983. Yolk formation and differentiation in teleost fishes. In W. S. Hoar, D. J. Randall, and E. M. Donaldson (eds.). Fish physiology, vol. IX, Reproduction pt. A, p. 373- 403. Academic Press, New York, NY. Nimrod, A. C., and W. H. Benson. 1996. Environmental estrogenic effects of alkylphenol ethoxylates. Crit. Rev. Toxicol. 26(3):335-364. Ozon, T. 1972. Androgens in fishes, amphibians, reptiles and birds. In D. R. Idler (ed.), Steroids in nonmammalian vertebrates, p. 329-389. Academic Press, New York, NY. Scott, A. P., and A.V. M. Canario. 1987. Status of oocyte maturation-inducing steroids in teleosts. In D. R. Idler, L. W. Crim, and J. M. Walsh, (eds.), Proc. 3th int. symp. repro. physiol, fish, p. 224-234. Ma- rine Sciences Research Laboratory, St. John’s, Newfound- land, Canada. 1992. 17a, 20P-dihydroxy-4-pregnen-3-one 20-sulfate: a major new metabolite of the teleost oocyte maturation-in- ducing steroid. Gen. Comp. Endocrinol. 85:91-100. Scott, A. P., E. L. Sheldrick, and A. P. Flint. 1982. Measurement of 17a, 20P-dihydroxy-4-pregnen-3-one in plasma of trout (Salmo gairdneri Richardson): seasonal changes and response to salmon pituitary extract. Gen. Comp. Endocrinol. 46:444-451. 870 Fishery Bulletin 96(4), 1998 Smith, R.T. 1936. Report on the Puget Sound otter trawl investigations. State of Washington. Dep. Fisheries. Biological Report 36B, 42 p. Sower, S. A., and C. B. Schreck. 1982. Steroid and thyroid hormones during sexual matu- ration of coho salmon (Oncorhynchus kisutch) in seawater or freshwater. Gen. Comp. Endocrinol. 47:42-53. Swanson, P. 1991. Salmon gonadotropins: reconciling old and new ideas. In A. P. Scott, J. P. Sumpter, D. E. Kime, and M. S. Rolfe (eds ), Proc. 4th int. symp. repro. physiol, fish, p. 2— 7. Univ. East Anglia, Sheffield, UK. Truscott, B., D. R. Idler, and G. L. Fletcher. 1992a. Alteration of reproductive steroids of male winter flounder (Pleuronectes americanus) chronically exposed to low levels of crude oil in sediments. Can. J. Fish. Aquat. Sci. 49:2190-2195. Truscott, B., Y. So, J. Nagler, and D. Idler. 1992b. Steroids involved with final oocyte maturation in the winter flounder. J. Steroid Biochem. Mol. Biol. 42:351-356. Wingfield, J. C., and A. S. Grimm. 1977. Seasonal changes in plasma cortisol, testosterone and oestradiol-17(3 in the plaice, Pleuronectes platessa L. Gen. Comp. Endocrinol. 31:1—11. Xiong, F., R. Chin, Z-Y. Gong, K. Suzuki, R. Kitching, S. Majumdar-Sonnydal, H. Elsholtz, and C.L. Hew. 1993. Control of salmon pituitary hormone gene expression. Fish Physiol. Biochem. 11:63-70. Xiong, F., D. Liu, Y. Le Drean, H. Elsholtz, and C. L. Hew. 1994. Differential recruitment of steroid hormone response elements may dictate the expression of the pituitary gona- dotropin 11(3 subunit gene during salmon maturation. Mol. Endocrinol. 8:782-793. 871 Abstract.-Aiong the continental slope of the eastern Bering Sea, two species of Careproctus snailfishes de- posit eggs within the branchial cham- bers of the commercially important golden king crab ( Lithodes aequi- spinus). The larger of the two species is the pink snailfish (C. furcellus) ; the smaller species is an undescribed spe- cies referred to as the red snailfish. According to 1982 trawl survey data, incidence of snailfish eggs and larvae within crab branchial chambers in- creases with carapace length (CL) and is greater for male than female crabs. A logistic model fitted to the incidence data predicts that a 140-mm-CL male will have an incidence of 0.52, approxi- mately 1.9 times greater than the inci- dence for a 100-mm male. Incidence for a 100-mm-CL male is approximately 1.9 times greater than that for a female of equal size. On the basis of develop- mental stages of embryos carried by female golden king crab and the devel- opmental stages of snailfish embryos within a female’s branchial chambers, snailfish appear to deposit eggs prefer- entially in crabs that are early in their molt cycle. The presence of the egg masses results in gill compression, lo- calized necrosis of gill tissue and, in extreme cases, total loss of gill tissue on one side of the body. For crabs of commercial size, the presence of eggs and larvae increases mortality within the holds of fishing vessels by 35%. The current incidence of eggs and larvae in commercial sized males, however, is so low that the effect on the commercial fishery is considered to be small. Manuscript accepted 7 January 1998. Fish. Bull. 96:871-884 (1998). Parasitism of the golden king crab, Lithodes aequispinus, by two species of snailfish, genus Careproctus David A. Somerton Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 7600 Sand Point Way NE, Seattle, Washington 98115 E-mail address: David.Somerton@noaa.gov William Donaldson Alaska Department of Fish and Game 2 1 1 Mission Road Kodiak, Alaska 996 1 5 Liparid snailfish in the genus Careproctus extrude eggs through an anteriorly positioned ovipositor into the branchial chambers of large lithodid crabs. This relationship was first described, independently, by Rass (1950) and Vinogradov (1950) for C. sinensis and Para- lithodes camtschatica in the north- west Pacific. Subsequent studies, however, have shown that addi- tional species both within Para- lithodes as well as within three other lithodid genera (Lithodes, Lopholithodes and Paralomis ), also serve as hosts to various Careproctus species and that the association ap- pears to be widespread according to reports from the North and South Pacific and South Atlantic (Hunter, 1969; Parrish, 1972; Peden and Corbett, 1973; Anderson and Cailliet, 1974; Balbontin et ah, 1979; Melville- Smith and Louw, 1987; and Love and Shirley, 1993). The advantages that snailfish gain by placing their eggs within an en- closed, constantly aerated space have been recognized since the association between Careproctus and lithodid crabs was discovered (Rass, 1950). The disadvantages to the crab, how- ever, remain equivocal, with reports of no obvious damage (Hunter, 1969; Parrish, 1972), minor gill compres- sion by the egg mass (Anderson and Cailliet, 1974; Melville-Smith and Louw 1987), and gill bleeding (Love and Shirley, 1993). Knowledge of the disadvantages is important because most affected lithodid species support commercial fisheries. In this paper, we report on the deposition of egg masses by two spe- cies of Careproctus into the bran- chial chambers of the golden king crab ( Lithodes aequispinus), a large, commercially harvested lithodid occurring along the continental slope of the North Pacific (Somerton and Otto, 1986). The identity of the two Careproctus species is uncer- tain because the taxonomy of the family Liparididae is problematic and incomplete (Allen and Smith, 1988). The larger species (Fig. 1), henceforth called the pink snailfish, is certainly within the C. melanurus group of morphologically similar species (Allen and Smith, 1988) and is likely to be C. furcellus which has been observed throughout the Bering Sea and Aleutian Islands (Kido, 1988). The smaller species (Fig. 1), henceforth called the red snailfish, is tentatively identified as a member of the C. mederi group described by Kido (1988), who re- 872 Fishery Bulletin 96(4), 1998 Figure 1 (Upper) pink snailfish ( Careproctus furcellus ); (lower) red snailfish (C. species). Both specimens are ripe females with their ovipositors extended. Note that the red snailfish has extremely long lobes on the pectoral fins that, in the picture, have been positioned in front of the body for clarity. ferred to it as Careproctus species B.1 Here we first 1 Stein, D. 1996. National Marine Fisheries Service, 1315 East- West Hwy., Silver Springs, MD, 20910. Personal commun. examine the characteristics of this association, then consider how Careproctus snailfish choose their host crabs and what the consequences may be to the crabs. Somerton and Donaldson: Parasitism of Lithodes aequispinus by two species of Careproctus 873 Figure 2 64°N 62° 60° 58° 56° 54° 52° 50° Collecting sites for the 1982 trawl survey (shaded area) and the 1996 commercial catch sampling areas in the Bering Sea and Aleutian Islands (unshaded areas). Materials and methods Golden king crab were sampled 1 1 Julv-25 August 1982 aboard the FV Rujin Maru no. 8 while it was conducting a bottom trawl survey of the benthic re- sources along the continental slope of the eastern Bering Sea (Fig. 2). All golden king crab were first sorted to sex, then measured for carapace length (CL) from the base of the eye to the posterior midline of the carapace. Females were evaluated for maturity, on the basis of relative width of their abdomens, and if mature, their pleopods were examined to deter- mine whether they were carrying uneyed embryos, eyed embryos, or remnants of recently hatched em- bryos (egg cases and funiculi), or whether they were bare. During this sampling, the senior author discovered that many of the crabs contained Careproctus- like egg masses or larvae within their branchial cham- bers. Further inspection of the egg masses and lar- vae revealed that there appeared to be two distinct types, one with egg diameters or larvae lengths con- siderably larger than the other. Several species of Careproctus were also captured in the trawls but two species, pink and red snailfish, predominated. Gentle pressure applied to the abdomens of some individu- als of both species resulted in eggs being extruded through the ovipositor (Fig. 1) or milt being extruded out the genital papilla (Stein, 1980). The eggs of pink snailfish were visibly larger than those of red snailfish and the egg sizes of both species clearly matched the egg sizes found within the crabs. Con- sequently, starting on 27 July, 48 consecutive trawl hauls were sampled as follows for Careproctus spp. and their eggs and larvae. All golden king crabs were examined for the pres- ence of Careproctus egg masses or larvae clusters by removing their carapaces. Each mass was evaluated to type (uneyed embryos, eyed embryos, or larvae) and species (large eggs and larvae were assigned to pink snailfish; small eggs and larvae were assigned to red snailfish), and for location (i.e. which side of the crab it was obtained). Snailfish were first sorted to species, then externally sexed, if possible, by the presence of either an ovipositor or a genital papilla. Snailfish were then gently squeezed about the abdo- men and classified as ripe if either eggs or milt were extruded. Thus, the sex could be either known or unknown and, if known, the individual could be ei- ther ripe or unripe. The 1996 incidence of snailfish eggs and larvae within golden king crab were assessed by Alaska Department of Fish and Game at-sea observers and port samplers during the golden king crab fishing season in the Aleutian Islands and eastern Bering Sea. Female and small (<135 mm-CL) male golden 874 Fishery Bulletin 96(4), 1998 king crab, which are not retained by the fishery, were sampled aboard commercial vessels by occasionally collecting all individuals from one randomly chosen trap. Sampled crabs were first measured for cara- pace length, then examined for the presence of snailfish eggs and larvae, without regard to species, by removing their carapaces. Legal-size male crab were sampled from the commercial catches when they were landed at the processing plant. Catches were first separated into groups of live and dead crabs, then counted and subsampled by examining up to 100 individuals from each group for the pres- ence of snailfish eggs. Port samples were collected for crabs caught either in the Bering Sea (three land- ings) or the Aleutian Islands (six landings; Fig. 2), but at-sea samples were collected only in the Aleu- tian Islands. Thirteen golden king crabs found, during commer- cial sampling, to contain snailfish eggs or larvae were frozen whole, with the eggs or larvae left in place for later processing. Eighteen of 23 egg and larvae samples were nonhatched egg masses that were placed in water overnight to fully rehydrate, then weighed to the nearest 1 mg. Displacement volume of the egg masses (mL) was measured in a gradu- ated cylinder. Egg masses were then teased apart and the total number of eggs counted. Diameters of approximately 200 eggs randomly chosen from each egg mass were measured to the nearest 0.03 mm with an optical micrometer. Volumes of the gills on both the affected and unaffected side of the crab were measured by displacement after they were excised with a scalpel at their bases. Volume of the empty branchial chamber on one side of the crab was then measured by securing the carapace in its proper place on the crab and then injecting Polycel insulating foam into the branchial cavity through a small hole. After any excess was removed, the volumes of the foam casts were then measured by displacement to the nearest milliliter. The 23 egg and larvae samples were then subjected to restriction fragment length polymorphism (RFLP) analysis (Dowling et al., 1996) by using six restriction enzymes to determine spe- cies identity.2 Species identification patterns were established by using the RFLP analysis on six Careproctus rastrinus, eight C. furcellus, and one C. cypselurus collected in crab pots during commer- cial sampling. The incidence (e.g. probability of occurrence) of snailfish eggs and larvae as a function of several pre- dictor variables was described with a generalized lin- ear model, with a logit link function and a binomial 2 Lopez, A. 1996. Marine Molecular Biological Laboratory, Uni- versity of Washington, Seattle, WA. Personal commun. variance function (Venables and Ripley, 1994). For each crab, model inputs included a binary response variable (presence or absence of eggs or larvae). For the 1982 survey data, predictors included two con- tinuous variables (carapace length and depth) and a discrete variable (sex). For the 1996 port sampling data, predictors included two discrete variables (area and whether the crab was alive or dead). Model se- lection was accomplished iteratively by first testing the significance of each interaction term with a like- lihood ratio test, then by discarding the least signifi- cant term. This process was then repeated on each main effect and any associated interaction terms. Differences in the depth distributions of male and female crab and of spawning and nonspawning snailfish were tested by first calculating a Cramer- von Mises statistic from catch-per-tow and depth data, then by determining the significance of the sta- tistic with randomization (Syrjala, 1996). Association between the developmental stage of eggs carried by a female golden king crab and the developmental stage of snailfish eggs within a crab’s branchial chambers was tested with a Spearman’s rank order correlation coefficient. Because king crabs extrude eggs onto their pleopods soon after molting (Powell and Nickerson, 1965; Sloan, 1985), the stage of embryonic development is a crude measure of the relative time since molting. Positive correlation be- tween the developmental stages of crab and snailfish embryos would indicate that snailfish preferentially choose newly molted crabs as hosts. To simplify in- terpretation, 3 of 26 females were not used because they contained more than a single egg mass or lar- vae cluster. Embryonic developmental stages for snailfish (uneyed, eyed, and larval) and crabs (uneyed, eyed, and hatched) were considered as or- dered variables and were assigned numeric codes (i.e. 1, 2, 3). Mature female crabs without any obvious egg remnants attached to the pleopods were grouped with females carrying hatched eggs because egg rem- nants may have been missed in the macroscopic ex- amination performed at-sea. The mortality rate of male crabs in the holding tanks of commercial vessels was estimated separately for crabs with and without snailfish eggs and larvae on the basis of 1996 port sampling data for commer- cial crab landings. For each landing, the total num- ber of live crabs that were infested was estimated by multiplying the number of live crabs landed by the infested proportion in the catch subsample. The to- tal number of infested dead crabs was estimated simi- larly. The mortality rate of infested and uninfested crabs was then estimated as the number dying di- vided by the total in each category. Average mortality rate was then estimated as the mean over all landings. Somerton and Donaldson: Parasitism of Lithodes aequispinus by two species of Careproctus 875 Results Depth distributions of golden king crabs and snaiffish The 43 hauls sampled on the 1982 survey were ran- domly distributed between 192 and 900 m. Over this range, golden king crab catch-per-tow (Fig. 3) and carapace length (Fig. 4) declined with increasing depth. Male and female crab occurred at significantly different depths (randomization test, P=0. 005), with the median depth of males (203 m) being less than that of females (336 m). Pink snailfish abundance declined with depth; most of the population occurred at depths <450 m (Fig. 51. Conversely, red snailfish abundance in- creased with depth; most of the population occurred at depths >450 m. (Fig. 5). Red snailfish was approxi- mately six times more abundant, averaged over the entire survey area, than pink snailfish (mean catch per haul: red snailfish=4.57, pink snailfish=0.75, to- tal number of both species=226). As a result, red snailfish were predominate at depths as shallow as 350 m. The proportion of the individuals that were ripe (both sexes combined) did not differ significantly between species (chi-square test, df=l, P=0.341). Ripe pink snailfish (both sexes combined) were found at significantly shallower depths than unripe individuals (randomization test, P400 m (Fig. 8). In con- trast to this, red snailfish masses increased in abun- dance with depth until reaching a peak at between 300-400 m, then declined until none were encoun- tered at depths >600 m. Despite the difference in depth distribution for each species, the species were combined to simplify analysis. Because the combined incidence of eggs and larvae increased with depth to a peak at 250 m, then subsequently declined (Fig. 8), the depth effect in the model included a quadratic term. The best fit of a logistic model to the combined incidence data indicated that sex (P<0.001), length (P=0.003), and depth (P<0. 001) were all highly significant (Table 2 ). The fitted model predicts 1 ) incidence increases with size (Fig. 9), 2) incidence is greater for males them fe- males (Fig. 9), and 3) incidence is greater at middepth (i.e. significant quadratic depth effect, Table 2). The best fit of a logistic model to the 1996 inci- dence in commercial males indicated that area (P<0.001) and whether the crabs were landed dead or alive (P=0.046) were both significant (Table 2). Incidence in the Bering Sea was over three times greater than in the Aleutian Islands. Incidence in dead crabs was 1.9 times greater than in live crabs in the Bering Sea and 1.6 times greater in the Aleu- tian Islands (Table 3). Incidences in live commercial males from the Aleutian Islands were 2.5 times greater than in sublegal males and 12.5 times greater than in females (Table 3). Incidence in 1996 of live commercial males in the Bering Sea was less than 20% of the incidence in 1982 commercial size males Somerton and Donaldson: Parasitism of Lithodes aequispinus by two species of Careproctus 877 Figure 6 (Upper) A male golden king crab with part of the carapace cut away to show two Careproctus egg masses in different stages of development. (Lower) An early uneyed egg mass (right), a late un- eyed egg mass (center), and a larvae cluster (left) taken from a single golden king crab. (CL>135 mm; Table 3). The probability of a male dying in the holding tank of a commercial vessel was 0.035 if the crab contained snailfish eggs and larvae and 0.025 if the branchial chambers were empty, in- dicating that snailfish infestion increases the hold- ing mortality of crabs. Association between crab and snailfish embryonic development stages Mature female golden king crab carried snailfish embryos in all three stages of development (uneyed, eyed, and larval) only when they themselves were 878 Fishery Bulletin 96(4), 1 998 nj E CD OJ LU 60 50 40 30 20 10 100 150 200 250 300 Branchial volume (mL) Figure 7 Volume of single egg masses as a function of the volume of the branchial chambers of the host crab. ro E cr> O) a> 200 400 600 800 200 400 600 800 Depth (m) Figure 8 Number of egg masses by 100-m depth intervals for pink snailfish (solid line) and red snailfish (dotted line, up- per). Incidence, or proportion of all crabs having snailfish eggs or larvae of either species, by 100-m depth interval (lower). Table 2 Summary statistics of the fitted logistic models. Model 1 (1982 survey data) logit( incidence) = intercept + sex + length h depth + depth2 + sex*depth + sex x depth2 Coefficient Value Intercept -8.6724639 Sex -2.2823274 Length 0.0261672 Depth 0.0310400 Depth2 -0.0000432 Sex x depth 0.0187658 Sex x depth2 -0.0000512 Significance tests of main effects Effect Likelihood ratio df Probability Length 8.87 1 0.003 Sex 25.02 3 <0.001 Depth 20.92 4 <0.001 Model 2 (1996 port sampling of commercial catch) logit( incidence) = intercept + area + dead or alive Coefficient Value Intercept -2.744489 Area -0.667367 Dead or alive -0.300833 Significance tests of main effects Effect Likelihood ratio df Probability Area 22.39 1 <0.001 Dead or alive 3.96 1 0.046 carrying uneyed embryos (Table 4). As crab embryos became more developed, so too did the snailfish em- bryos. Thus, crabs with eyed embryos were not found with uneyed snailfish embryos, and crabs with hatched embryos were not found with uneyed or eyed snailfish embryos. Considering the three develop- ment stages of both species as an ordered random variable, the crab developmental stages were highly correlated (rho=0.68, P=0.005) with the snailfish developmental stages. Genetic determination of species identity All 14 of the fish samples, but only 10 of the 23 egg and larvae samples, collected by commercial fishery observers were sufficiently well preserved to allow Somerton and Donaldson. Parasitism of Lithodes aequispinus by two species of Careproctus 879 Tabie 3 Incidence of snailfish (Careproctus spp. ) egg clusters within golden king crab (Lithodes aequispinus). Included are the total landed catch in numbers of live and dead crab, number of crab in each category examined, and the number with egg masses. Sampling type Area Condition Catch Crabs examined Crabs with fish eggs Percent infested Incidence in male crabs >135 mm CL 1996 port Bering Sea live 12788 302 25 8.3 dead 351 44 7 15.9 Aleutian Is. live 46206 603 15 2.5 dead 1508 343 14 4.1 1982 survey Bering Sea live 55 24 43.6 Incidence in female crabs and male crabs <135 mm CL 1996 at-sea Aleutian Is. male 1641 12 0.7 female 1631 4 0.2 1982 survey Bering Sea male 183 44 24.0 female 274 26 9.5 40 60 80 100 120 140 160 Carapace length (mm) Figure 9 Incidence of Careproctus eggs and larvae, predicted by the general- ized linear model, for male (solid line) and female (dotted line) golden king crab as a function of carapace length. DNA amplification. Careproctus furcellus, C. rastrinus, and C. cypselurus fish samples could be readily distinguished in the RFLP analysis. Of the 10 usable egg and larvae samples, two were identified as C. furcellus and eight clearly differed from all of the fish samples. Effect of the egg masses on gill function The presence of egg masses within the bran- chial chamber of a crab was associated with three distinct pathological conditions of the gills. First, egg masses compressed the gills so strongly that a distinct impression of the egg mass was left on the gill surface (Fig. 10, upper, p. 880). Second, gills on the infested side of a golden king crab were often darker in color than those on the opposing side (Fig 10, lower, p. 880). Third, gills on the infested side often had areas of blackened, necrotic tissue (Fig. 10, p. 881). Several gills with necrotic tissue were examined microscopically after standard histological preparation.3 A stage- wise progression of tissue damage was evi- dent, beginning as small lesions with hemo- cyte encapsulations. As the lesions increased in size, they were more likely to be melanized, indicating a chronic condition, and often oc- cluded one or more gill lamellae. Small (occlud- ing one or more lamellae) to medium size (occluding the gill stem) lesions typically possessed new cuticle 3 Morado, F. 1996. Alaska Fisheries Science Center, 7600 Sand Point Way NE, Seattle, WA. Personal commun. and normal respiratory epithelium underneath the af- fected area, indicating that the gill tissue would be re- generated during the next molt. In extreme cases, all of the gills on the affected side of the body were re- duced to blackened stubs (Fig. 10, p. 881). 880 Fishery Bulletin 96(4), 1 998 Figure 1 0 Pathological conditions associated with the presence of a snailfish egg mass in the branchial chamber of a golden king crab. Compression of the left gills and impression of individual eggs on the gill surface (upper). Blackened necrotic tissue on the right gills (lower). Complete loss of left gills (facing page). The fifth pereiopods, which act as gill cleaning appendages, are shown in their folded state at the posterior junction of the carapace and abdomen (lower). Discussion The branchial chambers of golden king crab provide Careproctus embryos and larvae with a well aerated environment and protection from predators. This advantage would be jeopardized if they were ejected prematurely by crab molting. To minimize the risk of this happening, snailfish must either have the Somerton and Donaldson: Parasitism of Lithodes aequispinus by two species of Careproctus 881 ability to choose host crabs that are early in their molt cycle or that have embryonic and larval resi- dence times that are short in relation to the intermolt period of the crab. Evidence that snailfish have both attributes is provided by the association that we found between the embryonic developmental stages of crabs and those of the snailfish found wdthin the crabs. Preference for newly molted crabs is indicated by the observed presence of young (uneved) snailfish embryos in female golden king crab carrying uneyed embryos but not in those carrying eyed embryos or remnants of hatched larvae (Table 4), because king crabs deposit their eggs soon after molting. Snailfish may locate newly molted crabs olfactorally, by de- tecting minute traces of molting hormones in the same way that some male crabs locate premolt fe- males (Ryan, 1966). A relatively short embryonic and larval residence time, compared with the intermolt period of the crab, is indicated by the presence of late snailfish devel- opmental stages (eyed embryos and larvae) in newly molted (uneyed) female golden king crab (Table 4). The length of the embryonic periods for pink and red snailfish are unknown, but the large size of the eggs and low temperatures of the water in which they in- cubate (average bottom temperature on the 1982 survey was 3.8°C) indicate that the embryonic peri- ods are probably quite long. Incubation times are Table 4 Number of female golden king crab observed categorized by the development stage of their own embryos and the snailfish embryos within their branchial chambers. Snailfish stages Crab stages Uneyed Eyed Larvae Uneyed 9 4 1 Eyed 0 1 0 Hatched 0 0 3 greater in colder water and for fish with larger eggs (Pauley and Pullin, 1988), but there are no studies applicable to a large egg-producing deep slope spe- cies such as Careproctus. If chinook salmon {Onco- rhynchus tshawytscha) are used as an approximate model for a relatively large egg, cold water species, then 4.5 months would be required for hatching at 3.8°C. (Alderdice and Velsen, 1978). Although this rate of embryonic development may be long for a fish, it is considerably shorter than that of golden king crab which perhaps exceeds 1 year (Somerton and Otto, 1986). If snailfish can preferentially choose newly molted crabs as hosts, then their spawning season must oc- cur at the same time as the molting season of the 882 Fishery Bulletin 96(4), 1998 crab. However, the variety of snailfish embryonic stages that occurred during the 1982 sampling (Table 1), coupled with the probable slow develop- mental rate, indicates that pink and red snailfish have either a protracted spawning season, or per- haps lack spawning seasonality. Previous studies of snailfish reproduction have reported that aseasonal spawning was typical of abyssal and slope snailfishes, although C. malanurus, a close relative to pink snailfish, has a seasonal peak in spawning off Or- egon (Stein, 1980). The likelihood for preferential selection of early molt-stage crabs is not necessarily diminished by aseasonal spawning in red and pink snailfish because golden king crab also have aseasonal reproduction and molting (Somerton and Otto, 1986). The incidence of snailfish eggs and larvae was greater in male than in female crabs and increased with crab size in both the 1982 survey (Table 2; Fig. 9) and the 1996 commercial sampling (Table 3). The preference for male crabs as hosts over females is quite pronounced. For example, at the median 1982 sampling depth, a 100-mm male is 1.9 times more likely to contain eggs and larvae than an equal-size female (Fig. 9). As in our study, male Lithodes tropicalis had a higher incidence of snailfish eggs than females (Melville-Smith and Louw, 1986), but the apparent sex selection was attributed to size se- lection and to a large sexual dimorphism in crab size. In our case, we believe sex itself is important in host choice because sex was a significant predictor of in- cidence even when size and depth effects were in- cluded in the model. Such sex selection may not be universal, however, because a previous study of L. aequispinus reported that incidence was higher in females (Love and Shirley, 1993). Why a preference for males should occur is not obvious. One possible explanation, based on aquarium observations (Love and Shirley, 1993), is that female golden king crab aggressively defend the embryos attached to their pleopods. Perhaps this aggressiveness can discour- age the attempts of a snailfish to extrude her eggs into the branchial chambers of the crab. The apparent preference for large crabs is also quite pronounced. For example, at the median 1982 sampling depth, a 140-mm-CL male is 1.9 times more likely to contain snailfish eggs or larvae than a 100- mm-CL male (Fig. 9). The apparent preference of snailfish for large crabs is likely due to two distinct attributes associated with size in lithodid crabs. First, large king crabs molt less frequently than small king crabs. Although the molting frequency of golden king crabs is unknown, for male red king crabs ( Para - lithodes camtschaticus ), molting frequency dimin- ishes continuously with increasing age (McCaughran and Powell, 1977). Since a lower molt frequency would result in a lower probability of premature re- lease of eggs and larvae, it is an advantageous fea- ture for a perspective host to have. Second, larger crabs have larger branchial chambers to contain snailfish eggs. In our case, egg mass volume increased with branchial volume (Fig. 10), indicating that the size of an egg mass is limited by the size of the bran- chial chamber. This increase, however, diminished with crab size, indicating that in large crabs the size of an egg mass may be determined more by snailfish fecundity than by the availability of space. It is not clear what effect the space limitation would have on the searching behavior of snailfish. If a female snailfish is capable of partitioning a batch of ripe eggs among several spawning events, then she might be able to reduce the problem of space limitation by depositing eggs in several crabs. If instead female snailfish must deposit their entire batch in one spawning event, then considerably more searching would be required to find a crab with sufficient vol- ume. To help provide some indication of whether a female partitions a batch of eggs, we measured egg batch volume for a single 40-cm pink snailfish, con- sidering only eggs that were free in the ovarian lu- men. Because the measured volume (48 mL) is about equal to the median volume of the egg masses found in crabs (Fig. 7), it is possible that a female spawns an entire batch in one event. If this is true, then a female would have to determine, perhaps by prob- ing with her ovipositor, whether a prospective host has sufficient branchial volume for her batch of eggs. Besides size and sex, other aspects of crab host choice by snailfish have been postulated. Melville- Smith and Louw (1987) suggested that Careproctus might preferentially choose only one side of host L. tropicalus to deposit egg masses. In our case, nei- ther pink snailfish nor red snailfish chose one side of the crab over the other (binomial test, red snailfish, P= 0.39, pink snailfish, P=0.76). Love and Shirley (1993) and Melville-Smith and Louw (1987) sug- gested, after finding no crabs with egg and larvae masses in both branchial chambers, that snailfish are inhibited from depositing egg masses in the un- affected branchial chamber of previously infested crabs, presumably to reduce host mortality. In our case, this feature was not true because of 23 crabs that had at least two egg or larvae masses, 13 had masses on both sides. Because large crabs occur in shallower waters than small crabs and males occur in shallower waters than females, the preferred hosts of both snailfish species occur in the shallower portion of the 1982 depth range. Pink snailfish occur at approximately the same depths as their preferred hosts, but red Somerton and Donaldson: Parasitism of Lithodes aequispinus by two species of Careproctus 883 snailfish live in deeper waters and must migrate into shallower water to find suitable hosts. For red snailfish, a spawning migration is apparent by the shallower peak in the depth distribution of its egg masses (350 m; Fig. 8) compared with the depth dis- tribution of the fish themselves (650 m; Fig. 5). For pink snailfish, by comparison, the depth distributions of the egg masses and fish nearly coincide. If the spawning migration is undertaken only by ripe fish, then a difference in depth distribution between ripe and unripe fish would be expected for red snailfish but not for pink snailfish. We are unable to account for our observation that the expected shift in depth distribution was found for pink snailfish but not red snailfish. One shortcoming of the 1982 sampling of snailfish eggs and larvae is that the species identification was based on an apparent species difference in egg and larvae size. Although larvae were collected in an at- tempt to establish identification by linking larval characteristics to adult fish, the larvae were poorly ossified and their identities could not be determined.1 In a second attempt at species identification, we sub- jected the eggs and larvae collected in the 1996 sam- pling to RFLP analysis. On this basis, it was pos- sible to establish that pink snailfish do deposit eggs in golden king crab. Unfortunately, we had no red snailfish tissue available to establish an identifica- tion pattern for the RFLP analysis because red snailfish were not captured during the 1996 sam- pling (the commercial crab fishery does not extend into sufficiently deep water). Therefore, it was not possible to determine if the 8 of 10 specimens not matching any of the three sampled Careproctus spe- cies were, as we suspect, red snailfish. Damage to the gills from the presence of egg masses may result from two causes. First, gill com- pression could be strong enough to restrict blood flow, resulting in localized necrosis. Second, egg masses could interfere with the functioning of the fifth pereio- pod (Fig. 10), which is covered with setae and func- tions as a gill-cleaning appendage (Pohle, 1989). Experiments with other lithodid crabs have demon- strated that gill fouling similar to the discoloration observed in golden king crab could be experimentally produced either by restricting the movement of the fifth pereiopod or removing the setae from this ap- pendage. The fouling is due not only to the accumu- lation of detritus on the gill surface, but also the ac- cumulation of a host of organisms that feed on the detritus or directly on the gill tissue itself (Pohle, 1989). In golden king crabs, the degeneration of the gill tissue can proceed to the point that all of the gill tissue on the infested side is missing (Fig 10). In cases of small to medium lesions, it was clear from histo- logical examinations that gill regeneration would occur at the next molt, but crabs with extreme cases of gill degeneration were not examined histologically and it is not known whether gill regeneration would occur. However, in at least one case, a newly molted golden king crab was encountered with no gills on one side of the body. Perhaps, this crab was an ex- treme case of gill damage which did not regenerate. Gill damage not only influences a crab’s respiration but also its capability for ion exchange. Lithodid crabs that died as the result of the restriction of their clean- ing appendages displayed considerable abdominal swelling which is indicative of ion imbalance ( Pohle, 1989). The 35% higher mortality of infested male crabs, compared with uninfested crabs, within the live tanks of commercial vessels (Table 4) indicates that the presence of egg masses hinders the ability of crabs to withstand the stress of capture and confinement. The impact on the fishery, however, depends on the incidence of snailfish eggs and larvae as well as the additional mortality they induce. In the case of the 1996 incidence in the Bering Sea (8.5%, Table 3, cor- rected for deadloss), for example, live-tank mortal- ity increased by only 0.08%. Even at the higher inci- dence found in 1982 (44%), additional live-tank mor- tality would have been only 0.40%. Although the mortality that is induced on wild crabs is unknown, it appears that the impact of snailfish parasitism on the golden king crab fishery is small. Acknowledgments We thank Brad Stevens, Jay Orr, Morgan Busby, and Doug Pengilly for reviewing the manuscript; Susie Byersdorfer for laboratory analysis; and A1 Shimada for helping the senior author with the 1981 collec- tion of crab and snailfish samples and for providing the photograph of the red snailfish. Literature cited Alderdice, D. F., and F. P. J. Velsen. 1978. Relation between temperature and incubation time for eggs of chinook salmon ( Oncorhynchus tshawytscha ). J. Fish. Res. Board Can. 35:69-75. Allien J. M., and G. B. Smith. 1988. Atlas and zoography of common fishes in the Bering Sea and northeastern Pacific. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 66, 151 p. Anderson, E. M., and G. M. Cailliet. 1974. Crab and snailfish commensalism in Monterey Bay. Underwater Nat. 8:29-31 . Balbontin, F., I. G. Campodonico, and L. M. Guzman. 1979. Descripcion de huevos y larvae de especies de 884 Fishery Bulletin 96(4), 1998 Careproctus (Pisces: Liparidae) comensales de Paralomis granulosa y Lithodes antarctica (Crustacea: Lithodidae). An. Inst. Patagonia 10:235-243. Dowling, T. E., C. Moritz, J. D. Palmer, and L. H. Rieseberg. 1996. Nucleic acids III: analysis of fragments and restric- tion sites. In D. M. Hillis, C. Moritz, and B. K. Mable (eds.). Molecular systematics, 2nd ed., p. 249-320. Sinauer, MA. Hunter, C. J. 1969. Confirmation of symbiotic relationship between liparid fishes (Careproctus spp.) and male king crab ( Paralithodes camtschatica). Pac. Sci. 23:546-547. Kido, K. 1988. Phylogeny of the family Liparididae, with the tax- onomy of the species found around Japan. Mem. Fac. Fish. Hokkaido Univ. 35:125-256. Love, D. C., and T. C. Shirley. 1993. Parasitism of the golden king crab, Lithodes aequispinus Benedict, 1895 (Decapoda, Anomura, Litho- didae) by a liparid fish. Crustaceana 65:97-104. McCaughran, D. A., and G. C. Powell. 1977. Growth model for Alaskan king crab (Paralithodes camtschatica). J. Fish. Res. Board Can. 34:989-995. Melville-Smith R., and E. Louw. 1987. An association between a liparid fish species and the stone crab Lithodes tropicalis (Decapoda, Anomura). S.-Afr. Tydskr. Dierk. 22:140-142. Parrish, R. 1972. Symbiosis in the blacktail snailfish, Careproctus melanurus, and the box crab Lopholithodes foraminatus. Calif. Fish Game 68:239-240. Pauly, D., and R. S. V. Pullin. 1988. Hatching time in spherical, pelagic, marine fish eggs in response to temperature and egg size. Environ. Biol. Fishes 22:261-271. Peden, A. E., and C. A. Corbett. 1973. Commensalism between a liparid fish, Careproctus sp., and the lithodid box crab Lopholithodes foraminatus. Can. J. Zool. 51:555-556. Pohle, G. 1989. Gill and embryo grooming in lithodid crabs: compara- tive functional morphology based on Lithodes maja. In B. E. Felgenhauer, L. Watling, and A. B. Thistle (eds.), Functional morphology of feeding and grooming in Crusta- cea, p 75-94. A. A. Balkema, Rotterdam. Powell, G. C., and R. B. Nickerson. 1965. Reproduction of king crabs, Paralithodes camt- schatica (Tilesius). J. Fish. Res. Board Can. 22:101-111. Rass, T. S. 1950. Zamechatel’nyi sluchai biologicheskoi svyazi ryby i kraba [An unusual instance of a biological relation between fish and crabs]. Priroda 39(7):68-69. In Russian. Translated by Paul Macy, U. S. Fish and Wildlife Service, Translation Series No. 41. [Available from the Northwest Fisheries Science Center, 2725 Montlake Blvd. E. Seattle, WA 98102.] Ryan, E. P. 1966. Pheromone: evidence in a decapod crustacean. Science 151:340-341. Sloan, N. A. 1985. Life history characteristics of fjord-dwelling golden king crab Lithodes aequispina . Mar. Ecol. Prog. Ser. 22:219-228. Somerton, D. A., and R. S. Otto. 1986. Distribution and reproductive biology of the golden king crab, Lithodes aequispina , in the eastern Bering Sea. Fish. Bull. 84:571-584. Stein, D. L. 1980. Aspects of reproduction of liparid fishes from the con- tinental slope and abyssal plain off Oregon, with notes on growth. Copeia 1980:687-699. Syrjala, S. E. 1996. A statistical test for a difference between the spatial distributions of two populations. Ecology 77:75-80. Venables, W. N., and B. D. Ripley. 1994. Modern applied statistics with S-plus. Springer- Verlag , New York, NY, 462 p. Vinogradov, K. A. 1950. K biologii Tikhookeanskogo pinagora v Kamchatskikh vodakh [Biology of the Pacific lumpfishes in Kamchatka waters]. Priroda (Leningrad) 39( 7 ):69— 70. 885 Settlement and recruitment of queen conch, Strombus gigas, in seagrass meadows: associations with habitat and micropredators Allan W, Stoner Melody Ray-Culp Sheila M. O'Connell Caribbean Marine Research Center 805 E. 46th Place Veto Beach, Florida 32963 Present address (for A. W. Stoner) Northeast Fisheries Science Center National Marine Fisheries Service, NOAA 74 Magruder Road, Highlands, New Jersey 07732 E-mail address (for A. W. Stoner): al.stoner@noaa.gov Abstract.-A suction dredge survey was conducted in the Bahamas in a tidal flow field system which contained a nursery ground for the economically significant gastropod Strombus gigas (queen conch). Settlement of larval conch within the system was associated with the specific location of the nurs- ery and positively correlated with sub- sequent recruitment to the juvenile population ( <45 mm shell length). Settlement was relatively independent of habitat features including depth, sediment characteristics, and macro- phytes. Conversely, densities of micro- predators (small crabs, shrimp, and predaceous gastropods) capable of con- suming early postsettlement conch were often correlated with habitat fea- tures such as seagrass shoot density, seagrass detritus, and organic content of the sediment. The density of small xanthid crabs (mode=1.5 mm carapace width) was positively correlated with density of live postsettlement conch (mean <4/m2), suggesting that conch settle in predator-prone areas or that the crabs respond numerically to small conch (or both). Densities of xanthids were very high (to >200/m2), and the crabs probably represent an important source of mortality for small conch in the primary nursery ground. Shells of dead conch indicated that molluscan and asteroid predators probably caused most of the predatory mortality on young conch that settled outside the nursery. Because critical settlement and recruitment habitats for queen conch are associated with particular hydrographic conditions, these habitats cannot be iden- tified or predicted simply by mapping obvious features such as seagrass cover, depth, or sediment type. An understand- ing of dynamic processes, such as larval transport and retention, selective settle- ment mechanisms, and trophic ecology, will be required to identify critical nurs- ery habitats. Manuscript accepted 13 January 1998. Fish. Bull. 96: 885-899 < 1998). Distribution of marine species with planktonic larvae is a function of both pre- and postsettlement pro- cesses. Location of reproductive sources, physical and oceanographic processes, duration and survivor- ship of larval stages, and larval be- havior combine to determine settle- ment location, whereas predation, suitable habitat, and the behavior of early juveniles affect the number of individuals that survive to repre- sent juvenile recruitment (Keough and Downes, 1982; Luckenbach, 1984; Connell, 1985). Experiments conducted in the laboratory with a large variety of species have shown that settlement in the field is prob- ably a nonrandom process for the majority of marine invertebrates, and settlement behavior is usually assumed to place juveniles in habi- tats most suitable for growth or sur- vival (or both) (Hadfield and Scheuer, 1985; Butman and Grassle, 1992; Grassle et al., 1992; Davis and Stoner, 1994; Stoner et al., 1996a). Despite an abundance of data and experiments on settlement, rela- tively little is known about the rela- tive importance of settlement rate and postsettlement predation in controlling local recruitment of most species to juvenile stages (Osman and Whitlatch, 1995). Be- cause early postsettlement juve- niles of many benthic species, espe- cially those associated with soft sedi- ments, are difficult to survey, quan- titative data related to recruitment often depend upon observations of the smallest juveniles that can be de- tected easily. The result is a relatively poor understanding of the crucial period between settlement and first detection in the benthos ( Keough and Downes, 1982; Luckenbach, 1984). Queen conch ( Strombus gigas) is a large gastropod that has great economic significance throughout the Caribbean and adjacent regions (Appeldoorn, 1994). Despite the wealth of information now available for conch larvae (Davis, 1994; Stoner and Davis, 1997), juveniles in year class 1 (e.g. Marshall, 1992; Iversen et al., 1987; Appeldoorn, 1994), and adults (Appeldoorn, 1988; Stoner and Sandt, 1992), little is known about larval settlement and the early postsettlement stage. Newly settled queen conch ( <10 mm ) are cryptic, often buried (Iversen et al., 1989; Sandt and Stoner, 1993), and are readily preyed upon by micropredators such as xanthid 886 Fishery Bulletin 96(4), 1 998 crabs and certain polychaetes (Ray-Culp et al., 1997). Queen conch <50 mm shell length have rarely been encountered in the field (Iversen et al., 1994; Ray and Stoner, 1995; Iversen and Jory, 1997), and the only density data extant for these small conch were collected by Sandt and Stoner (1993) near Lee Stock- ing Island, Bahamas. However, queen conch settle at a large size (1.2 mm), compared with many other mollusks, and even the smallest juvenile, unless crushed completely, leaves an identifiable shell record of its settlement in the sediment (our study). This record can be used to examine the settlement-recruitment relationship and spatial variation in settlement. Our study had three primary objectives. First, we conducted an extensive dredge survey for newly settled queen conch in and around a well-studied nursery area in the central Bahamas during the sum- mer recruitment season in 1992. The survey was designed to test the hypothesis that the long-term distribution pattern of year-class 1 and 2 conch was associated with settlement pattern. The sum of live and dead individuals was used as an index of settle- ment at each sampling station, and the number of surviving (live) conch served as an index of recruit- ment. Through this survey, we gathered the first quantitative data on the distribution of queen conch <45 mm shell length (1.5-44 mm). Second, we used observations of shell damage sustained by dead in- dividuals to determine the primary predatory forms on newly settled queen conch. We also examined spa- tial variation in predation type. Third, we examined the effects of environmental characteristics, such as depth, sediments, macrophytes, older conspecifics, and distribution of potential predators, on the ob- served distribution of newly settled queen conch. Study area The Exuma Cays are an important source of queen conch for a large fishery in the Bahamas; the island chain is 250 km long, bordered to the east by the Exuma Sound, and to the west by the Great Bahama Bank. Water exchange occurs through numerous tidal inlets separating the islands, creating exten- sive tidal flow fields on the shallow bank. Nurseries of juvenile conch (1- and 2-yr-old, 70-150 mm SL) are located primarily on the bank side of the Exuma Cays near these inlets and are typically associated with seagrass meadows (2-4 m deep) that are flushed with clear, oligotrophic water from the Sound dur- ing flood tide (Stoner et al., 1996b). We elected to survey postlarval conch in a tidal flow field located west of Lee Stocking Island and south of Norman’s Pond Cay because the conch nurs- ery near Shark Rock (Fig. 1) is well studied. Prior to sampling, five years of data had been collected on the distribution and abundance of 1- and 2-year-old juvenile conch, and environmental characteristics such as seagrass biomass, depth, and tidal currents had been mapped (Jones, 1996; Stoner et al., 1996b). Drogue studies have shown that, on flood tide, wa- ter from the Sound enters the inlet north of Lee Stock- ing Island, passes close to Shark Rock, and flows west of the nursery for a distance that is dependent on wind conditions and tidal phase (e.g. spring, neap) (Stoner et al., 1994). Tidal currents, which reach 100 cm/s at midtide, flow between sand bars in an S-shaped pattern following the bank bathymetry (Fig. IB). The middle of the tidal channel is ~3 m deep and veg- etated primarily with the seagrass Thalassia testudinum. Depth and seagrass density gradually decrease from midchannel to bare sand on both sides of the channel. Tidal range is ~1 m. Annual surveys conducted between 1988 and 1992 showed that aggregations of juvenile queen conch always grazed within a 2-km long section of the tidal flow field close to Shark Rock (Fig. 1) and, at any given time, occupied only portions of the suitable habitat available (Stoner et al., 1996b). Stations for this study were established with reference to long- term distribution of conch in this area (Stoner and Waite, 1990; Stoner and Ray, 1993; Ray and Stoner, 1994) and to represent both down-channel and across- channel dimensions of the flow field. One line of sta- tions was established in midchannel down the flow field from Adderly Cay (A) to Cook’s Cay (F) with an attempt to locate all of the stations in similar depth and moderate seagrass shoot density. Stations A and F were each located ~4 km from the geographic center of the traditional nursery ground (Fig. 1), and stations B and E were located -250 m outside the northeastern and southwestern ends of the nursery, respectively. Stations C3 and D3 were each established -500 m from the geographic center of the nursery (Fig. 1C). To represent the across-flow field dimension, five more stations were selected along two transects that lay perpendicular to the main axis of the tidal cur- rent at stations C3 and D3, as well as across the seagrass gradient. Thalassia testudinum density ranged from 4 to 704 shoots/m2 and from 0.5 to 178 g dry wt/m2 across this gradient in 1991 (Ray and Stoner, 1994) (Fig. 1C). Three stations were estab- lished along the D transect in addition to D3, increas- ing in macrophyte cover from bare sand at D1 to high seagrass and detrital biomass at D4. Along transect C, two stations were established in addition to C3, ranging from sand to moderate seagrass shoot den- sity and macrophyte biomass. There were no areas of high seagrass biomass present near transect C and, Stoner et a I.: Recruitment of Strombus gigas 887 Figure 1 (A) Study area in the southern Exuma Cays, central Bahamas. (B) On the flood tide, water from Exuma Sound passes between Adderly Cay and Lee Stocking Island, flowing in an S-shaped pattern past Shark Rock to- wards Cook’s Cay (dotted arrow). Stations A and F represented ends of the flow field. (C) Nine more stations were selected with respect to the long-term (1988-92) location of the juvenile queen conch aggregation, delin- eated by dashed line. In July 1992, just before dredge sampling was conducted, the aggregation occupied the areas shown in hatched polygons. Star indicates center of aggregation. consequently, no station was directly comparable to D4. Other environmental characteristics, such as depth and certain sediment characteristics, also varied across the flow field (see “Results” section, Tables 1 and 2). Methods Dredge sampling Given that maximum densities of conch veligers are observed during midsummer (June through August) and that the larval period lasts 3-4 weeks (Davis et al., 1993), dredge sampling for newly settled conch was conducted at the end of the summer, 24 August to 1 September 1992. Scuba equipment was used for all underwater work. At each of the 11 stations, wa- ter depth was measured and corrected to mean low water (MLW), and duplicate sediment samples were collected with a PVC core (diameter=40 mm, depth=5 cm) prior to dredging. Sediment samples were rinsed with freshwater and dried at 80°C to constant weight. A subsample ( ~ 15- 20 g) was incinerated at 550°C in a muffle furnace 888 Fishery Bulletin 96(4), 1 998 for 4 h to determine organic content, calculated as the percent difference between dry weight and ash- free dry weight. A second rinsed subsample ( —20 g) was analyzed by using standard dry sieve procedures (Folk, 1966), and product moment statistics were used to calculate mean grain size (McBride, 1971). The silt-clay fraction (>4.0 0, <62 mm), always < 9% of sediment dry weight, was not fractionated. Replicate dredge sample plots (n-5 for all stations, except C3, where n= 6) were delineated by a circular enclosure made of aluminum sheet metal (area=0.5 m2, height=0.3 m) that was placed haphazardly at each station. The enclosure was pushed securely into the sediment to prevent escape of motile fauna. In the middle of each sample plot, the number of Thalassia testudinum shoots was counted in a quad- rat (25 x 25 cm). A gasoline-powered suction dredge (modified from Brook, 1979) was used to collect year- class 0 queen conch, other macrofauna, and associ- ated macrophytes from each plot. The dredge cre- ated high-pressure water flow which, as a result of Venturi principle, drew algae, detritus, sediments, and macrofauna through a PVC intake tube (dia- meter=7.6 cm) into a mesh bag (40 x 70 cm, 1.2 mm mesh). Preliminary sampling in the study area and observations of year-class 0 conch ( <45 mm SL) indi- cated that conch in the nursery ground buried no deeper than 5 cm into the sediment. Therefore, sedi- ments from plots with seagrass were removed to the depth at which T. testudinum rhizomes occurred (usu- ally 8-15 cm); in bare sand, dredging was 8-10 cm deep. Unlike the other macrophytes, living seagrass did not detach easily and was not collected with the dredge material. The mesh bags holding the dredged materials were tied securely underwater and later fixed in a 10% formalin-seawater mixture contain- ing rose bengal as a staining agent. After 24 hours, each sample was rinsed onto a sieve (1.2 mm) and preserved in 70% ethanol until sorted. Macrophytes were sorted into three components: T. testudinum detritus (senescent blades and frag- ments), the green macroalga Batophora oerstedi, and the red algae Laurencia spp. Occasionally fronds of calcareous green algae (including Halimeda spp., Penicillus capitatus, Udotea spp., and Rhipocephalus phoenix) were collected, but they were sparsely dis- tributed and not quantified. Each fraction was rinsed with freshwater to remove salts and dried at 80°C to constant weight (~24 h) so that biomass (g dry wt/ m2) could be calculated. For ease in extracting newly settled queen conch and their potential predators, sediments were divided into two fractions, those retained on 1.2- and 2.0- mm sieves. A conch was classified as alive if its soft tissue and operculum were intact. Shortly after death, the operculum detaches from the foot, and soft tissues decompose quickly. Therefore, dead and liv- ing conch were easily distinguished. We do not be- lieve that conch shells were damaged during collec- tion because the dredge apparatus lifted samples off the bottom by suction that could be controlled and the materials collected did not pass through an impellor. Care was also taken so that the fauna were not damaged in sieving. None of the conch classified as alive at collection had damaged shells, and most other taxa, such as polychaetes and crabs, were in good (i.e. whole) condition. To gain insight into modes of predation, shells of dead conch were classified as 1) whole and undam- aged, 2) drilled, 3) peeled back along the spire line, or 4) crushed. Whole shells of dead conch were at- tributed to predation by mollusks or asteroids ( Jory, 1982; Iversen et ah, 1986; Ray and Stoner, 1995), drilled shells were probably the result of mollusk kills (Vermeij, 1987), and peeled and crushed shells were attributed to crustaceans (Randall, 1964; Vermeij, 1982, 1987; Davis, 1992). Whole shells can also re- sult from nonpredatory mortality. The proportion of dead individuals was used as an indicator of post- settlement mortality. Whole shells (from both live and dead conch) were measured for shell length. When only a shell spire or shell aperture was found, total shell length for the dead animal was calculated on the basis of regression formulae derived from measurements of 20 whole shells ranging in size from 3 to 40 mm total length: Length = ( spire length x 2.4) - 1.8; [r2=0.991] Length = ( aperture length x 1.6) + 0.5. [r2=0.998] Given that queen conch settle into nursery grounds during a distinct season, it was possible to determine if an individual had indeed settled in 1992 on the basis of its size, color, and the amount of biological encrustation. Settlement of queen conch can occur at the beginning of May, and growth rates during the early postsettlement period may be as great as 0.45 mm/d (Ray and Stoner, 1994); therefore, we con- sidered conch <45 mm in total shell length to be members of year-class 0. Also, conch shells lose their pink interior color within a few weeks after death. Animals that settled in 1992 were easily discerned on the basis of shell size and color. Xanthid and portunid crabs and alpheid shrimps were extracted from samples because they were abundant and known to be significant predators of postsettlement conch (Ray-Culp et al., 1997). Olivid and marginellid gastropods were also removed and counted as potential predators. Carapace width (includ- Stoner et a\. : Recruitment of Strombus gigas 889 ing lateral spines) of crabs was measured, carapace length for shrimps, and shell length of gastropods. Data analysis To discern station differences, density data from the dredge sampling were log-transformed (log10 (n+1)) to improve homogeneity of variance (Cochran’s test, P>0.05) and analysed by using 1-way AN OVA follow- ing the guidelines of Day and Quinn (1989). Tukey’s multiple comparison test was performed to determine pairwise relationships. Seven measures of density were examined: live conch, dead conch, total conch, alpheids, portunids, xanthids, and total predators. Relationships between these seven density mea- sures and eight environmental variables (water depth, distance of the station from the center of ju- venile queen conch aggregation, sediment grain size, sediment organics, Thalassia testudinum shoot den- sity, weight of T. testudinum detritus, and biomass of the macroalgae Batophora oerstedi and Laurencia spp.) were examined with pairwise correlation. Log- transformation improved the relationships, and Pearson correlation coefficients are reported for the log-transformed variables. The relationships between conch density (live, dead, and total) and each of the five predator fami- lies were also examined with correlation. Variables were not transformed in the analysis because trans- formation did not improve the correlations. Results Habitat characteristics Depth down the flow field was relatively uniform, ranging from 2.8 to 3.3 m at MLW, except at E where depth was 2 m (Table 1). Across the flow field at transect D, depth increased from bare sand (1.3 m) to high seagrass density (3.3 m), and, at transect C, greatest depth (3.5 m) occurred in low density seagrass (C2). Sediments were fine to medium sands (1.4-2. 6 0), with mean grain size decreasing slightly (increasing 0) with depth over both transects C and D (Table 1). Organic content of the sediments across transect D also increased with depth, ranging from 2.7% (station Dl) to 5.2% (station D4). The range of organic content in the down flow-field dimension was 3. 0-4.6%. Thalassia testudinum shoot density decreased down the flow field from 784 to 320 shoots/m2 (Table 2). Across the channel, shoot density increased rapidly from 0 to >500 shoots/m2 in both transects, as had been in- Table 1 Habitat characteristics for 11 dredge stations in the Shark Rock flow field. Depth was at mean low water and distance was measured from each station to the center of the juvenile queen conch nursery (see Fig. 1). Means and the range of values (paren- theses) are given for sediment grain size and organics ( n-2 for all stations except C3, where n= 3). Data for stations C3 and D3 are given twice for ease of comparison in both flow-field dimensions. Station Depth (m) Distance (km) Grain size (0) Sediments Organics (% dry wt) Down flow field, midchannel A 2.8 3.60 2.57 (2.56-2.58) 4.08 (3.68-4.50) B 3.0 1.00 1.81 (1.78-1.84) 4.05 (3.46-4.65) C3 3.1 0.55 1.84 (1.63-1.95) 3.54 (3.17-3.75) D3 2.8 0.55 2.07 (2.04-2.10) 4.56 (4.43-4.69) E 2.0 1.10 1.89 (1.86-1.92) 2.95 (2.95-2.95) F 3.3 4.50 2.04 (1.97-2.12) 3.99 (3.98-4.00) Across flow field Transect C Cl 2.0 0.70 1.51 (1.27-1.74) 2.18 (2.16-2.21) C2 3.5 0.70 1.78 (1.73-1.83) 3.31 (3.09-3.53) C3 3.1 0.55 1.84 (1.63-1.95) 3.54(3.17-3.75) Transect D Dl 1.3 0.60 1.44 ( 1.27-1.60) 2.72 (2.60-2.84) D2 2.1 0.55 1.73 (1.72-1.74) 3.12 (3.06-3.19) D3 2.8 0.55 2.07 (2.04-2.10) 4.56 (4.43-4.69) D4 3.3 0.70 1.87 (1.60-2.14) 5.24 (5.12-5.37) 890 Fishery Bulletin 96(4), 1998 Table 2 Macrophyte characteristics for 1 1 dredge stations in the Shark Rock flow field (see Fig. 1). Shoot count and detritus values are for Thalassia testudinum. All values are mean ±SE (n= 5 for all stations except C3, where n= 6). Data for stations C3 and D3 are given twice for ease of comparison in both flow-field dimensions. Station Shoot density (no./m2) Biomass (g dry wt/m2) Detritus B. oerstedi Laurencia spp. Down flow field, midchannel A 784 ± 36 353 ± 49 0.02 ± 0.01 0.20 ± 0.05 B 640 ± 65 144 ± 20 0.80 ± 0.40 0.04 ± 0.03 C3 536 ± 26 30 ± 11 0.007 ± 0.004 0.20 ± 0.18 D3 528 ± 36 139 ± 20 24.08 ± 6.00 0.84 ± 0.27 E 352 ± 22 29 ± 10 44.76 ± 4.77 0.66 ± 0.13 F 320 ± 17 34 ± 8 2.18 ± 0.57 0 ± 0 Across flow field Transect C Cl 0 ± 0 0.06 ± 0.04 0.02 ± 0.03 0 ± 0 C2 288 ± 44 37 ± 14 1.76 ± 0.45 0.34 ± 0.18 C3 536 ± 26 30 ± 11 0.007 ± 0.004 0.20 ± 0.18 Transect D Dl 0 ± 0 0.08 ± 0.04 0.06 ± 0.04 0 ± 0 D2 240 ± 36 14 ± 4 3.68 ± 0.32 1.60 ± 0.04 D3 528 ± 36 139 ± 20 24.08 ± 6.00 0.84 ± 0.27 D4 544 ± 22 180 ± 17 0.22 ± 0.23 1.58 ± 0.83 tended in the sampling design. A similar increase occurred with T. testudinum detritus, with the ex- ception of a relatively low value at station C3 (Table 2), where the aggregation of year-class 1 and 2 juve- nile conch undoubtedly had a grazing effect (Fig. 1). As was intended, shoot density and detritus biom- ass increased across transect D, and values were highest at station D4. Algal biomass was noticeably high only at stations D3 and E (where standing crops of Batophora oerstedi were 24 and 45 g dry wt/m2, respectively), and particularly low at station C3 (grazed by conch). Biomass of Laurencia spp. was low at all stations (<1.6 g dry wt/m2). Newly settled conch Newly settled queen conch were collected in dredge samples at all 11 stations except F (Fig. 2). Down the flow field, mean total density was relatively high (8-12 conch/m2) at stations B, C3, and D3, and low (0-2.4 conch/m2) at stations A, E, and F, although differences were only significant between C3 and F (Fig. 2). There was a significant negative correlation between total conch and distance from the geographic center of the long-term aggregation (Table 3); both total density and density of dead conch decreased with distance from C3 (Fig. 2). All of the conch col- lected at stations A and E were dead, as were most (60-86%) from the other down flow field stations. Live conch were collected at stations B, C3, and D3, with maximum density (4 conch/m2) observed at B. Across the flow field, mean total conch density appeared to increase with seagrass density in transect C (Fig. 2), although differences were not sig- nificant. Values were relatively uniform at all of the stations in transect D, ranging from 4.0 to 8.4/m2. Live newly settled conch were most abundant at sta- tions C3 (1.7/m2) and D3 (1.6/m2), where seagrass density was moderate. Significant numbers of live individuals were also collected at the other stations with at least some seagrass (C2, D2, D4). However, all of the conch dredged from the bare sand stations (Cl, Dl) were dead. When data for 11 stations were included in the analysis, live conch density (the index of recruitment) had a significant positive correlation with total conch density (the index of settlement) (coefficient of cor- relation [r]=0.703, P=0.016). The percentage of the total conch that were dead provides an index of mor- tality. This index was negatively correlated with settlement (r=0.654, P=0.04). Live individuals ranged in size from 3.3 mm to 38.5 mm with the mode at 10-14.9 mm (Fig. 3). Dead in- dividuals ranged in shell length from 1.5 to 44 mm with the mode at 1.0-4. 9 mm. The percentage of dead conch with whole, undamaged shells decreased across the flow field along both transects in the direction of increasing seagrass density and increasing density of Stoner et al.: Recruitment of Strom bus gigas 891 Table 3 Pearson correlation coefficients (n=ll) for the pairwise relationship between each of seven dependent variables (i.e. newly settled queen conch and predators) and eight independent variables. Depth was at mean low water; distance was measured from station to center of juvenile queen conch aggregation (see Fig. 1); grain size and organics are for sediments; shoot density and detritus values are for T. testudinum. * 0.01 < P < 0.05; ** 0.001 < P < 0.01; *** P < 0.001. All of the dependent variables were transformed (log10(x+D) prior to analysis. Live conch Total conch Xanthids Alpheids Portunids Olivids Marginellids Depth 0.474 0.123 0.710* 0.428 -0.081 0.063 0.231 Distance -0.393 -0.770** 0.141 0.358 -0.158 0.165 -0.346 Grain size -0.027 -0.312 0.533 0.821** 0.337 -0.254 -0.233 Organics 0.346 0.196 0.868*** 0.694* 0.041 -0.398 0.304 Shoots 0.454 0.168 0.812** 0.924*** 0.496 -0.460 0.273 Detritus 0.085 0.038 0.537 0.947*** 0.515 -0.707* 0.103 B. oerstedi -0.106 -0.260 0.120 -0.091 -0.013 0.164 -0.020 Laurencia spp. -0.026 0.182 0.237 0.164 -0.311 -0.205 0.082 Table 4 Number of dead newly settled queen conch dredged from 11 stations for each of four shell conditions (see text for definitions). Values are station percent- ages followed by (n). Data for stations C3 and D3 are given twice for ease of comparison in both flow field dimensions. Station Whole Drilled Peeled Crushed Total Down flow field, midchannel A 33.3 (2) 16.7 (1) 0 (0) 50.0(3) (6) B 6.7 (1) 0 (0) 13.3 (2) 80.0 (12) (15) C3 9.7 (3) 0 (0) 25.8 (8) 64.5 (20) (31) D3 18.7 (3) 0 (0) 12.5 (2) 68.8(11) (16) E 33.3 (1) 0 (0) 33.3 (1) 33.3 ( 1) (3) F 0 (0) 0 (0) 0 (0) 0 (0) (0) Total 14.1 (10) 1.4(1) 18.3 (13) 66.2 (47) (71) Across flow field Transect C Cl 57.1 (4) 0 (0) 28.6 (2) 14.3 (1) (7) C2 17.6(3) 0 (0) 5.9 (1) 76.5 (13) (17) C3 9.7 (3) 0 (0) 25.8 (8) 64.5 (20) (31) Total 18.2 (10) 0 (0) 20(11) 61.8 (34) (55) Transect D D1 64.7(11) 0 (0) 5.9 (1) 29.4 (5) (17) D2 55.6 (5) 111 (1) 22.2 (2) 11.1 (1) (9) D3 18.7 (3) 0 (0) 12.5 (2) 68.8(11) (16) D4 25.0 (5) 10.0 (2) 0 (0) 65.0 (13) (20) Total 38.7 (24) 4.8 (3) 8.1 (5) 48.4 (30) (62) Grand total7 27.0 (38) 2.8 (4) 13.5 (19) 56.7 (80) (141) 1 Stations C3 and D3 are included only once in grand total. year-class 1 and 2 conch (Table 4). The pattern of crushed shells was reverse. In the down flow-field dimension, the percentage of whole empty shells was lowest at stations within the nursery ground and at station F. Crushed shells made up the highest proportion of dead, newly settled conch within the nursery area. Overall, few shells were drilled (2.8%) or peeled (13.5%), and most were crushed (56.7%). Conch predators Xanthid crabs (mostly Micropanope spp.) composed the predator group with the greatest densities and gov- erned the density distribution of to- tal predators in both flow-field dimen- sions (Fig. 4). Xanthids were most abundant (102-286 per m2) at sta- tions B, D3, and D4, and were <36/m2 at all other stations. Alpheid density decreased down the flow field (Fig. 3), increased with seagrass density across the flow field (Figs. 3 and 4), and had a high positive correlation with detritus (r=0.947, Table 3). Xanthid and alpheid densities also had high positive correlations with Thalassia testudinum shoot density and sediment organics (Table 3). Portunid crab densi- ties were relatively low, compared with the two other crustacean families. They decreased down the flow field from a maximum at A (7.6/m2) to 0 at F (Fig. 4). Of the two predaceous mollusc families observed, the Olividae were most abundant at stations C2 (18/m2) and F (25/m2) (ANOVA, F1045=5.69, PcO.OOl). Val- ues were <7. 2/m2 at all other stations and the data were not plotted. Olivids were negatively correlated with detritus (Table 3). The Marginellidae were most abundant at B (9.2/m2) and D4 (6.8/m2) and densities were <2/m2 at all other stations, but the differences 892 Fishery Bulletin 96(4), 1998 16 12 8 4 0 a 40% lb 0% ab ab 114% 20% Live conch (F(1045) = 3.9; P= 0.0001) 16 12 8 b b 4 0% 0% ‘ < 0 B C3 D3 E Cl C2 C3 D1 D2 D3 D4 c o "O JZ o c o o 16 12 - Dead conch (F(10 45) = 2.0; P= 0.05) 16 12 A B C3 D3 E Cl C2 C3 D1 D2 D3 D4 Total conch (F(1045) = 2.5; P = 0.02) A B C3 D3 E F Down flow field Cl C2 C3 D1 D2 D3 D4 Across flow field Stations Figure 2 Density of live, dead, and total queen conch collected from 11 dredge stations located down and across the Shark Rock flow field. Stations C3 and D3 are illustrated in both dimensions for comparison. Values are mean ± SE (n= 5 for all stations except C3, where n= 6). Percentage of the total count represented by live and dead conch at each station is also given. F and P values are for 1-way ANOVAs made for each variable and flow-field dimension. Means that were not statisti- cally different (P>0.05) are designated by similar letters (Tukey multiple comparison test on log-transformed data). were not significant (F1045=1.92, 0.067). Neither olivids nor marginellids were collected at station A. Because size-frequency distribution of each preda- tor group varied little among the 11 stations, mea- surements from all stations were considered together (Table 5). The majority of xanthids were very small, with a mode of just 1.5 mm carapace width. The larg- est predator collected was a portunid (36.1 mm cara- pace width); all others were <15 mm. Modal size of alpheids was 4.2 mm carapace length. Stoner et al.: Recruitment of Strombus gigas 893 T3 CD Dead conch Z <5 5 10 15 20 25 30 35 40 40 30 20 10 0 <5 5 10 15 20 25 30 35 40 Shell length (mm) Figure 3 Length-frequency distribution for newly settled queen conch derived from all individuals, live and dead, col- lected from 11 dredge stations. Each value on the x axis represents a range of sizes starting with the lower end of the size interval (e.g. 5 represents 5. 0-9. 9 mm). Live conch density had a high positive correlation with xanthid density (r=0.825, P=0.002), marginellid density (r=0.695, P=0.02), and total predator den- sity (r=0.794, P=0.004), whereas relationships with densities of alpheids, portunids, and olivids were not significant (P>0.38). There were no significant rela- tionships with any of the predator families (P>0.15) for dead or total conch densities. Discussion Age of postlarvae and settlement dates The shell lengths of dredged newly settled conch pro- vide insight into when settlement occurred. Using the middle of the week-long sampling period as the Table 5 Size range and mode for three crustacean and two mollus- can families of potential predators of newly settled conch collected at 11 dredge stations. Family Measurement Range (mm) Mode (mm) Xanthidae carapace width 1.0-14.5 1.5 Alpheidae carapace length 1.0-7. 8 4.2 Portunidae carapace width 1.5-36.1 3.4 Marginellidae shell length 1. 7-6.1 3.2 Olividae shell length 2.0-11.6 3.2 endpoint (28 August 1992), we estimated a summer average growth rate of 0.39 mm/day (Ray and Stoner, 1995), and considering settlement at 1.2 mm shell length (Davis, 1994), we determined that animals in the 10-15 mm modal size class would have settled 3-4 weeks earlier, between 24 July and 5 August. The largest live individual (38.5 mm) would have settled on about 24 May, and the smallest (3.3 mm) would have settled on 23 August. Dead conch ranged from 1.5 to 44 mm SL. According to the same assump- tions, they would have been in the benthos for 1-110 days at the time of death. Therefore, settlement re- corded in our samples began in early May and con- tinued through at least late August. It should be pointed out that our intent was to collect young conch as close to the peak settlement period as possible, near the end of August. However, it is known that larvae are present in the water column near Lee Stocking Island until at least late September (Stoner and Davis, 1997), and our collections probably do not represent total settlement at Shark Rock in 1992. There are three possible explanations for the fact that dredging yielded relatively few live conch out- side the 10-14.9 mm modal class. First, there could have been a major settlement event between late- July and early August. Queen conch larvae were col- lected at the Shark Rock nursery ground on 13 dates between late May and mid-September 1992 (Stoner and Davis, 1997); however, relatively few late-stage larvae were found in these collections, and it is im- possible with the available data to determine if the modal size of survivors resulted from a period of high settlement rate. Nevertheless, this is the most par- simonious explanation. Second, the high number of year-class 0 conch in the 10-14.9 mm range may also be related to high survivorship in conch settling be- tween late July and early August, compared to conch settling at other times. However, the steep decline in numbers of individuals with size could be a reflec- tion of the natural, high mortality of small conch (dis- 894 Fishery Bulletin 96(4), 1998 JE 6 £ (0 o to "O a> a. 150 125 100 75 50 25 0 25 20 15 10 5 0 Alpheidae (F(I0 45) = 14.9; P< 0.001) A B C3 D3 E F Cl C2 C3 D1 D2 D3 D4 Portunidae (F(10 45) = 6.5: P< 0.001) Down flow field Across flow field Stations Figure 4 Density of three families of queen conch predators and total predators collected from 11 dredge stations located down and across the Shark Rock flow field. Total predators include alpheids, portunids, xanthids, as well as two mollusc families — olivids and marginellids (not shown individually). Stations C3 and D3 are illustrated in both dimensions for comparison. Values are mean ±SE (n= 5 for all stations except C3, where n- 6). Note different y scale on each graph. F and P values are for 1-way ANOVAs made for each variable and flow-field dimension. Means that were not statistically different (P> 0.05) are designated by similar letters (Tukey multiple comparison test on log-transformed data). cussed below). This latter conclusion is supported by the fact that the modal size for dead conch was small (1-5 mm). Future studies should examine the inter- action of settling date with food quality and quan- tity, growth, and predation rates. Few data of this type exist for sediment-dwelling invertebrates. Third, the smallest postlarval conch were probably under- sampled. Despite the use of small mesh (1.2-mm), some of the smallest individuals may have passed through the dredge bags. Furthermore, shells of newly settled queen conch are frequently broken into fragments that would not be retained on a 1.2-mm mesh. Nevertheless, our study provides the first quantitative data on newly settled conch ( <45 mm SL, <6 months postlarval age) and important insights into their early life history. Role of settlement and habitat on conch recruitment In this study, large volumes of sediment had to be sorted to detect relatively low settlement densities (<12/m2) of queen conch over the traditional nurs- ery. These densities were not surprising given that 1- and 2-year-old queen conch typically occur in ag- gregations with just 0. 2-1.0 individual/m2 (Stoner and Ray, 1993). The most intriguing findings are that settlement was concentrated in the traditional nurs- ery ground and that recruitment to early benthic stages was directly correlated with settlement den- sity over a large spatial scale. This is compatible with the prediction (Connell, 1985) that settlement will be most influential in predicting the recruitment of Stoner et at: Recruitment of Strombus gigas 895 350 300 250 200 150 100 50 0 350 300 250 200 150 1 00 50 0 Xanthidae (F(1045) = 36.1; P < 0.001) A B C3 D3 E F Cl C2 C3 D1 D2 D3 D4 Total predators (F(1045) = 22.5; P < 0.001) A B C3 D3 E F Cl C2 C3 D1 D2 D3 D4 Down flow field Across flow field Stations Figure 4 (continued) benthic invertebrates when settlement occurs in low density After an extensive review of the literature, Butman (1987) concluded that settlement patterns in inver- tebrates associated with soft sediments are a func- tion of passive accumulation and deposition of lar- vae over spatial scales of kilometers and that active habitat selection occurs primarily over smaller scales (centimeters to meters). Data on the abundance of queen conch veligers over the Shark Rock tidal flow field (Stoner and Davis, 1997) may support this con- clusion . Although newly hatched larvae ( 300-500 p m ) were collected well beyond the Shark Rock nursery, at station F on 12 of 13 sampling dates in 1992, nei- ther midstage larvae (500-900 pm), competent lar- vae (>900 pm), nor newly settled conch were ever collected there, suggesting that late-stage larvae are somehow concentrated at the nursery location. At maximum flood-tide current, velocity near the sur- face decreases by approximately 75% between sta- tion A, north of Lee Stocking Island, and Shark Rock. Near the bottom, the decrease in velocity is more than 95% (Stoner and Ray-Culp1). We also know that maximum flood tidal excursion in this flow field oc- curs near the Shark Rock nursery area on neap tide; therefore, the nursery is bathed in oligotrophic wa- ter (and perhaps larvae) from the Exuma Sound on every tide, whereas areas farther out on the bank are not (Jones, 1996; Stoner et al., 1996b). These hydrographic characteristics may result in the depo- sition of larvae in the long-term nursery area, or the lower botom-water velocities may allow them to settle. Density of newly settled conch had a strong nega- tive correlation with distance from the center of the traditional nursery ground but was relatively inde- pendent from all other environmental characteris- tics tested (i.e. depth and various qualities of sedi- ments, seagrass, and macroalgae). This finding in- dicates that conch settlement cannot be explained 1 Stoner, A. W., and M. Ray-Culp. 1997. Northeast Fisheries Science Center, Natl. Mar. Fish Serv., NOAA, 74 Magruder Road, Highlands, NJ 07732. Unpubl. data. 896 Fishery Bulletin 96(4), 1 998 simply by fine-scale hydrodynamic relationships as- sociated with seagrasses and macroalgae, as observed for certain other invertebrates including mollusks (Eckman, 1987; Harvey et al., 1993, 1995). In fact, conch larvae settled in approximately equal densi- ties across the seagrass gradient at transect D which spanned only 250 m. Therefore, the structure pro- vided by macrophytes appears not to influence settle- ment of conch larvae even though older juveniles prefer seagrass habitats and are associated with an optimal shoot density (Stoner and Waite, 1990). Although accumulation of competent larvae near the Shark Rock nursery is the most parsimonious explanation for the observed settlement and recruit- ment patterns, earlier experiments indicate that seemingly similar seagrass beds offer different quali- ties that have significant effects on larval and juve- nile conch. Laboratory experiments have shown that macrophytes collected from stations within the Shark Rock nursery (B, C3, D3, D4) induced significantly higher metamorphosis than the same types of sub- strata collected outside the general nursery area (sta- tions A, F) (Davis and Stoner, 1994). Growth rates of newly settled conch (1.2 mm) fed seagrass detritus from the different sources reflected metamorphic responses on the same substrata (Stoner et al., 1996a), and when 1 -yr-old juvenile conch were trans- planted to stations A and F, growth rates were low in comparison with those in the Shark Rock nursery (Stoner et al., 1994). It is clear, therefore, that the nursery area is trophically unique, despite visual similiarity to surrounding areas. Micro-organism films that coat the sediment in soft-bottom communities are known to be important inducers of metamorphosis in conch and other in- vertebrates, most likely because they are associated with favorable nutritional requirements for postlar- vae (Scheltema, 1961; Gray, 1974; Davis and Stoner, 1994; Stoner et al., 1996a). Characteristics that make the Shark Rock nursery an attractive location for settling larvae and an ecologically suitable habitat for juveniles probably stem from hydrodynamic prop- erties of the location that affect nutrient cycling and productivity patterns in certain algal foods for conch (Stoner et al., 1994, 1996b). Field manipulations will be needed to distinguish direct effects of hydrody- namics (i.e. larval transport and retention) from in- direct effects such as hydrographic mediation of bio- logical productivity, habitat choices, and postsettle- ment processes. Postlarval conch recruited to the same habitats traditionally occupied by 1- and 2-yr-old animals. This association is probably not the result of conspe- cific attraction because settlement was equally high at stations with and without older conspecifics. This result corroborates an earlier laboratory study show- ing that cues associated with previously settled ju- veniles (slime trails, feces, and the older conch them- selves) did not elicit larval metamorphosis (Davis and Stoner, 1994). The role of micropredators on conch recruitment Although settlement of queen conch was relatively independent of habitat features other than location, predator distributions were highly correlated with seagrass shoot density, detrital abundance, and sedi- ment organics and grain size. Numerous studies have shown that animal abundances in seagrass beds are correlated with certain measures of habitat complex- ity such as seagrass blade density or biomass, detri- tal biomass, leaf characteristics, or rhizome struc- ture (Orth et al., 1984; Stoner and Lewis, 1985). The association between benthic macrofauna and physi- cal structure may be related to food abundance or predation, or both. There is abundant experimental evidence that the physical structure provided in seagrass beds reduces predation rates on inverte- brates (Heck and Wilson, 1987; Heck and Crowder, 1991). Potential conch predators dredged in this study were prey species themselves, and they un- doubtedly derived some measure of protection or nutrition from the habitat, or both. Predation on early postsettlement stages can be an important process affecting the number of inver- tebrate settlers that survive to recruit into a popula- tion (Thorson, 1966; Keough and Downes, 1982; Osman and Whitlach, 1995). Heavy losses to micropredators during the first days or weeks after settlement can severely diminish or eliminate a prey species, and even regulate community composition (Osman et al., 1992; Osman and Whitlatch, 1995). It is apparent from the length frequency of dead conch collected in this study that very high mortality oc- curs immediately after settlement, when conch are <5-mm shell length. Queen conch have many preda- tors at this size (Ray-Culp et al., 1997), and one of the most important is probably the xanthid crab Micropanope sp., which was the most abundant in- vertebrate counted in dredge samples. The crab is capable of killing conch that are up to 0.5 times its own carapace width ( Ray-Culp et al.2 ). Although a large proportion of the xanthids collected were too small (mode=1.5 mm) to kill even newly settled conch 2 Ray-Culp, M., M. Davis, and A. W. Stoner. 1998. Escaping the xanthid crab gauntlet — the role of size, density and habitat for newly-settled queen conch. Caribbean Marine Research Cen- ter, 805 E. 46th Place, Vero Beach, FL 32963. Unpubl. manuscr. Stoner et a I.: Recruitment of Strombus gigas 897 ( 1.2 mm), xanthids up to 10 mm were far more abun- dant than conch, and they undoubtedly play a major role in conch mortality. The high positive correlation between live conch and xanthids suggests that conch settle in areas prone to high predator abundance. However, with typical summer growth rates, men- tioned earlier, conch would escape predation by xanthids in about 10 days when they reach 5 mm shell length. Consequently, predator-prey relation- ships associated with the early postsettlement stages may be highly dynamic, with suites of predators shift- ing rapidly over time. There was an increase in conch survivorship from bare sand to moderate density seagrass over both cross-channel transects. A similar trend was obtained experimentally when year-class 0 and 1 juveniles were tethered over an analogous seagrass gradient near transect D (Ray and Stoner, 1994, 1995). High- est mortality occurred on bare sand and in highest seagrass biomass whereas lowest mortality occurred in moderate biomass. It appears that a certain amount of seagrass structure provides protection, but too much is detrimental. However, in the down flow field dimension, survivorship was highest within the traditional nursery ground, and distribution is clearly related to both settlement and survivorship in this dimension. Identifying critical habitats Our findings have important management implica- tions for queen conch and other species associated with seagrass beds. First, it may not be possible to determine the value of a particular site for a species on the basis of simple habitat maps. Even descrip- tions of the beds that include seagrass species com- position, biomass, and shoot density do not provide adequate information to identify critical habitats for conch. Value of specific seagrass bed locations also depends upon hydrography, larval retention, larval settlement, predator abundance, and the related survivorship. Persistence in the locations of queen conch nurseries near Lee Stocking Island (Stoner et ah, 1996b; Jones, 1996) indicates that these factors are relatively constant over periods of several years, and that the specific locations may be as “critical” as habitat type. Only certain seagrass beds are suitable for queen conch; these must be identified and pro- tected. Furthermore, in many areas of the Caribbean, conch populations have been devastated by overfish- ing, and there is an intense interest in rehabilitat- ing them through releases of hatchery-reared juve- niles. Transplant experiments have shown that ju- venile conch will survive and grow only in very spe- cific locations, and releases must be made in loca- tions that ensure economically acceptable survivor- ship (Stoner, 1994). Clearly, thorough knowledge of distributional mechanisms is necessary to make pre- dictions on the habitat requirements of queen conch and other managed species. Acknowledgments This research was supported by a grant from the National Undersea Research Program of NOAA (U.S. Department of Commerce). We thank C. Bolton, D. Carlin, L. Cowell, M. Davis, C. Kelso, J. Lally, and P. Monaghan for assistance in the field, laboratory sort- ing, and sediment analysis. B. Bower-Dennis drafted Figure 1. J. Lin and anonymous reviewers helped to improve the manuscript. Literature cited Appeldoorn, R S. 1988. Age determination, growth, mortality and age of first reproduction in adult queen conch, Strombus gigas L., off Puerto Rico. Fish. Res. 6:363-378. 1994. Queen conch management and research: status, needs and priorities. In R. S. Appeldoorn and B. Rodriguez (eds.), Queen conch biology, fisheries and mariculture, p. 301-319. Fundacion Cientifica Los Roques, Caracas, Ven- ezuela. Brook, I. M. 1979. A portable suction dredge for quantitative sampling in difficult substrates. Estuaries 2:54-58. Butman, C. A. 1987. Larval settlement of soft-sediment invertebrates: the spatial scales of pattern explained by active habitat selec- tion and the emerging role of hydrodynamical processes. Oceanogr. Mar. Biol. Ann. Rev. 25:113-165. Butman, C. A., and J. P. Grassle. 1992. Active habitat selection by Capitella sp. I larvae. I. Two-choice experiments in still water and flume flows. J. Mar. Res. 50:669-715. Connell, J. H. 1985. The consequences of variation in initial settlement vs. post-settlement mortality in rocky intertidal communi- ties. J. Exp. Mar. Biol. Ecol. 93:11-45. Davis, M. 1992. Predation of hatchery-reared juvenile queen conch, Strombus gigas (L.) by juvenile spiny lobsters, Panulirus argus ( L. ). M.S. thesis, Florida Institute of Technology, Melbourne, FL, 37 p. 1994. Short-term competence in larvae of queen conch Strombus gigas: shifts in behavior, morphology and meta- morphic response. Mar. Ecol. Prog. Ser. 104:101-108. Davis, M., C. A. Bolton, and A. W. Stoner. 1993. A comparison of larval development, growth, and shell morphology in three Caribbean Strombus species. Veliger 36:236-244. Davis, M., and A. W. Stoner. 1994. Trophic cues induce metamorphosis of queen conch larvae (Strombus gigas Linnaeus). J. Exp. Mar. Biol. Ecol. 180:83-102. 898 Fishery Bulletin 96(4), 1 998 Day, R. W., and G. P. Quinn. 1989. Comparisons of treatments after an analysis of vari- ance in ecology. Ecol. Monogr. 59:433-463. Eckman, J. E. 1987. The role of hydrodynamics in recruitment, growth, and survival of Argopecten irradians (L.) and Anomia sim- plex (D’Orbigny) within eelgrass meadows. J. Exp. Mar. Biol. Ecol. 106:165-191. Folk, R. L. 1966. A review of grain-size parameters. Sedimentology 6:73-93. Grassle, J. P., C. A. Butman, and S. W. Mills. 1992. Active habitat selection by Capitella sp. I larvae. II. Multiple-choice experiments in still water and flume flows. J. Mar. Res. 50:717-743. Gray, J. S. 1974. Animal-sediment relationships. Oceanogr. Mar. Biol. Ann. Rev. 12:223-261. Hadfield, M. G., and D. Scheuer. 1985. Evidence for a soluble metamorphic inducer in Phestilla: ecological, chemical and biological data. Bull. Mar. Sci. 37:556-566. Harvey, M., E. Bourget, and R. G. Ingram. 1995. Experimental evidence of passive accumulation of marine bivalve larvae on filamentous epibenthic structures. Limnol. Oceanogr. 40:94-104. Harvey, M., E. Bourget, and G. Miron. 1993. Settlement of Iceland scallop Chlamys islandiea spat in response to hydroids and filamentous red algae: field observations and laboratory experiments. Mar. Ecol. Prog. Ser. 99:283-292. Heck, K. L., Jr., and L. B. Crowder. 1991. Habitat structure and predator-prey interactions in vegetated aquatic systems. In S. S. Bell, E. D. McCoy, and H. R. Mushinsky (eds.), Habitat structure: the physi- cal arrangement of objects in space, p. 281-299. Chapman and Hall, New York, NY. Heck, K. L., Jr., and K. A. Wilson. 1987. Predation rates on decapod crustaceans in latitudi- nally separated seagrass communities: a study of spatial and temporal variation using tethering techniques. J. Exp. Mar. Biol. Ecol. 107:87-100. Iversen, E. S., and D. E. Jory. 1997. Mariculture and enhancement of wild populations of queen conch ( Strombus gigas ) in the western Atlantic. Bull. Mar. Sci. 60:929-941. Iversen, E. S., S. P. Bannerot, and D. E. Jory. 1 989. Evidence of survival value related to burying behavior in queen conch Strombus gigas. Fish. Bull. 88:383-387. Iversen, E. S., D. E. Jory, and S. P. Bannerot. 1986. Predation on queen conchs, Strombus gigas, in the Bahamas. Bull. Mar. Sci. 39:61-75. Iversen, E. S., D. E. Jory, and D. J. DiResta. 1994. Research on the first year queen conch, Strombus gigas, relevant to fisheries management. Proc. Gulf Caribb. Fish. Inst. 43:498-505. Iversen, E. S„ E. S. Rutherford, S. P. Bannerot, and D. E. Jory. 1987. Biological data on Berry Islands (Bahamas) queen conchs, Strombus gigas, with mariculture and fisheries management implications. Fish. Bull. 85:299-310. Jones, R. L. 1996. Spatial analysis of biological and physical features associated with the distribution of queen conch, Strombus gigas , nursery habitats. M.S. thesis, Florida Institute of Technology, Melbourne, FL, 98 p. Jory, D. E. 1982. Predation by tulip snails, Fasciolaria tulipa, on queen conchs, Strombus gigas. M.S. thesis, Univ. of Miami, FL, 73 p. Keough, M. J., and B. J. Downes. 1 982. Recruitment of marine invertebrates: the role of active larval choices and early mortality. Oecologia 54:348-352. Luckenbach, M. W. 1984. Settlement and early post-settlement survival in the recruitment of Mulinia lateralis (Bivalvia). Mar. Ecol. Prog. Ser. 17:245-250. Marshall, L. S. 1992. Survival of juvenile queen conch, Strombus gigas, in natural habitats: impact of prey, predator and habitat features. Ph.D. diss., College of William and Mary, Williamsburg, VA, 144 p. McBride, E. F. 1971. Mathematical treatment of size distribution data. In R. E. Carver ( ed. ), Procedures in sedimentary petrology, p. 9-27. Wiley, New York, NY. Orth, R. J., K. L. Heck Jr., and J. van Montfrans. 1984. Faunal communities in seagrass beds: a review of the influence of plant structure and prey characteristics on predator-prey relationships. Estuaries 7:339-350. Osman, R. W., and R. B. Whitlatch. 1995. Predation on early ontogenetic life stages and its ef- fect on recruitment into a marine epifaunal community. Mar. Ecol. Prog. Ser. 117:111-126. Osman, R. W., R. B. Whitlatch, and R. J. Malatesta. 1992. Potential role of micro-predators in determining re- cruitment into a marine community. Mar. Ecol. Prog. Ser. 83:35-43. Randall, J. E. 1964. Contributions to the biology of the queen conch, Strombus gigas. Bull. Mar. Sci. Gulf. Caribb. 14:246-295. Ray-Culp, M., M. Davis, and A. W. Stoner. 1997. The micropredators of settling and newly-settled queen conch (Strombus gigas Linnaeus). J. Shellfish Res. 16:423-428. Ray, M., and A. W. Stoner. 1994. Experimental analysis of growth and survivorship in a marine gastropod aggregation: balancing growth with safety in numbers. Mar. Ecol. Prog. Ser. 105:47-59. 1995. Growth, survivorship, and habitat choice in a newly settled seagrass gastropod, Strombus gigas. Mar. Ecol. Prog. Ser. 123:83-94. Sandt, V. J., and A. W. Stoner. 1993. Ontogenetic shift in habitat by early juvenile queen conch, Strombus gigas: patterns and potential mechanisms. Fish. Bull. 91:516-525. Scheltema, R. S. 1961. Metamorphosis of the veliger larva of Nassarius obsoletus (Gastropoda) in response to bottom sediment. Biol. Bull. 120:92-109. Stoner, A. W. 1994. Significance of habitat and stock pre-testing for en- hancement of natural fisheries: experimental analyses with queen conch Strombus gigas. J. World Aquacult. Soc. 25:155-165. Stoner, A. W., and M. Davis. 1997. Abundance and distribution of queen conch veligers (Strombus gigas Linne) in the Central Bahamas: I. Hori- zontal patterns in relation to reproductive and nursery grounds. J. Shellfish. Res. 16:7-18. Stoner, A. W., M. D. Hanisak, N. P. Smith, and R. A. Armstrong. 1994. Large-scale distribution of queen conch nursery habi- Stoner et a I.: Recruitment of Strombus gigas 899 tats: implications for stock enhancement. In R. S. Appeldoorn and B. Rodriguez (eds.), Queen conch biology, fisheries and mariculture, p. 169-189. Fundacion Cienttfica Los Roques, Caracas, Venezuela. Stoner, A. W., and F. G. Lewis III. 1985. The influence of quantitative and qualitative aspects of habitat complexity in tropical sea-grass meadows. J. Exp. Mar. Biol. Ecol. 94:19-40. Stoner, A. W., P. A. Pitts, and R. A. Armstrong. 1996b. The interaction of physical and biological factors in the large-scale distribution of queen conch populations in seagrass meadows. Bull. Mar. Sci. 58:217-233. Stoner, A. W., and M. Ray. 1993. Aggregation dynamics in juvenile queen conch (Strombus gigas): population structure, mortality, growth, and migration. Mar. Biol. 116:571-582. Stoner, A. W., M. Ray, R. A. Glazer, and K. J. McCarthy. 1996a. Metamorphic responses to natural substrata in a gastropod larva: decisions related to postlarval growth and habitat preference. J. Exp. Mar. Biol. Ecol. 205:229-243. Stoner, A. W., and V. J. Sandt. 1992. Population structure, seasonal movements and feed- ing of queen conch, Strombus gigas , in deep-water habi- tats of the Bahamas. Bull. Mar. Sci. 51:287-300. Stoner, A. W., and J. M. Waite. 1990. Distribution and behavior of queen conch, Strombus gigas , relative to seagrass standing crop. Fish. Bull. 88:573-585. Thorson, G. 1966. Some factors influencing the recruitment and estab- lishment of marine benthic communities. Neth. J. Sea Res. 3:267-293. Vermeij, G. J. 1982. Gastropod shell form, breakage, and repair in relation to predation by the crab Calappa. Malacologia 23:1-12. 1987. Evolution and escalation: an ecological history of life. Princeton Univ. Press, Princeton, NJ, 527 p. 900 Abstract .—Late larvae (15-30 mm TL) of the Japanese sardine, Sardinops melanostictus, are commercially ex- ploited in fishing grounds along the Pacific coast of western and central Japan. Concentrated shoals of late lar- vae in the shallow (15-30 m deep) coastal (4-6 miles from the coast) fish- ing grounds enable fishermen to catch as much as several hundred metric tons (t) (several billion larvae in number) per month. Growth trajectories of sardine larvae caught in the fishing ground off Atsumi Peninsula in central Japan were individually backcalculated by using the biological intercept method based on the allometric relationship between otolith radius and fish length. Growth rates for larvae up to 13-21 d were high, ranging from 0.79 to 0.85 mm/d, but declined after reaching size of immigration ( 13-19 mm TL) from the offshore waters to the coastal fishing grounds. The decline of growth rate in the late larval stage seemed to be re- lated to the concentration of late lar- vae in the fishing grounds, the result of onshore intrusions of offshore Kuroshio waters. Total lengths at age 20 d were significantly smaller in 1990 (total catch of larval sardine was 720 t) than in 1991 (total catch 300 t) in spite of a higher sea surface temperature in 1990 in the coastal habitat. This may have resulted from a larger population of late larvae on the fishing ground in 1990 than in 1991. Manuscript accepted 6 February 1998. Fish. Bull. 96:900-907 (1998). Growth trajectory of the larval Japanese sardine, Sardinops melanostictus , transported into the Pacific coastal waters off central Japan Yoshiro Watanabe Ocean Research Institute, University of Tokyo 1- 15-1 Minamidai, Nakano-ku, Tokyo 164, Japan E-mail address : ywatanab@ori.u-tokyo.ac.jp Motohiko Nakamura Marine Resources Research Center, Aichi Fisheries Research Institute 2- 1 Toyoura, Toyohama, Minami-Chita, Chita-gun, Aichi 470-34, Japan Late larvae of the Japanese sardine Sardinops melanostictus and the Japanese anchovy, Engraulis japon- icus, termed shirasu in Japan, are fished by boat seiners in waters along the Pacific coast of western and central Japan. The coastal wa- ters (4-6 miles from the coast) off Atsumi Peninsula (Fig. 1) are one of the major fishing grounds for sh irasu. Depth of the sea bed of the fishing grounds is about 20 m, with a range from 15 to 30 m. The shirasu fishery for sardine larvae begins in March and continues to December, with the major effort shifting to anchovy in early sum- mer. The annual catch of sardine larvae in the waters ranged from 700 to 2000 metric tons (t), 7-20 billion larvae, 1980-88. However, between 1989 and 1991 the catch declined to 300-700 t (Fig. 2). Since 1992, catches have been as low as several tens of tons, with an excep- tion of about 400 t in 1993. Growth rates of fish larvae affect their survival and recruitment to the adult population (Anderson, 1988). From life-stage table analy- ses, growth rate has been shown to be an important determinant of year-class strength by delimiting the duration of a particular life stage with high instantaneous mor- tality (Lo et al., 1995; Butler et al., 1996). Watanabe et al. (1995) showed that recruitment failures of the Japanese sardine in 1988-91 could not be explained by mass mor- tality at the first feeding stage. Cu- mulative mortality after the first feeding stage of the sardine, which may be a function of growth rate, is likely to have determined the re- cruitment in these years. Meekan and Fortier (1996) examined early life growth and survival of the At- lantic cod, Gadus morhua, and showed that fast-growing pelagic larvae survived better through the larval stage and dominated in the cohort of demersal juveniles. Cam- pana (1996) found a positive corre- lation between growth rates in ju- venile Atlantic cod and subsequent year-class strengths, which enabled him to predict recruitment from the juvenile growth rates. Available population size of late larval sardine in the shirasu fish- ing grounds could be affected by lar- val growth. The size range of lar- vae caught in the shirasu fishery is usually from 15 to 30 mm TL. Be- cause the duration time of this size Watanabe and Nakamura: Growth of Sardinops melanostictus 901 5 . 34.0 °N 137.0 °E ’ 137.5 °E Figure 1 Location of shirasu fishing grounds in the coastal waters along Atsumi Peninsula. Minor fishing grounds are found in Ise Bay at the tip of Chita peninsula. Dots indicate locations of monthly SST observations. A solid line along the Pacific coast of western and central Japan denotes an example of the meandering Kuroshio current axis. range is a function of growth rate, slower growth results in longer duration time and an increase in the available number of larvae. Watanabe and Kuroki (1997) backcalculated the growth history of larval sardine Sardinops melanostictus in the coastal wa- ters off Miyazaki in western Japan. They found that growth rates reached a maximum (0.80-0.85 mm/d) at around 10-12 mm in total length (TL) but slowed thereafter, exhibiting asymptotic growth trajectories. According to this growth pattern, duration time from 15 to 30 mm TL was calculated to be 35 days. If, however, growth rates of 0.80-0.85 mm/d were main- tained in the late larval stage, duration time of the same size range could be as short as 18 days. Growth rate in early life stages is thus important as a potential determinant of year-class strength as well as of available stock sizes of shirasu larvae of the Japanese sardine. The daily growth rate of fish can be estimated by regressing size-at-age data to a growth model (Campana and Jones, 1992). When we apply this method, we need data points throughout an age range, from first feeding up to the fish size of concern (Watanabe et al., 1997). Because the size range of sardine larvae fished by the boat seine fish- ery is usually from 15 to 30 mm, we do not have a complete range of data points for early larvae, and therefore the regression method is not applicable to describe the growth history from first feeding to size at capture. Instead, backcalculation of size at age from the relationship between otolith radius and fish length makes it possible to draw a growth trajectory for individual fish (Campana, 1990; Campana and Jones, 1992). We backcalculated growth trajectories of individual larval sardines caught in coastal waters off Atsumi Peninsula, 1990 and 1991, and compared them with growth trajectories of larvae from other waters. Materials and methods Larval sampling Two shirasu fishing boats, towing a seine net, were used to sample waters off Atsumi Peninsula and in Ise Bay (Fig. 1). They usually departed before dawn from the fishing port, together with a catch-loading boat, set the net several times in the morning, and returned to the port for offloading. Catches were stored with ice before landing. In our study, we ran- domly sampled shirasu larvae from catches in the ports ofMorozaki and Toyohama (tip of Chita Penin- sula) four times in 1990 (11, 16, 27 April and 7 May) and four times in 1991 (15, 23 April and 7, 14 May). 902 Fishery Bulletin 96(4), 1998 The larvae were classified as Japanese sardine, S. melanostictus, Japanese anchovy, E. japonicus, and round herring, Etrumeus teres. Sardine larvae were preserved in 80% ethanol for otolith examinations. Otolith measurement Total lengths (TL) of sardine larvae were measured to the nearest 0.1 mm with an optical comparator. Sagittal otoliths were dissected and cleaned under a binocular microscope, mounted on a glass plate with enamel resin, and used for measurements and counts of daily growth rings. We used the otolith mea- surement system (Ratoc System Engineering Inc.) composed of a light microscope, a video camera and monitor, and an image analyzer controlled by com- puter. Because the otolith had not yet developed a rostrum and because we were not able to determine its orientation at the larval stage, we measured ra- dii of all daily rings along the maximum radius of the otolith. Growth backcalculation The relationship between larval TL and maximum otolith radius (OR) was considered for larvae sampled. Plots of TL against OR can be expressed by an allometric relationship (see “Results” section). An allometric OR-TL relationship has previously been demonstrated in larval S. melanostictus from differ- ent waters (Watanabe and Kuroki, 1997). Otolith growth rings have been found to be deposited on a daily basis in S. melanostictus , with the first ring being formed on the third day of hatching (the day of first feeding) when larvae are reared at 18°C (Hayashi et al., 1989). The relationship of the zth otolith ring radius iORt) and TL on the day of the ith ring formation (TL;) is considered to be expressed by an allometric formula, TLt -ax ORb, for individual larvae. We determined the allometric parameters a and b for each larva by using the biological intercept method (Campana, 1990; Campana and Jones, 1992) and a size of first feeding larvae (first otolith daily ring deposition) of 5.0 mm TL (Watanabe and Kuroki, 1997). The solution of the following two equations gives us a and b for each larva: TL^ ax OR b and TLcapture = a x ORcaptureb, where TL] = total length (mm) at the first ring deposition which was fixed at 5.0 mm; ORx = the measured radius of the first daily ring; TL capture = the measured total length (pm) at cap- ture; and ORcapture = otolith radius (pm) at capture. TLt of each larva was thus calculated from the for- mula for each larva independently. Mean ±SD (stan- dard deviation) of TL at ages from 4 d up to a certain age, with at least 10 backcalculated TLs, was calcu- lated for the March- and April-hatched cohorts for the two years of study. Differences in mean TLs at ages 10, 15, 20, and 25 d were examined between the same month of hatch cohorts in 1990 and 1991 by using Student’s t-test, when variances were equal, or by Welch’s t-test when they were not equal. Sea surface temperature distribution Sea surface temperatures (SST) in and around the shirasu fishing ground were measured monthly at fixed stations off Atsumi Peninsula (Fig. 1). We used data from April and May, 1990 and 1991, to describe temperature distributions in coastal waters. Results The SST in coastal waters off Atsumi Peninsula was in the range of 16-19°C during 23-25 April and 8-9 May 1990. From 23 to 25 April, the isotherms ran nearly parallel to the coast line but from 8 to 9 May offshore waters warmer than 18°C intruded into the coastal fishing ground (Fig. 3). In 1991, the SST dur- ing 2-3 April and 8-9 May ranged from 11 to 18°C. Cold waters extruded from Ise Bay covered part of the coastal fishing ground. The SST in the fishing grounds were 1.5-2.5°C lower in 8-9 May 1991 than in the corresponding season in 1990. Local government permits the shirasu fishery to operate year round in these fishing grounds, but catches were zero in January-February 1990 and January-March 1991 (Fig. 4). Japanese sardine lar- vae were caught mainly in April in these years. The major target species of the shirasu fishery shifted to Japanese anchovy after May. A monthly catch of sar- dine larvae was 430 t in April 1990, declining to 175 t in April 1991. Annual catches of sardine larvae were 724 t and 298 t in 1990 and 1991, respectively. Total length of larval sardines caught in the shirasu fishery were 15-27 mm in mid-April 1990 (Fig. 5). Frequency of larvae smaller than 15 mm TL in- creased after late April in this year, with sizes of 11- 20 mm. In 1991, the modal size and range of the lar- vae remained relatively constant at 22-25 and 20-27 mm TL, respectively, mid-April to mid-May (Fig. 5). Sardine larvae fished on 11 and 16 April 1990 were aged from 14 to 39 d after hatching. All of them hatched in March (Fig. 6). Larvae caught on 27 April Watanabe and Nakamura: Growth of Sardinops melanostictus 903 137.0 E 137.5 E 34.0 N 34.5 N 18 34.0 N 137.0 E 137.5 E 34.5 N 34.0 N 137.0 E 137.5 E 16 > 17 . \ __ — 34.5 N 18 • 137.0 E 137.5 E 18 -f- 34.0 N Figure 3 Sea surface temperature distribution off Atsumi Peninsula in April and May of 1990 and 1991. and 7 May were young, from 10 to 23 d. Except for one larva hatched on 31 March (not used for backcalculation), all larvae caught on 27 April and 7 May hatched in April 1990. In 1991, ages of sardine lar- vae were 19-30 d. Larvae caught on 15 April were all hatched in March. Those caught on 23 April were composed of March- and April-hatched fish. All larvae caught on 5 and 14 May were hatched in April except one which was hatched on 24 March (not used for backcalculation). Plots of TL (mm) on OR (pm) of all indi- viduals could be expressed in an allomet- ric formula (Fig. 7). Size at age of indi- vidual larva was backcalculated on the basis of allometric OR-TL relationship for each larva from first feeding to capture. The 1990 March-hatched cohort grew lin- early to 25 d when fish reached 20.5 mm TL (Fig. 8). Mean growth rate of the co- hort from 4 d (6.3 mm TL) to 25 d was 0.68 mm/d. Growth in the April-hatched cohort in 1990 was linear to about 13 d (13.1 mm TL, 0.83 mm/d), then slowed down. In the 1991 March- and April-hatched cohorts, growth was linear to 21 d (21.2 mm TL, 0.85 mm/d), and 20 d (19.2 mm TL, 0.79 mm/d), respectively, declining thereafter. Larval TL at which growth started to de- cline occurred at the approximate size when immigration into the coastal fishing grounds occurred in both months. Total lengths at 15 d in the 1990 March- and April-hatched cohorts were 14.2 ±1.2 and 14.9 ±1.4 mm, respectively, whereas in 1991 the March-hatched cohort reached 16.3 ±1.5, and the April cohort 15.8 ±1.3 mm. Backcalculated TLs of March- and April-hatched cohorts were significantly smaller in 1990 than in corresponding hatching month in 1991 at 15, 20, and 25 d (Table 1 ). Discussion Larvae of S. melanostictus and E. japonieus have been reported to be transported from the offshore Kuroshio area to the coastal fishing grounds along the Pacific coast in central Japan by onshore intru- sions of Kuroshio waters (Tsuji, 1983; Muranaka, 1984; Mitani, 1990). This is the case in the coastal fishing grounds off Atsumi Peninsula, because great densities of S. melanostictus eggs were detected in the offshore waters along the Kuroshio Current in 1990 and 1991 (Ishida and Kikuchi, 1992; Zenitani et al., 1995; Watanabe et al., 1996). Onshore intru- sions of Kuroshio waters often develop when the Kuroshio meanders in the waters off central Japan (Kobayashi et al., 1986; Kasai, 1995). The available population size of S. melanostictus larvae in coastal fishing grounds off Atsumi Peninsula and adjacent waters is a positive function of the latitudinal dis- tance from Cape Omaezaki to the Kuroshio axis (Fig. 1) (Kishida et al., 1994). This distance in April and May 1991 measured approximately 125 nautical miles (n mi) (Maritime Safety Agency, 1991), which was less than that in 1990 (200 n mi) (Maritime Safety Agency, 1990). Oceanic conditions were less favorable for onshore larval transport in 1991 than in 1990. As seen in Figure 3, SSTs in coastal waters were lower in 1991 than in 1990, a feature that was indicative of limited intrusion of the warm Kuroshio waters to the coastal area. Annual catch of sardine shirasu in 1991 in the waters off Atsumi Peninsula (including small catches in Ise Bay) decreased to 41% 904 Fishery Bulletin 96(4), 1998 of the 1990 catch. This was due partly to unfavor- able oceanic conditions. Growth rates of sardine larvae estimated in this study were 0.79-0.85 mm/d to a size when they im- migrated to the fishing grounds as 1990 April- and 1991 March- and April-hatched cohorts, but later declined to 0.6-0. 7 mm/d (Fig. 8). This decline in growth was similar to the asymptotic growth of lar- val S. melanostictus in the shirasu fishing grounds in western Japan (Watanabe and Kuroki, 1997). Fish- ermen for shirasu locate a concentrated larval shoal by echo sounder (Mitani, 1987) and catch large num- bers of larvae in relatively narrow fishing grounds (Fig. 1). In mid-April 1990, 100 t (ca. one billion in number) per day of sardine larvae were caught in the fishing ground in our study.1 Turbidity is a fac- tor that helps to retain concentrated larval shoals in the coastal fishing grounds (Funakoshi, 1988). Uotani et al. (1993) demonstrated experimentally that E. japonicus larvae showed a strong positive taxis to turbidity and tended to stay in the turbid water. Reduction of larval growth, after reaching the size when immigration to the shirasu fishing grounds occurred, is likely to be related to concentrations of large numbers of larval sardines in the narrow fish- ing grounds as a result of intrusions of offshore wa- ters (Muranaka, 1984). 20 10 April 11 20 - N=34 _,J HdfM n - April 15 | N=35 JJ 20 10 <1) | | | | | | I rr^ T T 1 II 1 1 1 1 I 1 1 1 1 1 1 Ur 5 15 25 April 16 a 20 - N=35 J J 10 . iLEVi o 5 15 25 April 23 || N=37 JL| Percental o o o c 1 i i i i 1 i i i n i TT f 1 1 1 1 II i i i i i u 5 15 25 April 27 g| 20 - Twp- i i i ™ i fill i i i i i i I i 'T~ r i U 5 15 25 May 7 M *" JL. i i i i i i i i i i i i i i i i i 5 15 25 5 15 25 20 10 May 7 N=35 drtlft || 5 15 25 20 - May 14 10- N=61 0 1 i ■ i i i i i 5 15 25 Total length (mm) Figure 5 Frequency distributions of TL of sardine larvae caught in 1990 (left) and 1991 (right). Sampling date and sample size (n) are shown in each panel. The modal size of sardine larvae caught on April 27 and May 7 (April-hatched cohort) in 1990 was excep- tionally smaller ( 15-16 mm TL) than those in the other larval groups (Fig. 5). The mesh aperture of the net of the shirasu fishery is 2.1 mm at the codend. The di- agonal of the mesh is 3.0 mm, which is much larger than the body depth of 15- 16 mm sardine larvae (about 1.3 mm). A large proportion of larvae smaller than 15 mm could, therefore, be extruded from the mesh at the codend (Smith and Richardson, 1970). Nevertheless, many larvae smaller than 15 mm 1 1990. Marine Resources Re- search Center of Aichi Fisheries Research Institute, 2-1 Toyoura, Toyohama, Minami-Chita, Chita- gun, Aichi 47034, Japan. Unpubl. data Watanabe and Nakamura: Growth of Sardinops melanostictus 905 0) CL 20 - 10 - 0 March 29 TTT 10 T A-rrA April 1 1 March 9 *TT -Wh rii i i i i i i 20 30 40 20 10 -I March 26 CD o> 10 I 20 10 0 10 - 0 April 16 | March 6 | v 30 40 I March 31 T ■ W . , . B 10 20 30 40 20 10 0 March 31 1 B [March 13 | i i i i i ^ ^ ‘ ft ft i April 15 1 l l l l M l l 10 20 30 40 20 - 10 1 | April 5 April 23 March 22 fg ^ ^ W W . W I i i i i m i i l l i i i i M ' i i ' n i i i i i i i i n i i 10 20 30 40 Aprii 1 7 | April 27 20 - 10 - 0 - - 1 April 6 | i T Ti; Mis i i i i 10 20 30 40 1 | April 27 | May 7 20 - | ApnM6|B May 7 [March 24 | faiBi 1W1 30 40 May 14 10 -| I April 23 0 10 20 30 40 Age in days Figure 6 Frequency distribution of age of sardine larvae caught on different dates in 1990 (left) and 1991 (right). Boxes show the earliest or latest hatch date of the group. Table 1 Back-calculated TLs (in mm) of sardine larvae at ages 10, 15, 20, and 25 d after hatching. Hatching month 10 15 20 25 Mean SD n Mean SD n Mean SD n Mean SD n 1990 March 10.8** 0.9 69 14.1** 1.2 68 17.5** 1.4 65 20.5** 1.3 35 April 11.5 1.0 69 14.9** 1.4 39 17.9* 2.0 14 1991 March 12.3 1.1 62 16.3 1.5 62 20.4 2.1 61 23.1 2.6 33 April 11.7 0.9 110 15.8 1.3 110 19.2 1.7 110 21.6 1.8 74 * Significantly smaller (P<0.01) than the same hatching month cohort in 1991. ** Significantly smaller PcO.001) than the same hatching month cohort in 1991. (April-hatched cohort) were caught from late April to early May. This finding indicates that large num- bers of small larvae were carried to the coastal fish- ing grounds after April 16 in 1990. The SST distri- bution in 8-9 May indicated that there was a sub- stantial amount of intrusion of the offshore warm water after 23-25 April. Because the size of immi- gration to the fishing ground was smaller in this April-hatched cohort than in others, decline of growth in this cohort started earlier, at 13 d old (13.1 mm TL), compared with other cohorts from 20 to 21 d (19-20 mm TL) (Fig. 8). Larval TLs at 20 d were 20.4 ±2.1 and 19.2 ±1.7 mm, as March- and April-hatched cohorts in 1991, respec- tively, which were comparable to the sizes at 20 d (19.1 ±0.6-19.4 ±1.6 mm) of the January- to March- hatched cohorts of S. melanostictus in the waters off Miyazaki in 1991 (Watanabe and Kuroki, 1997). In our current study, however, TLs of the 1990 larvae at 20 d ranged from 17.5 to 17.9 mm, significantly 906 Fishery Bulletin 96(4), 1 998 smaller than the sizes in 1991. The difference in growth rates between 1990 and 1991 could not be explained by SST because in our study area, slower growth was recorded in 1990, when SST was higher, than in 1991. Relative abundance of sardine shirasu in the coastal waters of central Japan in 1991 de- creased to about 50% of the 1990 (Kishida et al., 1994). The catch of sardine larvae in 1991 declined to 41% of that in 1990 (Fig. 2). Funakoshi (1996) dem- onstrated that abundance of macrozooplankton was negatively correlated with the biomass of young-of- the-year sardines in Ise Bay (Fig. 1). He considered that large biomasses of young-of-the-year sardines resulted in a reduced abundance of macrozoo- plankton, followed by a decline in growth of the sar- dines through density-dependent processes. The greater population of sardine larvae in the shirasu fishing grounds in 1990 than in 1991 may have re- sulted in the slower growth of the sardine larvae in 1990. We need to study the density of food items for sardine larvae in the shirasu fishing grounds to ex- amine if sardine larval growth is limited by food availability. We also need to know interspecific com- petition for food between sardine and anchovy lar- vae in a coastal ecosystem, because they coincided in April and May in our study area (Fig. 4). The offshore Kuroshio frontal waters provide sar- dine larvae with sufficient foods. Nakata et al. (1995) calculated that the total food requirement of carnivo- rous macrozooplankters and sardine larvae was about 11% of the total production of small copepods ( < 1 .0 mm prosome length) in Kuroshio frontal wa- ters. This resulted in a higher feeding incidence in the early larval stage ofS. melanostictus in the fron- tal waters compared with the inshore and offshore waters of the frontal area (Nakata, 1995). Juvenile S. melanostictus (32-48 mm fork length) collected in 30 20 10 March 1990 / 0 0 0 0 0 T 1 1 1 1 1 0 10 20 30 Age in days Figure 8 Backcalculated growth trajectories of sardine larvae hatched in March and April in 1990 and 1991. Small dia- mond represents mean TL at age and vertical bar one SD of the m the Kuroshio frontal waters had stomach contents (mostly copepods and larvaceans) of 7-10% of wet body weight and were backcalculated to have grown at 0.8-0. 9 mm/d in the larval stage (Watanabe and Saito, 1998). Perhaps the coastal waters in and around the shirasu fishing grounds are a less favor- able feeding area for sardine larvae than the Kuroshio frontal waters. Acknowledgments We thank Shigeo Funakoshi for comments on ecol- ogy of larval sardine and anchovy in the study area. Masao Bando sampled the sardine larvae in the fish- ing ports. Hisae Furukawa prepared and read otolith specimens. This work was supported in part by Grants-in-Aid from the Ministry of Agriculture, For- estry, and Fisheries (Biocosmos project, BCP-98-IV- A-8) and from the Ministry of Education, Science and Culture. Literature cited Anderson, J. T. 1988. A review of size dependent survival during pre-re- cruit stages of fishes in relation to recruitment. J. North- west Atl. Fish. Sci. 8:55-66. Watanabe and Nakamura: Growth of Sardinops melanostictus 907 Butler, J. L., K. A. Dahlin, and H. G. Moser. 1996. Growth and duration of the planktonic phase and a stage based population matrix of Dover sole, Microstomus pacificus. Bull. Mar. Sci. 58:29-43. Campana, S. E. 1990. How reliable are growth back-calculations based on otoliths? Can. J. Fish. Aquat. Sci. 47:2219-2227. 1996. Year-class strength and growth rate in young Atlan- tic cod Gadus morhua. Mar. Ecol. Prog. Ser. 135:21-26. Campana, S. E., and C. M. Jones. 1992. Analysis of otolith microstructure data. Can. Spec. Publ. Fish. Aquat. Sci. 117:73-100. Funakoshi, S. 1988. Biological production mechanism of the Japanese anchovy and Japanese sardine shirasu in Suruga Bay and Enshu Nada. Bull. Japan. Soc. Fish. Oceanogr. 52:240-243. 1996. Variations in zooplankton abundance in Ise and Mikawa Bay. Kaiyo Monthly 28:142-149. Hayashi A, Y. Yamashita, K. Kawaguchi, and T. Ishii. 1989. Rearing method and daily otolith ring of Japanese sardine larvae. Nippon Suisan Gakkaishi 55: 997-1000. Ishida, M., and H. Kikuchi. 1992. Monthly egg productions of the Japanese sardine, an- chovy, and mackerels off the southern coast of Japan by egg censuses: January 1989 through December 1990. National Research Institute of Fisheries Science, Yokohama, 86 p. Kasai, A. 1995. Effect of variations in the Kuroshio and Oyashio cur- rents on the egg and larval transport and recruitment of Japanese sardine ( Sardinops melanostictus). Doctoral diss.. University of Tokyo, Tokyo, 139 p. Kishida, T., Y. Katsumata, M. Nakamura, S. Yanagibashi, and S. Funakoshi. 1994. An attempt for assessment on relative abundance of shirasu, larval Japanese sardine, off the Pacific coast of Japan. Bull. Natl. Res. Inst. Fish. Sci. 6:57-66. Kobayashi, M., T. Sugimoto, and T. Hirano. 1986. Surface current patterns in the Kumano-nada and the Enshu-nada seas for different types of the Kuroshio paths based on GEK data - II for periods with large mean- der of the Kuroshio. Bull. Japan. Soc. Fish. Oceanogr. 50:2-11. Lo, N. C. H., P. E. Smith, and J. L. Butler. 1995. Population growth of northern anchovy and Pacific sardine using stage-specific matrix model. Mar. Ecol. Prog. Ser. 127:15-26. Maritime Safety Agency. 1990. Current profile. Quick bulletin of ocean conditions. No. 8-10. Maritime Safety Agency, Tokyo, 4 p. 1991. Current profile. Quick bulletin of ocean conditions, 8-10. Maritime Safety Agency, Tokyo, 4 p. Meekan, M. G., and L. Fortier. 1996. Selection for fast growth during the larval life of At- lantic cod Gadus morhua on the Scotian Shelf. Mar. Ecol. Prog. Ser. 137:25-37. Mitani, I. 1987. Echo sounder information on size range of anchovy larvae and juveniles forming a school in the shirasu fish- eries ground in Sagami Bay. Bull. Japan. Soc. Fish. Oceanogr. 51:120-123. 1990. The biological studies on the larvae of Japanese an- chovy, Engraulis japonica HOUTTUYN, in Sagami Bay. Spec. Rep. Kanagawa Prefect. Fish. Exp. Stn. 5:1-140. Muranaka, F. 1984. Environmental variability and catch fluctuations of shirasu. In S. Funakoshi ( ed. ), Biological study on alter- ation of dominant larval species — Investigation of feeding conditions of early larvae, p. 195-220. Fisheries Agency, Tokyo. Nakata, K. 1995. Feeding conditions of Japanese sardine in and near the Kuroshio examined from their gut contents. Bull. Natl. Res. Inst. Fish. Sci. 7:265-275. Nakata, K., H. Zenitani, and D. Inagake. 1995. Differences in food availability for Japanese sardine larvae between the frontal region and the waters on the offshore side of Kuroshio. Fish. Oceanogr. 4:68-79. Smith, P. E., and S. L. Richardson. 1970. Standard techniques for pelagic fish egg and larva surveys. FAO Fish. Tech. Pap. 175:1-100. Tsuji, S. 1983. Study on recruitment mechanism of larval anchovy to shirasu fishing grounds in Sagami Bay based upon otolith daily increment analyses. Ph.D. diss., Univ. To- kyo, Tokyo, 157 p. Uotani, I., T. Iwakawa, and K. Kawaguchi. 1993. Experimental study on the formation mechanisms of shirasu (postlarval Japanese anchovy) fishing grounds with special reference to turbidity. Nippon Suisan Gakkaishi 60:73-78. Watanabe, Y., and T. Kuroki. 1997. Asymptotic growth trajectories of larval sardine in the coastal waters off western Japan. Mar. Biol. 127:369- 378. Watanabe, Y., Y. Oozeki, and D. Kitagawa. 1997. Larval parameters determining pre-schooling juve- nile production of saury Cololabis saira in the northwest- ern Pacific. Can. J. Fish. Aquat. Sci. 54:1067-1076. Watanabe, Y., and Saito, H. 1998. Growth and feeding of early juvenile Japanese sar- dines in the Pacific waters off central Japan. J. Fish Biol. 51:519-533. Watanabe, Y., H. Zenitani, and R. Kimura. 1995. Population decline of the Japanese sardine Sardinops melanostictus owing to recruitment failures. Can. J. Fish. Aquat. Sci. 52:1609-1616. 1996. Offshore expansion of spawning of the Japanese sar- dine, Sardinops melanostictus, and its implication for egg and larval survival. Can. J. Fish. Aquat. Sci. 53:55-61. Zenitani, H., M. Ishida, H. Konishi, T. Goto, Y. Watanabe, and R. Kimura. 1995. Distribution of eggs and larvae of pelagic fish spe- cies around Japan. Resources Management Research Report, ser. A- 1:1-368. 908 Transition from pelagic to benthic prey for age group 0-1 Atlantic cod, Gadus morhua Tammy M. Lomond Department of Biology Memorial University of Newfoundland St. John's, Newfoundland, Canada AIC 3X9 Present address: Greenhouse and Processing Crops Research Centre Agriculture and Agri-Food Canada Highway 20, Harrow, Ontario, Canada NOR 1G0 E-mail address: lomondt@em.agr.ca David C. Schneider Ocean Sciences Centre, Memorial University of Newfoundland St. John's, Newfoundland, Canada AIC 5S7 David A. IWHethven Department of Biology, Memorial University of Newfoundland St. John's, Newfoundland, Canada A1B 3X9 Atlantic cod, Gadus morhua L., settle to the bottom early in life at standard lengths of 30 to 40 mm (Bowman, 1981; Hawkins et ah, 1985; Hop et ah, 1994; Methven and Bajdik, 1994). After settlement into benthic habitats, Atlantic cod con- tinue to feed on pelagic prey but then shift to benthic prey. The rate at which cod make this transition has not been quantified at a level usable in trophic dynamic studies. To quantify the rate at which this shift in diet occurs, the volumetric proportion of pelagic prey in the diet of 98 juvenile Atlantic cod (age group 0 and 1) was measured. The data were used to evaluate a method that enables the calculation of food quan- tities with high accuracy for incor- poration in trophic studies. Materials and methods Juvenile Atlantic cod (ages 0 and 1) were collected at Bellevue, Trinity Bay, Newfoundland, from July to December 1989 and from August to October 1991. Refer to Methven and Bajdik (1994) for additional information on fish collection and the collection site. A sample of 98 group-0 (40.2-100.1 mm standard length [SL] ) and group-1 (85.1- 192.0 mm SL) cod were used. Stom- ach contents were analyzed by us- ing prey volume to quantify the transition from pelagic to benthic prey. For each individual, prey were identified to the lowest taxon pos- sible, then counted. Displacement volume was measured for each taxon with a 5-mL graduated cyl- inder or a 200-pL micropipette. Prey volume was estimated as a proportion of the total volume of the pipette by using a millimeter ruler (200 pL measured 90.5 mm, i.e. 1 mm = approx. 2.2 pL). If a prey group was too small for volumetric displacement, measurements of length, width, and depth were taken for each individual prey item by using a dissecting microscope with an ocular micrometer. Mea- surements were made by using eye piece units and converted to milli- meters to calculate volumes with geometric formulae (Table 1). To evaluate the accuracy of cal- culated volumes, both displacement volume and calculated volume were measured for various prey items. Calculated volume tended to be higher than displacement volume by a constant rate, therefore linear regression was used to calibrate calculated volume against displace- ment volume. The intercept was not significantly different from zero (P=0.1453,F[1 33j= 1 16.43 ), therefore the parameters were re-estimated without the intercept (P=0.0001, F|i 33j = 258.02). The regression equation was Displacement volume = (0.72 )(calculated volume). A correction factor of 0.72 was ap- plied to all calculated volumes to obtain estimated displacement vol- umes. The following references were used to identify invertebrate prey items and to determine whether they were benthic or pelagic with respect to habitat: Smith, 1964; Allen, 1967; Russell-Hunter, 1969; Schultz, 1969; Feeley and Wass, 1971; Meglitsch, 1972; Naylor, 1972; Bousfield, 1973; and Gardner and Szabo, 1982. Scott and Scott (1988) was used to identify verte- brate prey. Percent pelagic prey in the diet was plotted against standard length to determine the relation between the two. Results Both group-0 and group- 1 cod fed on a broad range of prey. However, few occurred in large amounts Manuscript accepted 7 January 1988, Fish. Bull. 96:908-911 (1998), NOTE Lomond et a I.: Transition from pelagic to benthic prey of group 0-1 Gadus Morhua 909 Prey shapes and volume formulae used in Table 1 the study, spl = species length; spw = species width; spd = species depth Prey type and species Shape Formula Copepoda, Polychaeta, and shrimp and crab zoea cylinder V= 3.14 r2xspl r = spw/ 2 Amphipoda (straight) and Mysidicae cylinder V = 3.14 x r2 x spl r = [(spw + spd)/ 2]/2 or r = spw/ 2 Amphipoda (curved) 1/2 cylinder V = 3.14 x r2 x spw r = spl/ 2 Crab (adult) disc V = 3.14 x r2 x spd r = [(spl + spw)/ 2]/2 Crab (megalopa), Isopoda, and Ostracoda box V = spl x spw x spd Crustacean eggs and eyes sphere V = 4/3 x 3.14 r3 r = spl/ 2 Snail circular V= 1/3 x 3.14 r 2 x spd cone r = [(spl + spw)/ 2]/2 Table 2 Prey of age group 0-1 cod based on two measurement methods. B = Benthic, P = Mean number = number of individuals per stomach; mean volume = microliters Pelagic, U = Unknown, Unid. = Unidentified, per stomach. Method Group 0 Group 1 Mean number Calanoida (P) 22.7 Jaera marina (B) 37.9 Harpacticoida (B) 3.6 Eggs (U) 17.7 Unid. Copepoda (U) 3.2 Gammarus oceanicus (B) 3.0 Jaera marina (B) 2.7 Pontogeneia inermis (P) 3.0 Pontogeneia inermis (P) 2.1 Harpacticoida (B) 1.2 Other (U) 5.7 Other (U) 5.8 Mean volume Unid. Gammaridae (B) 27.4 Crangon septemspinosa (B) 297.7 Gammarus oceanicus (B) 24.9 Gammarus oceanicus (B) 98.4 Unid. Mysidacea (U) 12.9 Jaera marina (B) 84.0 Calliopius laeviusculus (P) 12.3 Urophycis tenuis (B) 62.5 Pontogeneia inermis (P) 11.2 Unid. Teleostei (U) 42.6 Unid. Copepoda (U) 10.7 Unid. Decapoda (U) 40.9 Calanoida (P) 6.5 Unid. Crustacea (U) 31.1 Jaera marina (B) 6.1 Unid Polychaeta 28.9 Other (U) 100.0 Other (U) 317.1 (Table 2). Group-0 cod fed predominately on amphi- pod taxa ( Pontogeneia inermis, Gammarus oceanicus , Calliopius laeviusculus, and unidentified Gammaridae) and copepod taxa (Calanoida and Harpacticoida). Pre- dominate prey of group- 1 cod were more diverse, in- cluding two amphipod taxa, an isopod (Jaera marina), invertebrate eggs, a shrimp ( Crangon septemspinosa ), white hake (Urophycis tenuis ), and other unidentified teleosts, decapods, and polychaete worms. Figure 1 indicates a rapid shift from pelagic to benthic prey at standard lengths of 60 to 100 mm. Four group-0 fish, represented by triangles in Fig- ure 1, had <5% pelagic prey. Prey of these fish con- sisted of a large volume of unidentified prey and a single benthic prey item (e.g. one ostracod or one am- phipod head). These fish were omitted in the calcu- lation of the mean proportion of pelagic prey for the three size groups of cod in Table 3. The rapid shift to benthic prey by juvenile Atlan- tic cod was related to body size. This change was quantified for use in food web computations. The diet of cod 40-59.9 mm SL was 98% pelagic prey, that of 910 Fishery Bulletin 96(4), 1998 cod 60-100 mm SL was 39% pelagic prey, and the diet of cod greater than 100 mm SL was 13% pelagic prey by volume. A simple explanation for this rapid shift is that the mouth opening of smaller size classes is too small to en- able predation upon many benthic prey species. However, it is important for juvenile cod to settle to benthic habi- tats as soon as possible to find protec- tion against predators. Therefore, they quickly shift to a benthic diet as soon as their gape size is large enough. A series of hypotheses were examined to decide whether the rapid shift in diet was associated with conditions of the study: time of day, time of year, and breakdown of the thermocline. The pat- tern of change in Figure 1 was not due to time of day, because cod were cap- tured primarily at night (18:10 to 04:30 hours). Data were then replotted by month of capture, which proved not to be responsible for the rapid shift in diet. Data were then replotted by date to in- vestigate relation to the breakdown of the ther- mocline in mid-October. It was hypothesized that the thermocline keeps pelagic and benthic prey well sepa- rated (i.e. pelagic prey are unavailable to demersal juveniles). After the breakdown of the thermocline, the pelagic prey may occur lower in the water col- umn. However, the change in percentage of pelagic prey did not change suddenly at the time of ther- mocline breakdown. Conclusions The diets of group-0 and group- 1 Atlantic cod at Trin- ity Bay, Newfoundland, are comparable to those of other areas (Daan, 1973 [northeastern Atlantic]; Arntz, 1974 [western Baltic]; Palsson, 1980 [Iceland]; Bowman, 1981 [western Atlantic]; Keats et al., 1987 [eastern Newfoundland]; Paz et al., 1991 [Flemish Cap]; Keats and Steele, 1992 [eastern Newfound- land]; Hop et al., 1992, 1994 [northeastern Atlantic]). Demersal group-0 cod fed on both pelagic and benthic organisms, whereas group-1 cod fed more on benthic organisms as the principal components of the diet. The diets of group-0 and group- 1 cod overlapped within a narrow range of body size (85.1 to 100.1 mm SL). Small pelagic prey, such as calanoid copepods, became less important in the diet of group- 1 cod, whereas larger benthic prey items, such as decapods, amphipods, te- leost fish, and polychaetes, became more important. Table 3 Percent pelagic prey in the diet of three size groups of ju- venile Atlantic cod. Size group (mm) Mean % pelagic prey Standard error Number of stomachs 40-59 97.56 2.44 8 60-100 38.5 5.32 45 >100 13.06 4.43 41 Results of the present study indicate that the on- togenetic shift from pelagic to benthic prey occurs within a narrow size range (from 60 to 100 mm SL) and hence more rapidly than previously believed (Daan, 1973; Bowman, 1981). This rapid shift may have been overlooked in previous studies that did not examine the diet of small cod in as much detail as in our study (i.e. volume in microliters of indi- vidual prey taxa). However, results of our study do not indicate that all cod less than 100 mm rely solely on benthic food. For example, it is well documented (e.g. Kohler and Fitzgerald, 1969) that, at particular times of the year, large juvenile and adult cod can forage far off the bottom on pelagic prey. One impli- cation of this finding is that any investigation of cod recruitment variability, in relation to food supply, should include postsettlement stages. NOTE Lomond et a\. : Transition from pelagic to benthic prey of group 0-1 Gadus Morhua 91 I Acknowledgments We thank Don Deibel, Sing Hoi Lee, Lori Meade, Kristine Miller, and Don H. Steele for assistance in prey identification; Rhonda Ford and Susan Forsey for providing the stomach-content analysis method and the majority of the data used in this study; Gavin Crutcher, John Horne, Daniel Ings, Dave Pinsent, and Tom Therriault for their help with various sta- tistical packages; and Derek Keats for comments on the manuscript. Financial support was provided by the Natural Sciences and Engineering Research Council of Canada (NSERC) and by the Northern Cod Science Program (Department of Fisheries and Oceans). Literature cited Allen, J. A. 1967. Crustacea: Euphausiacea and Decapoda. Scott. Mar. Biol. Assoc. Ann Rep., Millport, 116 p. Arntz, W. E. 1974. A contribution to the feeding of juvenile cod ( Gadus morhua) in the western Baltic. Rapp. R-V. Reun. Cons. Int. Explor. Mer 166:13-19. Bousfield, E. L. 1973. Shallow water Gammaridean amphipoda of New England. Cornell Univ. Press, New York. NY, 312 p. Bowman, R. E. 1981. Food of 10 species of northwest Atlantic juvenile groundfish. Fish. Bull. 79:201-206. Daan, N. 1973. A quantitative analysis of the food intake of North Sea cod, Gadus morhua. Neth. J. Sea Res. 6:479-517. Feeley, J. B., and M. L. Wass. 1971. The distribution and ecology of the Gammaridea (Crustacea: Amphipoda) of the lower Chesapeake estu- aries. Special Papers in Marine Science 2, Virginia Insti- tute of Marine Science, Gloucester Point, VA, 58 p. Gardner, G. A., and I. Szabo. 1982. British Columbia pelagic marine Copepoda: an iden- tification manual and annotated bibliography. Gov. of Canada Fisheries and Oceans, Ottawa, 536 p. Hawkins, A.D., N. M. Soofiani, and G. W. Smith. 1985. Growth and feeding of juvenile cod ( Gadus morhua L.). J. Cons. Int. Explor. Mer 42:11-32. Hop, H., J. Gjpsceter, and D. S. Danielssen. 1992. Seasonal feeding ecology of cod ( Gadus morhua L.) on the Norwegian Skagerrak coast. ICES J. Mar. Sci. 49:453-461. 1994. Dietary composition of sympatric juvenile cod, Ga- dus morhua L., and juvenile whiting, Merlangius merlangus L., in a fjord of southern Norway. Aquacult. Fish. Manage. 25:49-64. Keats, D. W., and D. H. Steele. 1992. Diurnal feeding of juvenile cod ( Gadus morhua) which migrate into shallow waters at night in eastern New- foundland. J. Northwest Atl. Fish. Sci. 13:7-14. Keats, D. W., D. H. Steele, and G. R. South. 1987. The role of fleshy macroalgae in the ecology of juve- nile cod (Gadus morhua L.) in inshore waters off eastern Newfoundland. Can. J. Zool. 65:49-53. Kohler, A. C., and D. N. Fitzgerald. 1969. Comparisons of food of cod and haddock in the Gulf of St. Lawrence and on the Nova Scotia Banks. J. Fish. Res. Board Canada 26:1273-1287. Meglitsch, P. A. 1972. Invertebrate zoology. Oxford Univ. Press, New York, NY, 834 p. Methven, D. A., and C. Bajdik. 1994. Temporal variation in size and abundance of juve- nile Atlantic cod (Gadus morhua ) at an inshore site off eastern Newfoundland. Can. J. Fish. Aquat. Sci. 51: 78-90. Naylor, E. 1972. British marine isopods. Academic Press Inc., Lon- don, 86 p. Palsson, O.K. 1980. On the biology of juvenile gadoids (age groups 0, I and II) in Icelandic waters. Meeresforschung Rep. Mar. Res. 28 (2-31:101-145. In German. Paz, J., M. Casas, and G. Perez-Gandaras. 1991. The feeding of cod ( Gadus morhua ) on Flemish Cap. Northwest Atlantic Fisheries Organization ( NAFO ) SCR Document 91/31, serial 1911, 19 p. Russell-Hunter, W. D. 1969. A biology of higher invertebrates. The Macmillan Company, New York, NY, 224 p. Schultz, G. A. 1969. How to know the marine isopod crustaceans. W.C. Brown Company, IA, 359 p. Scott, W. B., and M. G. Scott. 1988. Atlantic fishes of Canada. Can. Bull. Fish. Aquat. Sci. 219, 713 p. Smith, R. I. 1964. Keys to the marine invertebrates of the Woods Hole region. Spaulding Company, MA, 208 p. 912 Metazoan parasites as potential markers for selected Gulf of Alaska rockfishes Adam IVSoSes Jonathan Heifetz David C. Love Auke Bay Laboratory, Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 1 1305 Glacier Highway, Juneau, Alaska 99801-8626 E-mail address (for A. Moles): Adam.Moles@noaa.gov Rockfishes (family Scorpaenidae) of the genus Sebastes constitute some of the most important groundfish resources in the northeastern Pa- cific Ocean. In Alaska waters, shortraker rockfish, S. borealis , and rougheye rockfish, S. aleutianus, are especially prized because their large size and red color make them valuable on the Asian market. Both species reside in offshore waters of the upper continental slope, and they often co-occur. Commercial catches of these two species in the Gulf of Alaska (GOA) have aver- aged 1700 metric tons per year dur- ing this period, and the estimated exvessel value of the fishery in 1991 ( the most recent year with good eco- nomic data) was $3 million. Knowledge of stock structure and the degree of mixing among popu- lations is important for the ratio- nal management of commercially important marine species. At present, such information is lacking for shortraker and rougheye rockfish, and questions, such as whether these species perform lengthy migrations or whether discrete populations exist, are unanswered. Traditional tag-and-release experiments, which are often used to investigate stock structure, are not feasible for physoclistic fish such as rockfish because they cannot survive the barotrauma of initial capture. One alternative to human-implanted tags is the use of parasites as naturally occurring tags. Because necropsies of rockfishes are time consuming, we examined 100 fish of each species from the GOA to identify the feasi- bility of targeting selected parasites as tags for these species. Methods Adult shortraker and rougheye rockfish were captured in the GOA in summer 1990 during an annual longline survey conducted on the upper continental slope by the Na- tional Marine Fisheries Service. Sampling covered five statistical management areas established by the International North Pacific Fisheries Commission (Zenger and Sigler, 1992): Shumagin, Chirikof, Kodiak, Yakutat, and Southeast (Fig. 1). The survey area extended from the Island of the Four Moun- tains (52°50'N, 170°W) eastward to Dixon Entrance (54°29'N, 134°W). Forty-five stations were sampled along the upper continental slope at depths from 150 to 1000 m. Of the 7836 rockfishes of various spe- cies caught in the cruise, a sub- sample of 20 rougheye and 20 shortraker rockfish (21 from Shu- magin) from each management area were frozen on board for later necropsy. Total sample size was 101 shortraker and 100 rougheye rock- fish from 21 stations. In the laboratory, we conducted complete necropsies for metazoan parasites on an initial sample of 54 shortraker and 69 rougheye rock- fish. As a result of preliminary analysis of the parasite data from these fish, we restricted the necrop- sies on the remaining 47 shortraker and 31 rougheye rockfish to exami- nations for the presence of gill and fin copepods, monogenetic gill trematodes, and visceral acantho- cephalans. Identities of representa- tive parasite specimens were veri- fied by D. J. Whitaker of the Cana- dian Department of Fisheries and Oceans. Both prevalence (the pro- portion of fish with a given para- site) and intensity (the mean num- ber of parasites per infected fish) were calculated for each manage- ment area. Geographic trends were exam- ined by grouping samples into the five management areas. Categori- cal analysis of variance (SAS pro- cedure CATMOD; SAS Inst., 1989) was used to determine whether parasite prevalence differed among areas. For this analysis, the preva- lence of a particular parasite was used as the dependent variable, and area and fish size as indepen- dent variables. For each fish spe- cies, median fork length was used to divide the areawide sample into two groups, hereafter referred to as large and small. Data on parasite intensity for each fish were divided by fish length to account for the possible influence of fish size on parasite intensity. Distribution of the intensity data was highly non- normal, even after transformation. Therefore, intensity data were ana- lyzed by using the nonparametric Mann-Whitney U test. These analy- ses allowed us to account for differ- ences in fish size by area while test- Manuscript accepted 7 January 1998. Fish. Bull. 96:912-916 (1998). NOTE Moles et a I.: Metazoan parasites as markers for Sebastes 913 Northeastern Pacific Ocean, showing management areas of the Gulf of Alaska. Asterisks show sampling stations along the continental slope. ing for dependence of parasite prevalence and intensity by area. Fish size was cho- sen as a variable because the size of fish sampled in some areas differed, and fish size may influence parasite prevalence and intensity (Sekerak, 1975). A probabil- ity of 0. 10 or less was judged to be statis- tically significant. Results Seventeen species of parasites were found in our preliminary examinations: six species of copepod on the gills and fins and in the cephalic canal and nasal cavi- ties; two species of monogenetic trema- todes on the gills; three species of dige- netic trematodes, two species of acantho- cephalans, one species of cestode, and three species of nematodes in the viscera and mesenteries. All these species had been described previously from north- eastern Pacific rockfishes, although para- sites of shortraker and rougheye rockfishes in the western GOA had not been described previously. Many of the fish contained both larval nematodes Contracaecum spp. and Hysterothylacium spp. Be- cause it is time consuming to separate these two gen- era, they were lumped together as a single species as Contracaecum- type spp. On the basis of these pre- liminary results, nematodes, digenes, and cestodes were eliminated from further examination. These helminths were infrequent in the samples or showed no difference between management areas. Further examinations focused on enumeration of copepods, acanthocephalans, and monogenes. Prevalence (Table 1) and intensity (Table 2) of sev- eral parasite species differed distinctly among shortraker and rougheye rockfish populations in the GOA; they varied between areas, rather than being present in one area and absent in another. Three of these — Neobrachiella robusta (copepod), Trochopus trituba (monogenetic trematode), and Corynosoma sp. (acanthocephalan) — had the most distinctive changes in prevalence between areas. For shortraker rockfish, the highest prevalence of N. robusta and T. trituba parasites was in Kodiak samples. In rougheye rockfish, prevalence of all three parasites was sig- nificantly reduced in the Southeast. Several species of parasites were scarce in some areas and absent in others. Of the six species of copepods in our survey, the gill copepod N. robusta was sufficiently prevalent to be potentially useful in stock discrimination stud- ies. Prevalence of N. robusta differed significantly between areas for both shortraker (P=0.076) and rougheye (P=0.087) rockfish. Neobrachiella robusta was present in 80% of the Kodiak shortraker rock- fish sampled as opposed to 35-55% of those sampled in the other four areas. Larger rougheye rockfish also had a significantly (P=0.028) greater prevalence of N. robusta than smaller specimens. Among shortraker rockfish, intensities ranged from a mean of 2.4 in Shumagin to 7.6 in Chirikof; the differences were also significant between areas (P=0.029), especially when corrected for the size of the fish (P=0.016). Among rougheye rockfish, however, intensity of N. robusta infection did not differ between areas (P= 0.667). Mean infection levels were lower among rougheye rockfish than among shortraker rockfish for N. ro- busta but still averaged 1-5 parasites per fish. Other copepods were prevalent in the GOA samples. Naobranchia occidentalis was found in shortraker rockfish from Chirikof, Kodiak, and Yakutat in low prevalence (10%) and in rougheye rockfish from Yakutat (20%), but not from other ar- eas. Prevalence of Chondracanthus pinguis was above 40% in Yakutat and Southeast rougheye rock- fish, but below 10% in all other samples. This differ- ence, however, was a function of size rather than area. Prevalence was significantly greater (P=0.069) in large rougheye rockfish than in small ones. Colobo- matus kyphosus was present in one shortraker rock- fish from Yakutat and in two rougheye rockfish from Yakutat and Southeast. Chrondracanthus triventri- 914 Fishery Bulletin 96(4), 1998 Table 1 Comparison of parasite prevalence (proportion of fish having the parasite) by area for shortraker and rougheye rockfish in the Gulf of Alaska, 1990. The number of fish examined in each area is given in parentheses. Significance probability is given for a categorical analysis of variance model with area and fish size as main effects. Asterisks indicate P < 0.10. Parasite prevalence Significance probability Shumagin Chirikof Kodiak Yakutat Southeast Area Size Shortraker rockfish (n=21) (72 = 20) (77=20) o CM II (77=20) Neobrachiella robusta 0.48 0.55 0.80 0.35 0.40 0.076* 0.724 Naobranchia occidentalis 0.00 0.10 0.10 0.10 0.00 0.980 0.493 Chondracanthus pinguis 0.10 0.05 0.10 0.10 0.10 0.971 0.301 Colobomatus kyphosus 0.00 0.00 0.00 0.05 0.00 0.954 0.583 Trochopus trituba 0.62 0.60 0.86 0.55 0.45 0.081* 0.262 Microcotyle sebastis 0.14 0.10 0.00 0.00 0.00 0.927 0.655 Corynosoma sp. 0.81 0.75 0.60 0.65 0.65 0.606 0.914 Echinorhynchus gadi 0.10 0.00 0.00 0.05 0.00 0.929 0.633 Clavella parva 0.62 0.80 0.90 0.30 0.20 0.013* 0.370 Rougheye rockfish (72=20) O CM II (77 = 20) (77=20) (77=20) Neobrachiella robusta 0.25 0.40 0.30 0.50 0.10 0.087* 0.028* Naobranchia occidentalis 0.00 0.00 0.00 0.20 0.00 0.897 0.823 Chondracanthus pinguis 0.10 0.00 0.00 0.40 0.50 0.152 0.069* Colobomatus kyphosus 0.00 0.00 0.00 0.10 0.10 0.960 0.165 Trochopus trituba 0.70 0.60 0.55 0.50 0.20 0.045* 0.822 Microcotyle sebastis 0.15 0.00 0.00 0.15 0.00 0.906 0.895 Corynosoma sp. 0.80 0.75 0.75 0.75 0.20 0.001* 0.283 Echinorhynchus gadi 0.00 0.00 0.15 0.00 0.05 0.674 0.356 Clavella parva 0.48 0.10 0.10 0.70 0.55 0.105 0.704 cosus, a common parasite in the nasal cavities of Canadian rockfishes, was largely absent from our samples, occurring in only a single specimen. The fin copepod Clavella parva was present in all areas, with increased prevalence among shortraker rock- fish in the western GOA; this difference was signifi- cant (P=0.013) between areas. Although we exam- ined all parasites in our sample, C. parva was never a candidate for separation, because fin parasites could be lost in capture and sampling. The parasites showing the most potential for pro- viding insights into the population structure for both species of rockfishes were the monogenetic trema- todes T. trituba and Microcotyle sebastis, because of their high and low prevalences, respectively. Preva- lence and intensity of T. trituba were significantly lower for rougheye rockfish in the Southeast (20% prevalence, mean intensity of 1.8) than in other man- agement areas (50-70% prevalence, mean intensity of 2-9). For shortraker rockfish, prevalence and in- tensity of T. trituba were highest in Kodiak (86%, intensity of 16) and declined farther southward (45% and 8 in the Southeast, 62% and 4 in Shumagin). Both prevalence and intensities of T. trituba differed significantly among areas and did not depend on fish size. Microcotyle sebastis, common among many spe- cies of British Columbia rockfishes, was rare in the GOA. The only internal parasite showing potential as a tag was Corynosoma sp., which, with the less com- mon acanthocephalan Echinorhynchus gadi, infected nearly every Shumagin shortraker rockfish (90% in- fected with an acanthocephalan at an average of three Corynosoma sp. per fish). Most other shortraker rockfish were also infected (60-75%). Although prevalences of Corynosoma sp. in shortraker rock- fish did not differ among areas, intensities and in- tensities/cm differed significantly (P=0.024 and 0.019, respectively) by area, largely owing to high numbers of Corynosoma sp. among Yakutat fish. Rougheye rockfish were also infected with Cory- nosoma sp. at 75-80%, except in the Southeast (20%; significantly lower [P=0.001] than in other areas). Discussion Williams et al. (1992) proposed six criteria to assess the value of a parasite as a marker for stock separa- tion. The parasite should differ in levels of infection across geographic regions, not be detached easily with handling, be easily assessed, have no harmful effect NOTE Moles et a L Metazoan parasites as markers for Sebastes 915 Table 2 Comparison of mean parasite intensity by area for shortraker and rougheye rockfishes in the Gulf of Alaska, 1990. Range of intensity is in parentheses. Significance probability for differences in intensity and for intensity per cm fish length between areas is also given (Mann-Whitney U test). Asterisks indicate P < 0.10. Dash means insufficient number of infected fish to perform statistical tests. Int = intensity. Significance Mean parasite intensity (no. of parasites per infected fish) probability Shumagin Chirikof Kodiak Yakutat Southeast Intensity Int/cm Shortraker rockfish Neobrachiella robusta 2.4 (1-9) 7.6 (1-18) 4.9 (1-12) 4.4 (1-17) 3.8 (1-8) 0.029* 0.016* Naobranchia occidentalis 0 1 (1) 1 (1) 4.5 (4-5) 0 0.333 0.156 Chondracanthus pinguis 1 (1) 3 (3) 3.5 (3-4) 1 (1) 1.5 (1-2) 0.131 0.189 Colobomatus kyphosus 0 0 0 1 (1) 0 — — Trochopus trituba 3.8 (1-9) 4.8 (1-9) 16 (1-106) 17 (2-89) 8.3 (1-15) 0.001* 0.002* Microcotyle sebastis 1.3 (1-2) 2.5 (2-3) 0 0 0 0.133 0.083* Corynosoma sp. 3.3 (1-6) 2.3 (1-12) 2.7 (1-7) 5.6 (1-26) 3.7 (1-7) 0.024* 0.019* Echinorhynchus gadi 1 (1) 0 0 3 (3) 0 — 0.221 Clavella parva 1.2 (1-2) 1.5 (1-5) 1.5 (1-3) 1.5 (1-2 1.3 (1-2) 0.259 0.357 Rougheye rockfish Neobrachiella robusta 2.6 (1-4) 2.7 (1-7) 1.8 (1-4) 5.2 (1-32) 1 (1) 0.667 0.478 Naobranchia occidentalis 0 0 0 1 (1) 0 — — Chondracanthus pinguis 1 (1) 0 0 9.6 (1-61) 2.7 (1-5) 0.235 0.211 Colobomatus kyphosus 0 0 0 3 (2-4) 1.5 (1-2) 0.333 0.121 Trochopus trituba 4.6 (1-19) 1.7 (1-3) 2.1 (1-8) 9.1 (1-50) 1.8 (1-3) 0.023* 0.069* Microcotyle sebastis 1(1) 0 0 1 (1) 0 — — Corynosoma sp. 6 (1-17) 5.1 (1-15) 4.6 (1-8) 5 (1-17) 3.8 (1-7) 0.973 0.916 Echinorhynchus gadi 0 0 1 (1) 0 1(1) 1.000 0.180 Clavella parva 1.4 (1-3) 1 (1) 1 (1) 1.8 (1-3) 2.2 (1-5) 0.427 0.443 on the host, have a long life span, and be present in only one part of the host. Based on coincident samples, it is clear that the monogenetic trematodes and copepods on the gills of the rockfishes in our study meet all these criteria, as do the visceral acan- thocephalans. Gill parasites are protected by the operculum during capture, are sampled easily, and can be identified aboard ship. Notably, Leaman and Kabata ( 1987) previously proposed using N. robusta as a marker for separating stocks of S. alutus in Brit- ish Columbia. Acanthocephalans, such as Coryno- soma sp., also serve as excellent markers because they can be easily enumerated with a pepsin enzyme. In contrast, the lack of digenes and cestodes in the preliminary survey (probably due to regurgitation through barotrauma) makes these parasites poor candidates as tags for deepwater rockfishes. The nematodes are difficult to identify and their prevalences were similar throughout the Gulf of Alaska. In addition to their potential use in determining stock structure, differences in prevalence between areas among the parasites in this study also give insight on differences in diet and parasite distribu- tions. Corynosoma sp. is widely distributed in the northeastern Pacific Ocean as a parasite of marine mammals, birds, and fishes (Margolis, 1958); thus the different prevalences of Corynosoma sp. in our study were likely due to differences in diet rather than to parasite distribution. The intermediate hosts for Corynosoma sp. are amphipods; hence, the per- centage of amphipods in the diet may be higher in the western part of the GOA than in the eastern part. Alternatively, rougheye and shortraker rockfishes may be more likely to consume infected amphipods than other prey items. In contrast, both copepods and monogenes have no intermediate host stage, and dif- ferences in parasitism between management areas probably reflect differences in parasite distribution or host habitat, rather than differences in diet. Some of the parasite species reported in this study had very different prevalences than those reported for other species or locations of rockfishes. For ex- ample, Corynosoma sp. was less prevalent (<10%) in most species of British Columbia rockfishes (Sekerak, 1975; Stanley et al., 1992) than in our study. Corynosoma sp. may be more common in the GOA than in Canada or simply more prevalent in short- raker and rougheye rockfishes than in some other species. Sekerak (1975) examined 536 rockfishes of 916 Fishery Bulletin 96(4), 1998 26 species from British Columbia waters and the GOA: of the 40 Corynosoma sp. recovered, 12 were from rougheye rockfish and 23 were from the GOA. Size did not appear significant in parasite prevalences except for N. robusta and C. pinguis in- fection in rougheye rockfish. Prevalences of several species of parasites, including N. robusta, differed significantly among fish of the same size. Nor was size a factor in parasite intensities, despite the ob- servation of Sekerak (1975) that intensities of cope- pods and monogenetic trematodes increase with in- creasing fish size, largely due to increased gill sur- face area. Although the value of parasite tags would be enhanced by sampling similar-size fish, the sta- tistically significant differences among areas, despite size variation, provide an indication of the potential power of parasite markers for these species. In summary, the prevalence or intensity of N. ro- busta, T. trituba, and Corynosoma sp. may prove to be useful markers for population studies of GOA shortraker and rougheye rockfishes. It is interest- ing to note the major reduction in prevalence of all three parasites among Southeast rougheye rockfish. Possibly these fish may constitute a separate stock and support a localized fishery. More studies are needed to evaluate the effects of temporal and spa- tial variability on parasite prevalence and intensity, particularly among GOA rougheye rockfishes. Acknowledgments The authors wish to thank the crew of the vessel FV Ocean Prowler and the scientific staff that collected the samples, especially Harold Zenger and Mike Sigler. Ryan Scott provided laboratory assistance in sorting parasites, and helpful reviews were provided by Dave Clausen and Mike Sigler. Literature cited Leaman, B. M., and Z. Kabata. 1987. Neobraehiella robusta (Wilson, 1912) (Copepoda: Lernaepodidae) as a tag for identification of stocks of its host, Sebastes alutus (Gilbert, 1890) (Pisces: Teleostei). Can. J. Zool. 65: 2579-2582. Margolis, L. 1958. The occurrence of juvenile Corynosoma (Acantho- cephala) in Pacific salmon ( Oncorhynchus spp.). J. Fish. Res. Board Can. 15:983-990. SAS Institute, Inc. 1989. SAS/STAT user’s guide, version 6, 4th ed. SAS In- stitute, Inc., Cary, NC, 1,686 p. Sekerak, A. D. 1975. Parasites as indicators of populations and species of rockfishes ( Sebastes : Scorpaenidae) of the northeastern Pacific Ocean. Ph.D. diss., Univ. of Calgary, Alberta, 251 p. Stanley, R. D., D. L. Lee, and D. J. Whitaker. 1992. Parasites of yellowtail rockfish, Sebastes flavidus (Ayres, 1862) (Pisces: Teleostei), from the Pacific coast of North America as potential biological tags for stock identification. Can. J. Zool. 70:1086-1096. Williams, H. H., K. MacKenzie, and A. M. McCarthy. 1992. Parasites as biological indicators of the population biology, migrations, diet, and phylogenetics of fish. Rev. Fish Biol. Fish. 2:144-176. Zenger, H. H., Jr., and M. F. Sigler. 1992. Relative abundance of Gulf of Alaska sablefish and other groundfish based on National Marine Fisheries Ser- vice longline surveys, 1988-90. U.S. Dep. Commer., NOAATech. Memo. NMFS F/NWC-216, 103 p. 917 Catchability and retention of larval European anchovy, Engraulis encrasicolus, with bongo nets Stylianos Somarakis*' ** Barbara Catalano* Nikolaos Tsimenides*' ** * Institute of Marine Biology of Crete PO. Box 2214, 7 1 0 03 Iraklion, Crete, Greece ** University of Crete, Department of Biology PO. Box 1470, 711 10 Iraklion, Crete, Greece E-mail address (for S. Somarakis): somarak@talos.cc.uch.gr Whenever data on larval abun- dance are used in producing esti- mates or indices of stock size or re- cruitment, accurate length distri- butions are required. Factors that might bias sample distributions must be taken into account and, in broadscale ichthyoplankton sur- veys, the effect of environmental and behavioral factors on the con- tent of the samples or collections must be considered (Zweifel and Smith, 1981). The objective is to standardize sampling gear by ap- plying correction factors to make samples comparable (Smith and Richardson, 1977). A factor contributing to a possi- bly serious source of bias in larval fish collections is net avoidance. Larvae may be agile and capable of avoiding nets, with the general ef- fect of inaccurate abundance and mortality estimates (Clutter and Anraku, 1968). Catchability varies mainly with light regime and lar- val length. The visual stimulus of the sampling device is believed to be of crucial importance. This is indicated by numerous investiga- tions showing diel variation in avoidance, catches being signifi- cantly smaller during daylight than at night (Morse, 1989, and refer- ences therein). A second factor affecting larval fish collections is extrusion of cap- tured material through the net mesh. According to the “diagonal rule” (Saville, 1958; Smith et al., 1968), the maximum cross-sec- tional diameter of an organism must be greater than the mesh di- agonal if it is to be fully retained. However, the “diagonal rule” is of- ten too conservative (Lenarz, 1972; Colton et al., 1980). We examined the effects of time of day (day-night-twilight), fish length, and ontogeny on the catcha- bility of European anchovy (En- graulis encrasicolus), by using a 60- cm bongo net. We also investigated biases resulting from differential re- tention of larvae of different lengths, and, further, we examined the effect of correction of catch for net avoid- ance on estimation of mortality. Information on catchability and retention of anchovy with plankton nets is very limited and results are often contrasting. Catchability of European anchovy with bongo nets is unknown. Regarding retention, there is just a single study indicat- ing full retention of European an- chovy larvae with the 0.333-mm net (Aldebert et al., 1975), whose data are contradictory to Lo ( 1983) who has estimated a 0.63 retention rate for northern anchovy, Engraulis mordax, larvae <4.0-mm with a 0.333-mm net. Materials and methods From 1992 through 1995, five ichthyoplankton surveys were car- ried out in the Aegean Sea. A total of 474 stations were occupied in continental shelf and slope waters and areas characteristic of larval anchovy habitat (Fig. 1). Plankton and hydrographic sampling were performed at each station and cruise information is given in Table 1. Most stations were positive (i.e. at least one larva was captured). A 60-cm bongo net sampler (Hy- drobios) was used during all cruises. Mesh sizes on the sampler were 0.250-mm, and 0.500- or 0.335-mm, depending on the cruise (Table 1). The 0.250-mm mesh net is consid- ered to retain clupeoid eggs and larvae completely (Aldebert et al., 1975; Colton et al., 1980; Leslie and Timmins, 1989). Tows at two knots were double- oblique, within 5 m of the bottom to the surface, or from 120 m depth to the surface at deep stations. All tows consisted of a wire released to the desired depth and retrieved to the surface at standard speeds. The depth of the sampler could be moni- tored onboard at any time during the tow by means of a recording depthmeter attached to the sam- pler. Plastic codend buckets, with side windows of 155 cur covered with net gauze, were used in an ef- fort to minimize damage to larvae. Volumes filtered were calculated from a calibrated flowmeter in the mouth of each net. All samples were sorted in the laboratory and larvae were identi- fied to the lowest possible taxo- nomic level. Larval anchovies were counted and measured (notochord Manuscript accepted 14 January 1998. Fish. Bull. 96: 917-925 ( 1998). ' 918 Fishery Bulletin 96(4), 1998 Figure 1 Map of the study area showing the location of sampling stations ( black dots). A = Thracian Sea; B = Chalkidiki gulfs; C = Thermaikos Gulf; D = Gulfs of central and southeastern Greece. or standard length to the nearest 0.1 mm). If more than 100 specimens of anchovy were captured in a tow, 100 randomly selected larvae were measured, and subsequent length frequencies were raised to the total number of larvae. Lengths were rounded to 1-mm length groups (e.g. 2-2.99: 2.5 mm). Each station was assigned to day, night, or twilight hours according to the recorded time at the beginning of the tow. Twilight was designated as one hour before or after sunrise and sunset (Morse, 1989). The analysis of retention and catchability was made after pooling all available data and, subsequently, estimating mean catch per length group. For example, mean catch per m2 in each length category was calcu- lated for 227 day, 146 night, and 84 twilight positive stations (Table 1). Pooling data over large temporal and spatial regimes inte- grates areal and temporal heterogeneity in distribution and abundance, and results of the analyses represent average conditions (Hewitt, 1981; Morse, 1989). Negative stations (where no anchovy lar- vae were caught by either of two nets) were not included in our analysis ( sensu Hewitt, 1981); we assumed that they represent samples drawn from outside larval habitat. Given that the number of negative stations was very low in our data set ( 17 out of 474, of which 10 were from one cruise, 92ANC2, Table 1), inclusion or exclusion of such nega- tive stations was not expected to affect results. Catches were standardized to numbers per m3. This standardization is sufficient for comparing larval retention in paired differ- ent mesh-size nets but, in comparing day- time or twilight and nighttime catches, data had to be standardized to numbers per m2, by using maximum tow depth and volume of water filtered (Houde, 1977). Mean stan- dardized catches per length group were used to calculate the following ratios: day:night and twilight:night as well as 0.500-:0.250- mm mesh and 0.335-:0.250- mm mesh. Vari- ances of ratios were approximated as in Somerton and Kobayashi (1989). The calculation of mean catch and its variance fol- lowed the methods of Pennington ( 1983) for the delta- distribution of catch frequencies (see also Morse ( 1989)). The estimators based on the lognormal model (delta distribution) are more efficient for marine data than the usual sample estimators (Lo et al., 1992; Pennington, 1996). In particular, they provide rea- sonable estimates for data sets that contain isolated large catches. Anchovies collected during 1995 were further sorted into yolksac, preflexion, flexion, and postflexion larvae. Eye pigmentation, functional mouth, and the forma- tion of intestine (Regner, 1985; Palomera et al., 1988; Clarke, 1989) were used to distinguish yolksac from preflexion larvae. The flexion stage begins at initial notochord flexion and ends (postflexion stage starts) NOTt Somarakis et al.: Catchability and retention of larval Engraulis encrasicolus 919 Table 1 Cruise data. Subregions A, B, C (northern Aegean Sea), and D (central-southern Aegean Sea) are indicated in Figure 1 . n = number of stations. np = number of positive stations. D = daytime stations. N = nighttime stations. T = twilight stations. Cruise Date Subregion n nP D N T Mesh size (mm) Larval abundance/m2 92ANC1 16 Jul-3 Aug 1992 A-B-C 117 117 54 44 19 0.250, 0.500 97.25 92ANC2 3-9 Sep 1992 D 80 70 30 26 14 0.250, 0.500 21.02 93ANC3 7-14 Jul 1993 A-B-C 110 110 62 30 18 0.250, 0.500 233.91 94ANC4 19-24 Jan 1994 A-C 46 43 16 18 9 0.250, 0.335 114.55 95EP1 15-30 Jun 1995 A-B-C 121 117 65 28 24 0.250, 0.335 69.17 Total 474 457 227 146 84 when the posterior margin of the upper hypural plate is at 90° from the notochord axis (Moser, 1996). Additionally, to determine whether the inferences drawn regarding retention rates were plausible in terms of the “diagonal rule,” we plotted maximum head width values against standard length based on measurements of 74 staged anchovy larvae (for a justification of using head width, see Colton et al., 1980). In postyolksac larvae, maximum head width was the maximum body width and was greater than any body-depth measurement (Somarakis, unpubl. data). This is not the case for yolksac larvae because of the bulk of the yolk sac. Yolk sacs, however, may be easily compressed or crushed, hence, we consid- ered head width to be the significant dimension. Correction for net avoidance and mortality estimates In studies on mortality of European anchovy the usual practice has been to correct catches for net avoidance by using daymight catch ratios and the methods available for northern anchovy (e.g. Palomera and Lleonart, 1989) because catchability of European anchovy has been unknown. As a final step of this study we estimated mortality during 1994 and 1995 cruises without correcting catches for net avoidance, as well as after correction using three different methods: Method 1 The correction for avoidance of the net during day-light was calculated by the sinusoidal function (Hewitt and Methot, 1982): ( 1 + DN L ) (1 -DNl) (2nt\ f 1= — + — cos , 2 2 l 24 J where DNL = the midday-to-midnight catch ratio of L-length larvae; and t = the hour of the tow. DN j data used were those available for E. mordax.1 Method 2 The same as method 1, DNr data used were those calculated for E.encrasicolus in the present study. Method 3 The sinusoidal function was not used. Length-specific ratios of daymight and twilightmight catches calculated in the present study were used to correct catches of day and twilight stations. The correction for duration of each size class (Hewitt and Methot, 1982; Lo et al., 1989) was cal- culated from growth of postyolksac larvae in the sea measured by daily increments in otoliths (see also Somarakis et al., 1997a). Age-( micro-increment count )-at-length data and respective growth curves were available from 336 postyolksac larvae collected during the 1994 cruise, and 294 postyolksac larvae from the 1995 cruise. Significant spatial or inter- annual differences in growth of the European an- chovy larvae in the Aegean Sea can be found, which cannot be explained by temperature differences (Somarakis et al., 1997a; Somarakis et al., 1997b). Thus, we estimated mortality for only the 1994 and 1995 cruises for which otolith data were available. Mean corrected catches at length (i.e. length-spe- cific daily production of larvae per m2) were calcu- lated based on the lognormal model (delta distribu- tion— see above). Mortality was estimated for larvae >4- mm and <10-mm. The 3-3.99 size class was not used because it includes both yolksac and feeding larvae (the yolksac stage is characterized by differ- ent mortality rates than the feeding stage [Lo, 1985a; Lo, 1985b]). Mean length-specific daily production of larvae per m2 with its age (calculated from the 1 Lo, N. C. H. 1996. Southwest Fisheries Center, National Ma- rine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. Personal commun. 920 Fishery Bulletin 96(4), 1998 respective growth curve) constituted the database for nonlinear regression estimation of instantaneous mortality rate of postyolksac larvae in a simple ex- ponential model, Pt = P0 exp(-zt), where P = the daily production of larvae at age t; t = the age in days from the end of the yolksac stage; P() = the daily larval production at t = 0; and z = the daily instantaneous mortality rate. The simple exponential mortality model fitted the data very well (Watanabe and Lo, 1988). Mortality models resulting from the application of different methods of correcting (or not correcting) catches for light-induced net avoidance were com- pared by an analysis of the residual sum of squares (Chen et al., 1992). Results Estimates of mean catch No larvae >12-mm were caught in any of the 227 day-time tows. Estimates of mean catch per length group (Table 2) indicated that catches of >10-mm larvae were very low and their coefficients of varia- tion too great (generally >20%) to be useful in this study. Thus, our analysis was restricted to larvae <10-mm (see also Lenarz, 1972). Retention of anchovy larvae in the 0.335- and 0.500-mm mesh nets Results of the catch analysis indicate that the 0.500: 0.250-mm mesh catch ratio was less than one for lar- vae <7 mm but the 0. 335:0. 250-inm mesh ratio was not significantly different from one, at any length (Fig. 2). Data of maximum head width against standard length of the 74 anchovy larvae, along with the mesh diagonals of 0.500-, 0.335-, and 0.250-mm, are pre- sented in Figure 3. A comparison of head-width mea- surements with mesh diagonals indicates that the minimum lengths for complete retention in the 0.500- and 0.335-mm mesh nets were approximately 7.5 and 3.5 mm, respectively. The head-width measurements further suggest that to ensure full retention of lar- vae, it would be necessary to use netting with a mesh aperture of 0.250 mm or less. In terms of ontogenetic stages, the 0.335-mm mesh net is expected to retain all except the yolksac larvae, whereas the 0.500-mm mesh net seems to be inefficient for yolksac and Figure 2 Catch ratios of 0.500:0.250 mm (A) and 0.335:0.250 mm (B) mesh nets as a function of larval length. Bars indicate 95% confidence intervals. preflexion larvae. It therefore appears that the “diagonal rule” is conservative for yolksac larvae of the European anchovy, when head width is used as the maximum cross-sectional diameter of the larva. Differences in day versus night and twilight versus night catches Results of the catch analysis show a change in the daymight and twilightmight catch ratios at L - 6.5 mm (Fig. 4). In general, both ratios are practically one for larvae <6 mm, at which point they drop sub- stantially, continuing to decrease more or less lin- early thereafter. This decrease is stronger for the daymight than for the twilightmight catch ratio, the daymight ratio being significantly less than one from 6 mm onwards and greater than the corresponding twilightmight ratio. Thus, for larvae >6 mm, it seems NOTE Somarakis et a I.. Catchability and retention of larval Engraulis encrasicolus 921 Table 2 Estimates of mean catch and their coefficients of variation (in parentheses), by length class. Length (mm) 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.5 11.5 0.500:0.250 0.500 0.419 (0.079) 0.378 (0.067) 0.250 0.836 (0.084) 0.607 (0.071) 0.335:0.250 0.335 0.581 (0.067) 0.453 (0.067) 0.250 0.614 (0.068) 0.498 (0.076) day:night and twilight:night day 4.008 (0.067) 2.900 (0.072) night 4.002 (0.085) 2.775 (0.087) twilight 3.864 (0.115) 3.098 (0.113) 0.323 0.264 0.187 0.136 (0.062) (0.064) (0.072) (0.082) 0.408 0.329 0.230 0.139 (0.069) (0.072) (0.080) (0.101) 0.438 0.325 0.225 0.182 (0.068) (0.064) (0.077) (0.089) 0.427 0.332 0.241 0.184 (0.068) (0.069) (0.074) (0.085) 2.047 1.582 0.890 0.587 (0.075) (0.082) (0.093) (0.107) 2.090 1.512 1.352 1.003 (0.088) (0.102) (0.101) (0.111) 2.322 1.683 1.146 0.748 (0.114) (0.114) (0.129) (0.156) 0.078 0.054 0.020 0.012 (0.115) (0.131) (0.200) (0.248) 0.080 0.056 0.021 0.012 (0.128) (0.156) (0.233) (0.306) 0.098 0.062 0.023 0.013 (0.109) (0.132) (0.201) (0.249) 0.109 0.082 0.024 0.015 (0.110) (0.124) (0.232) (0.287) 0.326 0.201 0.059 0.043 (0.132) (0.167) (0.297) (0.335) 0.623 0.496 0.221 0.099 (0.129) (0.145) (0.191) (0.280) 0.454 0.319 0.161 0.101 (0.186) (0.217) (0.319) (0.397) that catchability changes with varying light condi- tions and length. On the other hand, the “sudden” change in catchability at 6.5 mm coincides with the onset of the development of the tail, i.e. the flexion stage (Fig. 5). Correction for avoidance and its effect on mortality estimates Estimates of length-specific larval production for the 1994 and 1995 cruises obtained by using the three different methods of adjusting data for net avoidance, or without any correction, were very similar (Table 3). Method 1, which was based on length-specific day:night catch ratio values calculated for Engraulis mordax, gave slightly greater values for larvae 4-6 mm, although differences were not statistically sig- nificant (overlapping confidence intervals). Consequently, analyses of the residual sum of squares showed that the resulting mortality curves (Table 4) were not statistically different (1994: F= 0.361, P> 0.5; 1995: F=0.780, P> 0.5). 1 1 1 1 1 r~ — i 1 1 1 1 0 1 23456789 10 11 12 Length (mm) Figure 3 Relation between standard length and maximum head width for larval anchovy. The horizontal broken lines indi- cate the mesh diagonals of 0.500-, 0.335-, and 0.250-mm nets. Solid squares = yolksac larvae. Open squares = preflexion larvae. Solid circles = flexion larvae. Open circles = postflexion larvae. Discussion Retention of larvae The results of this study indicate that the “diagonal rule” (maximum head width being considered as the maximum cross-sectional diameter of the larva) may adequately “predict” retention of anchovy larvae in the 0.500-mm mesh net but appears to be conservative in the case of the 0.335-mm mesh net. Specifically, yolksac larvae have a smaller head width than the diagonal of the 0.335-mm mesh net, but the latter seems to fully 922 Fishery Bulletin 96(4), 1998 1.6 1.4 JZ o 1.2 C3 O 1 JZ CD C 0.8 £ U) 0.6 I H 0.4 0.2 iri v~i iri m >n >n n ir, ro t" ir. vo (-•' ad o\ Length (mm) Figure 4 Day:night (A) and twilight:night (B) catch ratios as a func- tion of larval length. Bars indicate 95% confidence intervals. retain them. This is attributable to the fact that maxi- mum head width is not the maximum cross-sectional diameter of the larva, because of the bulk of the yolk sac. Our method of towing (at low speed, with plastic codend buckets) did not seem to cause damage to yolk sacs and resulted in full retention of yolksac larvae by the 0.335-mm mesh. However, the degree of extrusion strongly depends on filtration velocity (Smith and Richardson, 1977; Colton et al., 1980). Thus, at high towing speeds yolk sacs may easily be crushed. In his study on anchovy, Lenarz (1972) used depth of body at the insertion of the pectoral fin as the criti- cal dimension to compare with mesh diagonal and concluded that the mesh “diagonal rule” was too con- servative in the case of slowly towed nylon nets. However, body depth is not the critical dimension for clupeoids (Colton et al., 1980). The latter found no differences related to mesh size (0.253- and 0.333-mm mesh nets) in the length frequency distributions and abundances of larval herring in their collections at 1.5-knot towing speed. The mesh “diagonal rule” (with skull width as the critical dimension) was also conservative in their case, but for the length inter- val including only yolksac larvae. They found that most yolksac larvae caught in their slowly towed bongo nets, which were also equipped with plastic codend buckets, were undamaged. We feel that at a low, constant towing speed (1.5— 2.0 knots), bongo nets equipped with a 0.335-mm mesh net and plastic codend buckets are very effi- cient in retaining larvae of clupeoids. Net avoidance Changes in catchability with varying light conditions and larval length are demonstrated in this study. These changes begin at notochord flexion. If visual detection of the net is the primary cue for net avoid- ance, flexion and postflexion European anchovy lar- vae show the expected relationship of night>twilight> day catches (Morse, 1989). In another study, Murphy and Clutter ( 1972) com- pared catches of larval Hawaiian anchovy ( Stole - phorus purpureus ) taken by conventional towed coni- cal nets with catches taken by a miniature purse seine constructed of the same netting. The latter was considered to be a more effective sampler of the full size range of anchovy larvae than towed nets. In a way similar to the present study, larvae <6 mm long were captured with approximately equal efficiency by both the towed net and the seine (i.e. these larvae did not seem to avoid the towed net). For larger lar- NOTE Somarakis et a I.: Catchability and retention of larval Engraulis encrasicolus 923 Table 3 Estimates of length-specific larval production for the 1994 and 1995 cruises (and their standard errors in parentheses) with different methods of adjusting data for net avoidance. Method 1 is based on daymight catch ratio values available for Engraulis mordax, and methods 2 and 3 are based on daymight and twilightmight values calculated in the present study (see text for details). Cruise Correction Length (mm) 4.5 5.5 6.5 7.5 8.5 9.5 1994 none 1.984 1.655 1.169 0.893 0.518 0.413 (0.198) (0.198) (0.139) (0.127) (0.110) (0.093) method 1 2.142 1.793 1.281 0.972 0.569 0.450 (0.214) (0.212) (0.152) (0.138) (0.121) (0.102) method 2 1.984 1.655 1.231 0.944 0.553 0.445 (0.198) (0.198) (0.146) (0.134) (0.118) (0.101) method 3 1.984 1.655 1.258 0.979 0.573 0.475 (0.198) (0.198) (0.150) (0.140) (0.122) (0.109) 1995 none 1.537 1.235 0.928 0.690 0.472 0.341 (0.111) (0.090) (0.080) (0.071) (0.052) (0.046) method 1 1.721 1.409 1.076 0.722 0.559 0.399 (0.029) (0.015) (0.058) (0.045) (0.046) (0.047) method 2 1.537 1.235 1.009 0.681 0.531 0.390 (0.111) (0.090) (0.086) (0.062) (0.057) (0.051) method 3 1.537 1.235 1.026 0.706 0.549 0.412 (0.111) (0.090) (0.087) (0.064) (0.059) (0.054) vae, the efficiency of the towed net decreased with larval length relative to the seine. The change in catchability at flexion, observed in our study, can be attributed to a change in the swim- ming ability of larvae, which is associated with flex- ion. Batty ( 1984) found that Atlantic herring larvae, as they grow, change from an anguilliform mode of swimming to a subcarangiform mode of swimming. This change of swimming mode occurs as the caudal fin develops (at a body length of about 22 mm). Sub- sequently, Heath and Dunn (1990) using the large (5 m2), high-speed LOCHNESS sampler, found that the daymight catch differential increased with body length for larval herring in the North Sea, with a maximum fivefold difference between day and night catches for larvae >25mm. In another study, Osse and van den Boogaart (1995) reported results similar to those of Batty (1984) for common carp ( Cyprinus carpio). The morphological differentiation of the cau- dal fin, which begins at flexion and is accompanied by ossification of the caudal-fin rays and the caudal part of the notochord, closely parallels a change in the swimming mode from anguilliform to carangi- form. Burst-speed capability, which is believed to determine, in part, the ability of a larva to avoid plankton nets (Hunter, 1976), might also change with the development of the tail. Table 4 Estimated parameters of the exponential mortality mod- els (P,=P0exp(-zf )) for the 1994 and 1995 cruises. A differ- ent model was fitted for each different method of adjust- ing data for net avoidance (see text for details), r2 = coeffi- cient of determination. Cruise Correction Po z r2 1994 none 2.931 0.143 0.980 method 1 3.156 0.142 0.979 method 2 2.872 0.135 0.979 method 3 2.832 0.131 0.978 1995 none 2.195 0.136 0.992 method 1 2.463 0.135 0.989 method 2 2.139 0.126 0.989 method 3 2.109 0.121 0.989 Alternatively, changes in catchability may have a behavioral component. A substantial change in daymight catches of northern anchovy ( Engraulis tnordax ) with bongo nets occurs at approximately 11 mm (Fig. 2 in Hewitt and Methot, 1982). Flexion in this species occurs at around 11 mm (Watson and Sandknop, 1996). Laboratory and field studies (Hunter and Sanchez, 1976; Hewitt, 1981; Hunter 924 Fishery Bulletin 96(4), 1 998 and Coyne, 1982) have shown that, at about 11 mm, swimbladder inflation and schooling begins, where- upon patchiness increases rapidly. Our field obser- vations of European anchovy indicate initiation of swimbladder inflation at flexion. It is therefore pos- sible that schooling behavior also begins during this stage in E. enci'asicolus. School formation during the day and dispersion of schools during the night could result in low daymight catches for two main reasons: 1) larvae are less vul- nerable to plankton nets when in a school than when reacting individually, owing to the reduction in the reaction distance of the organisms in the school to the approaching net; and 2) increased patchiness resulting from schooling during the day produces a larger number of negative tows. In summary, field data on the European anchovy suggest the existence of an ontogenetic change in catchability with slowly towed plankton nets, which can be attributed to a respective change in the swim- ming ability, or a change in behavior (i.e. onset of schooling), or both. To our knowledge, the present study is the first to examine retention and catchability in a larval fish not only in terms of length but also of ontogeny. Sur- prisingly, this study also shows that, when sampling is undertaken on a 24-h basis, bias that results from light-induced avoidance of the net may not be great enough to affect significantly estimates of daily lar- val production or larval mortality. However, when only daytime sampling is undertaken, mean catches of postflexion larvae may well be biased and correc- tion factors may have to be applied. Size selectivity due to net avoidance by larvae is expected to lead to an overestimation of mortality rate, because older larvae are underrepresented in relation to younger larvae. In a recent simulation study, Somerton and Kobayashi (1992) have shown that approaches usually taken to eliminate the selection bias from sampled length frequencies (i.e. division of sampled length frequencies by length-specific es- timates of capture probability or elimination of the biased portion of the length distribution) may be only partially effective in reducing the bias in estimated mortality rates. The latter, as well as the present paper, suggests that the effectiveness of methods that attempt to correct for net avoidance is largely un- known. Their effectiveness has to be fully addressed. Acknowledgments This study was partially funded by an EU Study Project DG XIV (MED/91/011). We gratefully ac- knowledge the useful comments of three anonymous referees which substantially improved an initial ver- sion of this manuscript. We also thank. B. Nafpaktitis for his valued help and discussions and N. Lo for providing daymight ratios for Engraulis mordax. Literature cited Aldebert, Y., A. J. Dicenta, Y. Marinaro, and C. Piccinetti. 1975. Engins de peches pour F ichthyoplancton: essais comparatifs. Rev. Trav. Peches Marit. 39 (3):261— 277. Batty, R. S. 1984. Development of the swimming movements and mus- culature of larval herring (Clupea harengus). J. Exp. Biol. 110:217-229. Chen, Y., D. A. Jackson, and H. H. Harvey. 1992. A comparison of von Bertalanffy and polynomial func- tions in modeling fish growth data. Can. J. Fish. Aquat. Sci. 49:1228-1235. Clarke, T. A. 1989. Seasonal differences in spawning, egg size, and early development time of the Hawaiian anchovy or nehu, Encrasicholina purpurea. Fish Bull. 87:593-600. Clutter, R.I., and M. Anraku. 1968. Avoidance of samplers. In D. J. Tranter (ed. ), Zoop- lankton sampling, p. 57-76. UNESCO Monogr. Oceanogr. Methodol., vol. 2. Colton J. B., Jr., J. R. Green, R. R. Byron, and J .L. Frisella. 1980. Bongo net retention rates as affected by towing speed and mesh size. Can. J. Fish. Aquat. Sci. 37:606-623. Heath, M., and J. Dunn. 1990. Avoidance of a midwater frame trawl by herring larvae. J. Cons. Int. Explor. Mer 47:140-147. Hewitt, R. 1981. The value of pattern in the distribution of young fish. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 178:229-236. Hewitt R. P., and R. D. Methot. 1982. Distribution and mortality of northern anchovy lar- vae in 1978 and 1979. CalCOFI Rep. 23:226-245. Houde, E. D. 1977. Abundance and potential yield of the round herring Etrumeus teres and aspects of its life history in the east- ern Gulf of Mexico. Fish. Bull. 75:61-89. Hunter, J. R. 1976. Behavior and survival of northern anchovy Engraulis mordax larvae. CalCOFI Rep. 19:138-146. Hunter, J. R., and K. M. Coyne. 1982. The onset of schooling in northern anchovy larvae. CalCOFI Rep. 23:246-251. Hunter, J. R., and C. Sanchez. 1976. Diel changes in swim bladder inflation of the larvae of the northern anchovy, Engraulis mordax. Fish. Bull. 74:847-855. Lenarz, W.H. 1972. Mesh retention of Sardinops caerulea and Engraulis mordax by plankton nets. Fish. Bull. 70:839-848. Leslie, J. K., and C. A. Timmins. 1989. Double nets for mesh aperture selection and sam- pling in ichthyoplankton studies. Fish. Res. 7:225-232. Lo, N. C. H. 1983. Re-estimation of three parameters associated with anchovy egg and larval abundance: Temperature depen- dent incubation time, yolk-sac growth rate and egg and NOTE Somarakis et a I.: Catchability and retention of larval Engraulis encrasicolus 925 larval retention in mesh nets. NOAATech. Memo. NMFS- SWFC-31, 32 p. 1985a. Egg production of the central stock of northern an- chovy, Engraulis mordax, 1951-82. Fish. Bull. 83:137-150. 1985b. Modeling life-stage-specific instantaneous mortal- ity rates, an application to northern anchovy, Engraulis mordax , eggs and larvae. Fish. Bull. 84:395-407. Lo, N. C. H., J. R. Hunter, and R. P. Hewitt. 1989. Precision and bias of estimates of larval mor- tality. Fish. Bull. 87:399-416. Lo, N. C. H., L. D. Jacobson, and J. L. Squire. 1992. Indices of relative abundance from fish spotter data based on delta-lognormal models. Can. J. Fish. Aquat. Sci. 49:2515-2526. Morse, W. W. 1989. Catchability, growth, and mortality of larval fishes. Fish. Bull. 87:417-446. Moser, G. H. (editor). 1996. The early stages of fishes in the California Current region. CalCOFI Atlas 33. 1505 p. Murphy, G. I., and R. I. Clutter. 1972. Sampling anchovy larvae with a plankton purse seine. Fish. Bull. 70:789-798. Osse, J. W. M., and J. G. M. van den Boogaart. 1995. Fish larvae, development, allometric growth, and the aquatic environment. ICES Mar. Sci. Symp. 201:21-34. Palomera, I., and J. Lleonart. 1989. F ield mortality estimates of anchovy larvae, Engraulis encrasicolus, in the western Mediterranean. J. Fish Biol. 35 (suppl. A):133-138. Palomera, I., B. Morales-Nin, and J. Lleonart. 1988. Larval growth of anchovy, Engraulis encrasicolus , in the western Mediterranean Sea. Mar. Biol. 99:283-291. Pennington, M. 1983. Efficient estimators of abundance, for fish and plank- ton surveys. Biometrics 39: 281-286. 1996. Estimating the mean and variance from highly skewed marine data. Fish. Bull. 94:498-505. Regner, S. 1985. Ecology of planktonic stages of the anchovy, Engraulis encrasicolus (Linnaeus, 1758), in the central Adriatic. Acta Adriat. 26:5-113. Saville, A. 1958. Mesh selection of plankton nets. J. Cons. Int. Explor. Mer 23:192-201. Smith, P. E., R. C. Counts, and R. I. Clutter. 1968. Changes in filtering efficiency of plankton nets due to clogging under tow. J. Cons. Int. Explor. Mer 32:232-248. Smith, P. E., and S. L. Richardson. 1977. Standard techniques for pelagic fish egg and larva surveys. FAO Fish. Tech. Pap. 175, 100 p. Somarakis, S., I. Kostikas, N. Peristeraki, and N. Tsimenides. 1997a. Fluctuating asymmetry in the otoliths of larval an- chovy ( Engraulis encrasicolus ) and the use of developmen- tal instability as an indicator of condition in larval fish. Mar. Ecol. Progr. Ser. 151:191-203. Somarakis, N. Peristeraki, and N. Tsimenides. 1997b. Otolith microstructure and growth of larval anchovy ( Engraulis encrasicolus) in the northern Evoikos Gulf(east- em Mediterranean). Proc. 5th Hellenic Symp. Ocean. Fish. Natl. Cent. Mar. Res., Athens, vol. 2, p. 51-54. Somerton, D. A., and D. R. Kobayashi. 1989. A method for correcting catches of fish larvae for the size selection of plankton nets. Fish. Bull. 87:447-455. 1992. Inverse method for mortality and growth estimation: a new method for larval fishes. Fish. Bull. 90:368-375. Watanabe, Y., and N. C. H. Lo. 1988. Larval production and mortality of Pacific saury, Cololabis saira , in the northwestern Pacific ocean. Fish. Bull. 78:601-613. Watson, W., and E. M. Sandknop. 1996. Engraulidae: Anchovies. In H. G. Moser (editor), The early stages of fishes in the California Current region, p. 173-183. CalCOFI Atlas 33. Zweifel, J. R., and P. E. Smith. 1981. Estimates of abundance and mortality of larval an- chovies ( 1951-75): application of a new method. Rapp. R- V. Reun. Cons. Int. Explor. Mer 178:248-259. 926 Fishery Bulletin 96(4), 1998 1 998 Reviewers The editorial staff of Fishery Bulletin would like to thank the following referees for their time and efforts in providing reviews of the manuscripts published in 1998. Their contributions have helped ensure the publication of quality science. Dr. Peter B. Adams Dr. Dean W. Ahrenholz Dr. C. E. Alexander Dr. Larry G. Allen Dr. M. James Allen Dr. Sabine Alshuth Dr. John Anderson Dr. Neil L. Andrew Dr. Bernard Angers Dr. Richard S. Appeldoorn Dr Michael P. Armstrong Dr. Freddy Arocha Dr. Francisco Arreguin-Sanchez Mr. Steven M. Atran Dr. Whitlow W. L. Au Dr. Susan G. Ayvazian Dr. C. Scott Baker Dr. Luiz R. Barbieri Mr. Robert E. Bayley Dr. William H. Bayliff Dr. Richard J. Beamish Dr. Allen J. Bejda Dr. James R. Bence Dr. Daniel D. Benetti Dr. R. A. Bergstedt Dr. Steve Berkeley Dr. M. Bertignac Dr. Keith A. Bigelow Dr. Gustavo A. Bisbal Dr. Arne Bjorge Dr. Karen A. Bjorndal Dr. Barbara Block Dr. George W. Boehlert Dr. Anne A. Boettcher Dr. Christofer H. Boggs Dr. James A. Bohnsack Dr. George R. Bolz Dr. John Boreman Dr. S. A. Bortone Dr. Susan M. Bower Ms. Carole Bradbury Dr. Rod Bradford Dr. Steven Branstetter Dr. Richard W. Brill Dr. Matt K. Broadhurst Dr. Richard E. Brock Dr. William B. Brodie Dr. Edward B. Brothers Dr. M. L. Brown Dr. Nancy Brown-Peterson Dr. Barry D. Bruce Dr. Christopher Bublitz Dr. Jeffrey A. Buckel Dr. Lawrence J. Buckley Dr. Ronald S. Burton Dr. John L. Butler Dr. Mark J. Butler Dr. John F. Caddy Dr. Steven X. Cadrin Dr. Gregor M. Cailliet Dr. Charles W. Caillouet Dr. John Calambokidis Dr. Steven E. Campana Dr. Enrique de Cardenas Dr. Mark H. Carr Dr. Jose I. Castro Mr. Albert E. Caton Mr. Salvatore Cerchio Dr. Yong Chen Dr. Robert Cheney Dr. Anthony R. Child Dr. Seinen Chow Dr. Dezhang Chu Dr. Michael Clancy Dr. Malcolm R. Clark Dr. V. G. Cockcroft Dr. Richard P. Cody Dr. Pat L. Colin Dr. Bruce B. Collette Dr. James A. Colvocoresses Dr. M. Comeau Dr. Bruce H. Comyns Dr. Justin G. Cooke Dr. John E. Cooper Dr. Enric Cortes Dr. Charles C. Coutant Dr. Roy E. Crabtree Dr. Peter C. Craig Dr. Jean Cramer Dr. Laurence W. Crim Dr. Matthew A. Cronin Dr. Jeffrey N. Cross Dr. Nancie J. Cummings Dr. Michael J. Dagg Dr. Erdmann Dahm Dr. Louis B. Daniel Mr. Andrew W. David Dr. Earl G. Dawe Dr. J. T. DeAlteris Dr. Omar Defeo Dr. Donald J. Degan Dr. Edward E. DeMartini Dr. D. Demer Dr. M. Demestre Dr. J. Brian Dempson Dr. C. Braxton Dew Dr. W.J. Diehl Dr. Andrew E. Dizon Dr. P. J. Doherty Dr. Michael L. Domeier Dr. Ken F. Drinkwater Dr. Brett R. Dumbauld Dr. Dennis J. Dunning Dr. Nicholas G. Elliott Dr. Denise M. Ellis Dr. Charles E. Epifanio Dr. Bruce T. Estrella Dr. Mary Fabrizio Dr. Beatrice Padovani Ferreira Dr. Joseph P. Fisher Dr. William L. Fisher Dr. Gary R. Fitzhugh Dr. Clark T. Fontaine Dr. Michael Foreman Dr. Karin A. Forney Dr. Richard B. Forward Dr. Clive Fox Dr. David R. Franz Dr. Kathryn J. Frost Dr. Alejandro Gallego Dr. Charles E. Gates Dr. Stratis Gavaris Dr. Frances P. Gel wick Dr. Joseph R. Geraci Dr. F. Gerlotto Dr. Tony Gharrett Dr. M. F. Gil Dr. James R. Gilbert Dr. Kenneth J. Goldman Dr. A. D. Goodson Dr. K. V. Gorchinsky Dr. Harry K. Gorfine Dr. C.A. Goudey Dr. John J. Govoni Dr. John E. Graves Dr. Churchill B. Grimes Dr. E. Grunwald Dr. Donald R. Gunderson Dr. Wayne R. Haight Dr. Lewis J. Haldorson Dr. Martin Hall Mr. Norm G. Hall Dr. Harlyn O. Halvorson Dr. Philip S. Hammond Dr. John Hampton Dr. Michael J. Hansen Dr. Ronald W. Hardy Dr. Patrick J. Harris Mr. Karsten Hartel Dr. Jonathan Heifetz Dr. Edward J. Heist Dr. Thomas E. Helser Dr. Andrew Hendry Dr. Steven W. Hewett Dr. Roger P. Hewitt Mr. Y. Hiyama Dr. Wayne Hoggard Dr. Aleta A. Hohn Dr. Kim N. Holland Dr. David B. Holts Dr. Peter B. Hood Dr. Donald E. Hoss Dr. Edward D. Houde Dr. W. Huntting Howell Dr. John R. Hunter Dr. Gene R. Huntsman Dr. James N. Ianelli Dr. Elizabeth A. Irlandi Dr. Larry D. Jacobson Dr. Alan Jamieson 927 Dr. Thomas Jefferson Dr. Gregory P. Jenkins Dr. Simon Jennings Dr. A. L. Jensen Dr. Allyn G. Johnson Dr. James H. Johnson Dr. Lyndal L. Johnson Dr. I. A. Johnston Dr. Christopher D. Jones Dr. Cynthia M. Jones Dr. Susana Junquera Dr. A. Kanazawa Dr. Joseph Kane Dr. R.A. Kastelein Dr. Steven K. Katona Dr. Derek W. Keats Dr. Arthur W. Kendall, Jr. Dr. Robert D. Kenney Dr. Michael J. Kennish Dr. Daniel K. Kimura Dr. Stephen T. Kinsey Dr. G. P. Kirkwood Dr. Shuichi Kitada Dr. Elin Kjorsvik Dr. Grace Klein-Macphee Dr. Brian Knights Dr. T. Kobayashi Dr. Christopher C. Koenig Dr. Nancy E. Kohler Dr. E. T. Koutrakis Dr. David I. Kusher Dr. Marc Labelle Dr. Thomas E. Laidig Lt. Cdr. John T. Lamkin Dr. Geoff Lang Dr. Richard W. Langton Dr. Wayne A. Laroche Dr. Mark A. Lazzari Dr. Kenneth M. Leber Dr. Christopher M. Legault Dr. William H. Lenarz Dr. Phillip S. Levin Dr. Colin D. Levings Dr. John Lien Dr. George Lilly Dr. Karin E. Limburg Dr. Junda Lin Dr. Romuald Lipcius Dr. Robert Gregory Lough Dr. D. C. Love Dr. Milton S. Love Dr. Enrique Lozano-Alvarez Dr. Molly E. Lutcavage Dr. D.L. Mackas Dr. William K. Macy Dr. Gary L. Maillet Ms. Nancy E. Maloney Dr. Charles S. Manooch, III Dr. Daniel Margulies Dr. Douglas F. Markle Dr. Michael H. Martin Dr. Ann C. Matarese Dr. C. P. Mathews Dr. Michael Maxwell Dr. Tim McClanahan Dr. Gordon A. McFarlane Dr. Jeffrey S. McKinnon Dr. Charmion B. McMillan Dr. A. L. McMillen-Jackson Dr. Paul E. McShane Dr. Richard D. Methot Dr. Bill Michaels Dr. Russell B. Millar Dr. John M. Miller Dr. Lisa R. Mills Dr. Ole A. Misund Mr. Raymond Mojica Dr. Beatriz Morales-Nin Dr. H. Geoffrey Moser Dr. David Mountain Mr. Karl W. Mueller Dr. Ashley Mullen Dr. Timothy J. Mulligan Dr. M. Munekiyo Dr. Debra J. Murie Mr. Michael D. Murphy Dr. Ransom A. Myers Dr. Motohiko Nakamura Dr. T. Nakamura Dr. Hideki Nakano Dr. Lisa J. Natanson Dr. Kjell Harald Nedreaas Dr. Joseph S. Nelson Dr. Russell S. Nelson Mr. Daniel G. Nichol Dr. David L. Nieland Dr. Jun Nishikawa Dr. Yasuo Nishikawa Dr. Sergey A. Nizyaev Dr. G. Nonnotte Dr. Loretta O’Brien Dr. Ron K. O’Dor Dr. Debra L. Oda Dr. Kenneth Oliveira Dr. Robert J. Olson Dr. R.W. Osman Dr. William J. Overholtz Dr. Hazel A. Oxenford Dr. Debbie Palka Dr. Isabel Palomera Dr. N. V. Parin Dr. Richard O. Parker, Jr. Dr. James D. Parrish Dr. Santiago Pascual Dr. Daniel Pauly Dr. William G. Pearcy Mr. Peter C. Perkins Dr. William F. Perrin Dr. R. Ian Perry Dr. Edward J. Pfeiler Dr. Bruce F. Phillips Dr. Grant H. Pogson Dr. Thomas Polacheck Dr. Jeffrey J. Polovina Dr. Clay E. Porch Dr. Julie M. Porter Dr. Michael H. Prager Dr. Harold L. Pratt, Jr. Dr. Eric D. Prince Dr. Andre E. Punt Dr. Thomas P. Quinn Dr. Richard L. Radtke Dr. Stephen Ralston Dr. Peter S. Rand Dr. John E. Randall Dr. Robert W. Rangeley Dr. Andrew J. Read Dr. Marcel J. Reichert Dr. Stewart B. Reid Dr. Stephen B. Reilly Dr. Maurice L. Renaud Dr. Victor R. Restrepo Dr. Jake Rice Dr. James A. Rice Dr. William J. Richards Dr. B. E. Rieman Dr. Roger E. Robbins Dr. Denice Robertson Dr. Kelly M. Robertson Dr. Paul G. Rodhouse Dr. Stuart Rogers Dr. Patricia E. Rosel Dr. Andrew A. Rosenberg Dr. Steven W. Ross Dr. Brian Rothschild Dr. A. A. Rowden Dr. Gary R. Russ Dr. Thomas L. Rutecki Dr. Clifford H. Ryer Dr. Yvonne Sadovy Dr. Susan E. Safford Dr. Gary T. Sakagawa Dr. David B. Sampson Dr. H. C. Schaefer Dr. David R. Schiel Dr. Carl James Schwarz Dr. Jake F. Schweigert Dr. Michael D. Scott Dr. Kim Scribner Dr. David H. Secor Dr. George R. Sedberry Dr. James E. Seeb Dr. Joseph E. Serafy Dr. Fredric M. Serchuk Dr. Douglas Y. Shapiro Dr. Richard F. Shaw Dr. Jonathan M. Shenker Dr. Michiyo Shima Dr. Fred Short Dr. Colin Simpfendorfer Dr. Carl Sindermann Dr. Thomas R. Sminkey Dr. Steve J. Smith Dr. Tim D. Smith Dr. Rachel Smolker Dr. Susan M. Sogard Dr. David A. Somerton Dr. R. D. Stanley Dr. D. H. Steele Dr. John Steele Dr. Gretchen Steiger Dr. Frank W. Steimle Dr. Carol A. Stepien Dr. John D. Stevens Dr. Heath H. Stone Dr. Patrick J. Sullivan Dr. Steve Swartz Dr W. Mark Swingle 928 Fishery Bulletin 96(4), 1998 Dr. Stephen T. Szedlmayer Dr. Barbara L. Taylor Dr. Mark Terceiro Dr. Grant G. Thompson Dr. Ronald E. Thresher Dr. Brian N. Tissot Dr. David W. Townsend Dr. E. A. Trippel Dr. Peter L. Tyack Dr. Albert V. Tyler Dr. James H. Uchiyama Dr. Franz Uiblein Dr. Yuji Uozumi Dr. Fred M. Utter Dr. Douglas S. Vaughan Dr. Maria B. Santos Vazquez Dr. Michael Vecchione Dr. Roger Villanueva Dr. I. H. Von Herbing Dr. W. Waldo Wakefield Dr. Randal L. Walker Dr. William A. Walker Dr. Julie E. Wallin Dr. Stephen J. Walsh Dr. Carl J. Walters Dr. Robert D. Ward Dr. Gordon T. Waring Dr. William G. Warren Dr. R. M. Warwick Dr. Reg A. Watson Mr. William Watson Dr. Christopher R. Weidman Dr. James R. Weinberg Dr. Vidar G. Wespestad Mr. Andrew J. Westgate Dr. Jerry A. Wetherall Dr. Jeanne B. Wexler Ms. Susan E. Wigley Dr. David N. Wiley Dr. Stuart J. Wilk Dr. Mark E. Wilkins Dr. Howel Williams Dr. Charles A. Wilson Dr. Christopher D. Wilson Dr. George H. Winters Dr. Sabine Petra Wintner Mr. Robert L. Wisner Dr. Duncan G. Worthington Dr. Scott D. Wright Dr. Tina Wyllie Echeverria Dr. Xucai Xu Dr. Mei-Sun Yang Dr. Zhen Ye Dr. Mary M. Yoklavich Dr. D. D. Young Dr. A. Zebdi Dr. H. Zenitani Dr. Mark Zimmermann 929 Fishery Bulletin Index Volume 96 (1-4), 1998 List of titles 96 (1) 1 Systematics and ecology of tonguefishes of the genus Symphurus (Cynoglossidae: Pleuronectiformes) from the western Atlantic Ocean, by Thomas A. Munroe 96 (2) 185 Radiocarbon from nuclear testing applied to age validation of black drum, Pogonias cromis, by Steven E. Campana and Cynthia M. Jones 193 Discerning patterns in patchy data: a categorical approach using gulf menhaden, Brevoortia patronus, bycatch, by Janaka A. de Silva and Richard E. Condrey 210 Estimated tuna discard from dolphin, school, and log sets in the eastern tropical Pacific Ocean, 1989-1992, by Eliza- beth F. Edwards and Peter C. Perkins 223 Reproductive dynamics of southern bluefin tuna, Thunnus maccoyii, by Jessica H. Farley and Tim L.O. Davis 237 Dolphin prey abundance determined from acoustic back- scatter data in eastern Pacific surveys, by Paul C. Fiedler, Jay Barlow, and Tim Gerrodette 248 Feeding habits of pelagic summer flounder, Paralichthys dentatus, larvae in oceanic and estuarine habitats, by Jill J. Grover 258 Age composition, growth, reproductive biology, and recruit- ment of King George whiting, Sillaginodes punctata, in coastal waters of southwestern Australia, by Glenn A. Hyndes, Margaret E. Platell, Ian C. Potter, and Rodney C. J. Lenanton 271 Estimates of marine mammal, turtle, and seabird mortal- ity for two California gillnet fisheries: 1990-1995, by Fred Julian and Marilyn Beeson 285 Feeding habits of juvenile Pacific salmon in marine waters of southeastern Alaska and northern British Columbia, by Joyce H. Landingham, Molly V. Sturdevant, and Richard D. Brodeur 303 Interspecific comparisons ofsearobin ( Prionotus spp.) move- ments, size structure, and abundance in the temperate west- ern North Atlantic, by Richard S. McBride, Joseph B. O’Gorman, and Kenneth W. Able 315 Abundance, growth, and mortality of young-of-the-year pin- fish, Lagodon rhomboides, in three estuaries along the gulf coast of Florida, by Gary A. Nelson 329 Low-frequency acoustic measurements of Pacific hake, Merluccius productus, off the west coast of the United States, by Redwood W. Nero, Charles H. Thompson, and Richard H. Love 344 Population dynamics and stock size prediction for the sunray surfclam, Mactra chinensis, at Tomakomai, south- west Hokkaido, Japan, by Izumi Sakurai, Takashi Horii, Osamu Murakami, and Shigeru Nakao 352 Reproductive biology, growth, and natural mortality of Puget Sound rockfish, Sebastes emphaeus (Starks, 1911), by Andreas T. Beckmann, Donald R. Gunderson, Bruce S. Miller, Raymond M. Buckley, and Betty Goetz 357 Age, growth, and calving season of bottlenose dolphins, Tursiops truncatus, off coastal Texas, by Stephanie Fernandez and Aleta A. Hohn 366 Diet of dusky dolphins, Lagenorhynchus obscurus, in wa- ters off Patagonia, Argentina, by Mariano Koen Alonso, Enrique Alberto Crespo, Nestor Anibal Garcia, Susana Noemi Pedraza, and Mariano Alberto Coscarella 375 Preliminary estimate of spawning frequency and batch fe- cundity of striped weakfish, Cynoscion striatus, in coastal waters off Buenos Aires province, by Gustavo J. Macchi 382 Direct validation of ages determined for adult black drum, Pogonias cromis, in east- central Florida, with notes on black drum migration, by Michael D. Murphy, Douglas H. Adams, Derek M. Tremain, and Brent L. Winner 388 The influence of spear fishing on species composition and size of groupers on patch reefs in the upper Florida Keys, by Robert D. Sluka and Kathleen M. Sullivan 96(3) 395 A retrospective (1979-1996) multispecies assessment of coral reef fish stocks in the Florida Keys, by Jerald S. Ault, James A. Bohnsack, and Geoffrey A. Meester 415 Reproductive patterns, sex ratio, and fecundity in gag, Mycteroperca microlepis (Serranidae), a protogynous grouper from the northeastern Gulf of Mexico, by L. Alan Collins, Allyn G. Johnson, Christopher C. Koenig, and M. Scott Baker Jr. 428 Autumn food habits of harbor porpoises, Phocoena phocoena, in the Gulf of Maine, by Damon P. Gannon, James E. Craddock, and Andrew J. Read 438 A comparison of two underwater census methods for esti- mating the abundance of the commercially important blacklip abalone, Haliotis rubra, by Harry K. Gorfine, David A. Forbes, and Anne S. Gason 451 Age, growth, and mortality of black drum, Pogonias cromis, in the Chesapeake Bay region, by Cynthia M. Jones and Brian Wells 462 Stock structure and movement of tagged sablefish, Anoplopoma fimbria, in offshore northeast Pacific waters and the effects of El Nino-Southern Oscillation on migra- tion and growth, by Daniel K. Kimura, Allen M. Shimada, and Franklin R. Shaw 930 Fishery Bulletin 96(4), 1998 482 Age and growth estimates of the bigeye thresher shark, Alopias superciliosus, in northeastern Taiwan waters, by Kwang-Ming Liu, Po-Jen Chiang, and Che-Tsung Chen 492 Declines in nearshore rockfish recruitment and populations in the southern California Bight as measured by impinge- ment rates in coastal electrical power generating stations, by Milton S. Love, Jennifer E. Caselle, and Kevin Herbinson 502 Geographic variation in genetic and growth patterns of Atka mackerel Pleurogrammus monopterygius (Hexagrammidae), in the Aleutian archipelago, by Sandra A. Lowe, Donald M. Van Doornik, and Gary A. Winans 516 A population profile for hagfish, Myxine glutinosa, in the Gulf of Maine. Part 2: Morphological variation in popula- tions of Myxine in the North Atlantic Ocean, by Frederic H. Martini, Michael P. Lesser, and John Heiser 525 Low levels of genetic diversity in highly exploited popula- tions of Alaskan Tanner crabs, Chionoecetes bairdi, and Alaskan and Atlantic snow crabs, C. opilio , by Susan E. Merkouris, Lisa W. Seeb, and Margaret C. Murphy 538 A decision rule based on the mean square error for correct- ing relative fishing power diffferences in trawl survey data, by Peter T. Munro 547 Annual and between-sex variability of yellowfin sole, Pleuronectes asper, spring-summer distributions in the east- ern Bering Sea, by Daniel G. Nichol 562 Age, growth, mortality, and population characteristics of the Pacific red snapper, Lutjanus peru , off the southeast coast of Baja California, Mexico, by Axayacatl Rocha-Olivares 575 Enhancing diet analyses of piscivorous fishes in the North- west Atlantic through identification and reconstruction of original prey sizes from ingested remains, by Frederick S. Scharf, Richard M. Yetter, Adam P. Summers, and Francis Juanes 589 Marine turtle populations on the west-central coast of Florida: results of tagging studies at the Cedar Keys, Florida, 1986-1995, by Jeffrey R. Schmid 603 Effects of hypoxia and temperature on survival, growth, and respiration of juvenile Atlantic sturgeon, Acipenser oxyrinch us, by David H. Secor and Troy E. Gunderson 614 Multiple population bottlenecks and DNA diversity in popu- lations of wild striped bass, Morone saxatilis, by John R. Waldman, Reese E. Bender, and Isaac I. Wirgin 621 Can scales be used to sex winter flounder, Pleuronectes americanus ? by Allen J. Bejda and Beth A. Phelan 624 Evaluation of toxicity of oxytetracycline on growth of cap- tive nurse sharks, Ginglymostoma cirraturn, by James Gelsleichter, Enric Cortes, Charles A. Manire, Robert E. Hueter, and John A. Musick 628 Electrotaxis in American lobsters, Homarus americanus, and its potential use in sampling early benthic-phase ani- mals, by Peter Koeller and Gregory Crowell 633 Changes in the probability density function of larval fish body length following preservation, by Pierre Pepin, John F. Dower, and William C. Leggett 641 Sampling juvenile skipjack tuna, katsuwonus pelamis, and other tunas, Thunnus spp., using midwater trawls in the tropical western Pacific, by Toshiyuki Tanabe and Kodo Niu 647 Entanglement and mortality of bottlenose dolphins, Tursiops truncatus, in recreational fishing gear in Florida, by Randall S. Wells, Suzanne Hofmann, and Tristen L. Moors 96(4) 653 Stock discrimination of school mackerel, Scomberomorus queenslandicus, and spotted mackerel, Scomberomorus munroi, in coastal waters of eastern Australia by analysis of minor and trace elements in whole otoliths, by Gavin. A. Begg, Mike Cappo, Darren S. Cameron, Steve Boyle, and Michelle J. Sellin 667 Pelagic sharks associated with the swordfish, Xiphias gladius, fishery in the eastern North Atlantic Ocean and the Strait of Gibraltar, by Valentin Buencuerpo, Santiago Rios, and Julio Moron 686 Monophyly and intrarelationships of the family Pleuro- nectidae (Pleuronectiformes), with a revised classification, by J. Andrew Cooper and Fran?ois Chapleau 727 Documenting the bycatch of harbor porpoises, Phocoena phocoena, in coastal gillnet fisheries from stranded car- casses, by Tara M. Cox, Andrew J. Read, Susan Barco. Joyce Evans, Damon P. Gannon, Heather N. Koopman, William A. McLellan, Kimberly Murray, John Nicolas, D. Ann Pabst, Charles W. Potter, W. Mark Swingle, Victoria G. Thayer, Kathleen M. Touhey, and Andrew J. Westgate 735 Age, growth, and reproduction of black grouper, Mycter- operca bonaci, in Florida waters, by Roy E. Crabtree and Lewis H. Bullock 754 Feeding habits of bonefish, Albula vulpes, from the waters of the Florida Keys, by Roy E. Crabtree, Connie Stevens, Derke Snodgrass, and Fredrik J. Stengard 767 Population structure in greater amberjack, Seriola dumerili, from the Gulf of Mexico and the western Atlantic Ocean, by John R. Gold and Linda R. Richardson 779 Age, growth, and reproduction of the tropical squid Nototodarus hawaiiensis (Cephalopoda: Ommastrephidae) off the North West Slope of Australia, by George D. Jack- son and Vicki A. Wadley 788 Description of pelagic larval and juvenile grass rockfish, Sebastes rastrelliger (family Scorpaenidae), with an exami- nation of age and growth, by Thomas E. Laidig and Keith M. Sakuma 797 Changes in the sex ratio and size at maturity of gag, Mycteroperca microlepis, from the Atlantic coast of the southeastern United States during 1976-1995, by John C. McGovern, David M. Wyanski, Oleg Pashuk, Charles S. Manooch II, and George R. Sedberry 931 808 Distribution and abundance of and habitat use by harbor porpoise, Phocoena phocoena, off the northern San Juan Islands, Washington, by Kimberly L. Raum-Suryan and James T. Harvey 823 Global phylogeography of mackerels of the genus Scomber, by Daniel R. Scoles, Bruce B. Collette, and John E. Graves 843 Population structure and stock identification of British Columbia coho salmon, Oncorhynchus kisutch, based on microsatellite DNA variation, by Maureen P. Small, Ruth E. Withler, and Terry D. Beacham 859 Gonadal development and associated changes in plasma reproductive steroids in English sole, Pleuronectes vetulus, from Puget Sound, Washington, by Sean Y. Sol, O. Paul Olson, Daniel P. Lomax, and Lyndal L. Johnson 871 Parasitism of the golden king crab, Lithodes aequispinus, by two species of snailfish, genus Careproctus, by David A. Somerton and William Donaldson 885 Settlement and recruitment of queen conch, Strombus gi- gas, in seagrass meadows: associations with habitat and micropredators, by Allan W. Stoner, Melody Ray-Culp, and Sheila M. O’Connell 900 Growth trajectory of the larval Japanese sardine, Sardinops melanostictus, transported into the Pacific coastal waters off central Japan, by Yoshiro Watanabe and Motohiko Nakamura 908 Transition from pelagic to benthic prey for age group 0-1 Atlantic cod, Gadus morhua , by Tammy M. Lomond, David C. Schneider, and David A. Methven 912 Metazoan parasites as potential markers for selected Gulf of Alaska rockfishes, by Adam Moles, Jonathan Heifetz, and David C. Love 917 Catchability and retention of larval European anchovy, Engraulis encrasicolus, with bongo nets, by Stylianos Somarakis, Barbara Catalano, and Nikolaos Tsimenides 932 Fishery Bulletin 96(4), 1 998 Fishery Bulletin Volume 96 (1-4), List of authors Index 1998 Moles, Adam 912 Moors, Tristen L. 647 Moron, Julio 667 Munro, Peter T. 538 Munroe, Thomas A. 1 Murakami, Osamu 344 Murphy, Margaret C. 525 Able, Kenneth W, 303 Gerrodette, Tim 237 Murray, Kimberly 727 Musick, John A. 624 Adams, Douglas H. 382 Goetz, Betty 352 Nakamura, Motohiko 900 Ault, Jerald S. 395 Gold, John R. 767 Gorfine, Harry K. 438 Nakao, Shigeru 344 Nelson, Gary A. 315 Baker, M. Scott, Jr. 415 Graves, John E. 823 Nero, Redwood W. 329 Barco, Susan 727 Grover, Jill J. 248 Nichol, Daniel G. 547 Barlow, Jay 237 Gunderson, Donald R. 352 Nicolas, John 727 Beacham, Terry D. 843 Beckmann, Andreas T. 352 Gunderson, Troy E. 603 Niu, Kodo 641 Beeson, Marilyn 271 Harvey, James T. 808 O’Connell, Sheila M. 885 Begg, Gavin A. 653 Heifetz, Jonathan 912 O’Gorman, Joseph B. 303 Bejda, Allen J. 621 Bender, Reese E. 614 Heiser, John B, 516 Herbinson, Kevin 492 Olson, O. Paul 859 Bohnsack, James A. 395 Heuter, Robert E. 624 Pabst. D. Ann 727 Boyle, Steve 653 Hofmann, Suzanne 647 Pashuk, Oleg 797 Brodeur, Richard D. 285 Hohn.AletaA. 357 Pedraza, Susana Noeim 366 Buckley, Raymond M. 352 Horii, Takashi 344 Pepin, Pierre 633 Buencuerpo, Valentin 667 Bullock, Lewis H. 735 Hyndes, Glenn A. 258 Jackson, George D. 779 Perkins, Peter C. 210 Phelan, BethA. 621 Platell, Margaret E. 258 Cameron, Darren S. 653 Johnson, Allyn G. 415 Potter, Charles W. 727 Campana, Steven E. 185 Cappo, Mike 653 Johnson, Lyndal L. 859 Jones, Cynthia M. 185, 451 Potter, Ian C. 258 Caselle, Jennifer E. 492 Juanes, Francis 575 Raum-Suryan, Kimberly L. 808 Catalano, Barbara 917 Chapleau, Francois 686 Julian, Fred 271 Ray-Culp, Melody 885 Read, Andrew J. 428, 727 Chen, Che-Tsung 482 Kimura, Daniel K. 462 Richardson, Linda R. 767 Chiang, Po-Jen 482 Koeller, Peter 628 Rios, Santiago 667 Collette, Bruce B. 823 Collins, L. Alan 415 Koen Alonso, Mariano 366 Koenig, Christopher C. 415 Rocha-Olivares, Axayacatl 562 Condrey, Richard E. 193 Cooper, J. Andrew 686 Koopman, Heather N. 727 Sakuma, Keith M. 788 Sakurai, Izumi 344 Cortes, Enric 624 Laidig, Thomas E. 788 Scharf, Frederick S. 575 Coscarella, Mariano Alberto 366 Landingham, Joyce H. 285 Schmid, Jeffrey R. 589 Cox, Tara M. 727 Leggett, William C. 633 Schneider, David C. 908 Crabtree, Roy E. 735, 754 Lenanton, Rodney C.J. 258 Scoles, Daniel R. 823 Craddock, James E. 428 Lesser, Michael P. 516 Secor, David H. 603 Crespo, Enrique Alberto 366 Liu, Kwang-Ming 482 Sedberry, George R. 797 Crowell, Gregory 628 Lomax, Daniel P. 859 Lomond, Tammy M. 908 Seeb, Lisa W. 525 Sellin, Michelle J. 653 Davis, Tim L.O. 223 Love, David C. 912 Shaw, Franklin R. 462 de Silva, Janaka A. 193 Love, Milton S. 492 Shimada, Allen M. 462 Donaldson, William 871 Love, Richard H. 329 Sluka, Robert D. 388 Dower, John F. 633 Lowe, Sandra A. 502 Small, Maureen P. 843 Snodgrass, Derke 754 Edwards, Elizabeth F. 210 Macchi, Gustavo J. 375 Sol, Sean Y. 859 Evans, Joyce 727 Manire, Charles A. 624 Somarakis, Stylianos 917 Farley, Jessica H. 223 Manooch II, Charles S. 797 Somerton, David A. 871 Fernandez, Stephanie 357 Martini, Frederic H. 516 Stengard, Fredrik J. 754 Fiedler, Paul C. 237 McBride, Richard S. 303 Stevens, Connie 754 Forbes, David A. 438 McGovern, John C. 797 McLellan, William A. 727 Stoner, Allan W. 885 Sturdevant, Molly V. 285 Gannon, Damon P. 428, 727 Meester, Geoffrey A. 395 Sullivan, Kathleen M. 388 Garcia, Nestor Anibal 366 Merkouris, Susan E. 525 Summers, Adam P. 575 Gason, Anne S. 438 Gelsleichter, James 624 Methven, David A. 908 Miller, Bruce S. 352 Swingle, W. Mark 727 933 Tanabe, Toshiyuki 641 Thayer, Victoria G.T. 727 Thompson, Charles H. 329 Touhey, Kathleen M. 727 Tremain, Derek M. 382 Tsimenides, Nikolaos 917 Van Doornik, Donald M. 502 Wadley, Vicki 779 Waldman, John R. 614 Watanabe, Yoshiro 900 Wells, Brian 451 Wells, Randall S. 647 Westgate, Andrew J. 727 Winans, Gary A. 502 Winner, Brent L. 382 Wirgin, Isaac I. 614 Withler, Ruth E. 843 Wyanski, David M. 797 Yetter, Richard M. 575 934 Fishery Bulletin 96(4), 1998 Fishery Bulletin Index Volume 96 (1-4), 1998 List of subjects Abalone blacklip 438 Abundance abalone 438 porpoise, harbor 808 searobin 303 Acipenser oxyrhynchus — see Sturgeon, Atlantic Acipenseridae — see Sturgeon Acoustic data 237, 329 Aegean Sea 917 Age and growth drum, black 185, 451 elasmobranch 624 grouper, black 735 rockfish, grass 788 snapper, red 562 squid, tropical 779 thresher, bigeye 482 whiting, King George 258 Age determination dolphin, bottlenose 357 snapper, red 562 Age structure whiting 258 Age validation drum, black 185, 382 shark, nurse 624 Alaska Aleutian islands 502 Bering Sea 547 Southeastern 285 Albula vulpes — see Bonefish Albulidae — see Bonefish Alopias superciliosus — see Thresher, bigeye Amberjack greater 767 Anoplopoma fimbria — see Sablefish Anchovy 917 Annuli 185 Argentina 366, 375 Atlantic Ocean eastern North 667 North 516, northeast 575 northwest 575 outheastern 797 western 303. 767 western North 303 Australia eastern 653 North West Slope 729 southwestern 258 Victoria 438 Bahia de La Paz 562 Bass, striped 614 Batch fecundity weakfish 375 Bering Sea 547 Biomass 547 Bomb radiocarbon 185 Bonefish 754 Bongo net 917 Brevoortia patronus — see Menhaden, gulf Bycatch tuna 210 gillnet 271 menhaden, gulf 193 porpoise, harbor 428, 727 shark 667 California 271 Calving season dolphin, bottlenose 357 Canada British Columbia 285, 843 Carcharhinidae — see Shark Carangidae — see Amberjack, greater Careproctus — see Snailfish Caretta caretta — see Turtle, loggerhead Catchability anchovy 917 Cephalopoda 779 Cetaceans 271, 366 Chelonia mydas — see Turtle, green Chesapeake Bay 451, 603 Chionoecetes bairdi — see Crab, Tanner opilio — see Crab, snow Cladistics flatfishes 686 Clam sunray surf 344 Classification flatfishes 686 mackerel 823 Clupeidae — see Menhaden Cod Atlantic 908 Conch queen 885 Crab golden king 871 snow 525 Tanner 525 Cynoglossidae — see Tonguefish Cynoscion striatus — see Weakfish, striped Decapoda 525 Decision rule 538 Delphinidae — see Dolphin Demography turtle 589 Density estimate 329 Description rockfish, grass 788 Diagnostic bones 575 Diet — see Food habits Diet overlap 285 Distribution dolphin 237 porpoise, harbor 808 searobin 303 sole, yellowfin 547 Diver surveys 438 DNA microsatellite 843 mitochondrial bass, striped 614 amberjack, greater 767 mackerel 823 nuclear bass, striped 614 Dolphin 210, 237 bottlenose 357, 647 calving season 357 dusky 366 mortality, fishery-related 210 prey 237 Drift net 271 Drum, black 185, 382 East-central Florida 382 Eastern North Atlantic 667 Eastern Pacific 210, 237 Ecology tonguefish 1 El Nino-Southern Oscillation (ENSO) 462 Electrotaxis 628 Electrophoresis 502, 525 Engraulidae — see Anchovy Engraulis encrasicolus — see Anchovy Epinephelus — see Grouper Essential habitat sturgeon, Atlantic 603 Evolution mackerel 823 Fallout 185 Fecundity gag 415 rockfish 352 tuna 223 Feeding bonefish 754 flounder, summer 248 Fishery management 258, 395, 843 Fishing power differences 538 Flatfishes 686 — see also Flounder, Sole, and Tonguefish Florida 315, 382, 589, 735, 754 Florida Keys 395, 754 Flounder summer 248 winter 621 935 Food habits dolphin, dusky 366 piscivorous fishes 575 porpoise, harbor 428 turtle 589 Founder population 614 Gadus morhua - see Cod, Atlantic Gag 415, 797 Gear bongo net 917 entanglement in 647 recreational fishing 647 trawl 641 Genetics mackerel, Atka 502 crab, snow 525 crab, Tanner 525 Geographic differentiation crab, snow 525 crab. Tanner 525 Gillnet fisheries 271 Ginglymostoma cirratum — see Shark, nurse Gonads sole, English 859 Grouper 388 black 735 gag 415 spearfishing on 388 Growth dolphin, bottlenose 357 mackerel, Atka 502 rockfish, grass 788 sablefish 452 sardine, larval 900 shark, nurse 624 turtle 589 whiting 258 Gulf of Alaska 912 Gulf of Maine 428, 516 Gulf of Mexico 767 Habitat use porpoise, harbor 808 Hagfish 516 Hake Pacific 329 Haliotidae — see Abalone Haliotis rubra — see Abalone, blacklip Hermaphroditism 415, 614, 735 Hexigrammidae — see Mackerel Homarus americanus — see Lobster, American Hybrids 525 Hypoxia 603 Identification rockfish, grass 788 Inbreeding 614 Incidental capture 271, 727 Incidental mortality 647 Indian Ocean 223 Japan southwest Hokkaido 344 Juvenile lobster, American 628 salmon, Pacific 285 tuna, skipjack 641 Katsuwonus pelamis — see Tuna, skipjack Lagodon rhomboides — see Pinfish Lagenorhynchus obscurus — see Dolphin, dusky Larvae anchovy 917 flounder, summer 248 Larval fish length, effects of preservation on 633 rockfish, grass 788 Length distribution 633 Lepidochelys kempii — see Turtle, Kemp’s ridley sea Life history grouper, black 735 Lithodes aequispinus — see Crab, golden king Lobster, American 628 Logit models 193 Loglinear models 193 Longevity rockfish 352 Lutjanidae — see Snapper Lutjanus peru — see Snapper, red Mackerel Atka 502 school 653 Scomber genus 823 spotted 653 Mactra chinensis — see surfclam, sunray Majidae — see Crab Marine fishery reserve 388 Mark-recapture turtle 589 Maturity gag 797 Menhaden gulf 193 Merluccius productus — see Hake, Pacific Metamorphosis 248 Mexico Baha California 562 Gulf of 415 Mid-Atlantic 727 Mid-Atlantic Bight 303 Midwater trawl net 641 Migration sablefish 452 Morone saxatilis — see Bass, striped Morphology tonguefishes 1 Morphometries hagfish 516 Mortality drum, black 451 marine mammal 271 pinfish 315 seabird 271 snapper, red 562 turtle 271 Mortality estimate snapper, red 562 Movement sablefish 462 searobin 303 whiting 258 MULTIFAN 482 Mycteroperca — see Grouper Mycteroperca bonaci — see Grouper, black microlepis — see Gag Myxine glutinosa — see Hagfish Myxinidae — see Hagfish Nomenclature flatfishes 686 Nototodarus hawaiiensis — see Squid Ommastrephidae — see Squid Oncorhynchus gorbuscha 285 keta 285 kisutch 285, 843 nerka 285 tshawytscha 285 Otolith analysis , for stock discrimination 653 drum, black 185, 451 grouper, black 735 sardine, larval 900 whiting, King George 258 Overfishing 395 Oxytetracycline toxicity of 624 Pacific Ocean eastern 210, 237 North 285 northeast 462 off central Japan 900 western 641 Paralichthys dentatus — see Flounder, summer Parasites rockfish 912 Parasitism 912, 917 Patagonia 366 Patchy distribution menhaden, gulf 193 Phocoena phocoena — see Porpoise, harbor Phylogeny flatfishes 686 Phylogeography mackerel 823 Pinfish 315 Pinniped 271 Piscivore 575 Pleurogrammus monopterygius — see Mackerel, Atka Pleuronectes americanus — see Flounder, winter asper — see Sole, yellowfin revised classification 686 vetulus — see Sole, English Pleuronectidae — see Flounder and Sole 936 Fishery Bulletin 96(4), 1998 Pleuronectiformes — see Flatfish Pogonias cromis — see Drum, black Population dynamics clam 344 Population structure amberjack, greater 767 salmon, coho 843 turtles, marine 589 Porpoise harbor 428, 727, 808 Predation 575, 885 Preservation larval fish 633 Prey assemblages 285 benthic 908 ingested 575 of cod, Atlantic 908 pelagic 908 Pnonotus — see Searobin Probability density function 633 Protogyny 415 Purse seiners eastern tropical Pacific 210 Recreational fisheries 647 Recruitment conch, queen 885 whiting, King George 258 Reef fisheries 395 Regression 575 Reproduction crab, golden king 871 gag 415 grouper, black 735 rockfish, Puget Sound 352 sole, English 859 squid, tropical 779 tuna, southern bluefin 223 whiting. King George 258 Respiration sturgeon, Atlantic 603 Restoration 603 Rentention, larval anchovy 917 Rockfish grass 788 Gulf of Alaska 912 Puget Sound 352 rougheye 912 shortraker 912 Sablefish 462 Salmon coho 843 Pacific 285 Salmonidae — see Salmon San Juan Islands 808 Sardine larval Japanese 900 Sardinops melanostictus - see Sardine Scales flounder, winter 621 Sciaenidae — see Drum and Weakfish Scomberomorus queenslandicus - see Mackerel, school Scomberomorus munroi - see Mackerel, spotted Scombridae — see Mackerel and Tuna Scorpaenidae — see Rockfish Seabird 271 Searobin 303 Seasonality 315 Seasonal studies 303 Sebastes aleutianus — see Rockfish, rougheye borealis — see Rockfish, shortraker emphaeus — see Rockfish, Puget Sound rastrelliger — see Rockfish, grass Seriola dumerili — see Amberjack, greater Serranidae — see Bass and Grouper Set net 271 Settlement conch, queen 885 Sexing flounder, winter 621 Sex ratio gag 415, 797 sole, yellowfin 547 Sex variability 547 Sexual maturity whiting 258 Shark nurse 624 pelagic 667 Shirasu fishery 900 Shrinkage 633 Sillaginodes punctata — see Whiting, King George Size at maturity gag 797 Size structure searobin 303 Snailfish 871 Snapper red 562 Sole English 859 yellowfin 547 Southeastern Atlantic 797 Southwestern Atlantic 366, 375 Spatial distribution pinfish 315 Spawning dynamics 223 gag 415, 797 rockfish 352 tuna 223 weakfish, striped 375 whiting 258 Spawning frequency weakfish 375 Spermiogenesis 859 Squid 779 Statoliths Squid, tropical 779 Steroids, plasma sole, English 859 Stock assessment coral reef fishes 395 Stock discrimination mackerel 653 Stock management surfclam 344 Stock size surfclam 344 Stock structure amberjack, greater 797 sablefish 462 Stomach contents piscivorous fishes 575 porpoise, harbor 428 Stock size prediction surfclam 344 Strait of Gibraltar 667 Stranding harbor porpoise 727 Strombas gigas — see Conch, queen Sturgeon Atlantic 603 Sunda Islands 223 Surfclam sunray 344 Surveys 628 Swimbladder resonance 329 Swordfish 667 Symphurus — see Tonguefishes Systematics hagfish 516 tonguefishes 1 Tag analysis 462 Taiwan waters 482 Taxonomy tonguefish 1 Temperature sturgeon, Atlantic 603 Thresher, bigeye 482 Thunnus maccoyii — see Tuna, southern bluefin Tonguefishes 1 Transplantation 614 Trawl survey data 538 Triglidae — see Searobin Trophic model 908 Tropical western Pacific 641 Tuna 210 skipjack 641 southern bluefin 223 Tursiops truncatus — see Dolphin, bottlenose Turtle discard 210 green 589 Kemp’s ridley 589 loggerhead 589 United States Alaska 525 Aleutian Islands 502 Bering Sea 547 California 271, 492 east coast 727 Florida 315, 388, 395, 589, 647, 735, 754 Gulf of Alaska 912 Gulf of Maine 428, 516 Gulf of Mexico 415, 767 937 Puget Sound 352, 859 San Juan Islands 808 Texas 357 west coast 329 Validation, age drum, black 185 Vertebrae shark, nurse 624 Vitellogenesis 859 Volume scattering 329 Washington Sound 808 Weakfish striped 375 Western Atlantic 1, 767 Whiting King George 258 Xiphias glaclius — see Swordfish Xiphiidae — see Swordfish Young of the year pinfish 315 938 Fishery Bulletin 96(4), 1998 Superintendent of Documents Publications Order Form *5178 □yes, please send me the following publications: Subcriptions to Fishery Bulletin for $35.00 per year ($43.75 foreign) The total cost of my order is $ . Prices include regular domestic postage and handling and are subject to change. (Company or Personal Name) (Please type or print) (Additional address/attention line) (Street address) (City, State, ZIP Code) (Daytime phone including area code) (Purchase Order No.) Charge your order. It’s Easy! P3S Please Choose Method of Payment: I I Check Payable to the Superintendent of Documents I | GPO Deposit Account | | | | | | | ~| — Q | | VISA or MasterCard Account To fax your orders (202) 512-2250 (Credit card expiration date) (Authorizing Signature) Mail To: Superintendent of Documents P.O. Box 371954, Pittsburgh, PA 15250-7954 Thank you for your order! This statement is required by the Act of Au- gust 12, 1970, Section 3685, Title 39, U.S. Code, showing ownership, management, and circu- lation of the Fishery Bulletin, publication num- ber 366-370, and was filed on 10 September 1998. The Bulletin is published quarterly (four issues annually) with an annual subscription price of $35.00 (sold by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402). The complete mailing address of the office of publication is: NMFS Scientific Publications Office, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115. The complete mailing address of the headquarters of the publishing agency is: Na- tional Marine Fisheries Service, NOAA, De- partment of Commerce, 1335 East-West High- way, Silver Spring, MD 20910. The name of the publisher is Willis Hobart and the managing editor is Sharyn Matriotti; their mailing ad- dress is: NMFS Scientific Publications Office, 7600 Sand Point Way NE, BIN 15700, Seattle, WA 981 15. The owner is the U.S. Department of Commerce, 14th St. N.W., Washington, DC 20230; there are no bondholders, mortgages, or other security holders. The purpose, func- tion, and nonprofit status of the organization (agency) and the exempt status for Federal in- come tax purposes has not changed during the preceding 12 months. The extent and nature of circulation is as follows: Total number of cop- ies (A) (average number of copies of each issue during the preceding 12 months) was 2,038 and the actual number of copies of the single issue published nearest to the filing dates was 2,038. Paid circulation ( B ) is handled by the U.S. Gov- ernment Printing Office, Washington, DC 20402, and the total number printed for their sales (mail subscriptions and individual sales) was 600 for both the average number of copies each issue during the preceding 12 months and the actual number of copies of the single issue published nearest to the filing date. Free dis- tribution (D) by mail; samples, complimentary and other free copies (average number of cop- ies each issue during the preceding 12 months) was 0 and the actual number of copies of the single issue published nearest to the filing date was 0. Free distribution outside the mail (E) by carriers or other means was 0 for both aver- age number of copies and actual number of cop- ies. Total free distribution (F) was 0 for both average number of copies and actual number of copies of the single issue published nearest the filing date. The total distribution (G: sum of C and G) (average number of copies each is- sue during the preceding 12 months) was 2,038 and the actual number of copies of the single issue published nearest to the filing date was 2,038. There were no copies not distributed or returned from news agents (H). The total (I: sum of G and H) is equal to the net press run figures shown in Item A: 2,038 and 2,038 cop- ies, respectively. I certify that the statements made by me above are correct and complete: (Signed) Willis Hobart, Publisher. Fishery Bulletin Guide for Contributors Content Articles published in Fishery Bulletin de- scribe original research in fishery marine science, engineering and economics, and the environmental and ecological sciences, in- cluding modeling. Articles may range from relatively short to extensive; notes are re- ports of 5 to 10 pages without an abstract and describing methods or results not sup- ported by a large body of data. Although all contributions are subject to peer review, responsibility for the contents of papers rests upon the authors and not upon the editor or the publisher. It is therefore im- portant that the contents of the manuscript are carefully considered by the authors. Submission of an article is understood to imply that the article is original and is not being considered for publication elsewhere. Manuscripts should be written in English. Authors whose native language is not En- glish are strongly advised to have their manuscripts checked by English-speaking colleagues prior to submission. Preparation Title page should include authors’ full names and mailing addresses, the corre- sponding author’s telephone, FAX number, and E-mail address, and a list of key words to describe the contents of the manuscript. Abstract should not exceed one double- spaced typed page. It should state the main scope of the research but emphasize its con- clusions and relevant findings. Because ab- stracts are circulated by abstracting agencies, it is important that they represent the research clearly and concisely. Text must be typed double-spaced through- out. A brief introduction should portray the broad significance of the paper; the remain- der of the paper should be divided into the following sections: Materials and meth- ods, Results, Discussion (or Conclusions), and Acknowledgments. Headings within each section must be short, reflect a logical sequence, and follow the rules of multiple subdivision (i.e. there can be no subdivision without at least two items). The entire text should be intelligible to interdisciplinary readers; therefore, all acronyms, abbrevia- tions, and technical terms should be spelled ☆ U.S. Gov. Printing Office 1998 689015/80004 out the first time they are mentioned. The scientific names of species must be written out the first time they are mentioned; sub- sequent mention of scientific names may be abbreviated. Follow the U.S. Government Printing Office Style Manual (1984 ed.) and the CBE Scientific Style and Format (6th ed.) for editorial style, and the most current issue of the American Fisheries Society’s Common and Scientific Names of Fishes from the United States and Canada for fish nomenclature. Dates should be written as follows: 11 November 1991. Measurements should be expressed in metric units, e.g., metric tons as (t); if other units of measure- ment are used, please make this fact explicit to the reader. The numeral one (1) should be typed as a one, not as a lower-case el (1). Use of appendices is discouraged. Text footnotes should be numbered with Arabic numerals. Footnote all personal communications, unpublished data, and un- published manuscripts with full address of the communicator or author, or, as in the case of unpublished data, where the data are on file. Authors are advised to avoid ref- erences to nonstandard (gray) literature, such as internal, project, processed, or ad- ministrative reports. Where these refer- ences are used, please include whether they are available from NTIS (National Techni- cal Information Service) or from some other public depository. Literature cited comprises published works and those accepted for publication in peer- reviewed literature (in press). Follow the name and year system for citation format. In the text, cite as follows: Smith and Jones (1977) or (Smith and Jones, 1977). If there is a sequence of citations, list by year: Smith, 1932; Smith and Jones, 1985; Smith and Allen, 1986. Abbreviations of serials should conform to abbreviations given in Serial Sources for the BIOSIS Previews Database. Authors are responsible for the accuracy and completeness of all citations. Tables should not be excessive in size and must be cited in numerical order in the text. Headings should be short but ample enough to allow the table to be intelligible on its own. All unusual symbols must be ex- plained in the table legend. Other inciden- tal comments may be footnoted with italic numerals. Use the asterisk only to indicate probability in statistical data. Because ta- bles are typeset, they need only be submit- ted typed and formatted, with double-spaced legends. Zeros should precede all decimal points for values less than one. Figures must be cited in numerical order in the text. The senior author’s last name and the figure number should be written on the back of each one. Hand-drawn illustrations should be submitted as originals and not as photocopies. Submit photographs as glossy prints or slides that show good contrast, otherwise we cannot guarantee a good final printed copy. Graphic illustrations should be submitted as laser-printed copies, not as photocopies. Label all figures with Helve- tica typeface and capitalize the first letter of the first word in axis labels. Do not use boldface. Italicize species name and vari- ables in equations. Use zeros before all deci- mal points. Use uppercase Times Roman bold typeface to label the parts of a figure, e.g. A, B, C, etc. Send original figures to the Scientific Editor when the manuscript has been accepted for publication. Each figure legend should explain all symbols and ab- breviations in the figure and should be dou- ble-spaced and placed at the end of the manuscript. Copyright law does not cover government publications; they fall within the public do- main. If an author reproduces any part of a government publication in his work, refer- ence to source is appreciated. Submission Send printed copies (original and three cop- ies without staples) to the Scientific Editor: Dr. John B. Pearce, Scientific Editor Northeast Fisheries Science Center National Marine Fisheries Service 166 Water Street Woods Hole, MA 02543-1097 Once the manuscript has been accepted for publication, you will be asked to submit a software copy of your manuscript to the Man- aging Editor. The software copy should be submitted in WordPerfect or Microsoft Word text format and placed on a 3.5-inch disk that is double-sided, double or high density, and that is compatible with either DOS or Apple Macintosh systems. A copy of page proofs will be sent to the au- thor for final approval prior to publication. Copies of published articles and notes are available free of charge to the senior author (50 copies) and to his or her laboratory (50 copies). Additional copies may be purchased in lots of 100 when the author receives page proofs. Heckman BINDERY, INC. Bound-To-Please® FEB 00 N. MANCHESTER, INDIANA 46962